- Email: [email protected]
Quantitative analysis of visna virus replication in Vivo
VIROLOGY
141, 148-154 (1985)
Quantitative
Analysis of Visna Virus Replication in Viva
ADAM P. GEBALLE,' PETERVENTURA,LINDA STOWRING,ANDASHLEY T. HAASE' Section of Iqfkctious Diseases, Veterans Administration Medical Center, &.SO Clement Street, San Fmwiaco, California 94121 Received July 26, 1984 accepted October 24, 1984 Visna virus is the prototype of the lentivirus subfamily, a group of nontransforming retroviruses that cause slow infections in sheep and goats. In nature, virus is acquired primarily by the respiratory route and subsequently spreads to several organ systems. These viruses persist for years in their hosts despite a vigorous immune response because of a block in virus gene expression. This report continues the analysis of persistence in viuo, and specifically examines a gene dosage hypothesis that has been advanced as an explanation for the decrease in transcription and virus production in the cells in infected animals. For this analysis a new pulmonary model has been developed that, in conjunction with quantitative in situ hybridization, provides an opportunity to examine in animals the molecular events that occur in the course of the viral life cycle. We establish the feasibility of such a longitudinal analysis in vivo, document restriction in gene expression in alveolar macrophages and provide evidence that this restriction cannot be accounted for simply by gene dosage. The approach illustrated with visna should be of general applicability to other dynamic and molecular investigations of VirUS infeCtion. 0 1985 Academic press, Inc.
For the past several years, we have been investigating the life cycles of viruses in viva, using a sensitive and quantitative method of in situ hybridization to define biochemical parameters of replication in tissues (1). This analysis has been largely focused on the role of viruses in chronic diseases and the pathogenesis of slow virus infections, and thus far has been a static one based on observations at a single point in time. In this article we establish the feasibility of a dynamic analysis of virus replication in viva in studies of the synthesis and transcription of the visna virus genome in pulmonary alveolar macrophages (PAMS). The particular issue that persuaded us to develop an animal model amenable to sequential analysis is the question of how visna virus gene expression is regulated * Current address: Department of Medical Microbiology, Stanford University, Stanford, Calif. 94305. * Current address: Department of Microbiology, University of Minnesota, Minneapolis, Minn. 55455. To whom correspondence and requests for reprints should be addressed. 0042-6822/85 $3.00 Copyright All rights
0 1985 by Academic Press, Inc. of reproduction in any form reserved.
in vivo. In tissue culture, visna virus replicates productively and lytically in a matter of days. By contrast, in the infected animal virus titers are low as a consequence of a block in gene expression imposed at the level of transcription (2). Two explanations have been proposed for the decrease in viral RNA, one analogous to lysogeny, and the other a gene dosage model that in its most narrow construction attributes transcriptional regulation to the concentration of extrachromosomal DNA in the cell (3). The gene dosage hypothesis is supported thus far by observations of the life cycle of visna virus in tissue culture where most, if not all, viral DNA molecules are extrachromosomal (4) and the extent of early DNA synthesis governs transcription and virus production (5). To see if a similar correlation obtains in infected sheep between initial concentration of DNA and subsequent level of transcription, we needed to be able to measure the concentration of viral DNA in cells at one point in time, and RNA content at a later time. To conduct studies 148
SHORT
COMMUNICATIONS
at more than one time point in a single animal, we developed a model in which visna virus is introduced into one lobe of the lung of a sheep and pulmonary alveolar macrophages (PAMS) are collected by lavage at frequent intervals. Although only a fraction of the population of PAMS is infected, we can assess the synthesis and transcription of the visna genome in this population by in situ hybridization. In this report, we show that molecular studies of this kind can be done in an intact animal; provide evidence that the transcription of the visna genome is restricted in PAMS just as it is in the central nervous system; and demonstrate that the relationship between genomic synthesis and transcription does not conform to a simple gene dosage model. Visna virus was introduced into the lung as follows: After sedation with ketamine, 3- to B-month-old American lambs were intubated with a g-mm endotracheal tube (Mallinckrodt). A fiberoptic bronchoscope (Pentax FB19A) was passed through the endotracheal tube and wedged into the cranial