VIROLOGY
g&265-268
(1979)
Detection
DENISE
of Herpes
A. GALLOWAY,*,‘**
Simplex
RNA in Human
CECILIA FENOGLIO,+ K. MCDOUGALL*,*
Sensory
MARIA
Ganglia
SHEVCHUK,+
AND JAMES *Cold
Spring Harbor Laboratory, P.O. Box 100, Cold Spring Harbor, New York 11724, and ‘Departmnt Pathology, Columbia University, College of Physicians and Surgeons, New York, New York 10052 Accepted
February
of
18,/979
Human paravertebral ganglia were examined for the presence of viral-specific RNA by cytological hybridization using a nick-translated 3H-labeled herpes simplex virus-2 DNA probe. Ganglia from several locations (lumbar, thoracic, etc.) were obtained from seven patients at autopsy. In two cases HSV RNA was detected in all of the ganglia assayed. The autoradiographic grains were localized in the neurons (ganglions cells), and only a fraction of the neurons in any one ganglion were synthesizing viral-specific RNA. These results argue that the HSV genome can be active in the latent state.
in situ
As a consequence of an initial infection, herpes simplex virus (HSV) has been shown to persist in a latent state in the sensory ganglia of mice (1, 2), rabbits (3), monkeys (4), and humans (5, 6) as demonstrated by the recovery of virus. Viral DNA has been detected in association with neurons by in situ hybridization using as probe radioactive complementary RNA made from HSV DNA (7, 8). The state of the viral genome during a latent infection is still an unsettled question. For heuristic reasons it is convenient to consider two hypotheses: either the virus or its genome persists in a nonreplicating state, or it replicates so slowly that infectious virus, or viral gene products, cannot be detected by standard techniques. Several lines of evidence support the nonreplicating model. First, temperature-sensitive mutants are able to establish latent infections, suggesting that viral DNA replication is not necessary for a latent infection (9). Second, extensive searches for infectious virus or other viral specific products detectable by immunofluorescence or ultrastructural analysis, have almost consistently yielded negative results (10); and third, when the presence of viral-specific nucleic acids was ’ To whom reprint requests should be sent. * Present address: Fred Hutchinson Cancer Research Center, 1124 Columbia St., Seattle, Washington 98194. 265
assayed by renaturation kinetics, only 0.11 * 0.03 genome equivalents of viral DNA per cell and no viral-specific RNA was detected in latently infected ganglia (11). These latter approaches may not be sufficiently sensitive to detect viral gene expression in only a few cells of the ganglion. In order to establish whether transcription of the latent HSV genome occurs in individual neurons we turned to in situ cytological hybridization. By using a radioactive DNA probe to detect viral RNA, the target sequences represent an amplification of the persistent viral genome. We have hybridized either iodinated DNA (12) or tritiated nick-translated DNA (13) to a variety of cells and tissues. Viral RNA could be detected in HSV infected BHK cells, in clones of the HSV-2 transformed cell line 333-8-9, in sections of tumors induced in hamsters by the transformed cells, and in HSV transfected-thymidine kinase transformed mouse cells. No autoradiographic grains were seen when hybridized to uninfected or untransformed rodent cells, or to a variety of tissues of rodent or human origin. The hybridization probe was prepared using HSV-2 DNA extracted from infected hamster (BHK) cells, purified by centrifugation in sodium iodide equilibrium density gradients (14) and labeled with tritium in vitro by the process of nick-translation 0042~6822/‘79/070265-04$02.00/O Copyright Q 1979 by Academic All rights of reproduction in
Press,
Inc.
any form reserved.
