- Email: [email protected]
qPCR increases sensitivity to detect cytomegalovirus in formalin-fixed, paraffin-embedded tissue of gastrointestinal biopsies
Human Pathology (2014) 45, 48–53
www.elsevier.com/locate/humpath
Original contribution
qPCR increases sensitivity to detect cytomegalovirus in formalin-fixed, paraffin-embedded tissue of gastrointestinal biopsies☆,☆☆ Morgan H. McCoy PhD, MD a , Kristin Post MPH b , Joyashree D. Sen MD b , Hsim Y. Chang BS b , Zijin Zhao MD a , Rong Fan MD, PhD a , Shaoxiong Chen MD, PhD a , Diane Leland PhD a , Liang Cheng MD a , Jingmei Lin MD, PhD a,⁎ a
Department of Pathology and Laboratory Medicine, Indiana University School of Medicine, Indianapolis, IN 46202, USA Indiana University Health, Indianapolis, IN 46202, USA
b
Received 7 June 2013; revised 29 July 2013; accepted 31 July 2013
Keywords: qPCR; Cytomegalovirus; Colon; Formalin-fixed, paraffin-embedded tissue
Summary Histopathologic diagnosis of gastrointestinal (GI) tract cytomegaloviral (CMV) infection relies on hematoxylin and eosin (H&E)–stained tissue, along with the aid of immunohistochemistry (IHC). However, non-classic appearing inclusions or atypical IHC staining patterns remain an ongoing concern for pathologists. We reported the use of real-time polymerase chain reaction (qPCR) on nucleic acid extracted from paraffin-embedded, formalin-fixed tissue of GI biopsies from cases of CMV infection (n = 91) diagnosed by H&E and IHC. Seventy-nine biopsies, including normal colon biopsies (n = 35), active colitis (n = 25), and active duodenitis (n = 19), were used as negative controls. Of 91 CMV-positive biopsies diagnosed by histology, 88 tested positive by qPCR, with a sensitivity of 96.7%. Of 79 negative controls, 78 were negative and 1 positive by qPCR, resulting in a specificity of 98.7%. Of the cases that were positive for CMV by histopathology, there were an additional 40 biopsies taken from these patients either during the same or previous procedures, some taken just days prior, which were negative for CMV by histology. Interestingly, 22 (55%) of these biopsies tested positive by qPCR, which correlated well with additional clinical CMV results. By defining qPCR as the “gold standard” for a CMV result, histology (H&E and/or IHC) had a sensitivity and specificity of 79% and 97%, respectively. Eighteen biopsies were found negative by H&E and equivocal by IHC. Among them, 14 (78%) tested positive for CMV by qPCR, which also correlated well with additional clinical results. qPCR is a sensitive, specific, and rapid molecular tool that may be helpful to aid in early diagnosis of CMV infection on equivocal or clinically highly suspicious small GI biopsies. © 2014 Elsevier Inc. All rights reserved.
1. Introduction ☆
Funding: This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. ☆☆ Conflict of Interest: The authors have no conflicts of interest to disclose. ⁎ Corresponding author. Department of Pathology and Laboratory Medicine, Indiana University School of Medicine, Indianapolis, IN 46202. E-mail address: [email protected] (J. Lin). 0046-8177/$ – see front matter © 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.humpath.2013.07.040
Cytomegalovirus (CMV), a member of the Betaherpesvirinae subfamily of herpesviruses, is frequently identified in a variety of clinical specimens from many body sites. Similar to other herpesviruses, CMV maintains the ability to establish persistent or latent infection in its host [1]. These dormant stages
qPCR increases sensitivity to detect cytomegalovirus
49
are interrupted by periods of active viral replication known as reactivation, characterized by virion release and which may, or may not, result in disease manifestations [2–4]. Reactivation may occur during periods of stress or other causes of immunosuppression, resulting in detectable virus by either viral culture or molecular methods from whole blood or plasma. It is crucial to identify active CMV infections in immunosuppressed patients, such as those with HIV/AIDS, on chemotherapy, or receiving organ/stem cell transplants, given their greater risk for developing severe infection. The diagnosis of CMV colitis can be made by multiple methods, including viral culture from biopsy material and histopathologic examination including immunohistochemical (IHC) staining. Identification of CMV infection by hematoxylin and eosin (H&E)–stained tissue sections relies on classic CMV viral inclusions (Fig. 1A). Unfortunately viral inclusions are not always apparent, even when CMV infection is clinically anticipated. Thus, IHC is frequently employed to aid H&E (Fig. 1B). However, IHC can also be hampered by rare non–classic appearing cellular staining patterns, leading to an equivocal interpretation (Fig. 1C). Such equivocal interpretations puzzle pathologists and clinicians. Therefore, we sought to employ molecular methods to detect CMV in gastrointestinal (GI) biopsies as an adjunct to atypical H&E and IHC. In addition, we compared the molecular results to traditional histologic evaluation by correlating with other available clinical test results, ie, CMV viral culture and molecular viremia assays.
