An Algorithm of Quantification, Genotyping, and ... - Clinical Chemistry

Clinical Chemistry 52:12 2236 –2242 (2006)

Evidence-Based Laboratory Medicine and Test Utilization

Diagnosis of ␣-1-Antitrypsin Deficiency: An Algorithm of Quantification, Genotyping, and Phenotyping Melissa R. Snyder,1 Jerry A. Katzmann,1,2* Malinda L. Butz,3 Ping Yang,4 D. Brian Dawson,3 Kevin C. Halling,3 W. Edward Highsmith,3 and Stephen N. Thibodeau3

Background: Laboratory testing in suspected ␣-1-antitrypsin (A1AT) deficiency involves analysis of A1AT concentrations and identification of specific alleles by genotyping or phenotyping. The purpose of this study was to define and evaluate a strategy that provides reliable laboratory evaluation of A1AT deficiency. Methods: Samples from 512 individuals referred for A1AT phenotype analysis were analyzed by quantification, phenotype, and genotype. A1AT concentrations were measured by nephelometry. Phenotype analysis was performed by isoelectric focusing electrophoresis. The genotype assay detected the S and Z deficiency alleles by a melting curve analysis. Results: Of the 512 samples analyzed, 2% of the phenotype and genotype results were discordant. Among these 10 discordant results, 7 were attributed to phenotyping errors. On the basis of these data we formulated an algorithm, according to which we analyzed samples by genotyping and quantification assays, with a reflex to phenotyping when the genotype and quantification results were not concordant. Retrospective analyses demonstrated that 4% of samples submitted for genotype and quantitative analysis were reflexed to phenotyping. Of the reflexed samples, phenotyping confirmed the genotype result in 85% of cases. In the remaining 15%, phenotyping provided further information, including identifying rare deficiency alleles and suggesting

1 Division of Clinical Biochemistry & Immunology, Department of Laboratory Medicine & Pathology, 2 Division of Hematology, Department of Internal Medicine, 3 Division of Laboratory Genetics, Department of Laboratory Medicine & Pathology, 4 Department of Health Sciences Research, Mayo Clinic College of Medicine, Rochester, MN. * Address correspondence to this author at: Mayo Clinic, 200 First St SW, Rochester, MN 55905. Fax 507-266-4088; e-mail [email protected]. Received May 2, 2006; accepted September 27, 2006. Previously published online at DOI: 10.1373/clinchem.2006.072991

the presence of a null allele, and allowed for a more definitive interpretation of the genotype result. Conclusions: The combination of genotyping and quantification, with a reflex to phenotyping, is the optimal strategy for the laboratory evaluation of A1AT deficiency. © 2006 American Association for Clinical Chemistry

␣-1-Antitrypsin (A1AT)5 is a member of the serine protease inhibitor family (1, 2 ). A1AT is produced in the liver, released into the circulation, and enters the lungs by diffusion, where it inhibits neutrophil elastase (3 ). This inhibition involves proteolytic cleavage of the reactive loop of A1AT by elastase. The cleavage of this peptide bond causes the N-terminal portion of the loop to insert into the main ␤-sheet structure of A1AT. The elastase becomes trapped against the inhibitor, leading to internalization and lysosomal degradation of the protease/ inhibitor complex (4 –7 ). The gene for A1AT displays a significant degree of heterogeneity (2 ). Many of these heterogeneous variations do not affect the expression or function of A1AT (8 ). However, there are variations within this gene associated with decreased A1AT concentrations or dysfunctional protein. These disease-associated alleles can be classified according to their pathogenic mechanism (1, 2 ). One group, deficiency alleles, leads to intracellular accumulation or degradation of A1AT such that little is released into the circulation. A second group, the null alleles, either do not produce A1AT transcript, produce truncated protein, or produce unstable proteins that are essentially completely degraded before secretion. A third group, the

