Isolation and Characterization of a Neuropathogenic Simian Immunodeficiency Virus Derived from a Sooty Mangabey

Abstract

Transfusion of blood from a simian immunodeficiency virus (SIV)- and simian T-cell lymphotropic virus-infected sooty mangabey (designated FGb) to rhesus and pig-tailed macaques resulted in the development of neurologic disease in addition to AIDS. To investigate the role of SIV in neurologic disease, virus was isolated from a lymph node of a pig-tailed macaque (designated PGm) and the cerebrospinal fluid of a rhesus macaque (designated ROn2) and passaged to additional macaques. SIV-related neuropathogenic effects were observed in 100% of the pig-tailed macaques inoculated with either virus. Lesions in these animals included extensive formation of SIV RNA-positive giant cells in the brain parenchyma and meninges. Based upon morphology, the majority of infected cells in both lymphoid and brain tissue appeared to be of macrophage lineage. The virus isolates replicated very well in pig-tailed and rhesus macaque peripheral blood mononuclear cells (PBMC) with rapid kinetics. Differential replicative abilities were observed in both PBMC and macrophage populations, with viruses growing to higher titers in pig-tailed macaque cells than in rhesus macaque cells. An infectious molecular clone of virus derived from the isolate from macaque PGm (PGm5.3) was generated and was shown to have in vitro replication characteristics similar to those of the uncloned virus stock. While molecular analyses of this virus revealed its similarity to SIV isolates from sooty mangabeys, significant amino acid differences in Env and Nef were observed. This virus should provide an excellent system for investigating the mechanism of lentivirus-induced neurologic disease.

While the induction of immunosuppression and AIDS are the major pathogenic effects associated with human immunodeficiency virus type 1 (HIV-1) infection in humans, other, clinically significant, debilitating sequelae are often observed. These include hematologic abnormalities (5, 17), gastrointestinal disease (24, 53), and, perhaps most significantly, neurologic disease (10, 15). Manifestations of AIDS-related neurologic dysfunction include peripheral neuropathy and myelopathy (9, 14), but most often patients present with a condition known as HIV-associated dementia. HIV-associated dementia, characterized by motor slowness, slowness of cognitive functioning, and disturbances in memory and language, is not caused by secondary opportunistic infections but rather results directly or indirectly from the presence of HIV in the nervous system (45, 54). Although studies of human cognition have clearly produced much information on the neurological sequelae of AIDS, it has been argued that many factors which correlate with HIV infection (e.g., substance abuse) may in fact be responsible for some behavioral manifestations (31). For this reason, an animal model of HIV-associated neurologic disease is extremely useful in permitting the isolation of the specific role of viral infection in the etiology of dementia.

HIV-1 apparently enters the brain soon after infection, as evidenced by isolation of virus from cerebrospinal fluid (CSF) during the acute phase (4, 16, 20, 47). The mechanism of HIV entry into the brain has not been fully elucidated, but a number of hypotheses exist and HIV may gain entry by any or all of these. Probably the most widely accepted hypothesis of the means by which HIV gains entry into the brain is through infected monocytes (42). However, an alternative mechanism, for which there is reasonable data, is the direct infection of microvascular endothelial cells (36, 44), which may either destroy the integrity of the blood-brain barrier or pass on infection to migrating lymphocytes or monocytes (or both).

Simian immunodeficiency virus (SIV) is a lentivirus which is closely related to HIV. SIV isolates from macaques (which are given the prefix SIVmac) and from sooty mangabeys (which are given the prefix SIVsmm) induce a disease in Asian macaques that is remarkably similar to AIDS in humans (28, 34, 37). This disease process is characterized by a decline in CD4⁺ cell counts, development of immunosuppression, and opportunistic infections, ultimately leading to death.