2. Methods With the approval of the institutional review board, a search of our electronic laboratory database was performed to identify formalin-fixed, paraffin-embedded blocks representing cases of CMV infection in GI biopsies (n = 91) diagnosed by histopathology (H&E and/or IHC). Seventy-nine cases, including normal colon (n = 35), active colitis (n = 25), and active duodenitis (n = 19), negative for viral inclusions on H&E and negative for CMV by IHC, were used as negative controls. Two pathologists reviewed all cases (M.M. and J.L.). Of the cases that were positive for CMV by histopathology, there were an additional 40 biopsies taken from these patients either during the same (n = 25) or previous procedures (n = 15), some taken just days prior, which were interpreted as negative for CMV by H&E and IHC. Additionally, eleven equivocal cases were identified. Equivocal biopsies were defined as biopsies with no viral inclusions on H&E and with rare isolated cells demonstrating weak or incomplete cellular CMV IHC staining (Fig. 1C). In addition, seven equivocal biopsies, unrelated to the 91 CMV colitis cases, were also identified in our database. Therefore, in total 18 equivocal cases for CMV were included in the study. Totally, 228 blocks were identified and subjected to realtime polymerase chain reaction (qPCR) for CMV. Briefly,
Fig. 1 CMV colitis. A, Hematoxylin and eosin stained section of a colon biopsy demonstrating extensive involvement with typical CMV inclusions (original magnification ×400; indicated by arrows). B, A colon biopsy stained with a monoclonal antibody against CMV demonstrating a typical IHC staining pattern (original magnification ×400). C, A colon biopsy demonstrating equivocal IHC staining, characterized by a single, small weakly stained cell (original magnification ×400; indicated by arrow).
from each block, 5-10-micron-thick scrolls of tissue were cut, and DNA extractions were performed according to the manufacturer’s specifications (QIAamp DNA FFPE Tissue
50 Kit, Qiagen, Valencia, CA). The extracted DNA was then spiked with an internal reaction control and analyzed by realtime PCR using artus CMV LightCycler PCR reagents and primers (Qiagen) on the LightCycler 2.0 (Roche, Indianapolis, IN). This assay is intended to quantify CMV DNA by amplifying a 105 base-pair fragment of CMV immediateearly DNA polymerase gene. Detection occurs by the specific probe hybridization/dual probe FRET. The result was reported as positive or negative for CMV. For IHC staining, a commercially available monoclonal antibody specific for CMV was used per the manufacturer’s specifications (Dako, Carpinteria, CA; 1:1 dilution). To correlate with clinical findings, a search through the electronic medical record was performed on each case, including CMV identification by whole blood/plasma molecular methods or by viral culture from biopsy material or whole blood performed on clinical specimens submitted within 7 days of the biopsy collection.
3. Results 3.1. Sensitivity and specificity Sensitivity and specificity were calculated in two ways: (1) by using traditional histology, that is, typical viral inclusions on H&E and/or positive for typical staining by CMV IHC, as the “true positives” with “true negatives” lacking both, and (2) by using the CMV qPCR result as determining the “true positives and negatives.” 3.1.1. Traditional histology as the “gold standard” Of 91 biopsies positive for CMV by histopathology, 88 were positive by qPCR (Table 1). With a true positive definition based on traditional histology, that is, typical CMV viral inclusions and/or typical CMV IHC, these data resulted in a sensitivity of 96.7%. Of these 91 cases, 76 had CMV testing on additional specimens submitted at or within 7 days of the time the biopsy was taken. Among them, 62 (82%) had associated clinical specimens that also tested positive for CMV. Of 79 negative control biopsies, that is, negative for viral inclusions and negative by CMV IHC, 78 tested negative and 1 tested positive by qPCR, resulting in a specificity of 98.7%. This biopsy that tested positive by qPCR showed active duodenitis and was collected from a patient with a long standing history of HIV with poor highly active antiretroviral therapy compliance, end-stage liver disease secondary to hepatitis C virus cirrhosis, and type II diabetes mellitus. Upon presentation to the hospital, this patient was found to have sepsis with a duodenal ampullary mass, which was biopsied. In the year prior to admission, the patient’s documented CD4 T-cell counts were in the mid 100s/mm3 and 129 cells/mm3 one month following discharge from this hospitalization. Importantly, none of the true normal biopsies (n = 35) tested positive by qPCR.
M. H. McCoy et al. Table 1 qPCR results of 91 positive CMV cases in correlation with other clinical tests qPCR
Other CMV specimen b
n
+ + + − − Total
+ − N/A c − + 88 a qPCR+
61 12 15 2 1 91
a
The sensitivity of CMV by qPCR is 96.7%. Whole blood/plasma molecular methods and viral culture from gastrointestinal biopsy material or whole blood. Each was performed on clinical specimens submitted within 7 days of the biopsy collection date. c N/A refers to those biopsies with no additional specimens submitted for CMV testing within the defined 7-day time frame. b
3.1.2. qPCR as the “gold standard” Of 111 biopsies positive for CMV by qPCR, 88 tested positive by traditional histology (H&E and CMV IHC), resulting in a sensitivity of 79% for traditional histology at detecting CMV in GI biopsies. Of the 99 biopsies negative for CMV by qPCR, 96 tested negative by traditional histology resulting in a specificity of 97%.
3.2. qPCR results on additional histologically negative specimens Of 91 cases that were positive for CMV by histopathology, there were an additional 40 biopsies taken from these patients either during the same (n = 25) or previous procedures (n = 15), some taken just days prior, which were interpreted as negative for CMV by H&E and IHC. Interestingly, 22 (55%) of these biopsies tested positive for CMV by qPCR. As shown in Table 2, within the group of 40 biopsies, 15 were from eight patients submitted on a previous day that were negative for CMV by H&E and IHC. Of these 15 biopsies, 9 (60%) tested positive for CMV by qPCR. Among these 15 biopsies, 13 had more information regarding CMV (cases 26, 8) on additional clinical specimens (viral culture or molecular methods) submitted at or within 7 days of the time the biopsy was taken. Seven of these 8 patients (9 out of 15 biopsies) had biopsy material that tested positive by qPCR. Among these 9 qPCR positive biopsies, 7 were from 5 patients who had additional clinical specimens tested. Four patients had additional negative tests for CMV (Table 2; cases 3, 5, 6 and 8). One patient had two biopsies that tested positive for CMV by qPCR with additional clinical specimens testing positive (case 4). One patient had three biopsies that tested negative for CMV by qPCR with additional clinical specimens testing positive (case 2). Thus, qPCR results, compared to conventional histology, of the prior biopsies correlated better with the subsequent CMV diagnosis. Therefore, qPCR results likely represent the true CMV status in these histologically negative specimens.
