Circulating Ghrelin in Patients Undergoing Elective Cholecystectomy

2 downloads 0 Views 1MB Size Report
Beggs AH, Koenig M, Boyce FM, Kunkel LM. Detection of 98% of DMD/ ..... Watkins S, Madison L, Galliano M, Minchiotti L, Putnam FW. Analbumin- emia: three ...
Technical Briefs Collection and Storage of Human Blood Cells for mRNA Expression Profiling: A 15-Month Stability Study, Jean-Brice Marteau, Steve Mohr, Miche`le Pfister, and Sophie Visvikis-Siest* (Institut National de la Sante´ et de la Recherche Me´dicale, U525 Equipe 4, Faculte´ de Pharmacie, Nancy, France; * address correspondence to this author at: INSERM U525 Equipe 4, 30 rue Lionnois, 54000 Nancy, France; fax 33-3-83321322, e-mail [email protected]) Gene expression profiling is increasingly important in human health research and applications (1 ). Unfortunately, the human tissue samples required for this process, particularly those from healthy individuals, are not safely and easily available. Therefore, the use of peripheral blood mononuclear cells (PBMCs) as surrogate material for high-throughput analysis of gene expression is currently being explored. These cells are involved in a large variety of immune-related diseases, including infection and cancer. Moreover, in recent studies, characteristic sets of transcriptional changes in PBMCs were associated with physiologic or pathologic states (2– 4 ). Thus, a PBMC transcriptome may be used as an individual’s health sensor, a concept referred to as the sentinel principle (5 ). In large multicenter studies, the reliable and reproducible detection of transcript concentrations from PBMCs requires standardization of blood sampling and an efficient method of conservation. Indeed, many preanalytical factors during collection, processing, and storage of blood specimens may affect RNA and its subsequent use as a biomarker (6 ). Although numerous technical and clinical aspects of blood sampling have been addressed (7–9 ), comprehensive data on the long-term storage and stability of RNA from PBMCs are needed. Here we report experiments performed over 15 months to test the preanalytical conditions involved in blood collection procedures and PBMC storage. Blood samples were collected from healthy donors into EDTA and sodium heparin tubes. RNA extractions were performed on isolated PBMCs stored at ⫺80 °C up to 15 months in 4 different lysis buffers. Using spectrometry and real-time PCR, we compared the concentration, purity, integrity, and stability of the total RNA. We propose a quality-assured and controlled protocol of PBMC banking for further mRNA expression analysis. After informed consent was obtained from 12 unrelated adult volunteers, whole blood (10 mL) was collected by standardized venipuncture in EDTA (EDTA-blood; n ⫽ 12) and heparin (heparin-blood; n ⫽ 12) tubes (VacutainerTM; Becton Dickinson) and processed for PBMC preparation. Briefly, blood samples were homogenized with 10 mL of Hanks Balanced Salt Solution, Modified (SigmaAldrich). PBMCs were prepared by Ficoll density-gradient centrifugation (Ficoll-PaqueTM PLUS; Amersham) at 300g for 30 min at 20 °C. The ring of high-density PBMCs was isolated and washed twice in 50 mL of Hanks buffer. After cell survival was determined with the trypan-blue exclusion test, the PBMC concentration was normalized to 106 cells/mL in Hanks buffer. PBMC populations were evaluated by microscopic observation after May–Grun-

1250

Clinical Chemistry 51, No. 7, 2005

wald–Giemsa staining. After centrifugation for 6 min at 700g (4 °C), 200 ␮L of cell lysis buffer was immediately added to the PBMC pellet. Four different commercially available buffers were tested (n ⫽ 24 for each buffer, including 12 EDTA-blood and 12 heparin-blood samples): RNA InstaPure [Eurogentec (E)], lysis/binding buffer from the MagNA Pure LC RNA Isolation Kit I [Roche Diagnostics (R)], buffer RTL from the QIAamp® RNA Blood Mini Kit [Qiagen (Q)], and RNA lysis buffer from the StrataPrep® Total RNA Miniprep Kit [Stratagene (S)]. The PBMC lysates were used to extract total RNA immediately (t0) or after storage at ⫺80 °C for 3, 5, 10, and 15 months (t3, t5, t10, and t15, respectively). For total RNA extraction, we used an automated process with the MagNA Pure LC instrument (Roche Diagnostics) and the MagNa Pure LC RNA Isolation Kit I according to the manufacturer’s instructions, including DNase treatment. Purified RNA was eluted at 70 °C with 100 ␮L of low-salt buffer. Total RNA quality and quantity were assessed in 2 ways. In the first method, we estimated the RNA concentration by ultraviolet absorbance at 260 nm (1 absorbance unit at 260 nm ⫽ 40 ng/␮L RNA) and the RNA purity by measuring the ratio of absorbance at 260 nm and 280 nm (1.8 ⬍ A260/A280 ⬍ 2.1 for pure RNA). Total RNA was run on 1% agarose gels to check size and integrity. In the second method, RNA from each sample was assessed for integrity by a 2-step reverse transcription-PCR assay for glyceraldehyde-3-phosphate dehydrogenase (GAPDH; NM_002046). cDNA synthesis was performed at 37 °C for 60 min with 80 ng of extracted RNA, 0.25 ␮g of oligo(dT)15 primer, and 200 U of Moloney murine leukemia virus (M-MLV) reverse transcriptase (Promega). After each reverse transcription, cDNA was amplified by a standard real-time protocol in the LightCycler instrument (Roche Diagnostics). The 20-␮L PCR reaction for the GAPDH gene included 1⫻ LightCycler FastStart DNA Master SYBR Green I (Roche Diagnostics), 4 mM MgCl2, 0.2 ␮M each of 5⬘-ACGGATTTGGTCGTATTGGG-3⬘ (forward primer) and 5⬘-CGCTCCTGGAAGATGGTGAT-3⬘ (reverse primer; Sigma-Aldrich), and 2 ␮L of 1:100-diluted template cDNA. Cycling settings were an initial denaturation step at 95 °C for 8 min and 45 cycles of 95 °C for 6 s, 60 °C for 5 s, and 72 °C for 15 s, followed by melting curve analysis to evaluate the specificity of PCR products. To ensure amplification of the appropriately sized fragment, PCR products were analyzed by electrophoresis in 3% agarose gels. PCR controls included a separate reaction in which either cDNA template or the reverse transcriptase was eliminated. None of these controls showed detectable amplification product (data not shown). GAPDH mRNA detection was estimated by use of the crossing point (CP), defined by the LightCycler software (Ver. 3.5) as the cycle at which each reaction reaches the logarithmic portion of the PCR curve. PCR determinations in each sample were performed in triplicate. To improve our results, we analyzed the ADM (NM_001124), and IL-8 (NM_ 000584) genes, which are known to quickly change expression in response to technical influences such as temperature. The PCR conditions for these reactions are shown in Table 1

Clinical Chemistry 51, No. 7, 2005

of the Data Supplement that accompanies the online version of this Technical Brief at http://www.clinchem. org/content/vol51/issue7/. Statistical analysis was performed with StatView software (Ver. 5; SAS Institute Inc.). Differences in mean RNA concentration or CP value according to either anticoagulant (EDTA, heparin) or lysis buffer (E, R, Q, and S) were estimated by paired Student t-test. Variation of RNA stability according to different storage times was calculated by the Friedman test. Values are reported as the mean (SE). A P value ⱕ0.05 was considered statistically significant. As determined by trypan-blue counting, PBMC viability was 97.5 (1.3)%. According to results obtained with May–Grunwald–Giemsa staining, PBMC populations usually contained ⬎97% mononuclear cells. To compare the purity and yield of total RNA obtained from the whole-blood collection under the conditions being investigated, we performed spectrometric analysis on the RNA samples extracted at t0. Regardless of the cell lysis buffer used, we observed no difference in the A260/A280 ratio between anticoagulants [1.95 (0.12) and 1.94 (0.11) for EDTA and heparin, respectively; P ⫽ 0.24], suggesting that RNA purity was equivalent for all samples. In contrast, the yield of total RNA at t0 was higher in EDTA-blood than in heparin-blood: 23.9 (15.2) vs 19.1 (10.5) ng/␮L (P ⬍0.0001; Fig. 1A). The CP value of the GAPDH gene determined by real-time PCR indicated

1251

that RNA integrity was better in EDTA-blood than in heparin-blood (Fig. 1A in the online Data Supplement). The measured mean CP was 38.0 (2.0) cycles in EDTAblood and 40.4 (2.3) cycles in heparin-blood (P ⬍0.0001; n ⫽ 24). Melting curve analysis confirmed the presence of a single GAPDH-specific amplicon for all samples (data not shown). These results indicate that the use of EDTA as anticoagulant in blood collection ensures the best total RNA yield and quality. This conclusion was further supported by other findings involving blood-based RNA assays (10, 11 ). To further improve the yield of total RNA, we tested 4 different cell lysis buffers combined with automated MagNA Pure LC extraction. Typical results obtained for 12 donors at t0 are shown in Fig. 1B and in Fig. 1B of the online Data Supplement. RNA concentrations were significantly higher with use of the E buffer than with the other buffers (P ⬍0.01; Fig. 1B). The RNA yield obtained in buffers Q and S did not differ significantly from the yield obtained in buffer R. Neither EDTA nor heparin affected this result (Fig. 1B). In addition, RNA integrity was best in EDTA-blood with buffer E (Fig. 1B of the online Data Supplement). Results were similar for the ADM and IL-8 genes (panels A and B, respectively, of Fig. 2 of the online Data Supplement). We therefore propose that buffer E rather than buffer R be used to constitute PBMC banks for the purpose of RNA expression analysis.

Fig. 1. Impact of preanalytical factors on PBMC RNA concentrations. Impact of anticoagulants (EDTA and heparin; n ⫽ 48 for both anticoagulants; A), cell lysis buffers (Eurogentec, Roche, Qiagen, and Stratagene; n ⫽ 12 for all groups; B), and storage time (3, 5, 10, and 15 months; n ⫽ 8 for all time points) in Eurogentec buffer (EDTA-blood; C) on RNA concentration. The RNA concentrations in all panels are in ng/␮L. The results are shown as box-plots with medians (lines inside boxes), 25th and 75th percentiles (limits of boxes), and the 10th and 90th percentiles (whiskers). The F in panel A indicate outliers. NS, not significant.

1252

Technical Briefs

Finally, to determine whether buffer E would preserve the RNA of PBMCs over time, we analyzed total RNA obtained from 8 individuals at t3, t5, t10, and t15 of PBMC storage time with reference to their t0 values. For all samples examined, RNA purity, determined by the A260/ A280 ratio, was 1.8 –2.1, indicating that highly pure RNA could be obtained with storage in buffer E. In addition, for all 3 genes, we observed no significant variations in either the amount (Fig. 1C; Friedman test, P ⫽ 0.0812) or the integrity (Fig. 1C of the online Data Supplement; Friedman test, P ⫽ 0.3682) of RNA at t3, t5, t10, and t15 compared with t0. Thus, buffer E provides advantageous conditions for PBMC storage at ⫺80 °C. In addition, it is noteworthy that buffer E is also suitable for manual extraction (see the manufacturer’s instructions), which does not necessarily require specific and costly automation. The kinetics of gene expression and RNA stability in blood samples processed ex vivo (preanalytical variables) are currently under intense investigation (6, 12, 13 ). To guarantee optimal conditions, samples reach our laboratory for processing after a transportation time of ⬍2 h at 4 °C. Indeed, gene expression was shown to change 2 h after blood collection (13 ), probably because of significant stress responses (12 ) or possible contact activation of cells in the collection tubes (9 ). These findings underscore the extreme sensitivity of blood cells to ex vivo handling. One study has suggested that EGTA is preferable to EDTA as an anticoagulant to prevent the irreversible loss of antigen-specific lymphoproliferative responses in instances in which unclotted blood may be stored for long periods (14 ). Methods such as the PAXgene tube system or addition of acidic phenol and guanidine isothiocyanate have been proposed to stabilize blood cell gene expression after prolonged incubation of blood ex vivo, but new methods and protocols are needed (8, 12 ). Our studies demonstrate that PBMC RNA is stable over time (up to 15 months) when blood samples are collected into EDTA tubes and when PBMCs are stored at ⫺80 °C in buffer E. We provide here a simple assay for rapid, efficient, and standardized banking of PBMCs, a key step in large gene expression studies. To our knowledge, this is the first comprehensive report on RNA stability over 15 months. Moreover, preliminary observations suggest that our protocol provides successful results in gene expression assays such as quantitative reverse transcription-PCR or microarrays (data not shown). Finally, our recommendations may be helpful for clinicians and researchers involved in experimentation with PBMC transcriptomes in molecular medicine and epidemiologic surveillance, an emerging field of noninvasive health applications.

This work was supported by an INSERM Grant CRB “Collection de Ressources Biologiques” 4CH07H. We thank G. Siest, the staff of the Centre of Preventive Medicine of Vandoeuvre-le`s-Nancy, France, and the participants who made this study possible.

References 1. Mohr S, Leikauf GD, Keith G, Rihn BH. Microarrays as cancer keys: an array of possibilities [Review]. J Clin Oncol 2002;20:3165–75. 2. Chon H, Gaillard CA, van der Meijden BB, Dijstelbloem HM, Kraaijenhagen RJ, van Leenen D, et al. Broadly altered gene expression in blood leukocytes in essential hypertension is absent during treatment. Hypertension 2004; 43:947–51. 3. Lampe JW, Stepaniants SB, Mao M, Radich JP, Dai H, Linsley PS, et al. Signatures of environmental exposures using peripheral leukocyte gene expression: tobacco smoke. Cancer Epidemiol Biomarkers Prev 2004;13: 445–53. 4. Whitney AR, Diehn M, Popper SJ, Alizadeh AA, Boldrick JC, Relman DA, et al. Individuality and variation in gene expression patterns in human blood. Proc Natl Acad Sci U S A 2003;100:1896 –901. 5. Liew CC. Expressed genome signatures of cardiovascular disease [Abstract]. Clin Chem Lab Med 2004;42:A25. 6. Debey S, Schoenbeck U, Hellmich M, Gathof BS, Pillai R, Zander T, et al. Comparison of different isolation techniques prior gene expression profiling of blood derived cells: impact on physiological responses, on overall expression and the role of different cell types. Pharmacogenomics J 2004;4:193–207. 7. Holland NT, Smith MT, Eskenazi B, Bastaki M. Biological sample collection and processing for molecular epidemiological studies [Review]. Mutat Res 2003;543:217–34. 8. Pahl A, Brune K. Stabilization of gene expression profiles in blood after phlebotomy [Technical Brief]. Clin Chem 2002;48:2251–3. 9. Rainen L, Oelmueller U, Jurgensen S, Wyrich R, Ballas C, Schram J, et al. Stabilization of mRNA expression in whole blood samples. Clin Chem 2002;48:1883–90. 10. Dickover RE, Herman SA, Saddiq K, Wafer D, Dillon M, Bryson YJ. Optimization of specimen-handling procedures for accurate quantitation of levels of human immunodeficiency virus RNA in plasma by reverse transcriptase PCR. J Clin Microbiol 1998;36:1070 –3. 11. Ginocchio CC, Wang XP, Kaplan MH, Mulligan G, Witt D, Romano JW, et al. Effects of specimen collection, processing, and storage conditions on stability of human immunodeficiency virus type 1 RNA levels in plasma. J Clin Microbiol 1997;35:2886 –93. 12. Baechler EC, Batliwalla FM, Karypis G, Gaffney PM, Moser K, Ortmann WA, et al. Expression levels for many genes in human peripheral blood cells are highly sensitive to ex vivo incubation. Genes Immun 2004;5:347–53. 13. Tanner MA, Berk LS, Felten DL, Blidy AD, Bit SL, Ruff DW. Substantial changes in gene expression level due to the storage temperature and storage duration of human whole blood. Clin Lab Haematol 2002;24:337– 41. 14. Kumar P, Satchidanandam V. Ethyleneglycol-bis-(␤-aminoethylether)tetraacetate as a blood anticoagulant: preservation of antigen-presenting cell function and antigen-specific proliferative response of peripheral blood mononuclear cells from stored blood. Clin Diagn Lab Immunol 2000;7:578 – 83. DOI: 10.1373/clinchem.2005.048546

Denaturing HPLC Coupled with Multiplex PCR for Rapid Detection of Large Deletions in Duchenne Muscular Dystrophy Carriers, Chia-Cheng Hung,1† Yi-Ning Su,2† Chia-Yun Lin,1 Chih-Chao Yang,3 Wang-Tso Lee,4 ShuChin Chien,2 Win-Li Lin,1 and Chien-Nan Lee5* (1 Institute of Biomedical Engineering, National Taiwan University, Taipei, Taiwan; Departments of 2 Medical Genetics, 3 Neurology, 4 Pediatrics, and 5 Obstetrics and Gynecology, National Taiwan University Hospital, Taipei, Taiwan; † C-C. Hung and Y-N. Su contributed equally to this study; * address correspondence to this author at: Department of Obstetrics and Gynecology, National Taiwan University Hospital,Taipei,Taiwan;fax886-2-2392-0470,e-mailleecn@ ha.mc.ntu.edu.tw) Duchenne muscular dystrophy (DMD) and Becker muscular dystrophy (BMD) affect 1 in 3500 newborn male infants (1 ). DMD and BMD are both inherited in an

Clinical Chemistry 51, No. 7, 2005

X-linked recessive pattern resulting from mutations in the dystropin gene on Xp21.1. In approximately two-thirds of cases, the disease is inherited from female carriers, and the remaining cases come from de novo mutations in individuals without a family history of muscular dystrophy (2, 3 ). Approximately 60% of DMD cases are associated with large intragenic deletions of 1 or more exons in 2 hotspot regions in the proximal and central regions of the gene (exons 3–7 and exons 44 –55) (4 – 8 ). Approximately 6% of DMD mutations are associated with duplications of large segments (9 ), and the remaining cases result from point mutations, small deletions, or insertions (10 –12 ). Affected males can be readily detected by the absence of an amplification product in a multiplex PCR, as described by Chamberlain et al. (13 ) and Beggs et al. (14 ). Up to 98% of all frequent deletions can be detected, but in carriers, the nondeleted X chromosome hampers detection, making identification of DMD carrier status difficult. Several DMD carrier identification strategies based on quantitative Southern blotting (15 ), fluorescent in situ hybridization (16 ), linkage analysis (17 ), fluorescencebased strategies, microchip electrophoresis, capillary electrophoresis (18 –27 ), multiplex amplifiable probe hybridization (28 ), and multiplex ligation-dependent probe amplification (29 ) have been described. Denaturing HPLC (DHPLC) is a simple, rapid, non-gel– based, non-fluorescence– based method that uses ion-pair reversed-phase liquid chromatography for detection of DNA variations; the method is sensitive and specific (30 ). We describe a new application combining DHPLC with multiplex PCR to efficiently and accurately identify carrier and noncarrier status in DMD families. We analyzed 84 DNA samples from the National Taiwan University Hospital, including DNA from 11 patients with dystropin gene deletions previously detected by gel electrophoresis, 1 patient with a duplication detected by the multiplex PCR/DHPLC detection method and confirmed by quantitative real-time PCR (27 ), 23 obligate carriers and noncarriers from families with DMD patients, and 50 unaffected females from the general population. Genomic DNA was collected from peripheral whole blood by use of a Puregene DNA Isolation Kit (Gentra Systems) according to the manufacturer’s instructions. To amplify the dystropin gene, multiplex PCRs of the dystropin gene were carried out as described in Chamberlain et al. (13 ) and Beggs et al. (14 ). The reaction described by Beggs et al. (14 ) uses multiplex I (exons 3, 50, 6, and 60) and multiplex II [muscle promoter (pm), exons 43, 13, 47, and 52]. The reaction described by Chamberlain et al. (13 ) uses multiplex III (exons 48, 17, 8, 44, and 46) and multiplex IV (exons 45, 19, 51, 12, and 4). Each multiplex PCR for the specific DNA fragments was performed in a total volume of 50 ␮L containing 200 ng of genomic DNA, 0.04 – 0.4 ␮M each primer, 200 ␮M deoxynucleotide triphosphates, 1 U of AmpliTaq Gold enzyme (PE Applied Biosystems), and 5 ␮L of GeneAmp 10⫻ buffer II [10 mM Tris-HCl (pH 8.3), 50 mM KCl] in 2 mM MgCl2 as provided by the manufacturer. Because

1253

amplification efficiencies differed among primer pairs in multiplex PCR, different amounts of the primers were used to obtain comparable signals within different exons. PCR amplification was performed on an MBS thermocycler (ThermoHybaid) with an initial denaturation step at 95 °C for 10 min followed by 24 cycles of denaturation at 94 °C for 1 min, annealing at 59 °C for 1 min, and extension at 72 °C for 3 min, with a final extension step at 72 °C for 10 min. For multiplex detection, we loaded 40 ␮L of crude PCR product on a DNASep column (Transgenomic) and conducted DHPLC analysis in a WAVE DHPLC instrument (Transgenomic) as described previously (30 ). The initial and final concentrations of solvent B in the 12-min gradient were 46% and 70%, the flow rate was 0.9 mL/min, and the temperature was 50 °C. For each sample, we measured the heights and the areas of the different peaks from each exon, using the WAVE MAKER software, which labeled the data automatically. To identify female carrier status, as indicated by the presence of large deletions or duplications in the gene dosage analysis, the copy number of a specific test exon in the unknown samples was determined by the formula: Copy number Peak height (area) of test exon (U)/ Peak height (area) of reference exon (U) ⫻2 ⫽ Peak height (area) of test exon (C)/ Peak height (area) of reference exon (C) where U indicates the unknown sample and C the control sample. The reference exon is an undeleted or unduplicated exon in the same multiplex group as a deleted or duplicated/test exon. We determined the linearity of the PCR reaction by testing the relationships among the number of PCR cycles (22–28 cycles), the DNA concentration (50 – 400 ng), and the amount of PCR product (31 ), and then standardizing the number of quantitative PCR reaction cycles to 24 cycles with a DNA concentration of 200 ng. Using this technique to detect common exon deletions in the dystropin gene, we assigned 19 DNA fragments to 4 sets of multiplex PCR reactions according to their sizes and amplification efficiencies. To detect gene deletions, we analyzed the PCR products by both gel electrophoresis and DHPLC (Fig. 1 in the Data Supplement that accompanies the online version of this Technical Brief at http:// www.clinchem.org/content/vol51/issue7/). Deleted exons are indicated by the absence of corresponding signals in the affected cases compared with controls. The results of multiplex PCR/DHPLC analysis of affected males, carriers, and noncarriers from different DMD families are shown in Fig. 1. All results were analyzed by use of the ratios of deleted and undeleted exons. Every sample was analyzed at least 3 times, and the results were reproducible. Individuals were identified with decreased amplification of signals that were absent in affected individuals with deleted exons. In family C, no

1254

Technical Briefs

Fig. 1. DHPLC chromatography of 4 multiplex PCR sets for 4 different families. Family A, affected male and the carrier individual (his mother) with deleted exons 48 –52; Family B, affected male and the carrier individual (his mother) with deleted exon 51; Family C, affected male with deleted exons 50 –52 and his noncarrier mother; Family D, affected male and the carrier individual (his mother) with deleted exons 45– 60; E, affected male with duplicated exons 12–19.

dosage change is apparent in the mother; therefore, de novo mutation can be deduced. In family D, more than 1 exon deletion exists in the same multiplex PCR reaction.

In this situation, carrier state can still be correctly identified by calculating the ratios of tested exons to reference exons.

Clinical Chemistry 51, No. 7, 2005

Measured copy numbers of the dystropin gene from carriers, noncarriers, and unaffected females, obtained by multiplex PCR/DHPLC analysis and expressed in testexon to reference-exon ratios, are shown in Fig. 2 of the online Data Supplement. The measured copy numbers between deleted and undeleted exons can be differentiated unambiguously with the use of this analytical tool, enabling successful identification of unaffected females (Table 1 of the online Data Supplement). In our study, the peak height and peak area for PCR amplification results from each exon were divided by the reference exon, and the ratios of the unknown samples were compared with the control samples to determine the copy number of each exon. Theoretically, the copy number should be 1 for a single copy and 2 for a double copy. The mean values measured by peak height and peak area for a single and double copy are shown in Table 1 of the online Data Supplement. The values never overlapped in our controls. Large deletions in all of the DMD carriers were clearly identified with the multiplex PCR/DHPLC assay. Furthermore, 1 study patient with gene duplications was identified (Fig. 1). Recently, a similar strategy was used to detect RB1 gene rearrangement (32 ). Unlike other available techniques, including capillary gel electrophoresis, multiplex amplifiable probe hybridization, and multiplex ligation-dependent probe amplification, the multiplex PCR/DHPLC assay uses ultraviolet detection and automated instruments that can speed up carrier detection without the use of radioisotopes or fluorescence-labeled or gel-based materials. For clinical applications, a specific advantage of this detection system is the procedure for interpreting carrier status. Another advantage is that the crude PCR products do not require further purification before DHPLC analysis. Our method took ⬃3 h for PCR amplification and ⬃48 min for DHPLC analysis for 4 multiplex sets. The DHPLC analysis cost was less than US $4.00 for each sample. This technique thus is well suited for routine diagnostics. In conclusion, the multiplex PCR/DHPLC detection system is a simple, rapid, and powerful assay enabling direct detection of deletions and duplications in DMD patients and carriers. This system may also be adapted for diagnostic use in other genetic diseases involving deletion and duplication mutations.

6.

7.

8.

9.

10. 11. 12.

13.

14. 15. 16.

17.

18.

19.

20.

21.

22.

23.

24.

This work was supported by grants from the National Science Council of Taiwan (NSC 93-2314-B-002-067). 25.

References 1. Emery AE. Population frequencies of inherited neuromuscular diseases–a world survey. Neuromuscul Disord 1991;1:19 –29. 2. Mukherjee M, Chaturvedi LS, Srivastava S, Mittal RD, Mittal B. De novo mutations in sporadic deletional Duchenne muscular dystrophy (DMD) cases. Exp Mol Med 2003;35:113–7. 3. Sinha S, Mishra S, Singh V, Mittal RD, Mittal B. High frequency of new mutations in North Indian Duchenne/Becker muscular dystrophy patients. Clin Genet 1996;50:327–31. 4. Baumbach LL, Chamberlain JS, Ward PA, Farwell NJ, Caskey CT. Molecular and clinical correlations of deletions leading to Duchenne and Becker muscular dystrophies. Neurology 1989;39:465–74. 5. Darras BT, Blattner P, Harper JF, Spiro AJ, Alter S, Francke U. Intragenic

26.

27.

28.

1255

deletions in 21 Duchenne muscular dystrophy (DMD)/Becker muscular dystrophy (BMD) families studied with the dystrophin cDNA: location of breakpoints on HindIII and BglII exon-containing fragment maps, meiotic and mitotic origin of the mutations. Am J Hum Genet 1988;43:620 –9. Den Dunnen JT, Grootscholten PM, Bakker E, Blonden LA, Ginjaar HB, Wapenaar MC, et al. Topography of the Duchenne muscular dystrophy (DMD) gene: FIGE and cDNA analysis of 194 cases reveals 115 deletions and 13 duplications. Am J Hum Genet 1989;45:835– 47. Gillard EF, Chamberlain JS, Murphy EG, Duff CL, Smith B, Burghes AH, et al. Molecular and phenotypic analysis of patients with deletions within the deletion-rich region of the Duchenne muscular dystrophy (DMD) gene. Am J Hum Genet 1989;45:507–20. Lindlof M, Kiuru A, Kaariainen H, Kalimo H, Lang H, Pihko H, et al. Gene deletions in X-linked muscular dystrophy. Am J Hum Genet 1989;44:496 – 503. Hu XY, Ray PN, Murphy EG, Thompson MW, Worton RG. Duplicational mutation at the Duchenne muscular dystrophy locus: its frequency, distribution, origin, and phenotype-genotype correlation. Am J Hum Genet 1990; 46:682–95. Rininsland F, Reiss J. Microlesions and polymorphisms in the Duchenne/ Becker muscular dystrophy gene. Hum Genet 1994;94:111– 6. Roberts RG, Bobrow M, Bentley DR. Point mutations in the dystrophin gene. Proc Natl Acad Sci U S A 1992;89:2331–5. Bennett RR, den Dunnen J, O’Brien KF, Darras BT, Kunkel LM. Detection of mutations in the dystrophin gene via automated DHPLC screening and direct sequencing. BMC Genet 2001;2:17. Chamberlain JS, Gibbs RA, Ranier JE, Nguyen PN, Caskey CT. Deletion screening of the Duchenne muscular dystrophy locus via multiplex DNA amplification. Nucleic Acids Res 1988;16:11141–56. Beggs AH, Koenig M, Boyce FM, Kunkel LM. Detection of 98% of DMD/BMD gene deletions by polymerase chain reaction. Hum Genet 1990;86:45– 8. Mao YP, Cremer M. Detection of Duchenne muscular dystrophy carriers by dosage analysis using the DMD cDNA clone 8. Hum Genet 1989;81:193–5. Ried T, Mahler V, Vogt P, Blonden L, van Ommen GJ, Cremer T, et al. Direct carrier detection by in situ suppression hybridization with cosmid clones of the Duchenne/Becker muscular dystrophy locus. Hum Genet 1990;85: 581– 6. Clemens PR, Fenwick RG, Chamberlain JS, Gibbs RA, de Andrade M, Chakraborty R, et al. Carrier detection and prenatal diagnosis in Duchenne and Becker muscular dystrophy families, using dinucleotide repeat polymorphisms. Am J Hum Genet 1991;49:951– 60. Ferrance J, Snow K, Landers JP. Evaluation of microchip electrophoresis as a molecular diagnostic method for Duchenne muscular dystrophy. Clin Chem 2002;48:380 –3. Yau SC, Bobrow M, Mathew CG, Abbs SJ. Accurate diagnosis of carriers of deletions and duplications in Duchenne/Becker muscular dystrophy by fluorescent dosage analysis. J Med Genet 1996;33:550 – 8. Fortina P, Cheng J, Shoffner MA, Surrey S, Hitchcock WM, Kricka LJ, et al. Diagnosis of Duchenne/Becker muscular dystrophy and quantitative identification of carrier status by use of entangled solution capillary electrophoresis. Clin Chem 1997;43:745–51. Gelfi C, Orsi A, Leoncini F, Righetti PG, Spiga I, Carrera P, et al. Amplification of 18 dystrophin gene exons in DMD/BMD patients: simultaneous resolution by capillary electrophoresis in sieving liquid polymers. Biotechniques 1995;19:254 – 8, 60 –3. Shen Y, Xu Q, Han F, Ding K, Song F, Fan Y, et al. Application of capillary nongel sieving electrophoresis for gene analysis. Electrophoresis 1999;20: 1822– 8. Pastore L, Caporaso MG, Frisso G, Orsini A, Santoro L, Sacchetti L, et al. A quantitative polymerase chain reaction (PCR) assay completely discriminates between Duchenne and Becker muscular dystrophy deletion carriers and normal females. Mol Cell Probes 1996;10:129 –37. Allingham-Hawkins DJ, McGlynn-Steele LK, Brown CA, Sutherland J, Ray PN. Impact of carrier status determination for Duchenne/Becker muscular dystrophy by computer-assisted laser densitometry. Am J Med Genet 1998;75:171–5. Mansfield ES, Robertson JM, Lebo RV, Lucero MY, Mayrand PE, Rappaport E, et al. Duchenne/Becker muscular dystrophy carrier detection using quantitative PCR and fluorescence-based strategies. Am J Med Genet 1993;48:200 – 8. Frisso G, Carsana A, Tinto N, Calcagno G, Salvatore F, Sacchetti L. Direct detection of exon deletions/duplications in female carriers of and male patients with Duchenne/Becker muscular dystrophy. Clin Chem 2004;50: 1435– 8. Joncourt F, Neuhaus B, Jostarndt-Foegen K, Kleinle S, Steiner B, Gallati S. Rapid identification of female carriers of DMD/BMD by quantitative real-time PCR. Hum Mutat 2004;23:385–91. White S, Kalf M, Liu Q, Villerius M, Engelsma D, Kriek M, et al. Comprehensive detection of genomic duplications and deletions in the DMD gene, by use of multiplex amplifiable probe hybridization. Am J Hum Genet 2002;71: 365–74.