266
SHORT COMMUNICATIONS
(15). To determine whether the in vitrolabeled DNA represented a faithful copy of the viral genome, a 32P-nick-translated probe (spec. act. 1 x lOa cpm/ug) was prepared and then hybridized to nitrocellulose filters which contained restriction endonuclease produced fragments of HSV-2 DNA transferred from agarose gels (16). The pattern obtained was consistent with all regions of the genome having been copied (data not shown). To determine the specificity of the probe, DNA from a variety of sources (human, monkey, hamster, rat, adenovirus 2, SV40, Escherichia coli; and others) was cleaved with restriction enzymes, fractionated by electrophoresis, and transferred to nitrocellulose. Using a 32P-labeled probe under conditions which detected 5 x lop5 pg of HSV-2 DNA/pg of cell DNA, no hybridization to sequences of human DNA was detected. It is worth noting that viral DNA prepared from BHK cells showed significant hybridization to hamster DNA and likewise that viral DNA isolated from virus grown in BSC-1 cells hybridized to monkey cell DNA. (For details see McDougall et al., manuscript in preparation.) This rigorous characterization of the probe gives us confidence that autoradiographic grains observed with specimens of human origin represent hybridization to HSV-2-specific sequences. At autopsy, within a few hours of death, paravertebral ganglia from several locations (lumbar, thoracic, etc.) were removed from seven patients, sectioned with a cryostat onto glass slides and fixed in -20” ethanol. The sections were hybridized to 3H-labeled HSV-2 DNA and exposed for autoradiography for 4 weeks. A summary of the detection of autoradiographic grains is given in Table 1. In two cases HSV RNA was detected in all of the ganglia assayed. Ganglion tissue is composed primarily of support cells and only a few of the total cells are neurons or ganglion cells. The autoradiographic grains are associated only with the neurons not with any of the surrounding tissue. And, only a small percentage of the neurons (0.04-&O%) are expressing HSV RNA. When ganglia are stained with Giemsa, Nissl bodies which are located in the body of the neuron and
TABLE
1
DETECTIONOF HSV RNA INHUMANSENSORY GANGLIABYINSITU HYBRIDIZATION Case No. 1 1 2 2
3 3 4 4 4 5 5 6 7
Location of ganglia
HSV RNA
Percentage of positive neurons
Sacral Thoracic Sacral Thoracic Sacral Thoracic Sacral Thoracic Lumbar Sacral
+ + + + + -
8 8 0 0 0 0 0.5 0.4 0.4 0
Lumbar Sacral Sacral
-
0 0 0
consist of endoplasmic reticulum and ribosomes appear as dark, granular organelles. One must be careful to distinguish these structures from autoradiographic grains. A high power magnification (Fig. 1A) of a region of the ganglion shows the location of authentic grains over the Nissl bodies indicating that the neurons are elaborating HSV-specific RNA. Figure 1B shows a ganglion in which no autoradiographic grains were detected. The radioactive signal could be abolished by pretreatment of the tissue with alkali indicating that the target sequences are RNA. To further convince ourselves that the observed autoradiographic grains using HSV probes represented true viral RNA, other 3H-labeled DNA probes (SV40, adenovirus 2 and 12, and mouse satellite) prepared in the same manner, were hybridized to various sections of human tissues. These results were negative. Because of the paucity of neurons within a ganglion which express HSV RNA, biochemical analysis of the viral nucleic acids present in ganglia is generally not feasible. Puga et al. (11) acknowledged that under the hybridization conditions they used, as many as 400 mRNA molecules per ganglion could be present and escape detection. In our study, only 0.4 to 8.0% of the neurons were positive for HSV RNA. It is not yet
SHORT COMMUNICATIONS
267
FIG. 1. Localization of HSV RNA in neurons of human sensory ganglia. A sacral ganglion was removed at autopsy, flash frozen, sectioned onto acid-washed glass slides, and placed in -20” ethanol. The hybridization reaction was in a total volume of 10 ~1 containing 6 x SSC and 5 x lo4 of 3H-labeled HSV-2 DNA (spec. act. 1 x 10’ cpm/pg DNA) denatured in alkali and neutralized. A siliconized coverslip was placed over the mixture, the slides were placed in a tray containing Whatman paper saturated with 20 x SSC and incubated at 68” for 18 hr. The slides were washed extensively with 2 x SSC at 4”, dried, dipped in Kodak NTB photographic emulsion, and exposed for autoradiography for 4 weeks at 4”. The emulsion was developed and the cells stained with Giemsa. (A) A high power magnification (800x) shows autoradiographic grains covering the body of a neuron with very low background in surrounding tissue. (B) In a neuron which does not show any hybridization to HSV-2 DNA the Nissl body is seen as a darkly staining granular structure.