bronchus. On Day 0, visna virus was injected into the cranial lobe through the lumen of the bronchoscope. In various experiments 2 X lOa to 1 X 10’ PFU of virus in 2 to 20 ml of phosphatebuffered saline or maintenance medium were used. To ensure delivery of the inoculum, two rinses of 20 ml or normal saline followed by 20 cc of air were injected through the same lumen used to inject the virus. On subsequent days, PAMS were obtained by again inserting the bronchoscope into the cranial bronchus, injecting 20 ml of normal saline followed by 20 cc of air, and immediately aspirating the bronchial lavage containing PAMS. After two or three lavages, PAMS were collected by centrifugation (500 Q, 5 min at room temperature) and washed once with phosphate-buffered saline. Aliquots were assayed for infectious centers and cytocentrifuged onto treated glass slides for analysis by in situ hybridization. In situ hybridization was carried out as described (I). Briefly, after fixation in
149
ethanol/acetic acid and ethanol alone, slides were pretreated with acid, heat, and proteinase K. For RNA detection, slides were hybridized for 3 days with 0.7 ng of probe in 5 ~1 of a hybridization medium containing 50% formamide, 10% dextran sulfate, and 0.6 M NaCl in buffer with 1X Denhardt solution. Excess probe was washed off and slides were coated with nuclear track emulsion. After appropriate exposure times, the emulsion was developed and the cells were stained with Giemsa. For DNA detection, RNA was digested with RNAse, and DNA was postfixed with paraformaldehyde and denatured. Slides were dehydrated and hybridized with 0.3 ng of probe in 5 ~1 for 3 days. Washing and autoradiography were carried out as described for RNA. In these experiments, either l%I-labeled nicktranslated probe of cloned visna DNA (sp act = 9 X lo* dpm/pg) or 3H-labeled probe reverse transcribed from purified viral RNA with random priming (sp act = 4.8 X lo* dpm/pg) was used. In both cases, sequences from the entire genome were represented in the probe. The rationale and procedures to quantitate the number of copies of viral nucleic acid per cell also followed previously established protocols (I). (1) The infected cell population was defined by first determining the background level of binding of probe to uninfected PAMS. From the mean number of silver grains over 100 randomly selected PAMS obtained prior to infection, and the Poisson distribution, the number of grains was calculated that would occur by chance in 1% or fewer cells in a population of uninfected cells. Any cell with this number or more grains was considered infected. The mean number of grains per cell in the infected population was then determined. Grain counts were converted to copy numbers using relationships previously established in SCP cultures infected in vitro. The percentage of infected cells was determined by counting the total number of cells in all microscopic fields required to accumulate 300 or more infected cells.
150
SHORT COMMUNICATIONS
The main objective of these studies was to develop an animal model in which virus gene expression is restricted, and in which restriction could be analyzed sequentially. Pulmonary infection with visna virus fully
satisfied these requirements. Macrophages could be recovered for weeks from the same animal, and contained sufficient viral DNA and RNA for detection by in situ hybridization (Figs. lA, B) even when the
FIG. 1. Detection of visna DNA and RNA in PAMS by in situ hybridization. (A) Viral DNA in a pulmonary macrophage (\) 1 day after injection with l@ PFU of visna virus. Hybridization as described in the text, ‘H-radiolabeled probe, ‘I-day exposure; original magnification X500. (B) Viral RNA in PAMS (-). Hybridization as in (A). Five-day exposure.
SHORT
COMMUNICATIONS
infected cells were only a minority species in the population. Visna virus replication is restricted in PAMS (6) just as it is in choroid plexus (2). In one animal studied for several weeks (Fig. 2), the peak DNA copy number in PAMS on Day 1 was 9 copies per cell, and the peak RNA copy number on Day 3 was 80 copies per cell. These levels represent an order of magnitude decrease in synthesis compared to permissive infections in vitro (7’), and a somewhat greater restriction than we documented previously in choroid plexus (2). Virus production was even more limited than synthesis of viral DNA and RNA; only about 1 in 50 cells with viral genomes and transcripts produced enough virus for detection by infectious center assay (Fig. 2C). The maximal percentage of PAMS that scored as infectious centers (0.3%) also occurred in the first growth cycle and subsequently declined to 10m3to 10m4in the population. We detected neutralizing antibody to visna virus in low titer (1:lO) in plasma in the third to fourth week of infection (Fig. 2C). Although we did not detect antibody in lavage fluid at this time, it may have been present as the small amount of cell-free virus in