- Email: [email protected]
Detection of rabies virus RNA in the central nervous system of experimentally infected mice using in situ hybridization with RNA probes
Journal of Virological Methods, 2.5 (1989) 1-12
Elsevier JVM 00887
Detection of rabies virus RNA in the central nervous system of experimentally infected mice using in situ hybridization with RNA probes Alan C. Jacksoni**,
Dorothy
L. Reimer2
and William
H. Wunner3
Departments of ‘Medicine and ‘Microbiology d Immunology, Queen’s University, Kingston, Ontario, Canada and 3The Wistar Institute of Anatomy and Biology, Philadelphia, Pennsylvania, U.S.A.
(Accepted 8 February 1989)
Summary
Rabies virus is usually demonstrated in human or animal tissues using antigendetection or viral isolation techniques. Rabies virus RNA can be demonstrated in paraffin-embedded tissues using in situ hybridization. Negative (-) sense 35S- and ‘H-labeled RNA probes, specific for rabies virus nucleocapsid protein mRNA, were used for the detection of rabies virus RNA in the nervous system of mice experimentally infected with fixed and street strains of rabies virus. In situ hybridization signals were compared with rabies virus antigen demonstrated with immunoperoxidase staining. Rabies virus RNA and antigen were.also demonstrated in the same neurons using a double-labeling technique. In situ hybridization has potential applications as a diagnostic test for rabies and in studies of rabies pathogenesis. Encephalitis;
Immunohistochemistry;
In situ Hybridization;
Rabies
Introduction
Despite the presence of dramatic clinical signs in rabies, the neuropathologic findings are usually quite mild. In street rabies virus infection there is often a nonspecific inflammatory response. Although the hallmark of rabies is eosinophilic cytoplasmic inclusions called Negri bodies (Perl, 1975), these are not always present and more sophisticated techniques are usually required to determine whether tisCorrespondence to: A.C. Jackson, Department Kingston, Ontario, Canada K7L 357.
of Medicine, Queen’s University, 78 Barrie Street,
0166-0934/89/$03.50 @ 1989 Elsevier Science Publishers B.V. (Biomedical Division)
2
sues are infected by rabies virus. Viral proteins of rabies virus have been demonstrated in tissues for many years using antigen-detection techniques. Specimens are routinely submitted to diagnostic laboratories for the rapid detection of rabies virus antigens using immunofluorescence (Kissling, 1975). Immunoperoxidase staining techniques have also been used, particularly when good morphology is important in addition to requirements for the sensitive and specific detection of antigens (Bourgon and Charlton, 1987; Feiden et al., 1985; Fekadu et al., 1988). Rabies virus can also be isolated from tissues in cell culture (Wiktor, 1973) or using the mouse inoculation test (Koprowski, 1973). Advances in recombinant DNA technology have resulted in the ability to detect viral nucleic acids in tissues using hybridization techniques (Haase et al., 1984). In situ hybridization allows the localization of viral nucleic acids in tissues with preservation of the morphology of cells and tissues. In the present study, in situ hybridization was used for the localization of rabies virus RNA in the central nervous system (CNS) of expe~ment~y infected mice. Radiolabeled RNA probes were used because of their high sensitivity and low background (Angerer et al., 1987). In situ hybridization has potential applications for the diagnosis of rabies in selected cases and in future studies of rabies pathogenesis.
Materials and Methods Viruses
The CVS-11 strain of fixed rabies virus (Wistar Institute, Philadelphia, Pennsylvania) was used. Stock CVS virus was grown in BHK-21 cells to a titer of 4.2 x 10’ PFU/ml. A salivary gland homogenate (10% suspension) of an Ontario fox isolate of street rabies virus was obtained from Dr. K.M. Charlton (Animal Diseases Research Institute, Nepean, Ontario). The titer was 105.0mouse intracerebral 50% lethal doses (L~~~O.03 ml), calculated by the method of Reed and Muench (1938). Animals
Six-week-old female ICR mice (Charles River Canada, Inc., St-Constant, bec) were used.
Que-
Mice were inoculated subcutaneously in the Ieft hindlimb footpad or intracerebrally with 9.3 X lo5 PFU of CVS virus (10 subcutaneous LD,,) in 0.03 ml of PBS with 2% fetal bovine serum or in the left hindlimb footpad with 0.03 ml of the 10% suspension of street rabies virus. Neurologic signs developed 5 to 6 days after inoculation with CVS virus, and the mean date of death was day 9 (intracerebral inoculation) and 13 (subcutaneous inoculation). Signs developed a mean of 10 days
3
after inoculation with street virus and, in fatal cases, death occurred a mean of 14 days after inoculation (unpublished observations). Preparation of tissue sections
Mice were sacrificed at daily intervals after inoculation with CVS virus, and on alternate days after inoculation with street rabies virus. Mice were anesthetized with methoxyflurane and perfused with buffered 4% paraformaldehyde. Brain and spinal cords were removed, immersion-fixed in the same fixative for 18 h at 4°C dehydrated, and embedded in paraffin. Coronal sections of brain and transverse sections of brainstem and spinal cord were cut at multiple levels on a microtome (6 pm). Immunohistochemistry