Seminal pathogenesis studies have shown that SIV, like HIV, is both neuroinvasive and neurovirulent (6, 28). Several SIV interactions with the brain parallel those of HIV, including the induction of pathologic lesions (6, 21, 25, 52, 56), upregulation of adhesion molecules (48, 49), infection of endothelial cells in vitro (33), induction of apoptosis (1), and induction of cognitive and motor impairments (38). Additionally, elevated levels of quinolinic acid appear to correlate with elevated viral loads in the brain (18, 22, 46). Results of more recent studies have suggested that neuroinvasion by SIV may be accompanied by the influx of infected monocytes into perivascular areas of the brain (27).

A major impediment to the study of SIV-induced neurologic disease has been the lack of a uniformly neurovirulent isolate. A virus that induces neuropathogenic effects in 100% of inoculated animals is an ideal model system for investigating mechanisms of neurovirulence, induction of neurologic disease, and the effectiveness of therapy on neuropathogenesis. Recently, two groups have described an increase in neurovirulence following the adaptation of SIV to the brain (12, 32, 55). However, the first system (12, 32), a variant of SIVmac17E, still does not induce 100% neurovirulent infection and the second system (55) relies on inoculation of macaques with microglial cells obtained from SIV-infected macaques instead of with an SIV isolate. This latter system is not very practical for neuropathogenic investigations. Thus, a more uniformly neuropathogenic model is still needed for detailed studies of the central nervous system (CNS).

We describe here the initial characterization of a highly neuropathogenic isolate of SIV from sooty mangabeys. This isolate, termed SIVsmmFGb, replicates well both in vitro and in vivo, is highly macrophage tropic, and is neurovirulent in 100% of infected pig-tailed macaques. This virus may serve as an important reagent for the analysis of lentiviral effects on the CNS as a model for HIV-associated dementia.

MATERIALS AND METHODS

Initial animal and transmission of SIV infection to macaques and other managbeys.

The sooty mangabey FGb was housed at the Yerkes Regional Primate Research Center until it was humanely sacrificed in 1989. A brief history of this animal is described in Results. At necropsy, blood was obtained from FGb and transfused directly into two rhesus macaques (ROn2 and RHo2), two pig-tailed macaques (PGm and PHm), and two sooty mangabeys (FRk and FIk).

Specimen collection from SIV-infected monkeys.

To obtain specimens, animals were anesthetized by intramuscular injection of ketamine (10 to 15 mg/kg of body weight). Blood was collected by venipuncture. Peripheral lymph nodes were collected by routine biopsy procedures. CSF, obtained from the cisterna magna, was collected aseptically in a conical centrifuge tube.

Peripheral blood mononuclear cells (PBMC) were prepared from blood samples by centrifugation over lymphocyte separation medium (Organon Teknika, Durham, N.C.). Cells at the interface were collected and washed before use. Single-cell suspensions of lymph node cells were prepared by mincing and passage through a 70-μm-pore-size nylon mesh. Cells were washed before use.

Virus isolations and growth of virus stocks.

Virus isolations from PBMC were performed by first stimulating PBMC for 3 to 5 days in RPMI 1640 containing 10% heat-inactivated fetal calf serum, 5 μg of concanavalin A (ConA) per ml, 5% interleukin 2 (IL-2), and antibiotics. At the end of the stimulation period, 10⁷ cells were cocultured with 10⁷ stimulated human PBMC or 2 × 10⁶ CEMx174 cells. Cocultures containing only PBMC were fed with fresh cells every 9 to 10 days, and medium was changed twice per week. Cocultures with CEMx174 cells were split once per week. Supernatants from cocultures were monitored for the presence of reverse transcriptase (RT) activity on a weekly basis. A culture was considered positive if two successive positive RT results were received.

Virus isolations from lymph nodes were conducted in the same manner as those from PBMC, except that lymph node cells (LNC) were used. Virus isolation from CSF was performed by inoculating stimulated human PBMC with 0.5 to 1.0 ml of CSF. Cultures of LNC or CSF were monitored for the development of RT activity on a weekly basis.