5 Nonstandard abbreviations: A1AT, ␣-1-antitrypsin; IEF, Isoelectric focusing.

2236

2237

Clinical Chemistry 52, No. 12, 2006

dysfunctional alleles, are those in which the variation leads to decreased elastase inhibitory activity. The 2 most common disease-associated alleles found in patients with A1AT deficiency are the S and Z alleles. The S allele, a substitution of valine for glutamate at codon 288 (E288V), is a deficiency allele attributable to intracellular degradation (9, 10 ). Patients homozygous for the S allele have A1AT concentrations ⬃10%–20% below those typically observed in individuals homozygous for nondeficiency alleles. A1AT concentrations in homozygous S allele patients are sufficient to prevent clinical manifestations of A1AT deficiency. In the presence of a Z allele, however, further decreases in A1AT concentrations occur and may lead to uncontrolled elastase activity and proteolytic damage in the lower respiratory tract (11, 12 ), increasing the risk for developing chronic obstructive pulmonary disease. The Z variation, which consists of a lysine-to-glutamate substitution at codon 366 (E366K), is classified as both a deficiency and dysfunctional allele (13 ). This variation causes conformational instability in the protein, which leads to polymerization of numerous A1AT molecules and formation of inclusions within the hepatocytes (14 ). This allele is also associated with decreased inhibitory activity due to decreased binding affinity for neutrophil elastase (15 ). Because the small portion of the intracellular A1AT that escapes polymerization and enters the circulation is also defective in its ability to inhibit elastase, Z homozygotes usually present with more severe pulmonary disease (16 ). In addition, the accumulation of polymerized A1AT protein in the hepatocytes leads to liver disease, which can progress to cirrhosis (17 ). The wild-type sequence of the A1AT6 gene, designated as the M allele, is present in serum at a concentration and activity sufficient to prevent clinical manifestations. A number of variations are not associated with a decrease in A1AT quantity or activity and are not linked to any clinical phenotype, for example the M1, M2, G, X, C, and D alleles defined by isoelectric focusing (IEF) migration patterns. There are some rare alleles, however, that are associated with A1AT deficiency, such as the I allele (R63C) (18 ). Current laboratory testing for A1AT deficiency involves quantification, phenotyping, genotyping, or a combination of these. The American Thoracic Society/European Respiratory Society Statement on standards for diagnosis and management of A1AT deficiency identifies phenotyping as the gold standard of genetic testing and recommends that if molecular diagnostic tests are performed, quantification may be necessary to identify null and rare deficiency alleles (19 ). A1AT quantification is typically done by immunoassay. With phenotyping, IEF gels are used to identify the A1AT alleles by protein migration patterns (20 ). Genotyping also distinguishes

between A1AT alleles, although at the DNA level. The most common genetic technique involves testing specifically for the S and Z alleles (21, 22 ). The purpose of this study was to define a laboratory-testing algorithm to assist clinicians in the diagnosis of A1AT deficiency and then to analyze the performance of this algorithm in our practice.

Materials and Methods study population Study patients were 512 consecutive Mayo Clinic patients who were referred for A1AT phenotypic analysis. Samples of whole blood (for genotyping) and serum (for quantification and phenotyping) were collected from each patient. This protocol was approved by the Mayo Clinic Institutional Review Board, and all patients gave informed consent.

a1at quantification The concentrations of A1AT were measured by nephelometry on a Behring Nephelometer II (Dade Behring, Inc.) with a commercially available standard and monospecific antisera (Dade Behring), according to the manufacturer’s instructions.

a1at phenotyping Phenotype analysis was performed by IEF on polyacrylamide gels with a pH of 3.5–5.0 (20 ). This technique separates the various isoforms of A1AT based on their migration in a pH gradient. Each isoform migrates to the position within the pH gradient at which the overall net charge of the protein is zero.

a1at genotyping Genomic DNA was isolated from whole blood samples with either a Capture Column (Gentra Systems, Inc) or a QIAamp 96-Well Column (Qiagen). PCR and subsequent melting curve analysis were used to identify the S and Z alleles (21, 22 ). PCR primers were designed to separately amplify the portions of the A1AT gene containing the S and Z variations (Table 1). For the melting curve analysis, detection probes (labeled with LC-Red640) and anchor

Table 1. PCR primers and melting curve probes used in the A1AT genotyping assay. Allele

Probe

Sequence

Sa

PCR forward PCR reverse Anchor Detection PCR forward PCR reverse Anchor Detection