Sections were stained for rabies virus antigen by the avidin-biotin-peroxidase method of Hsu et al. (1981) with minor modifications. Deparaffinized slides were successively reacted with 0.001% pepsin (Boehringer-Mannheim, Mannheim, F.R.G.) in 0.01 N HCl at 37°C for 30 min, 5% normal goat serum for 20 min, rabbit anti-rabies virus serum diluted 1:500 (obtained from Dr. K.M. Charlton) for 60 min, biotinylated goat anti-rabbit IgG diluted 1:lOO (Vector Laboratories, Burlingame, CA) for 30 min, 1% hydrogen peroxide in methanol for 30 min, avidin-biotinylated horseradish peroxidase complex (Vector Laboratories) for 30 min, 3,3’-diaminobenzidine tetrachloride (Polysciences, Inc., Warrington, PA) with 0.01% hydrogen peroxide for 8 min, 0.5% cupric sulfate in 0.15 M sodium chloride for 5 min, and the slides were counterstained with hematoxylin. Tissues from uninfected mice were used as a control. Normal rabbit serum diluted 1:500 was used as a primary antibody on tissues from infected mice as another control. In situ hybridization
In situ hybridization was performed by the method of Moench et al. (1988) with minor modifications. 35S- or 3H-labeled RNA probes were used for localization of rabies virus RNA in tissues. Deparaffinized slides were pretreated by sequential immersion in 0.2 N HCl for 20 min, 2 X SSC (1 x SSC = 0.15 M NaCl, 0.015 M sodium citrate, pH 7.4) for 30 min, 0.25% (v/v) acetic anhydride in 0.1 M triethanolamine HCl buffer (pH 8.0) for 10 min, and then they were rehydrated in graded alcohols and air dried. A cDNA clone containing 96% of the coding sequence for the nucleocapsid protein of the ERA strain of fixed rabies virus (J.K. Larson, W.H. Wunner, C. Rayssiguier, and P.J. Curtis, Wistar Institute, unpublished data) was used to prepare radiolabeled probes. The 1.28 kb fragment, representing 95% of the cDNA clone, was excised with PstI and AccI and subcloned into the promoter-containing pGEM-2 vector (Promega, Madison, WI). Radiolabeled probes were synthesized in the presence of [35S]UTP (New England Nuclear, Boston, MA) or [5,6-3H]UTP
4
(ICN Radiochemicals, Irvine, CA) using SP6 polymerase (Promega), which were specific for rabies virus nucleocapsid protein mRNA (and positive strand replicative intermediate). The probes were reduced in size by alkaline hydrolysis (Cox et al., 1984). The 35S-labeled probes had a specific activity of 2.1 x 10’ dpm/kg, and the 3H-labeled probes had specific activities of 4.8 6.1 x 10’ dpm/pg. An irrelevant control template (Riboprobe Gemini positive control template, Promega) was used to prepare 35S- and 3H-labeled RNA transcripts as a control of the specificity of the hybridization. The hybridization mixture contained 0.2 pg/ml of 35S- or 3H-labeled RNA transcripts, 50 mM dithiothreitol (DTT), 0.3 M NaCl, 50% (v/v) deionized formamide, 10% (w/v) dextran sulfate, 0.2 mg/ml sheared salmon sperm DNA, 0.125 mg/ml tRNA/ml, 0.02% (w/v) Ficoll, 0.02% (w/v) polyvinylpyrrolidone, 10 mM Tris (pH 7.4), 1 mM EDTA, and 0.1% Triton X-100. The mixture was applied to the slides for 4 h at 45°C. After hybridization the slides were washed twice in 4 x SSC with 5 mM DTT for 10 min, once in 0.06254.25 ug/ml (3H) or 2 kg/ml (35S) RNase A (Boehringer-Mannheim) in 0.5 M NaCl and 10 mM Tris (pH 8.0) for 30 min at 37”C, twice in 0.5 M NaCl and 10 mM Tris (pH 8.0) for 15 min at 37”C, twice in 0.1 x SSC and 0.1% SDS for 10 min at 50°C 2 x SSC for 10 min, dehydrated in graded alcohols (each containing 0.3 M ammonium acetate), and air dried. Slides were dipped in NTB2 nuclear track emulsion (Eastman Kodak Company, Rochester, New York) diluted 1:l with 0.6 M ammonium acetate and exposed for 24 h (35S) or 8-11 days (3H) at 4”C, and then developed with D19 developer (Eastman Kodak Company) for 5 min, fixed with 30% sodium thiosulfate for 5 min, and counterstained with hematoxylin. Controls included sections pretreated with RNase A (Boehringer-Mannheim) or micrococcal nuclease (Boehringer-Mannheim) (Williamson, 1988), in situ hybridization on uninfected tissues with the rabies virus RNA probe, and on rabies virus-infected tissues with a RNA probe prepared using the irrelevant control template (Promega). Double-labeling
of infected ceils
Immunohistochemical staining for rabies virus antigen was performed before in situ hybridization for rabies virus RNA as reported by Gendelman et al. (1985). The immunohistochemical method described above was used. The slides were counterstained with hematoxylin after in situ hybridization.
Results Cellular distribution of rabies virus RNA
Rabies virus RNA was detected by in situ hybridization only in CNS cells with the morphological appearance of neurons (Fig. 1) over the first five days of the CNS infection. Backgrounds were low, and no definite signals were found in controls (Fig. 1C). Signals were not detected in glial cells. The cellular localization of
5 ?
‘ .a*
.‘* ,
6 Fig. 1. In situ hybridization with a 3H-labeled probe for rabies virus RNA on a ventral horn cell (A) and a dorsal root ganglion neuron (B) from day 5 in CVS virus infection, a ventral horn cell of an uninfected mouse (C), and a ventral horn cell (D) from day 7 and Purkinje cell (E) from day 9 in street virus infection. There are many grains over the perikarya of infected neurons and over dendritic processes (A), but not over a neuron from an uninfected mouse (C). The signals are greater in neurons infected with fixed than street virus. Comparisons of serial sections using in situ hybridization with a 3H-labeled probe for rabies virus RNA (F, H, and J) and antigen (G, I, and K) in a dorsal root ganglion (F and G), brainstem neurons (H and I), and Purkinje cells (J and K) 5 days after hindlimb footpad inoculation with CVS virus. The two techniques gave similar signals, although rabies virus was better demonstrated in dendritic processes with immunoperoxidase staining than in situ hybridization (H-K). Hematoxylin. A, x 420; B, x 560; C, x 525; D, x 800; E, x 1250; F and G, x 188; H and I, x 220; J and K, x 630.
