In order to assess the value of human cytomegalovirus (HCMV) DNA ... biopsies, we studied 57 human immunodeficiency virus-infected patients with and ...
Vol. 31, No. 8
JOURNAL OF CLINICAL MICROBIOLOGY, Aug. 1993, p. 2066-2069
0095-1137/93/082066-04$02.00/0 Copyright © 1993, American Society for Microbiology
Diagnostic Value of Amplification of Human Cytomegalovirus DNA from Gastrointestinal Biopsies from Human Immunodeficiency Virus-Infected Patients L.
CO1TE,1* E. DROUET,2 F. BISSUEL,1 G. A. DENOYEL,2 AND C. TREPO1
Hepatogastroenterology and AIDS Clinical and Research Unit, Hotel-Dieu Hospital and Institut National de la Sante et de la Recherche Medicale U271, and Virology Unit, Pasteur Institute,2 Lyon, France '
Received 14 December 1992/Accepted 11 May 1993
In order to assess the value of human cytomegalovirus (HCMV) DNA amplification of gastrointestinal biopsies, we studied 57 human immunodeficiency virus-infected patients with and without gastrointestinal HCMV diseases. After DNA extraction, a 406-bp fragment from the unique short region of the HCMV genome was amplified by 35 cycles of polymerase chain reaction (PCR) and semiquantified from 80 to 80,000 HCMV genomic copies. Among 12 non-AIDS patients, the PCR assay was negative for 11 of 12 duodenal and 8 of 8 colorectal samples. It was also negative for 28 of 31 duodenal and 12 of 15 colorectal samples from 31 AIDS patients without gastrointestinal HCMV diseases. Among 14 AIDS patients with gastrointestinal HCMV diseases, the PCR assay was positive for 12 of 12 patients with HCMV duodenitis and for 13 of 13 patients with HCMV colitis. Results were dichotomized between high and low HCMV-DNA copy numbers. For duodenitis, sensitivity was 92% and specificity was 100%. For colitis, sensitivity was 92% and specificity was 93%. Specificity and sensitivity were not influenced by shedding status for HCMV or by other gastrointestinal infections. HCMV DNA amplification of gastrointestinal biopsies is a sensitive and specific tool for the diagnosis of gastrointestinal HCMV diseases in AIDS patients. consent for endoscopies and gastrointestinal biopsies. A gastroscopy was performed for 56 patients, and duodenojejunal biopsies were obtained from each patient for histology, parasitology, bacteriology, and HCMV identification by culture and DNA amplification. Thirty-six patients also underwent rectosigmoidoscopies or colonoscopies with colorectal biopsies. Twelve patients were classified in Centers for Disease Control group II, III, or IVC2 and constituted the non-AIDS control group (group 1). Thirty-one patients presented with AIDS-defining conditions without evidence of gastrointestinal HCMV diseases (group 2). Fourteen patients had gastrointestinal HCMV diseases defined by clinical, endoscopical, and histological evidence of duodenitis or colitis and by either the presence of typical intranuclear inclusions on gastrointestinal biopsies or positive HCMV isolation from biopsies (group 3). All biopsies were obtained before the initiation of systemic anti-HCMV ther-
Human cytomegalovirus (HCMV) diseases are a major of morbidity and mortality in patients with AIDS. The most common manifestation of HCMV diseases in these patients is retinitis, but gastrointestinal involvement is reported for 2.2 to 13.1% of AIDS patients (3, 6, 8, 9) and disseminated diseases are found at autopsy in up to 90% of patients (4). Since both standard histology and virus isolation by culture may lack sensitivity (2, 6), the diagnosis of gastrointestinal HCMV diseases from endoscopical biopsies is often difficult. Since the efficacy of antiviral therapy is influenced by early administration, rapid and specific methods may be developed for an early diagnosis. Recent studies have demonstrated the high sensitivity of in situ hybridization for HCMV with gastrointestinal biopsies (1, 10, 15), but the method is technically complex and cannot be adapted to routine analysis. Detection of HCMV DNA in blood and urine by polymerase chain reaction (PCR) has been previously reported for immunosuppressed and AIDS patients (7, 11). We have therefore evaluated the diagnostic value of HCMV DNA amplification by PCR of gastrointestinal biopsies obtained from a well-defined population of 57 patients infected with human immunodeficiency virus (HIV) with and without gastrointestinal HCMV diseases. cause
apy.