Virus isolated from the LNC of a pig-tailed macaque (PGm) and from the CSF of a rhesus macaque (ROn2) were used for the preparation of viral stocks by inoculating stimulated human PBMC. At the first sign of positive RT activity, cultures were expanded by the addition of fresh, stimulated PBMC. At peak RT activity, cell-free virus stock was prepared from the supernatant of infected PBMC, aliquoted, and stored under liquid nitrogen. Stocks were titrated by limiting dilution on CEMx174 cells. Cells from the virus stock infections were used for the preparation of genomic DNA with a commercially available kit (Puregene; Gentra Systems, Minneapolis, Minn.).

In vivo infections and subsequent monitoring.

Anesthetized macaques (ketamine, 10 mg/kg) were inoculated intravenously with 10⁴ 50% tissue culture infectious doses of virus (either virus from cells of a mesenteric lymph node of PGm [PGm/MLN] or from the CSF of ROn2 [ROn2/CSF]). At biweekly-to-monthly intervals, animals were again anesthetized and given full physical examinations and blood was collected by venipuncture. Blood samples were used for complete blood counts, fluorescence-activated cell sorter analysis of lymphocyte subsets (3), and the preparation of PBMC and plasma.

To test for the presence of simian T-cell lymphotrophic virus (STLV) in animals, PBMC or LNC were used for genomic DNA preparation. Subsequently, this DNA was then used as the template in nested PCRs designed to amplify STLV sequences as previously described in detail (30).

Cloning strategy, PCR amplification, sequencing, and production of molecularly cloned virus.

The strategy for preparing full-length molecular clones of FGb-derived virus was similar to that used by us in the past: we generated 5′ and 3′-half clones by PCR and combined them (39, 40). For PCR amplification of 5′ and 3′ halves, genomic DNA prepared from the PGm virus culture was used as the template. Amplification reactions were performed with an Expand Long-Template PCR kit (Boehringer Mannheim, Indianapolis, Ind.), according to the manufacturer’s instructions. Primers for amplification were as follows: 5′-half the forward primer 443 (5′CGCTTT CGA ACA GTG GGA TGA CCC CTG GGG AGA GGT3′), the 5′-half reverse primer 426 (5′TTT TCT CGA GGT ATT TCT TGT TCT GTG GTG ATC A3′), the 3′-half forward primer 023 (5′ATG CAA GCT TAG GGG ATA TGA CTC CAG CAG A3′), and the 3′-half reverse primer 066 (5′AAT ACT CGA GAA AGG GTC CTA ACA GAC CA3′) (these primers contained restriction sites [underlined]) at their 5′ termini to facilitate cloning). Amplification products were gel purified, digested with the appropriate enzymes, and cloned into the plasmid vector pGEM7ZF (Promega, Madison, Wis.). For combining 5′ and 3′ halves, plasmids were passaged through DM1 bacteria (Dam⁻; Life Technologies, Gaithersburg, Md.) prior to being digested with the enzyme BclI. Plasmids were then digested with the enzyme XhoI, and products were ligated to generate a full-length molecular clone termed SIVsmmPGm5.3 (PGm5.3). Double-stranded plasmid DNA was sequenced by both the Sequenase method (Amersham Life Science, Arlington Heights, Ill.) and the fmol DNA Sequencing method (Promega). Sequence analyses were performed with Intelligenetics Suite software (Oxford Molecular, Beaverton, Oreg.) and the Lasergene software package (DNASTAR, Inc., Madison, Wis.).

To produce virus for in vitro analyses, the molecular clone PGm5.3 was transfected into CEMx174 cells with DEAE-dextran. The clone was determined to be infectious by the development of RT activity within a week after transfection. The culture was expanded, and at peak RT activity, cell-free supernatant was harvested, aliquoted, and stored under liquid nitrogen. Additionally, the molecular clone PGm5.3 was also used to transfect 293 cells with the reagent Fugene (Boehringer Mannheim). After 48 h, ConA-stimulated rhesus macaque PBMC were laid over the 293 cells. After an additional 24 h, the PBMC were removed to a new flask for growth of a PBMC-derived virus stock. This stock was prepared as described above.