1256

Technical Briefs

29. Schouten JP, McElgunn CJ, Waaijer R, Zwijnenburg D, Diepvens F, Pals G. Relative quantification of 40 nucleic acid sequences by multiplex ligationdependent probe amplification. Nucleic Acids Res 2002;30:e57. 30. Xiao W, Oefner PJ. Denaturing high-performance liquid chromatography: a review. Hum Mutat 2001;17:439 –74. 31. Cha RS, Thilly WG. Specificity, efficiency, and fidelity of PCR. PCR Methods Appl 1993;3:S18 –29. 32. Dehainault C, Lauge A, Caux-Moncoutier V, Pages-Berhouet S, Doz F, Desjardins L, et al. Multiplex PCR/liquid chromatography assay for detection of gene rearrangements: application to RB1 gene. Nucleic Acids Res 2004;32:e139. Previously published online at DOI: 10.1373/clinchem.2004.046144

Congenital Analbuminemia attributable to Compound Heterozygosity for Novel Mutations in the Albumin Gene, Filomena Campagna,1 Francesca Fioretti,1 Marco Burattin,1 Stefano Romeo,2 Federica Sentinelli,2 Maura Bifolco,1 Maria Isabella Sirinian,1 Maria Del Ben,3 Francesco Angelico,3 and Marcello Arca1* (1 Department of Clinical and Applied Medical Therapy, 2 Department of Clinical Sciences, and 3 IV Division of Internal Medicine, Department of Experimental Medicine and Pathology, University of Rome, La Sapienza, Italy; * address correspondence to this author at: Dipartimento di Clinica e Terapia Medica Applicata, Universita` di Roma “La Sapienza”, Policlinico Umberto I, Viale del Policlinico, 155, 00161 Rome, Italy; fax 39-064440290, e-mail [email protected]) Congenital analbuminemia is a rare autosomal recessive disorder characterized by the absence or very low concentrations of serum albumin (HSA) (1 ). The disorder is conventionally defined as an HSA ⬍1 g/L (as determined by immunoassay) associated with normal liver function and absence of protein loss (2 ). The incidence of congenital analbuminemia is estimated to be ⬍1 in 1 million births, without sex or race predilection. To date, 39 cases of congenital analbuminemia have been reported, 16 of which were diagnosed in children (3 ). Despite multiple functions of HSA (4 ), its absence is a relatively tolerable condition. Except for associated hyperlipidemia, minor edema, and mild fatigability, analbuminemic individuals suffer few adverse symptoms (3, 5, 6 ). Such relative mildness of symptoms is attributed to a compensatory increase in hepatic biosynthesis of other plasma proteins (7, 8 ). Congenital analbuminemia is attributable to defects in the gene coding for HSA. The HSA gene is located on chromosome 4 and is split into 15 exons by 14 intervening introns (9 ). The identification of multiple mutations that cause analbuminemia clearly indicates that it is a genetically heterogeneous disorder. Six of these mutations introduced stop codons (10 –13 ), and 2 caused splicing defects and determined a premature termination of the translation (14, 15 ). All were found to be present in the homozygous state. In the present study we characterized another case of congenital analbuminemia caused by 2 newly identified mutations in an Italian family. The proband was a 29-year-old Italian man with an

HSA of 10 –12 mg/L as determined by immunoassay. His medical history has been reported elsewhere (16, 17 ). The proband’s relatives had HSA concentrations in the lower end of the reference interval (father, 35 g/L; mother, 33 g/L; sister, 42 g/L). All gave informed consent before participating in the study. We used single-strand conformation polymorphism analysis to screen the coding region of the HSA gene in samples from the proband and a control individual. Briefly, the 14 exons and their intron– exon junctions were PCR-amplified from genomic DNAs by use of specific primer pairs as already reported (10 ). After amplification, 2 ␮L of PCR products was mixed with 8 ␮L of singlestrand conformation polymorphism analysis buffer containing 950 mL/L formamide, 10 mmol/L NaOH, and 0.5 g/L of both xylene cyanol and bromphenol blue. The mixtures were denatured at 95 °C for 3 min and chilled on ice before electrophoretic separation. Denatured and nondenatured samples were then loaded on a 1⫻ Mutation Detection Enhancement gel (FMC BioProducts) and visualized with silver staining. No abnormal bands were detected in the proband sample, except in exon 13. Direct sequencing of this fragment revealed a silent C-to-T substitution at nucleotide 15229. This variation, which introduced a recognition site for the restriction enzyme SacI (New England Biolabs), was demonstrated by restriction fragment length polymorphism (RFLP) analysis to be a common single-nucleotide polymorphism in samples from 50 control individuals (allele frequency, 0.585). We next investigated the 14 PCR products by direct sequencing. Both strands were sequenced with the fluorescent dideoxy termination method (BigDye Terminator Cycle Sequencing Kit, Ver. 3.1; Applied Biosystems) on an ABI 310 automated sequencer according to the manufacturer’s protocol (Applied Biosystems). Electropherograms were analyzed with the ABI Prism 310 Data Collector software. The first mutation was identified in exon 10 and consisted of a C3 T transition at nucleotide 11999 (Fig. 1A). This mutation changes codon CAG for Gln-385 to stop codon TAG (Gln385Stop). The putative protein product should have a length of 384 amino acid residues. The proband was heterozygous for the defect as demonstrated by the sequencing electrophoretogram (Fig. 1A). As this mutation abolished a restriction site for the BseMII enzyme (MBI Fermentas), we used RFLP analysis to evaluate its segregation in the proband’s family. As shown in Fig. 1B, the digestion pattern revealed the presence of an abnormal 340-bp fragment in the proband and his father, whereas the mother and the proband’s sister were normal. This confirmed that the proband was heterozygous for Gln385Stop and demonstrated that he inherited this mutation from his father. The Gln385Stop mutation was named Roma2, for the city of origin of the father. This mutation was not found in 50 DNA samples obtained from healthy individuals. We were unable to evaluate whether the putative truncated protein produced by the Roma2 allele was present in serum. In all previous cases with a stop codon

Clinical Chemistry 51, No. 7, 2005

within an exon, no evidence was found for the presence in the serum of the truncated albumin form. The second mutation was found in exon 11. The penultimate nucleotide of exon 11 had undergone an A3 G transition at position 13378 (Fig. 1C). This mutation changes the codon TAT for Tyr-452 to codon TGT (see Fig. 1A of the Data Supplement that accompanies the online version of this Technical Brief at http://www. clinchem.org/content/vol51/issue7/). This mutation was also confirmed by restriction analysis using the HpyCH4 III enzyme (New England Biolabs). We found that it was inherited from the mother (Fig. 1D). We named this mutation Fondi for the small town in Central Italy where she was born.

1257

As the GT dinucleotide is known to be a site for intron splicing, we predicted that the Fondi mutation introduced a novel, anticipated donor splicing site replacing the first 2 bases of intron 12. To confirm this hypothesis, we examined HSA cDNA obtained from proband’s leukocytes. Briefly, total RNA was extracted with use of TRIzol (Invitrogen) according to the manufacturer’s protocol. The RNA was reverse-transcribed by use of MultiScribe reverse transcriptase (Applied Biosystems) with random hexamers to obtain total cDNA. The primer pair 5⬘ATTGTGAGCAAAAAGAGCAGCTTG-3⬘ (forward) and 5⬘-TTTTGTTGCCTTGGGCTTGTGTTT-3⬘ (reverse) was designed from mRNA sequences in GenBank by use of Primer3. We used this pair of primers, spanning exon 10

Fig. 1. Sequence of the HSA gene of the analbuminemic proband and RFLP analysis of samples from the proband’s family. (A), DNA sequence of exon 10 showing the C3 T transition at position 11999 (Roma2 allele). The arrow indicates the nucleotide that is mutated in the patient. The proband is heterozygous for the defect, as seen by the presence of the 2 superimposed peaks representing the normal and abnormal base. (B), RFLP analysis of the Roma2 mutation in the proband’s family and in a control individual (lane C); the BseMII enzyme digested the 406-bp PCR product, giving 3 fragments of 245, 95, and 66 bp. The Roma2 mutation eliminates 1 restriction site for the enzyme, thus producing an abnormal 340-bp fragment. The proband (indicated by the arrow) and his father were heterozygous for this mutation. (C), DNA sequence of exon 11 showing the A3 G transition at position 13378 (Fondi allele). The proband is heterozygous for the defect, as seen by the presence of the 2 superimposed peaks. (D), RFLP analysis of the Fondi mutation in the proband’s family and in a healthy control (lane C); the HpyCH4 III enzyme digested the 343-bp PCR product into 2 fragments of 288 and 55 bp. The Fondi mutation introduces 1 restriction site for the enzyme, thus producing 2 abnormal fragments of 178 and 110 bp. The proband (indicated by the arrow) and his mother are heterozygous for the Fondi mutation.

1258

Technical Briefs

to exon 13, to PCR-amplify the region of HSA cDNA where the mutation was located. All reactions were performed on a T Personal (Biometra) thermocycler in a 20-␮L reaction volume. Amplified products were then sequenced using the reported primers as described above. The cDNA electropherogram from the proband showed a double sequence starting from nucleotide 13378, a finding that indicates the presence of both wild-type and mutated alleles (see Fig. 1B of the online Data Supplement). These results confirmed that the mutation disrupts the normal codon TAT for Tyr-452, leaving the residual T nucleotide to join the first 2 CT nucleotides of exon 12 (see Fig. 1A of the online Data Supplement). This effect leads to a reading frameshift that introduces a premature stop codon 12 amino acids downstream in exon 12 (see Fig. 1A of the online Data Supplement). The translation product from the Fondi allele is expected to be a truncated protein consisting of 463 instead of the 585 amino acid residues usually found in mature HSA. In previously reported cases of analbuminemia attributable to aberrant splicing, the consequences of the defect at the mRNA level were not evaluated (13, 14 ). We were able to obtain the mutated HSA mRNA from circulating leukocytes and thus directly confirmed that the Fondi allele caused a defect in the ligation of exon 11– exon 12 sequences. How aberrant splicing of the HSA gene affects mRNA processing is poorly known. Studies on Nagase analbuminemic rats, with a splicing error in the HSA gene (18, 19 ), demonstrated a marked decrease in hepatic mRNA (20 ). Although we were unable to measure the specific amount of mutated mRNA in our study participants, a substantial amount of cDNA was obtained by reverse transcription-PCR from our proband’s leukocytes, suggesting that the Fondi allele did not cause, at least in leukocytes, a complete degradation of the mutant mRNA. In summary, we report the first case of congenital analbuminemia attributable to compound heterozygosity for 2 new mutations in the HSA gene. We also demonstrated that circulating leukocytes may be used to investigate the effects of analbuminemia-causing mutations on mRNA processing, making it possible to reevaluate all analbuminemic patients in whom molecular characterization of the HSA gene was carried out only at the genomic level. References 1. Russi E, Weigand K. Analbuminemia. Klin Wochenschr 1983;61:541–5. 2. Lyon AW, Meinert P, Bruce GA, Laxda VA, Salkie ML. Influence of methodology on the detection and diagnosis of congenital analbuminemia. Clin Chem 1998;44:2365–7. 3. Koot BG, Houwen R, Pot DJ, Nauta J. Congenital analbuminaemia: biochemical and clinical implications. A case report and literature review. Eur J Pediatr 2004;163:664 –70. 4. Peters T Jr. All about albumin: biochemistry, genetics and medical applications. Orlando, FL: Academic Press, 1996:432pp. 5. Gossi B, Kleinert D, Gossi U. A further case of analbuminemia. Schwiez Med Wochenschr 2000;130:583–9. 6. Peters T. Albumin website. http//www.albumin.org (accessed March 5, 2005). 7. Frohlich J, Pudek MR, Cormode EJ, Sellers EM, Abel JG. Further studies on plasma proteins, lipids, and dye- and drug-binding in a child with analbuminemia. Clin Chem 1981;27:1213– 6. 8. Boman H, Hermodson M, Hammond CA, Motulsky AG. Analbuminemia in an American Indian girl. Clin Genet 1976;9:513–26.

9. Minghetti PP, Ruffner DE, Kuang WJ, Dennison OE, Hawkins JW, Beattie WG, et al. Molecular structure of the human albumin gene is revealed by nucleotide sequence within q11-22 of chromosome 4. J Biol Chem 1986; 261:6747–57. 10. Watkins S, Madison L, Galliano M, Minchiotti L, Putnam FW. Analbuminemia: three cases resulting from different point mutations in the albumin gene, Proc Natl Acad Sci U S A 1994;91:9417–21. 11. Watkins S, Madison J, Galliano M, Minchiotti L, Putnam FW. A nucleotide insertion and frameshift cause analbuminemia in an Italian family. Proc Natl Acad Sci U S A 1994;91:2275–9. 12. Galliano M, Campagnoli M, Rossi A, Wirsing von Konig CH, Lyon AW, Cefle K, et al. Molecular diagnosis of analbuminemia: a novel mutation identified in two Amerindian and two Turkish families. Clin Chem 2002;48:844 –9. 13. Campagnoli M, Sala A, Romano A, Rossi A, Nauta J, Koot BG, et al. Novel nonsense mutation causes analbuminemia in a Moroccan family. Clin Chem 2005;51:227–9. 14. Ruffner DE, Dugaiczyk A. Splicing mutation in human hereditary analbuminemia Proc Natl Acad Sci U S A 1988;85:2125–9. 15. Campagnoli M, Rossi A, Palmqvist L, Flisberg A, Niklasson A, Minchiotti L, et al. A novel splicing mutation causes an undescribed type of analbuminemia. Biochim Biophys Acta 2002;1586:43–9. 16. Papi N, Castiglioni A, Reale A. Su di un caso di analbuminemia congenita. Riv Ital Pediatr 1983;9:85–7. 17. Del Ben M, Burattin M, Arca M, Ceci F, Violi F, Angelico F. Treatment of severe hypercholesterolemia with atorvastatin in congenital analbuminemia. Am J Med 2004;117:803– 4. 18. Nagase S, Shimamune K, Shumiya S. Albumin-deficient rat mutant. Science 1979;205:590 –1. 19. Esumi H, Takahashi Y, Sato S, Nagase S, Sugimura T. A seven-base-pair deletion in an intron of the albumin gene of analbuminemic rats. Proc Natl Acad Sci U S A 1983;80:95–9. 20. Shalaby F, Shafritz DA. Exon skipping during splicing of albumin mRNA precursors in Nagase analbuminemic rats. Proc Natl Acad Sci U S A 1990;87:2652– 6. DOI: 10.1373/clinchem.2005.048561

Circulating Ghrelin in Patients Undergoing Elective Cholecystectomy, Claudio Chiesa,1,3* John F. Osborn,2 Lucia Pacifico,1,3 Guglielmo Tellan,4 Pier Michele Strappini,3 Roberto Fazio,4 and Giovanna Delogu4 (1 National Research Council, Rome, Italy; Departments of 2 Public Health Science, 3 Pediatrics, and 4 Anesthesia and Intensive Care, “La Sapienza” University of Rome, Rome, Italy; * address correspondence to this author at: Department of Pediatrics, “La Sapienza” University of Rome, Viale Regina Elena, 324 00161 Rome, Italy; fax 39-06-4997-9215, e-mail [email protected]) Secreted predominantly from the stomach (1 ), ghrelin is a peptide identified in 1999 as an endogenous ligand of the growth hormone (GH) secretagogue receptor located on the pituitary gland, thus fulfilling the criteria of a brain– gut peptide (2, 3 ). The brain– gut axis serves as an effector of anabolism by regulating growth, feeding, and metabolism via vagal afferent-mediating ghrelin signaling (2, 4 ). The role of ghrelin as a brain– gut peptide emphasizes the significance of afferent vagal fibers as a major pathway to the brain, serving the purpose of maintaining physiologic homeostasis (2, 4 ). The importance of ghrelin as a “hunger hormone” with orexigenic effects mediated by the hypothalamic peptides, agouti-related peptide, and neuropeptide Y, and the fact that it is the most potent peripheral signal of diminishing energy stores, implies that ghrelin release might be the most important of the

1259

Clinical Chemistry 51, No. 7, 2005

min before induction of anesthesia (sampling time t1); after induction of anesthesia (t2); at 30 min intraoperatively (t3); at the end of surgery, i.e., at exsufflation or incision closure (t4); in the first postoperative hours [t5 and t6: 6 h (afternoon) and 10 h (evening) after the start of surgery]; and on the morning (0800) of the first postoperative day (t7). Serum concentrations of immunoreactive ghrelin were measured in duplicate with an RIA that recognizes both the acylated and des-acyl forms (Phoenix Pharmaceuticals; lower limit of detection, 10 ng/L; interand intraassay CVs, 9.0%–13.6% and 4.5%–5.3%, respectively, as reported by the manufacturer). A frequency distribution of the 30 measured ghrelin concentrations at each of the 7 time points was constructed and inspected. Given that every frequency distribution was positively skewed, the natural logarithm of each observed value was calculated, and these logarithms were used in all subsequent calculations. To investigate how the mean ghrelin concentration changed between the first and the seventh and last observation, we applied 2-way ANOVA (in which the rows represented the patients and the columns were the 7 time points) to the logarithms of the ghrelin concentrations to determine whether there were significant differences between the mean concentrations at times t1 to t7. We compared the differences between the observed values at times t2 to t7 with the baseline value (t1) by paired t-test and used regression analysis to investigate whether the patients’ sex and the type of operation influenced the mean logarithmic values of the ghrelin concentrations. The 2-way ANOVA revealed highly significant (P ⬍0.00005) differences between the mean values at times t1 to t7. The geometric mean concentrations of ghrelin at sampling times t1 to t7 are shown in Table 1. After induction of anesthesia, at t2, the geometric mean ghrelin concentrations decreased by 5%, but this difference was not statistically significant compared with baseline values. By time t3, ghrelin had decreased by 23% (P ⫽ 0.035) compared with baseline values, and by time t4, the concentration had decreased by 30% (P ⫽ 0.01), reaching a minimum at t5 with a decrease of 39% (P ⬍0.00005) compared with baseline values. At time t6, the concentrations had increased slightly, and by t7 they had returned to baseline values. In regression analysis, sex, duration of surgery, and type of surgical procedure had no significant effect on log ghrelin concentrations when adjusted for patients and times.

many redundant mechanisms ensuring human survival in times of famine (4 ). However, the wide tissue distribution of ghrelin suggests that it may have other functions as well (5 ). Further characterization of the functions of ghrelin is fundamental to discovering new approaches to the diagnosis and treatment of different disease entities, including those related to the catabolic response to surgical trauma (2 ). In this preliminary study, we report the pattern of ghrelin secretion in different perioperative periods in patients undergoing elective cholecystectomy. Thirty patients [17 females; age range, 20 – 45 years; body mass index (BMI), 18 –25 kg/m2] with ultrasoundconfirmed cholecystolithiasis undergoing elective cholecystectomy via laparoscopy or laparotomy at the Surgical Department at the Hospital Umberto I° of Rome were recruited to the study. All patients presented a low surgical risk (American Society of Anesthetists’ score 1 or 2). Patients with a diagnosis of choledocholitiasis, jaundice, acute cholecystitis, and pancreatitis; those with a history of metabolic, endocrine, hepatic, cardiac, or renal disease; those who, before surgery, received medication known to interfere with hormonal responses to stress; and those with a history of tobacco use or substance abuse were not eligible for the study. The study was approved by the Hospital Ethical Committee, and each participant gave informed consent to be included in the study. After an overnight fast, between 0830 and 0900 in the morning, 20 study patients (13 females) underwent laparoscopic cholecystectomy (LC) by a 4-trocar technique, whereas 10 (4 females) underwent open cholecystectomy (OC) via a right subcostal incision. No LC patient required conversion to OC. There were no significant differences in the mean duration of surgery (time elapsed from the end of the induction of anesthesia until exsufflation or incision closure) between LC [mean (SD), 91.9 (36.8) min; range, 40 –190 min] and OC [108 (30.9) min; range, 80 –165 min] groups. All patients had an uneventful postoperative clinical course. Induction of anesthesia consisted of intravenous fentanyl (2 ␮g/kg), thiopental (5 mg/kg), and vecuronium (0.08 mg/kg). Anesthesia was maintained with 1% sevoflurane in a mixture of 60% nitrous oxide and 40% oxygen and intravenous fentanyl and vecuronium. Fluid replacement was performed with normal saline solution during surgery (7 mL 䡠 kg⫺1 䡠 h⫺1) and throughout the postoperative course (1 mL 䡠 kg⫺1 䡠 h⫺1) until the next morning after surgery (see below) when the seventh blood sample was taken. Blood samples were collected from each patient at ⬃60

Table 1. Ghrelin values in the cohort of 30 patients undergoing elective cholecystectomy at the sampling times (t1 to t7) defined in the text. Sampling times

Ghrelin, ng/L Geometric mean 95% confidence interval a

Compared with t1:

a

P ⬍0.05;

b

t1

t2

t3

t4

t5

t6

t7

231 197–271

221 188–258

178a 152–208

161a 137–188

142b 121–166

156c 133–182

250 213–293

P ⬍0.0001;

c

P ⬍0.01.

1260

Technical Briefs

To our knowledge, this is the first report on the characteristics of ghrelin secretion in different perioperative periods of a cohort of adult patients undergoing elective cholecystectomy. The novel finding of our study is that ghrelin may be added to the list of substances whose concentrations remain stable until after the induction of anesthesia, at time t2 (6 ). However, the finding of a significant decrease in ghrelin concentrations from baseline during the intraoperative period, as well as over the course of several hours after surgery, was somewhat surprising. In general, ghrelin concentrations are inversely correlated with positive energy balance, BMI, body fat mass, adipocyte size, and leptin concentrations (7–9 ), whereas they are lower in obese persons than in controls (9 ). Pima Indians, known for their propensity to develop type II diabetes and obesity, also have lower circulating ghrelin concentrations, independent of BMI, compared with matched controls (9 ). Patients with anorexia nervosa have higher plasma ghrelin concentrations than age- and sex-matched controls, and weight gain lowers their increased ghrelin concentrations (10 ). Thus, fluctuations in plasma ghrelin concentrations may reflect physiologic adaptation to long-term alterations in energy balance (4 ). Although ghrelin concentrations may increase in response to short-term fasting, the effects of a more prolonged fast on ghrelin have not been well characterized (11, 12 ). Only a very few studies have evaluated ghrelin concentrations in humans who have fasted completely and therefore have lost the meal-related pattern. Using single measurements, Norrelund et al. (13 ), in a study of healthy individuals and GH-deficient patients during a 36-h fast, found no significant changes in ghrelin concentrations in either group. In contrast, Muller et al. (14 ) reported that healthy nonobese individuals who fasted for 3 days developed a diurnal rhythm in ghrelin characterized by low concentrations in the morning with subsequent increases in the afternoon and at midnight. This study was, however, difficult to evaluate because a synthetic GH secretagogue was administered before and during fasting and plasma ghrelin was measured every 6 h in a relatively small number of individuals (n ⫽ 10). Recently, Chan et al. (11 ) described the circadian profiles of ghrelin in healthy lean persons who had fasted for 72 h (n ⫽ 6) and who had blood samples taken every 15 min for 24 h; the authors found minimal variation in ghrelin during the day before concentrations increased in the evening, followed by a significant decrease between 0200 and 0400. More recently, Espelund et al. (12 ) studied 33 healthy young adults (17 lean and 16 obese) with blood sampling every 3 h from 12 to 84 h of fasting. In contrast to the data from Chan et al. (11 ), Espelund et al. (12 ) found that serum ghrelin showed a marked diurnal rhythm with a nadir in the morning (0800), maximum values in the afternoon, and a gradual decrease during the night. This pattern was preserved during the entire fasting period and was independent of sex and obesity. Considering this background, the ghrelin kinetics observed in our clinical setting involving a combination of

fasting and uncomplicated surgical injury were quite different from any of the above-mentioned ghrelin patterns observed in response to the metabolic stress of fasting by itself in healthy individuals. Patients who undergo elective cholecystectomy do not appear to follow the circadian meal-independent rhythm of ghrelin secretion. The ghrelin response to surgical injury may be interpreted as an appropriate adaptation to an acute stress to help maintain homeostasis (15 ). Indeed, adaptation to starvation or caloric restriction may be the primary physiologic need for which ghrelin evolved as a peripheral regulator of energy balance (16, 17 ). On the other hand, it may be an adaptive response to inflammation. Consistent with the latter possibility are the recent findings by Dixit et al. (18 ) in a murine model of lipopolysaccharide (LPS)-induced endotoxemia, a well-recognized model associated with anorexia resulting from excessive production of proinflammatory mediators and a refractory catabolic state. The authors demonstrated that ghrelin infusion in LPS-challenged mice leads to a significant inhibition of the proinflammatory cytokines in circulation as well as in various organs. They also demonstrated that LPS-induced inflammatory anorexia is also significantly reduced in ghrelin-treated mice. Their data complement those reported previously by Basa et al. (19 ) and Hataya et al. (20 ) in rats. Whereas Basa et al. (19 ) showed that LPS-induced endotoxemia leads to inhibition of ghrelin secretion, Hataya et al. (20 ) found that ghrelin infusion increases body weight in septic animals. Considering these data, it seems plausible that inhibition of ghrelin secretion after LPS challenge might exacerbate the ongoing inflammatory insult and promote the development of a catabolic state. Thus, the decrease in ghrelin that we have found in the present study may be associated with the transition from an inflammatory immune response to an adaptive immune response, and future studies of the prognostic role of ghrelin after elective and nonelective surgery, with and without a complicated postoperative course, are warranted.

We thank Gino Davı` for excellent technical assistance. References 1. Date Y, Kojima M, Hosoda H, Sawaguchi A, Mondal MS, Suganuma T, et al. Ghrelin, a novel growth hormone-releasing acylated peptide, is synthesized in a distinct endocrine cell type in the gastrointestinal tracts of rats and humans. Endocrinology 2000;141:4255– 61. 2. Meier U, Gressner AM. Endocrine regulation of energy metabolism: review of pathobiochemical and clinical chemical aspects of leptin, ghrelin, adiponectin, and resistin. Clin Chem 2004;50:1511–25. 3. Kojima M, Hosoda H, Date Y, Nakazato M, Matsuo H, Kangawa K. Ghrelin is a growth-hormone-releasing acylated peptide from stomach. Nature 1999; 402:656 – 60. 4. Wu JT, Kral JG. Ghrelin: integrative neuroendocrine peptide in health and disease. Ann Surg 2004;239:464 –74. 5. Gnanapavan S, Kola B, Bustin SA, Morris DG, McGee P, Fairclough P, et al. The tissue distribution of the mRNA of ghrelin and subtypes of its receptor, GHS-R, in humans. J Clin Endocrinol Metab 2002;87:2988 –91. 6. Giesecke K, Hamberger B, Jarnberg PO, Klingstedt C, Persson B. High- and low-dose fentanyl anaesthesia: hormonal and metabolic responses during cholecystectomy. Br J Anaesth 1988;61:575– 82. 7. Ravussin E, Tscho¨p M, Morales S, Bouchard C, Heiman ML. Plasma ghrelin

Clinical Chemistry 51, No. 7, 2005

8.

9.

10.

11.

12.

13.

14.

15. 16.

17. 18.

19.

20.

concentration and energy balance: overfeeding and negative energy balance studies in twins. J Clin Endocrinol Metab 2001;86:4547–51. Tscho¨p M, Wawarta R, Riepl RL, Friedrich S, Bidlingmaier M, Landgraf R, et al. Post-prandial decrease of circulating human ghrelin levels. J Endocrinol Invest 2001;24:RC19 –21. Tscho¨p M, Weyer C, Tataranni PA, Devanarayan V, Ravussin E, Heiman ML. Circulating ghrelin levels are decreased in human obesity. Diabetes 2001; 50:707–9. Otto B, Cuntz U, Fruehauf E, Wawarta R, Folwaczny C, Riepl RL, et al. Weight gain decreases elevated plasma ghrelin concentrations of patients with anorexia nervosa. Eur J Endocrinol 2001;145:669 –73. Chan JL, Bullen J, Lee JH, Yiannakouris N, Mantzoros CS. Ghrelin levels are not regulated by recombinant leptin administration and/or three days of fasting in healthy subjects. J Clin Endocrinol Metab 2004;89:335– 43. Espelund U, Hansen TK, Hojlund K, Beck-Nielsen H, Clausen JT, Hansen BS, et al. Fasting unmasks a strong inverse association between ghrelin and cortisol in serum: studies in obese and normal-weight subjects. J Clin Endocrinol Metab 2005;90:741– 6. Norrelund H, Hansen TK, Orskov H, Hosoda H, Kojima M, Kangawa K, et al. Ghrelin immunoreactivity in human plasma is suppressed by somatostatin. Clin Endocrinol (Oxf) 2002;57:539 – 46. Muller AF, Lamberts SW, Janssen JA, Hofland LJ, Koetsveld PV, Bidlingmaier M, et al. Ghrelin drives GH secretion during fasting in man. Eur J Endocrinol 2002;146:203–7. Stryjewski G, Dalton HJ. Circulating leptin: mediator or marker of the neuroendocrinological stress response? Crit Care Med 2001;29:2397– 8. Kolaczynski JW, Considine RV, Ohannesian J, Marco C, Opentanova I, Nyce MR, et al. Responses of leptin to short-term fasting and refeeding in humans: a link with ketogenesis but not ketones themselves. Diabetes 1996;45:1511–5. Schwartz MW, Seeley RJ. Neuroendocrine responses to starvation and weight loss. N Engl J Med 1997;336:1802–11. Dixit VD, Schaffer EM, Pyle RS, Collins GD, Sakthivel SK, Palaniappan R, et al. Ghrelin inhibits leptin- and activation-induced proinflammatory cytokine expression by human monocytes and T cells. J Clin Invest 2004;114:57– 66. Basa NR, Wang L, Arteaga JR, Heber D, Livingston EH, Tache Y. Bacterial lipopolysaccharide shifts fasted plasma ghrelin to postprandial levels in rats. Neurosci Lett 2003;343:25– 8. Hataya Y, Akamizu T, Hosoda H, Kanamoto N, Moriyama K, Kangawa K, et al. Alterations of plasma ghrelin levels in rats with lipopolysaccharideinduced wasting syndrome and effects of ghrelin treatment on the syndrome. Endocrinology 2003;144:5365–71. DOI: 10.1373/clinchem.2005.050104

Clinical Applications of Plasma Circulating mRNA Analysis in Cases of Gestational Trophoblastic Disease, Hideaki Masuzaki,1† Kiyonori Miura,1†* Kentaro Yamasaki,1 Shoko Miura,1 Koh-ichiro Yoshiura,2 Shuichiro Yoshimura,1 Daisuke Nakayama,1 Christophe K. Mapendano,2 Norio Niikawa,2,3 and Tadayuki Ishimaru1 (Departments of 1 Obstetrics and Gynecology, and 2 Human Genetics, Nagasaki University Graduate School of Biomedical Sciences, Nagasaki, Japan; 3 CREST, Japan Science and Technology Agency (JST), Kawaguchi, Japan; † these authors contributed equally to this work; * address correspondence to this author at: Department of Obstetrics and Gynecology, Nagasaki University Graduate School of Biomedical Sciences, 1-7-1 Sakamoto, Nagasaki 852-8501, Japan; fax 81-95-849-7365, e-mail [email protected]) The finding of tumor-related mRNA (tyrosinase mRNA, telomerase component mRNA, various tumor-derived mRNAs, and viral RNAs) in plasma from a patient with cancer suggests that real-time quantitative reverse transcription-PCR (RT-PCR) study of circulating plasma

1261

mRNAs may be used as a very simple assay for residual tumor after surgery or therapy (1–3 ). Gestational trophoblastic diseases (GTDs) are a heterogeneous group of diseases that include partial and complete hydatidiform moles (CHMs), invasive moles, choriocarcinomas, placental site trophoblastic tumors, and epithelioid trophoblastic tumors. Human chorionic gonadotropin (hCG) is generally used as a tumor marker in the clinical management of all GTDs. Recently, hCG-␤ mRNA from placenta has been detected in maternal plasma by real-time quantitative RT-PCR (4 ), suggesting that this assay may be useful in the management of GTDs. In this study, we measured plasma mRNA concentrations by real-time RT-PCR in 3 GTD patients with CHMs and 1 with choriocarcinoma. Blood samples (10 mL) were collected before and after suction evacuation or chemotherapy. Plasma mRNA was extracted, and a 1-step real-time RT-PCR assay was performed as described by Ng et al. (4 ). We selected hCG-␤ mRNA as a tumor marker of GTD and glyceraldehyde-3phosphate dehydrogenase (GAPDH) mRNA as a housekeeping gene. Primer sets and TaqMan probes for the 2 genes selected were prepared as described previously (4 ). We prepared calibration curves for quantification of each mRNA by assaying serial dilutions of HPLC-purified single-strand synthetic DNA oligonucleotides from each PCR amplicon (R2 ⫽ 0.99; slope, ⫺2.9 to 3.3). The absolute concentration of each mRNA is reported as copies/mL of maternal plasma, based on the formula described by Farina et al. (5 ). Interpolation of obtained results on the calibration curves indicated that hCG-␤ mRNA concentrations ranged from 1 ⫻ 107 to 1 ⫻ 101 copies/mL, and GAPDH mRNA concentrations ranged from 1⫻ 1010 to 1 ⫻ 104 copies/mL. Each sample was analyzed in triplicate under thermal conditions as described previously (4 ). All study protocols were approved by the Committee for the Ethical Issues on Human Genome and Gene Analysis in Nagasaki University, and written informed consent was obtained from all of the women. None of the data regarding plasma mRNA concentrations influenced the clinical management of patients. Three cases with CHM, diagnosed by pathology examinations, were treated twice with suction evacuation. Chest x-rays indicated no pulmonary lesion. The day of the first evacuation was called day 0, and the second evacuation was on day 7. The plasma hCG-␤ mRNA in all 3 cases decreased rapidly to undetectable (⬍10 copies/ mL) by day 7 (Fig. 1A), and the corresponding hCG protein in serum, as measured by an IRMA, also decreased to ⬍1000 IU/L until 35 days after the first evacuation (Fig. 1B) (6 ). A patient with clinical choriocarcinoma, which was diagnosed by the diagnostic score used in Japan (6 ), showed a ⬍3 cm pulmonary lesion detected by chest x-ray and a biphasic pattern of basal body temperature (BBT). Etoposide and actinomycin D therapy was given on days 0 –3, 15–18, 29 –32, and 43– 46. Her plasma hCG-␤ mRNA concentration showed a decreasing tendency similar to the pattern for hCG protein in serum, but exhibited

1262

Technical Briefs

Fig. 1. Changing concentrations of hCG-␤ and GAPDH mRNA in plasma of cases with GTD. Plasma mRNA concentrations are in copies/mL, and the serum hCG protein concentration is in IU/L. (A), plasma hCG-␤ mRNA in cases with CHM. (B), serum hCG protein (measured by IRMA) in cases with CHM. (C), plasma GAPDH mRNA in cases with CHM. f, F, and Œ in panels A, B, and C indicate cases 1, 2, and 3, respectively. Arrows indicate days 0 and 7, the days on which suction evacuations were performed. (D), plasma hCG-␤ mRNA (f) and serum hCG protein (Œ; measured by IRMA) in a case with choriocarcinoma. (E), plasma hCG-␤ (f) and GAPDH mRNA (Œ) in a case with choriocarcinoma. EA, etoposide and actinomycin D therapy.