qPCR increases sensitivity to detect cytomegalovirus
51
Table 2 The qPCR result of 15 prior negative biopsies from eight patients who had positive CMV biopsies in correlation with other clinical tests Case number
Other CMV specimen a
Fraction of CMV qPCR positive biopsies in previous submission
Blood/plasma molecular 1 2 3 4 5 6 7 8 Percent positive
1/1 0/3 1/1 2/2 1/2 2/4 1/1 1/1 9/15 (60%)
N/A + − + − N/A N/A −
b
Biopsy viral culture N/A − − N/A − − N/A N/A
a
Whole blood/plasma molecular methods and viral culture from gastrointestinal biopsy material or whole blood. Each was performed on clinical specimens submitted within 7 days of the biopsy collection date. b N/A refers to those biopsies with no additional specimens submitted for CMV testing within the defined 7-day time frame.
3.3. qPCR results on histologically equivocal specimens As illustrated in Table 3, 14 of the 18 IHC equivocal biopsies tested positive for CMV by qPCR (78%). Among these 18 equivocal cases, eleven had additional biopsies taken the same day that were positive for CMV by histopathology. Among these 14 qPCR positive samples, 10 had biopsies taken the same day that were positive for CMV on histopathology. Of the 18 equivocal biopsies tested, 11 (cases 1-5, 10, 12, 13, 15-17) had additional clinical information regarding CMV testing on other specimens submitted at or within 7 days of the time the biopsy was taken. Among these 11 cases, 8 (73%) tested positive for CMV by qPCR. Of these 11 biopsies, 5 (45%) tested positive by qPCR with additional positive CMV specimens (cases 1-5); 3 (27%) tested positive by qPCR with no positive CMV specimens (cases 10, 12, 13); 3 (27%) tested negative by qPCR with additional negative CMV specimens (cases 15-17). As more specifically demonstrated in Table 3, of the 11 equivocal biopsies with additional clinical specimens submitted (cases 1-5, 10,12, 13, 15-17), 10 were GI biopsies collected on the same day and submitted for CMV early antigen detection by viral culture (cases 1-5, 12, 13, 15-17). Five of those 10 samples had positive CMV early antigen viral cultures, all of which had corresponding biopsies that tested positive by qPCR (cases 1-5). Of the remaining 5 negative CMV early antigen viral cultures: 2 had negative CMV molecular results from whole blood/ plasma and tested positive for CMV by qPCR; the other 3 had negative CMV molecular results from whole blood/ plasma and also negative by qPCR. Therefore, qPCR results correlate better with concurrent clinical data compared to conventional histology. Thus, qPCR results likely represent the true CMV status in these IHC equivocal biopsies.
4. Discussion Detection of CMV has changed greatly as medicine continuously demands faster and more accurate diagnostic tools and technologies. Serology, once used frequently, is of limited use since the virus is so prevalent in the population (at least 40% seropositivity) [5]. However, serology is still helpful in immunocompromised hosts in determining whether patients
Table 3 Equivocal biopsies by histology in correlation with qPCR results and other clinical specimens tested for CMV Case number
qPCR
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
+ + + + + + + + + + + + + + − − − −
Other CMV specimen a Blood/plasma molecular
Biopsy culture
+ N/A b N/A + N/A N/A N/A N/A N/A − N/A − − N/A − − − N/A
+ + + + + N/A N/A N/A N/A N/A N/A − − N/A − − − N/A
a Whole blood/plasma molecular methods and viral culture from gastrointestinal biopsy material or whole blood. Each was performed on clinical specimens submitted within 7 days of the biopsy collection date. b N/A refers to those biopsies with no additional specimens submitted for CMV testing within the defined 7-day time frame.
52 are at risk for acquiring a new CMV infection (seronegative) or reactivation (seropositive). Of note, seropositivity does not preclude the possibility of a new infection with an antigenically dissimilar CMV virus strain. Serology specific for IgM antibody is useful when there is a concern for acute infection, although IgM levels may persist for months following exposure [6]. Culture for CMV was once the gold standard for detection of antigen from many body sites including respiratory specimens, whole blood, urine, or biopsy material. However, traditional culture is problematic due to the length of time (1-6 weeks) required to observe typical cytopathic effect on fibroblast monolayer. The advent of early antigen detection by shell vial culture has shortened the time to 2 to 3 days, and its use has become standard [7]. Viremia assays that test for either antigen or nucleic acid in leukocytes, whole blood, or plasma specimens have become clinically useful. Quantitative nucleic acid assays are essential in laboratory diagnostics, particularly in immunocompromised hosts. These techniques provide not only the most sensitive methods to detect CMV in blood specimens, but also provide a quantitative measurement for disease progression and treatment effect [8–10]. Clinically, we have encountered examples of patients who lacked viremia by qPCR, but the following tissue biopsies showed convincing CMV infection. Generally speaking, blood is a good source to detect infection if patients have systemic involvement. However, in localized infections, CMV virus concentration might be higher in infected tissue than that found in blood. Although quantitative nucleic acid assays are extremely sensitive for infections involving viremia, their use in localized infections, such as CMV colitis, has not been adequately evaluated. In this study, we employed a qPCR method to identify CMV in GI biopsies. qPCR was utilized as opposed to conventional PCR methods, due to the availability of commercial reagents including primers and standards, the ability to automate many steps in the procedure, as well as the intrinsic increased sensitivity of qPCR compared to conventional PCR methods. We demonstrate that qPCR is a rapid, sensitive, and specific method for detecting CMV in formalin-fixed, paraffin-embedded GI biopsy tissues. This is most evident by calculating the sensitivity for traditional histology (79%) in detecting CMV in GI biopsies with a “true positive” defined as positive by qPCR. Specificity calculations were similar whether a true negative was defined by traditional histology or qPCR. qPCR demonstrated a specificity of 98.7% using traditional histology as a “true negative” definition. Traditional histology demonstrated a 97% specificity using qPCR as a “true negative” definition. Causes for the failure of qPCR to detect CMV in three of the histopathologically positive CMV biopsies are uncertain. The effect of PCR reaction inhibitors in the specimens was largely ruled out due to the use of an internal control added during the nucleic acid extraction step. Too little tissue for nucleic acid extraction or laboratory error during processing are potential explanations for these false negative results.