the lavage fluid declined coincident with the appearance of neutralizing antibody in plasma (data not shown). The restriction in viral nucleic acid synthesis, however, was evident in the first growth cycle (Fig. 2A), 3 weeks prior to the humoral response in the animal, and before the humoral and cellular immune response in infected animals reported by others (6, 8, 9). These results add further support to the notion that restriction reflects some intrinsic aspect of the interaction of virus and host cell, rather than limitations in virus growth imposed externally by the immune response (6, 10). The availability on a recurring basis of cells with reduced levels of viral RNA encouraged us to examine the gene dosage hypothesis in vivo. Moreover, the results obtained in sheep A were not inconsistent with gene dosage: after the first growth cycle (Fig. 2, Days l-3), there were secondary cycles (Fig. 2, Days 16, 1’7) in
151
which the RNA copy had decreased to barely detectable levels at a time when the level of viral DNA in the population also had fallen dramatically. The major prediction of the gene dosage hypothesis is that there should be a good correlation between the number of copies of viral DNA and the number of transcripts that accumulate in the succeeding 48 hr. We sought two kinds of evidence to test this prediction experimentally: (1) We administered lower and higher doses of virus in an attempt to vary the number of virus particles per cell. At higher particle-to-cell ratios, more viral genomes should enter the cells to give higher DNA copy numbers, and, according to gene dosage, higher levels of RNA. (2) At the higher dose of virus, we hoped to find a logically decisive refutation of gene dosage, e.g., high levels of DNA and low levels of RNA, or low levels of DNA and high RNA copy numbers. Both experimental approaches provided data that are not easily reconciled with simple gene dosage effects. We inoculated five additional sheep with 2 X lo8 PFU to 1 X 10’ PFU of virus, and measured (1) the number of copies of DNA per cell on Day 1 reverse transcribed from input RNA, and (2) the maximum number of copies of RNA produced from these templates in 48-72 hr. Although we were unable to manipulate the DNA copy number per cell as precisely and extensively as we hoped, we were able to vary it over about a threefold range (Table 1). However, there was little correlation between differences in DNA content and the amount of viral RNA transcribed in the cell. Most significantly, one exceptional animal (sheep 2, Table 1) had nearly permissive levels of viral RNA in PAMS on Day 3, from DNA concentrations comparable to those that produce lo-fold less RNA in other animals. The greater output of RNA by PAMS in sheep 2 was not due to superinfection, to increase DNA content over Day 1 levels. In the frequency distributions of DNA copy number for the animal (Fig. 3), there was a decline in the number of copies of DNA in the infected
SHORT
152
COMMUNICATIONS 80
C 60
I
o----o DNA Copies m Infected Ceils
40
-
RNA Copies I” Infected Cells
t 20
-
r c,o
1,”“I
o---o
t 0
15
p:$EJ~~~~=~“-:
!
10
Percent Cells Contammg Vlral RNA
Centers 0.1 - i
:... b.C.&
0 0
I 4
I 8
‘..
.. . 1-1 12
. . . . . . . . . .. . . . . . . . . . . . . . 16
20
24
28
Day 01lnlectlon FIG. 2. Sequential analysis of visna replication in viva in PAMS. Visna (10’ PFU) were inoculated into the cranial lobe of sheep A and PAMS were obtained as described in the text. The number of viral genomes in infected cells and the percentage of all cells which contain viral nucleic acid are shown in the panels A and B. Viral DNA was detected with the ‘%I virus-specific probe; RNA was detected with a ‘H probe. Various autoradiographic exposure times were selected to produce numbers of grains which could be accurately counted. The percentage of PAMS which were infected as measured by an infectious center assay are shown in panel C (A - - - A). For this assay, lo6 to 10’ PAMS were suspended in 24 ml of L15 medium with 5% heat-inactivated lamb serum. Low gel temperature (LGT) agarose (Bio-Rad) was melted, cooled to 37”. and added to make a 0.5% solution. The PAMS in LGT agarose were then poured onto four lOO-mm tissue culture dishes (Falcon) containing confluent SCP cells. In the ensuing 90-min incubation at 37’ the PAMS settled onto the SCP indicator cells. The plates were carefully removed from the incubator, and the low gel temperature agarose was allowed to solidify at room temperature for 30 min. To stabilize the LGT agarose layer, 10 ml of L15 plaquing medium with 1% agarose was added to each plate, and solidified at room temperature for 20 min. The plates were then incubated at 37“ for 12 days. Plaques produced by infected PAMS were counted after the SCP cell indicator monolayer had been fixed in formalin and stained with crystal violet (1.2). The arrow in panel C indicates the time at which neutralizing antibody was first detected at a 1:lO dilution.