Clinical Chemistry 51, No. 7, 2005

Table 1. Changes in plasma hCG-␤ and GAPDH mRNA concentrations with therapy in a case with choriocarcinoma.a Day

hCG-␤ mRNA, copies/mL

0 3 10 17 24 31 38 45 52 59 66

1440 901 954 44 301 45 166 26 23 96 17

GAPDH mRNA, copies/mL

1.51 2.63 8.65 2.59 4.00 3.60 8.73 5.74 5.72 5.18 9.38

⫻ ⫻ ⫻ ⫻ ⫻ ⫻ ⫻ ⫻ ⫻ ⫻ ⫻

106 107 106 107 106 107 106 106 107 106 106

a Therapy with etoposide and actinomycin D was given on days 0 –3, 15–18, 29 –32, and 43– 46.

a transient increase from 901 to 954 copies/mL after the first course of therapy and then decreased to 44 copies/mL (Fig. 1D). The same transient tendency was detected after all subsequent courses of therapy (Table 1). The cases with CHM had no metastatic lesions, whereas the case with choriocarcinoma had lung metastasis. Rapid clearance of plasma hCG-␤ mRNA may reflect complete elimination of CHM, and the transiently increased concentration of plasma hCG-␤ mRNA after every each course of therapy may reflect apoptotic activity induced by the therapy (3, 7 ). The concentration of plasma GAPDH mRNA in all 3 cases with CHM also increased after both evacuations (Fig. 1C), and that in a case with choriocarcinoma also increased after therapy with etoposide and actinomycin D (Fig. 1E and Table 1). Thus, cell/tissue damage by both evacuations and chemotherapy may be associated with increased circulating concentrations of GAPDH mRNA (3, 7 ). Our results suggest that measurement of plasma mRNA by real-time quantitative RT-PCR can be used as a noninvasive diagnostic, prognostic, and follow-up test for GTD (1–3 ). The mRNA is cleared rapidly and provides information different from that obtained with hCG immunoassays. The measurements may be particularly useful in patients whose serum contains substances (such as human anti-mouse antibodies) that interfere in immunoassays. As our data were based on only a few cases, it remains to be seen in additional studies whether the method is sensitive enough to monitor changes in mRNA concentrations in GTD.

We thank Dr. Joseph Wagstaff for assistance and valuable advice. This work was supported in part by Grants-in-Aid for Scientific Research (15591761, 16591670, and 13854024) from the Ministry of Education, Sports, Culture, Science and Technology of Japan (to H.M., K.M., and N.N.) and CREST from the Japan Science and Technology Agency (to N.N.)

1263

References 1. Dasi F, Lledo S, Garcia-Granero E, Ripoll R, Marugan M, Tormo M, et al. Real-time quantification in plasma of human telomerase reverse transcriptase (hTERT) mRNA: a simple blood test to monitor disease in cancer patients. Lab Invest 2001;81:767–9. 2. Silva JM, Dominguez G, Silva J, Garcia JM, Sanchez A, Rodriguez O, et al. Detection of epithelial messenger RNA in the plasma of breast cancer patients is associated with poor prognosis tumor characteristics. Clin Cancer Res 2001;7:2821–5. 3. Ng EK, Tsui NB, Lam NY, Chiu RW, Yu SC, Wong SC, et al. Presence of filterable and nonfilterable mRNA in the plasma of cancer patients and healthy individuals. Clin Chem 2002;48:1212–7. 4. Ng EK, Tsui NB, Lau TK, Leung TN, Chiu RW, Panesar NS, et al. mRNA of placental origin is readily detectable in maternal plasma. Proc Natl Acad Sci U S A 2003;100:4748 –53. 5. Farina A, Chan CW, Chiu RW, Tsui NB, Carinci P, Concu M, et al. Circulating corticotropin-releasing hormone mRNA in maternal plasma: relationship with gestational age and severity of preeclampsia. Clin Chem 2004;50:1851– 4. 6. Sasaki S. Management of gestational trophoblastic diseases in Japan—a review. Placenta 2003;24:S28 –32. 7. Wataganara T, LeShane ES, Chen AY, Borgatta L, Peter I, Johnson KL, et al. Plasma ␥-globin gene expression suggests that fetal hematopoietic cells contribute to the pool of circulating cell-free fetal nucleic acids during pregnancy. Clin Chem 2004;50:689 –93. Previously published online at DOI: 10.1373/clinchem.2005.050666

Double-Gradient–Denaturing-Gradient Gel Electrophoresis for Mutation Screening of the BCR-ABL Tyrosine Kinase Domain in Chronic Myeloid Leukemia Patients, Nathalie Sorel, Florence Chazelas, Andre´ Brizard,‡ and Jean-Claude Chomel* (Laboratoire d’He´matologie et EA 3805, CHU de Poitiers, France; ‡ deceased; * address correspondence to this author at: Laboratoire d’He´matologie, CHU de Poitiers, 2 rue de la Mile´trie, 86021 Poitiers Cedex, France; fax 335-49444095, e-mail [email protected]) Chronic myeloid leukemia (CML) is a rare hematopoietic stem cell disorder characterized by the t(9;22) translocation. This somatic event leads to the formation of the BCR-ABL fusion gene, which is translated in a functional protein (1 ). The Bcr-Abl oncoprotein differs from the endogenous c-Abl protein in both its subcellular localization and its tyrosine kinase (TK) activity. The deregulated Abl TK activity of the Bcr-Abl protein initiates the oncogenic process and represents a target to TK inhibitors. Among a series of compounds exhibiting TK inhibition, STI 571 (imatinib) was found to be highly effective against Abl and its derivative Bcr-Abl (2, 3 ). This molecule targets the inactive conformation of the kinase, which leads to stabilization of the protein in its inactive form and impairs ATP binding (4 ). Approved for the treatment of CML, imatinib induces hematologic and cytogenetic remissions in the chronic phase as well as in the blast phase (5–7 ). A significant proportion of patients, however, become resistant to the treatment. Different mechanisms of imatinib resistance have been described, in particular point mutations within the sequence of the BCR-ABL gene coding for the TK domain (8 –10 ). Approximately 30 missense mutations have been identified in CML patients. Within the Bcr-Abl TK domain, these mutations are

1264

Technical Briefs

located in the nucleotide binding loop (P-loop), in the active site (imatinib contact site), or in the activation loop (A-loop; Fig. 1A). Among these mutations, some prevent the binding of imatinib, and others make this binding more difficult. In the latter case, a dose increase can overcome the resistance. Recently, a novel TK inhibitor

A

Tyrosine kinase domain

Bcr

Abl

M244V L248V G250E G250R Q252E Q252H Q252R Y253H Y253F E255K E255V D276G T277A

V289A F311L T315I T315N F317L M343T M351T E355G F359V F359A

V379I F382L L387M L387F H396P H396R A397P

P-loop

Active site

A-loop

RT-PCR 5’

RT-PCR 3’ RT-PCR 5’

B h

C

ix M

Y2 53 H

h

H3 96 P

V3 79 I

17 L F3

T3 15 I

M

W T

H 35 1T

RT-PCR 3’

D

W T E2 5 Y2 5V 53 H

E2 55 K

Q2 52 H

Q2 52 E

G2 50 E

W T

M

24 4V

H

M244V mutation (RT-PCR 5’) E255K mutation (RT-PCR 5’) T315I mutation (RT-PCR 3’) [50%] [20%] [10%] [5%]

[2%] [1%]

Fig. 1. Analysis of the BCR-ABL TK domain: results of single-step DG-DGGE experiments. (A), schematic representation of the 2 amplicons within the TK domain and localization of known mutations. The 2 black lines above RT-PCR 5⬘ and RT-PCR 3⬘ represent the regions of amplicons in which all mutations can theoretically be detected. (B), results of DG-DGGE screening of the 5⬘ fragment: WT, wild-type control; M244V, G250E, Q252E, Q252H, and E255K, mutants; E255V:Y253H, double mutant; Mix, mixture of E255V and Y253H PCR products in equal amounts; Y253H, mutant. H and h, homo- and heteroduplex bands, respectively. The mixture pattern includes, starting from the band of lower molecular weight, 1 Y253H mutant homoduplex, 1 E255V mutant homoduplex, 1 WT homoduplex, 2 WT/Y253H heteroduplexes, 2 nonseparated WT/E255V heteroduplexes, and 2 Y253H/E255V heteroduplexes. (C), results of DG-DGGE screening of the 3⬘ fragment: WT, wild-type control; M351T, T315I, F317L, V379I, and H396P, mutants. (D), results of DG-DGGE analysis of serial dilutions of M244V (top gel), E255K (middle gel), or T315I (bottom gel) mutated cDNA (from 50% to 1%) in WT cDNA. The arrows indicate the detection thresholds of the method.

was described (11 ). In vitro, this molecule remains active for the majority of Bcr-Abl mutants, with the exception of T315I. Molecular monitoring of CML patients is based on the quantification of BCR-ABL transcripts in peripheral blood (or bone marrow) samples by real-time reverse transcription-PCR at regular intervals. It has recently been proposed that during imatinib therapy, a more than 2-fold increase in the BCR-ABL transcript concentration is a primary indicator for an emerging mutation within the Bcr-Abl kinase domain (12 ). Detection of mutations in the ABL TK domain of the BCR-ABL oncogene by several methods has been reported. Some of these focus on known mutations (e.g., allele-specific oligonucleotide PCR and PCR with restriction fragment length polymorphism analysis). Others, such as denaturing HPLC (13, 14 ) can identify unknown single-base mutations in a given sequence. We describe a mutation screening method based on double-gradient– denaturing-gradient gel electrophoresis (DG-DGGE) (15, 16 ), which relies on wild-type (WT)/mutant heteroduplex formation. The entire BCR-ABL TK domain covers 4 exons of the ABL gene; therefore, the analysis was performed at the RNA level by means of 2 overlapping PCR fragments (5⬘ and 3⬘ in Fig. 1A). DG-DGGE is generally performed on the ABL sequence by a single-step PCR. In this case, both BCR-ABL and ABL tyrosine kinase domains were amplified. Nested PCR was used when the analysis was restricted to BCR-ABL. Total RNA was extracted from blood or bone marrow samples with RNABle reagent (Eurobio) according to the manufacturer’s instructions. RNA pellets were resuspended in 40 ␮L of RNase-free water. cDNA was synthesized from 1 ␮g of total RNA in a 20-␮L reaction mixture by use of Moloney murine leukemia virus reverse transcriptase (Invitrogen). For a nested PCR analysis, 2 ␮L of cDNA was subjected to a first round of PCR (25 cycles) with the forward primer located in exon 13 of the BCR gene (5⬘-ACAGCATTCCGCTGACCATC-3⬘) and the reverse primer located in exon 8 of the ABL gene (5⬘-GAACGGTCAATTCCCGGG3⬘) under the following conditions: denaturation at 95 °C for 15 min, followed by 25 cycles of denaturation at 94 °C for 30 s, annealing at 60 °C for 30 s, and extension at 72 °C for 90 s. Amplifications were carried out in 50 ␮L with 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 2 mM MgCl2, 200 ␮M each deoxynucleotide triphosphate, 0.5 ␮M each primer, and 2.5 U of Sure Prime DNA polymerase (Q-Biogene). An aliquot of the first PCR product (5 ␮L of a 1:100 dilution) or 5 ␮L of cDNA (in the case of a single-step PCR) was used for the ABL TK amplification. For this purpose, 2 pairs of primers, 5F (5⬘-CTGTCTATGGTGTGTCCCCC-3⬘) and 5R [5⬘-(40GC)-GCAGTTTCGGGCAGCAAGAT-3⬘] and 3F (5⬘-AAAGAGATCAAACACCCTAACC-3⬘) and 3R [5⬘-(40GC)-TCCCAAAGCAATACTCCAAATG-3⬘] were designed using the MELT94 program (http://web.mit.edu/osp/www/melt.html). A 40-base GC-rich fragment (40GC; 5⬘-GCCCGCCGTCCCGGCCCGACCCCCGCGCGTCCGGCGCCCG-3⬘) was added to the 5⬘ end of each

Clinical Chemistry 51, No. 7, 2005

reverse primer (5R and 3R). The 2 amplicons, of 437 and 422 bp, respectively, overlapped on more than 200 bp. However, MELT programs predict the detection of any mutations in a sequence corresponding to amino acids 236 – 417 with little redundancy between the 2 amplicons. PCR was performed with the following conditions: denaturation at 95 °C for 15 min, followed by 40 cycles of denaturation at 94 °C for 30 s, annealing at 62 °C for 30 s, and extension at 72 °C for 30 s. PCR was terminated by a final extension at 72 °C for 7 min. Amplifications were carried out as described above. The heteroduplexes were generated during the last cycles of PCR. DG-DGGE analysis was performed on a DGGE-2001 system (CBS Scientific). An aliquot (15 ␮L) of PCR sample was electrophoresed in a colinearly increasing double gradient of 50% to 95% denaturant (100% denaturant: 7 mol/L urea, 400 mL/L formamide) and 6.5% to 12% polyacrylamide (acrylamide/bisacrylamide, 37.5/1). Electrophoresis was performed in 1⫻ Tris-acetate-EDTA buffer (40 mmol/L Tris-acetate, 20 mmol/L sodium acetate, 1 mmol/L EDTA, pH 8.0) at 1800 V-h (e.g., 15 h at 120 V) and at a constant temperature of 60 °C. Gels were stained with ethidium bromide and visualized through ultraviolet transillumination. In the presence of a shift in the acrylamide gel, the nucleotide change was characterized by direct sequencing of the PCR product carried out in both directions, using the BigDye Terminator (Ver. 1.1) Cycle Sequencing Kit and the ABI PRISM 310 sequencer (Applied Biosystems) according to the manufacturer’s recommendations. To avoid false-negative results, PCR products that showed no mobility shift in DG-DGGE were mixed with an equal amount of WT amplicon, denatured (5 min at 95 °C), annealed to form heteroduplexes (15 min at 56 °C), and analyzed again. To evaluate the efficiency of the DG-DGGE methods, we analyzed cDNA samples from patients carrying mutations in the BCR-ABL TK domain that had already been characterized by direct sequencing. Compared with the WT controls, all samples showed an abnormal DGGE pattern (Fig. 1, B and C). In the presence of a mutation, up to 4 bands were visualized: the WT homoduplex, the mutant homoduplex, and the 2 heteroduplexes (M244V, G250E, E255K, Y253H, M351T, T315I, F317L, and V379I). In these cases, the difference in the intensities of the homoduplex bands reflected the ratio between mutated and WT sequences. In a few samples (Q252E, Q252H, and H396P), the 2 homoduplexes were not separated. A double mutant with a dominant (E255V) and a minority mutation (Y253H) was detected by DG-DGGE. The E255V mutation, but not the minority mutation, was identified by direct sequencing of the corresponding PCR product. The 2 faint bands of high molecular weight were excised from the gel, purified, and amplified for direct sequencing. The 2 mutations (E255V and Y253H) were then characterized. To confirm these results, amplified products from E255V and Y253H mutants were mixed and subjected to DG-DGGE. Three homoduplexes (Y253H, E255V, and WT) and 5 bands corresponding to the WT/

1265

Y253H, WT/E255V, and Y253H/E255V heteroduplexes were clearly distinguishable in this assay. Thus, all previously identified mutations were unambiguously detected within the entire ABL TK domain. Different point mutations leading to amino acid substitution were detected in 26 of 59 imatinib-resistant CML patients (M351T in 6; M244V and T315I in 4; G250E in 3; Q252E in 2; and Q252H, Y253H, E255K, E255V, F317L, V379I, and H396P in 1 each). To define the detection threshold of the method, we serially diluted cDNAs carrying the M244V, E255K, or T315I mutations to concentrations ranging from 50% to 1% in a WT cDNA (Fig. 1D). The heteroduplex bands remained clearly detectable until the 5% (M244V and T315I) or the 2% dilution (E255K). Mutated sequences can thus be detected when they account for at least 5% of total cDNA. When correctly designed and applied, the percentage of point mutations detectable by DGGE is theoretically 100% in a given sequence. The method described here appears to be robust, inexpensive, and sensitive enough to detect a small proportion of mutant with an excess of WT sequences. Moreover, the detection of minority mutations is possible. Screening of the entire TK domain of the BCR-ABL gene enables the classification of imatinib-resistant CML patients into 3 groups based on the amino acid change. The first group consists of patients for whom an imatinib dose increase can overcome the resistance (e.g., M244V and F359V mutants). The second group includes patients with a permanent resistance to imatinib (e.g., G250E, E255V, and H396P mutants); they could benefit from secondgeneration TK inhibitors. Finally, CML patients carrying the T315I mutation (found in 15%–20% of imatinib-resistant cases) seem to show a universal resistance toward all known TK inhibitors. Alternative treatments are currently in evaluation for patients carrying this mutation. Thus, a screening procedure such as DG-DGGE can be used to ensure the best therapeutic strategy as targeted therapies are developed for diseases such as CML.

This work was supported by grants from the Ligue Contre le Cancer (Comite´ de la Vienne et de la Charente). References 1. Shtivelman E, Lifshitz B, Gale RP, Canaani E. Fused transcript of abl and bcr genes in chronic myelogenous leukaemia. Nature 1985;315:550 – 4. 2. Buchdunger E, Zimmermann J, Mett H, Meyer T, Muller M, Druker BJ, et al. Inhibition of the Abl protein-tyrosine kinase in vitro and in vivo by a 2-phenylaminopyrimidine derivative. Cancer Res 1996;56:100 – 4. 3. Druker BJ, Lydon NB. Lessons learned from the development of an abl tyrosine kinase inhibitor for chronic myelogenous leukemia [Review]. J Clin Invest 2000;105:3–7. 4. Dorey K, Engen JR, Kretzschmar J, Wilm M, Neubauer G, Schindler T, et al. Phosphorylation and structure-based functional studies reveal a positive and a negative role for the activation loop of the c-Abl tyrosine kinase. Oncogene 2001;20:8075– 84. 5. Kantarjian H, Sawyers C, Hochhaus A, Guilhot F, Schiffer C, GambacortiPasserini C, et al. Hematologic and cytogenetic responses to imatinib mesylate in chronic myelogenous leukemia. N Engl J Med 2002;346:645– 52. 6. Talpaz M, Silver RT, Druker BJ, Goldman JM, Gambacorti-Passerini C, Guilhot F, et al. Imatinib induces durable hematologic and cytogenetic responses in

1266

7.

8.

9.

10.

11. 12.

13.

14.

15.

16.

Technical Briefs

patients with accelerated phase chronic myeloid leukemia: results of a phase 2 study. Blood 2002;99:1928 –37. Sawyers CL, Hochhaus A, Feldman E, Goldman JM, Miller CB, Ottmann OG, et al. Imatinib induces hematologic and cytogenetic responses in patients with chronic myelogenous leukemia in myeloid blast crisis: results of a phase II study. Blood 2002;99:3530 –9. Shah NP, Nicoll JM, Nagar B, Gorre ME, Paquette RL, Kuriyan J, et al. Multiple BCR-ABL kinase domain mutations confer polyclonal resistance to the tyrosine kinase inhibitor imatinib (STI571) in chronic phase and blast crisis chronic myeloid leukemia. Cancer Cell 2002;2:117–25. von Bubnoff N, Peschel C, Duyster J. Resistance of Philadelphia-chromosome positive leukemia towards the kinase inhibitor imatinib (STI571, Glivec): a targeted oncoprotein strikes back [Review]. Leukemia 2003;17: 829 –38. Nardi V, Azam M, Daley GQ. Mechanisms and implications of imatinib resistance mutations in BCR-ABL [Review]. Curr Opin Hematol 2004;11:35– 43. Shah NP, Tran C, Lee FY, Chen P, Norris D, Sawyers CL. Overriding imatinib resistance with a novel ABL kinase inhibitor. Science 2004;305:399 – 401. Branford S, Rudzki Z, Parkinson I, Grigg A, Taylor K, Seymour JF, et al. Real-time quantitative PCR analysis can be used as a primary screen to identify patients with CML treated with imatinib who have BCR-ABL kinase domain mutations. Blood 2004;104:2926 –32. Deininger MW, McGreevey L, Willis S, Bainbridge TM, Druker BJ, Heinrich MC. Detection of ABL kinase domain mutations with denaturing highperformance liquid chromatography. Leukemia 2004;18:864 –71. Soverini S, Martinelli G, Amabile M, Poerio A, Bianchini M, Rosti G, et al. Denaturing-HPLC-based assay for detection of ABL mutations in chronic myeloid leukemia patients resistant to Imatinib. Clin Chem 2004;50:1205– 13. Cremonesi L, Firpo S, Ferrari M, Righetti PG, Gelfi C. Double-gradient DGGE for optimized detection of DNA point mutations. Biotechniques 1997;22: 326 –30. Cremonesi L, Carrera P, Fumagalli A, Lucchiari S, Cardillo E, Ferrari M, et al. Validation of double gradient denaturing gradient gel electrophoresis through multigenic retrospective analysis. Clin Chem 1999;45:35– 40. DOI: 10.1373/clinchem.2004.047274

Relationship between Serum Folate and Plasma Nitrate Concentrations: Possible Clinical Implications, Mohammad A. Mansoor,1,2* Ole Kristensen,1 Tor Hervig,3 Jacob A. Stakkestad,4 Thor Berge,5 Per A. Drabløs,6 Svanhild Rolfsen,5 and Tore Wentzel-Larsen7 (1 Division of Medical Biochemistry, Stavanger University Hospital, Stavanger, Norway; 2 Agder University College, Department of Natural Sciences, Kristiansand, Norway; 3 Blood Bank and 7 Center for Clinical Research, Haukeland University Hospital, Bergen, Norway; 4 Cecor AS, Haugesund, Norway; 5 Elf Petroleum Stavanger AS, Stavanger, Norway; 6 Norsk Hydro Aluminum, Karmøy, Norway; * address correspondence to this author at: Agder University College, Department of Natural Sciences, N-4604 Kristiansand, Norway; e-mail [email protected]) Endothelial-cell– derived nitric oxide (NO) relaxes vascular smooth muscle cells, causes vasodilation, and inhibits platelet aggregation. It has been demonstrated that NO3 is the predominant and stable form of NO present in plasma and represents the bioavailability of NO (1, 2 ). NO generation is depressed in blood vessels affected by atherosclerosis, and a loss of NO activity is associated with impaired vasoreactivity, enhanced platelet aggregation, and increased endothelial cell–leukocyte interactions (3 ). Recently, evidence has accumulated that suggests that

increased concentrations of the amino acid homocysteine (Hcy) also increase the risk for cardiovascular disease (CVD) (4 ). Hcy concentrations can be lowered by supplementation with folate, vitamin B6, vitamin B2, and vitamin B12 (5, 6 ). In particular, folate may also contribute in the prevention of CVD because folate seems to restore impaired NO metabolism (7 ). In the present investigation, we included apparently healthy persons. Individuals on all types of medications were excluded from this cross-sectional study. Blood samples were collected from participants who had been fasting for 12 h. The blood samples were centrifuged at 4 °C within 1 h and stored at ⫺80 °C until analysis. Informed written consent was obtained from all participants. The study was approved by the Regional Ethics committee, University of Bergen, Norway. Total homocysteine (tHcy) concentrations in plasma were measured by an HPLC method (8 ). Plasma NO3 concentrations were measured by capillary electrophoresis (9 ), and total serum cholesterol, HDL-cholesterol, and serum triglyceride concentrations were measured by a routine enzymatic colorimetric method (Roche/Hitachi). The concentrations of erythrocyte folate, serum folate, and vitamin B12 were measured on an Access Immuno Assay System (Sanofi Pasteur Diagnostics). The program “Power and Precision” was used for the power calculations. Power computations were performed for the comparison of smokers vs nonsmokers with serum folate as the primary endpoint, based on a 2-sample t-test for ln-transformed data. All P values given are 2-tailed. The statistical program StatView for the Macintosh (Abacus Concepts) was used for all calculations. The characteristics of the study participants are given in Table 1. In a simple regression analysis, we found a significant positive relationship between the concentraTable 1. Characteristics and biochemical values in plasma/ serum from the study participants.a Characteristic/Analyte

n M/F, n Age, years BMI,b kg/m2 Smokers/nonsmokers, n Cigarettes smoked per day, n Smoking, number of years Plasma nitrate, ␮mol/L (n ⫽ 205) Serum folate, nmol/L (n ⫽ 200) Erythrocyte folate, nmol/L (n ⫽ 148) Hcy, ␮mol/L (n ⫽ 205) Vitamin B12, pmol/L (n ⫽ 204) Total cholesterol, mg/L (n ⫽ 205) HDL, mg/L (n ⫽ 109) LDL, mg/L (n ⫽ 109) Triglycerides, mg/L (n ⫽ 203) a

205 156/49 41 (26–60) 24.9 (17.2–34.3) 112/93 20 (5–40) 22.5 (7.0–43.0) 25.4 (12.3–135.5) 11.0 (4.0–39.0) 685 (230–1300) 10.8 (6.0–29.2) 350 (146–924) 5.8 (3.0–13.7) 1.4 (0.6–2.3) 3.8 (2.2–6.9) 1.2 (0.5–6.9)

All data except for the numbers of males/females and the numbers of smokers/nonsmokers are presented as the median (range). b BMI, body mass index.

Clinical Chemistry 51, No. 7, 2005

1267

Fig. 1. Proposed metabolic relationships between folate and NO3 and folate and Hcy (1, 5 ). Enzymes: 1, methionine adenosyltransferase; 2, methyltransferase; 3, S-adenosylhomocysteine hydrolase; 4, betaine-homocysteine methyltransferase; 5, methionine synthase; 6, dihydrofolate reductase; 7, serine hydroxymethyltransferase; 8, 5,10-methylenetetrahydrofolate reductase; 9, dihydropterin reductase; 10, NOS. Abbreviations: SAM, S-adenosylmethionine; SAH, S-adenosylhomocysteine; DHF, dihydrofolate; THF, tetrahydrofolate; qBH2, quininoid 7,8-dihydrobiopterin; BH2, 7,8-dihydrobiopterin; Hcy-S-NO, homocysteine-nitrosothiol; ONOO⫺, peroxynitrate; Hcy-SH, reduced Hcy; Hcy-S-S-, oxidized Hcy; Hcy-S-S-P, protein-bound Hcy; Hcy-Tlc, Hcy thiolactone.

tions of serum folate and plasma NO3 (R ⫽ 0.17; P ⫽ 0.017) and a significant inverse relationship between the concentrations of serum folate and tHcy (R ⫽ 0.244; P ⫽ 0.0005). In a multiple regression analysis that included age, sex, body mass index, NO3, tHcy, and vitamin B12 as independent variables and serum folate as dependent variable, we found a significant positive relationship between the concentrations of serum folate and plasma NO3 (P ⫽ 0.0037), and a significant negative relationship between the concentrations of serum folate and tHcy (P ⫽ 0.0005; Table 1B of the Data Supplement that accompanies the online version of this Technical Brief at http:// www.clinchem.org/content/vol51/issue7/). The present findings may indicate that folate is involved in the intracellular formation and release of NO into plasma. There are at least 2 mechanisms by which NO could be converted to NO3. The first possibility is the direct and rapid oxidative conversion of NO to NO3 in the presence of oxy-hemoproteins in the extracellular fluids. The second possibility is that the oxidation of NO to NO3 takes place through the formation of an NO2 intermediate, also in the presence of oxy-hemoproteins (10 ). 5-Methyltetrahydrofolate (5-MTHF) is the principal form (perhaps as much as ⬃95%) of folate in serum; therefore, most studies have used serum folate as an appropriate marker for folate status in their aim to explore the relationship between serum folate and tHcy concen-

trations (11 ). As suggested in Fig. 1, 5-MTHF is related to the formation of both NO and Hcy in the cell. In the presence of 5-MTHF, the enzyme dihydrobiopterin reductase can reduce quininoid dihydrobiopterin to tetrahydrobiopterin (BH4) (12 ). 5-MTHF also stabilizes BH4, enhances its binding to NO synthase (NOS), increases the activity of NOS, and thus regulates the ratio of NO to superoxide (O2. ) (13 ). Therefore, our findings may indicate that the intracellular concentrations of 5-MTHF may regulate the bioavailability of NO. A decrease in the folate concentration leads to an accumulation of tHcy. Oxidation of Hcy in the presence of trace elements generates O2. and hydrogen peroxide (H2O2) (14 ); it has therefore been suggested that folate deficiency may contribute to an increase in the concentration of O2. . Furthermore, increased concentrations of Hcy also reduce the availability of BH4 in vitro and thus increase the formation of O2. (15 ). In a recent study it was suggested that folic acid supplementation improves endothelial function in coronary artery disease (CAD) by a mechanism largely independent of Hcy (16 ). It is reasonable to assume that the improvement in endothelial function was attributable to increased formation of NO during the folic acid intervention. A different study showed that folic acid restored endothelial dysfunction in patients with hypercholester-

1268

Technical Briefs

olemia who had tHcy concentrations within the reference interval (17 ). Our observation that smokers have higher plasma concentrations of NO3 than do nonsmokers is in accordance with previously published data (Fig. 1B of the online Data Supplement) (18 ), but are contradictory to the findings of Node et al. (19 ). Oxidants, free radicals, and trace elements in cigarette smoke may activate macrophages and neutrophils in the circulation, which may generate increased amounts of NO and nitrite (20 ). It is also possible that cigarette smoke may prevent reaction of NO with thiols to form nitrosothiols, important reservoirs for NO in blood (21 ). From our findings in the present investigation, we postulate the following possible clinical implications: Increased dietary consumption of fruits and green vegetables may increase the concentrations of folate in serum/ cells and thus may increase the ratios NO/Hcy and NO/ O2. . Other possible benefits from increased concentrations of serum folate may be decreases in the concentrations of soluble circulating adhesion molecules. Increased concentrations of circulating adhesion molecules seem to be associated with the process of inflammation.

