Comparison of in situ hybridization and immunohistochemistry for detection of cytomegalovirus and herpes simplex virus

Comparison of In Situ Hybridization and Immunohistochemistry for Detection of Cytomegalovirus and Herpes Simplex Virus JOHN G. STRICKLER, MD, J. CARLOS MANIVEL, MD, CEDITH MaCOPENHAVER, MS, AND VIRGINIA L. KUBIC. MD In situ hybridization (ISH) and immunohistochemistry (IHC) were compared for detection of cytomegalovirus (CMV) and herpes simplex virus (HSV) in routinely processed tissue. Fiftyfour formalin-fixed paraffin-embedded tissue samples infected with CMV (36 tissues) or HSV (18 tissues) from 30 autopsies were studied. All tissues had either positive viral cultures (38 of 54) or characteristic viral inclusions on hematoxylin and eosin examination (39 of 54). The tissues examined included lung (28), liver (nine), kidney (five), heart (three), adrenal (two), spleen (two), and thymus, pancreas, appendix, esophagus, and duodenum (one each). Studies by ISH were performed with two detection systems, using biotinylated probes to CMV and HSV (Enzo B&hem, New York, NY). Using ISH with an alkaline phosphatase detection system, infected cells were detected in 33 of 54 tissues (CMV: 23 of 36, HSV: 10 of 18). Using ISH with a peroxidase detection system, infected cells were identified in 30 of 54 tissues (CMV: 22 of 36, HSV: eight of 18). With IHC, antibodies to CMV and HSV stained the infected cells in 34 of 54 tissues (CMV: 24 of 36, HSV: 10 of 18). All infections detected with ISH were also detected with IHC. We conclude that these techniques for ISH and IHC are equally effective for detecting CMV and HSV in paraffin sections. The results of both techniques correlate better with viral inclusions than with culture results. The ISH stains are more difficult to prepare and in some cases are more difficult to interpret. Therefore, IHC may be preferable to ISH for detecting CMV and HSV in routine diagnostic work. HUM PATHOL 21:443448. 0 1990 by W.B. Saunders Company.

In situ hybridization probes to cytomegalovirus

(ISH) using biotinytated (CMV) and herpes simplex

Grus (HSV) is a useful technique for rapidly detecting viral nucleic acids in formalin-fixed. paraffinembedded tissue. I-’ ’ Immunohistochemistry (IHC) using antibodies to viral antigens allows detection of viral proceins in tissue sections. Several studies, based on s~nall nurnbers of patients with CMV infections, reveal cclnflicting data on the relative sensitivity of these techniques.‘;‘-‘H In the present study, we compared ISH and IHC in 54 tissue samples from 30 itutopsy patients with well-documented CMV or HSV

infections.

From l.he Department of Laboratory Medicine and Pathology. University of Minnecota Hospital, Minneapolis, MN. ;\c-cepted for Imbliratio1i August 31. 10X9. Supported in pnrt bv a grant from the Minnesota Medical Foundation. Kry rowcfr: in situ hybridization, immunohistochemia~r~, herpes \impleu virus. cytomegalovirus. Address correspondence and reprint requests to John G. Qrickler, MD. Box 1%. Department of Pathology, Llniversity of Minnesota llospital, 420 Delaware St. SE, Minneapolis, MN 5545.5. 6 I990 by W.B. Saunders Company. 0046-1~ I77/C~O1”194-00 I I $5.00/O 443

MATERIALS AND METHODS Case Selection Fifty-four formalin-fixed paraffin-embedded tissues infected with CMV (36 tissues) or HSV (18 tissues) from 30 autopsies were studied. All tissues had either positive viral cultures (38 of’ 54) or characteristic viral inclusions on hematoxylin and eosin examination (39 of 54). The tissues examined included lung (28), liver (nine)., kidney (five), heart (three), adrenal (two), spleen (two), and thymus, pancreas, appendix, esophagus, and duodenum (one each).