cohort of cells as RNA copy numbers increased. These studies demonstrate the feasibility of sequential biochemical investigations of a virus life cycle in animals. With sensitive and quantitative methods of in situ hybridization, we were able to ex-
amine viral nucleic acid synthesis in individual cells obtained in the course of the virus replicative cycle. We believe such a system has a special relevance unattainable in cell culture, where one cannot precisely mimic the natural milieu or fully maintain the differentiated state
SHORT
COMMUNICATIONS TABLE
153
1
RELATIONSHIP OF DNA AND RNA COPY NUMBER IN PAMS INFECTED WITH VISNA VIRUS
Inoculum (PFU)
Sheep A 1 2 3 4 5
Number of copies of Viral DNA per cell (Day 1)
Peak number of copies of viral RNA per cell (Day 3 or 4)
9 21 14 6 7 8
80 72 900 32 15 35
lo9 109 109 lo9 2 x lo* 2 x lo*
of the cell. This pulmonary model of visna also provides closer parallels to natural infections in Iceland, where most if not all animals acquired virus by the respiratory route, eventually became dyspneic, and died as a result of an interstitial pneumonitis called maedi. About 10% of the animals with disseminated disease
developed a paralytic condition called visna, but the same virus appears to be responsible for both the pulmonary and neurological disorders (11). These investigations continue the analysis of the restriction in visna gene expression in sheep that in our view provides the best explanation for the persis-
02
r-
DNA
RNA I Day 1
Day 3
r I
Day 3
0 1
5
10
Copy
Number
50
10010
100
1000 Copy
10.000
Number
FIG. 3. Frequency distribution of viral RNA and DNA copy number in sheep 2 (Table l), animal with high RNA copy number, low DNA copy number, Days 1, 3. Copy numbers determined in 50 cells as described in the text.
154
SHORT
COMMUNICATIONS
tence of the virus and slow evolution of infection. Because restriction is imposed at the level of transcription, and, in infected tissue culture cells, transcription is governed by the extent of early DNA synthesis (5), the gene dosage hypothesis was the first that we tested in the pulmonary model. This hypothesis was also attractive a priori, since in natural infections we expect inocula to be small. PAMS and other cells therefore would be infected with a few particles at most, and only a few copies of the viral RNA genome would be available for reverse transcription. According to the gene dosage hypothesis, this would provide insufficient concentrations of DNA to support high levels of transcription and virus production. However, we found little consistent change in RNA copy number with three-fold variations in viral DNA concentrations; and, in one animal, there were high RNA copy numbers in cells that contained DNA concentrations equivalent to those in animals with lower levels of viral RNA. These results do not support a simple gene dosage hypothesis. We currently are evaluating in this pulmonary model other explanations for restricted gene expression, particularly the relationship of viral DNA structure and integration to RNA synthesis.
REFERENCES I. HAASE, A., BRAHIC, M., STOWRING, L., and BLUM, H., In “Methods in Virology,” Vol. VII, pp. 189-226. Academic Press, New York, 1984. .s. BRAHIC, M., STOWRING, L., VENTURA, P., and HAASE, A. T., Nature (London) 292, 240-242 (1981). 5. HAASE, A. T., Curr. Top. Microbid 12, 101-156 (1975). 4. HARRIS, J. D., BLUM, H., SCOTT, J., TRAYNOR, B., VENTURA, P., and HAASE, A. T., Proc. Natl. Aead Sci USA, 81,7212-7215 (1984). 5. HAASE, A. T., STOWRING, L., HARRIS, J. D., TRAYNOR, B., VENTURA, P., PELUSO, R., and BRAHIC, M., Virdogy 119,399-410 (1982). 6. NARAYAN, O., GRIFFIN, D. E., and SILVERSTEIN, A. M., J. IMect. Dis 135,800~806 (1977). r. BRAHIC, M., FILIPPI, P., VIGNE, R., and HAASE, A. T., .I. Viro~ 24, 74-81 (1977). 8. PETURSSON, G., NATHANSON, N., GEORGSSON, G., PANITCH, H., and PALSSON, P., Lab. Investigat. 35,402-412 (1976). 9. GRIFFIN, D. E., NARAYAN, O., and ADAMS, R. J., J. Ir&d Dk. 138,340~350 (1978). 10. NATHANSON, N., PANITCH, H., PALSSON, Pw, P., PETURSSON, G., and GEORGSSON, G., Lab. Inwstigat. 35,444-451 (1976). fz. BRAHIC, M., and HAASE, A. T., In “Comparative Diagnosis of Viral Diseases (E. Kurstak and C. Kurstak, eds.), pp. 619-643. Academic Press, New York, 1981. 12. SCOTT, J. V., STOWRING, L., and HAASE, A. T., J. Vird 24. 74-81 (1977).