We thank Drs. Øyvind Hetland and Peter H. Evans for reading the manuscript and for useful comments. References 1. Cines DB, Pollak ES, Buck CA, Loscalzo J, Zimmerman GA, McEver RP, et al. Endothelial cells in physiology and in pathophysiology of vascular disorders. Blood 1998;91:3527– 61. 2. Zeballos GA, Bernstein RD, Thompson CI, Forfia PR, Seyedi N, Shen W, et al. Pharmacodynamics of plasma nitrate/nitrite as an indication of nitric oxide in conscious dogs. Circulation 1995;91:2982– 8. 3. Cai H, Harrison DG. Endothelial dysfunction in cardiovascular disease. The role of oxidant stress. Circ Res 2000;87:840 – 4. 4. Welch GN, Loscalzo J. Homocysteine and atherothrombosis. N Engl J Med 1998;338:1042–50. 5. Finkelstein JD. Methionine metabolism in mammals. J Nutr Biochem 1990; 1:228 –37. 6. Mansoor MA, Kristensen O, Hervig T, Bates CJ, Pentieva K, Berge T, et al. Homocysteine response to folic acid and vitamin B6 supplements. Oral doses of vitamin B6 decreases concentrations of serum folate. Scand J Clin Lab Invest 1999;59:139 – 46. 7. Verhaar MC, Stroes E, Rabelink TJ. Folates and cardiovascular disease. Arterioscler Thromb Vasc Biol 2002;22:6 –13. 8. Mansoor MA, Svardal AM, Ueland PM. Determination of the in vivo redox status of cysteine, cysteinylglycine homocysteine and glutathione in human plasma. Anal Biochem 1992;200:218 –29. 9. Bories PN, Scherman E, Dziedzic L. Analysis of nitrite and nitrate in biological fluids by capillary electrophoresis. Clin Biochem 1999;32:9 –14. 10. Ignarro LJ, Fukuto JM, Griscavage JM, Rogers NE, Byrns R. Oxidation of nitric oxide in aqueous solution to nitrite but not nitrate: comparison with enzymatically formed nitric oxide from L-arginine. Proc Natl Acad Sci U S A 1993;90:8103–7. 11. Drogan D, Klipstein-Grobusch K, Wans S, Luley C, Boeing H, Dierkes J. Plasma folate as marker of folate status in epidemiological studies: the European Investigation into Cancer and Nutrition (EPIC)-Potsdam study. Br J Nutr 2004;92:489 –96. 12. Mathews RG, Kaufman S. Characterization of the dihydropterin reductase activity of pig liver methylenetetrahydrofolate reductase. J Biol Chem 1980;255:6014 –7. 13. Stroes ESG, van Faassan EE, Yo M, Martasek P, Boer P, Govers R, et al. Folic acid reverts dysfunction of endothelial nitric oxide synthase. Circ Res 2000;86:1129 –34. 14. Starkebaum G, Harlan JM. Endothelial cell injury due to copper catalyzed hydrogen peroxide generation from homocysteine. J Clin Invest 1986;77: 1370 – 6. 15. Topal G, Brunet A, Millanvoye E, Boucher JL, Rendu F, Devynck MA, et al.

16.

17.

18.

19.

20. 21.

Homocysteine induces oxidative stress by uncoupling of NO synthase activity through reduction of tetrahydrobiopterin. Free Radic Biol Med 2004;36:1532– 41. Doshi SN, McDowell IFW, Moat SJ, Payne N, Durrant HJ, Lewis MJ, et al. Folic acid improves endothelial function in coronary artery disease via mechanisms largely independent of homocysteine lowering. Circulation 2000;105:22– 6. Verhaar MC, Wever RMF, Kastelein JJP. 5-Methyltetrahydrofolate, the active form of folic acid, restores endothelial function in familial hypercholesterolemia. Circulation 1998;97:237– 41. Yoon Y, Song J, Hong SH, Kim JQ. Plasma nitric oxide concentrations and nitric oxide synthase gene polymorphisms in coronary artery disease. Clin Chem 2000;46:1626 –30. Node K, Kitakaze M, Yoshikawa H, Kosaka H, Hori M. Reversible reduction in plasma concentration of nitric oxide induced by cigarette smoking in young adults. Am J Cardiol 1997;79:1538 – 41. Church DF, Pryor WA. Free radical chemistry of cigarette smoke and its toxicological implications. Environ Health Perspect 1985;64:111–26. Stamler JS, Jaraki O, Osborne J, Simon DI, Keaney J, Vita J, et al. Nitric oxide circulates in mammalian plasma primarily as an S-nitroso adduct of serum albumin. Proc Natl Acad Sci U S A 1992;89:7674 –7. DOI: 10.1373/clinchem.2004.046409

Liquid Chromatography–Tandem Mass Spectrometry Method for the Analysis of Asymmetric Dimethylarginine in Human Plasma, Edzard Schwedhelm,1* Jing TanAndresen,1 Renke Maas,1 Ulrich Riederer,2 Friedrich Schulze,1 and Rainer H. Bo¨ger1 (1 Institute of Experimental and Clinical Pharmacology, University Hospital HamburgEppendorf, and 2 Institute of Pharmacy, University of Hamburg, Hamburg, Germany; * address correspondence to this author at: Institute of Experimental and Clinical Pharmacology, University Hospital Hamburg-Eppendorf, 20246 Hamburg, Germany; fax 49-40-428039757, e-mail [email protected]) Nitric oxide (NO) is essential in numerous physiologic processes and may be involved in related pathologic processes. Asymmetric dimethylarginine (ADMA) is an endogenous inhibitor of isoforms of NO synthase in humans (1 ). ADMA originates from protein arginine methylation by protein arginine methyltransferases after protein hydrolysis (2 ). Enzymatic hydrolysis by dimethylarginine dimethylaminohydrolases to dimethylamine and citrulline is the major pathway for elimination of ADMA (3 ). Circulating ADMA is altered in patients with cardiovascular and neurologic diseases, erectile dysfunction, and many other disorders (4 – 6 ), and increased circulating ADMA independently predicts future cardiovascular events and mortality (7, 8 ). Short-time infusion of ADMA affects hemodynamics and cardiac function in humans (3, 9 ). Analytical methods for the measurement of ADMA include HPLC, capillary electrophoresis, ELISA, and mass spectrometry (MS). The commonly used HPLC methods with fluorescence detection (10 ) measure o-phthaldialdehyde derivatives of ADMA. These derivatives are not stable and must be analyzed on-line. Moreover, ADMA must be separated by chromatographic means from its biologically inactive isomer, symmetric dimethylamine

1269

Clinical Chemistry 51, No. 7, 2005

(SDMA). Thus, these HPLC methods have long analysis times of up to 45 min. Alternatively, ADMA and SDMA can be analyzed by gas chromatography–mass spectrometry (11, 12 ), which shows different fragmentation patterns for the respective ions. These methods include several extraction and derivatization steps and require considerable analysis times. Recently, Kirchherr and Ku¨hn-Velten (13 ) developed a mass spectrometric method that omits derivatization procedures; however, chromatographic separation of ADMA from SDMA is still required for this method. Here we report a simple and rapid liquid chromatography–tandem MS (LC-MS/MS)based method with 1 derivatization step and sample pretreatment consisting of protein precipitation. This is the first LC-MS– based method that separates methylated arginine derivatives by specific fragmentation patterns instead of by chromatographic separation. Analyses were performed on a Varian 1200L Triple Quadrupole MS equipped with 2 Varian ProStar model 210 HPLC pumps and a Varian analytical column [50 ⫻ 2.0 mm (i.d.)] packed with Polaris C18-Ether (3 ␮m bead size). The mobile phases consisted of methanol containing 1 mL/L formic acid (mobile phase A) and water containing 1 mL/L formic acid (mobile phase B). Chromatography was performed at 25 °C with a flow rate of 0.4 mL/min. The gradient started with 2% A for 0.5 min and increased linearly to 50% A over 1.5 min, with subsequent reequilibration with 2% A for 2 min. Nitrogen was used as the nebulizing and drying gas (380 °C) at 90 and 180 L/h, respectively. For ionization in the positive electrospray ionization mode (ESI⫹), the needle and shield voltages were set at 5850 and 400 V, respectively. The following transitions were observed after fragmentation with argon (1.5 mTorr): m/z 231370 [collision energy (CE), ⫺22 eV] for l-arginine (USP Reference Standard); m/z 238377 (CE, ⫺24 eV) for l-[2H7]-arginine (98 atom% 2H isotopic purity; 2,3,3,4,4,5,5-l-[2H7]-arginine; Euriso-top); m/z 259.33214 (CE, ⫺16 eV) for ADMA (⬎99% purity; Sigma-Aldrich); m/z 259.33228 (CE, ⫺14 eV) for SDMA (⬎99% purity; Sigma-Aldrich); and m/z 265.33220 (CE, ⫺16 eV) for [2H6]-ADMA (98 atom% 2H isotopic purity; 3,3,4,4,5,5[2H6]-ADMA). The synthesis and characterization of [2H6]-ADMA were performed as described previously (12 ). All compounds were analyzed as their butyl ester derivatives (see below). All parent ions (MS) represented the molecular protonated cations, i.e., [M⫹H]⫹. The major daughter ions formed, i.e., m/z 70 and m/z 77, for larginine and l-[2H7]-arginine, respectively, correspond to a pyrrolinium ion as a cyclization product (14 ). Major daughter ions of m/z 259.3 were m/z 214 [M⫹H ⫺ NHCH3CH3]⫹ and m/z 228 [M⫹H ⫺ NH2CH3]⫹ for authentic ADMA and SDMA, respectively. The m/z 228 and m/z 214 ions were absent in the product ion mass spectra of ADMA and SDMA, respectively. We checked ion suppression by comparing peak areas of aqueous calibrator solutions without matrix and calibrator added to plasma matrix (15 ). Ion suppression was ⬍10% for 5 different specimens.

We used 3 different calibrators, l-arginine, ADMA, and SDMA, at 6 different concentrations (n ⫽ 5 each): 0, 25, 50, 100, 250, and 500 ␮mol/L for l-arginine and 0, 0.25, 0.5, 1, 2, and 4 ␮mol/L for ADMA and SDMA, respectively. Aqueous stock solutions of l-arginine, ADMA, and SDMA were made by weighing authentic material supplied by the manufacturer. The calibrators were treated exactly the same as patient plasma samples. l-[2H7]Arginine and [2H6]-ADMA were added as internal standards at concentrations of 50 and 2 ␮mol/L, respectively. [2H6]-ADMA was used as internal standard for ADMA and SDMA. Quantitative determinations in biological samples (50 ␮L) were performed by dilution of 5 ␮L of aqueous internal standards (500 ␮mol/L l-[2H7]-arginine and 20 ␮mol/L [2H6]-ADMA) corresponding to concentrations of 50 and 2 ␮mol/L, respectively, in 50 ␮L of biological sample. Subsequently, proteins were precipitated with 100 ␮L of acetone, and dried supernatants were derivatized. Compounds were derivatized with 100 ␮L of 1 mol/L HCl in 1-butanol for 17 min at 65 °C. After evaporation, samples were reconstituted in 1 mL of water; 20 ␮L of the water extract was injected on the column. The mean (SD) retention times of the butyl ester derivatives of l-arginine, ADMA, and SDMA, respectively, were 1.03 (0.04) min (CV ⫽ 3.9%; n ⫽ 6), 1.84 (0.02) min (CV ⫽ 1.3%), and 2.03 (0.02) min (CV ⫽ 0.8%). Differences in retention times for unlabeled and labeled compounds were not statistically significant. Table 1. Precision of the LC-MS/MS method for the measurement of L-arginine, ADMA, and SDMA in human plasma.

Analyte L-Arginine

Concentration added, ␮mol/L (n ⴝ 5)

0 0.5 1.0 25 50 100 250

Mean (SD) measured concentration, ␮mol/L

69.4 (1.0) 70.0 (0.6) 70.4 (0.4) 95.5 (1.0) 122 (2.6) 173 (5.5) 320 (8.6)

Mean (SD) recovery of added analyte, %

CV, %

NAa 111 (22) 97.7 (19) 104 (5.5) 105 (2.3) 104 (2.5) 100 (1.5)

1.4 0.9 0.6 3.2 2.1 3.2 2.7

ADMA

0 0.05 0.1 0.5 1 2 4

0.448 (0.007) 0.496 (0.011) 0.553 (0.004) 0.949 (0.012) 1.464 (0.022) 2.422 (0.030) 4.443 (0.104)

NA 96.9 (10.1) 105 (3.9) 100 (1.1) 102 (1.0) 98.7 (1.4) 99.9 (1.2)

1.5 2.3 0.8 1.3 1.5 1.3 2.3

SDMA

0 0.05 0.1 0.5 1 2 4

0.424 (0.024) 0.479 (0.008) 0.515 (0.013) 0.924 (0.043) 1.449 (0.041) 2.441 (0.070) 4.480 (0.109)

NA 109 (14.6) 90.8 (11.4) 100 (3.9) 103 (3.7) 101 (3.1) 101 (2.4)

5.6 1.7 2.5 4.7 2.9 2.9 2.4

a

NA, not applicable.

1270

Technical Briefs

L-Arginine

L-[2H7]-Arginine

SDMA

ADMA

[2H6]-ADMA

Fig. 1. Representative LC-MS/MS chromatograms for L-arginine, L-[2H7]-arginine (internal standard), SDMA, ADMA, and [2H6]-ADMA (internal standard). Chromatograms from the analysis of a human plasma sample (50 ␮L) to which 50 ␮mol/L L-[2H7]-arginine and 2 ␮mol/L [2H6]-ADMA had been added as internal standards. Trace 1 (top), L-arginine; trace 2, L-[2H7]-arginine; trace 3, SDMA; trace 4, ADMA; trace 5 (bottom), [2H6]-ADMA. All of the above analytes could be separated based on their specific daughter ions in LC-MS/MS analysis: m/z 70 for L-arginine, m/z 77 for L-[2H7]-arginine, m/z 228 for SDMA, m/z 214 for ADMA, and m/z 220 for [2H6]-ADMA. Scan rate was 1 s. (⫹) ESI SRM, electrospray ionization single reaction monitoring.

The linear regression equations for peak-area ratio (LC-MS/MS; y) and ratio of injected calibrators (x) were as follows: y ⫽ 0.94x ⫺ 0.001 (r2 ⫽ 0.999) for l-arginine; y ⫽ 0.98x ⫹ 0.02 (r2 ⫽ 0.999) for ADMA; and y ⫽ 2.05x ⫹ 0.01 (r2 ⫽ 0.999) for SDMA. Ideally, regression analysis of peak-area ratios should give a slope of 1; however, the intensities of the daughter ions obtained for labeled and unlabeled compounds were not the same. Thus, only under MS conditions was a slope of 1 obtained: The linear regression equations for peak-area ratio (LC-MS; y) and ratio of injected calibrators (x) were as follows: y ⫽ 1.003x ⫺ 0.02 (r2 ⫽ 0.999) for l-arginine; y ⫽ 1.001x ⫹ 0.01 (r2 ⫽ 0.999) for ADMA; and y ⫽ 1.002x ⫹ 0.03 (r2 ⫽ 0.998) for SDMA. Nevertheless, calibration curves were calculated to account for these differences, and the slopes of the calibration curves were included in all plasma concentration calculations. The lower limits of detection, defined as a signal-to-noise ratio of 3, were 45 nmol/L for l-arginine, 3 nmol/L for ADMA, and 2 nmol/L for SDMA.

We validated our LC-MS/MS method by adding different concentrations of l-arginine, ADMA, and SDMA in quintuplicate to samples. l-Arginine was added at 0, 0.5, 1, 25, 50, 100, and 250 ␮mol/L, and ADMA and SDMA were added at 0, 0.05, 0.1, 0.5, 1, 2, and 4 ␮mol/L. Linear regression analysis between measured (y) and added (x) concentrations yielded the following slopes and y-intercepts: 1.01 and 70.3 ␮mol/L (r2 ⫽ 0.999) for l-arginine; 1.00 and 0.45 ␮mol/L (r2 ⫽ 0.999) for ADMA; and 1.01 and 0.42 ␮mol/L (r2 ⫽ 0.999) for SDMA. Data from these validation experiments are listed in Table 1. The observed recoveries of the added analytes were 97.7%–111% for l-arginine, 96.9%–105% for ADMA, and 90.8%–109% for SDMA. All CVs were ⬍6% for the imprecision and ⬍20% for recovery except for the addition of 0.5 ␮mol/L larginine. Thus, the analytical range of the method was 1–250 ␮mol/L for l-arginine and 50 nmol/L to 4 ␮mol/L for ADMA and SDMA, with 1 ␮mol/L and 50 nmol/L representing the lower limits of quantification for l-

Clinical Chemistry 51, No. 7, 2005

arginine and ADMA and SDMA, respectively. The addition of l-arginine, ADMA, or SDMA did not influence the measurement of the other analytes. We also determined within- and between-run imprecision (CVs; n ⫽ 10). The within-run CVs were 3.3% for l-arginine, 2.6% for ADMA, and 2.5% for SDMA. Between-run CVs were 4.7% for l-arginine, 4.1% for ADMA, and 3.9% for SDMA We also measured ADMA in plasma samples from healthy human volunteers (n ⫽ 22) by this LC-MS/MS method. Shown in Fig. 1 is a typical chromatogram of a human plasma sample. Samples were aliquoted (50 ␮L) and frozen at ⫺20 °C before analysis. The mean (SD) concentrations of l-arginine, ADMA, and SDMA were 65.6 (23.4), 0.55 (0.14), and 0.69 (0.23) ␮mol/L, respectively. The mean (SD) molar ratio of l-arginine to ADMA was 132 (55). Our LC-MS/MS– based method allows for the simultaneous determination of ADMA, SDMA, and l-arginine and has been validated for the measurement of these 3 analytes in human plasma. Values obtained for l-arginine, ADMA, and SDMA in human plasma are within previously reported ranges (11–13, 16 ). Major advantages over previous reported methods for the measurement of ADMA include a minimized sample volume and reduction in laboratory work. In particular, in comparison with other LC-MS– based methods (13, 14, 16, 17 ), sample run time has been reduced to 4 min including reequilibration. Chromatographic separation required only slightly more than 2 min. This considerable reduction in analysis time was achieved by use of the specific fragmentation patterns of the ester derivatives of ADMA and SDMA. Identical MS spectra were obtained from the ester derivatives of ADMA and SDMA, but the major fragments in the daughter ion spectra for ADMA and SDMA were at m/z 214 and m/z 228, respectively. These ions were specific for ADMA and SDMA: they were absent in the corresponding spectra of SDMA and ADMA, respectively. In contrast to previous reports, the proposed method allows, for the first time, spectrometric instead of chromatographic separation of these methylated arginines (see the Data Supplement that accompanies the online version of this Technical Brief at http://www.clinchem.org/content/ vol51/issue7/). References 1. Tsikas D, Bo¨ger RH, Sandmann J, Bode-Bo¨ger SM, Fro¨lich JC. Endogenous nitric oxide synthase inhibitors are responsible for the L-arginine paradox. FEBS Lett 2000;478:1–3. 2. McBride AE, Silver PA. State of the Arg: protein methylation at arginine comes of age. Cell 2001;106:5– 8. 3. Achan V, Broadhead M, Malaki M, Whitley G, Leiper J, MacAllister R, et al. Asymmetric dimethylarginine causes hypertension and cardiac dysfunction in humans and is actively metabolized by dimethylarginine dimethylaminohydrolase. Arterioscler Thromb Vasc Biol 2003;23:1455–9. 4. Bo¨ger RH, Bode-Bo¨ger SM, Szuba A, Tsao PS, Chan JR, Tangphao O, et al. Asymmetric dimethylarginine (ADMA): a novel risk factor for endothelial dysfunction: its role in hypercholesterolemia. Circulation 1998;98:1842–7. 5. Selley ML. Increased concentrations of homocysteine and asymmetric dimethylarginine and decreased concentrations of nitric oxide in the plasma of patients with Alzheimer’s disease. Neurobiol Aging 2003;24:903–7. 6. Maas R, Zabel M, Wenske S, Ventura R, Schwedhelm E, Steenpass A, et al. Erectile dysfunction in patients with cardiovascular disease and diabetes: evidence for disturbed relation of L-arginine and the endogenous inhibitor of

7.

8.

9.

10.

11.

12.

13.

14.

15. 16.

17.

1271

NO-synthase asymmetrical dimethylarginine (ADMA). Z Kardiol 2004; 93(Suppl 3):V1235. Zoccali C, Bode-Bo¨ger S, Mallamaci F, Benedetto F, Tripepi G, Malatino L, et al. Plasma concentration of asymmetrical dimethylarginine and mortality in patients with end-stage renal disease: a prospective study. Lancet 2001; 358:2113–7. Bo¨ger RH, Lenzen H, Hanefeld C, Bartling A, Osterziel KJ, Kusus M, et al. Asymmetric dimethylarginine: an endogenous inhibitor of NO synthase is a predictor of the risk for coronary heart disease—results of the multicenter CARDIAC study. Circulation 2003;108(Suppl IV):256. Kielstein J, Impraim B, Simmel S, Bode-Bo¨ger SM, Tsikas D, Fro¨lich JC, et al. Cardiovascular effects of systemic nitric oxide synthase inhibition with asymmetrical dimethylarginine in humans. Circulation 2004;109:172–7. Tsikas D, Junker W, Fro¨lich JC. Determination of dimethylated arginines in human plasma by high-performance liquid chromatography. J Chromatogr B 1998;705:174 – 6. Tsikas D, Schubert B, Gutzki FM, Sandmann J, Fro¨lich JC. Quantitative determination of circulating and urinary asymmetric dimethylarginine (ADMA) in humans by gas chromatography-tandem mass spectrometry as methyl ester tri(N-pentafluoropropionyl) derivative. J Chromatogr B 2003; 798:87–99. Albsmeier J, Schwedhelm E, Schulze F, Kastner M, Bo¨ger RH. Determination of NG,NG-dimethyl-L-arginine, an endogenous NO synthase inhibitor, by gas chromatography-mass spectrometry. J Chromatogr B 2004;809:59 – 65. Kirchherr H, Ku¨hn-Velten WN. HPLC-tandem mass spectrometric method for rapid quantification of dimethylarginines in human plasma. Clin Chem 2005;51:249 –52. Vishwanathan K, Tackett RL, Stewart JT, Bartlett MG. Determination of arginine and methylated arginines in human plasma by liquid chromatography-tandem mass spectrometry. J Chromatogr B 2000;748:157– 66. Annesley TM. Ion suppression in mass spectrometry. Clin Chem 2003;49: 1041– 4. Martens-Lobenhoffer J, Bode-Bo¨ger SM. Simultaneous detection of arginine, asymmetric dimethylarginine, symmetric dimethylarginine and citrulline in human plasma and urine applying liquid chromatography-mass spectrometry. J Chromatogr B 2003;798:231–9. Huang LF, Guo FQ, Liang YZ, Li BY, Cheng BM. Simultaneous determination of L-arginine and its mono- and dimethylated metabolites in human plasma by high-performance liquid chromatography-mass spectrometry. Anal Bioanal Chem 2004;380:643–9. Previously published online at DOI: 10.1373/clinchem.2004.046037

Relationship between Isoprostane Concentrations, Metabolic Acidosis, and Morbid Neonatal Outcome, Michael Scott Rogers,1* Chi Chiu Wang,1 Tze Kin Lau,1 Xin Xiao,2 Xiao Guang Zhou,3 Tai Fai Fok,4 Kai On Chu,1 and Chi Pui Pang5 (Departments of 1 Obstetrics & Gynaecology and 4 Paediatrics, The Chinese University of Hong Kong, Prince of Wales Hospital, Shatin, Hong Kong; 2 Department of Neonatology, Perinatal Medicine Center, Medical College of Jinan University, Guangzhou, China; 3 Department of Neonatology, The 2nd Affiliated Hospital of Guangzhou Medical College, Guangzhou, China; 5 Department of Ophthalmology & Visual Sciences, The Chinese University of Hong Kong, University Eye Centre, Hong Kong Eye Hospital, Kowloon, Hong Kong; * address correspondence to this author at: 1st Floor, Block E, Department of Obstetrics and Gynaecology, The Chinese University of Hong Kong, Prince of Wales Hospital, Shatin, New Territories, Hong Kong; fax 852-2636-0008, e-mail msrogers@ cuhk.edu.hk) When a fetus is subjected to a massive perinatal hypoxicischemic insult, it may suffer sufficient damage to cause intrauterine death and stillbirth (1–5 ). In less severe,

1272

Technical Briefs

Table 1. Areas under the ROC curves for prediction of HIE/PND. Asymptotic 95% confidence interval Group

Test result variable(s)

Area

SE

P

Lower boundary

Upper boundary

Infants admitted to the NICU for respiratory support (n ⫽ 40)

Apgar score at 1-min Apgar score at 5-min Umbilical arterial pH Arterial base excess Isoprostane

0.62 0.74 0.77 0.90 0.74

0.10 0.09 0.07 0.06 0.10

NS ⬍0.05 ⬍0.01 ⬍0.001 ⬍0.05

0.43 0.56 0.64 0.79 0.54

0.81 0.92 0.91 1.01 0.94

All 297 cases

Apgar score at 1-min Apgar score at 5-min Umbilical arterial pH Arterial base excess Isoprostane

0.84 0.85 0.89 0.99 0.70

0.07 0.08 0.03 0.01 0.11

⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 0.05

0.71 0.69 0.83 0.98 0.48

0.96 1.01 0.96 1.00 0.92

nonfatal cases, prolonged or severe intrauterine hypoxia may lead to serious neonatal complications such as hypoxic-ischemic encephalopathy (HIE), cerebral palsy, and impaired myocardial function (6 – 8 ). Hypoxia during labor leads to anaerobic respiration in the fetus and an accumulation of lactic acid in the tissues, producing metabolic acidosis. Damage to the brain and heart in asphyxia neonatorum is not a direct result of hypoxia, but rather of the toxic effects of reactive oxygen species generated after reperfusion of ischemic tissues. We have demonstrated that lipid peroxidation in the fetus increases during normal labor (9 ) and that umbilical cord plasma lipid peroxide concentrations are higher in situa-

Fig. 1. Scatter plot of isoprostane concentration (log scale) against base excess in umbilical cord arterial blood for all neonatal outcome categories. The horizontal line is the 90th centile of the cord arterial isoprostane concentration; the vertical line is the 10th centile for cord arterial base excess. Perinatal outcomes: v, prematurity only; ⽧, perinatal death; , poor outcome (asphyxia at birth); 䡺, sepsis (no asphyxia at birth); ‚, good outcome (asphyxia at birth); ✕, good outcome (no asphyxia); ——, lowess regression for the total sample.

tions known to lead to intrapartum hypoxia (10 –12 ) and much lower after elective cesarean delivery (13 ). Reperfusion of damaged organs with oxygenated blood after delivery may also be detrimental to the infant’s long-term survival. Prediction of long-term neurodevelopmental outcomes in infants with birth asphyxia remains challenging. Unfortunately, previous studies have shown that clinical and biochemical variables, such as umbilical artery blood gases or Apgar scores, are of limited value in predicting morbid neonatal outcome. We have proposed the use of lipid peroxidation products, acting as footprints of oxidative stress, as alternative measures of perinatal outcome (14 ). Increased concentrations of lipid peroxidation products in cord arterial blood are associated with clinical situations known to cause fetal distress, but they have not been shown to be associated with significant morbid outcomes such as HIE. This study aims to demonstrate an association between increased umbilical cord blood lipid peroxide concentrations and morbid neonatal outcomes. The study hypotheses are (a) that cord arterial plasma isoprostane (15 ) concentrations are higher in neonates who require immediate admission to the neonatal intensive care unit (NICU) for respiratory support than in healthy neonates, and (b) that cord arterial plasma isoprostane concentrations are higher in infants who will ultimately develop HIE than in those who will recover without neurologic sequelae. In total, 301 neonates had cord blood 8-isoprostane measured, and 297 cord blood samples were confirmed as arterial in origin (see the Materials and Methods section of the Data Supplement that accompanies the online version of this Technical Brief at http://www.clinchem.org/ content/vol51/issue7/). Of the infants for whom arterial cord blood samples were available, 13.1% (39 of 297) required immediate postdelivery admission to a NICU for respiratory support, and another neonate (1 of 297; 0.3%) died immediately after birth and was classified as a fresh stillbirth. Of these 40 (39 ⫹ 1) cases, 24 (60%) were originally classified as having asphyxia neonatorum, with 11 of the 24 (45.8%) ultimately classified as having poor outcome [8 with HIE and 3 with postnatal death (PND)]; the remaining 13 cases (54.2%) were classified as having a

Clinical Chemistry 51, No. 7, 2005

good outcome. Another 25% of infants (10 of 40) showed no evidence of asphyxia at birth but were treated in a NICU because of pneumonia (9 with bacterial sepsis and 1 with meconium aspiration syndrome), and 15% (6 of 40) were admitted solely because of prematurity. The maternal, obstetric, and neonatal characteristics for all 297 neonates who had cord blood samples available, grouped according to outcome, are shown in Table 1 of the online Data Supplement. Statistical significance was achieved (P ⬍0.01) in a group sequential analysis comparing cord arterial 8-isoprostane concentrations in the 40 sick neonates requiring immediate NICU admission for respiratory support (including the 1 fresh stillbirth) and the 257 healthy control infants (Fig. 1 of the online Data Supplement), and in the post hoc analysis comparing cord isoprostane in NICU cases that developed HIE with those that did not. Shown in Fig. 1 is a scatter plot of cord arterial blood 8-isoprostane plotted against cord arterial base excess according to actual neonatal outcome group for all cases. Lowess regression suggested a linear relationship between 8-isoprostane concentrations and base excess when base excess was below ⫺10. Linear regression analysis of this group (n ⫽ 23) gave a correlation coefficient (r) of ⫺0.673 (P ⫽ 0.002). The results of ROC curve analysis for predictions of poor outcome (HIE/PND) on the basis of 1- and 5-min Apgar score, cord arterial blood isoprostane concentration, cord arterial pH, and cord arterial base excess are shown in Table 1. The values for the 40 cases admitted to a NICU for respiratory support are listed first, followed by the values for all 297 cases. All infants who survived with evidence of significant brain damage (HIE) had severe metabolic acidosis (base excess ⫺12 or lower), as did one of the infants who subsequently died. The pattern of oxidative stress was similar, with 6 of these 9 cases (66.7%) having increased 8-isoprostane concentrations. The 8-isoprostane concentration was also increased in 25 nonasphyxiated cases (9.2%); 17 (68%) of these had nuchal cord entanglement detected at birth, with 12 (48%) having tight nuchal entanglement. Although the significant difference in isoprostane concentrations between healthy infants and neonates requiring NICU admission for respiratory support is of interest, an ability to identify those infants who will ultimately develop HIE or PND from among sick infants requiring NICU admission would be of greater value. In a post hoc analysis of the 39 infants admitted to a NICU for respiratory support, cord arterial isoprostane concentrations were shown to be significantly higher (P ⬍0.01) in neonates who subsequently were classified as having a poor outcome. The 8-isoprostane concentration performed significantly better than chance (area under the curve, 0.74), roughly on a par with cord arterial pH and 5-min Apgar score, but it significantly underperformed compared with base excess (DeLong statistic, P ⬍0.05). The 1-min Apgar score was not able to differentiate between infants who were later classified as HIE/PND and those with a good

1273

outcome. Isoprostane exhibited only a moderate ability to identify those infants with a poor outcome from among all neonates studied (area under the curve, 0.70), but it underperformed compared with the standard outcome measures (pH, base excess, and 1- and 5-min Apgar scores). This reflects the wider variance in 8-isoprostane concentrations among apparently healthy neonates, particularly when there has been evidence of cord compression during labor. Base excess appears to be a better predictor of HIE, suggesting that the duration of sublethal hypoxic insult, as well as its severity, plays an important role in the pathophysiology of HIE. The prediction of poor outcome cannot be improved by combining isoprostane with base excess because all 8 cases of HIE had base excess measurements between ⫺12 and ⫺20. For those neonates with metabolic acidosis, cord arterial blood isoprostane concentrations and base excess showed a linear relationship, suggesting that oxidative stress plays a role in a significant proportion of cases with hypoxic-ischemic encephalopathy. Only 1 of the 3 perinatal deaths had a base excess in this range. This suggests that the other 2 perinatal deaths were unrelated to chronic intrapartum hypoxia. In fact, one was a fresh stillbirth associated with a massive placental abruption, and the other was a neonatal death attributable to chorioamnionitis but with no evidence of fetal distress or asphyxia at birth (Apgar scores of 9 at 1 min and 10 at 5 min). It is not surprising, therefore, to find that these 2 cases also had 8-isoprostane concentrations within the reference interval. Despite the presence of metabolic acidosis in the pooroutcome group, only 2 cases actually had a cord arterial blood pH ⬍7.05 (1 stillbirth and 1 case with HIE). One healthy neonate also had a low pH and base excess of ⫺12. The links between obstetric complications, intrapartum fetal heart rate abnormalities, and perinatal outcome measures are tenuous, partly because of the therapeutic effect of intervention on outcome measurements (treatment paradox) (16 ). Obstetric intervention may remove the cause of fetal distress but does not necessarily represent an abrupt end to the pathologic processes. Unlike pH, which can be affected by intra- and postpartum resuscitation, lipid peroxide concentrations remain relatively stable immediately after delivery and may actually continue to increase after resuscitation if the hypoxic injury has been severe (17 ). In the pathogenesis of perinatal HIE, variable degrees of hypoxia occur over a period of several hours, usually during labor. During this time, the fetus compensates for the production of excess lactic acid by producing buffering hydrogen ions, causing an increase in base deficit: the pH therefore remains within the reference interval until these buffering systems decompensate. Similarly, excess oxygen free radicals are removed by various circulating and tissue antioxidants, causing depletion of total antioxidant capacity; evidence of oxidative stress (increased lipid peroxidation) manifests only after these antioxidant systems decompensate. Whether these 2 systems are damaged by recurrent asphyxia to an equal extent and

1274

Technical Briefs

during the same time frame is beyond the scope of this observational study. Our study highlights the fact that isoprostane concentrations have the same limitations as other measures of obstetric outcome, such as pH and Apgar score: i.e., they are increased in neonates with HIE, but may also be increased in apparently healthy infants and may be within reference values in infants who die from acute causes or in whom a pathology manifests after delivery.