Virus replication in PBMC.

To examine the kinetics of virus replication, PBMC isolated from rhesus or pig-tailed macaques were stimulated with ConA and IL-2 for 3 days. Infections were initiated by incubating virus (10 ng of p27) with 10⁷ PBMC overnight at 37°C. Following washes, PBMC were resuspended in medium containing IL-2. At various times after infection, 1 ml of supernatant fluid was removed from cultures, centrifuged to remove cells, and frozen at −70°C until use. At that time, the fluid in the culture was replenished with fresh medium. At the end of the study, the RT activities in the supernatants were quantitated.

Virus replication in blood-derived macrophages.

The ability of FGb-derived viruses to replicate in macrophage populations was evaluated as follows. Pig-tailed macaque PBMC were obtained from the blood of healthy animals (SIV⁻, STLV⁻, and type D simian retrovirus [SRV] negative). Cells were resuspended in macrophage medium (RPMI 1640 containing 15% human AB⁺ serum, 1.5 ng macrophage colony-stimulating factor [R & D Systems], 0.08 ng of granulocyte-macrophage colony-stimulating factor [R & D Systems], 10 mM HEPES, and antibiotics) at a concentration of 3 × 10⁶ cells/ml and distributed into the wells of a 24-well microtiter plate. After 4 days, the nonadherent cells were removed and cells were fed with fresh macrophage medium. Cells were incubated for an additional 3 to 4 days to allow full differentiation of macrophages. Virus infections (in quadruplicate) were initiated by adsorption of 10 ng of input virus stock overnight at 37°C. Following washes, cells were overlaid with macrophage medium. One-half of the volume was replaced with fresh macrophage medium every 3 to 4 days. At various times after infection, supernatants were harvested and used to determine the levels of RT activity.

Histopathologic analyses and in situ hybridization.

Tissues obtained at necropsy were fixed in buffered 10% formalin for at least 7 days before they were routinely processed into paraffin blocks. Blocks were sectioned (thickness, 6 μm), and tissue sections were stained with hematoxylin and eosin. For virus localization, productively infected cells in formalin-fixed, paraffin-embedded tissues were identified through localization of SIV RNA by in situ hybridization. Six-micrometer-thick sections of brain (cerebrum and cerebellum) and lymph node (mesenteric and axillary) were deparaffinized in xylene and rehydrated in graded ethanol to diethyl pyrocarbonate-treated water. Endogenous alkaline phosphatase activity was blocked by incubations in 5 mM levamisole and then in 0.2 N HCl. Protease digestion was accomplished by incubation in proteinase K for 10 min at 37°C. Tissues were acetylated in acetic anhydride, prehybridized in hybridization buffer (50% deionized formamide, 1× SSC [0.15 M NaCl, 0.015 M sodium citrate] 1× Denhardt’s solution, 5 mM NaPO₄, 0.1% sodium dodecyl sulfate, 0.25 mg of salmon sperm DNA per ml, 5% dextran sulfate, 0.25 mg of tRNA per ml, 7% diethyl pyrocarbonate-treated H₂O) for 30 min at 50°C, and hybridized overnight at 50°C with a digoxigenin-labeled antisense SIV riboprobe cocktail (which spans the length of the SIVmac239 genome). The following day, tissues were washed thoroughly in 2× SSC containing formamide, treated with RNase, washed, and blocked with 10% normal horse serum before being incubated in alkaline phosphatase-conjugated antidigoxigenin. Sections were then washed in 50 mM Tris-HCl–150 mM NaCl (Tris-buffered saline [pH 7.6]) and incubated for 6 h in NBT-BCIP (4-nitroblue tetrazolium chloride–5-bromo-4-chloro-3-indolylphosphate). Chromagen development was stopped in 10 mM Tris–EDTA solution, and sections were washed in distilled water, counterstained with nuclear fast red, dehydrated, cleared, and mounted with permanent mounting medium. Negative controls included anatomically matched tissues from uninfected animals processed in parallel with infected tissues (with antisense SIV riboprobe), as well as SIV-infected tissues probed with SIV sense riboprobe.