In Situ Hybridization with Alkaline Phosphatase Detection System Our protocol was similar to the protocol of’ I:nger et at” with minor modifications. Paraffin sections were mounted on cleaned 3-aminopropy,ltriethoxysitane-coated glass slides”’ and were baked overnight at WC. After deparaffinization, the tissues were treated with a fresh preparation ot pronase (0.3 mg/mL in 50 mmol/L tris-c I pH 7.4 with 5 mmol/L EDTA) in a 37°C water bath for 5 minutes (except esophagus-15 minutes). The reaction was stopped with Tris saline glycine (0.1 mot/L Tris-(11 pH 7.5. 0.1 moI/L NaCl, 2.0 mg/mL glycine) and the sections were dehydrated. Biorinylated (nick translation) CMV and HSV probes were obtained from Enzo Biochem, New York, NY. The CMV probe is a mixture of two clones of CMV sequences in the Barn Hl site of pBR322. The insert sizes are 17.2 Kb and 25.2 Kb. The HSV probe is a mixture of HSV 1 and HSV 2 probes. The HSV 1 probe is prepared from a fragment containing the thymidine kinase gene of HSV 1 cloned into the Barn H I site of pBR322. The insert size is 3.5 Kb including the thymidine kinase coding region and approximately ‘) ,.5 Kb of flanking sequences. The HSV 2 probe is prepared from a fragment containing the thymidine kinase gene of HSV 2 cloned into the BglII-Hind III fragment of pBR322. The insert size is 4 Kb. containing the rhymidine kinase coding region and approximately 2.2 Kb of flanking sequences. The biotinylated CMV and HSV probes were used at a concentration of 1.0 FgimL in the hybridization cocktail (45%; formamide, .iX SSC. 25 mmolli sodium phosphate pH 6.5. 1 X Denhardt’s, 250 ~g/mL sheared denatured salmon sperm DNA, and 10% dextran sulfate). The probe solution (30 FL) was placed on each tissue section, and a coverslip cut from autoctavable polypropylene bags (Ctavies) was used. The slides were placed in a 100°C convection oven for 10 minutes and then transferred to a moisture chamber in a 37°C incubation oven for 2 hours. Posthybridization washes consisted of removing the coverslips in 2~ SSC, followed by a series of 3 minute washes of increasing stringency (two washes in 2 X SSC 0.1% SDS, two washes in 0.2~ SSC 0.17,# SDS, and two washes in 0.16~ SSC 0.1% SDS), followed by a l-minute wash in 2x SSC O.l%, SDS. The sections were blocked for 5 minutes with 3% BSA

HUMANPATHOLOGY

Volume 21, No. 4 (April 1990)

in Tris saline and air dried. The slides were then incubated for 30 minutes with an avidin alkaline phosphatase complex, which had been mixed (30 minutes previously) using 40 IJ.L of avidin-DN (Vector, Burlingame, CA), 5 ~.LLof a biotin alkaline phosphatase complex (synthesized by the method of Leary et al),20 and 11 mL of 1% BSA in trissaline-triton (TST). The slides were rinsed with three washes of TST and one wash of Tris saline pH 9.5. Freshly prepared McGadey’s reagent (0.2 mL NBT 50 mg/mL in 50% DMF, and 0.1 mL of 50 mg/mL BCIP in DMF, in 30 mL Tris saline 9.5) was applied as the chromogen for 20 to 30 minutes in a 37°C moisture chamber. After a final rinse with ddH,O, the slides were counterstained with Gill’s hematoxylin and mounted with permount. (In contrast to other alkaline phosphatase detection systems, an aqueous mounting medium is not required.) Appropriate positive and negative controls were performed. Negative controls were obtained by using CMV and HSV probes on each tissue; therefore, the negative control for the HSV infected tissues was the CMV probe, and the negative control for the CMV infected tissues was the HSV probe. The CMV and HSV probes did not crossreact. Positive controls always included an esophagus with culture-proven HSV infection or a lung with documented CMV infection, both with characteristic intranuclear inclusions.