We are grateful to all medical and nursing staffs in the labor wards and neonatal units at Prince of Wales Hospital, The First Affiliation Hospital of Medical College of Jinan University, and The Women’s & Children’s Hospital of Guangdong Province for case recruitment and sample collection. The study was supported by the Hong Kong Special Administration Region Research Grant Council (CUHK4329/99M). All contributors and guarantors in this study are independent from the funding agency, and the views expressed here are those of the authors. This report was partially awarded (Asian-Pacific Congress of Clinical Biochemistry Regional Service Award) and represented at the 10th Asian-Pacific Congress of Clinical Biochemistry and 42nd Annual Scientific Conference of the Australasian Association of Clinical Biochemists in Perth, Australia, during September 2004. References 1. Ferguson SD, Gehrke T. Deaths related to intrapartum asphyxia. Audit in one unit found neonatal care to be suboptimal at weekends [Letter]. BMJ 1998;316:1318 –9. 2. Macintosh M, Lee C. Deaths related to intrapartum asphyxia. Intrapartum death rates in England 1993–5 did not show consistent peaks or troughs [Letter]. BMJ 1998;316:1319. 3. Murphy DJ. Deaths related to intrapartum asphyxia. Denominators are needed before conclusions can be drawn [Letter]. BMJ 1998;316:1319. 4. Luckas M, Thomson A, Aird I. Deaths related to intrapartum asphyxia. Consultant expansion in obstetrics and gynaecology is not fast enough [Letter]. BMJ 1998;316:1319. 5. Bastian H, Keirse MJNC, Lancaster PAL. Perinatal death associated with planned home birth in Australia: population based study. BMJ 1998;317: 384 – 8. 6. Barela TD, Johnson JD, Hayek A. Metabolic acidosis in the newborn period. Clin Endocrinol Metab 1983;12:429 – 46. 7. Low JA, Panagiotopoulos C, Derrick EJ. Newborn complications after intrapartum asphyxia with metabolic acidosis in the term fetus. Am J Obstet Gynecol 1994;170:1081–7. 8. Socol ML, Garcia PM, Riter S. Depressed Apgar scores, acid-base status, and neurologic outcome. Am J Obstet Gynecol 1994;170:991–9. 9. Rogers MS, Wang W, Mongelli M, Pang CP, Duley JA, Chang AMZ. Lipid peroxidation in cord blood at birth: a marker of fetal hypoxia during labour. Gynecol Obstet Invest 1997;44:229 –33. 10. Wang CC, Rogers MS. Lipid peroxides in cord blood: the effects of umbilical nuchal cord. Br J Obstet Gynaecol 1997;104:251–5. 11. Liu BY, Wang CC, Lau TK, Chu CY, Pang CP, Rogers MS, et al. Meconium stained liquor (MSL) during labor is associated with raised neonatal cord blood 8-iso-prostaglandin F2␣ concentration. Am J Obstet Gynecol 2004; 192:289 –94. 12. Mongelli JM, Wang CC, Wang W, Pang CCP, Rogers MS. Oxygen free radical activity in the second stage of labor. Acta Obstet Gynecol Scand 1997;76: 765– 8. 13. Rogers MS, Mongelli JM, Tsang KH, Wang CC, Law KP. Lipid peroxidation in cord blood at birth: the effect of labour. Br J Obstet Gynaecol 1998;105: 739 – 44. 14. Wang W, Pang CC, Rogers MS, Chang AMZ. Lipid peroxidation in cord blood at birth. Am J Obstet Gynecol 1996;174:62–5. 15. Hoffman SW, Rzigalinski BA, Willoughby KA, Ellis EF. Astrocytes generate isoprostanes in response to trauma or oxygen radicals. J Neurotrauma 2000;17:415–20.

16. Rogers MS, Chang AMZ. Perinatal asphyxia: a Bayesian analysis of prediction and prevention. J Obstet Gynaecol 1991;11:34 – 40. 17. Rogers MS, Murray HG, Wang CC, Pennell CE, Turner A, Yan P, et al. Oxidative stress in the fetal lamb brain following intermittent umbilical cord occlusions: a path analysis. Br J Obstet Gynaecol 2001;108:1283–90. DOI: 10.1373/clinchem.2004.047241

Comparison of a Fully Automated Immunoassay with a Point-of-Care Testing Method for B-Type Natriuretic Peptide, Concetta Prontera, Simona Storti, Michele Emdin, Claudio Passino, Luc Zyw, Gian Carlo Zucchelli, and Aldo Clerico* (Consiglio Nazionale delle Ricerche Institute of Clinical Physiology, Pisa, Italy; * address correspondence to this author at: Laboratory of Cardiovascular Endocrinology and Cellular Biology, CNR Institute of Clinical Physiology, Via Trieste 41, 56125 Pisa, Italy; fax 39-0585493601, e-mail [email protected]) The clinical relevance of B-type natriuretic peptide (BNP) as a diagnostic tool and prognostic marker in patients with cardiovascular diseases has been confirmed recently (1–3 ). Over the last 5 years, several immunoassay methods for the measurement of BNP, which share some analytical characteristics, such as full automation, turnaround time ⬍60 min, lower imprecision, and/or better analytical and functional sensitivity, have become commercially available (1, 4 ). The analytical performance and diagnostic accuracy of immunoassays for BNP were compared recently (5, 6 ). In the present study, we evaluated the analytical performance and diagnostic accuracy of an automated immunochemiluminescent assay for BNP (ACCESS System; Beckman Coulter). We also compared its performance and accuracy with the TRIAGE BNP test (Biosite). The 2 immunoassays use as antigen the intact BNP 1–32 peptide and the same mouse anti-human BNP antibodies. The capture antibody (Scios) recognizes the peptide ring, whereas the detection antibody (Biosite) recognizes an epitope between amino acids 5 and 10 at the NH2 terminus (6, 7 ). All statistical analyses were performed by use of parametric tests after log transformation of the original data. Nonparametric tests were also performed and showed identical statistical trends. We collected blood samples from 91 healthy individuals [51 women and 40 men; mean (SD) age, 43.2 (13.4) years; range, 16 –71 years] and 214 patients with idiopathic or secondary cardiomyopathy (only 21 with a left ventricular ejection fraction ⬎50%). Blood was collected between 0800 and 0900 after the individuals had fasted overnight and had rested for 20 min in a supine position. Immediately after withdrawal, blood samples (8 –10 mL) were placed in ice-chilled disposable polypropylene tubes containing EDTA (1 mg/mL of plasma). Plasma samples were obtained shortly after venipuncture by centrifugation for 15 min at 4 °C and then were frozen and stored at ⫺20 °C in 0.5-mL aliquots until being assayed. We found that plasma samples can be kept at 37 °C for

1275

Clinical Chemistry 51, No. 7, 2005

Fig. 1. Total imprecision profiles (A) and Bland–Altman difference plot (B) for the tested immunoassays. (A), the CV values corresponding to desirable imprecision (15% CV) and optimum imprecision (10% CV), respectively, are indicated with dashed lines. (B), the difference plot shows the difference between methods as a percentage (y axis) plotted vs the mean value (x axis). Solid horizontal lines indicate the mean difference and the ⫾ 3 SD values. The dashed line indicates no difference.

4 h without a significant decrease in measured (on an ACCESS system) BNP concentration [mean (SD) measured value, 97.8 (5.3)% of the original concentration; mean difference from the original concentration, 10.3 ng/L; 95% confidence interval (CI), ⫺20.4 to 41.1 ng/L (P ⫽ 0.1147); median measured value, 790 ng/L (range, 37–2904 ng/L); n ⫽ 15]. We observed a slight but significant difference when plasma samples were kept at 4 °C for 24 h [mean measured value 89.0 (5.0)% of the original concentration; mean difference from the original concentration, 27.5 ng/L; 95% CI, 13.7– 41.4 ng/L (P ⬍0.0001); median measured value, 160.5 ng/L (range, 12– 824 ng/L); n ⫽ 24]. We assessed the detection limit of the ACCESS by repeatedly measuring (n ⫽ 20) the calibrator with a BNP concentration of 0 ng/L in 3 different runs. The mean (SD) measured concentration was 0.3 (0.1) ng/L (range, 0.3– 0.4 ng/L). We confirmed the results of a previous study from our laboratory indicating that the detection limit of the TRIAGE is ⬍8 ng/L (manufacturer-reported detection limit, 5 ng/L) (8 ). The imprecision profiles of the TRIAGE and ACCESS, obtained by repeated measurements over

20 different days of several plasma samples (7 on the TRIAGE and 12 on the ACCESS) with BNP concentrations covering the analytical working range, are shown in Fig. 1A. According to the IFCC goals for desirable (15% total CV) and optimum (10% total CV) imprecision (7 ), the 2 methods had similar functional sensitivities (⬃10 ng/L and 30 ng/L, respectively). The imprecision for samples with BNP values ⬍10 ng/L was not satisfactory in either method. We assessed the within-run and total imprecision of the ACCESS according to the NCCLS EP5-A guideline (including experimental protocol, number of measurements, and data calculation) (9 ) by repeatedly assaying 2 plasma samples with BNP concentrations of 52.6 and 1095 ng/L in duplicate on 20 different working days (both in the morning and in the afternoon). For the 2 samples, within-run imprecision was 3.4% and 1.6%, respectively, and total imprecision was 8.4% and 5.9%. After exclusion by statistical analysis of the 17 samples with BNP values ⬍5 ng/L in the TRIAGE assay, we found a close linear relationship (R ⫽ 0.919; n ⫽ 287) between the 2 methods: ACCESS ⫽ 62.5 (14.1) ⫹ 0.932 (0.024) TRIAGE (P ⬍0.0001 for both). The Bland–Altman plot of the same data confirmed that there was a difference between the values measured with the 2 immunoassays; this difference increased proportionally as the measured concentration increased (Fig. 1B) (10 ). There was a slight but significant mean difference between the values [mean (SD) difference for TRIAGE ⫺ ACCESS, ⫺45.2 (217.8) ng/L (minimum– maximum, ⫺1152 to 1667 ng/L; P ⬍0.0001); mean % difference, ⫺0.2% (minimum–maximum, ⫺2.0% to 1.7%)]. All 91 healthy persons were nonobese, normotensive, and free from acute or chronic diseases. Moreover, all had values within the reference intervals for creatinine, urea, glucose, uric acid, albumin, electrolytes, cardiac markers, and hemoglobin as well as for erythrocyte and leukocyte

Table 1. Characteristics of the heart failure patients enrolled for the evaluation of diagnostic accuracy. NYHA class All

I-II

III-IV

No. of patients 193 122 71 Mean (SD) age, years 64 (13) 62 (13) 67 (12) F/M, n 45/148 28/94 16/55 Mean (SD) BMI,a kg/m2 26 (4) 27 (2) 25 (3) Idiopathic/postischemic/ 49/36/15 57/26/17 37/51/12 secondary cardiomyopathy, % Mean (SD) LVEF, % 34 (10) 37 (11) 27 (7) Diabetes, % 14 11 18 Treatment, % Frusemide 84 36 75 Beta-blockers 71 64 79 ACE inhibitor 78 79 77 ARB 7 4 12 Spironolactone 32 16 59 a BMI, body mass index; LVEF, left ventricular ejection fraction; ACE, angiotensinogen-converting enzyme; ARB, angiotensin receptor blocker.

1276

Technical Briefs

counts and basic urine analysis. All were asymptomatic and underwent a complete examination by a cardiologist, including a visit, standard 12-lead electrocardiogram, Doppler echocardiographic examination, and a bicycle stress test (when older than 50 years), which excluded silent heart disease. The median ACCESS and TRIAGE BNP concentrations were 10.5 ng/L (range, 1–53 ng/L; 97.5th percentile, 45 ng/L) and 8.9 ng/L (range ⬍5 to 52 ng/L; 97.5th percentile, 35 ng/L), respectively. The distribution of BNP values measured by the 2 methods in the healthy group is shown in Fig. 1 of the Data Supplement that accompanies this Technical Brief at http://www. clinchem.org/content/vol51/issue7/. To evaluate the diagnostic accuracy of the 2 methods, we measured only samples from the 193 patients with a clinically ascertained diagnosis of heart failure. The clinical characteristics of these patients are reported in Table 1. Cardiac morphology and function were assessed by 2-dimensional echocardiography, or cardiac catheterization when needed. The median (range) ACCESS and TRIAGE BNP concentrations measured were 171 (2– 4833) ng/L and 118 (⬍5 to 4930) ng/L, respectively. The distributions of BNP values measured by the 2 methods in these patients grouped according to the severity of heart failure [New York Heart Association (NYHA) functional classes I to IV] are shown in Figs. 1 and 2 of the online Data Supplement. However, the results obtained with both the TRIAGE and ACCESS indicated a significant difference between the mean BNP values in healthy persons vs heart failure patients (P ⬍0.0001 by Scheffe` test after ANOVA). We used ROC curve analysis to evaluate the diagnostic accuracy of the 2 methods in differentiating between healthy persons and patients with heart failure. We found no difference in diagnostic accuracy between the 2 methods for differentiating healthy persons from patients with mild (NYHA class I and II; n ⫽ 122; P ⫽ 0.196) or severe (NYHA class III and IV; n ⫽ 71; P ⫽ 0.697) heart failure. For the TRIAGE, the areas under the curves were 0.840 (SE, 0.027; 95% CI, 0.788 – 0.893) for patients with mild disease and 0.998 (SE, 0.002; 95% CI, 0.995–1.000) for patients with severe heart failure, whereas for the ACCESS system, the areas under the curves were 0.870 (SE, 0.023; 95% CI, 0.825– 0.916) for patients with mild disease and 0.997 (SE, 0.002; 95% CI, 0.993–1.000) for patients with severe heart failure. The BNP values corresponding to a sensitivity of 95% in differentiating healthy persons from patients with mild heart failure were 7.5 ng/L (corresponding specificity, 40%) for the ACCESS and 5.1 ng/L for the TRIAGE (corresponding specificity, 29%). The BNP values corresponding to a specificity of 95% were 41 ng/L (corresponding sensitivity, 65%) for the ACCESS and 29 ng/L (corresponding sensitivity, 63%) for the TRIAGE (Fig. 3 of the online Data Supplement). However, it is important to emphasize that the 95% sensitivity with the TRIAGE was obtained at a cutoff near the assay detection limit. The decision cutoff values from the current study are strictly related to our specific clinical setting, comparing healthy persons and patients with clinically ascertained heart failure. In routine clinical

practice, several groups of individuals/patients suspected of having a specific disease are usually compared. The present data confirm and extend previous results suggesting that BNP results are method dependent and that a single predefined common cutoff value cannot be used (5, 6 ). Furthermore, we demonstrated that immunoassays that use the same antibodies and calibration materials do not automatically give the same results. However, the cutoff and reference values of these 2 methods are similar. References 1. Clerico A, Emdin M. Diagnostic accuracy and prognostic relevance of the measurement of the cardiac natriuretic peptides: a review [Review]. Clin Chem 2004;50:33–50. 2. Galvani M, Ferrini D, Ottani T. Natriuretic peptides for risk stratification of patients with acute coronary syndromes [Review]. Eur J Heart Fail 2004;6: 327–33. 3. Doust JA, Glasziou PP, Pietrzak E, Dobson AJ. A systematic review of the diagnostic accuracy of natriuretic peptides for heart failure. Arch Intern Med 2004;164:1978 – 84. 4. Clerico A, Del Ry S, Giannessi D. Measurement of cardiac natriuretic hormones (atrial natriuretic peptide, brain natriuretic peptide, and related peptides) in clinical practice: the need for a new generation of immunoassay methods [Review]. Clin Chem 2000;46:1529 –34. 5. Clerico A, Prontera C, Emdin M, Passino C, Storti S, Poletti R, et al. Analytical performance and diagnostic accuracy of immunometric assays for the measurement of plasma B-type natriuretic peptide (BNP) and N-terminal proBNP. Clin Chem 2005;51:445–7. 6. Rawlins ML, Owen WE, Roberts WL. Performance characteristics of four automated natriuretic peptide assays. Am J Clin Pathol 2005;123:439 – 45. 7. Apple FS, Panteghini M, Ravkilde J, Mair J, Wu AHB, Tate J, et al., on behalf of the Committee on Standardization of Markers of Cardiac Damage of the IFCC. Quality specifications for B-type natriuretic peptide assays. Clin Chem 2005;51:486 –93. 8. Storti S, Prontera C, Emdin M, Passino C, Prati P, Fontani G, et al. Analytical performance and clinical results of a fully automated MEIA system for BNP assay: comparison with a POCT method. Clin Chem Lab Med 2004;42: 1178 – 85. 9. National Committee for Clinical Laboratory Standards. Evaluation of precision performances of clinical chemistry devices: approved guideline. NCCLS Document EP5-A, Vol. 19, No 2. Wayne, PA: NCCLS, February 1999:1– 43. 10. Bland JM, Altman DG. Applying the right statistics: analysis of measurement studies. Ultrasound Obstet Gynecol 2003;22:85–93. DOI: 10.1373/clinchem.2005.048496

Invasive Trophoblast Antigen (Hyperglycosylated Human Chorionic Gonadotropin) as a First-Trimester Serum Marker for Down Syndrome, Martin J.N. Weinans,1* Ulrich Sancken,2 Raj Pandian,3 Jody M.W. van de Ouweland,4 Henk W.A. de Bruijn,4 Jozien P. Holm,1 and Albert Mantingh1 (1 Antenatal Diagnosis Unit, Department of Obstetrics and Gynaecology, and 4 Department of Pathology and Laboratory Medicine, University Hospital, Groningen, The Netherlands; 2 Institut fu¨r Humangenetik der Universita¨t Go¨ttingen, Go¨ttingen, Germany; 3 Quest Diagnostics Nichols Institute, San Juan Capistrano, CA; * address correspondence to this author at: Department of Obstetrics and Gynaecology, University Hospital, PO Box 30.001, 9700 RB Groningen, The Netherlands; fax 31503611806, e-mail [email protected]) Hyperglycosylated human chorionic gonadotropin (hCG) is a variant of hCG with more asparagine (N)-linked

Clinical Chemistry 51, No. 7, 2005

Table 1. Median MoM values, means, and SDs of ITA, free ␤-subunit, and PAPP-A in unaffected (control) and Down syndrome-affected first-trimester pregnancies. Down syndrome-affected pregnancies

Controls

ITA

Median MoM Mean, log MoM SD, log MoM

2.56 0.822 0.489

1.03 0.00 0.701

Free ␤-subunit

Median MoM Mean, log MoM SD, log MoM

2.19 0.82 0.553

0.99 0.00 0.628

PAPP-A

Median MoM Mean, log MoM SD, log MoM

0.35 ⫺1.053 0.572

1.02 0.00 0.673

triantennary carbohydrates and a serine (O)-linked tetrasaccharide core structure in the ␤-subunit of hCG (1 ). Whereas nonhyperglycosylated hCG is secreted by differentiated syncytiotrophoblast cells, hyperglycosylated hCG is secreted solely by invasive cytotrophoblast cells and is therefore also called invasive trophoblast antigen (ITA) (2, 3 ). Before the sixth week of gestation, ITA appears to be the predominant form of hCG (2– 4 ). In Down syndrome pregnancies, differentiation of the cytotrophoblast into a syncytiotrophoblast may be delayed, leading to increased production of ITA (5 ). Urinary ITA is a promising candidate for use as a biochemical marker in Down syndrome screening. In one study conducted during the second trimester of pregnancy, ITA alone detected 78% of the Down syndrome cases at a 5% false-positive rate (6, 7 ). In a setting that simulated routine use, urinary ITA was reported to be the best single marker in the second trimester (8 ). In the first trimester of pregnancy, however, the performance of urinary ITA is lower, with a reported 63% Down syndrome detection rate at a 10% false-positive rate (9 ). ITA is detectable in serum as well as in urine (10 ). Studies have not been performed with serum ITA because of concerns about the stability of ITA in serum, possible loss of ITA when blood is collected in gel barrier tubes, and possible aggregation. A recently developed automated immunochemiluminometric assay measures ITA in various sample types, including serum (11 ). In the present study, we investigated the Down syndrome screening performance of serum ITA before 12 weeks of gestation and compared it with the performance of pregnancyassociated plasma protein A (PAPP-A) and free ␤-subunit in the same sample set. Sera from 24 women with Down syndrome-affected pregnancies and 320 unaffected pregnant women were used in this retrospective study. Samples were collected from women at 9 weeks and 5 days (9 ⫹ 5) of gestation to 11 weeks and 4 days (11 ⫹ 4). These samples were collected between 1999 and 2002, with permission, before chorionic villus sampling. The primary indication for chorionic villus sampling was advanced maternal age (ⱖ36 years). All samples were collected at the Antenatal

1277

Diagnosis Unit of the University Hospital Groningen, The Netherlands, into non– gel-barrier Vacutainer Tubes. An Institutional Review Board-approved protocol was followed. Control samples were matched for gestational age, maternal age, and length of storage. The mean gestational age was 76.4 days for cases and 76.0 days for controls. The mean (SD) maternal age was 38.8 (2.36) years for cases and 37.2 (3.04) years for controls. The median duration of sample storage was 2 years and 1 month for cases and 2 years and 2 months for controls. Blood samples were allowed to clot for 1–3 h at room temperature and were centrifuged at 2500g and 10 °C for 10 min. Serum fractions were frozen immediately and stored at ⫺20 °C and never thawed at room temperature except before being assayed. Serum samples were then thawed overnight at 4 °C and analyzed within 4 h. ITA was measured by an immunochemiluminometric assay on the Nichols Advantage® platform (Nichols Institute Diagnostics) with an acridinium-ester–labeled, antihCG␤ monoclonal antibody (B207) and a biotinylated ITA-specific monoclonal capture antibody (B152) (11 ). The assay has a calibration range up to 300 ␮g/L, with automatic dilution at higher concentrations. The assay has ⬍1% cross-reactivity with recombinant hCG (11 ). The reported intra- and interassay variations (as CV) are ⬍8% and 12%, respectively. PAPP-A and free ␤-subunit were both measured by a fluoroimmunoassay (AutoDELFIA® PAPP-A and Free hCG␤ reagent sets; Perkin-Elmer). The detection limits of the assays are 5 mIU/L for PAPP-A and 0.2 ␮g/L for free ␤-subunit. The intra- and interassay variations (CVs) for PAPP-A were ⬍2% and 4%, respectively, at a concentration of 1500 mIU/L. For free ␤-subunit, the CVs were ⬍4% and ⬍5% at 40 ␮g/L. We used STATISTICA for Windows, Ver. 6 (StatSoft Inc). Multivariate discriminant analysis was performed to calculate the risks. The multiples of the median (MoM) were derived from regressed medians, with gestational days used as the independent variable. All parametric statistical procedures were based on the natural logarithms of the concentrations or MoM values. A Monte Carlo model was applied to adapt detection and falsepositive rates to the present age-standardized population of The Netherlands. Each patient was assigned the mean of 10 randomized maternal age-related risks where the randomization was based on the proportion of maternal age rates according to the present birth frequencies in The Netherlands. Means, SD values based on logarithmic MoM values, and median MoM values for all 3 markers are summarized in Table 1. The squared Mahalanobis distance was 1.42 for ITA, 1.73 for free ␤-subunit, and 2.50 for PAPP-A. The predicted detection rates for Down syndrome for the combination of ITA and PAPP-A at 3%, 5%, and 10% false-positive rates were 62%, 71%, and 83%, respectively. For the combination of free ␤-subunit and PAPP-A, the predicted detection rates were 58%, 75%, and 79%, respectively.

1278

Technical Briefs

The associations among ITA, PAPP-A, and free ␤-subunit concentrations were calculated in both affected and control pregnancies. In controls, there was a significant correlation between ITA and free ␤-subunit. The Pearson correlation coefficients (r) between the log-transformed MoM values were as follows: ITA/free ␤-subunit, 0.63; ITA/PAPP-A, 0.15; and PAPP-A/free ␤-subunit, 0.27. The ROC curves for both ITA and free ␤-subunit in combination with PAPP-A, including maternal age after modeling against the age-standardized population of The Netherlands in 2002 (12 ), are shown in Fig. 1. From these curves, the predicted detection rate for a given screenpositive rate can be determined. We have shown that serum ITA is a useful firsttrimester marker for Down syndrome screening. In the present study the combination of PAPP-A and ITA detected 71% of the Down syndrome cases at a 5% falsepositive rate. The predicted detection rate for the combination PAPP-A/ITA, including maternal age after modeling against the age-standardized population of The Netherlands, was 63% at a 5% false-positive rate (Fig. 1). PAPP-A was the most powerful biochemical marker in this study, as evidenced by the highest Mahalanobis distance. Although the median MoM value in Down syndrome cases was slightly higher for ITA (2.6 MoM) than for free ␤-subunit (2.2 MoM), ITA was not a better marker than free ␤-subunit, perhaps because there was less variation in free ␤-subunit concentrations in the control population (see the SDs in Table 1). The higher Mahalanobis distance for free ␤-subunit (1.7) compared with ITA (1.4) indicates that free ␤-subunit was a slightly better marker in this study. The ROC curves (Fig. 1) indicate that at false-positive rates of 2%–7.5%, the PAPP-A/free ␤-subunit combination outperforms the PAPP-A/ITA combination, whereas

at false-positive rates ⬍2% and between 7.5% and 10%, the PAPP-A/ITA combination performed better. Because the gestational age window in the present study was only 2 weeks (9 ⫹ 5 to 11 ⫹ 4) and the sample size was small, we do not know whether ITA is a better marker than free ␤-subunit during the very early first trimester (⬃9 weeks of gestation). In late first trimester (12–14 weeks of gestation), serum ITA might be a better marker than free ␤-subunit because urinary ITA concentrations in affected pregnancies have been shown to be very high at those gestational ages (13 ). Moreover, because the correlation between ITA and PAPP-A is less than that of free ␤-subunit and ITA (i.e., ITA is more independent of PAPP-A), an ITA/PAPP-A combination may be a more effective screen than the free ␤-subunit/ PAPP-A combination. ITA should not be added as a third marker (i.e., added to free ␤-subunit and PAPP-A) because of the high correlation between ITA and free ␤-subunit in unaffected pregnancies. The results of this study are comparable to the early first-trimester (10 –11 weeks of gestation) study of ITA in maternal urine samples (9 ). Because urinary ITA studies during the second trimester of pregnancy show a greater discriminatory power (i.e., 78% detection at 5% falsepositive rate) (6, 7 ), it is expected that serum ITA will also have a higher Down syndrome detection rate in the second trimester. Currently, studies are in progress to establish the role of ITA in the second trimester of pregnancy as a serum marker for Down syndrome.

We gratefully acknowledge Robert de Vrij and Jurjen IJlstra for excellent technical assistance. We also thank Nichols Institute Diagnostics for providing ITA reagents. References

Fig. 1. ROC curves showing the utility of PAPP-A/ITA (solid line) and PAPP-A/free ␤-subunit (dashed line) combinations after modeling against the age-standardized population of The Netherlands (2002).

1. Elliott MM, Kardana A, Lustbader JW, Cole LA. Carbohydrate and peptide structure of the ␣- and ␤-subunits of human chorionic gonadotropin from normal and aberrant pregnancy and choriocarcinoma. Endocrine 1997;7: 15–32. 2. Kovalevskaya G, Genbacev O, Fisher SJ, Caceres E, O’Connor JF. Trophoblast origin of hCG isoforms: cytotrophoblasts are the primary source of choriocarcinoma-like hCG. Mol Cell Endocrinol 2002;194:147–55. 3. Cole LA, Khanlian SA, Sutton JM, Davies S, Stephens ND. Hyperglycosylated hCG (invasive trophoblast antigen, ITA) a key antigen for early pregnancy detection. Clin Biochem 2003;36:647–55. 4. Kovalevskaya G, Birken S, Kakuma T, Ozaki N, Sauer M, Lindheim S, et al. Differential expression of human chorionic gonadotropin (hCG) glycosylation isoforms in failing and continuing pregnancies: preliminary characterization of hyperglycosylated hCG epitope. J Endocrinol 2002;172:497–506. 5. Frendo JL, Vidaud M, Guibourdenche J, Luton D, Muller F, Bellet D, et al. Defect of villous cytotrophoblast differentiation into syncytiotrophoblast in Down’s syndrome. J Clin Endocrinol Metab 2000;85:3700 –7. 6. Cole LA, Shahabi S, Oz UA, Bahado-Singh RO, Mahoney MJ. Hyperglycosylated human chorionic gonadotropin (invasive trophoblast antigen) immunoassay: a new basis for gestational Down syndrome screening. Clin Chem 1999;45:2109 –19. 7. Cole LA, Shahabi S, Oz UA, Rinne KM, Omrani A, Bahado-Singh RO, et al. Urinary screening tests for fetal Down syndrome: II. Hyperglycosylated hCG. Prenat Diagn 1999;19:351–9. 8. Palomaki GE, Knight GJ, Roberson MM, Cunningham GC, Lee JC, Strom CM, et al. Invasive trophoblast antigen (hyperglycosylated human chorionic gonadotropin) in second-trimester maternal urine as a marker for Down syndrome: preliminary results of an observational study on fresh samples. Clin Chem 2004;50:182–9. 9. Weinans MJN, Butler SA, Mantingh A, Cole LA. Urinary hyperglycosylated

Clinical Chemistry 51, No. 7, 2005

10.

11.