5⬘-AGC GTT TAG GCA TGT TTA AC-3⬘ 5⬘-AAG TTT ATA CAG AGT AGC AGT G-3⬘ 5⬘-CCC ACG ATA TCA TCA CCA AGT TCC-3⬘ 5⬘-CAC CTG GTA AAT GAA CTC-3⬘ 5⬘-GTC CAC GTG AGC CTT GCT-3⬘ 5⬘-CAT GAA GAG GGG AGA CTT GG-3⬘ 5⬘-TGA AGC TGC TGG GGC CAT GTT T-3⬘ 5⬘-GAC CAT CGA CAA GAA AGG GA-3⬘

Zb

a 6

Human genes: A1AT, ␣-1-antitrypsin gene.

b

(21 ). Sequences kindly provided by Nicholas von Ahsen.

2238

Snyder et al.: A Diagnostic Algorithm for A1AT Deficiency Testing

probes (labeled with fluorescein) were also synthesized (Table 1). All primers were synthesized by Integrated DNA Technologies. LC probes were synthesized by Genset Corp. Each 10-␮L reaction contained ⬃25 ng genomic DNA, 1X PCR buffer (Applied Biosystems), 0.2 mmol/L each dNTP, 1 g/L bovine serum albumin (New England Biolabs), 10% glycerol, 0.4 ␮mol/L anchor probe, 0.2 ␮mol/L detection probe, and 0.6 U Taq polymerase (Applied Biosystems). In addition, the reaction for the S allele contained 2.5 mmol/L MgCl2, 0.2 ␮mol/L forward primer, and 0.8 ␮mol/L reverse primer, and the reaction for the Z variation contained 2 mmol/L MgCl2, 0.25 ␮mol/L forward primer, and 1.25 ␮mol/L reverse primer. Each assay included an S/Z heterozygote, S/S homozygote, and Z/Z homozygote control, as well as a no-DNA control. The assembled reaction mixtures were amplified and analyzed on a LightCycler™ (Roche Applied Science). The thermocycling program included a 30 s incubation at 95 °C, followed by 40 cycles of 95 °C for 0 s, 50 °C (S allele) or 58 °C (Z allele) for 10 s, and 72 °C for 8 s, all with ramp rates of 20 °C/s. The melting curve conditions involved heating the sample at 95 °C for 30 s and cooling to 30 °C for 90 s (S allele) or 40 °C for 60 s (Z allele), both at ramp rates of 20 °C/sec, then heating the sample to 65 °C (S allele) or 68 °C (Z allele) with a ramp rate of 0.1 °C/s. During the melting curve portion of the program, fluorescence was monitored constantly, with the gain set at 15 on channel 2. When required, PCR products were directly sequenced with a ThermoSequenase Radiolabeled Terminator Cycle Sequencing Kit (USB) according to manufacturer’s instructions.

Results comparison of laboratory results for a1at deficiency

Mayo Clinic patients (n ⫽ 512) were recruited from individuals referred to the laboratory for A1AT phenotype analysis. The serum A1AT concentration and phenotype were determined and reported to the medical record.

DNA was extracted from the whole blood sample, and genotyping was performed in batch mode. The genotype assay was interpreted with respect to the Z and S alleles. An individual with a Z/Z genotype is homozygous for the Z variation and homozygous for the wild-type allele at the codon associated with the S allele. A similar interpretation is made for an individual with an S/S genotype. If neither the Z nor S allele is detected, it is likely that the individual possesses 2 wild-type alleles. However, the result is termed a non-Z/non-S genotype because it is not possible from the genotype assay alone to determine if any rare or disease-causing variants are present outside the 2 regions that are assessed by S and Z genotyping. If 1 Z allele is detected, the individual is interpreted as having a Z/non-Z genotype, otherwise referred to as a Z heterozygote and consistent with carrier status. This individual is heterozygous for the Z variation at codon 288 and homozygous for the wild-type allele at the codon associated with the S allele. A similar interpretation is made for an individual in whom 1 S allele is detected. At the end of the study period, the phenotype and genotype results were compared. If the phenotype and genotype were not in agreement, the A1AT concentration was reviewed and all 3 assays were repeated to confirm the original results. Of the 512 individuals enrolled in this study, 427 were phenotyped as homozygous for the common M alleles (M, M1, and M2). Another 10 patients were phenotyped as heterozygous for an M allele and a rare, nondeficiency allele, such as G, X, C, or D. There were 65 individuals phenotyped as carriers of A1AT deficiency alleles, 35 as heterozygous for the Z allele, and 30 as heterozygous for the S allele. The remaining 10 individuals were phenotyped as being affected with A1AT deficiency, with 4 Z/Z homozygotes and 6 Z/S compound heterozygotes. The genotype and phenotype results were 98% concordant. Ten cases were discordant, with phenotypes of 1 M/M homozygote, 1 Z/Z homozygote, 4 M/Z heterozygotes, and 4 M/S heterozygotes (Table 2). When the 3 assays were repeated for these 10 cases, the discrepancies