M. H. McCoy et al. Since the “falsely positive” biopsy was from a patient who had well documented immunosuppression, the detection of CMV could represent reactivated CMV replication, as well as possible detection of latent virus. One important contribution of the current study is that we evaluated the prior negative biopsies from patients with subsequent positive CMV infection in a short time interval. Interestingly, over half of the histologically negative biopsies from 7 of the 8 patients tested positive for CMV by qPCR. Given that subsequent biopsies from these patients were histologically positive, the qPCR results from biopsy tissue might represent actual CMV infections that were not detected by histology, viral culture from GI biopsy tissue, or qPCR from blood/plasma. Another important contribution is that we evaluated the IHC equivocal biopsies, since they pose a diagnostic challenge to pathologists. In this study, the CMV results by qPCR better correlate with additional clinical data. Compared to conventional methods, qPCR is more sensitive in diagnosing CMV infection in GI biopsy material. In such cases, a positive qPCR result would have resulted in an earlier diagnosis.
Fig. 2 Proposed algorithm utilizing qPCR on formalin-fixed, paraffin-embedded GI biopsy tissue for working up of CMV infection. Abbreviations: POS, positive; NEG, negative.
qPCR increases sensitivity to detect cytomegalovirus The significantly increased sensitivity of qPCR, as demonstrated here and by others, raises the issue of the significance of detectable CMV in clinical settings. The clinical significance of a single CMV IHC-positive cell is debatable as a consesus has not been reached as to what extent detection of CMV in biopsies should be considered significant. For example, is a single IHC-positive cell sufficient for the diagnosis of CMV colitisq? The same question applies to CMV qPCR. Is the positive CMV qPCR with the negative histology sufficient for the diagnosis? The answer may differ depending on the clinical circumstances. It is critical to evaluate the significance of detectable CMV in correlation with the patients’ symptoms. If the patient has diarrhea or other related GI symptoms with the confirmation of CMV in the biopsy, either on H&E, IHC, or by qPCR, CMV is definitely highly considered as the cause of the disease, unless other etiologies are evident. On the other hand, the presence of rare CMV by histology or only detectable by PCR in an asymptomatic patient probably could a latent viral status and careful clinical follow-up is warranted. In summary, diagnosis of CMV infection by qPCR in formalin-fixed, paraffin-embedded GI biopsy tissue is highly sensitive and specific. Generally speaking, diagnosis of most CMV infections can be easily rendered by histology with occasional aid of IHC. qPCR is especially helpful in equivocal biopsies by H&E and IHC, but also in cases where clinical suspicion for CMV infection is high but histology is negative, and finally in any case where the pathologist feels H&E is suspicious but IHC is negative. An algorithm for managing such cases is illustrated in Fig. 2. Compared to traditional histology, qPCR testing of GI biopsy material may facilitate an accurate diagnosis, thus leading to earlier and more effective patient management.
53
Acknowledgments We acknowledge Amy Thomasson and Tracey Bender for their assistance with preparing this manuscript, and Fredrik Skarstedt and Ryan Christy for their efforts in the preparation of figures.
References [1] Hodinka RL. Human cytomegalovirus. In: Versalovic JC, Funke Guido KC, Jorgensen JH, Landry M, Warnock DW, editors. Manual of clinical microbiology. Washington, DC: ASM Press; 2011. p. 1558-74. [2] Murph JR, Baron JC, Brown CK, Ebelhack CL, Bale Jr JF. The occupational risk of cytomegalovirus infection among day-care providers. JAMA 1991;265:603-8. [3] Lasry S, Deny P, Asselot C, et al. Interstrain variations in the cytomegalovirus (CMV) glycoprotein B gene sequence among CMVinfected children attending six day care centers. J Infect Dis 1996;174: 606-9. [4] Pass RF, Zhang C, Evans A, et al. Vaccine prevention of maternal cytomegalovirus infection. N Engl J Med 2009;360:1191-9. [5] Krech U. Complement-fixing antibodies against cytomegalovirus in different parts of the world. Bull World Health Organ 1973;49:103-6. [6] Wreghitt TG, Teare EL, Sule O, Devi R, Rice P. Cytomegalovirus infection in immunocompetent patients. Clin Infect Dis 2003;37:1603-6. [7] Chou S. Newer methods for diagnosis of cytomegalovirus infection. Rev Infect Dis 1990;12(Suppl 7):S727-36. [8] Brytting M, Xu W, Wahren B, Sundqvist VA. Cytomegalovirus DNA detection in sera from patients with active cytomegalovirus infections. J Clin Microbiol 1992;30:1937-41. [9] Shinkai M, Bozzette SA, Powderly W, Frame P, Spector SA. Utility of urine and leukocyte cultures and plasma DNA polymerase chain reaction for identification of AIDS patients at risk for developing human cytomegalovirus disease. J Infect Dis 1997;175:302-8. [10] Revello MG, Zavattoni M, Sarasini A, Percivalle E, Simoncini L, Gerna G. Human cytomegalovirus in blood of immunocompetent persons during primary infection: prognostic implications for pregnancy. J Infect Dis 1998;177:1170-5.