In Situ Hybridization with Peroxidase Detection System Our protocol was similar to the protocol of Przepiorka and MyersonL2 and is identical to the above protocol with the following revisions. To block endogenous peroxidase, a 30-minute incubation in 3% H,O, in methanol was added after the deparaffinization step. A post-fixation step of 4% paraformaldehyde in phosphate-buffered saline (PBS, 5 minutes RT) followed the pronase. Following post-fixation, the slides were washed in 2.0 mg/mL glycine in PBS. A prehybridization incubation (10 minutes, RT) consisting of 2~ SSC with 45% formamide preceded the probe application. Following the washes of increasing stringency, the slides were developed with a peroxidase detection system. Avidin-DN was applied after a brief block with NGS in BSA/PBS (10% NGS, 2% BSA, in PBS). The slides were then incubated with biotinylated anti-avidin D (Vector) for 10 minutes at 37”C, followed by an ABC complex (Vector) for 30 minutes at 37°C. Finally, the stains were visualized with a nickel DAB substrate” for 5 to 15 minutes and enhanced with a silver reagent I2 for 5 to 15 minutes. Sections were counterstained with Gill’s hematoxylin and coverslipped using an aqueous mounting medium (Lerner). Alternatively, the hematoxylin-stained sections could be dehydrated and mounted with Permount.

tibodies as previously described. ” Visualization of the reactants was accomplished by immersion of the sections in 3,3’-diaminobenzidine solution (0.25 mg/mL) with 0.003% hydrogen peroxide: they were then counterstained with Harris’ hematoxylin, dehydrated, and coverslipped with Permount. Positive controls consisted of sections of stock tissues with documented CMV or HSV infection and characteristic viral inclusions. Negative controls in each case consisted of sections stained with mouse ascites fluid or nonimmune rabbit serum in lieu of primary antibodies; and by the use of a nonrelevant antibody (anti-CMV for HSV-infected cases, and anti-HSV for CMV-infected cases).

RESULTS The results of the ISH studies and the IHC studies are summarized in Table 1 and illustrated in Figs 1 and 2. In situ hybridization with an alkaline phosphatase detection system (ISH-AP) demonstrated infected cells in 33 of 54 tissues, including 23 of the 36 CMV tissues and 10 of the 18 HSV tissues.The infected cells were extremely variable in number and showed both nuclear and cytoplasmic staining with both probes, although nuclear staining predominated. Compared to ISH-AP, ISH with a peroxidase detection system (ISH-Px) produced more intense staining but was often associated with more background. This background staining sometimes made interpretation difficult and may be one reason why ISH-Px detected infected cells in fewer cases than ISH-AP (Table 1). Background staining was also related to tissue type and probe type. Liver, adrenal, and heart tissue had more background than lung or other tissues, and for any given tissue type, the HSV TABLE1. Results By Virus and Tissue Type ISH-AP-

CMV Lung Liver Kidney Heart Spleen Duodenum Appendix Pancreas Thymus Adrenal Total

Immunohistochemistry For immunohistochemical studies, 5 km sections were mounted on chrome-alum
HSV Lung Liver Heart

Esophagus Adrenal Total Total

ISH-Px-

IHC-

Total

positive

positive

positive

17

12

11

12

5 5 2 2

3

3

3

1

1

I

1

1

2

1

1

1

1

1

1

1

1 1

1 1

1

1

1

1 1

1

1

1 22

1 1 1 24

I

36

23

11 4 1

1

i

1

5 3 0 1

0 1

0 1

1 18

1 10

8

10

54

33

30

34

.5

I

1

Abbreviations: ISH-AP, in situ hybridization with an alkaline phosphatase detection system; ISH-Px, in situ hybridization with a peroxidase detection system.

444

ISH AND IHC FOR CMV AND HSV (Strickler et al)

FIGURE 1. Cytomegalovirus pneumonltis. An infected cell is detected by in situ hybridization with an alkaline phosphatase detection system [top left]. in situ hybridization with a peroxidase detection system [top right], and IHC [bottom left). A negative control [bottom right) is shown for comparison. (Hematoxylin counterstain, magnification x 1,000.)

probe tended to produce more background than the CMV probe. Studies with IHC using antibodies to CMV and HSV, detected infected cells in 34 of the 54 tissues,

including 24 of the 36 CMV tissues and 10 of the 18 HSV tissues. Both antibodies produced strong staining with nuclear and cytoplasmic distribution, although cytoplasmic staining often predominated. 445

HUMAN PATHOLOGY

Volume 21. No. 4 [April 1990)

FIGURE 2. Herpes simplex virus esophagitis. The infected cells are detected by in situ hybridization with an alkaline phosphatase detection system (top left], in situ hybridization with a peroxidase detection system [top right), and IHC (bottom left]. A negative control (bottom right) is shown for comparison. (Hematoxylin counterstain, magnification x 200.)