HCMV DNA amplification. On the basis of published the AD169 strain (14) and previous studies conservation (12), two 25-base oligonucleotide primers selected in the HindIII-X fragment were synthesized. The primer pair (sequences 5' to 3': upstream primer, GGATCCGCATGGCATTCACGTATGT; downstream primer, GAATTCAGTGGATAACCTGCGGCGA) frame a 406-bp fragment between restriction sites BamHI and EcoRI from the unique short region of the HCMV genome and gave consistantly negative results when assayed with related human herpesvirus DNA (herpes simplex types 1 and 2, Epstein-Barr virus, varicella-zoster virus, and human herpesvirus type 6). For amplification, DNA from fresh biopsies was prepared by RNase and proteinase K digestion, phenol extraction, and ethanol precipitation. After being diluted in 10 mM Tris-1 mM EDTA, DNA was quantitated with a 260-nm spectrophotometer and was adjusted for amplification at an optical density of 1. A fixed sequence data for on restriction site
MATERIALS AND METHODS Patients. From October 1991 to November 1992, 57 HIV type 1-infected patients attending the hepatogastroenterology and AIDS unit of the Hotel-Dieu Hospital (Lyon, France) for gastrointestinal investigations were prospectively enrolled. The major reason for gastrointestinal endoscopies was chronic diarrhea in 45 patients; this diarrhea was associated with fever in 30 patients. Other reasons included abdominal pain and portal hypertension. All patients gave *
Corresponding author. 2066
HCMV DNA AMPLIFICATION OF GASTROINTESTINAL BIOPSIES
VOL. 31, 1993
-
1 2 3
abc
FIG. 1. HCMV DNA amplification of gastrointestinal biopsies from an AIDS patient with HCMV colitis and a normal duodenum. Amplification of a 406-bp fragment from the unique short region of the HCMV genome and hybridization with an internal probe labelled with 11dUTP-digoxigenin after Southern blotting are shown. The amplified fragment is marked by the arrowhead. Lanes: 1, +3 PCR assay of colorectal biopsy; 2 and 3, negative PCR assay of duodenojejunal biopsies; a, b, and c, reference standards corresponding to 80,000, 8,000, and 800 genomic HCMV DNA copies. amount (0.1 ,ug) of extracted DNA was subjected to amplification in 100 Il of reaction buffer consisting of 50 mM KCl, 10 mM Tris HCl (pH 8.4), 1.5 mM MgCl2, 100 jg of gelatin per ml, 200 ,uM (each) deoxynucleotide triphosphate, 1 ,uM (each) primer, and 2.5 U of Taq DNA polymerase (Boehringer, Mannheim, Germany). After the first denaturation at 99°C for 10 min, the PCR was performed in a thermal cycler for 35 cycles (annealing, 1 min at 55°C; primer extension, 1 min at 72°C; denaturation, 30 s at 95°C). The PCR product was electrophoresed in 3% Nusieve agarose (FMC) and 1% SeaKem agarose (FMC) and transferred from the gel to a nylon membrane by the Southern technique. The hybridization probe was a 180-bp fragment between two Sau3AI sites within the 406-bp fragment labelled with 11-dUTP-digoxigenin. As a control of internal consistency and for semiquantitative analysis of the PCR product, six reference standards were included in each PCR experiment. These standards were prepared by adding known amounts of purified HCMV
2067
AD169 DNA (approximately equivalent to from 0.8 to 80,000 HCMV genomic copies) to DNA from 104 MRC-5 cells. Within this range, the linearity of the standard curves was stated for a total number of between 25 and 35 amplification cycles. For objective evaluation, Southern blots were analyzed by a computer image-processing system designed for densitometric measurements (Biocom 200, Les Ulis, France). The specific band resulting from each PCR-positive specimen was assigned a score of from + 1 to +4 on the basis of band intensity corresponding to four points on the standard dilution curve: 80, 800, 8,000, and 80,000 HCMV genomic copies (Fig. 1). The +1 and +2 results were defined as low-level HCMV DNA, while the +3 and +4 results were defined as high-level HCMV DNA. In parallel with DNA amplification, conventional viral cultures and shell vial cultures (13) were performed with gastrointestinal biopsies and with blood and urine samples. All specimens were processed blindly with respect to clinical data.