Table 2. Results and interpretations for samples with discordant phenotype and genotype results. Sample ID

A1AT concentration, g/L

Phenotype Result

Genotype Result

Reason for Discrepancy

121 234 258 420 472 264

0.80 1.83 0.98 1.39 1.12 1.31

M/Z M/S M/Z M/M M/Z M/S

M/M M/M M/M M/Z M/M S/S

327

1.57

M/S

M/M

46

0.32

Z/Z

M/Z

226 361

1.45 3.53

M/Z M/S

Z/Z M/M

Phenotype error due to poor sample quality Phenotype error due to poor sample quality Phenotype error due to poor sample quality Phenotype error due to poor sample quality Phenotype error due to poor sample quality Error in interpretation of initial phenotype IEF gel Error in interpretation of initial phenotype IEF gel and presence of nondeficiency allele not detected by genotype assay Presence of null allele not detected by genotype assay Patient receiving A1AT replacement therapy Undetermined


could be attributed to several different causes. Based on the original and repeat phenotype assay, there were 7 samples in which the discordant result was due solely to a phenotyping error. Of these 7 cases, 5 samples were difficult to interpret on the original and repeat IEF gels. We concluded that the samples were of poor quality and that the results were uninterpretable. In the 6th sample, the phenotype result of M/S was discordant with the S/S genotype result. The subsequent phenotype assay gave a result of S/S, consistent with the genotype result and indicative of a phenotype error. In the 7th case, the sample was originally phenotyped as an M/S, but was genotyped as an M/M. The repeat phenotype demonstrated that the M allele was present; however, the 2nd allele had a migration pattern similar to, but not identical to, the S allele. Direct sequencing of this sample revealed a variation at codon 47, not previously described, resulting in a change from serine to arginine (data not shown). Although this variation was missed by both techniques, the genotyping assay did provide the more clinically relevant result, which is that of a noncarrier, unaffected individual. The remaining discordant results were attributable to a variety of issues involving both the genotype and phenotype assays. One discrepancy resulted from the clinical management of a patient. This sample was phenotyped as M/Z, but was genotyped as Z/Z. The A1AT concentration of 1.45 g/L was consistent with the M/Z phenotype result; repeat analysis, however, confirmed the genotype result. DNA sequencing of exons 2 through 5 revealed no variations under the primer binding regions or other sequence variations (data not shown). On further investigation, we discovered that the patient was receiving A1AT replacement therapy for a known deficiency. Neither assay was technically in error, because the patient did have a Z/Z genotype, and the phenotype assay detected both the endogenous Z allele and the replacement M allele. Another discrepancy appeared to be due to limitations of genotyping for only the 2 most common deficiency alleles. This sample had a phenotype of Z/Z (confirmed on repeat testing), but was genotyped as M/Z. Although a non-Z/non-S allele was detected by genotyping, the A1AT concentration was 0.32 g/L, consistent with Z/Z homozygosity. This discrepancy between A1AT concentration, genotype, and phenotype is consistent with the presence of a Z allele and a null allele (Z/null) that is not detected by either assay. The phenotyping result of homozygous Z, however, is clinically more compatible with the low A1AT concentration and homozygous A1AT deficiency. The final discrepancy involved a sample that was phenotyped as M/S, but was genotyped as an M/M homozygote. The A1AT quantification of 3.53 g/L was consistent with the M/M genotype; however, the M/S result was confirmed by a 2nd phenotyping assay. No variation was detected by gene sequencing, and the reason for the discrepancy between the phenotype and genotype for this sample remains undetermined. A sam-