clear whether all of the neurons contain HSV viral DNA and only some cells express viral genes, or whether only some neurons are latently infected. Also because of the homology shared by the two genomes, we are unable to discriminate between HSV-land HSV-Z-specific RNA with our present hybridization conditions. None of the patients from whom ganglia were obtained had been undergoing a herpes infection nor suffered from recurrent herpetic infections, therefore, there is no reason to assume that the HSV RNA detected was related to reactivation of the virus. We had not expected that from
one patient, all of the ganglia would be either all positive or all negative. Baringer (6) had reported that virus was often present bilaterally and could be recovered from more than one site. These results might argue persons harboring HSV are susceptible to latent infections in many sites. Our results neither support nor argue against a model in which HSV is retained in a nonreplicating state. The viral RNA detected might represent extensive transcription of the viral genome necessary for viral replication. Alternatively, the RNA might represent a transcript from a limited
268
SHORT
COMMUNICATIONS
region of the genome, with the expression of one or more viral genes a necessary factor for the maintenance of the latent state. In latent infections established by other herpes viruses, e.g., Epstein-Barr virus and the Luck6 virus, viral-specific RNA, and antigens have been detected in association with cells harboring the latent virus (10). We have presented evidence which indicates that some transcription of the herpes simplex virus genome occurs in latently infected human sensory ganglia. Experiments are under way to determine whether the RNA represents a specific and limited set of viral transcripts which are essential for the maintenance of a latent infection. ACKNOWLEDGMENTS We thank Beth Stewart for preparing the HSV-2 DNA, Bob Yaffe for the illustrations, and Tammie Tsujikawa for typing the manuscript. D.A.G. was the recipient of a PHS research fellowship from NCI. C.F. is supported in part by Grant BRSG 10113-39966. This work was funded by a Cancer Center grant to C.S.H. and by NSF Grant PCM77-11835 to J.K.M. REFERENCES 1. STEVENS, 843-845
J. G., and COOK, (1971).
M.
L., Science
173,
2. COOK, J. L., and STEVENS, J. G., J. Gen. Viral. 31, 75-80 (1976). 3. STEVENS, J. G., NESBURN, .A. B., and COOK, M. L., Nature New Biol. 235, 216-217 (1972). L. REEVES, W. C., DIGIACOMO, R. F., ALEXANDER, E. R., and LEE, C. K., Proc. Sot. Exp. Biol. Med. 153, 258-261 (1976). 5. BASTIAN, F. O., RABSON, A. S., YEE, C. L., and TRALKA, T. S., Science 178, 366-307 (1972). 6. BARINGER, J. R., N. Engl. J. Med. 291,828-830 (1974). 7. STEVENS, J. G., and COOK, M. L., J. Immunol. 113, 1685-1693 (1974). 8. ZUR HAUSEN, H., and SCHULTE-HOLTHAUSEN, H.,ZAECSci.Publ. ll(PartI), 117-123(1975). 9. LOFGREN, R. W., STEVENS, J. G., MARSDEN, H. S., and SUBAK-SHARPE, J. H., Virology 76, 440-443 (1977). 10. STEVENS, J. G., Curr. Top. Microbial. Immunol. 40, 31-50 (1975). 11. PUGA, A., ROSENTHAL, J. D., OPENSHAW, H., and NOTKINS, A. L., Virology 89, 102-111 (1978). 12. COPPLE, C. D., and MCDOUGALL, J. K., ht. J. Cancer 17, 501-510(1976). 13. MCDOUGALL, J. K., and GALLOWAY, D. A., in “Persistent Viruses,” Vol. 11 (J. Stevens, G. Todaro, C. F. Fox, eds.) Academic Press, New York, pp. 181-188, 1978. J. M. M., and SCHEGGET, J. T., 14. WALBOOMERS, Virology 74, 256-258 (1976). 15. MANIATIS, T., KEE, S. G., EFSTRADIADIS, A., and KAFATOS, F. C., Cell 8, 163-182 (1976). 16. SOUTHERN, E. M., J. Mol. Biol. 98,593-517(1975).