12. 13.

hCG in first trimester screening for chromosomal abnormalities. Prenat Diagn 2000;20:976 – 8. Shahabi S, Oz UA, Bahado-Singh RO, Mahoney MJ, Omrani A, Baumgarten A, et al. Serum hyperglycosylated hCG: a potential screening test for fetal Down syndrome. Prenat Diagn 1999;19:488 –9. Pandian R, Lu J, Ossolinska-Plewnia J. Fully automated chemiluminometric assay for hyperglycosylated human chorionic gonadotropin (invasive trophoblast antigen). Clin Chem 2003;49:808 –10. Central Bureau of Statistics. Voorburg/Heerlen, The Netherlands: CBS, 2004. Cole LA, Omrani A, Cermik D, Bahado-Singh RO, Mahoney MJ. Hyperglycosylated hCG, a potential alternative to hCG in Down syndrome screening. Prenat Diagn 1998;18:926 –33. DOI: 10.1373/clinchem.2005.048751

Validating a Rapid Method for Detecting Common Polymorphisms in the APOA5 Gene by Melting Curve Analysis Using LightTyper, Francesc France´s,* Dolores Corella, Jose´ Vicente Sorlı´, Marisa Guille´n, Jose´ I. Gonza´lez, and Olga Portole´s (Genetic and Molecular Epidemiology Unit, Department of Preventive Medicine, School of Medicine, University of Valencia, Valencia, Spain; * address correspondence to this author at: Department of Preventive Medicine, School of Medicine, Avda. Blasco Iban˜ez, 15, 46010 Valencia, Spain; fax 34-963864166, e-mail francesc. [email protected]) The recently identified apolipoprotein A-V gene (APOA5) has been shown to play an important role in hypertriglyceridemia (1 ). Genetic variation in APOA5 has been consistently associated with plasma triglyceride concentrations in several studies (2– 4 ). Moreover, some studies have demonstrated additional associations with lipoprotein subclasses, remnant-like particles, and cardiovascular disease risk (4 – 6 ). Several single-nucleotide polymorphisms (SNPs) in the human APOA5 gene have been detected with differing frequencies depending on the population analyzed (7, 8 ), and Klos et al. (7 ) have also suggested context-dependent associations in different populations. Overall, 5 common SNPs, ⫺1131T⬎C, ⫺3A⬎G, 56C⬎G, IVS3 ⫹ 476G⬎A, and 1259T⬎C, have been widely reported. Apart from the 56C⬎G SNP, the other SNPs are in strong linkage dysequilibrium, producing 3 haplotypes (11111, 22122, and 11211) representing ⬃98% of the population in Caucasians (5, 9 ); therefore, 2 independent APOA5 SNPs (56C⬎G and ⫺1131T⬎C) can be analyzed in association studies as indicators of the corresponding haplotypes. The former consists of a nonsynonymous substitution, changing codon 19 from serine to tryptophan (S19W), whereas the latter is a T-to-C substitution 1131 nucleotides upstream of the initiation codon. To date, in most published reports, these SNPs have been genotyped by PCR with restriction fragment length polymorphism (RFLP) analysis (3 ). A system for high-throughput genotyping using fluorescence melting curve analysis, the LightTyperTM (Roche), has recently become commercially available. This instrument (10 ) offers higher throughput than the original

1279

LightCyclerTM (11 ). The LightTyper is designed explicitly for melting curve analysis to perform rapid, straightforward, reliable allelic discrimination. The system, which uses 384-well plates, provides postamplification genotyping within 10 –15 min and performs genotyping automatically. A variety of probe chemistries are compatible for genotyping, including single-labeled probes and fluorescence resonance energy transfer probes (10 ). SimpleProbesTM are designed to specifically hybridize to a target sequence that contains the SNP of interest (12 ). Once hybridized, the SimpleProbe emits a larger fluorescent signal than when it is not hybridized to its target sequence. SimpleProbes are more cost-effective than fluorescence resonance energy transfer probes and represent a major advance in decreasing the cost and complexity of SNP analysis (12 ). We report here the validation of an assay based on the LightTyper system and SimpleProbes to rapidly screen for the 56C⬎G and ⫺1131T⬎C SNPs in the APOA5 gene. In the validation study we genotyped 825 randomly selected individuals (age range, 18 –75 years) from the Spanish Mediterranean population. All participants gave informed consent, and the study protocol was approved by the Ethics Committee of the School of Medicine of the University of Valencia. DNA, isolated from blood, was first genotyped for the 56C⬎G SNP in the APOA5 gene by melting curve analysis with the LightTyper; the results were then compared with those obtained with the classic RFLP method. For the melting curve analysis, a 136-bp fragment containing the 56C⬎G SNP was amplified with primers 5⬘-AGAGCCCAGGCCCTGATTA-3⬘ and 5⬘-CATCTTCTGCTGATGGATCTGCT-3⬘ (TIB MOLBIOL) together with the SimpleProbe Flq-TCTCCACAGCGTTTTCGGCC-p (TIB MOLBIOL), where Flq represents a fluorescence quencher. PCR was carried out in 384-well plates with a total volume of 10 ␮L per well in a Thermocycler (Mastercycler-ep380®; Eppendorf). The reaction mixture used in the PCR consisted of 40 ng of genomic DNA, 0.2 ␮L of each primer (10 ␮M), 1 ␮L of the SimpleProbe (1.6 ␮M), 5 U of Fast Start Taq Polymerase (Roche Diagnostics GmbH), 1.2 ␮L of 25 mM MgCl2, and 200 ␮M each deoxynucleotide triphosphate (Roche Diagnostics). After an initial denaturation at 94 °C, 34 PCR cycles were performed with 30 s of denaturation at 94 °C, 45 s of annealing at 55 °C, and 45 s of extension at 72 °C; final extension was at 72 °C for 10 min. A final melting cycle was performed on the LightTyper by heating to 85 °C and cooling to 40 °C at a ramping rate of 0.2 °C/s, with fluorescence data collected continuously. The assay exploits the thermal properties of DNA, i.e., the melting temperature (Tm). We designed the probe to match the wild-type DNA (56C allele), so that the wild-type DNA is 100% complementary to the fluorescent probe, making this complex more stable and thus giving it a higher Tm (65.5 °C). The presence of the mutation gives a lower Tm (56 °C). This difference in Tm allows genotypes to be assigned. The fluorescence signal (F) is plotted in real time against the temperature (T) to generate melting curves for

1280

Technical Briefs

Fig. 1. Genotyping of the 56C⬎G (A) and ⫺1131T⬎C (B) SNPs in the APOA5 gene by melting curve analysis with the LightTyper and SimpleProbes. Shown are derivative melting curve plots (see the text for a description of the procedure for determining melting curves). The Tms and the corresponding genotypes are indicated.

each sample. From these curves, melting peaks (Fig. 1) are generated by plotting the negative derivative of F with respect to T against T (⫺dF/dT against T) (13 ). The melting peak for homozygous CC samples occurs at a higher temperature than the melting peak for homozy-

gous GG samples. In CG heterozygotes, both temperature peaks can be detected. The genotyping results for the 825 individuals from the Mediterranean population are shown in Table 1. To validate these results, we subsequently determined the

Clinical Chemistry 51, No. 7, 2005

Table 1. Genotype distribution for the 56C>G polymorphism in the APOA5 gene depending on method of analysis. Genotype

n

Melting curve analysis (LightTyper)

Method of analysis

CC CG GG

723 96a 6

RFLP analysis

CC CG GG

723 96b 6

Combination of melting curve analysis, RFLP analysis, and direct sequencing

CC CG GG

724 95 6

a One individual with the true genotype CC (by direct sequencing) was misclassified as CG by melting curve analysis. This individual was genotyped as CC by the RFLP method. b One individual with the true genotype CC (by direct sequencing) was misclassified as CG by the RFLP method. This individual was genotyped as CC by the melting curve method.

56C⬎G polymorphism in these 825 individuals by RFLP analysis. PCR was carried out in a total volume of 25 ␮L in the same thermocycler under standard conditions. The following forward and reverse primers were used for amplification (3 ): 5⬘-GGCTCTTCTTTCAGGTGGGTCTCCG-3⬘ and 5⬘-GCCTTTCCGTGCCTGGGTGGT-3⬘ (TIB MOLBIOL). This amplification was designed to force a G⬎A (T in the reverse primer, shown underlined), which introduced a TaqI restriction site in the rare allele (3 ). After restriction enzyme digestion with TaqI (Promega) at 65 °C for 2 h, the common C allele gave fragments of 134 and 23 bp, whereas the G allele gave a single 157-bp product. The fragments were separated by electrophoresis on a 4% Metaphor® agarose gel. The PCR conditions consisted of initial denaturing at 96 °C for 5 min, followed by 30 cycles at 96 °C for 30 s, 63 °C for 30 s, and 72 °C for 45 s, with final extension at 72 °C for 10 min. The genotype results obtained are shown in Table 1. Although the genotype distributions of the 825 samples successfully tested by both the LightTyper and RFLP analysis were identical and did not differ from the Hardy– Weinberg expectations (␹2 ⫽1.97; P ⫽ 0.161), we found 2 discrepant samples: 1 typed as GG by the LightTyper and GC by the RFLP method, and 1 typed as GC by the LightTyper and GG by the RFLP method. This represents a concordance rate of 99.75%. These 2 discordant samples were sequenced directly by standard methods (14 ) on an ABI Prism Automated DNA sequencer. Direct sequence analysis confirmed the melting curve analysis results for 1 of the 2 discrepant samples, and in the other sample direct sequencing confirmed the RFLP result. After individually checking the discrepant results, we found that the error in the RFLP method was the result of a partial digestion, which led to a spurious heterozygote result, whereas the error in the LightTyper was the result of background noise for the probe, which generated a “pseudopeak” corresponding to the mutant Tm. How-ever, the small amplitude of the peak and the deviation by 2 °C from the correct Tm suggested its

1281

spurious condition in the second revision of results. Regarding the RFLP method, although standard positive digestion controls were included, when hundred of samples are tested, inhibitory factors present in some samples can decrease the enzyme activity, leading to incorrect results. One possible solution is to generate a internal positive control for each sample, as proposed Danneberg et al. (15 ). Overall, our results indicate that the LightTyper system can successfully detect SNPs with an accuracy similar to that of the RFLP method. The cost in reagents is similar for both methods; therefore, because the LightTyper is quicker, it appears to be a better choice than the standard RFLP method. Having validated the melting curve analysis obtained with LightTyper for high-throughput genotyping under real biomedical laboratory conditions, using the 56C⬎G polymorphism as an example, we also developed an assay to detect the ⫺1131T⬎C polymorphism. We used the LightTyper and SimpleProbe approach with the primers 5⬘-CACATCCCTCTTTATGAAACAAT-3⬘ and 5⬘-GTAGACGGAGTGGGTGTGTCA-3⬘ (TIB MOLBIOL) to amplify a 229-bp fragment. PCR was carried out in 384-well plates in a total volume of 10 ␮L containing 40 ng of genomic DNA, 0.2 ␮L of each primer (10 ␮M), 1.6 ␮M SimpleProbe (Flq-AGGAACTGGAGCGAAAGTAAGATTTp; TIB MOLBIOL), 5 U of Fast Start Taq Polymerase (Roche Diagnostics), 1.75 ␮L of 25 mM MgCl2, and 200 ␮M each deoxynucleotide triphosphate. The PCR conditions were as follows: initial denaturation at 96 °C for 5 min, followed by 40 cycles of 96 °C for 45 s, 53.5 °C for 45 s, and 72 °C for 45 s, with final extension at 72 °C for 10 min. The melting curve analysis was performed as indicated above. The Tms were 63.5 °C for the wild-type allele (T) and 58 °C for the variant allele (C); accordingly, we found 736 individuals who were homozygous TT, 86 who were heterozygous CT, and 3 who were homozygous CC. Carriers of the less common allele were grouped, and a crude association analysis was carried out (results not shown). The 56C⬎G polymorphism was statistically associated (P ⫽ 0.04) with higher mean triglyceride concentrations, but the association for the ⫺1131T⬎C SNP did not reach statistical significance (P ⫽ 0.21). Nevertheless, a more detailed analysis including control for covariates and an increased sample size is needed in this population. In this regard, it is interesting to note that our allele frequencies for the variant alleles (particularly for the ⫺1131T⬎C polymorphism) were slightly lower than those reported in other populations (7, 8 ). Finally, to take a full advantage of this new technique, we also tested several primer concentrations in asymmetric PCR assays and found that forward/reverse primer ratios of 0.26 for 56C⬎G and 0.20 for ⫺1131T⬎C improved PCR efficiency, decreasing the failure rate near to 0%.

We gratefully acknowledge F. Gimenez-Ferna´ndez for laboratory assistance. F.F. is the recipient of a fellowship from the University of Valencia (V Segles). This work was

1282

Technical Briefs

supported by Grants PI02/1096 and PI042239 from the FIS, Instituto de Salud Carlos III, Spain. References 1. Oliva CP, Pisciotta L, Li Volti G, Sambataro MP, Cantafora A, Bellocchio A, et al. Inherited apolipoprotein A-V deficiency in severe hypertriglyceridemia. Arterioscler Thromb Vasc Biol 2005;25:411–7. 2. Pennacchio LA, Olivier M, Hubacek JA, Cohen JC, Cox DR, Fruchart JC, et al. An apolipoprotein influencing triglycerides in humans and mice revealed by comparative sequencing. Science 2001;294:169 –73. 3. Talmud PJ, Hawe E, Martin S, Olivier M, Miller GJ, Rubin EM, et al. Contribution of variation within the APOC3/A4/A5 gene cluster in determining plasma triglycerides. Hum Mol Genet 2002;11:3039 – 46. 4. Austin MA, Talmud PJ, Farin FM, Nickerson DA, Edwards KL, Leonetti D, et al. Association of apolipoprotein A5 variants with LDL particle size and triglyceride in Japanese Americans. Biochim Biophys Acta 2004;1688:1–9. 5. Lai CQ, Demissie S, Cupples LA, Zhu Y, Adiconis X, Parnell LD, et al. Influence of the APOA5 locus on plasma triglyceride, lipoprotein subclasses, and CVD risk in the Framingham Heart Study. J Lipid Res 2004;45:2096 – 105. 6. Bi N, Yan SK, Li GP, Yin ZN, Chen BS. A single nucleotide polymorphism ⫺1131T⬎C in the apolipoprotein A5 gene is associated with an increased risk of coronary artery disease and alters triglyceride metabolism in Chinese. Mol Genet Metab 2004;83:280 – 6. 7. Klos KL, Hamon S, Clark AG, Boerwinkle E, Liu K, Sing CF. The influence of APOA5 polymorphisms on variation in plasma triglycerides in young, healthy African-Americans and Whites of the CARDIA study. J Lipid Res 2004;46: 564 –71. 8. Lai CQ, Tai ES, Tan CE, Cutter J, Chew SK, Zhu YP, et al. The APOA5 locus is a strong determinant of plasma triglyceride concentrations across ethnic groups in Singapore. J Lipid Res 2003;44:2365–73. 9. Pennacchio LA, Olivier M, Hubacek JA, Krauss RM, Rubin EM, Cohen JC. Two independent apolipoprotein A5 haplotypes influence human plasma triglyceride levels. Hum Mol Genet 2002;11:3031– 8. 10. Bennett CD, Campbell MN, Cook CJ, Eyre DJ, Nay LM, Nielsen DR, et al. LightTyper: high-throughput genotyping using fluorescent melting curve analysis. Biotechniques 2003;34:1288 –92. 11. Wittwer CT, Ririe KM, Andrew RV, David DA, Gundry RA, Balis UJ. The LightCyclerTM: a microvolume multisample fluorimeter with rapid temperature control. Biotechniques 1997;22:176 – 81. 12. Wittwer CT, Crockett AO, Caplin BE, Stevenson W, Chen J, Kusukawa N, inventors. Single labeled oligonucleotide probes for homogeneous nucleic acid sequence analysis. US Patent 6,635,427; August 10, 2001. 13. Lay MJ, Wittwer CT. Real time fluorescence genotyping of factor V Leiden during rapid-cycle PCR. Clin Chem 1997;43:2262–7. 14. Guillen M, Corella D, Cabello M, Gonzalez J, Sabater A, Chaves J, et al. Identification of novel SLC3A1 gene mutations in Spanish cystinuria families and association with clinical phenotypes. Clin Genet 2005;67:240 –51. 15. Danneberg J, Abbes AP, Bruggeman BJ, Engel H, Gerrits J, Martens A. Reliable genotyping of the G-20210-A mutation of coagulation factor II (prothrombin). Clin Chem 1998;44:349 –51. DOI: 10.1373/clinchem.2005.049676

Two-Step Genetic Screening of Thrombophilia by Pyrosequencing, Annalisa Verri,1* Federico Focher,2 Guido Tettamanti,3 and Vittorio Grazioli1 (1 Istituto Clinico Humanitas, Rozzano, Italy; 2 Istituto di Genetica Molecolare, Consiglio Nazionale delle Ricerche, Pavia, Italy; 3 Dipartimento di Chimica e Biochimica Medica, Universita` di Milano, Segrate, Italy; * address correspondence to this author at: Istituto Clinico Humanitas, via Manzoni, 56, 20089 Rozzano, Italy; fax 39-02-82244790, e-mail [email protected]) Venous thrombotic events are quite common; they affect ⬃1 in every 1000 persons per year and have a lifetime clinical prevalence of ⬃5%. The pathogenesis of venous

thrombotic events is complex, involving the interaction of acquired risk factors with some genetic predisposition. A wide array of methods and technologies have been used for screening of prothrombotic mutations (1– 4 ). However, it is known that when mutation detection methods other than direct sequencing are used to identify a particular sequence change, there is always some risk that other sequence alterations occurring at the recognition site could lead to allele misclassification. This issue has been discussed, for example, for the silent A1692C polymorphism in the factor V gene, which is erroneously identified as factor V Leiden by restriction enzyme digest detection (5 ). Other genotyping methods could also be affected by adjacent sequence alterations, including allele-specific amplification (6 ), single-strand conformational polymorphism analysis (7 ), oligonucleotide ligation (8 ), heteroduplex analysis (9 ), and methods based on melting curve analysis, in which unexpected results should be clarified definitively by sequencing (10 ). Although DNA sequencing is still considered the “gold standard” for characterizing specific nucleotide alterations and improved technology has made automated DNA sequencing available to the clinical molecular diagnostics laboratory, DNA sequencing remains too expensive and time-consuming for most applications. Recent studies demonstrating the robustness and speed of pyrosequencing technology, as well as its possible use for multiplex genotyping (11 ), have led to its use in an increasing range of genetic research areas (12–15 ). The aim of our work was to establish a multiplex protocol for direct pyrosequencing analysis of a panel of coagulation factors mutations: the 3 single-nucleotide polymorphisms (SNPs) most commonly associated with thrombophilia—G1691A in factor V Leiden, G20210A in factor II, and C677T in methylenetetrahydrofolate reductase (MTHFR)—for a first-tier screening, and 3 additional polymorphisms—A1298C in MTHFR, Val34Leu in factor XIII, and 4G/5G in plasminogen activator inhibitor-1 (PAI-1)—which are believed not to have an independent effect on venous thrombosis but could be investigated in a second-tier screening because they may act synergistically with the previously mentioned factor mutations (16, 17 ) or, in the case of factor XIII Val34Leu, exert a protective effect (18 ). We used pyrosequencing to genotype 100 individuals, previously analyzed by LightCycler (Roche) or by direct sequencing, for all 6 polymorphisms. Pure genomic DNA from EDTA-anticoagulated blood was isolated either by use of the semiautomated Magna Pure instrument with the Magna Pure LC DNA Isolation Kit (Roche) or manually, with the High Pure PCR Template Preparation Kit (Roche). These extraction procedures gave the same yield and PCR performance. The 6 genomic segments containing the SNPs of interest were amplified in triplex PCR reactions with 3 pairs of primers (Eurogentec; see Table 1). PCR conditions were optimized in preliminary experiments in which amplified products were analyzed by electrophoresis on agarose gel (4%) to optimize template

1283

Clinical Chemistry 51, No. 7, 2005

concentration, magnesium concentration, and number of cycles to enhance PCR yield and specificity, which improve the success of the sequencing reaction. Optimal PCR conditions for triplex amplification of factor V Leiden, factor II G20210A, and MTHFR C677T were as follows: 10 ng of pure genomic DNA, 10 pmol of each primer, 200 ␮M each deoxynucleotide triphosphate (dNTP), 3 mM MgCl2 and 0.5 U of HotGoldStar Taq polymerase (Eurogentec) in 20 ␮L (final volume) of the buffer supplied (Eurogentec). Thermal cycling started with 10 min at 95 °C followed by 35 cycles of 95 °C for 30 s, 59 °C for 30 s, and 72 °C for 60 s, with a final step at 72 °C for 5 min. Triplex amplification of factor XIII Val34Leu, PAI-1 4G/5G, and MTHFR A1298C was performed with the following conditions: 10 ng of pure genomic DNA, 4.5 pmol of factor XIII primers, 25 pmol of MTHFR primers, 9 pmol of PAI-1 primers, 200 ␮M each dNTP, 1.5 mM MgCl2, and 0.5 U of HotGoldStar Taq polymerase in 30 ␮L (final volume) of the buffer supplied (Eurogentec). Thermal cycling started with 10 min at 95 °C, followed by 40 cycles of 95 °C for 30 s, 58 °C for 45 s, and 72 °C for 60 s, with a final step at 72 °C for 5 min. At the end of PCR, 20 –30 ␮L of each biotinylated PCR products was immobilized on 3 ␮L of streptavidin-coated Sepharose beads (Amersham Biosciences) to obtain singlestranded DNA suitable for sequencing. The immobilization, denaturation, washing, and primer annealing steps were performed with a vacuum preparation workstation according to the manufacturer’s instructions (Biotage AB). The single-stranded biotinylated PCR products were subjected to a multiplex minisequencing reaction on a PSQ96MA instrument (Biotage AB) to interrogate 3 polymorphic loci simultaneously. The multibase reading capability of pyrosequencing facilitates optimal positioning

of the sequencing primers (Table 1). Sequencing was performed as described previously (19 ). The dispensation order to analyze the 3 sequences at the same time was selected by use of the SNP Entry module of the SNP Analysis Software (Biotage AB). Parallel processing of 96 samples markedly reduced the handling time and pipetting steps needed. The maximum throughput was limited only by the number of PCR reactions because the multisequencing reaction lasted only 23 min for each 96-well microtiter plate. In addition, the results from the 96 completed sequencing reactions were analyzed by the pyrosequencing software in 2 min and were displayed as shown in Fig. 1 and in Fig. 1 of the Data Supplement that accompanies the online version of this Technical Brief at http://www.clinchem.org/content/ vol51/issue7/. Sequences are determined by computer-automated comparison of predicted patterns with raw data. Generally, samples did not require time-consuming manual interpretation. Failure to make a genotype call at the first attempt was infrequent (⬃5%) and was mostly attributable to insufficient signal-to-noise ratios caused by poor PCR amplification. The accuracy, robustness, and reproducibility of the assay were very high: 100% of the results obtained with pyrosequencing (Table 1 of the online Data Supplement) were confirmed by LightCycler or sequencing analysis. In conclusion, this method is rapid and cost-effective when compared with traditional sequencing; it is also suitable for the present challenge of high-throughput SNP genotyping: including all reagents and PCR, the cost per sample was 3 €. This study shows, as has been shown previously, how a technology originally introduced into the field of basic biomedical research can be successfully adapted to the clinical laboratory.

Table 1. Primers used for amplification and sequencing. Mutation

Primer

Sequence, 5ⴕ–3ⴕ

Factor II G20210A

PCR forward PCR reverse Sequencing

Biotin-CCTGAAGAAGTGGATACAGAAGG CAGTAGTATTACTGGCTCTTCCTGA ACTGGGAGCATTGAG

Factor V Leiden (G1691A)

PCR forward PCR reverse Sequencing

GGGCTAATAGGACTACTTCTAATC Biotin-TCTCTTGAAGGAAATGCCCCATTA AGCAGATCCCTGGAC

MTHFR C677T

PCR forward PCR reverse Sequencing

TTGAGGCTGACCTGAAGCAC Biotin-5GTGATGCCCATGTCGGTG GGTGTCTGCGGGAG

Factor XIII Val34Leu

PCR forward PCR reverse Sequencing

AGCAGTTCCACCCAATAACTCT Biotin-TCATACCTTGCAGGTTGACG CACAGTGGAGCTTCAG

MTHFR A1298C

PCR forward PCR reverse Sequencing

GTGGCACTGCCCTCTGTC Biotin-CTCCCGAGAGGTAAAGAACGA AGGAGTTGACCAGTGA

PAI-1 4G/5G

PCR forward PCR reverse Sequencing

Biotin-GGCACAGAGAGAGTCTGGACAC CGCCTCCGATGATACACG ACACGGCTGACTCCC

1284

Technical Briefs

Fig. 1. Results of triplex genotyping given as patterns of peaks called pyrograms (top), and genotype assignment (bottom). (Top), the x axis shows the dispensation order of dNTPs during pyrosequencing; the y axis shows peak heights representing the intensity of the light signal produced, which is proportional to the number of dNTPs incorporated into the DNA templates. (Bottom), genotypes are assigned by computer-automated comparison of peaks with predicted histograms. The individual shown in this figure was heterozygous for MTHFR C677T (u) and homozygous wild type for factor V G1691A ( ) and factor II G20210A (f).

We are grateful to Biosense for supplying the instrument and reagents and, in particular, to Dr. Elvira Meroni for excellent technical support. This work was partially supported by MURST-FIRB Grant RBAU01LSR4_001 (to F.F.).

14.

References

16.

1. Angelini A, Di Febbo C, Baccante G, Di Nisio M, Di Ilio C, Cuccurullo F, et al. Identification of three genetic risk factors for venous thrombosis using a multiplex allele-specific PCR assay: comparison of conventional and new alternative methods for the preparation of DNA from clinical samples. J Thromb Thrombolysis 2003;16:189 –93. 2. Blasczyk R, Ritter M, Thiede C, Wehling J, Hintz G, Neubauer A, et al. Simple and rapid detection of factor V Leiden by allele-specific PCR amplification. Thromb Haemost 1996;75:757–9. 3. Bortolin S, Black M, Modi H, Boszko I, Kobler D, Fieldhouse D, et al. Analytical validation of the Tag-It high-throughput microsphere-based universal array genotyping platform: application to the multiplex detection of a panel of thrombophilia-associated single-nucleotide polymorphisms. Clin Chem 2004;50:2028 –36. 4. Schrijver I, Lay M, Zehnder J. Diagnostic single nucleotide polymorphism analysis of factor V Leiden and prothrombin 20210G⬎A. A comparison of the Nanogen Eelectronic Microarray with restriction enzyme digestion and the Roche LightCycler. Am J Clin Pathol 2003;119:490 – 6. 5. Liebman H, Sutherland D, Bacon R, McGehee W. Evaluation of a tissue factor dependent factor V assay to detect factor V Leiden: demonstration of high sensitivity and specificity for a generally applicable assay for activated protein C resistance. Br J Haematol 1996;95:550 –3. 6. Patnaik M, Dlott JS, Fontaine RN, Subbiah MT, Hessner MJ, Joyner KA, et al. Detection of genomic polymorphisms associated with venous thrombosis using the Invader biplex assay. Mol Diagn 2004;6:137– 44. 7. Corral J, Iniesta J, Gonzalez-Conejero R, Vicente V. Detection of factor V Leiden from a drop of blood by PCR-SSCP. Thromb Haemost 1996;76: 735–7. 8. Zotz R, Maruhn-Debowski B, Scharf R. Mutation in the gene coding for coagulation factor V and resistance to activated protein C: detection of the genetic mutation by oligonucleotide ligation assay using a semi-automated system. Thromb Haemost 1996;76:53–5. 9. Bowen D, Standen G, Granville S, Bowley S, Wood N, Bidwell J. Genetic diagnosis of factor V Leiden using heteroduplex technology. Thromb Haemost 1997;77:119 –22. 10. Lay MJ, Wittwer CT. Real-time fluorescence genotyping of factor V Leiden during rapid-cycle PCR. Clin Chem 1997;43:2262–7. 11. Palmieri O, Toth S, Ferraris A, Andriulli A, Latiano A, Annese V, et al. CARD15 Genotyping in inflammatory bowel disease patients by multiplex pyrosequencing. Clin Chem 2003;49:1675–9. 12. Lindqvist M, Haglund S, Almer S, Peterson C, Taipalensu J, Hertervig E, et al. Identification of two novel sequence variants affecting thiopurine methyltransferase enzyme activity. Pharmacogenetics 2004;14:261–5. 13. Fakhrai-Rad H, Pourmand N, Ronaghi M. Pyrosequencing: an accurate

15.

17.

18. 19.

detection platform for single nucleotide polymorphisms. Hum Mutat 2002; 19:479 – 85. Zhang Z, Liu W, Jia X, Gao Y, Hemminki K, Lindholm B. Use of pyrosequencing to detect clinically relevant polymorphisms of genes in basal cell carcinoma. Clin Chim Acta 2004;342:137– 43. Ronaghi M, Elahi E. Pyrosequencing for microbial typing. J Chromatogr B Analyt Technol Biomed Life Sci 2002;782:67–72. Segui R, Estelles A, Mira Y, Espana F, Villa P, Falco C, et al. PAI-1 promoter 4G/5G genotype as an additional risk factor for venous thrombosis in subjects with genetic thrombophilic defects. Br J Haematol 2000;111:122– 8. Weisberg I, Tran P, Christensen B, Sibani S, Rozen R. A second genetic polymorphism in methylenetetrahydrofolate reductase (MTHFR) associated with decreased enzyme activity. Mol Genet Metab 1998;64:169 –72. Endler G, Mannhalter C. Polymorphisms in coagulation factor genes and their impact on arterial and venous thrombosis. Clin Chim Acta 2003;330:31–55. Ronaghi M, Uhlen M, Nyren P. A sequencing method based on real-time pyrophosphate. Science 1998;281:363–5. DOI: 10.1373/clinchem.2005.048124

Arrayed Primer Extension Resequencing of Mutations in the TP53 Tumor Suppressor Gene: Comparison with Denaturing HPLC and Direct Sequencing, Florence Le Calvez,1 Aune Ahman,2 Neeme Tonisson,2,3 Jeremy Lambert,1 Ste´phane Temam,4 Paul Brennan,1 David G. Zaridze,5 Andres Metspalu,2,3 and Pierre Hainaut1* (1 International Agency for Research on Cancer, Lyon, France; 2 Asper Biotech Ltd., Tartu, Estonia; 3 Institute of Molecular and Cell Biology, University of Tartu/Estonian Biocentre, Tartu, Estonia; 4 Department of Head and Neck Surgery, Institut Gustave-Roussy, Villejuif, France; 5 Institute of Carcinogenesis, Cancer Research Center, Moscow, Russia; * address correspondence to this author at: International Agency for Research on Cancer, 150, Cours Albert Thomas, F-69372 Lyon Cedex 08, France; fax 33472738322, e-mail [email protected]) Mutations of TP53 (17p13.1; OMIM 191170; PubMed accession number X54156) are common in cancers and are

1285

Clinical Chemistry 51, No. 7, 2005

A

Homoduplexes

Heteroduplexes

B

Codon 236, TA C>TG C

⇓ 7

Absorbance (mV)

6

5

4

3

2

1 0

C

1

2

3

4

D

Codon 236, TAC>TGC A

C

G

T

Average signals

Wild-type reference

Sense Antisense

Sense Antisense

A

C

G

T

Fig. 1. Example of detection of TP53 mutation at codon 236 (TAC⬎TGC) in an archived pathology specimen. DNA extracted from a formalin-fixed, paraffin-embedded lung squamous cell carcinoma was analyzed by DHPLC and direct sequencing (A and B) or by APEX (C and D). (A), DHPLC chromatograms of exon 5– 6, showing the superposition of profiles for wild-type (blue) and tumor (red) DNA. (B), direct sequencing of exon 5– 6. A portion of the sense sequence surrounding codon 236, containing an A-to-G mutation at the second base, is shown. (C), detection of a mutation at the second base of codon 236 by APEX. The signals for all 4 fluorescent dinucleotides are shown for the sense and antisense strands. Each dinucleotide is visualized in duplicate. (D), analysis of the signals shown in A by the GENORAMA software. Left, average of the signals for each dinucleotide on the sense and antisense strands, expressed as a percentage of the highest signal detected at each position. Right, average of wild-type signals at each position. Comparison of the panels indicates the presence of a mutant, G:C, base pair against a background of the wild-type, A:T, base pair.

typically missense within exons 4 –9, impairing the capacity of p53 to transactivate genes involved in cell cycle arrest, apoptosis, and DNA repair (1 ). Functionally, mutations may differ according to their nature and position, as well as to the presence of a common polymorphism at codon 72 (arginine or a proline) in the mutant allele (2 ). Knowing TP53 mutation status has potential applications for cancer prognosis (3, 4 ) and early diagnosis (5 ), identification of mutagen “fingerprints” (1, 6 ), and prediction of therapeutic outcomes (7, 8 ). To achieve this purpose, sensitive, fast, and cost-effective methods are needed to Table 1. DHPLC and APEX detection limits for TP53 mutations in 6 cell lines. Detection limit, % mutant DNA Cell line

Exon

Codon

Mutation

DHPLC

APEX

Hs578T T47D TE11 TE6 TE1 MDA-MB 231

5 6 7 7 8 9

157 194 237 248 272 280

GTC⬎TTC CTT⬎TTT ATG⬎ATT CGG⬎CAG GTG⬎ATG AGA⬎AAA

12.5 3.125 3.125 6.25 3.125 6.25

6.25 3.125 6.25 3.125 6.25 6.25

assess the whole coding sequence plus exon/intron boundaries. Current approaches are based on mutation prescreening with single strand conformational polymorphism analysis, temporal temperature gradient electrophoresis, or denaturing HPLC (DHPLC) combined with direct sequencing of relevant PCR fragments [reviewed in Ref. (9 )]. These methods are labor-intensive, difficult to standardize, and in some cases, of limited sensitivity. In recent years, 2 microarray methods for resequencing TP53 have been described: the Affymetrix p53 GeneChip array, described elsewhere (10, 11 ), and the Arrayed Primer Extension (APEX), based on incorporation of 4 dye terminators into oligonucleotide primers that each identify a base in the target sequence (12 ). In 2002, we described an APEX array for resequencing TP53 exons 2–9, which contain 95% of known mutations in TP53 (13 ). Here we compare the sensitivity and detection limits of APEX with a standard method, DHPLC/direct sequencing, and discuss the potential of APEX for application to cancer diagnostic or prognostic purposes. Specimens in the comparison set included 6 cell lines with mutations in different TP53 exons (see Table 1 of the Data Supplement that accompanies the online version of this Technical Brief at http://www.clinchem.org/content/