www.elsevier.com/locate/humpath
Original contribution
qPCR increases sensitivity to detect cytomegalovirus in formalin-fixed, paraffin-embedded tissue of gastrointestinal biopsies☆,☆☆ Morgan H. McCoy PhD, MD a , Kristin Post MPH b , Joyashree D. Sen MD b , Hsim Y. Chang BS b , Zijin Zhao MD a , Rong Fan MD, PhD a , Shaoxiong Chen MD, PhD a , Diane Leland PhD a , Liang Cheng MD a , Jingmei Lin MD, PhD a,⁎ a
Department of Pathology and Laboratory Medicine, Indiana University School of Medicine, Indianapolis, IN 46202, USA Indiana University Health, Indianapolis, IN 46202, USA
b
Received 7 June 2013; revised 29 July 2013; accepted 31 July 2013
Keywords: qPCR; Cytomegalovirus; Colon; Formalin-fixed, paraffin-embedded tissue
Summary Histopathologic diagnosis of gastrointestinal (GI) tract cytomegaloviral (CMV) infection relies on hematoxylin and eosin (H&E)–stained tissue, along with the aid of immunohistochemistry (IHC). However, non-classic appearing inclusions or atypical IHC staining patterns remain an ongoing concern for pathologists. We reported the use of real-time polymerase chain reaction (qPCR) on nucleic acid extracted from paraffin-embedded, formalin-fixed tissue of GI biopsies from cases of CMV infection (n = 91) diagnosed by H&E and IHC. Seventy-nine biopsies, including normal colon biopsies (n = 35), active colitis (n = 25), and active duodenitis (n = 19), were used as negative controls. Of 91 CMV-positive biopsies diagnosed by histology, 88 tested positive by qPCR, with a sensitivity of 96.7%. Of 79 negative controls, 78 were negative and 1 positive by qPCR, resulting in a specificity of 98.7%. Of the cases that were positive for CMV by histopathology, there were an additional 40 biopsies taken from these patients either during the same or previous procedures, some taken just days prior, which were negative for CMV by histology. Interestingly, 22 (55%) of these biopsies tested positive by qPCR, which correlated well with additional clinical CMV results. By defining qPCR as the “gold standard” for a CMV result, histology (H&E and/or IHC) had a sensitivity and specificity of 79% and 97%, respectively. Eighteen biopsies were found negative by H&E and equivocal by IHC. Among them, 14 (78%) tested positive for CMV by qPCR, which also correlated well with additional clinical results. qPCR is a sensitive, specific, and rapid molecular tool that may be helpful to aid in early diagnosis of CMV infection on equivocal or clinically highly suspicious small GI biopsies. © 2014 Elsevier Inc. All rights reserved.
1. Introduction ☆
Funding: This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. ☆☆ Conflict of Interest: The authors have no conflicts of interest to disclose. ⁎ Corresponding author. Department of Pathology and Laboratory Medicine, Indiana University School of Medicine, Indianapolis, IN 46202. E-mail address: [email protected] (J. Lin). 0046-8177/$ – see front matter © 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.humpath.2013.07.040
Cytomegalovirus (CMV), a member of the Betaherpesvirinae subfamily of herpesviruses, is frequently identified in a variety of clinical specimens from many body sites. Similar to other herpesviruses, CMV maintains the ability to establish persistent or latent infection in its host [1]. These dormant stages
qPCR increases sensitivity to detect cytomegalovirus
49
are interrupted by periods of active viral replication known as reactivation, characterized by virion release and which may, or may not, result in disease manifestations [2–4]. Reactivation may occur during periods of stress or other causes of immunosuppression, resulting in detectable virus by either viral culture or molecular methods from whole blood or plasma. It is crucial to identify active CMV infections in immunosuppressed patients, such as those with HIV/AIDS, on chemotherapy, or receiving organ/stem cell transplants, given their greater risk for developing severe infection. The diagnosis of CMV colitis can be made by multiple methods, including viral culture from biopsy material and histopathologic examination including immunohistochemical (IHC) staining. Identification of CMV infection by hematoxylin and eosin (H&E)–stained tissue sections relies on classic CMV viral inclusions (Fig. 1A). Unfortunately viral inclusions are not always apparent, even when CMV infection is clinically anticipated. Thus, IHC is frequently employed to aid H&E (Fig. 1B). However, IHC can also be hampered by rare non–classic appearing cellular staining patterns, leading to an equivocal interpretation (Fig. 1C). Such equivocal interpretations puzzle pathologists and clinicians. Therefore, we sought to employ molecular methods to detect CMV in gastrointestinal (GI) biopsies as an adjunct to atypical H&E and IHC. In addition, we compared the molecular results to traditional histologic evaluation by correlating with other available clinical test results, ie, CMV viral culture and molecular viremia assays.
2. Methods With the approval of the institutional review board, a search of our electronic laboratory database was performed to identify formalin-fixed, paraffin-embedded blocks representing cases of CMV infection in GI biopsies (n = 91) diagnosed by histopathology (H&E and/or IHC). Seventy-nine cases, including normal colon (n = 35), active colitis (n = 25), and active duodenitis (n = 19), negative for viral inclusions on H&E and negative for CMV by IHC, were used as negative controls. Two pathologists reviewed all cases (M.M. and J.L.). Of the cases that were positive for CMV by histopathology, there were an additional 40 biopsies taken from these patients either during the same (n = 25) or previous procedures (n = 15), some taken just days prior, which were interpreted as negative for CMV by H&E and IHC. Additionally, eleven equivocal cases were identified. Equivocal biopsies were defined as biopsies with no viral inclusions on H&E and with rare isolated cells demonstrating weak or incomplete cellular CMV IHC staining (Fig. 1C). In addition, seven equivocal biopsies, unrelated to the 91 CMV colitis cases, were also identified in our database. Therefore, in total 18 equivocal cases for CMV were included in the study. Totally, 228 blocks were identified and subjected to realtime polymerase chain reaction (qPCR) for CMV. Briefly,
Fig. 1 CMV colitis. A, Hematoxylin and eosin stained section of a colon biopsy demonstrating extensive involvement with typical CMV inclusions (original magnification ×400; indicated by arrows). B, A colon biopsy stained with a monoclonal antibody against CMV demonstrating a typical IHC staining pattern (original magnification ×400). C, A colon biopsy demonstrating equivocal IHC staining, characterized by a single, small weakly stained cell (original magnification ×400; indicated by arrow).