The number of infected cells per slide was variable and correlated with the number of infected cells detected with ISH. Background staining was usually insignificant and was often less than the background in

the corresponding ISH stains. The tissues which showed significant background staining with IHC also showed extensive background staining with ISH, indicating suboptimal tissue fixation or processing.

ISH AND IHC FOR CMV AND HSV (Strickleret al] TABLE 3.

Both ISH and IHC are equally effective for detecting CMV and HSV in paraffin sections (Tables 1 and 2). One heart specimen was the only tissue in our study where the ISH and IHC studies were not in agreement; I HC demonstrated rare cells infected with CMV which could not be identified by ISH using either detection system. Table 3 compares the culture and inclusion data to the ISH and IHC results. When inclusions were identified (39 cases), ISH and IHC usually, but not always, detected the infected cells. Both ISH and IHC were positive in all four culture-negative cases with inclusions’. When inclusions were not identified (15 cases, all confirmed with viral culture), ISH and IHC demonstrated the infected cells in only one case.

No.

One ‘objective of this study is to compare the sensitivity of two ISH protocols for detecting CMV and HSV in paraffin sections. One protocol (ISH-AP) uses a detection system based on alkaline phosphatase,’ 4 and the other protocol (ISH-Px) uses a detection system based on peroxidase.i2 The results show that ISH-AP is slightly more sensitive than ISHPx. The protocols use slightly different hybridization conditions, even though both protocols employ identical tissues, probes, and controls. Therefore, a direct comparison of alkaline phosphatase and horseradish peroxida,se as receptor enzymes is not possible. A second objective of this study is to compare these ISH protocols to a standard IHC protocol for detecting CMV and HSV in paraffin sections. The results show that these ISH and IHC techniques have similar sensitivities. However, IHC is easier to optimize and perform, since fewer steps are required. In addition, IHC stains are easier to interpret, since ISH may produce significant background staining, especially with the ISH-Px method. The increased background staining of the ISH-Px method is most likely due to the nickel followed by silver intensification, procedures which were not used in the peroxidase detection system of the IHC protocol. The sensitivity of ISH depends on a large number of variables, including characteristics of the probes. The CMV and HSV probes, each of which is a cocktail of two probes, were chosen for this study because they are commercially available in biotinylated form (Enzo Biochem). Although we have not conducted an extensive search, we are not aware of other commercial sources for biotinylated CMV or HSV probes. However, use of other probes or probe cocktails could theoretically increase the sensitivity of

Comparison of In Situ Hybridization With lmmunohistochemistry ISH-positive

IHC-positive IHC-negative

(34) (20)

33 0

(33)

ISH-negative

Inclusions Culture Culture Culture Total

present positive negative not done

Inclusions Culture (:ulture Culture Total

absent positive negative not done

ISH-positive

IHGpositive

23 4 12 39

1x 4 10 32

19 4 10 33

15 0 0 15

I

I

0 0 1

0 0

I

the ISH protocols. Radiolabeled probes also usually increase the sensitivity of ISH but require detection with autoradiography, which adds a minimum of 3 to 5 days to the procedure. We have successfully detected HSV with 95S-labeled probes, and others22-24 have used radioactive probes to detect CM\‘, HSV, and other viruses. The sensitivity of IHC depends on variables which include the antibodies and the detection system. The CMV and HSV antibodies in this study are available commercially from DakoPatts. The CMV antibody reacts with both early (nuclear staining) and late (nuclear and cytoplasmic staining) antigens. The sensitivity of this antibody is similar to that of a polyclonal human antiserum, used previously in our laboratory (data not shown). The HSV antiserum was prepared using immunogens obtained by sonication and extraction of HSV type I-infected rabbit cornea cells. The antiserum reacts with type-specific, as well as with type-common antigens. In previous studies in our laboratory (unpublished data), these DakoPatts antibodies have been shown to reliably detect CMV and HSV in paraffin sections. We have not compared the sensitivity of these antibodies to other commercially available antibodies or antibody cocktails. Previous studies using both ISH and IHC have yielded conflicting data about their relative sensitivities, due at least in part to the variable methodology used and the small number of cases studied. Keh and Gerber’” conclude that ISH is more sensitive than IHC for detecting CMV infections in acquired immunodeficiency syndrome patients. However, Linden et al” prefer IHC over ISH for diagnosis of CMV pneumonitis because IHC has increased staining intensity and greater specificity (using culture as the “gold standard”). Robey et a1*8 found that IHC was more useful than ISH for diagnosis of CMV colitis, mainly because ISH staining was not observed in cases with only a few inclusions. Borisch et all5 concluded that performing both ISH and IHC is the best approach for detecting CMV infection in various organs. Wolber and Lloyd25 showed that sequential ISH and IHC can be performed on the same histologic section, but for diagnostic purposes, this method appears to have no advantage over performing each technique on separate slides. A third objective of this study is to compare the