the 3H-labeled probes was much better than 35S-labeled probes. Grains were present over the perikarya, but not over the nuclei of infected neurons. Grains were also observed prominently over dendritic processes (Fig. 1A). Fixed rabies virus v’s street rabies virus infection
There were fewer neurons observed with rabies virus RNA in mice infected with street virus than with fixed virus at all time points. This was in keeping with the smaller number of neurons observed containing rabies virus antigen. In addition, infected neurons generally contained a smaller number of grains in street virus infection than fixed virus infection (Figs. 1A and 1B vs 1D and 1E). Distribution of rabies virus RNA vs antigen
In general, there was a good correlation with the distribution of rabies virus RNA and immunohistochemically-detected rabies virus antigen (Figs. lF-K and 2). There was a progressive increase in the number of neurons containing rabies virus RNA and antigen over the first five days of the CNS infection caused by CVS virus. In situ hybridization using 35S-labeled probes with short exposure times (24 h) had similar sensitivity for the detection of virus as immunoperoxidase staining for rabies virus antigen. However, the cellular localization of antigen was much better than RNA using the 35S-labeled probes. Signals with 3H-labeled probes with S-11 day exposures were, in general, less than after 24 h exposures with 35S-labeled probes. However, their cellular localization was much better than 35S-labeled probes, and approached that for antigen. Virus was better demonstrated in dendritic processes with antigen than RNA using in situ hybridization (Figs. lH-K and 2E, F) Double-labeling
of neurons for rabies virus RNA and antigen
Rabies virus RNA and antigen were also detected in the same cells (neurons) with the immunohistochemical detection of antigen followed by in situ hybridiza-
Fig. 2. In situ hybridization with a %-labeled probe on CVS virus-infected cervical spinal cord from day 6 with brightfield (A) and darkfield (C) optics, and uting the control template with brightfield optics (D). A serial section was stained for rabies virus antigen (B). The dist~bution of virus in the gray matter of the cord is well demonstrated with in situ hybridization, and compares favorabIy with the antigen-detection technique. In situ hybridization for rabies virus RNA with a 35S-labeled probe (E) and antigen (F) in a CVS virus-infected cerebellum from day 7. Both techniques show strong signals in the perikarya of Purkinje cells, but virus is better demonstrated with antigen than RNA in the dendrites of Purkinje ceils in the motecuiar layer. Hematoxylin. A-D, X 16; E and F, X 44.
with 35S- or 3H-labeled probes (Fig. 3). The immunoperoxidase stain was associated with a reduction in the hybridization signal, especially with 3H-labeled probes and when the amount of staining was high. Neurons were observed that contained antigen without any hybridization signal. However, in situ hybridization on adjacent sections demonstrated that signal was present, indicating that there was inhibition of the hybridization signal. tion
9 Fig. 3. Double-labeling of a ventral horn cell (top) from day 5 and dorsal root ganglion neuron (bottom) from day 6 in CVS virus infection. Immunoperoxidase staining for rabies virus antigen was followed by in situ hybridization for rabies virus RNA with ‘H-labeled (A) and ““S-labeled (B) probes. Signals for both RNA and antigen are shown in single cells. Immunoperoxidase-hematoxylin. A, x 1365; B, x 1630.
Discussion Specific nucleic acid sequences can be detected in tissues using in situ hybridization (Haase et al., 1984; Jones, 1973; Valentino et al., 1987). In situ hybridization can be used for the sensitive and specific localization of rabies virus RNA in infected tissues. Radiolabeled probes were used in the present study because of their improved sensitivity over non-isotopic methods by about two orders of magnitude (Haase, 1987). There are a number of advantages of single-stranded RNA probes over the more widely used DNA probes. RNA-RNA hybrids are more stable than DNA-RNA hybrids and, unlike double-stranded DNA probes, there is a lack of competition for hybridization in solution between different strands with the single-stranded RNA probes (Angerer et al., 1987; Moench, 1987). In addition, the background can be reduced with a post-hybridization RNase treatment that digests non-specifically bound single-stranded probe. The presence of high backgrounds in preliminary studies on rabies virus-infected mouse tissues with DNA probes resulted in using RNA probes in the present study. A cDNA clone encoding the rabies virus nucleocapsid protein was used as a probe because of the presumed high sequence homology between different rabies virus strains in this gene. Rabies virus RNA was detected in tissues infected by either fixed or street rabies virus strains using this probe. Ermine et al. (1988) have reported the detection of rabies virus RNA in infected tissues using a dot-blot hybridization assay. In situ hybridization offers the ability to assess infection in a small percentage of the total number of cells without loss of sensitivity due to a dilution effect from large quantities of cellular nucleic acids (Haase, 1986). This is an advantage in detecting early infection, when only a small number of cells are involved. This study has demonstrated that rabies virus RNA can be detected in paraffinembedded tissues in a sensitive and specific manner using in situ hybridization. The distribution of rabies virus RNA correlated well with the distribution of antigen revealed by immunoperoxidase staining. In addition, both viral RNA and antigen can be demonstrated in the same cells using a double-labeling technique. In situ hybridization is not likely to become a routine diagnostic test for rabies in the near future. However, in situ hybridization could be applied in selected cases when antigen-detection techniques (immunofluorescence or immunoperoxidase staining) are equivocal and viral isolation is negative or tissue is unavailable for isolation studies. Formalin-fixed paraffin-embedded tissues from routine autopsy material could also be examined, even after storage for many years (Moench et al., 1988). The development of a non-isotopic method (e.g. using biotin-labeled
10
using raprobes) (Singer et al., 1987), although less sensitive than this technique diolabeled probes, would be quicker, less expensive, and technically less difficult for wide application. Nevertheless, a number of controls would be required for demonstrating the specificity of the signals. In situ hybridization also has potential applications in the study of rabies pathogenesis. Infected tissues could be examined with strand-specific RNA probes complementary to different rabies virus genes in order to assess the molecular pathogenesis of infection in vivo. In situ hybridization could also be used in a doublelabeling technique (Haase, 1986) to determine the effect of rabies infection on transcription of important cellular genes. In situ hybridization has potential applications both as a diagnostic test in selected cases and in rabies pathogenesis studies .