RESULTS The major clinical and biological findings in the three groups are summarized in Tables 1 and 2. The patients' ages ranged from 20 to 61 years, with a mean age of 38 years. Fifty-four patients were male, and 3 were female. The major risk factor for acquiring HIV was male homo- or bisexuality for 44, intravenous drug addiction for 9, and heterosexual transmission for 4 patients. Control groups 1 and 2: non-AIDS patients and AIDS patients without gastrointestinal HCMV diseases. All of the non-AIDS patients and 28 AIDS patients without gastrointestinal HCMV diseases remained free of HCMV diseases after follow-up 4 months later. One patient presented with HCMV retinitis at the initial examination, and two others developed HCMV retinitis 3 and 4 months, respectively, after endoscopies. PCR was negative for duodenal biopsies from these three patients, while a rectal biopsy revealed low-level HCMV DNA (+1) in one patient. Gastrointestinal symptoms were related to other enteric pathogens in 13 AIDS patients (Cryptosporidium spp. in seven, Giardia intestinalis in two, Entamoeba histolytica in one, Isospora belli in one, Enterocytozoon bieneusi in one, and Mycobac-
TABLE 1. Clinical and biological data from 57 HIV-infected patients with and without gastrointestinal HCMV diseases Characteristic
CD4+/mm3
(range)
P24 Age (positive assay/total no. of patients)
HCMV shedding (positive assay/total no. of patients) Urine Blood Lungs Total
Symptoms Diarrhea Fever a Non-AIDS patients. b AIDS patients without gastrointestinal HCMV diseases. AIDS patients with gastrointestinal HCMV diseases. d Results are means + standard deviations. P24 antigenemia.
Group ia (n = 12)
Result for: Group 2" (n = 31)
Group 3C
(n =14)
325 ± 225 (73-809) 2/12
76 ± 98 (0-386) 12/31
28 + 27
3/10 0/10 1/2 3/10
8/28 7/29 6/16 12/30
8/8 12/13 8/8 12/13
8 3
23 16
14 11
5/13
2068
J. CLIN. MICROBIOL.
COTlFE ET AL.
TABLE 2. PCR results from gastrointestinal biopsies Result from biopsy
Histology Inflammation Inclusions
Culture (positive/total no. of biopsies)
Group 3a
Group 2
Group 1
Test or characteristic
Rectal (n = 15)
Duodenal (n = 12)
Rectal (n = 13)
Duodenal (n = 12)
Rectal (n = 8)
Duodenal (n = 31)
5 0
6 0
21 0
7 0
12 4
13 7
0/12
0/8
0/31
0/15
9/12
7/13
11 1 0 0
28 2 1 0 0 3/31
12 2 0 1 0 3/15
0 0 1 7
0 1 0 2 10
12/12
13/13
0/31
1/15
11/12
12/13
PCR resultsb 0 +1 +2 +3 +4 Total
0 1/12
8 0 0 0 0 0/8
HCMV DNA >8,000 copiesc
0/12
0/8
4
a All members of group 3 presented with duodenitis and/or colitis. b Scores of from +1 to +4 correspond to approximately 80 to 80,000 HCMV DNA genomic copies. c Results correspond to the highest HCMV DNA levels (+3 or +4 PCR results).
terium avium in one), including 1 patient with a +2 PCR result from the duodenal biopsy and 1 patient with a +3 PCR result from the rectal biopsy. The six control patients who gave positive PCR results from biopsies were all severely standard deviation CD4+ cell immunodepressed (mean count, 54 51 cells per mm3), and five of them also shedded HCMV in urine samples, blood samples, or bronchoalveolar fluid. Group 3: AIDS patients with gastrointestinal HCMV diseases. All 14 patients were male, and all but one were homoor bisexual. They all had previous AIDS-defining conditions before the onset of HCMV diseases. At the onset of gastrointestinal symptoms, six patients had typical HCMV retinitis and three others had evidence of disseminated HCMV diseases and encephalitis. All patients reported severe diarrhea, which was associated with fever above 38.5°C in 11 patients. Two patients subsequently developed peritonitis. Eleven patients had diffuse gastrointestinal HCMV diseases with endoscopical and histological signs of both duodenitis and colitis. One patient presented with isolated duodenitis, while two patients had severe colitis with normal duodena (Fig. 1) and were considered in the AIDS control group (group 2) for the calculation of the specificity of PCR assay for the diagnosis of HCMV duodenitis. The sensitivity of a positive PCR result from duodenal biopsies for the diagnosis of HCMV duodenitis was 100% (12 of 12 biopsies), and the specificity was 88% (29 of 33 biopsies). The sensitivity of a positive PCR result from colorectal biopsies for the diagnosis of HCMV colitis was 100% (13 of 13 biopsies), and the specificity was 80% (12 of 15 biopsies). To increase specificity, PCR results were dichotomized between those with negative and low-level HCMV DNA (negative, +1 and +2 results, corresponding to fewer than 800 HCMV DNA genomic copies) and those with high-level HCMV DNA (+3 and +4 results, corresponding to 8,000 or more HCMV DNA genomic copies). Specificity therefore increased to 100% (33 of 33 biopsies) for duodenitis and to 93% (14 of 15 biopsies) for colitis, while sensitivity decreased to 92% for both duodenitis (11 of 12 biopsies) and colitis (12 of 13 biopsies).