2239

ple mix-up cannot be excluded as the cause of this discrepancy.

an algorithm for a1at deficiency testing On the basis of this analysis of 512 patients, we devised a diagnostic algorithm for laboratory testing in A1AT deficiency (Fig. 1). We propose that the first level of testing include the genotype assay in conjunction with A1AT quantification. If serum concentrations of A1AT are ⱖ1.00 g/L (the lower limit of our reference interval) and the genetic assay indicates a non-Z/non-S genotype, the results would be reported without further testing. If the serum concentrations are below the lower reference limit and the genetic assay detects at least one copy of a Z or S allele, the laboratory results would be released as long as the A1AT quantification is in the expected range for the genotype result (Fig. 1). If the serum concentrations of A1AT are discordant with the genotype, however, then protein phenotyping by IEF would be used as a reflex test. In the clinical laboratory, reflex tests are initiated when results obtained during primary testing are within a specific, predetermined range. The phenotype assay is complementary to the genetic assay and may clarify cases that cannot be detected by genotyping.

evaluation of the a1at deficiency algorithm Data from samples referred for A1AT genotyping were collected to determine the frequency of reflexing to phenotyping and the frequency with which the phenotyping assay clarified discrepancies between genotyping and A1AT quantification. Analysis included a total of 1246 samples that were distributed among 6 possible genotype results (Table 3). In this cohort of samples, ⬃4% of the cases were reflexed for phenotyping. The most common genotype observed (63%) was non-Z/non-S. Of the 785 non-Z/non-S genotyped samples, 26 were reflexed to phenotyping because of A1AT concentrations ⬍1.00 g/L. In all of the 26 cases, the reflexed phenotype result confirmed the non-Z/non-S interpretation from the genotype assay (Table 3). The low A1AT concentrations but normal genotyping and phenotyping results suggested that the low A1AT concentrations for these patients were attributable to etiologies other than hereditary A1AT deficiency. In 25 of the samples, the phenotype result demonstrated the presence of only wild-type M alleles. In 1 case, the phenotype assay detected an E allele in conjunction with an M allele. The E allele is a rare wild-type allele and is not associated with A1AT deficiency. The next 2 most commonly identified genotypes were those heterozygous for either an S or a Z allele (Table 3). Two S heterozygotes, which had A1AT values of 0.59 g/L and 0.79 g/L, and 23 Z heterozygotes, all with A1AT concentrations ⬍0.70 g/L, were reflexed to phenotyping. Based on the phenotyping assay, the genotype results for 1 of the S heterozygotes and 17 of the Z heterozygotes were confirmed (Table 3). However, 7 results from the 2

2240


Fig. 1. A1AT deficiency-testing algorithm. When a patient presents with symptoms suggestive of A1AT deficiency disease, the physician would order “A1AT Deficiency Testing”. This testing would initially involve quantification of A1AT concentrations and genotyping. If the results from these 2 assays were concordant, the results and interpretation would be released. For an individual with a non-Z/non-S genotype, the A1AT concentration should be ⬎1.00 g/L. For a Z/non-S or S/non-Z heterozygote, a concordant A1AT quantification would be any value ⬎0.70 g/L. A patient with an S/S genotype would be expected to have an A1AT concentration ⬍1.00 g/L, and a Z/Z homozygote should have an A1AT quantification of ⬍0.70 g/L. For results that are not within these ranges for the genotype and quantification, the sample would then reflex to the phenotype assay. The results of all 3 assays, as well as an interpretation, would then be reported to the physician.

assays were discrepant (Table 4). For the S heterozygote with the A1AT value of 0.59 g/L, phenotyping detected only an S allele, which, in conjunction with the genotype

and A1AT quantification, suggested that the other chromosome carries a null allele. Similarly, for 2 of the Z heterozygotes, only the Z allele was detected by pheno-