Acknowledgements We thank Dr. Thomas R. Moench (The Johns Hopkins University) for helpful discussions, and Dr. K.M. Charlton (Animal Diseases Research Institute) for the street rabies virus and the anti-rabies virus serum. The secretarial assistance of Martha Steacy is gratefully acknowledged. This work was supported by Grant MA-10068 from the Medical Research Council of Canada, the Violet E. Powell Fund (Queen’s University), and Research Grants AI 18883 and AI 09706 from the National Institute of Allergy and Infectious Diseases, National Institutes of Health. References Angerer, L.M., Stoler, M.H. and Angerer, R.C. (1987) In situ hybridization with RNA probes: an annotated recipe. In: K.L. Valentino, J.H. Eberwine, J.D. Barchas (Eds), In Situ Hybridization: Applications to Neurobiology, pp. 42-70. Oxford University Press, New York. Bourgon, A.R. and Charlton, K.M. (1987) The demonstration of rabies antigen in paraffin-embedded tissues using peroxidase-antiperoxidase method: A comparative study. Can. J. Vet. Res. 51, 117-120. Cox, K.H., DeLeon, D.V., Angerer, L.M. and Angerer, R.C. (1984) Detection of mRNAs in sea urchin embryos by in situ hybridization using asymmetric RNA probes. Develop. Biol. 101, 485-502. Ermine, A., Tordo, N. and Tsiang, H. (1988) Rapid diagnosis of rabies infection by means of a dot hybridization assay. Mol. Cell. Probes 2, 75-82. Feiden, W., Feiden, U., Gerhard, L., Reinhardt, V. and Wandeler, A. (1985) Rabies encephalitis: immunohistochemical investigations. Clin. Neuropathol. 4, 156164. Fekadu, M., Greer, P.W., Chandler, F.W. and Sanderlin, D.W. (1988) Use of the avidin-biotin peroxidase system to detect rabies antigen in formalin-fixed paraffin-embedded tissues. J. Virol. Methods 19, 91-96. Gendelman, H.E., Moench, T.R., Narayan, O., Griffin, D.E. and Clements, J.E. (1985) A double labeling technique for performing immunocytochemistry and in situ hybridization in virus infected cell cultures and tissues. J. Virol. Methods 11, 93-103. Haase, A.T. (1986) Analysis of viral infections by in situ hybridization. .I. Histochem. Cytochem. 34, 27-32. Haase. A.T. (1987) Analysis of viral infections by in situ hybridization. In: K.L. Valentino, J.H. Eberwine and J.D. Barchas (Eds), In Situ Hybridization: Applications to Neurobiology, pp. 197-219. Oxford University Press, New York.
11 Haase, A., Brahic, M., Stowring, L. and Btum, H. (1984) Detection of viral nucleic acids by in situ hybridization. Methods Virol. 7, 189-226. Hsu, S.M., Raine, L. and Fanger, H. (1981) Use of avidin-biotin-peroxidase complex (ABC) in immunoperoxidase techniques: A comparison between ABC and unlabeled antibody (PAP) procedures. J. Histochem. Cytochem. 29, 577-580. Jones, K.W. (1973) The method of in situ hybridization. In: R.H. Pain and B.J. Smith (Eds), New Techniques in Biophysics and Cell Biology, Vol. 1, pp. 29-66. John Wiley & Sons, London. Kiss&g, R.E. (1975) The fluorescent antibody test in rabies. In: G.M. Baer (Ed.), The Natural History of Rabies, pp. 401-416. Academic Press, New York. Koprowski, H. (1973) The mouse inoculation test. In: M.M. Kaplan and H. Koprowski (Eds), Lahoratory Techniques in Rabies, Third edit. pp. 85-93. World Health Organization, Geneva. Moench, T.R. (1987) In situ hybridization. Mol. Cell. Probes 1, 195-205. Moench, T.R., Griffin, D.E., Obriecht, C.R., Vaisberg, A.J. and Johnson, R.T. (1988) Acute measles in patients with and without neurological involvement: Distribution of measles virus antigen and RNA. J. Infect. Dis. 158, 433-442. Perl, D.P. (1975) The pathology of rabies in the central nervous system. In: G.M. Baer (Ed), The Natural History of Rabies, Vol. 1, pp. 235-272. Academic Press, New York. Reed, L.J. and Muench, H. (1938) A simple method of estimating fifty per cent endpoints. Am. J. Hyg. 27, 493-497. Singer, R.H., Lawrence, J.B. and Rashtchian. R.N. (1987) Toward a rapid and sensitive in situ hybridization methodology using isotopic and nonisotopic probes. in: K.L. Valentino, J.H. Eberwine and J.D. Barchas (Eds), In Situ Hybridization: Applications to Neurobiology, pp. 71-96. Oxford University Press, New York. Valentino, K.L., Eberwine, J.H. and Barchas, J.D. (1987) In Situ Hybridization: Applications to Neurobiology. Oxford University Press, New York. Wiktor, T.J. (1973) Tissue culture methods. In: M.M. Kaplan and H. Koprowski (Eds). Laboratory Techniques in Rabies, Third edit. pp. 101-123. World Health Organizations Geneva. Williamson, D.J. (1988) Specificity of riboprobes for intracellular RNA in hybridization histochemistry. J. Histochem. Cytochem. 36, 811-813.