1286

Technical Briefs

vol51/issue7/) and 60 tumor samples with TP53 mutations identified by DHPLC (exons 4 –9) and direct sequencing (29 from paraffin-embedded lung cancers and 31 frozen specimens of oral cancers). The persons performing the APEX screening (exons 2–9) were blinded to mutation status. Mutations in these specimens were representative of the diversity of TP53 mutations in human cancers. Four specimens positive by DHPLC but with no mutation found by sequencing were also tested. DNA extractions, PCR, DHPLC, and sequencing were performed according to described protocols (see Tables 2 and 3 of the online Data Supplement) (14 ). APEX probes were 25mer oligonucleotides with 12carbon amino linkers at their 5⬘ end covering the sense and antisense wild-type TP53 sequence (Genset) and were spotted on 24 ⫻ 60 mm SAL-type aminosilane/ phenylene diisothiocyanate-coated microarray slides (15 ). PCR products were concentrated and purified by use of the Jet quick PCR purification Spin Kit (Genomed). Hybridization and primer extension reactions were performed as published previously (13 ). Data were analyzed with GenoramaTM 4.2 genotyping software, using clustered signal patterns from 20 wild-type TP53 DNA sequences as statistical reference. For each position, the software calculates the distance (difference) between the sample signal and the signal pattern of the wild-type reference cluster database [algorithm described by Tonisson et al. (13 )]. The distance value is used as a measure for calling the given base. Zero distance indicates a perfect match between the given base and the wild-type reference. The default cutoff distance value of the software with cluster analysis is 30 but can be adapted by the user. A representative analysis of a TP53 exon 7 mutation at codon 236 (TAC⬎TGC) is shown in Fig. 1. To evaluate APEX detection limits, defined amounts of PCR products of DNA from 6 cell lines homo- or hemizygous for known mutations were mixed with wild-type PCR products and analyzed in parallel by DHPLC and APEX. An example of DHPLC and APEX detection limits is given in Fig. 1 of the online Data Supplement. With DHPLC, detection limits showed variations depending on mutation type and mutation sequence context (Table 1). With APEX, all mutations were detected with detection limits of 3.125%– 6.25%, consistent values reported previously for codons 273 (CGT⬎CAT) and 248 (CGG⬎TGG) (13 ). In comparison, direct sequencing does not detect most mutations in samples containing less than 25%–30% mutated DNA (16, 17 ). Of 66 TP53 mutations identified by DHPLC/ sequencing in 60 tumors, APEX identified 62 but failed to detect 4 mutations (sensitivity, 94%; see Table 4 of the online Data Supplement). All mutations occurred between exons 4 and 9. Among substitutions, 49 of 52 (94.2%) were correctly identified by APEX. Two of the mutations that escaped APEX detection (codon 91, TGG⬎TGA and codon 174 AGG⬎TGG) corresponded to positions generating background signals in one DNA strand. The third one (codon 173 GTG⬎ATG) was identified on the antisense strand only because of inefficient

hybridization at this position. APEX specificity and sensitivity may be further improved by experimentally establishing cutoff values for mutants at each position of the APEX array or by modifying probe design to minimize the formation of secondary structures interfering with DNA hybridization. The 14 other mutations detected by DHPLC/sequencing were deletions or insertions. APEX correctly identified 1 insertion (⫹ 1 bp; codons 62– 63) and detected the base change at one extremity in 12 other cases (85.7%), including 4 insertions ranging from 1 to 8 bp, 7 deletions ranging from 1 to 4 bp, and 1 rearrangement at codon 184 that was not clearly identified as an insertion or deletion by sequencing. However, in these 12 cases, APEX did not allow complete identification of the change in DNA sequence. Overall, APEX had a sensitivity of 93% in detecting insertion/deletions, but only 7% in fully identifying their positions and sizes. These limitations are intrinsic to APEX design, which is based on hybridization of wildtype sequences to complementary, immobilized oligonucleotides (18 ). In theory, the presence of a deletion should lead to incorporation of an unexpected nucleotide at the 5⬘ border of the deletion and corresponding to the first nucleotide flanking the 3⬘ border of the deletion. Presence of an insertion should generate an unexpected signal corresponding to the first inserted base at the 5⬘ border of the insertion. However, these events will be detectable only if the signals they generate differ from expected wild-type signals at these positions. Assuming equal distribution of insertions and substitutions at all base positions, there is a 1 in 4 chance for either strand that the border may be undetectable and a 1 in 16 chance that the sequence alteration escapes detection. However, in practice, the odds for nondetection are higher because insertions and deletions often occur within singlebase repeats, making it likely that the modified signal may be identical to the expected one. The nonrandom distribution of insertions/deletions in TP53 may explain the relatively poor performance of APEX in mapping both borders of insertions/deletions. Despite these limitations, APEX compares well with the Affymetrix GeneChip p53 in detecting deletions and insertions (11, 17, 19, 20 ). Indeed, the latter does not allow detection of sequence alterations unless representative mutant oligonucleotides have been spotted on the array (10, 11 ). Of the 4 specimens positive by DHPLC but negative by sequencing, APEX detected base changes in 3, including 1 specimen with 2 mutations (a 1-bp substitution in intron 5 and a silent mutation at codon 147; see Table 4 of the online Data Supplement). Finally, in 5 specimens, APEX generated several signals (2 or 3 per specimen), only one of which was confirmed by sequencing (see Table 4 of the online Data Supplement). The unconfirmed signals by DHPLC/sequencing were detected only on 1 DNA strand by APEX, suggestive of an insertion or deletion as discussed above. Thus, none of these discordant changes were single-base substitutions. We interpreted the discordant signals as “false-positive” APEX signals, which gave

Clinical Chemistry 51, No. 7, 2005

an overall positive predictive value for APEX of 92.5% (62 of 67). We found that both fresh-frozen and formalin-fixed, paraffin-embedded tissues are suitable for mutation detection by APEX. The positive predictive values for APEX were 92% (33 of 36) and 93.5% (29 of 31) for frozen and paraffin-embedded tissues, respectively. Thus, the use of paraffin archives did not lead to a higher number of ambiguous or artifactual signals. The second base of codon 72 harbors a common polymorphism in TP53 that affects a BstUI digestion site. The concordance between APEX and either restriction digestion or DHPLC/sequencing data was perfect for all 64 specimens analyzed (see Fig. 2 of the online Data Supplement). This polymorphism may play a role in cancer susceptibility (21, 22 ) and response to therapy (2 ). Recently, Bonafe et al. (23 ) reported that preferential loss of the codon 72P allele in breast tumors of heterozygous patients was associated with a significant decrease in disease-free and overall survival. Additional studies have shown an improved response to chemo-radiotherapy and a longer overall survival in patients whose squamous cell carcinomas tumor retained a wild-type 72R allele rather than a wild-type 72P allele (24 ). Thus, genotyping of TP53 at codon 72, combined with mutation detection, may be relevant for the therapeutic management of cancer. APEX also genotyped 2 other, less frequent polymorphisms at codons 213 (CGA/CGG) and 36 (CCG/CCA) for which concordance between DHPLC and APEX was also perfect. In conclusion, APEX offers a flexible, sensitive, and low-cost resequencing alternative for large-scale studies involving retrospective analysis of pathology collections as well as for application to studies for which fresh-frozen materials are available.

5. 6.

7.

8.

9.

10. 11.

12.

13.

14.

15.

16.

17.

18.

Florence Le Calvez was supported by Special Training Awards from the International Agency for Research on Cancer; Neeme Tonisson was supported as Junior Fellow by Wellcome Trust International Senior Research Grant for Central Europe no. 070191/Z/03/Z and Andres Metspalu by targeted funding from EMER (no. 0518). This research project was supported by European Community (EC) FP6 funding. This publication reflects the authors’ views and not necessarily those of the EC. The Community is not liable for any use that may be made of the information contained herein.

19.

20.

21.

22.

23.

References 1. Hainaut P, Hollstein M. p53 and human cancer: the first ten thousand mutations. Adv Cancer Res 2000;77:81–137. 2. Bergamaschi D, Gasco M, Hiller L, Sullivan A, Syed N, Trigiante G, et al. p53 polymorphism influences response in cancer chemotherapy via modulation of p73-dependent apoptosis. Cancer Cell 2003;3:387– 402. 3. Schneider PM, Stoeltzing O, Roth JA, Hoelscher AH, Wegerer S, Mizumoto S, et al. P53 mutational status improves estimation of prognosis in patients with curatively resected adenocarcinoma in Barrett’s esophagus. Clin Cancer Res 2000;6:3153– 8. 4. Samowitz WS, Curtin K, Ma KN, Edwards S, Schaffer D, Leppert MF, et al.

24.

1287

Prognostic significance of p53 mutations in colon cancer at the population level. Int J Cancer 2002;99:597– 602. Sidransky D. Emerging molecular markers of cancer. Nat Rev Cancer 2002;2:210 –9. Greenblatt MS, Bennett WP, Hollstein M, Harris CC. Mutations in the p53 tumor suppressor gene: clues to cancer etiology and molecular pathogenesis. Cancer Res 1994;54:4855–78. Borresen AL, Andersen TI, Eyfjord JE, Cornelis RS, Thorlacius S, Borg A, et al. TP53 mutations and breast cancer prognosis: particularly poor survival rates for cases with mutations in the zinc-binding domains. Genes Chromosomes Cancer 1995;14:71–5. Geisler S, Lonning PE, Aas T, Johnsen H, Fluge O, Haugen DF, et al. Influence of TP53 gene alterations and c-erbB-2 expression on the response to treatment with doxorubicin in locally advanced breast cancer. Cancer Res 2001;61:2505–12. Kristensen VN, Kelefiotis D, Kristensen T, Borresen-Dale AL. High-throughput methods for detection of genetic variation. Biotechniques 2001;30: 318 –22, 324, 326. Hacia JG. Resequencing and mutational analysis using oligonucleotide microarrays. Nat Genet 1999;21:42–7. Wikman FP, Lu ML, Thykjaer T, Olesen SH, Andersen LD, Cordon-Cardo C, et al. Evaluation of the performance of a p53 sequencing microarray chip using 140 previously sequenced bladder tumor samples. Clin Chem 2000;46: 1555– 61. Kurg A, Tonisson N, Georgiou I, Shumaker J, Tollett J, Metspalu A. Arrayed primer extension: solid-phase four-color DNA resequencing and mutation detection technology. Genet Test 2000;4:1–7. Tonisson N, Zernant J, Kurg A, Pavel H, Slavin G, Roomere H, et al. Evaluating the arrayed primer extension resequencing assay of TP53 tumor suppressor gene. Proc Natl Acad Sci U S A 2002;99:5503– 8. Dai M, Clifford GM, Le Calvez F, Castellsague X, Snijders PJ, Pawlita M, et al. Human papillomavirus type 16 and TP53 mutation in oral cancer: matched analysis of the IARC multicenter study. Cancer Res 2004;64:468 – 71. Guo Z, Guilfoyle RA, Thiel AJ, Wang R, Smith LM. Direct fluorescence analysis of genetic polymorphisms by hybridization with oligonucleotide arrays on glass supports. Nucleic Acids Res 1994;22:5456 – 65. Rosenblum BB, Lee LG, Spurgeon SL, Khan SH, Menchen SM, Heiner CR, et al. New dye-labeled terminators for improved DNA sequencing patterns. Nucleic Acids Res 1997;25:4500 – 4. Ahrendt SA, Halachmi S, Chow JT, Wu L, Halachmi N, Yang SC, et al. Rapid p53 sequence analysis in primary lung cancer using an oligonucleotide probe array. Proc Natl Acad Sci U S A 1999;96:7382–7. Tonisson N, Kurg A, Kaasik K, Lohmussaar E, Metspalu A. Unravelling genetic data by arrayed primer extension. Clin Chem Lab Med 2000;38: 165–70. Bosch FX, Ritter D, Enders C, Flechtenmacher C, Abel U, Dietz A, et al. Head and neck tumor sites differ in prevalence and spectrum of p53 alterations but these have limited prognostic value. Int J Cancer 2004;111:530 – 8. Wen WH, Bernstein L, Lescallett J, Beazer-Barclay Y, Sullivan-Halley J, White M, et al. Comparison of TP53 mutations identified by oligonucleotide microarray and conventional DNA sequence analysis. Cancer Res 2000;60: 2716 –22. McGregor JM, Harwood CA, Brooks L, Fisher SA, Kelly DA, O’nions J, et al. Relationship between p53 codon 72 polymorphism and susceptibility to sunburn and skin cancer. J Invest Dermatol 2002;119:84 –90. Buller RE, Sood A, Fullenkamp C, Sorosky J, Powills K, Anderson B. The influence of the p53 codon 72 polymorphism on ovarian carcinogenesis and prognosis. Cancer Gene Ther 1997;4:239 – 45. Bonafe M, Ceccarelli C, Farabegoli F, Santini D, Taffurelli M, Barbi C, et al. Retention of the p53 codon 72 arginine allele is associated with a reduction of disease-free and overall survival in arginine/proline heterozygous breast cancer patients. Clin Cancer Res 2003;9:4860 – 4. Sullivan A, Syed N, Gasco M, Bergamaschi D, Trigiante G, Attard M, et al. Polymorphism in wild-type p53 modulates response to chemotherapy in vitro and in vivo. Oncogene 2004;23:3328 –37.

DOI: 10.1373/clinchem.2005.048348

1288

Technical Briefs

Rapid, Simultaneous Genotyping of 10 Southeast Asian Glucose-6-Phosphate Dehydrogenase Deficiency–Causing Mutations and a Silent Polymorphism by Multiplex Primer Extension/Denaturing HPLC Assay, Grant Wu,1 Wei-Hua Liang,2 Jim Zhu,1 Hong Ouyang,2 Paul Pearson,1 Ren Cai,3 Can Liao,4 Qiu-Hua Mo,2 Dong-Lin Zhu,3 and Xiang-Ming Xu2* [1 Transgenomic, Inc., Omaha, NE; 2 Department of Medical Genetics, Southern Medical University, Guangzhou, Guangdong, Peoples Republic of China; 3 Liugzhou Maternal and Neonatal Hospital, Liuzhou, Guangxi, Peoples Republic of China; 4 Guangzhou Maternal and Neonatal Hospital, Guangzhou, Guangdong, Peoples Republic of China; * address correspondence to this author at: Department of Medical Genetics, Southern Medical University (Formerly: First Military Medical University), Tonghe 510515, Guangzhou, Guangdong, Peoples Republic of China; fax 86-020-87278766, e-mail gzxuxm@ pub.guangzhou.gd.cn] Deficiency of glucose-6-phosphate dehydrogenase (G6PD) is the most common inherited metabolic enzyme disorder in the world, with a very high incidence throughout the tropics and subtropics as a result of malarial selection (1–3 ). More than 140 mutations or combinations of mutations of the X-linked gene for G6PD have been characterized at the DNA level from different ethnic populations worldwide (4, 5 ). In areas of Southeast Asia, including southern China, at least 25 deficiency-causing point mutations have been identified in the human G6PD gene. Ten mutations (95A3 G, 392G3 T, 487G3 A, 493A3 G, 592C3 T, 871G3 A, 1024C3 T, 1360C3 T, 1376G3 T, and 1388G3 A) account for ⬎80% of all known mutations in Southeast Asian countries (4 – 8 ). The current molecular approaches to identifying G6PDdeficiency– causing mutations rely on various methods for allele differentiation (7–10 ). Each of these approaches has advantages and limitations, but the technical aspects and/or cost have limited their routine use in most laboratories. Rapid and accurate genotyping of G6PD-deficiency– causing mutations is necessary to meet the requirements for genetic counseling, genetic epidemiology, and clinical molecular diagnosis of this disorder. We have developed a multiplex primer extension/fully denaturing HPLC (PE/DHPLC) assay capable of simultaneously detecting the above 10 G6PD mutations and 1 common polymorphism (1311C3 T) (11 ). Eleven genomic DNA samples with different known genotypes, including the above 11 G6PD mutations and the silent polymorphism as previously determined by direct sequencing, were used to validate this application. A total of 209 blood samples with various G6PD-deficiency– causing mutations, the silent polymorphism, or the wild-type G6PD gene sequence identified by direct sequencing were obtained to test the specificity of this assay by blind analysis. The genotypes of these DNA samples are shown in Table 1 of the Data Supplement that accompanies the online version of this Technical Brief at http://www.clinchem. org/content/vol51/issue7/. Multiplex PCRs were designed to produce sequences of

interest in the G6PD gene, to be used for detection by the PE reaction, and generated 370-, 967-, and 1041-bp fragments (GenBank accession number X55448), respectively. Eleven oligonucleotides for specific genotyping of the 11 different mutations were designed according to the previously published sequence [(4, 12 ); Table 1 and Fig. 1 in the online Data Supplement] and were used in 2 separate groups for multiplex PE. The basic principle for identification of point mutations by this method is PE-based specific detection (13 ). Because the separation of oligonucleotides by DHPLC is both size and sequence dependent under fully denaturing conditions, modification of some of the oligonucleotides was necessary to provide an easily interpretable genotype pattern (14 ). The multiplex PCR was performed in a total volume of 25 ␮L containing 12.5 ␮L of 2⫻ PCR Master Mix [including deoxynucleotide triphosphates (dNTPs), buffer, MgCl2, and Taq DNA polymerase (Promega)], 0.2 ␮g of genomic DNA, 0.2 ␮M each of the primers for fragment 1 (5⬘-GACCCCAGAGGAACTCTCAA-3⬘ and 5⬘-CAGGTAGAGCCGGGATGAT-3⬘), 0.6 ␮M each of the primers for fragment 2 (5⬘-ACGATGATGCAGCCTCCTAC-3⬘ and 5⬘GTTCTGCACCATCTCCTTGC-3⬘), and 0.8 ␮M each of the primers for fragment 3 (5⬘-CCTGAGGGCTGCACATCT-3⬘ and 5⬘-CCTTTCCTCACCTGCCATAA-3⬘). The final volume was adjusted to 25 ␮L with water. For amplification, an initial 2-min denaturation at 95 °C was followed by 14 “touchdown” cycles of 94 °C for 30 s, 62 °C for 30 s (with a temperature decrease of 0.5 °C/cycle for 14 cycles to 55 °C), and 72 °C for 1 min 30 s; 20 cycles of 94 °C for 30 s, 55 °C for 30 s, and 72 °C for 1 min 30 s; and a final extension at 72 °C for 5 min. PCR was carried out in an Applied Biosystems 9700 thermal cycler (PerkinElmer). Excess primers and dNTPs were removed from the multiplex PCR reaction by use of exonuclease I (US Biochemicals) and shrimp alkaline phosphatase (Roche Diagnostics) digestion. Two units of each enzyme were added directly to 20 ␮L of PCR reaction mixture, mixed, and incubated at 37 °C for 25 min followed by 94 °C for 5 min to inactivate the enzymes. The multiplex PE reactions were carried out in a final volume of 20 ␮L containing the following: 50 ␮M each dideoxynucleotide triphosphate (ddNTP) except for the replacement of ddGTP by dGTP (Amersham Biosciences) in the group 1 reaction, 0.5 ␮M each primer, 0.5 U of Thermo Sequenase (Amersham), 4 ␮L of purified PCR product, and 2 ␮L of 10⫻ reaction buffer provided by the manufacturer. The PE reactions were performed on an Applied Biosystems 9700 thermal cycler under the following conditions: an initial 1-min denaturation at 96 °C was followed by 50 cycles of 96 °C for 5 s and 65 °C for 15 s. This step for groups 1 and 2 used the same PE thermocycling profile. DHPLC analysis was conducted on the automated WAVE DNA Fragment Analysis System (Transgenomic). Briefly, 5 ␮L of each PE product was injected into the mobile phase (buffer A, 0.1 mol/L triethyammonium acetate; buffer B, 250 mL/L acetonitrile in 0.1 mol/L triethylammonium acetate) at 0.9 mL/min flow rate. The

Clinical Chemistry 51, No. 7, 2005

1289

Fig. 1. Representative DHPLC profiles for the detection of 11 common Southeast Asian mutations in the G6PD gene. W, wild type; M, mutant. The negative control illustrates the elution times of the oligonucleotides with no PE products present. The positive control is a mixture of all 6 or 5 mutant G6PD allele samples as a template for PCR followed by PE. Genotypes and sample numbers are labeled to the left of the chromatogram. Group 1 for detection of 6 mutations and group 2 for detection of 5 mutations are illustrated separately. There are samples from 7 males (hemizygotes) and 9 females (heterozygotes or homozygotes) in these 2 DHPLC profiles, including 5 samples for compound heterozygotes (samples 77, 110, 143, 195, and m55), 1 sample from a homozygote (sample 43), and 1 sample from a compound hemizygote (sample 160), of which 4 showed positive specific mutant peaks in both groups for their detection (samples 143, 160, 195, and m55).

1290

Technical Briefs

Table 1. Primer sequences used for multiplex PCR and extension products for the 11 frequent Southeast Asian G6PD mutations.a Mutations

Name

Group 1 1376G3T

Primers, 5ⴕ–3ⴕ

J

CCCCTCAGCGACGAGCTCC

592C3T

E

GTCCGTGAGGACCAGATCTAC

1311C3T

H

CCTCACAGAACGTGAAGCTCCCTGACGCCTA

392G3T

B

TTTAGGTAGAAGAGGCGGTTGGCCTGTGACc

1388G3A

K

TTTAATCTGGTGCAGCAGTGGGGTGAAAATAc

95A3G

A

TTTTTTTTGCCTTCCATCAGTCGGATACAC

Group 2 487G3A

C

CAGCTCCGGGCTCCCAGCAGA

493A3G

D

ACTTCTCCACGATGATGCGGTc

1360C3T

I

TTTTGCGGGAGCCAGATGCACTTCGTG

871G3A

F

TTTTTTGGCTTTCTCTCAGGTCAAG

1024C3T

G

TTTTTTTCCACCTCTCATTCTCCACATAGAc

Extension

W:b -GddT M: -ddT W: -ddC M: -ddT W: -ddC M: -ddT W: -ddC M: -ddA W: -ddC M: -ddT W: -ddA M: -GddC W: M: W: M: W: M: W: M: W: M:

-ddG -ddA -ddT -ddC -ddC -ddT -ddG -ddA -ddG -ddA

a

The primers are listed in the order of elution of PE products by DHPLC. The artificial bases added at the 5⬘ end are in italics. W, wild-type; M, mutant. c Reverse primer. b

primer extension products were eluted from the solid phase (DNASep cartridge; Transgenomic) by a linear gradient (18%–39% buffer B) in an 11-min sample run under fully denaturing conditions (70 °C). The eluted products were detected by their ultraviolet absorbance at 260 nm. We established 2 groups of multiplex PE reactions for detection of 10 G6PD-deficiency– causing mutations and 1 silent polymorphism; these reactions can simultaneously detect 6 (group 1) or 5 mutations (group 2). Representative DHPLC profiles for genotyping these 11 G6PD mutations are shown in Fig. 1. In this technique, there are 3 possible DHPLC profiles for each potential mutation, which have been described previously (14 ), except that a primer peak plus a mutant peak (M in Fig. 1) indicates a hemizygous mutation in men, because the G6PD gene is X-linked, or a homozygous mutation in women. In addition, we noted that the heterozygous mutation was identified only in female samples. In a blind analysis, 208 of the 209 samples tested (99.5%) were found to be in concordance with the DNA sequencing data; the exception involved 1 sample that gave a false-positive (genotype of 1376/1388) result in the PE assay but no 1376G3 T mutation in sequence analysis. The combination of extension primers for each group was chosen to obtain well-distributed and easily interpretable DHPLC profiles. We assembled primers A, B, E,

H, J, and K in group 1 with substrates of dGTP and 3 ddNTPs and primers C, D, F, G, and I in group 2 with the substrates being the 4 ddNTPs (Table 1) so that highquality profiles for genotyping each of the mutations were produced. In our experience, the highest numbers of multiplex oligonucleotide products that can be detected in 1 DHPLC test is 8 because of limitations of retention time and other conditions. We propose that “building blocks” of 5 mutations be a general rule for designing multiplex PE assays using fully denaturing DHPLC analysis on the Wave system. To establish the identities of grouped PE products without overlapping of peaks, the design of primers and their combination based on substrates (ddNTPs and dNTPs) should be the main considerations for PE/DHPLC optimization. In addition to this, one needs to pay close attention to column equilibration in the PE/DHPLC application. Nonequilibration of a column may cause a shift in retention time, low peak intensity, broader peak patterns, or even split peaks. To avoid this, use only the “Normal Clean” setting on the WAVE DNA Fragment Analysis System. Therefore, buffer B is used for column cleaning. A couple of blank injections between sample injections are also recommended to ensure column equilibration. We noticed that some samples from hemizygous and homozygous individuals had similar DHPLC profiles that were missing the wild-type peak. It therefore is necessary

Clinical Chemistry 51, No. 7, 2005

to determine the patient’s sex when genotyping samples as hemizygous or homozygous. Nine male and 9 female samples were randomly selected for sex detection to test the accuracy of genotyping results by PE/DHPLC in combination with data on sex. All matched perfectly in all of the comparative tests, confirming the accuracy of this strategy. This assay thus enables accurate differentiation of the G6PD mutant or wild-type alleles from hemizygous and wild-type males from those of female heterozygotes and female mutant homozygotes, although the latter was rarely detected in our population. In our blind analysis, all different genotypes were well characterized by the PE/DHPLC assay (Fig. 1), including a few homozygous mutant samples from G6PD-deficient females, e.g., genotypes of 1388A/1388A (sample 43), 1376T/1388A (sample 110), 392T/1024T (sample 143), and 487A/1376T (sample m55), indicating the reliability of this technology for genotyping X-linked G6PD mutations. Our results, along with the relatively low reagent costs (approximately US $0.50/genotype) and short processing time (1 day for testing for 11 mutations in 65 samples) for our assay, suggest a potential use of this technology for screening programs. This approach has been applied to the detection of G6PD deficiency (present study) and ␤-thalassemia (14 ), and may be expanded to other common diseases, such as nondeletional ␣-thalassemia, Wilson disease, or cystic fibrosis.

We thank Shiqi Jia for editorial assistance. This study was partially supported by the National Science Fund for Distinguished Young Scholars of NSFC (30325037 to Xiangmin Xu), by the National Key Technologies R&D Program (2004BA720A04), and by the Natural Science Fund from the Department of Science and Technology of Guangdong Province (32824). References 1. Beutler E. G6PD: population genetics and clinical manifestations. Blood Rev 1996;10:45–52. 2. Johns Hopkins University. Online Mendelian Inheritance in Man. http:// www.ncbi.nlm.nih.gov/entrez/dispomim.cgi?id⫽305900 (accessed January 2005). 3. Sabeti PC, Reich DE, Higgins JM, Levine HZ, Richter DJ, Schaffner SF, et al. Detecting recent positive selection in the human genome from haplotype structure. Nature 2002;419:832–7. 4. Beutler E, Vulliamy TJ. Hematologically important mutations: glucose-6phosphate dehydrogenase. Blood Cells Mol Dis 2002;28:93–103. 5. Kwok CJ, Martin AC, Au SW, Lam VM. G6PDdb, an integrated database of glucose-6-phosphate dehydrogenase (G6PD) mutations. Hum Mutat 2002; 19:217–24. 6. Iwai K, Hirono A, Matsuoka H, Kawamoto F, Horie T, Lin K, et al. Distribution of glucose-6-phosphate dehydrogenase mutations in Southeast Asia. Hum Genet 2001;108:445–9. 7. Ainoon O, Yu YH, Amir Muhriz AL, Boo NY, Cheong SK, Hamidah NH. Glucose-6-phosphate dehydrogenase (G6PD) variants in Malaysian Malays. Hum Mutat 2003;21:101. 8. Du CS, Ren X, Chen L, Jiang W, He Y, Yang M. Detection of the most common G6PD gene mutations in Chinese using amplification refractory mutation system. Hum Hered 1999;49:133– 8. 9. Gemignani F, Perra C, Landi S, Canzian F, Kurg A, Tonisson N, et al. Reliable detection of ␤-thalassemia and G6PD mutations by a DNA microarray. Clin Chem 2002;48:2051– 4. 10. Zhao F, Ou XL, Xu CC, Cai GQ, Wu XY, Huang YM, et al. Rapid detection of six common Chinese G6PD mutations by MALDI-TOF MS. Blood Cells Mol Dis 2004;32:315– 8.

1291

11. Ren X, Du C, Lin Q. Studies on a G6PD polymorphic site, cDNA C1311T. Zhonghua Xue Ye Xue Za Zhi 1999;20:197–9. 12. Beutler E, Vulliamy T, Luzzatto L. Hematologically important mutations: glucose-6-phosphate dehydrogenase. Blood Cells Mol Dis 1996;22:49 –56. 13. Sokolov BP. Primer extension technique for the detection of single nucleotide in genomic DNA. Nucleic Acids Res 1990;18:3671. 14. Wu G, Hua L, Zhu J, Mo QH, Xu XM. Rapid, accurate genotyping of ␤-thalassaemia mutations using a novel multiplex primer extension/denaturing high-performance liquid chromatography assay. Br J Haematol 2003; 122:311– 6. DOI: 10.1373/clinchem.2005.049114

CYP2D6 Genotyping by a Multiplex Primer Extension Reaction, Johanna Sistonen,1* Silvia Fuselli,1,2 Antti Levo,1 and Antti Sajantila1 (1 Department of Forensic Medicine, University of Helsinki, Helsinki, Finland; 2 Department of Biology, University of Ferrara, Ferrara, Italy; * address correspondence to this author at: Department of Forensic Medicine, PO Box 40, University of Helsinki, 00014 Helsinki, Finland; fax 358-9-19127518, e-mail johanna.sistonen@ helsinki.fi) Great interindividual variability in drug response leads to variation in drug safety and efficacy. Although this can be the result of environmental and physiologic factors as well as drug– drug interactions, in many cases the response is inherited, arising from a polymorphism in genes encoding drug transporters, drug receptors, and especially, drug-metabolizing enzymes. The polymorphic cytochrome P450 2D6 (CYP2D6; OMIM 124030) is one of the most widely studied drug-metabolizing enzymes, being responsible for the metabolism of many commonly used drugs belonging to classes such as antidepressants, neuroleptics, beta-blockers, and antiarrhythmics (1 ). The CYP2D6 gene spans a 4.2-kb region located on chromosome 22q13.1 and is part of the CYP2D cluster together with CYP2D8P and CYP2D7 pseudogenes (2 ). At present, more than 50 different major polymorphic CYP2D6 alleles are known (3 ). These include variants produced by major rearrangements (whole-gene deletion or duplication), point mutations, and single or multiple base deletions or insertions. The phenotypic consequences of this variation are considerable: the CYP2D6 enzyme activity ranges from complete deficiency, possibly giving rise to profound toxicity of medication in the case of poor-metabolizing individuals, to ultrarapid metabolism, which can lead to therapeutic failure at recommended drug dosages. Individuals with a normal or slightly below normal rate of metabolism related to the CYP2D6 enzyme are usually defined as extensive or intermediate metabolizers, respectively. CYP2D6 genotyping to predict metabolic status is considered a valid alternative to traditional phenotyping methods (4 ). Assessing the CYP2D6 genotype also offers several distinct advantages over the experimental determination of a CYP2D6 phenotype (5 ). The characteristics of the gene do not change during the lifetime, and genetic

1292

Technical Briefs

status is uninfluenced by environmental or physiologic factors. Genotyping requires only 1 sample and can be done before a drug is given to a patient. It therefore may facilitate improved drug efficacy and diminished risk for adverse drug reactions (6 ). At present, there are specific guidelines available on how genetic information could be taken into account in clinical dose adjustments for specific antidepressants and antipsychotics (7 ). Furthermore, in forensic medicine, CYP2D6 genotyping may provide important information on whether a polymorphism may have contributed to a drug fatality. Because CYP2D6 genotyping has a wide range of applications in both routine and research investigations, methods for rapid and cost-effective genotyping are necessary. Different genotyping methods for the CYP2D6 gene have been developed, including multiplex allelespecific PCR (8 ), PCR with restriction fragment length polymorphism (PCR-RFLP) analysis (9 ), real-time PCR (10 ), pyrosequencing (11 ), and oligonucleotide microarray technology (12 ). Despite allele-specific PCR and RFLP analysis being laborious and low throughput, these methods are still widely used (13, 14 ) because the newer, high-throughput methods often require special laboratory facilities (15 ). Our aim was to develop a rapid and technically feasible genotyping method that is suitable for moderate-throughput laboratories and includes all clinically important CYP2D6 mutations that are known to alter the enzyme activity. In this study, we describe a CYP2D6 genotyping protocol based on a combination of PCR and multiplex extension of unlabeled oligonucleotide primers with fluorescently labeled dideoxynucleotide triphosphates (SNaPshotTM; Applied Biosystems). The method has been designed to screen whole-gene deletion and duplication and 11 of the most relevant polymorphic positions of the gene (Table 1) as well as the allelic composition of the gene duplication. It allows the identification of CYP2D6 alleles highly represented in different human populations (i.e., *2, *4, *10, *17, *29, *39, and *41) or alleles, even if rare, known to be responsible for low or null metabolic activity