Table 3. Retrospective analysis of samples submitted for genotype analysis. Phenotype/genotype comparison Genotype

Number of cases, n (% of total)

A1AT concentration, g/L mean (SD) (Range)

Cases reflexed to phenotyping, n (% of genotype)

Concordant

Discrepanta

Non-Z/non-S Z/non-S S/non-Z Z/Z S/S Z/S Total

785 (63) 253 (20) 110 (8.8) 55 (4.4) 9 (0.7) 34 (2.7) 1246 (100)

1.47 (0.44) (0.67–4.29) 0.85 (0.20) (0.19–1.59) 1.20 0.37 (0.59–2.70) 0.28 (0.22) (0.14–1.79) 0.88 (0.18) (0.61–1.18) 0.60 (0.11) (0.45–0.96)

26 (3.3) 23 (9.1) 2 (1.8) 1 (1.8) 1 (11) 0 (0) 53 (4.3)

26 17 1 0 1 NA 45

0 6 1 1 0 NA 8

a

See Table 4

2241


Table 4. Summary of findings for samples with discordant phenotype and genotype results. Discrepant genotype

A1AT concentration, g/L

S/non-Z Z/non-S Z/non-S Z/non-S

0.59 0.19 0.27 0.64

S Z Z I/Z

Z/non-S

0.67

I/Z

Z/non-S

0.69

I/Z

Z/non-S

0.64

M

Z/Z

1.79

M/Z

typing, also suggesting the presence of a null allele. This corresponded to the A1AT concentration, which was ⬍0.30 g/L in both patients. In 3 of the Z heterozygotes, both the Z and the I allele, which is a partial deficiency allele, were identified by IEF (Table 4), leading to a diagnosis of A1AT deficiency (18 ). For the final Z heterozygote discrepant result, only the M allele was detected by phenotyping, suggesting that the Z allele is not produced or is not released into the circulation at quantities sufficient for detection by the phenotyping assay. The final group of patients in this analysis included those who were identified as having 2 copies of the Z and S alleles, either as homozygotes or compound heterozygotes. This group accounted for 8% of the total cases (Table 3). In each of the homozygote groups, only 1 sample was reflexed to phenotyping. These 2 samples both had A1AT concentrations significantly ⬎1.00 g/L. In the S/S case, phenotyping detected the presence of only the S allele and confirmed the genotyping result. For the Z/Z homozygote, phenotyping identified the presence of an M and a Z allele, which correlated with the A1AT quantification (Table 4). This case is another for which the patient was receiving A1AT replacement therapy.

Discussion In this study, we compared the performance of genotyping and phenotyping in the diagnosis of A1AT deficiency (19, 23 ). We have proposed an algorithm to be used for laboratory testing in A1AT deficiency. This algorithm uses genotyping and A1AT quantification as the first level of testing, with phenotyping as a reflex test to be used in cases in which the A1AT concentration does not fall into the range expected for the given genotype. After establishing the algorithm, we then evaluated its performance. Quantification of A1AT serum concentrations alone is not sufficient to diagnose genetic causes of A1AT deficiency. Variations in the A1AT gene are not the only causes of A1AT deficiency. Secondary causes of A1AT deficiency include liver damage or conditions such as protein-losing enteropathies that cause a general decrease in serum proteins. In cases in which a low concentration of A1AT is attributed to a genetic cause, it is critical to the

Phenotype

Interpretation

Presence of null allele (S/null) Presence of null allele (Z/null) Presence of null allele (Z/null) Presence of deficiency allele not detected by genotype assay Presence of deficiency allele not detected by genotype assay Presence of deficiency allele not detected by genotype assay Z allele not produced or not released into circulation Patient receiving A1AT replacement therapy