Elsevier JVM 00887
Detection of rabies virus RNA in the central nervous system of experimentally infected mice using in situ hybridization with RNA probes Alan C. Jacksoni**,
Dorothy
L. Reimer2
and William
H. Wunner3
Departments of ‘Medicine and ‘Microbiology d Immunology, Queen’s University, Kingston, Ontario, Canada and 3The Wistar Institute of Anatomy and Biology, Philadelphia, Pennsylvania, U.S.A.
(Accepted 8 February 1989)
Summary
Rabies virus is usually demonstrated in human or animal tissues using antigendetection or viral isolation techniques. Rabies virus RNA can be demonstrated in paraffin-embedded tissues using in situ hybridization. Negative (-) sense 35S- and ‘H-labeled RNA probes, specific for rabies virus nucleocapsid protein mRNA, were used for the detection of rabies virus RNA in the nervous system of mice experimentally infected with fixed and street strains of rabies virus. In situ hybridization signals were compared with rabies virus antigen demonstrated with immunoperoxidase staining. Rabies virus RNA and antigen were.also demonstrated in the same neurons using a double-labeling technique. In situ hybridization has potential applications as a diagnostic test for rabies and in studies of rabies pathogenesis. Encephalitis;
Immunohistochemistry;
In situ Hybridization;
Rabies
Introduction
Despite the presence of dramatic clinical signs in rabies, the neuropathologic findings are usually quite mild. In street rabies virus infection there is often a nonspecific inflammatory response. Although the hallmark of rabies is eosinophilic cytoplasmic inclusions called Negri bodies (Perl, 1975), these are not always present and more sophisticated techniques are usually required to determine whether tisCorrespondence to: A.C. Jackson, Department Kingston, Ontario, Canada K7L 357.
of Medicine, Queen’s University, 78 Barrie Street,
0166-0934/89/$03.50 @ 1989 Elsevier Science Publishers B.V. (Biomedical Division)
2
sues are infected by rabies virus. Viral proteins of rabies virus have been demonstrated in tissues for many years using antigen-detection techniques. Specimens are routinely submitted to diagnostic laboratories for the rapid detection of rabies virus antigens using immunofluorescence (Kissling, 1975). Immunoperoxidase staining techniques have also been used, particularly when good morphology is important in addition to requirements for the sensitive and specific detection of antigens (Bourgon and Charlton, 1987; Feiden et al., 1985; Fekadu et al., 1988). Rabies virus can also be isolated from tissues in cell culture (Wiktor, 1973) or using the mouse inoculation test (Koprowski, 1973). Advances in recombinant DNA technology have resulted in the ability to detect viral nucleic acids in tissues using hybridization techniques (Haase et al., 1984). In situ hybridization allows the localization of viral nucleic acids in tissues with preservation of the morphology of cells and tissues. In the present study, in situ hybridization was used for the localization of rabies virus RNA in the central nervous system (CNS) of expe~ment~y infected mice. Radiolabeled RNA probes were used because of their high sensitivity and low background (Angerer et al., 1987). In situ hybridization has potential applications for the diagnosis of rabies in selected cases and in future studies of rabies pathogenesis.
Materials and Methods Viruses
The CVS-11 strain of fixed rabies virus (Wistar Institute, Philadelphia, Pennsylvania) was used. Stock CVS virus was grown in BHK-21 cells to a titer of 4.2 x 10’ PFU/ml. A salivary gland homogenate (10% suspension) of an Ontario fox isolate of street rabies virus was obtained from Dr. K.M. Charlton (Animal Diseases Research Institute, Nepean, Ontario). The titer was 105.0mouse intracerebral 50% lethal doses (L~~~O.03 ml), calculated by the method of Reed and Muench (1938). Animals
Six-week-old female ICR mice (Charles River Canada, Inc., St-Constant, bec) were used.
Que-
Mice were inoculated subcutaneously in the Ieft hindlimb footpad or intracerebrally with 9.3 X lo5 PFU of CVS virus (10 subcutaneous LD,,) in 0.03 ml of PBS with 2% fetal bovine serum or in the left hindlimb footpad with 0.03 ml of the 10% suspension of street rabies virus. Neurologic signs developed 5 to 6 days after inoculation with CVS virus, and the mean date of death was day 9 (intracerebral inoculation) and 13 (subcutaneous inoculation). Signs developed a mean of 10 days
3
after inoculation with street virus and, in fatal cases, death occurred a mean of 14 days after inoculation (unpublished observations). Preparation of tissue sections
Mice were sacrificed at daily intervals after inoculation with CVS virus, and on alternate days after inoculation with street rabies virus. Mice were anesthetized with methoxyflurane and perfused with buffered 4% paraformaldehyde. Brain and spinal cords were removed, immersion-fixed in the same fixative for 18 h at 4°C dehydrated, and embedded in paraffin. Coronal sections of brain and transverse sections of brainstem and spinal cord were cut at multiple levels on a microtome (6 pm). Immunohistochemistry