g&265-268
(1979)
Detection
DENISE
of Herpes
A. GALLOWAY,*,‘**
Simplex
RNA in Human
CECILIA FENOGLIO,+ K. MCDOUGALL*,*
Sensory
MARIA
Ganglia
SHEVCHUK,+
AND JAMES *Cold
Spring Harbor Laboratory, P.O. Box 100, Cold Spring Harbor, New York 11724, and ‘Departmnt Pathology, Columbia University, College of Physicians and Surgeons, New York, New York 10052 Accepted
February
of
18,/979
Human paravertebral ganglia were examined for the presence of viral-specific RNA by cytological hybridization using a nick-translated 3H-labeled herpes simplex virus-2 DNA probe. Ganglia from several locations (lumbar, thoracic, etc.) were obtained from seven patients at autopsy. In two cases HSV RNA was detected in all of the ganglia assayed. The autoradiographic grains were localized in the neurons (ganglions cells), and only a fraction of the neurons in any one ganglion were synthesizing viral-specific RNA. These results argue that the HSV genome can be active in the latent state.
in situ
As a consequence of an initial infection, herpes simplex virus (HSV) has been shown to persist in a latent state in the sensory ganglia of mice (1, 2), rabbits (3), monkeys (4), and humans (5, 6) as demonstrated by the recovery of virus. Viral DNA has been detected in association with neurons by in situ hybridization using as probe radioactive complementary RNA made from HSV DNA (7, 8). The state of the viral genome during a latent infection is still an unsettled question. For heuristic reasons it is convenient to consider two hypotheses: either the virus or its genome persists in a nonreplicating state, or it replicates so slowly that infectious virus, or viral gene products, cannot be detected by standard techniques. Several lines of evidence support the nonreplicating model. First, temperature-sensitive mutants are able to establish latent infections, suggesting that viral DNA replication is not necessary for a latent infection (9). Second, extensive searches for infectious virus or other viral specific products detectable by immunofluorescence or ultrastructural analysis, have almost consistently yielded negative results (10); and third, when the presence of viral-specific nucleic acids was ’ To whom reprint requests should be sent. * Present address: Fred Hutchinson Cancer Research Center, 1124 Columbia St., Seattle, Washington 98194. 265
assayed by renaturation kinetics, only 0.11 * 0.03 genome equivalents of viral DNA per cell and no viral-specific RNA was detected in latently infected ganglia (11). These latter approaches may not be sufficiently sensitive to detect viral gene expression in only a few cells of the ganglion. In order to establish whether transcription of the latent HSV genome occurs in individual neurons we turned to in situ cytological hybridization. By using a radioactive DNA probe to detect viral RNA, the target sequences represent an amplification of the persistent viral genome. We have hybridized either iodinated DNA (12) or tritiated nick-translated DNA (13) to a variety of cells and tissues. Viral RNA could be detected in HSV infected BHK cells, in clones of the HSV-2 transformed cell line 333-8-9, in sections of tumors induced in hamsters by the transformed cells, and in HSV transfected-thymidine kinase transformed mouse cells. No autoradiographic grains were seen when hybridized to uninfected or untransformed rodent cells, or to a variety of tissues of rodent or human origin. The hybridization probe was prepared using HSV-2 DNA extracted from infected hamster (BHK) cells, purified by centrifugation in sodium iodide equilibrium density gradients (14) and labeled with tritium in vitro by the process of nick-translation 0042~6822/‘79/070265-04$02.00/O Copyright Q 1979 by Academic All rights of reproduction in
Press,
Inc.
any form reserved.