141, 148-154 (1985)
Quantitative
Analysis of Visna Virus Replication in Viva
ADAM P. GEBALLE,' PETERVENTURA,LINDA STOWRING,ANDASHLEY T. HAASE' Section of Iqfkctious Diseases, Veterans Administration Medical Center, &.SO Clement Street, San Fmwiaco, California 94121 Received July 26, 1984 accepted October 24, 1984 Visna virus is the prototype of the lentivirus subfamily, a group of nontransforming retroviruses that cause slow infections in sheep and goats. In nature, virus is acquired primarily by the respiratory route and subsequently spreads to several organ systems. These viruses persist for years in their hosts despite a vigorous immune response because of a block in virus gene expression. This report continues the analysis of persistence in viuo, and specifically examines a gene dosage hypothesis that has been advanced as an explanation for the decrease in transcription and virus production in the cells in infected animals. For this analysis a new pulmonary model has been developed that, in conjunction with quantitative in situ hybridization, provides an opportunity to examine in animals the molecular events that occur in the course of the viral life cycle. We establish the feasibility of such a longitudinal analysis in vivo, document restriction in gene expression in alveolar macrophages and provide evidence that this restriction cannot be accounted for simply by gene dosage. The approach illustrated with visna should be of general applicability to other dynamic and molecular investigations of VirUS infeCtion. 0 1985 Academic press, Inc.
For the past several years, we have been investigating the life cycles of viruses in viva, using a sensitive and quantitative method of in situ hybridization to define biochemical parameters of replication in tissues (1). This analysis has been largely focused on the role of viruses in chronic diseases and the pathogenesis of slow virus infections, and thus far has been a static one based on observations at a single point in time. In this article we establish the feasibility of a dynamic analysis of virus replication in viva in studies of the synthesis and transcription of the visna virus genome in pulmonary alveolar macrophages (PAMS). The particular issue that persuaded us to develop an animal model amenable to sequential analysis is the question of how visna virus gene expression is regulated * Current address: Department of Medical Microbiology, Stanford University, Stanford, Calif. 94305. * Current address: Department of Microbiology, University of Minnesota, Minneapolis, Minn. 55455. To whom correspondence and requests for reprints should be addressed. 0042-6822/85 $3.00 Copyright All rights
0 1985 by Academic Press, Inc. of reproduction in any form reserved.
in vivo. In tissue culture, visna virus replicates productively and lytically in a matter of days. By contrast, in the infected animal virus titers are low as a consequence of a block in gene expression imposed at the level of transcription (2). Two explanations have been proposed for the decrease in viral RNA, one analogous to lysogeny, and the other a gene dosage model that in its most narrow construction attributes transcriptional regulation to the concentration of extrachromosomal DNA in the cell (3). The gene dosage hypothesis is supported thus far by observations of the life cycle of visna virus in tissue culture where most, if not all, viral DNA molecules are extrachromosomal (4) and the extent of early DNA synthesis governs transcription and virus production (5). To see if a similar correlation obtains in infected sheep between initial concentration of DNA and subsequent level of transcription, we needed to be able to measure the concentration of viral DNA in cells at one point in time, and RNA content at a later time. To conduct studies 148
SHORT
COMMUNICATIONS
at more than one time point in a single animal, we developed a model in which visna virus is introduced into one lobe of the lung of a sheep and pulmonary alveolar macrophages (PAMS) are collected by lavage at frequent intervals. Although only a fraction of the population of PAMS is infected, we can assess the synthesis and transcription of the visna genome in this population by in situ hybridization. In this report, we show that molecular studies of this kind can be done in an intact animal; provide evidence that the transcription of the visna genome is restricted in PAMS just as it is in the central nervous system; and demonstrate that the relationship between genomic synthesis and transcription does not conform to a simple gene dosage model. Visna virus was introduced into the lung as follows: After sedation with ketamine, 3- to B-month-old American lambs were intubated with a g-mm endotracheal tube (Mallinckrodt). A fiberoptic bronchoscope (Pentax FB19A) was passed through the endotracheal tube and wedged into the cranial bronchus. On Day 0, visna virus was injected into the cranial lobe