Since group 2 and group 3 were not comparable in terms of immunosuppression (CD4+ cell count [mean + standard deviation], 76 ± 98 and 28 ± 27 cells per mm3, respectively) and in terms of shedding status for HCMV (patients with positive blood, urine, or bronchoalveolar fluid culture for HCMV out of total number of patients, 12 of 30 and 12 of 13), we compared the 14 patients from group 3 with the 12 patients from group 2 who shedded HCMV. The two groups were therefore similarly immunosuppressed (CD4+ cell count, [mean ± standard deviation], 28 ± 27 and 31 ± 29, P = 0.7 by Student's t test). For dichotomized data, specificities were therefore 100% (12 of 12 biopsies) for duodenitis and 83% (5 of 6 biopsies) for colitis. DISCUSSION HCMV has the remarkable property of remaining in a latent phase in many tissues following acute infection. Almost all AIDS patients are susceptible to HCMV reactivation and clinical diseases according to the progression of the immune defect. Because of the difficulties encountered in differentiating the carrier state of HCMV and HCMV diseases, there is no "gold standard" for the diagnosis of organ involvement in HCMV diseases, and histology alone cannot be considered a reliable standard. The discrepancy between the high prevalence of disseminated HCMV diseases demonstrated by autopsy studies (4) and the relatively low prevalence of organ involvement diagnosed before death (6, 8) may be explained by the small amount of tissue examined with endoscopical biopsy samples. Even with patients with documented gastrointestinal diseases, 37.5% of biopsies fail to demonstrate inclusions (6). The location of the biopsies may also affect the diagnosis, since preferential cecal involvement has been previously reported in some studies (3, 9). Immunocytochemical techniques have also demonstrated that typical intranuclear inclusions (Cowdry type A) are far less common than atypical cytopathic effect (6) and can be confused with herpes simplex virus pathological effect. Compared with HCMV isolation from gastrointestinal biop-
VOL. 31, 1993
HCMV DNA AMPLIFICATION OF GASTROINTESTINAL BIOPSIES
sies, histology has a sensitivity of 30 to 54% (1, 5). The sensitivities of both histology and culture are also low (38 to 58 and 55%, respectively) compared with that of in situ hybridization (1, 10, 15). The diagnosis of gastrointestinal HCMV diseases relies, therefore, on the conjunction of clinical, histological, and virological evidence rather than on the result of a single assay. We have used these criteria to define our study and control groups, and we have therefore studied the value of PCR assay to diagnose gastrointestinal HCMV diseases in these populations. The sensitivity and the specificity of the technique are very high, especially when only the highest levels of HCMV DNA are considered. Factors which could induce false-positive results are the severity of immunodeficiency and the shedding status for HCMV. However, these two factors did not affect sensitivity and specificity in this study. Another parameter which could explain false-positive results is gastrointestinal inflammation caused by other pathogens. In our study, two patients from group 2 with G. intestinalis and I. belli infections presented with low-level HCMV DNA from biopsies. The significance of such results is unclear, since these patients had no evidence of gastrointestinal HCMV diseases during the follow-up visit. However, 11 other AIDS patients with other causes of gastrointestinal inflammation had negative PCR results, suggesting that this phenomenon was of minor importance. The specificity of the assay was also enhanced by a negative result from one patient who presented with HCMV retinitis and no evidence of gastrointestinal HCMV diseases. Similarly, two patients presenting with colitis and normal duodena had high-level HCMV DNA from colorectal biopsies, while duodenal biopsies gave negative or low-level HCMV DNA results. In conclusion, HCMV DNA amplification of gastrointestinal biopsies appears a promising, highly sensitive, and specific tool for the diagnosis of gastrointestinal HCMV diseases. Results may be obtained within a few days, allowing the prompt initiation of antiviral therapy. Further studies are required to assess its value for monitoring antiviral therapy. REFERENCES 1. Clayton, F., E. B. Klein, and D. P. Kotler. 1989. Correlation of in situ hybridization with histology and viral culture in patients with acquired immunodeficiency syndrome with cytomegalovirus colitis. Arch. Pathol. Lab. Med. 113:1124-1126. 2. Culpepper-Morgan, J. A., D. P. Kotler, J. V. Scholes, and A. R. Tierney. 1987. Evaluation of diagnostic criteria for mucosal cytomegalic inclusion disease in the acquired immune deficiency syndrome. Am. J. Gastroenterol. 82:1264-1270.