from each block, 5-10-micron-thick scrolls of tissue were cut, and DNA extractions were performed according to the manufacturer’s specifications (QIAamp DNA FFPE Tissue
50 Kit, Qiagen, Valencia, CA). The extracted DNA was then spiked with an internal reaction control and analyzed by realtime PCR using artus CMV LightCycler PCR reagents and primers (Qiagen) on the LightCycler 2.0 (Roche, Indianapolis, IN). This assay is intended to quantify CMV DNA by amplifying a 105 base-pair fragment of CMV immediateearly DNA polymerase gene. Detection occurs by the specific probe hybridization/dual probe FRET. The result was reported as positive or negative for CMV. For IHC staining, a commercially available monoclonal antibody specific for CMV was used per the manufacturer’s specifications (Dako, Carpinteria, CA; 1:1 dilution). To correlate with clinical findings, a search through the electronic medical record was performed on each case, including CMV identification by whole blood/plasma molecular methods or by viral culture from biopsy material or whole blood performed on clinical specimens submitted within 7 days of the biopsy collection.
3. Results 3.1. Sensitivity and specificity Sensitivity and specificity were calculated in two ways: (1) by using traditional histology, that is, typical viral inclusions on H&E and/or positive for typical staining by CMV IHC, as the “true positives” with “true negatives” lacking both, and (2) by using the CMV qPCR result as determining the “true positives and negatives.” 3.1.1. Traditional histology as the “gold standard” Of 91 biopsies positive for CMV by histopathology, 88 were positive by qPCR (Table 1). With a true positive definition based on traditional histology, that is, typical CMV viral inclusions and/or typical CMV IHC, these data resulted in a sensitivity of 96.7%. Of these 91 cases, 76 had CMV testing on additional specimens submitted at or within 7 days of the time the biopsy was taken. Among them, 62 (82%) had associated clinical specimens that also tested positive for CMV. Of 79 negative control biopsies, that is, negative for viral inclusions and negative by CMV IHC, 78 tested negative and 1 tested positive by qPCR, resulting in a specificity of 98.7%. This biopsy that tested positive by qPCR showed active duodenitis and was collected from a patient with a long standing history of HIV with poor highly active antiretroviral therapy compliance, end-stage liver disease secondary to hepatitis C virus cirrhosis, and type II diabetes mellitus. Upon presentation to the hospital, this patient was found to have sepsis with a duodenal ampullary mass, which was biopsied. In the year prior to admission, the patient’s documented CD4 T-cell counts were in the mid 100s/mm3 and 129 cells/mm3 one month following discharge from this hospitalization. Importantly, none of the true normal biopsies (n = 35) tested positive by qPCR.
M. H. McCoy et al. Table 1 qPCR results of 91 positive CMV cases in correlation with other clinical tests qPCR
Other CMV specimen b
n
+ + + − − Total
+ − N/A c − + 88 a qPCR+
61 12 15 2 1 91
a
The sensitivity of CMV by qPCR is 96.7%. Whole blood/plasma molecular methods and viral culture from gastrointestinal biopsy material or whole blood. Each was performed on clinical specimens submitted within 7 days of the biopsy collection date. c N/A refers to those biopsies with no additional specimens submitted for CMV testing within the defined 7-day time frame. b
3.1.2. qPCR as the “gold standard” Of 111 biopsies positive for CMV by qPCR, 88 tested positive by traditional histology (H&E and CMV IHC), resulting in a sensitivity of 79% for traditional histology at detecting CMV in GI biopsies. Of the 99 biopsies negative for CMV by qPCR, 96 tested negative by traditional histology resulting in a specificity of 97%.
3.2. qPCR results on additional histologically negative specimens Of 91 cases that were positive for CMV by histopathology, there were an additional 40 biopsies taken from these patients either during the same (n = 25) or previous procedures (n = 15), some taken just days prior, which were interpreted as negative for CMV by H&E and IHC. Interestingly, 22 (55%) of these biopsies tested positive for CMV by qPCR. As shown in Table 2, within the group of 40 biopsies, 15 were from eight patients submitted on a previous day that were negative for CMV by H&E and IHC. Of these 15 biopsies, 9 (60%) tested positive for CMV by qPCR. Among these 15 biopsies, 13 had more information regarding CMV (cases 26, 8) on additional clinical specimens (viral culture or molecular methods) submitted at or within 7 days of the time the biopsy was taken. Seven of these 8 patients (9 out of 15 biopsies) had biopsy material that tested positive by qPCR. Among these 9 qPCR positive biopsies, 7 were from 5 patients who had additional clinical specimens tested. Four patients had additional negative tests for CMV (Table 2; cases 3, 5, 6 and 8). One patient had two biopsies that tested positive for CMV by qPCR with additional clinical specimens testing positive (case 4). One patient had three biopsies that tested negative for CMV by qPCR with additional clinical specimens testing positive (case 2). Thus, qPCR results, compared to conventional histology, of the prior biopsies correlated better with the subsequent CMV diagnosis. Therefore, qPCR results likely represent the true CMV status in these histologically negative specimens.