through the lumen of the bronchoscope. In various experiments 2 X lOa to 1 X 10’ PFU of virus in 2 to 20 ml of phosphatebuffered saline or maintenance medium were used. To ensure delivery of the inoculum, two rinses of 20 ml or normal saline followed by 20 cc of air were injected through the same lumen used to inject the virus. On subsequent days, PAMS were obtained by again inserting the bronchoscope into the cranial bronchus, injecting 20 ml of normal saline followed by 20 cc of air, and immediately aspirating the bronchial lavage containing PAMS. After two or three lavages, PAMS were collected by centrifugation (500 Q, 5 min at room temperature) and washed once with phosphate-buffered saline. Aliquots were assayed for infectious centers and cytocentrifuged onto treated glass slides for analysis by in situ hybridization. In situ hybridization was carried out as described (I). Briefly, after fixation in
149
ethanol/acetic acid and ethanol alone, slides were pretreated with acid, heat, and proteinase K. For RNA detection, slides were hybridized for 3 days with 0.7 ng of probe in 5 ~1 of a hybridization medium containing 50% formamide, 10% dextran sulfate, and 0.6 M NaCl in buffer with 1X Denhardt solution. Excess probe was washed off and slides were coated with nuclear track emulsion. After appropriate exposure times, the emulsion was developed and the cells were stained with Giemsa. For DNA detection, RNA was digested with RNAse, and DNA was postfixed with paraformaldehyde and denatured. Slides were dehydrated and hybridized with 0.3 ng of probe in 5 ~1 for 3 days. Washing and autoradiography were carried out as described for RNA. In these experiments, either l%I-labeled nicktranslated probe of cloned visna DNA (sp act = 9 X lo* dpm/pg) or 3H-labeled probe reverse transcribed from purified viral RNA with random priming (sp act = 4.8 X lo* dpm/pg) was used. In both cases, sequences from the entire genome were represented in the probe. The rationale and procedures to quantitate the number of copies of viral nucleic acid per cell also followed previously established protocols (I). (1) The infected cell population was defined by first determining the background level of binding of probe to uninfected PAMS. From the mean number of silver grains over 100 randomly selected PAMS obtained prior to infection, and the Poisson distribution, the number of grains was calculated that would occur by chance in 1% or fewer cells in a population of uninfected cells. Any cell with this number or more grains was considered infected. The mean number of grains per cell in the infected population was then determined. Grain counts were converted to copy numbers using relationships previously established in SCP cultures infected in vitro. The percentage of infected cells was determined by counting the total number of cells in all microscopic fields required to accumulate 300 or more infected cells.
150
SHORT COMMUNICATIONS
The main objective of these studies was to develop an animal model in which virus gene expression is restricted, and in which restriction could be analyzed sequentially. Pulmonary infection with visna virus fully
satisfied these requirements. Macrophages could be recovered for weeks from the same animal, and contained sufficient viral DNA and RNA for detection by in situ hybridization (Figs. lA, B) even when the
FIG. 1. Detection of visna DNA and RNA in PAMS by in situ hybridization. (A) Viral DNA in a pulmonary macrophage (\) 1 day after injection with l@ PFU of visna virus. Hybridization as described in the text, ‘H-radiolabeled probe, ‘I-day exposure; original magnification X500. (B) Viral RNA in PAMS (-). Hybridization as in (A). Five-day exposure.
SHORT
COMMUNICATIONS
infected cells were only a minority species in the population. Visna virus replication is restricted in PAMS (6) just as it is in choroid plexus (2). In one animal studied for several weeks (Fig. 2), the peak DNA copy number in PAMS on Day 1 was 9 copies per cell, and the peak RNA copy number on Day 3 was 80 copies per cell. These levels represent an order of magnitude decrease in synthesis compared to permissive infections in vitro (7’), and a somewhat greater restriction than we documented previously in choroid plexus (2). Virus production was even more limited than synthesis of viral DNA and RNA; only about 1 in 50 cells with viral genomes and transcripts produced enough virus for detection by infectious center assay (Fig. 2C). The maximal percentage of PAMS that scored as infectious centers (0.3%) also occurred in the first growth cycle and subsequently declined to 10m3to 10m4in the population. We detected neutralizing antibody to visna virus in low titer (1:lO) in plasma in the third to fourth week of infection (Fig. 2C). Although we did not detect antibody in lavage fluid at this time, it may have been present as the small amount of cell-free virus in the lavage fluid declined coincident