DISCUSSION

TABLE 2.

Culture and Inclusions v In Situ Hybridization and lmmunohistochemistry

(21)

I 20

447

HUMAN PATHOLOGY

Volume 21, No. 4 [April 19901

8. Myerson D, Hackman RC, Meyers JD: Diagnosis of cytomegaloviral pneumonia by in situ hybridization. J Infect Dis 150:272-277. 1984 9. Myerson D, Hackman RC, Nelson JA, et al: Widespread presence of histologically occult cytomegalovirus. HUM PATHOL 15:430-439, 1984 10. Naoumov NV, Alexander GJM, Eddleston ALWF, et al: In situ hybridisation in formalin fixed, paraffin wax embedded liver specimens: Method for detecting human and viral DNA using biotinylated probes. J Clin Path01 41:793-798, 1988 11. Naoumov NV, Alexander GJM, O’Grady JG, et al: Rapid diagnosis of cytomegalovirus infection by in-situ hybridisation in liver grafts. Lancet 1:1361-1364, 1988 12. Przepiorka D, Myerson D: A single-step silver enhancement method permitting rapid diagnosis of cytomegalovirus infection in formalin-fixed. paraffin-embedded tissue sections by in situ hybridization and immunoperoxidase detection. J Histochem Cytochem 34: 1731-1734, 1986 13. Roberts WH, Hammond S, Sneddon JM, et al: In situ DNA hybridization for cytomegalovirus in colonoscopic biopsies. Arch Path01 Lab Med 112:1106-1109, 1988 14. Unger ER, Budgeon LR, Myerson D, et al: Viral diagnosis by in situ hybridization. Description of a rapid simplified colorimetric method. Am J Surg Path01 IO: l-8, 1986 15. Borisch B. Jahn G, Scholl BC, et al: Detection of human cytomegalovirus DNA and viral antigens in tissues of different manifestations of CMV infection. Virchows Arch [B] 55:93-99, 1988 16. Keh WC, Gerber MA: In situ hybridization for cytomegalovirus DNA in AIDS patients. Am J Path01 131:490-496, 1988 17. Linden MD, Tubbs RR, Bauer TW. et al: Detection of cytomegalovirus infections: Comparison of in situ hybridization with immunohistology. Am J Clin Path01 90:493, 1988 (abstr) 18. Robey SS. Gage WR, Kuhajda FP: Comparison of immunoperoxidase and DNA in situ hybridization techniques in the diagnosis of cytomegalovirus colitis. Am J Clin Path01 89:666-671, 1988 19. Rentrop M. Knapp B, Winter H, et al: Aminoalkylsilanetreated glass slides as support for in situ hybridization of keratin cDNAs to frozen section; ;nder varying fixation and pretreatment conditions. Histochem 1 18:271-276. 1986 20. Leary JJ, Brigiti DJ, Ward DC: Rapid and sensitive colorimetric method for visualizing biotin-labeled DNA probes hybridized to DNA or RNA immobilized on nitrocellulose bio-blots. Proc Nat1 Acad Sci USA 80:4045-4049, 1983 21. Hsu SM, Raine L, Fanger H: Use of avidin-biotinperoxidase complex (ABC) in immunoperoxidase techniques: A comparison between ABC and unlabeled antibody (PAP) procedures. J Histochem Cytochem 29:577-580, 1981 22. Gnann JW, Ahlmen J, Svalander C, et al: Inflammatory cells in transplanted kidneys are infected by human cytomegalovirus. Am J Path01 132:239-248, 1988 23. Kennedy PGE, Newsome DA, Hess J, et al: Cytomegalovirus but not human T lymphotropic virus type III/ lymphadenopathy associated virus detected by in situ hybridisation in retinal lesions in patients with the acquired immune deficiency syndrome. Br Med J [Clin Res] 293:162-164, 1986 24. Saldanha J, Sutton RNP, Gannicliffe A, et al: Detection of HSVI DNA by in situ hybridisation in human brain after immunosuppression. J Neurol Neurosurg Psychiatry 49:613-619, 1986 25. Wolber RA, Lloyd RV: Cytomegalovirus detection by nonisotopic in situ DNA hybridization and viral antigen immunostaining using a two-color technique. HUM PATHOL 19:736-741, 1988 26. Hilborne LH, Nieberg RK, Cheng L, et al: Direct in situ hybridization for rapid detection of cytomegalovirus in bronchoalveolar lavage. Am J Clin Path01 87:766-769, 1987 27. Masih AS, Linder J, Shaw BW, et al: Rapid identification of cytomegalovirus in liver allograft biopsies by in situ hybridization. Am J Surg Path01 12:362-367, 1988