266
SHORT COMMUNICATIONS
(15). To determine whether the in vitrolabeled DNA represented a faithful copy of the viral genome, a 32P-nick-translated probe (spec. act. 1 x lOa cpm/ug) was prepared and then hybridized to nitrocellulose filters which contained restriction endonuclease produced fragments of HSV-2 DNA transferred from agarose gels (16). The pattern obtained was consistent with all regions of the genome having been copied (data not shown). To determine the specificity of the probe, DNA from a variety of sources (human, monkey, hamster, rat, adenovirus 2, SV40, Escherichia coli; and others) was cleaved with restriction enzymes, fractionated by electrophoresis, and transferred to nitrocellulose. Using a 32P-labeled probe under conditions which detected 5 x lop5 pg of HSV-2 DNA/pg of cell DNA, no hybridization to sequences of human DNA was detected. It is worth noting that viral DNA prepared from BHK cells showed significant hybridization to hamster DNA and likewise that viral DNA isolated from virus grown in BSC-1 cells hybridized to monkey cell DNA. (For details see McDougall et al., manuscript in preparation.) This rigorous characterization of the probe gives us confidence that autoradiographic grains observed with specimens of human origin represent hybridization to HSV-2-specific sequences. At autopsy, within a few hours of death, paravertebral ganglia from several locations (lumbar, thoracic, etc.) were removed from seven patients, sectioned with a cryostat onto glass slides and fixed in -20” ethanol. The sections were hybridized to 3H-labeled HSV-2 DNA and exposed for autoradiography for 4 weeks. A summary of the detection of autoradiographic grains is given in Table 1. In two cases HSV RNA was detected in all of the ganglia assayed. Ganglion tissue is composed primarily of support cells and only a few of the total cells are neurons or ganglion cells. The autoradiographic grains are associated only with the neurons not with any of the surrounding tissue. And, only a small percentage of the neurons (0.04-&O%) are expressing HSV RNA. When ganglia are stained with Giemsa, Nissl bodies which are located in the body of the neuron and
TABLE
1
DETECTIONOF HSV RNA INHUMANSENSORY GANGLIABYINSITU HYBRIDIZATION Case No. 1 1 2 2
3 3 4 4 4 5 5 6 7
Location of ganglia
HSV RNA
Percentage of positive neurons
Sacral Thoracic Sacral Thoracic Sacral Thoracic Sacral Thoracic Lumbar Sacral
+ + + + + -
8 8 0 0 0 0 0.5 0.4 0.4 0
Lumbar Sacral Sacral
-
0 0 0
consist of endoplasmic reticulum and ribosomes appear as dark, granular organelles. One must be careful to distinguish these structures from autoradiographic grains. A high power magnification (Fig. 1A) of a region of the ganglion shows the location of authentic grains over the Nissl bodies indicating that the neurons are elaborating HSV-specific RNA. Figure 1B shows a ganglion in which no autoradiographic grains were detected. The radioactive signal could be abolished by pretreatment of the tissue with alkali indicating that the target sequences are RNA. To further convince ourselves that the observed autoradiographic grains using HSV probes represented true viral RNA, other 3H-labeled DNA probes (SV40, adenovirus 2 and 12, and mouse satellite) prepared in the same manner, were hybridized to various sections of human tissues. These results were negative. Because of the paucity of neurons within a ganglion which express HSV RNA, biochemical analysis of the viral nucleic acids present in ganglia is generally not feasible. Puga et al. (11) acknowledged that under the hybridization conditions they used, as many as 400 mRNA molecules per ganglion could be present and escape detection. In our study, only 0.4 to 8.0% of the neurons were positive for HSV RNA. It is not yet
SHORT COMMUNICATIONS
267
FIG. 1. Localization of HSV RNA in neurons of human sensory ganglia. A sacral ganglion was removed at autopsy, flash frozen, sectioned onto acid-washed glass slides, and placed in -20” ethanol. The hybridization reaction was in a total volume of 10 ~1 containing 6 x SSC and 5 x lo4 of 3H-labeled HSV-2 DNA (spec. act. 1 x 10’ cpm/pg DNA) denatured in alkali and neutralized. A siliconized coverslip was placed over the mixture, the slides were placed in a tray containing Whatman paper saturated with 20 x SSC and incubated at 68” for 18 hr. The slides were washed extensively with 2 x SSC at 4”, dried, dipped in Kodak NTB photographic emulsion, and exposed for autoradiography for 4 weeks at 4”. The emulsion was developed and the cells stained with Giemsa. (A) A high power magnification (800x) shows autoradiographic grains covering the body of a neuron with very low background in surrounding tissue. (B) In a neuron which does not show any hybridization to HSV-2 DNA the Nissl body is seen as a darkly staining granular structure.