with the appearance of neutralizing antibody in plasma (data not shown). The restriction in viral nucleic acid synthesis, however, was evident in the first growth cycle (Fig. 2A), 3 weeks prior to the humoral response in the animal, and before the humoral and cellular immune response in infected animals reported by others (6, 8, 9). These results add further support to the notion that restriction reflects some intrinsic aspect of the interaction of virus and host cell, rather than limitations in virus growth imposed externally by the immune response (6, 10). The availability on a recurring basis of cells with reduced levels of viral RNA encouraged us to examine the gene dosage hypothesis in vivo. Moreover, the results obtained in sheep A were not inconsistent with gene dosage: after the first growth cycle (Fig. 2, Days l-3), there were secondary cycles (Fig. 2, Days 16, 1’7) in
151
which the RNA copy had decreased to barely detectable levels at a time when the level of viral DNA in the population also had fallen dramatically. The major prediction of the gene dosage hypothesis is that there should be a good correlation between the number of copies of viral DNA and the number of transcripts that accumulate in the succeeding 48 hr. We sought two kinds of evidence to test this prediction experimentally: (1) We administered lower and higher doses of virus in an attempt to vary the number of virus particles per cell. At higher particle-to-cell ratios, more viral genomes should enter the cells to give higher DNA copy numbers, and, according to gene dosage, higher levels of RNA. (2) At the higher dose of virus, we hoped to find a logically decisive refutation of gene dosage, e.g., high levels of DNA and low levels of RNA, or low levels of DNA and high RNA copy numbers. Both experimental approaches provided data that are not easily reconciled with simple gene dosage effects. We inoculated five additional sheep with 2 X lo8 PFU to 1 X 10’ PFU of virus, and measured (1) the number of copies of DNA per cell on Day 1 reverse transcribed from input RNA, and (2) the maximum number of copies of RNA produced from these templates in 48-72 hr. Although we were unable to manipulate the DNA copy number per cell as precisely and extensively as we hoped, we were able to vary it over about a threefold range (Table 1). However, there was little correlation between differences in DNA content and the amount of viral RNA transcribed in the cell. Most significantly, one exceptional animal (sheep 2, Table 1) had nearly permissive levels of viral RNA in PAMS on Day 3, from DNA concentrations comparable to those that produce lo-fold less RNA in other animals. The greater output of RNA by PAMS in sheep 2 was not due to superinfection, to increase DNA content over Day 1 levels. In the frequency distributions of DNA copy number for the animal (Fig. 3), there was a decline in the number of copies of DNA in the infected
SHORT
152
COMMUNICATIONS 80
C 60
I
o----o DNA Copies m Infected Ceils
40
-
RNA Copies I” Infected Cells
t 20
-
r c,o
1,”“I
o---o
t 0
15
p:$EJ~~~~=~“-:
!
10
Percent Cells Contammg Vlral RNA
Centers 0.1 - i
:... b.C.&
0 0
I 4
I 8
‘..
.. . 1-1 12
. . . . . . . . . .. . . . . . . . . . . . . . 16
20
24
28
Day 01lnlectlon FIG. 2. Sequential analysis of visna replication in viva in PAMS. Visna (10’ PFU) were inoculated into the cranial lobe of sheep A and PAMS were obtained as described in the text. The number of viral genomes in infected cells and the percentage of all cells which contain viral nucleic acid are shown in the panels A and B. Viral DNA was detected with the ‘%I virus-specific probe; RNA was detected with a ‘H probe. Various autoradiographic exposure times were selected to produce numbers of grains which could be accurately counted. The percentage of PAMS which were infected as measured by an infectious center assay are shown in panel C (A - - - A). For this assay, lo6 to 10’ PAMS were suspended in 24 ml of L15 medium with 5% heat-inactivated lamb serum. Low gel temperature (LGT) agarose (Bio-Rad) was melted, cooled to 37”. and added to make a 0.5% solution. The PAMS in LGT agarose were then poured onto four lOO-mm tissue culture dishes (Falcon) containing confluent SCP cells. In the ensuing 90-min incubation at 37’ the PAMS settled onto the SCP indicator cells. The plates were carefully removed from the incubator, and the low gel temperature agarose was allowed to solidify at room temperature for 30 min. To stabilize the LGT agarose layer, 10 ml of L15 plaquing medium with 1% agarose was added to each plate, and solidified at room temperature for 20 min. The plates were then incubated at 37“ for 12 days. Plaques produced by infected PAMS were counted after the SCP cell indicator monolayer had been fixed in formalin and stained with crystal violet (1.2). The arrow in panel C indicates the time at which neutralizing antibody was first detected at a 1:lO dilution.