qPCR increases sensitivity to detect cytomegalovirus
51
Table 2 The qPCR result of 15 prior negative biopsies from eight patients who had positive CMV biopsies in correlation with other clinical tests Case number
Other CMV specimen a
Fraction of CMV qPCR positive biopsies in previous submission
Blood/plasma molecular 1 2 3 4 5 6 7 8 Percent positive
1/1 0/3 1/1 2/2 1/2 2/4 1/1 1/1 9/15 (60%)
N/A + − + − N/A N/A −
b
Biopsy viral culture N/A − − N/A − − N/A N/A
a
Whole blood/plasma molecular methods and viral culture from gastrointestinal biopsy material or whole blood. Each was performed on clinical specimens submitted within 7 days of the biopsy collection date. b N/A refers to those biopsies with no additional specimens submitted for CMV testing within the defined 7-day time frame.
3.3. qPCR results on histologically equivocal specimens As illustrated in Table 3, 14 of the 18 IHC equivocal biopsies tested positive for CMV by qPCR (78%). Among these 18 equivocal cases, eleven had additional biopsies taken the same day that were positive for CMV by histopathology. Among these 14 qPCR positive samples, 10 had biopsies taken the same day that were positive for CMV on histopathology. Of the 18 equivocal biopsies tested, 11 (cases 1-5, 10, 12, 13, 15-17) had additional clinical information regarding CMV testing on other specimens submitted at or within 7 days of the time the biopsy was taken. Among these 11 cases, 8 (73%) tested positive for CMV by qPCR. Of these 11 biopsies, 5 (45%) tested positive by qPCR with additional positive CMV specimens (cases 1-5); 3 (27%) tested positive by qPCR with no positive CMV specimens (cases 10, 12, 13); 3 (27%) tested negative by qPCR with additional negative CMV specimens (cases 15-17). As more specifically demonstrated in Table 3, of the 11 equivocal biopsies with additional clinical specimens submitted (cases 1-5, 10,12, 13, 15-17), 10 were GI biopsies collected on the same day and submitted for CMV early antigen detection by viral culture (cases 1-5, 12, 13, 15-17). Five of those 10 samples had positive CMV early antigen viral cultures, all of which had corresponding biopsies that tested positive by qPCR (cases 1-5). Of the remaining 5 negative CMV early antigen viral cultures: 2 had negative CMV molecular results from whole blood/ plasma and tested positive for CMV by qPCR; the other 3 had negative CMV molecular results from whole blood/ plasma and also negative by qPCR. Therefore, qPCR results correlate better with concurrent clinical data compared to conventional histology. Thus, qPCR results likely represent the true CMV status in these IHC equivocal biopsies.
4. Discussion Detection of CMV has changed greatly as medicine continuously demands faster and more accurate diagnostic tools and technologies. Serology, once used frequently, is of limited use since the virus is so prevalent in the population (at least 40% seropositivity) [5]. However, serology is still helpful in immunocompromised hosts in determining whether patients
Table 3 Equivocal biopsies by histology in correlation with qPCR results and other clinical specimens tested for CMV Case number
qPCR
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
+ + + + + + + + + + + + + + − − − −
Other CMV specimen a Blood/plasma molecular
Biopsy culture
+ N/A b N/A + N/A N/A N/A N/A N/A − N/A − − N/A − − − N/A
+ + + + + N/A N/A N/A N/A N/A N/A − − N/A − − − N/A
a Whole blood/plasma molecular methods and viral culture from gastrointestinal biopsy material or whole blood. Each was performed on clinical specimens submitted within 7 days of the biopsy collection date. b N/A refers to those biopsies with no additional specimens submitted for CMV testing within the defined 7-day time frame.
52 are at risk for acquiring a new CMV infection (seronegative) or reactivation (seropositive). Of note, seropositivity does not preclude the possibility of a new infection with an antigenically dissimilar CMV virus strain. Serology specific for IgM antibody is useful when there is a concern for acute infection, although IgM levels may persist for months following exposure [6]. Culture for CMV was once the gold standard for detection of antigen from many body sites including respiratory specimens, whole blood, urine, or biopsy material. However, traditional culture is problematic due to the length of time (1-6 weeks) required to observe typical cytopathic effect on fibroblast monolayer. The advent of early antigen detection by shell vial culture has shortened the time to 2 to 3 days, and its use has become standard [7]. Viremia assays that test for either antigen or nucleic acid in leukocytes, whole blood, or plasma specimens have become clinically useful. Quantitative nucleic acid assays are essential in laboratory diagnostics, particularly in immunocompromised hosts. These techniques provide not only the most sensitive methods to detect CMV in blood specimens, but also provide a quantitative measurement for disease progression and treatment effect [8–10]. Clinically, we have encountered examples of patients who lacked viremia by qPCR, but the following tissue biopsies showed convincing CMV infection. Generally speaking, blood is a good source to detect infection if patients have systemic involvement. However, in localized infections, CMV virus concentration might be higher in infected tissue than that found in blood. Although quantitative nucleic acid assays are extremely sensitive for infections involving viremia, their use in localized infections, such as CMV colitis, has not been adequately evaluated. In this study, we employed a qPCR method to identify CMV in GI biopsies. qPCR was utilized as opposed to conventional PCR methods, due to the availability of commercial reagents including primers and standards, the ability to automate many steps in the procedure, as well as the intrinsic increased sensitivity of qPCR compared to conventional PCR methods. We demonstrate that qPCR is a rapid, sensitive, and specific method for detecting CMV in formalin-fixed, paraffin-embedded GI biopsy tissues. This is most evident by calculating the sensitivity for traditional histology (79%) in detecting CMV in GI biopsies with a “true positive” defined as positive by qPCR. Specificity calculations were similar whether a true negative was defined by traditional histology or qPCR. qPCR demonstrated a specificity of 98.7% using traditional histology as a “true negative” definition. Traditional histology demonstrated a 97% specificity using qPCR as a “true negative” definition. Causes for the failure of qPCR to detect CMV in three of the histopathologically positive CMV biopsies are uncertain. The effect of PCR reaction inhibitors in the specimens was largely ruled out due to the use of an internal control added during the nucleic acid extraction step. Too little tissue for nucleic acid extraction or laboratory error during processing are potential explanations for these false negative results.