clear whether all of the neurons contain HSV viral DNA and only some cells express viral genes, or whether only some neurons are latently infected. Also because of the homology shared by the two genomes, we are unable to discriminate between HSV-land HSV-Z-specific RNA with our present hybridization conditions. None of the patients from whom ganglia were obtained had been undergoing a herpes infection nor suffered from recurrent herpetic infections, therefore, there is no reason to assume that the HSV RNA detected was related to reactivation of the virus. We had not expected that from
one patient, all of the ganglia would be either all positive or all negative. Baringer (6) had reported that virus was often present bilaterally and could be recovered from more than one site. These results might argue persons harboring HSV are susceptible to latent infections in many sites. Our results neither support nor argue against a model in which HSV is retained in a nonreplicating state. The viral RNA detected might represent extensive transcription of the viral genome necessary for viral replication. Alternatively, the RNA might represent a transcript from a limited
268
SHORT
COMMUNICATIONS
region of the genome, with the expression of one or more viral genes a necessary factor for the maintenance of the latent state. In latent infections established by other herpes viruses, e.g., Epstein-Barr virus and the Luck6 virus, viral-specific RNA, and antigens have been detected in association with cells harboring the latent virus (10). We have presented evidence which indicates that some transcription of the herpes simplex virus genome occurs in latently infected human sensory ganglia. Experiments are under way to determine whether the RNA represents a specific and limited set of viral transcripts which are essential for the maintenance of a latent infection. ACKNOWLEDGMENTS We thank Beth Stewart for preparing the HSV-2 DNA, Bob Yaffe for the illustrations, and Tammie Tsujikawa for typing the manuscript. D.A.G. was the recipient of a PHS research fellowship from NCI. C.F. is supported in part by Grant BRSG 10113-39966. This work was funded by a Cancer Center grant to C.S.H. and by NSF Grant PCM77-11835 to J.K.M. REFERENCES 1. STEVENS, 843-845
J. G., and COOK, (1971).
M.
L., Science
173,
2. COOK, J. L., and STEVENS, J. G., J. Gen. Viral. 31, 75-80 (1976). 3. STEVENS, J. G., NESBURN, .A. B., and COOK, M. L., Nature New Biol. 235, 216-217 (1972). L. REEVES, W. C., DIGIACOMO, R. F., ALEXANDER, E. R., and LEE, C. K., Proc. Sot. Exp. Biol. Med. 153, 258-261 (1976). 5. BASTIAN, F. O., RABSON, A. S., YEE, C. L., and TRALKA, T. S., Science 178, 366-307 (1972). 6. BARINGER, J. R., N. Engl. J. Med. 291,828-830 (1974). 7. STEVENS, J. G., and COOK, M. L., J. Immunol. 113, 1685-1693 (1974). 8. ZUR HAUSEN, H., and SCHULTE-HOLTHAUSEN, H.,ZAECSci.Publ. ll(PartI), 117-123(1975). 9. LOFGREN, R. W., STEVENS, J. G., MARSDEN, H. S., and SUBAK-SHARPE, J. H., Virology 76, 440-443 (1977). 10. STEVENS, J. G., Curr. Top. Microbial. Immunol. 40, 31-50 (1975). 11. PUGA, A., ROSENTHAL, J. D., OPENSHAW, H., and NOTKINS, A. L., Virology 89, 102-111 (1978). 12. COPPLE, C. D., and MCDOUGALL, J. K., ht. J. Cancer 17, 501-510(1976). 13. MCDOUGALL, J. K., and GALLOWAY, D. A., in “Persistent Viruses,” Vol. 11 (J. Stevens, G. Todaro, C. F. Fox, eds.) Academic Press, New York, pp. 181-188, 1978. J. M. M., and SCHEGGET, J. T., 14. WALBOOMERS, Virology 74, 256-258 (1976). 15. MANIATIS, T., KEE, S. G., EFSTRADIADIS, A., and KAFATOS, F. C., Cell 8, 163-182 (1976). 16. SOUTHERN, E. M., J. Mol. Biol. 98,593-517(1975).