Sections were stained for rabies virus antigen by the avidin-biotin-peroxidase method of Hsu et al. (1981) with minor modifications. Deparaffinized slides were successively reacted with 0.001% pepsin (Boehringer-Mannheim, Mannheim, F.R.G.) in 0.01 N HCl at 37°C for 30 min, 5% normal goat serum for 20 min, rabbit anti-rabies virus serum diluted 1:500 (obtained from Dr. K.M. Charlton) for 60 min, biotinylated goat anti-rabbit IgG diluted 1:lOO (Vector Laboratories, Burlingame, CA) for 30 min, 1% hydrogen peroxide in methanol for 30 min, avidin-biotinylated horseradish peroxidase complex (Vector Laboratories) for 30 min, 3,3’-diaminobenzidine tetrachloride (Polysciences, Inc., Warrington, PA) with 0.01% hydrogen peroxide for 8 min, 0.5% cupric sulfate in 0.15 M sodium chloride for 5 min, and the slides were counterstained with hematoxylin. Tissues from uninfected mice were used as a control. Normal rabbit serum diluted 1:500 was used as a primary antibody on tissues from infected mice as another control. In situ hybridization
In situ hybridization was performed by the method of Moench et al. (1988) with minor modifications. 35S- or 3H-labeled RNA probes were used for localization of rabies virus RNA in tissues. Deparaffinized slides were pretreated by sequential immersion in 0.2 N HCl for 20 min, 2 X SSC (1 x SSC = 0.15 M NaCl, 0.015 M sodium citrate, pH 7.4) for 30 min, 0.25% (v/v) acetic anhydride in 0.1 M triethanolamine HCl buffer (pH 8.0) for 10 min, and then they were rehydrated in graded alcohols and air dried. A cDNA clone containing 96% of the coding sequence for the nucleocapsid protein of the ERA strain of fixed rabies virus (J.K. Larson, W.H. Wunner, C. Rayssiguier, and P.J. Curtis, Wistar Institute, unpublished data) was used to prepare radiolabeled probes. The 1.28 kb fragment, representing 95% of the cDNA clone, was excised with PstI and AccI and subcloned into the promoter-containing pGEM-2 vector (Promega, Madison, WI). Radiolabeled probes were synthesized in the presence of [35S]UTP (New England Nuclear, Boston, MA) or [5,6-3H]UTP
4
(ICN Radiochemicals, Irvine, CA) using SP6 polymerase (Promega), which were specific for rabies virus nucleocapsid protein mRNA (and positive strand replicative intermediate). The probes were reduced in size by alkaline hydrolysis (Cox et al., 1984). The 35S-labeled probes had a specific activity of 2.1 x 10’ dpm/kg, and the 3H-labeled probes had specific activities of 4.8 6.1 x 10’ dpm/pg. An irrelevant control template (Riboprobe Gemini positive control template, Promega) was used to prepare 35S- and 3H-labeled RNA transcripts as a control of the specificity of the hybridization. The hybridization mixture contained 0.2 pg/ml of 35S- or 3H-labeled RNA transcripts, 50 mM dithiothreitol (DTT), 0.3 M NaCl, 50% (v/v) deionized formamide, 10% (w/v) dextran sulfate, 0.2 mg/ml sheared salmon sperm DNA, 0.125 mg/ml tRNA/ml, 0.02% (w/v) Ficoll, 0.02% (w/v) polyvinylpyrrolidone, 10 mM Tris (pH 7.4), 1 mM EDTA, and 0.1% Triton X-100. The mixture was applied to the slides for 4 h at 45°C. After hybridization the slides were washed twice in 4 x SSC with 5 mM DTT for 10 min, once in 0.06254.25 ug/ml (3H) or 2 kg/ml (35S) RNase A (Boehringer-Mannheim) in 0.5 M NaCl and 10 mM Tris (pH 8.0) for 30 min at 37”C, twice in 0.5 M NaCl and 10 mM Tris (pH 8.0) for 15 min at 37”C, twice in 0.1 x SSC and 0.1% SDS for 10 min at 50°C 2 x SSC for 10 min, dehydrated in graded alcohols (each containing 0.3 M ammonium acetate), and air dried. Slides were dipped in NTB2 nuclear track emulsion (Eastman Kodak Company, Rochester, New York) diluted 1:l with 0.6 M ammonium acetate and exposed for 24 h (35S) or 8-11 days (3H) at 4”C, and then developed with D19 developer (Eastman Kodak Company) for 5 min, fixed with 30% sodium thiosulfate for 5 min, and counterstained with hematoxylin. Controls included sections pretreated with RNase A (Boehringer-Mannheim) or micrococcal nuclease (Boehringer-Mannheim) (Williamson, 1988), in situ hybridization on uninfected tissues with the rabies virus RNA probe, and on rabies virus-infected tissues with a RNA probe prepared using the irrelevant control template (Promega). Double-labeling
of infected ceils
Immunohistochemical staining for rabies virus antigen was performed before in situ hybridization for rabies virus RNA as reported by Gendelman et al. (1985). The immunohistochemical method described above was used. The slides were counterstained with hematoxylin after in situ hybridization.
Results Cellular distribution of rabies virus RNA
Rabies virus RNA was detected by in situ hybridization only in CNS cells with the morphological appearance of neurons (Fig. 1) over the first five days of the CNS infection. Backgrounds were low, and no definite signals were found in controls (Fig. 1C). Signals were not detected in glial cells. The cellular localization of
5 ?
‘ .a*
.‘* ,
6 Fig. 1. In situ hybridization with a 3H-labeled probe for rabies virus RNA on a ventral horn cell (A) and a dorsal root ganglion neuron (B) from day 5 in CVS virus infection, a ventral horn cell of an uninfected mouse (C), and a ventral horn cell (D) from day 7 and Purkinje cell (E) from day 9 in street virus infection. There are many grains over the perikarya of infected neurons and over dendritic processes (A), but not over a neuron from an uninfected mouse (C). The signals are greater in neurons infected with fixed than street virus. Comparisons of serial sections using in situ hybridization with a 3H-labeled probe for rabies virus RNA (F, H, and J) and antigen (G, I, and K) in a dorsal root ganglion (F and G), brainstem neurons (H and I), and Purkinje cells (J and K) 5 days after hindlimb footpad inoculation with CVS virus. The two techniques gave similar signals, although rabies virus was better demonstrated in dendritic processes with immunoperoxidase staining than in situ hybridization (H-K). Hematoxylin. A, x 420; B, x 560; C, x 525; D, x 800; E, x 1250; F and G, x 188; H and I, x 220; J and K, x 630.