clinical management of the patient that the specific variation, particularly the S or Z allele, be identified. Patients with variations that lead to lowered concentrations of A1AT in the serum are at increased risk for developing chronic obstructive pulmonary disease, and individuals with variations that lead to A1AT intracellular polymerization are also at risk for developing liver disease. In addition, the identification of a genetic etiology can be used to screen other at-risk family members for disease or carrier status. Genotyping and phenotyping for A1AT are complementary assays, each with advantages and disadvantages. Phenotyping identifies many different alleles, both common and rare, that are secreted into the blood, but this technique is particularly time-consuming and requires a high degree of technical skill to interpret results. In addition, there are no commercially available controls or reagent sets for the IEF assay. An additional shortcoming is that phenotyping alone cannot distinguish between an individual who is homozygous for a single allele and an individual who is heterozygous for that allele in trans to a null allele. With S- and Z-directed genotyping, the assay is less technically demanding, but it does not directly detect common wild-type alleles or rare alleles such as the I allele, which may have a clinically relevant phenotype (21, 22 ). In our initial analysis of 512 patients, we observed more errors based on phenotyping than genotyping. This finding and the relative ease of performing genotyping prompted us to use genotyping as the first level of testing in our A1AT deficiency algorithm. In our algorithm, the serum concentrations of A1AT are used in conjunction with genotyping assays for interpretation, and in certain situations, to trigger further analysis by phenotyping. Approximately 4% of samples submitted for A1AT genotyping were reflexed to phenotyping. These were all cases in which the serum concentration of A1AT did not correlate with the determined genotype. In 85% of these reflexed cases, the phenotyping result confirmed the accuracy of the genotype assay. This confirmation step allows for more confidence in the resulting diagnosis. In the remaining 15% of the reflexed samples (e.g., 6 per 1000

2242


cases), the genotype result was clarified by phenotyping. In 3 cases, the phenotype assay detected the presence of a rare deficiency allele, the I allele, which was present in conjunction with a Z allele. This finding helped to explain the lower concentrations of A1AT in the serum and to establish a diagnosis of A1AT deficiency (18 ). In certain other cases, the phenotype results indicated the presence of only 1 allele. These data, in conjunction with the A1AT concentration and genotype, suggested the presence of a null allele. In each of the cases that were limited by the genotyping approach, the A1AT serum concentration alerted the laboratory to perform phenotyping, and the results were clarified. Techniques other than A1AT phenotyping that can be considered as reflex tests include whole-gene sequencing and the addition of other alleles to the melting curve genotype assay. Although the sequencing approach would detect any alteration from the wild-type sequence, including novel variations, it is substantially more expensive than IEF analysis. For the A1AT gene, which includes 7 exons, the cost for whole-gene sequencing would be ⬃50-fold higher than the current cost of the phenotyping assay. Adding additional variations to the current melting-curve analysis would be more cost-efficient than a sequencing assay. However, because more than 50 variations have been identified in the A1AT gene, most of them quite rare, this alternative is not very attractive. The ability to detect novel variations (such as the novel variation S47R found in this study), as can be done with the current phenotype assay, would also be lost. The phenotyping assay is the method of choice for the reflex test in our A1AT deficiency algorithm because of its ability to detect a wide variety of alleles in a cost-effective manner. The use of genotyping simplifies the laboratory assessment of A1AT deficiency. Because of the limitations of the genotyping strategy, however, serum concentrations should be measured as a quality-control tool, and phenotype analysis is needed to clarify discordant results. We propose that the use of the described algorithm optimizes the accuracy of laboratory evaluation for A1AT deficiency. This algorithm is consistent with the American Thoracic Society/European Respiratory Society Statement on standards for diagnosis and management of A1AT deficiency.

3.

4. 5. 6. 7.

8. 9.

10.

11.

12.

13.

14. 15.

16.

17. 18.

19.

20.

This work has been partially supported by the National Institute of Health grants NIH-CA-77118 and NIH-CA80127.

21.

22.

References 1. Stoller JK and Aboussouan LS. ␣1-antitrypsin deficiency. Lancet 2005;365:2225–36. 2. DeMeo DL and Silverman EK. ␣1-antitrypsin deficiency. 2: Genetic

23.