cohort of cells as RNA copy numbers increased. These studies demonstrate the feasibility of sequential biochemical investigations of a virus life cycle in animals. With sensitive and quantitative methods of in situ hybridization, we were able to ex-
amine viral nucleic acid synthesis in individual cells obtained in the course of the virus replicative cycle. We believe such a system has a special relevance unattainable in cell culture, where one cannot precisely mimic the natural milieu or fully maintain the differentiated state
SHORT
COMMUNICATIONS TABLE
153
1
RELATIONSHIP OF DNA AND RNA COPY NUMBER IN PAMS INFECTED WITH VISNA VIRUS
Inoculum (PFU)
Sheep A 1 2 3 4 5
Number of copies of Viral DNA per cell (Day 1)
Peak number of copies of viral RNA per cell (Day 3 or 4)
9 21 14 6 7 8
80 72 900 32 15 35
lo9 109 109 lo9 2 x lo* 2 x lo*
of the cell. This pulmonary model of visna also provides closer parallels to natural infections in Iceland, where most if not all animals acquired virus by the respiratory route, eventually became dyspneic, and died as a result of an interstitial pneumonitis called maedi. About 10% of the animals with disseminated disease
developed a paralytic condition called visna, but the same virus appears to be responsible for both the pulmonary and neurological disorders (11). These investigations continue the analysis of the restriction in visna gene expression in sheep that in our view provides the best explanation for the persis-
02
r-
DNA
RNA I Day 1
Day 3
r I
Day 3
0 1
5
10
Copy
Number
50
10010
100
1000 Copy
10.000
Number
FIG. 3. Frequency distribution of viral RNA and DNA copy number in sheep 2 (Table l), animal with high RNA copy number, low DNA copy number, Days 1, 3. Copy numbers determined in 50 cells as described in the text.
154
SHORT
COMMUNICATIONS
tence of the virus and slow evolution of infection. Because restriction is imposed at the level of transcription, and, in infected tissue culture cells, transcription is governed by the extent of early DNA synthesis (5), the gene dosage hypothesis was the first that we tested in the pulmonary model. This hypothesis was also attractive a priori, since in natural infections we expect inocula to be small. PAMS and other cells therefore would be infected with a few particles at most, and only a few copies of the viral RNA genome would be available for reverse transcription. According to the gene dosage hypothesis, this would provide insufficient concentrations of DNA to support high levels of transcription and virus production. However, we found little consistent change in RNA copy number with three-fold variations in viral DNA concentrations; and, in one animal, there were high RNA copy numbers in cells that contained DNA concentrations equivalent to those in animals with lower levels of viral RNA. These results do not support a simple gene dosage hypothesis. We currently are evaluating in this pulmonary model other explanations for restricted gene expression, particularly the relationship of viral DNA structure and integration to RNA synthesis.
REFERENCES I. HAASE, A., BRAHIC, M., STOWRING, L., and BLUM, H., In “Methods in Virology,” Vol. VII, pp. 189-226. Academic Press, New York, 1984. .s. BRAHIC, M., STOWRING, L., VENTURA, P., and HAASE, A. T., Nature (London) 292, 240-242 (1981). 5. HAASE, A. T., Curr. Top. Microbid 12, 101-156 (1975). 4. HARRIS, J. D., BLUM, H., SCOTT, J., TRAYNOR, B., VENTURA, P., and HAASE, A. T., Proc. Natl. Aead Sci USA, 81,7212-7215 (1984). 5. HAASE, A. T., STOWRING, L., HARRIS, J. D., TRAYNOR, B., VENTURA, P., PELUSO, R., and BRAHIC, M., Virdogy 119,399-410 (1982). 6. NARAYAN, O., GRIFFIN, D. E., and SILVERSTEIN, A. M., J. IMect. Dis 135,800~806 (1977). r. BRAHIC, M., FILIPPI, P., VIGNE, R., and HAASE, A. T., .I. Viro~ 24, 74-81 (1977). 8. PETURSSON, G., NATHANSON, N., GEORGSSON, G., PANITCH, H., and PALSSON, P., Lab. Investigat. 35,402-412 (1976). 9. GRIFFIN, D. E., NARAYAN, O., and ADAMS, R. J., J. Ir&d Dk. 138,340~350 (1978). 10. NATHANSON, N., PANITCH, H., PALSSON, Pw, P., PETURSSON, G., and GEORGSSON, G., Lab. Inwstigat. 35,444-451 (1976). fz. BRAHIC, M., and HAASE, A. T., In “Comparative Diagnosis of Viral Diseases (E. Kurstak and C. Kurstak, eds.), pp. 619-643. Academic Press, New York, 1981. 12. SCOTT, J. V., STOWRING, L., and HAASE, A. T., J. Vird 24. 74-81 (1977).