the 3H-labeled probes was much better than 35S-labeled probes. Grains were present over the perikarya, but not over the nuclei of infected neurons. Grains were also observed prominently over dendritic processes (Fig. 1A). Fixed rabies virus v’s street rabies virus infection
There were fewer neurons observed with rabies virus RNA in mice infected with street virus than with fixed virus at all time points. This was in keeping with the smaller number of neurons observed containing rabies virus antigen. In addition, infected neurons generally contained a smaller number of grains in street virus infection than fixed virus infection (Figs. 1A and 1B vs 1D and 1E). Distribution of rabies virus RNA vs antigen
In general, there was a good correlation with the distribution of rabies virus RNA and immunohistochemically-detected rabies virus antigen (Figs. lF-K and 2). There was a progressive increase in the number of neurons containing rabies virus RNA and antigen over the first five days of the CNS infection caused by CVS virus. In situ hybridization using 35S-labeled probes with short exposure times (24 h) had similar sensitivity for the detection of virus as immunoperoxidase staining for rabies virus antigen. However, the cellular localization of antigen was much better than RNA using the 35S-labeled probes. Signals with 3H-labeled probes with S-11 day exposures were, in general, less than after 24 h exposures with 35S-labeled probes. However, their cellular localization was much better than 35S-labeled probes, and approached that for antigen. Virus was better demonstrated in dendritic processes with antigen than RNA using in situ hybridization (Figs. lH-K and 2E, F) Double-labeling
of neurons for rabies virus RNA and antigen
Rabies virus RNA and antigen were also detected in the same cells (neurons) with the immunohistochemical detection of antigen followed by in situ hybridiza-
Fig. 2. In situ hybridization with a %-labeled probe on CVS virus-infected cervical spinal cord from day 6 with brightfield (A) and darkfield (C) optics, and uting the control template with brightfield optics (D). A serial section was stained for rabies virus antigen (B). The dist~bution of virus in the gray matter of the cord is well demonstrated with in situ hybridization, and compares favorabIy with the antigen-detection technique. In situ hybridization for rabies virus RNA with a 35S-labeled probe (E) and antigen (F) in a CVS virus-infected cerebellum from day 7. Both techniques show strong signals in the perikarya of Purkinje cells, but virus is better demonstrated with antigen than RNA in the dendrites of Purkinje ceils in the motecuiar layer. Hematoxylin. A-D, X 16; E and F, X 44.
with 35S- or 3H-labeled probes (Fig. 3). The immunoperoxidase stain was associated with a reduction in the hybridization signal, especially with 3H-labeled probes and when the amount of staining was high. Neurons were observed that contained antigen without any hybridization signal. However, in situ hybridization on adjacent sections demonstrated that signal was present, indicating that there was inhibition of the hybridization signal. tion
9 Fig. 3. Double-labeling of a ventral horn cell (top) from day 5 and dorsal root ganglion neuron (bottom) from day 6 in CVS virus infection. Immunoperoxidase staining for rabies virus antigen was followed by in situ hybridization for rabies virus RNA with ‘H-labeled (A) and ““S-labeled (B) probes. Signals for both RNA and antigen are shown in single cells. Immunoperoxidase-hematoxylin. A, x 1365; B, x 1630.
Discussion Specific nucleic acid sequences can be detected in tissues using in situ hybridization (Haase et al., 1984; Jones, 1973; Valentino et al., 1987). In situ hybridization can be used for the sensitive and specific localization of rabies virus RNA in infected tissues. Radiolabeled probes were used in the present study because of their improved sensitivity over non-isotopic methods by about two orders of magnitude (Haase, 1987). There are a number of advantages of single-stranded RNA probes over the more widely used DNA probes. RNA-RNA hybrids are more stable than DNA-RNA hybrids and, unlike double-stranded DNA probes, there is a lack of competition for hybridization in solution between different strands with the single-stranded RNA probes (Angerer et al., 1987; Moench, 1987). In addition, the background can be reduced with a post-hybridization RNase treatment that digests non-specifically bound single-stranded probe. The presence of high backgrounds in preliminary studies on rabies virus-infected mouse tissues with DNA probes resulted in using RNA probes in the present study. A cDNA clone encoding the rabies virus nucleocapsid protein was used as a probe because of the presumed high sequence homology between different rabies virus strains in this gene. Rabies virus RNA was detected in tissues infected by either fixed or street rabies virus strains using this probe. Ermine et al. (1988) have reported the detection of rabies virus RNA in infected tissues using a dot-blot hybridization assay. In situ hybridization offers the ability to assess infection in a small percentage of the total number of cells without loss of sensitivity due to a dilution effect from large quantities of cellular nucleic acids (Haase, 1986). This is an advantage in detecting early infection, when only a small number of cells are involved. This study has demonstrated that rabies virus RNA can be detected in paraffinembedded tissues in a sensitive and specific manner using in situ hybridization. The distribution of rabies virus RNA correlated well with the distribution of antigen revealed by immunoperoxidase staining. In addition, both viral RNA and antigen can be demonstrated in the same cells using a double-labeling technique. In situ hybridization is not likely to become a routine diagnostic test for rabies in the near future. However, in situ hybridization could be applied in selected cases when antigen-detection techniques (immunofluorescence or immunoperoxidase staining) are equivocal and viral isolation is negative or tissue is unavailable for isolation studies. Formalin-fixed paraffin-embedded tissues from routine autopsy material could also be examined, even after storage for many years (Moench et al., 1988). The development of a non-isotopic method (e.g. using biotin-labeled
10
using raprobes) (Singer et al., 1987), although less sensitive than this technique diolabeled probes, would be quicker, less expensive, and technically less difficult for wide application. Nevertheless, a number of controls would be required for demonstrating the specificity of the signals. In situ hybridization also has potential applications in the study of rabies pathogenesis. Infected tissues could be examined with strand-specific RNA probes complementary to different rabies virus genes in order to assess the molecular pathogenesis of infection in vivo. In situ hybridization could also be used in a doublelabeling technique (Haase, 1986) to determine the effect of rabies infection on transcription of important cellular genes. In situ hybridization has potential applications both as a diagnostic test in selected cases and in rabies pathogenesis studies .