(i.e., *3, *6, and *9) (16 ). Alleles *10 and *17, in particular, have been included to extend the usability of the method because these are considered to be the major decreasedfunction variants in Asian and African populations, respectively. Alleles not carrying detected mutations were classified as *1. The entire CYP2D6 gene (5.1 kb) was amplified in long-PCR reaction using primers CYP2D6-F (5⬘-CCAGAAGGCTTTGCAGGCTTCA-3⬘) (17 ) and CYP2D6-R (5⬘-ACTGAGCCCTGGGAGGTAGGTA-3⬘) (17 ) to separate the gene from the flanking highly homologous CYP2D8P and CYP2D7 pseudogenes. The 50-␮L reaction mixture contained 1.5 U of GeneAmp® rTth DNA Polymerase XL (Applied Biosystems), 1⫻ XL Buffer II, 1.25 mM Mg(OAc)2, 0.2 mM each deoxynucleotide triphosphate, 0.4 ␮M each primer, and 20 –100 ng of genomic DNA. The PCR was conducted as follows: denaturation at 94 °C for 1 min; 10 cycles of 94 °C for 30 s and 68 °C for 10 min; 25 cycles of 94 °C for 30 s and 68 °C for 10 min and 15 s, plus 15 s per cycle; and a final extension at 72 °C for 30 min. Two additional long-PCR reactions were used to analyze the major rearrangements, i.e., duplication or deletion, of the entire CYP2D6 gene. We amplified a duplication-specific 3.2-kb fragment, using primers CYP207-F (5⬘-CCCTCAGCCTCGTCACCTCAC-3⬘) (18 ) and CYP-32-R (5⬘-CACGTGCAGGGCACCTAGAT-3⬘) (18 ). To rule out false-negative results, we added primer CYP13-F (5⬘-ACCGGGCACCTGTACTCCTCA-3⬘) (19 ) to the reaction mixture, which yielded an internal control fragment of 3.8 kb. To detect the gene deletion, we again used a set of 3 primers; we amplified a deletion-specific 3.5-kb fragment and a 3.0-kb control fragment, using primers CYP-13-F, CYP-24-R (5⬘-GCATGAGCTAAGGCACCCAGAC-3⬘) (19 ), and CYP-207-F. The 50-␮L reaction contained 1.0 U of DyNAzymeTM EXT DNA Polymerase (Finnzymes), 1⫻ magnesium-free EXT buffer, 1.0 mM MgCl2, and 0.2 mM each deoxynucleotide triphosphate. The primer concentrations were as follows: for the duplication-specific reaction, we used 0.6 ␮M CYP-207-F, 0.4 ␮M CYP-32-R, and 0.2 ␮M CYP-13-F; and for the deletion-

Table 1. Primers used in the SNaPshot reaction and detected CYP2D6 alleles. Primera

Sequence, 5ⴕ–3ⴕ

Concentration in the SNaPshot reaction, ␮M

Detected polymorphism

Related CYP2D6 allele(s)

p19-100 p24-1023 p29-1661 p34-1707 p39-1846 p44-2549 p49-2615 p54-2850 p59-4180 p64-2988 p69-3183

ACGCTGGGCTGCACGCTAC (T)3ACCGCCCGCCTGTGCCCATCA (T)8CGAGCAGAGGCGCTTCTCCGT (T)14GCAAGAAGTCGCTGGAGCAG (T)21CCGCATCTCCCACCCCCA (T)23GATGAGCTGCTAACTGAGCAC (T)28GCCTTCCTGGCAGAGATGGAG (T)33AGCTTCAATGATGAGAACCTG (T)38GTGTCTTTGCTTTCCTGGTGA (T)45AGTGCAGGGGCCGAGGGAG (T)47TGTCCAACAGGAGATCGACGAC

0.06 0.06 0.02 0.12 0.02 0.04 0.04 0.14 0.08 0.04 0.02

100C⬎T 1023C⬎T 1661G⬎C 1707delT 1846G⬎A 2549delA 2613–15delAGA 2850C⬎T 4180G⬎C 2988G⬎A 3183G⬎A

*4, *10 *17 *2, *4, *10, *17, *29, *39, *41 *6 *4 *3 *9 *2, *17, *29, *41 *2, *4, *10, *17, *29, *39, *41 *41b *29

a

The first number in the primer name describes the length of the primer, and the second refers to the SNP position. Intronic position 2988 has been considered as defining allele *41 because the predictivity of this mutation for impaired enzyme function has been demonstrated recently (27 ). b

Clinical Chemistry 51, No. 7, 2005

1293

Fig. 1. Multiplex single-base extension genotyping of the CYP2D6 gene. (A), agarose gel electrophoresis (1%) of the long-PCR products. Lanes 2 and 3, 5.1-kb fragment (wt) used as a template in the SNaPshot reaction. Lanes 4 and 5 show detection of gene duplication (dup): lane 4, duplication-negative sample; lane 5, duplication-positive sample. Lanes 6 and 7 show detection of gene deletion (del): lane 6, deletion-negative sample; lane 7, deletion-positive sample. Lanes 8 and 9 show results for a 9.3-kb fragment obtained in an additional experiment to confirm the duplicated genotype identification (dup*). (B), structure of the CYP2D6 gene showing the polymorphic positions assayed by the method. Top scheme shows positions in 11-plex reaction; bottom scheme shows positions in an additional duplication-specific reaction (9.3-kb fragment). f indicate exons. (C), electropherograms of the SNaPshot reaction products for 3 different individuals. Peaks correspond to the 11 polymorphic sites screened (numbers above the peaks indicate their positions in the gene), and the genotypes are indicated in the top right-hand corners. The top electropherogram represents the CYP2D6*1/*4 genotype. Stars indicate the mutated positions defining the CYP2D6*4 allele. Two other samples were found to be duplication-positive by long-PCR detection. By comparing these electropherograms with the CYP2D6*1/*4 genotype and considering the ratio between the intensities of the 2 peaks at the same position, the duplicated allele can readily be identified as the one displaying higher signals (arrows). Electropherograms presenting genotypes CYP2D6*2/*4, CYP2D6*2xN/*4, and CYP2D6*2/*4xN are shown in Fig. 2 of the online Data Supplement.

specific reaction, 0.6 ␮M CYP-13-F, 0.4 ␮M CYP-24-R, and 0.2 ␮M CYP-207-F. The temperature cycling profile was the same as described above. The long-PCR products were analyzed on 1% agarose gels (Fig. 1A), and the 5.1-kb fragments were purified by use of 1 ␮L of ExoSAPIT® (USB Corporation) per 5 ␮L of PCR product, with incubation at 37 °C for 15 min and at 80 °C for 15 min to successively deactivate the enzymes. The purified 5.1-kb product was used as a template to detect 11 polymorphic positions of the CYP2D6 gene (Fig. 1B and Table 1). In the following single-base extension reaction, the detection primers annealed adjacent to the

single-nucleotide polymorphism (SNP) position. A difference of 5 nucleotides between primer lengths was chosen to avoid overlapping between the fluorescent signals of the final products in the capillary electrophoresis. To achieve the final length, 5⬘ poly(dT) tails were added to the primers. The sequences of the detection primers and their final concentrations in the reaction are listed in Table 1. The SNaPshot reaction contained 2.0 ␮L of purified PCR product, 2.0 ␮L of pooled detection primers, and 2.5 ␮L of SNaPshot ready reaction mixture in a final volume of 10 ␮L. The temperature cycling profile was 25 cycles of 96 °C for 10 s, 55 °C for 10 s, and 60 °C for 30 s. After the

1294

Technical Briefs

reaction, 5⬘-phosphoryl groups of unincorporated dideoxynucleotide triphosphates were removed by addition of 1.0 U of calf intestinal phosphatase (Finnzymes) and incubation of the samples at 37 °C for 1 h and at 75 °C for 15 min to successively deactivate the enzyme. Capillary electrophoresis of samples was performed on the ABI PRISM® 3100 genetic analyzer, and the results were analyzed with GeneMapper ID, Ver. 3.1 (Applied Biosystems; Fig. 1C). Electropherograms presenting different CYP2D6 alleles are shown in Fig. 1 in the Data Supplement that accompanies the online version of this Technical Brief at http://www.clinchem.org/content/vol51/ issue7/. We have used the method to genotype the HGDPCEPH Human Genome Diversity Cell Line Panel, which includes 1064 individuals [Ref. (20 ); our unpublished data]; 247 samples extracted from fresh blood (21 ), and ⬎200 postmortem cases. Altogether, these samples represent a worldwide distribution of populations; thus each detected mutation was found in at least 8 individuals. For method validation, we selected a subset of 50 individuals from the above-mentioned samples, representing all different detected alleles, and genotyped them by PCR-RFLP analysis (9, 22 ). The minimum allelic frequency in the subset was 3% (at least 3 chromosomes bearing the mutation). Polymorphic positions 2988G⬎A and 3183G⬎A, which were not included in our PCR-RFLP protocol, were verified by sequencing 5 carriers and 2 noncarriers according to the SNaPshot profile. In addition, we performed a SNaPshot genotyping to reanalyze 33 postmortem cases typed previously by PCR-RFLP analysis (22 ). Concordance was 100% between the new method and the conventional methods. A new and interesting feature of our genotyping protocol is the possibility of determining the phase of gene duplication. We observed that the SNaPshot reaction described above presents as a byproduct a quantitative aspect, and we believe that it provides adequate information to determine which of the 2 alleles is actually duplicated (Fig. 1C; also see Fig. 2 in the online Data Supplement). To confirm the previous result, we selected 11 samples from the CEPH dataset that were found duplication-positive (2 *1xN/*2, 2 *1/*2xN, 1 *1xN/*4, 2 *1/*4xN, 2 *2xN/*4, and 2 *2/*4xN). We amplified a 9.3-kb fragment (Fig. 1A), using primers Lx2F (5⬘-GCCACCATGGTGTCTTTGCTTTC-3⬘) (23 ) and Lx2R (5⬘-ACCGGATTCCAGCTGGGAAATG-3⬘) (23 ). The reaction conditions and the temperature cycling profile were the same as described for the 5.1-kb fragment. The amplified region starts from exon 9 and ends at intron 2 of the 2 subsequent CYP2D6 genes. Three polymorphic positions, namely 4180 of the first gene and 100 and 1023 of the second gene, were examined by nested PCR-RFLP analysis. The duplicated genotype identifications obtained through the original 11-plex SNaPshot reaction were confirmed in all cases. From our experience, the SNaPshot method is very robust, accurate, and easy to perform. Advantages of the SNaPshot technique include the detection of high num-

bers of mutations in a single reaction and the possibility to increase the number of detected SNPs without substantially increasing the costs of analysis. With our method the cost per genotype is ⬃5 € (from long PCR to the capillary electrophoresis), which is less than one-half the cost of the corresponding PCR-RFLP analysis. Samples can be processed in 96-well plates, and after the overnight PCR, the final genotypes can be obtained in 5 h. In summary, we designed a cost-effective and technically feasible CYP2D6 genotyping technique through a 2-step assay that provides a straightforward interpretation of results. It allows detection of 11 of the most relevant polymorphic positions, the assessment of wholegene deletion and duplication, and unlike most available typing methods, the allele composition of the gene duplication. The latter feature may be useful in the prediction of phenotype; a clear example is the difference between the genotype CYP2D6*1/*4xN, which produces a single full-function allele, and genotype CYP2D6*1xN/*4, which will produce at least twice the amount of enzyme. Methods using the same principle have recently been published (24, 25 ), but to our knowledge this is the most exhaustive CYP2D6 SNaPshot assay in terms of number and type of polymorphic positions and validation. Covering the most frequent and clinically relevant CYP2D6 mutations, this method can be used in routine and research investigations worldwide. This feature is particularly important considering the mixed genetic background of modern societies (26 ). In addition, the results obtained by the proposed combination of PCR and SNaPshot were confirmed, with no exceptions, by conventional methods. This was true for both the definition of the 11 SNP haplotypes and the phase of gene duplication. The method is suitable for moderate-throughput laboratories and is easily extended to alternative or additional SNPs in the targeted CYP2D6 gene region. References 1. Ingelman-Sundberg M. Genetic polymorphisms of cytochrome P450 2D6 (CYP2D6): clinical consequences, evolutionary aspects and functional diversity [Review]. Pharmacogenomics J 2005;5:6 –13. 2. Kimura S, Umeno M, Skoda RC, Meyer UA, Gonzalez FJ. The human debrisoquine 4-hydroxylase (CYP2D) locus: sequence and identification of the polymorphic CYP2D6 gene, a related gene, and a pseudogene. Am J Hum Genet 1989;45:889 –904. 3. Human Cytochrome P450 (CYP) Allele Nomenclature Committee. CYP2D6 allele nomenclature. www.imm.ki.se/CYPalleles/cyp2d6.htm (accessed April 2005). 4. Griese EU, Zanger UM, Brudermanns U, Gaedigk A, Mikus G, Morike K, et al. Assessment of the predictive power of genotypes for the in-vivo catalytic function of CYP2D6 in a German population. Pharmacogenetics 1998;8:15– 26. 5. McElroy S, Sachse C, Brockmoller J, Richmond J, Lira M, Friedman D, et al. CYP2D6 genotyping as an alternative to phenotyping for determination of metabolic status in a clinical trial setting. AAPS PharmSci 2000;2:E33. 6. Lindpaintner K. Pharmacogenetics and the future of medical practice. Br J Clin Pharmacol 2002;54:221–30. 7. Kirchheiner J, Nickchen K, Bauer M, Wong ML, Licinio J, Roots I, et al. Pharmacogenetics of antidepressants and antipsychotics: the contribution of allelic variations to the phenotype of drug response [Review]. Mol Psychiatry 2004;9:442–73. 8. Stu¨ven T, Griese EU, Kroemer HK, Eichelbaum M, Zanger UM. Rapid detection of CYP2D6 null alleles by long distance- and multiplex-polymerase chain reaction. Pharmacogenetics 1996;6:417–21. 9. Sachse C, Brockmo¨ller J, Bauer S, Roots I. Cytochrome P450 2D6 variants in a Caucasian population: allele frequencies and phenotypic consequences. Am J Hum Genet 1997;60:284 –95.

Clinical Chemistry 51, No. 7, 2005

10. Stamer UM, Bayerer B, Wolf S, Hoeft A, Stu¨ber F. Rapid and reliable method for cytochrome P450 2D6 genotyping. Clin Chem 2002;48:1412–7. 11. Zackrisson AL, Lindblom B. Identification of CYP2D6 alleles by single nucleotide polymorphism analysis using pyrosequencing. Eur J Clin Pharmacol 2003;59:521– 6. 12. Chou WH, Yan FX, Robbins-Weilert DK, Ryder TB, Liu WW, Perbost C, et al. Comparison of two CYP2D6 genotyping methods and assessment of genotype-phenotype relationships. Clin Chem 2003;49:542–51. 13. Gaikovitch EA, Cascorbi I, Mrozikiewicz PM, Brockmo¨ller J, Frotschl R, Kopke K, et al. Polymorphisms of drug-metabolizing enzymes CYP2C9, CYP2C19, CYP2D6, CYP1A1, NAT2 and of P-glycoprotein in a Russian population. Eur J Clin Pharmacol 2003;59:303–12. 14. Scordo MG, Caputi AP, D’Arrigo C, Fava G, Spina E. Allele and genotype frequencies of CYP2C9, CYP2C19 and CYP2D6 in an Italian population. Pharmacol Res 2004;50:195–200. 15. Syva¨nen AC. Accessing genetic variation: genotyping single nucleotide polymorphisms [Review]. Nat Rev Genet 2001;2:930 – 42. 16. Bradford LD. CYP2D6 allele frequency in European Caucasians, Asians, Africans and their descendants [Review]. Pharmacogenomics 2002;3:229 – 43. 17. Lundqvist E, Johansson I, Ingelman-Sundberg M. Genetic mechanisms for duplication and multiduplication of the human CYP2D6 gene and methods for detection of duplicated CYP2D6 genes. Gene 1999;226:327–38. 18. Lovlie R, Daly AK, Molven A, Idle JR, Steen VM. Ultrarapid metabolizers of debrisoquine: characterization and PCR-based detection of alleles with duplication of the CYP2D6 gene. FEBS Lett 1996;392:30 – 4. 19. Steen VM, Andreassen OA, Daly AK, Tefre T, Borresen AL, Idle JR, et al. Detection of the poor metabolizer-associated CYP2D6(D) gene deletion allele by long-PCR technology. Pharmacogenetics 1995;5:215–23. 20. Cann HM, de Toma C, Cazes L, Legrand MF, Morel V, Piouffre L, et al. A human genome diversity cell line panel. Science 2002;296:261–2. 21. Fuselli S, Dupanloup I, Frigato E, Cruciani F, Scozzari R, Moral P, et al. Molecular diversity at the CYP2D6 locus in the Mediterranean region. Eur J Hum Genet 2004;12:916 –24. 22. Levo A, Koski A, Ojanpera¨ I, Vuori E, Sajantila A. Post-mortem SNP analysis of CYP2D6 gene reveals correlation between genotype and opioid drug (tramadol) metabolite ratios in blood. Forensic Sci Int 2003;135:9 –15. 23. Johansson I, Lundqvist E, Dahl ML, Ingelman-Sundberg M. PCR-based genotyping for duplicated and deleted CYP2D6 genes. Pharmacogenetics 1996;6:351–5. 24. Bender K. SNaPshot for pharmacogenetics by minisequencing. Methods Mol Biol 2004;297:243–52. 25. Knaapen AM, Ketelslegers HB, Gottschalk RW, Janssen RG, Paulussen AD, Smeets HJ, et al. Simultaneous genotyping of nine polymorphisms in xenobiotic-metabolizing enzymes by multiplex PCR amplification and single base extension [Technical Brief]. Clin Chem 2004;50:1664 – 8. 26. Schultz J. FDA guidelines on race and ethnicity: obstacle or remedy? J Natl Cancer Inst 2003;95:425– 6. 27. Raimundo S, Toscano C, Klein K, Fischer J, Griese EU, Eichelbaum M, et al. A novel intronic mutation, 2988G⬎A, with high predictivity for impaired function of cytochrome P450 2D6 in white subjects. Clin Pharmacol Ther 2004;76:128 –38. Previously published online at DOI: 10.1373/clinchem.2004.046466

Distinguishing Different DNA Heterozygotes by HighResolution Melting, Robert Graham,1 Michael Liew,2 Cindy Meadows,2 Elaine Lyon,2 and Carl T. Wittwer1,2* (1 Department of Pathology, University of Utah School of Medicine, Salt Lake City, UT; 2 Institute for Clinical and Experimental Pathology, ARUP, Salt Lake City UT; * address correspondence to this author at: Department of Pathology, University of Utah School of Medicine, Salt Lake City, UT 84132; fax 801-581-4517, e-mail carl.wittwer@ path.utah.edu) High-resolution melting was recently introduced as a technique to genotype single-nucleotide polymorphisms (SNPs) within small amplicons (1 ). This closed-tube

1295

method (including rapid-cycle PCR) can be completed in ⬍15 min and does not require real-time PCR instruments (2 ), allele-specific PCR (3 ), or fluorescently labeled oligonucleotides (4 – 6 ). The process is made possible by heteroduplex-detecting DNA dyes that can be used at saturating concentrations without inhibiting PCR (7 ). Wildtype and homozygous mutant samples are distinguished by melting temperature (Tm) shifts. Heterozygous samples are best distinguished from homozygotes, not by Tm, but by altered curve shape. Heterozygous samples produce heteroduplexes that melt at lower temperatures than homoduplexes. Melting curves of amplified heterozygotes include 2 homoduplexes and 2 heteroduplexes, giving a skewed composite melting curve easily distinguished from the curves for homozygotes by curve shape (1 ). However, it was not clear whether different heterozygotes within the same amplicon could be distinguished from each other based on curve shape differences. Four different classes of SNPs have been defined based on the homo- and heteroduplexes that are produced after amplification (1 ). Because the heteroduplex mismatches of SNPs in different classes are different, their melting curves should be distinguishable if the resolution of the instrumentation is sufficient. It is also possible that different heterozygotes in the same class may be distinguishable. Although the mismatches are the same, nearestneighbor stability parameters depend on the bases adjacent to the mismatches; therefore, the predicted stabilities are often different. We selected DNA samples retrospectively from those submitted to ARUP Laboratories for standard clinical genotyping of HFE, factor V Leiden, and factor II polymorphisms. DNA was extracted from blood samples by use of the MagNa Pure instrument (Roche) and genotyped by use of adjacent hybridization probe (HybProbeTM) methods (4, 5 ) on a LightCycler®. One advantage of probe melting analysis for genotyping is that sequence alterations other than those expected can often be identified under the probes (8 ). During routine clinical analysis at ARUP, when aberrant melting peaks are observed at unexpected temperatures, samples are sequenced to identify the sequence alteration. Several heterozygous sequence variants near the expected mutations were thus identified. Three factor II heterozygotes (all SNPs), 4 factor V heterozygotes (2 SNPs, 1 single-base deletion, and 1 compound heterozygote), and 6 HFE heterozygotes (4 SNPs and 2 compound heterozygotes) were available for comparison. After selection and deidentification of samples, DNA concentrations were checked on an ND-1000 (NanoDrop Technologies, Inc.). When available, 3 individuals of each heterozygous genotype were processed. Previously described primers were used (1 ) unless it was necessary to redesign them to flank all sequence variants. SNPWizard (http://DNAWizards.path.utah.edu) was used for primer design. Primers were synthesized by Integrated DNA Technologies and used without further purification. PCR was performed as described previously

1296

Technical Briefs

(1 ), except that 1X LCGreen ® PLUS (Idaho Technology) replaced LCGreen I. LCGreen PLUS is brighter than LCGreen I and can be used on a wider variety of melting instruments. The amplicon sequences, primer regions, and specific PCR conditions used for amplification are listed in Table 1 of the Data Supplement that accompanies the online version of this Technical Brief at http:// www.clinchem.org/content/vol51/issue7/. After PCR, capillaries were heated to 90 °C at 20 °C/s in the LightCycler and then cooled to 40 °C at 20 °C/s to favor heteroduplex formation (9 ). Samples were removed from the LightCycler and analyzed on a high-resolution melting instrument (HR-1; Idaho Technology). Each capillary was inserted into the instrument at 59 °C, the temperature was increased at 0.3 °C/s, and fluorescence was acquired from 65 to 95 °C, which required only 1–2 min. Heteroduplex detection in small amplicons is favored by rapid cooling before melting, rapid heating during melting, and low Mg2⫹ concentrations (9 ). Melting curves were analyzed by custom software with normalized, temperature-shifted curves displayed as difference plots (1, 7 ). Difference plots display the difference between each sample and a “standard” (often the average of multiple wild-type curves) and should not be confused with derivative plots (⫺dF/dT), which are often used to analyze low-resolution melting data. Difference plots are a convenient way of viewing high-resolution melting data because slight differences in curve shape and Tm become obvious. Predicted duplex Tms were calculated with previously described nearest-neighbor thermodynamic models (1, 10 ) as implemented in MeltingWizard (available at http://DNAWizards.path.utah.edu). No corrections for dUTP incorporation instead of dTTP (tends to lower Tm) or influence of LCGreen PLUS (tends to increase Tm) were used. To maximize differences in Tm, amplicon lengths were kept short (1 ), although SNP typing has been demonstrated in products more than 500 bp long (7 ). Calculated Tm differences between homozygous wildtype and homozygous mutant amplicons varied from 0.0 to 1.3 °C for the different SNPs (Table 2 of the online Data Supplement). For single SNPs, heteroduplex destabilization varied from 1.1 to 3.0 °C compared with the most stable homoduplex, or 0.7–2.2 °C compared with the least stable homoduplex. For compound heterozygotes, the respective destabilizations were 2.8 – 4.0 °C and 2.0 –3.5 °C. High-resolution melting techniques are necessary to differentiate different heterozygotes within the same amplicon. Melting curves generated on the LightCycler did not reliably separate heterozygotes (data not shown), in contrast to high-resolution HR-1 analysis. As seen in Fig. 1, each heterozygote traces a unique melting curve path according to its 4 duplex Tms. These unique melting curve shapes are best seen with normalized, temperatureshifted data, shown in the form of difference plots. In Fig. 1A, factor II and 2 rare heterozygotes (all class 1 SNPs) are clearly separated from the wild type and from each other. In Fig. 1B, factor V Leiden and 3 rare heterozygotes (one SNP, one 1-bp deletion, and one compound heterozygote)

Fig. 1. Melting analysis of several heterozygous sequence variants near common mutations in factor II (20210G⬎A), factor V (Lieden; 1691G⬎A), and HFE (187C⬎G). The melting curves were normalized and temperature-shifted, and the difference between each sample and the average of the wild types is displayed at each temperature (1, 7 ). Such “difference” plots should not be confused with the commonly used derivative plots. Each mutation is indicated by its location (20210) followed by the base pair change (G⬎A). All samples are heterozygous or compound heterozygous, unless followed by a WT, which signifies the wild type. Three individuals of each genotype are shown, except for factor II 20218G⬎A and factor V compound heterozygote 1696A⬎G/1690G⬎A, which are in duplicate, and the HFE compound heterozygote 187C⬎G/189T⬎C, for which only 1 sample was available.

Clinical Chemistry 51, No. 7, 2005

also followed distinct paths. As expected, single SNPs and the 1-bp deletion (single unpaired loci) were more similar to each other than the compound heterozygote (2 unpaired loci). In Fig. 1C, all heterozygotes tested were clearly separated from each other and the wild type, including the targeted HFE mutation, 3 rare SNPs, and 2 rare compound heterozygotes. In all cases studied, different heterozygotes could be distinguished, including those within the same SNP class (1 ). At the factor V and HFE loci, compound heterozygotes were easily distinguished from “single” heterozygotes, as might be expected because of 2, rather than 1, region of destabilization. It is interesting that the predicted Tms of the homoduplexes and/or heteroduplexes are different depending on the cis/trans orientation of the sequence variants (Table 2 of the online Data Supplement). This suggests that melting analysis may allow haplotyping, at least when the variations to be haplotyped are in the same melting domain. The presence of unexpected polymorphisms close to a SNP of interest may or may not interfere with genotyping, depending on the analysis method (10 ). All PCR-based methods can be compromised if polymorphisms occur under the primers and lead to undesired allele-specific PCR. The same concern applies to internal primers used for sequencing or extension reactions. These caveats aside, some PCR-based methods, including sequencing, will identify and distinguish polymorphisms that are missed by restriction fragment length polymorphism assays and isothermal probe-based assays. Hybridization probe assays that interrogate over a range of temperatures (melting assays) often detect the presence of unexpected polymorphisms but may require further studies to identify them. In our study, high-resolution melting of small amplicons distinguished all heterozygotes studied. This included 21 pairwise comparisons, suggesting that most randomly selected heterozygotes within small amplicons can be distinguished. The power of melting analysis to distinguish multiple sequence variants has been controversial (11–14 ). When wild-type probes are used on standard low-resolution systems, 19%–52% of possible SNP heterozygotes may be confused with the expected heterozygous variant (11 ). Although better temperature resolution has been reported on these instruments (12, 13 ), it is clear that rare sequence variants under a probe may not be distinguished from the common targeted variant. Similar concerns for genotyping by amplicon melting temperature have been described and focus on the limitations of Tm to distinguish variants (14 ). High-resolution melting analysis introduces the use of melting curve shape, as well as Tm, to distinguish different variants. Because the composite melting curves of heterozygotes include 2 homoduplexes and 2 heteroduplexes, the meaning of Tm is ambiguous and less useful as a metric than is curve shape. Is the Tm of a composite melting curve the temperature at which one-half of the duplexes have melted (conventional definition), or is it the peak of the derivative plot (common derivation)?

1297

These are not identical because such curves are skewed at low temperatures from heteroduplex contributions. In either case, the Tm is only one point on the melting curve. Use of the complete melting curve, conveniently displayed as difference plots, allows differentiation of most heterozygotes (21 of 21 random pairwise comparisons in this study). High-resolution melting analysis is convenient because no processing or separation steps are required (7 ). In cases in which the complexity of the target exceeds the genotyping capabilities of amplicon melting, unlabeled oligonucleotide probes can be added (15 ). In addition to genotyping, high-resolution melting analysis is an accurate mutation scanning tool (16 ) that has been applied to the medium chain acyl-CoA dehydrogenase (17 ), c-kit (18 ), and SLC22A5 (19 ) genes, as well as HLA matching (20 ). The sensitivity and specificity are greater than denaturing HPLC (21 ), and no electrophoresis is required, as in temperature gradient (22 ), denaturing gradient (23 ), or conformation-sensitive (24 ) gel electrophoresis. References 1. Liew M, Pryor R, Palais R, Meadows C, Erali M, Lyon E, et al. Genotyping of single-nucleotide polymorphisms by high-resolution melting of small amplicons. Clin Chem 2004;50:1156 – 64. 2. Wittwer CT, Ririe KM, Andrew RV, David DA, Gundry RA, Balis UJ. The LightCycler: a microvolume multisample fluorimeter with rapid temperature control. Biotechniques 1997;22:176 – 81. 3. Germer S, Higuchi R. Single-tube genotyping without oligonucleotide probes. Genome Res 1999;9:72– 8. 4. Lay MJ, Wittwer CT. Real-time fluorescence genotyping of factor V Leiden during rapid-cycle PCR. Clin Chem 1997;43:2262–7. 5. Bernard PS, Ajioka RS, Kushner JP, Wittwer CT. Homogeneous multiplex genotyping of hemochromatosis mutations with fluorescent hybridization probes. Am J Pathol 1998;153:1055– 61. 6. Crockett AO, Wittwer CT. Fluorescein-labeled oligonucleotides for real-time pcr: using the inherent quenching of deoxyguanosine nucleotides. Anal Biochem 2001;290:89 –97. 7. Wittwer CT, Reed GH, Gundry CN, Vandersteen JG, Pryor RJ. High-resolution genotyping by amplicon melting analysis using LCGreen. Clin Chem 2003; 49:853– 60. 8. Gundry CN, Bernard PS, Herrmann MG, Reed GH, Wittwer CT. Rapid F508del and F508C assay using fluorescent hybridization probes. Genet Test 1999; 3:365–70. 9. Gundry CN, Vandersteen JG, Reed GH, Pryor RJ, Chen J, Wittwer CT. Amplicon melting analysis with labeled primers: a closed-tube method for differentiating homozygotes and heterozygotes. Clin Chem 2003;49:396 – 406. 10. Peyret N, Seneviratne PA, Allawi HT, SantaLucia J Jr. Nearest-neighbor thermodynamics and NMR of DNA sequences with internal A:A, C:C, G:G, and T:T mismatches. Biochemistry 1999;38:3468 –77. 11. Schutz E, von Ahsen N, Oellerich M. Genotyping of eight thiopurine methyltransferase mutations: three-color multiplexing, “two-color/shared” anchor, and fluorescence-quenching hybridization probe assays based on thermodynamic nearest-neighbor probe design. Clin Chem 2000;46:1728 –37. 12. Berrnard PS, Wittwer CT. Homogeneous amplification and variant detection by fluorescent hybridization probes. Clin Chem 2000;46:147– 8. 13. Lyon E, Millson A, Phan T, Wittwer CT. Detection and identification of base alterations within the region of factor V Leiden by fluorescent melting curves. Mol Diagn 1998;3:203–10. 14. von Ahsen N, Oellerich M, Schutz E. Limitations of genotyping based on amplicon melting temperature. Clin Chem 2001;47:1331–2. 15. Zhou L, Myers AN, Vandersteen JG, Wang L, Wittwer CT. Closed-tube genotyping with unlabeled oligonucleotide probes and a saturating DNA dye. Clin Chem 2004;50:1328 –35. 16. Reed GH, Wittwer CT. Sensitivity and specificity of single-nucleotide polymorphism scanning by high-resolution melting analysis. Clin Chem 2004; 50:1748 –54. 17. McKinney JT, Longo N, Hahn SH, Matern D, Rinaldo P, Strauss AW, et al. Rapid, comprehensive screening of the human medium chain acyl-CoA dehydrogenase gene. Mol Genet Metab 2004;82:112–20. 18. Willmore C, Holden JA, Zhou L, Tripp S, Wittwer CT, Layfield LJ. Detection of

1298

Technical Briefs

c-kit-activating mutations in gastrointestinal stromal tumors by high-resolution amplicon melting analysis. Am J Clin Pathol 2004;122:206 –16. 19. Dobrowolski SF, McKinney JT, Amat di San Filippo C, Giak Sim K, Wilcken B, Longo N. Validation of dye-binding/high-resolution thermal denaturation for the identification of mutations in the SLC22A5 gene. Hum Mutat 2005;25: 306 –13. 20. Zhou L, Vandersteen J, Wang L, Fuller T, Taylor M, Palais B, et al. High-resolution DNA melting curve analysis to establish HLA genotypic identity. Tissue Antigens 2004;64:156 – 64. 21. Chou LS, Lyon E, Wittwer CT. A comparison of high-resolution melting analysis to denaturing high performance liquid chromatography for mutation scanning: cystic fibrosis transmembrane conductance regulator gene as a model. Am J Clin Pathol 2005;in press.

22. Wartell RM, Hosseini S, Powell S, Zhu J. Detecting single base substitutions, mismatches and bulges in DNA by temperature gradient gel electrophoresis and related methods. J Chromatogr A 1998;806:169 – 85. 23. Abrams ES, Murdaugh SE, Lerman LS. Comprehensive detection of single base changes in human genomic DNA using denaturing gradient gel electrophoresis and a GC clamp. Genomics 1990;7:463–75. 24. Nataraj AJ, Olivos-Glander I, Kusukawa N, Highsmith WE Jr. Single-strand conformation polymorphism and heteroduplex analysis for gel-based mutation detection. Electrophoresis 1999;20:1177– 85. Previously published online at DOI: 10.1373/clinchem.2005.051516