aspects of ␣1-antitrypsin deficiency: phenotypes and genetic modifiers of emphysema risk. Thorax 2004;59:259 – 64. Lee WL and Downey GP. Leukocyte elastase: physiological functions and role in acute lung injury. Am J Respir Crit Care Med 2001;164:896 –904. Carrell RW and Lomas DA. Alpha1-antitrypsin deficiency—a model for conformational diseases. N Engl J Med 2002;346:45–53. Carrell RW. Cell toxicity and conformational disease. Trends Cell Biol 2005;15:574 – 80. Lomas DA and Parfrey H. ␣1-antitrypsin deficiency. 4: Molecular pathophysiology. Thorax 2004;59:529 –35. Perlmutter DH, Joslin G, Nelson P, Schasteen C, Adams SP, Fallon RJ. Endocytosis and degradation of ␣ 1-antitrypsin-protein complexes is mediated by the serpin-enzyme complex (SEC) receptor. J Biol Chem 1990;265:16713– 6. Frants RR and Eriksson AW. Alpha1-antitrypsin: common subtypes of Pi M. Hum Hered 1976;26:435– 40. Long GL, Chandra T, Woo SL, Davie EW, Kurachi K. Complete sequence of the cDNA for human ␣1-antitrypsin and the gene for the S variant. Biochemistry 1984;23:4828 –37. Curiel DT, Chytil A, Courtney M, Crystal RG. Serum ␣-1-antitrypsin deficiency associated with the common S-type (glu264-to-val) mutation results from intracellular degradation of ␣-1-antitrypsin prior to secretion. J Biol Chem 1989;264:10477– 86. Turino GM, Barker AT, Brantly ML, Cohen AB, Connelly RP, Crystal AG, et al. Clinical features of individuals with PI*SZ phenotype of ␣ 1-antitrypsin deficiency. Am J Respir Crit Care Med 1996;154: 1718 –25. Dahl M, Hersh CP, Ly NP, Berkey CS, Silverman EK, Nordestgaard BG. The protease inhibitor PI*S allele and COPD: a meta-analysis. Eur. Respir. J 2005;26:67–76. Nukiwa T, Satoh K, Brantly ML, Ogushi F, Fells GA, Courney M, et al. Identification of a second mutation in the protein-coding sequence of the Z type ␣1-antitrypsin gene. J Biol Chem 1986; 261:15989 –94. Lomas DA, Evans DL, Finch JT, Carrell RW. The mechanism of Z ␣-1-antitrypsin accumulation in the liver. Nature 1992;357:605–7. Ogushi F, Fells GA, Hubbard RC, Straus SD, Crystal RG. Z-type ␣ 1-antitrypsin is less competent than M1-type ␣ 1-antitrypsin as an inhibitor of neutrophil elastase. J Clin Invest 1987;80:1366 –74. Hersh CP, Dahl M, Ly NP, Berkey CS, Nordestgaard BG, Silverman EK. Chronic obstructive pulmonary disease in ␣ 1-antitrypsin PI MZ heterozygotes: a meta-analysis. Thorax 2004;59:843–9. Needham M and Stockley RA. ␣1-antitrypsin deficiency. 3: Clinical manifestations and natural history. Thorax 2004;59:441–5. Mahadeva R, Chang WW, Dafforn TR, Oakley DJ, Foreman RC, Calvin J, et al. Heteropolymerization of S, I, and Z ␣1-antitrypsin and liver cirrhosis. J Clin Invest 1999;103:999 –1006. American Thoracic Society/European Respiratory Society Statement: Standards for the diagnosis and management of individuals with alpha-1 antitrypsin deficiency. Am J Respir Crit Care Med 2003;168:818 –900. Kueppers F. Determination of ␣-1-antitrypsin phenotypes by isoelectric focusing in polyacrylamide gels. J Lab Clin Med 1976;88:151–5. von Ahsen N, Oellerich M, Schu¨tz E. Use of two reporter dyes without interference in a single tube rapid-cycle PCR: ␣1-antitrypsin genotyping by multiplex real-time fluorescence PCR with the LightCycler. Clin Chem 2000;46:156 – 61. Braun A, Meyer P, Cleve H, Roscher A. Rapid and simple diagnosis of the two common ␣1-proteinase inhibitor deficiency alleles Pi*Z and Pi*S by DNA analysis. Eur J Clin Chem Clin Biochem 1996; 34:761– 4. Stoller JK. Key current clinical issues in ␣-1 antitrypsin deficiency. Respir Care 2003;48:1216 –21.