Acknowledgements We thank Dr. Thomas R. Moench (The Johns Hopkins University) for helpful discussions, and Dr. K.M. Charlton (Animal Diseases Research Institute) for the street rabies virus and the anti-rabies virus serum. The secretarial assistance of Martha Steacy is gratefully acknowledged. This work was supported by Grant MA-10068 from the Medical Research Council of Canada, the Violet E. Powell Fund (Queen’s University), and Research Grants AI 18883 and AI 09706 from the National Institute of Allergy and Infectious Diseases, National Institutes of Health. References Angerer, L.M., Stoler, M.H. and Angerer, R.C. (1987) In situ hybridization with RNA probes: an annotated recipe. In: K.L. Valentino, J.H. Eberwine, J.D. Barchas (Eds), In Situ Hybridization: Applications to Neurobiology, pp. 42-70. Oxford University Press, New York. Bourgon, A.R. and Charlton, K.M. (1987) The demonstration of rabies antigen in paraffin-embedded tissues using peroxidase-antiperoxidase method: A comparative study. Can. J. Vet. Res. 51, 117-120. Cox, K.H., DeLeon, D.V., Angerer, L.M. and Angerer, R.C. (1984) Detection of mRNAs in sea urchin embryos by in situ hybridization using asymmetric RNA probes. Develop. Biol. 101, 485-502. Ermine, A., Tordo, N. and Tsiang, H. (1988) Rapid diagnosis of rabies infection by means of a dot hybridization assay. Mol. Cell. Probes 2, 75-82. Feiden, W., Feiden, U., Gerhard, L., Reinhardt, V. and Wandeler, A. (1985) Rabies encephalitis: immunohistochemical investigations. Clin. Neuropathol. 4, 156164. Fekadu, M., Greer, P.W., Chandler, F.W. and Sanderlin, D.W. (1988) Use of the avidin-biotin peroxidase system to detect rabies antigen in formalin-fixed paraffin-embedded tissues. J. Virol. Methods 19, 91-96. Gendelman, H.E., Moench, T.R., Narayan, O., Griffin, D.E. and Clements, J.E. (1985) A double labeling technique for performing immunocytochemistry and in situ hybridization in virus infected cell cultures and tissues. J. Virol. Methods 11, 93-103. Haase, A.T. (1986) Analysis of viral infections by in situ hybridization. .I. Histochem. Cytochem. 34, 27-32. Haase. A.T. (1987) Analysis of viral infections by in situ hybridization. In: K.L. Valentino, J.H. Eberwine and J.D. Barchas (Eds), In Situ Hybridization: Applications to Neurobiology, pp. 197-219. Oxford University Press, New York.
11 Haase, A., Brahic, M., Stowring, L. and Btum, H. (1984) Detection of viral nucleic acids by in situ hybridization. Methods Virol. 7, 189-226. Hsu, S.M., Raine, L. and Fanger, H. (1981) Use of avidin-biotin-peroxidase complex (ABC) in immunoperoxidase techniques: A comparison between ABC and unlabeled antibody (PAP) procedures. J. Histochem. Cytochem. 29, 577-580. Jones, K.W. (1973) The method of in situ hybridization. In: R.H. Pain and B.J. Smith (Eds), New Techniques in Biophysics and Cell Biology, Vol. 1, pp. 29-66. John Wiley & Sons, London. Kiss&g, R.E. (1975) The fluorescent antibody test in rabies. In: G.M. Baer (Ed.), The Natural History of Rabies, pp. 401-416. Academic Press, New York. Koprowski, H. (1973) The mouse inoculation test. In: M.M. Kaplan and H. Koprowski (Eds), Lahoratory Techniques in Rabies, Third edit. pp. 85-93. World Health Organization, Geneva. Moench, T.R. (1987) In situ hybridization. Mol. Cell. Probes 1, 195-205. Moench, T.R., Griffin, D.E., Obriecht, C.R., Vaisberg, A.J. and Johnson, R.T. (1988) Acute measles in patients with and without neurological involvement: Distribution of measles virus antigen and RNA. J. Infect. Dis. 158, 433-442. Perl, D.P. (1975) The pathology of rabies in the central nervous system. In: G.M. Baer (Ed), The Natural History of Rabies, Vol. 1, pp. 235-272. Academic Press, New York. Reed, L.J. and Muench, H. (1938) A simple method of estimating fifty per cent endpoints. Am. J. Hyg. 27, 493-497. Singer, R.H., Lawrence, J.B. and Rashtchian. R.N. (1987) Toward a rapid and sensitive in situ hybridization methodology using isotopic and nonisotopic probes. in: K.L. Valentino, J.H. Eberwine and J.D. Barchas (Eds), In Situ Hybridization: Applications to Neurobiology, pp. 71-96. Oxford University Press, New York. Valentino, K.L., Eberwine, J.H. and Barchas, J.D. (1987) In Situ Hybridization: Applications to Neurobiology. Oxford University Press, New York. Wiktor, T.J. (1973) Tissue culture methods. In: M.M. Kaplan and H. Koprowski (Eds). Laboratory Techniques in Rabies, Third edit. pp. 101-123. World Health Organizations Geneva. Williamson, D.J. (1988) Specificity of riboprobes for intracellular RNA in hybridization histochemistry. J. Histochem. Cytochem. 36, 811-813.