results of ISH and IHC with viral inclusions and culture data. Our results show that ISH and IHC correlate better with viral inclusions than with culture results; this is likely due to the fact that the same tissue sample is used for ISH and IHC, while different tissue samples are required for these techniques and for culture studies. Viral inclusions are usually closely correlated with the results of ISH and IHC, although both techniques occasionally do not react with cells containing viral inclusions, and both techniques may occasionally detect infected cells without viral inclusions.7*1” Viral cultures are the “gold standard” for detecting CMV and HSV infection in most studies, but Myerson et al” suggest that ISH may be as sensitive as culture. In contrast, our results suggest that ISH and IHC are not sensitive techniques for detecting culture-positive cases without inclusions, although we and others6,26.27 report rare culturenegative cases that are ISH-and/or IHC-positive. The obvious disadvantages of culture are the time required (several weeks) and the need for fresh tissue. We conclude that ISH and IHC are equally effective for detecting CMV and HSV in routinely embedded tissue and that neither technique detects all cases which have viral inclusions and/or positive viral cultures. IHC is easier to perform and interpret, and is less expensive, than ISH. Therefore, IHC may be preferable to ISH for detecting CMV and HSV in routine diagnostic work with the techniques currently available. Acknowledgment. The authors thank for secretarial assistance, and Kim Hagen for technical assistance.

Carol Ghandour and Amy Birney

REFERENCES 1. Burns J, Redfern DRM, Esiri MM, et al: Human and viral gene detection in routine paraffin embedded tissue by in situ hyEridisation with biotinylatkd probes: Viral localisatioh in herpes encenhalitis. I Clin Path01 39:1066-1073. 1986 2. Coated PJ, Hall PA, Butler MG, et al: Rapid technique of DNA-DNA in situ hybridisation on formalin fixed tissue sections using microwave irradiation. J Clin Pathol 40:865-869, 1987 3. Forghani B, Dupuis KW, Schmidt NJ: Rapid detection of herpes simplex virus DNA in human brain tissue by in situ hybridization. J Clin Microbial 22:656-658, 1985 4. Grady WW, Lewin Kr, Naeim F: Detection of cytomegalovirus DNA h classic and epidemic Kaposi’s sarcoma’by in-situ hvbridization. HUM PATHOL 19:524-528. 1988 5. Hashimoto H, Muller H, Muller F, et al: In situ hybridization analysis of cytomegalovirus lytic infection in Kaposi’s sarcoma associated with AIDS. A study of 14 autopsy cases. Virchows Arch [A] 411:441-448, 1987 6. Langenberg A, Smith D, Brake1 CL, et al: Detection of herpes simplex virus DNA from genital lesions by in situ hybridization. J Clin Microbial 26:933-937, 1988 7. Loning T, Milde K, Foss HD: In situ hybridization for the detection of cytomegalovirus (CMV) infection. Application of biotinylated CMV-DNA probes on paraffin-embedded specimens. Virchows Arch [A] 409:777-790, 1986

448