2069
3. Dieterich, D. T., and M. Rahmin. 1991. Cytomegalovirus colitis in AIDS: presentation in 44 patients and a review of the literature. J. Acquired Immune Defic. Syndr. 4:S29-S35. 4. Drew, W. L. 1992. Cytomegalovirus infection in patients with AIDS. Clin. Infect. Dis. 14:608-615. 5. Finet, J. F., B. Aymard, J. G. Georges, M. A. Bigard, M. Kessler, and J. Floquet. 1991. Interet (reel mais limite) de l'immunohistochimie pour la detection precoce du cytomegalovirus dans la muqueuse gastroduodenale des transplantes r6naux. Ann. Pathol. 11:141-142. 6. Francis, N. D., A. W. Boylston, A. H. G. Roberts, J. M. Parkin, and A. J. Pinching. 1989. Cytomegalovirus infection in gastrointestinal tracts of patients infected with HIV-1 or AIDS. J. Clin. Pathol. 42:1055-1064. 7. Hsia, K., D. H. Spector, J. Lawrie, and S. A. Spector. 1989. Enzymatic amplification of human cytomegalovirus sequences by polymerase chain reaction. J. Clin. Microbiol. 27:1802-1809. 8. Jacobson, M. A., and J. Mills. 1988. Serious cytomegalovirus disease in the acquired immunodeficiency syndrome (AIDS): clinical findings, diagnosis and treatment. Ann. Intern. Med. 108:585-594. 9. Rene, E., C. Marche, T. Chevalier, C. Rouzioux, B. Regnier, A. G. Saimot, Y. Negesse, S. Matheron, C. Leport, B. Wolff, B. Moriniere, C. Katlama, B. Godeberge, B. Vittecoq, F. Bricaire, C. Brun-Vesinet, B. Pangon, A. M. Deluol, J. P. Coulaud, J. Modai, J. Frottier, J. L. Vilde, F. Vachon, M. Mignon, and S. Bonfils. 1988. Cytomegalovirus colitis in patients with acquired immunodeficiency syndrome. Dig. Dis. Sci. 33:741-750. 10. Roberts, W. H., S. Hammond, J. M. Sneddon, J. Thesing, J. H. Calwell, and K. P. Clausen. 1988. In situ DNA hybridization for cytomegalovirus in colonoscopic biopsies. Arch. Pathol. Lab. Med. 112:1106-1109. 11. Shibata, D., W. J. Martin, M. D. Appleman, D. M. Causey, J. M. Leedom, and N. Arnheim. 1988. Detection of cytomegalovirus DNA in peripheral blood of patients infected with human immunodeficiency virus. J. Infect. Dis. 158:1185-1192. 12. Somogyi, T., R. Colimon, C. Bertrand, and S. Michelson. 1987. Conservation and map location of human cytomegalovirus strain AD 169 transforming sequences in the DNA clinical isolates. Microbiologica 98:125-131. 13. Stirk, P. R., and P. D. Griffiths. 1987. Use of monoclonal antibodies for the diagnosis of cytomegalovirus infection by the detection of early antigen fluorescent foci (DEAFF) in cell culture. J. Med. Virol. 21:329-337. 14. Weston, K., and B. G. Barrell. 1986. Sequence of the short unique region, short repeats and part of the long repeats of human cytomegalovirus. J. Mol. Biol. 192:177-208. 15. Wu, G. D., I. P. Shintaku, K. Chien, and S. A. Geller. 1989. A comparison of routine light microscopy, immunohistochemistry and in situ hybridization for the detection of cytomegalovirus in gastrointestinal biopsies. Am. J. Gastroenterol. 84:1517-1520.