M. H. McCoy et al. Since the “falsely positive” biopsy was from a patient who had well documented immunosuppression, the detection of CMV could represent reactivated CMV replication, as well as possible detection of latent virus. One important contribution of the current study is that we evaluated the prior negative biopsies from patients with subsequent positive CMV infection in a short time interval. Interestingly, over half of the histologically negative biopsies from 7 of the 8 patients tested positive for CMV by qPCR. Given that subsequent biopsies from these patients were histologically positive, the qPCR results from biopsy tissue might represent actual CMV infections that were not detected by histology, viral culture from GI biopsy tissue, or qPCR from blood/plasma. Another important contribution is that we evaluated the IHC equivocal biopsies, since they pose a diagnostic challenge to pathologists. In this study, the CMV results by qPCR better correlate with additional clinical data. Compared to conventional methods, qPCR is more sensitive in diagnosing CMV infection in GI biopsy material. In such cases, a positive qPCR result would have resulted in an earlier diagnosis.
Fig. 2 Proposed algorithm utilizing qPCR on formalin-fixed, paraffin-embedded GI biopsy tissue for working up of CMV infection. Abbreviations: POS, positive; NEG, negative.
qPCR increases sensitivity to detect cytomegalovirus The significantly increased sensitivity of qPCR, as demonstrated here and by others, raises the issue of the significance of detectable CMV in clinical settings. The clinical significance of a single CMV IHC-positive cell is debatable as a consesus has not been reached as to what extent detection of CMV in biopsies should be considered significant. For example, is a single IHC-positive cell sufficient for the diagnosis of CMV colitisq? The same question applies to CMV qPCR. Is the positive CMV qPCR with the negative histology sufficient for the diagnosis? The answer may differ depending on the clinical circumstances. It is critical to evaluate the significance of detectable CMV in correlation with the patients’ symptoms. If the patient has diarrhea or other related GI symptoms with the confirmation of CMV in the biopsy, either on H&E, IHC, or by qPCR, CMV is definitely highly considered as the cause of the disease, unless other etiologies are evident. On the other hand, the presence of rare CMV by histology or only detectable by PCR in an asymptomatic patient probably could a latent viral status and careful clinical follow-up is warranted. In summary, diagnosis of CMV infection by qPCR in formalin-fixed, paraffin-embedded GI biopsy tissue is highly sensitive and specific. Generally speaking, diagnosis of most CMV infections can be easily rendered by histology with occasional aid of IHC. qPCR is especially helpful in equivocal biopsies by H&E and IHC, but also in cases where clinical suspicion for CMV infection is high but histology is negative, and finally in any case where the pathologist feels H&E is suspicious but IHC is negative. An algorithm for managing such cases is illustrated in Fig. 2. Compared to traditional histology, qPCR testing of GI biopsy material may facilitate an accurate diagnosis, thus leading to earlier and more effective patient management.
53
Acknowledgments We acknowledge Amy Thomasson and Tracey Bender for their assistance with preparing this manuscript, and Fredrik Skarstedt and Ryan Christy for their efforts in the preparation of figures.
References [1] Hodinka RL. Human cytomegalovirus. In: Versalovic JC, Funke Guido KC, Jorgensen JH, Landry M, Warnock DW, editors. Manual of clinical microbiology. Washington, DC: ASM Press; 2011. p. 1558-74. [2] Murph JR, Baron JC, Brown CK, Ebelhack CL, Bale Jr JF. The occupational risk of cytomegalovirus infection among day-care providers. JAMA 1991;265:603-8. [3] Lasry S, Deny P, Asselot C, et al. Interstrain variations in the cytomegalovirus (CMV) glycoprotein B gene sequence among CMVinfected children attending six day care centers. J Infect Dis 1996;174: 606-9. [4] Pass RF, Zhang C, Evans A, et al. Vaccine prevention of maternal cytomegalovirus infection. N Engl J Med 2009;360:1191-9. [5] Krech U. Complement-fixing antibodies against cytomegalovirus in different parts of the world. Bull World Health Organ 1973;49:103-6. [6] Wreghitt TG, Teare EL, Sule O, Devi R, Rice P. Cytomegalovirus infection in immunocompetent patients. Clin Infect Dis 2003;37:1603-6. [7] Chou S. Newer methods for diagnosis of cytomegalovirus infection. Rev Infect Dis 1990;12(Suppl 7):S727-36. [8] Brytting M, Xu W, Wahren B, Sundqvist VA. Cytomegalovirus DNA detection in sera from patients with active cytomegalovirus infections. J Clin Microbiol 1992;30:1937-41. [9] Shinkai M, Bozzette SA, Powderly W, Frame P, Spector SA. Utility of urine and leukocyte cultures and plasma DNA polymerase chain reaction for identification of AIDS patients at risk for developing human cytomegalovirus disease. J Infect Dis 1997;175:302-8. [10] Revello MG, Zavattoni M, Sarasini A, Percivalle E, Simoncini L, Gerna G. Human cytomegalovirus in blood of immunocompetent persons during primary infection: prognostic implications for pregnancy. J Infect Dis 1998;177:1170-5.