Nucleotide sequence accession number.

The entire sequence of PGm5.3 has been submitted to GenBank under the accession no. AF077017.

RESULTS

An STLV- and SIV-infected sooty mangabey, FGb.

FGb, a Yerkes colony-born sooty mangabey naturally infected with SIV and STLV, was initially examined for a head wound sustained in 1987. Due to progressive weight loss and deteriorating clinical condition, FGb was euthanized. Because of the unique hematologic history (lymphocytosis followed by anemia and thrombocytopenia) and dual retroviral infection, blood was obtained on the day of euthanasia for transfusion into other monkeys (described below). In situ hybridization for SIV RNA in tissues of FGb (Fig. 1) revealed small numbers of SIV-positive cells within the paracortex and medullas of lymph nodes. However, productively infected cells were not found in the brain, spinal cord, or other nonlymphoid tissues.

FIG. 1.

In situ hybridization for SIV RNA in sooty mangabey FGb. An axillary lymph node from animal FGb was used for in situ hybridization studies to identify productively infected cells. Note that only a few infected cells (that stain bluish-purple with NBT [arrows]) were found within the lymph node parenchyma, typical of cells of naturally infected, SIV-positive sooty mangabeys.

Clinical disease in macaques transfused with blood from an SIV- and STLV-infected managabey.

The initial focus of our investigations was to examine the effects of concomitant STLV and SIV infection. For this purpose, two sooty mangabeys, two pig-tailed macaques, and two rhesus macaques were transfused with blood from FGb and monitored for development of disease (Table 1). The sooty mangabeys, while becoming infected, did not develop clinical disease and remain healthy to date. All four macaques developed simian AIDS, characterized by the loss of CD4⁺ lymphocytes and wasting, and were sacrificed by 13 months posttransfusion. The pig-tailed macaques in this study developed disease more rapidly than the rhesus macaques (survival times of 4 and 5 months versus 11 and 13 months, respectively). Histopathologic examinations revealed that all animals had lesions characteristic of SIV-related diseases, including opportunistic infections and/or giant-cell inflammation. Additionally, three of the four animals (PGm, PHm, and ROn2) had SIV-positive-giant-cell encephalitis. Prior to euthanasia, these three animals displayed clinical neurologic signs which included tremors, ataxia, head tilt, and anisocoria (differences in pupil size), suggesting an unusually high incidence of retroviral neurologic disease. However, because the animals had been given a transfusion, it was unknown whether the rapid disease development and neuropathogenic effects were associated with concomitant SIV and STLV infection or with SIV or STLV infection alone.

TABLE 1.

Outcome of transfusion of blood from mangabey FGb to macaques and mangabeys

Animal	Time to death (mo) or current status	SIV presence in CNSa	Neurologic signsb	Necropsy findings	CD4+ cell count
					At transfusion	At death or most recent
Pig-tailed macaques
PGm	4	+	+	Wasting, candidiasis, SIV encephalitis, cryptosporidiosis	2,530	1,070
PHm	5	+	+	Wasting, pneumocystis pneumonia, SIV encephalitis, giant-cell colitis	1,860	590
Rhesus macaques
RHo2	10.5	−	−	Wasting, giant-cell colitis	3,780	30
ROn2	13	+	+	Wasting, giant-cell colitis, SIV encephalitis	4,100	430
Mangabeysc
FRk	Alive	NAd	NA	NA	3,570	2,780e
FIk	Alive	NA	NA	NA	2,940	1,140e

^aAs detected by in situ hybridization for SIV RNA in the brains and spinal cords of the indicated animals. 

^bObservational only, i.e., tremors and ataxia. 

^cMangabeys became infected, as was determined by virus isolation and seroconversion. 

^dNA, not applicable. 

^eMost recent measurement of CD4⁺ level. 

Clinical disease in macaques infected with virus isolates.

To investigate the role of SIV in neuropathogenesis, virus was isolated from two of the macaques and passaged into new animals. Virus stocks were prepared by inoculation of human PBMC with ROn2/CSF or by coculture of human PBMC with PGm/MLN (both CSF and MLN were obtained at necropsy). These animals were chosen because they were thought to be the most severely affected based on clinical and histopathologic evaluations. The difference in choice of tissue for virus isolation was to investigate the ability of viruses isolated from brain or from lymphoid tissue to induce neuropathogenic infections and to examine whether viruses isolated from different subspecies could induce similar disease patterns. The fact that the virus stocks were prepared as cell-free supernatant precluded contamination with STLV, since this virus is highly cell associated and has not been shown to be infectious in vivo as cell-free stock. However, the virus stocks were tested for the presence of STLV antigen with a human T-cell lymphotrophic virus type 1 antigen kit (Coulter), which confirmed that STLV was not present.

Three pig-tailed macaques and three rhesus macaques were inoculated intravenously with 10⁴ 50% tissue culture infections doses of either virus (total of 12 animals). All pig-tailed macaques developed AIDS-like disease characterized by depletion of CD4⁺ cells and the development of opportunistic infections or lesions typically observed in SIV-infected macaques (26, 28, 29). In contrast, opportunistic infections and giant-cell inflammation were shown to occur less frequently in the infected rhesus macaques (Table 2). The development of disease in these animals was not dependent upon the origin of virus (CSF versus MLN), suggesting that both viruses were highly pathogenic. All animals were tested for the presence of STLV by nested-PCR amplification of genomic DNA according to published methodologies (30), and all were determined to be negative (data not shown). DNA isolated from the lymph node of FGb was used as a positive control.

TABLE 2.

Outcomes of infection with FGb-derived viruses

Animal (virus)	Time to death (mo)	SIV presence in CNSa	Neurologic signsb	Necropsy findings	CD4+ cell count
					At infection	At death
Rhesus macaques
RLt2 (ROn2)	4	+	−	SIV encephalitis, giant-cell colitis	710	290
RHd1 (PGm)	5.5	−	−	Wasting, anemia	790	510
RRy2 (PGm)	16.5	−	−	Streptococcal septicemia, anemia	1,390	790
RQc3 (PGm)	18	−	−	Wasting, adenoviral enteritis	910	300
RYt2 (ROn2)	20.5	−	−	Wasting, endocarditis, thrombocytopenia	1,640	330
N902 (ROn2)	36	−	−	Wasting, gastritis	890	246
Pig-tailed macaques
12058 (PGm)	3	+	+	Cytomegalovirus infection, SIV encephalitis, SIV pneumonia	1,170	60
PEs (PGm)	3.5	+	+	Wasting, SIV encephalitis	700	440
4290 (ROn2)	3.5	+	+	Cytomegalovirus infection, SIV encephalitis	1,400	0
PBt (PGm)	5	+	−	Wasting, SIV encephalitis, SIV pneumonia	1,100	135
PZm (ROn2)	7.5	+	−	SIV encephalitis, pneumocystis pneumonia, candidiasis	2,620	0
PGt (ROn2)	8	+	−	SIV encephalitis, candidiasis, SIV pneumonia	1,180	10

^aAs detected by in situ hybridization for SIV RNA in the brains and spinal cords of the indicated animals. 

^bObservational only, i.e., tremors and ataxia. 

As observed with the initial transfusion animals, the pig-tailed macaques in this second cohort developed disease more rapidly than the rhesus macaques, with the mean time to sacrifice being 4.8 months for the pig-tailed macaques and 16.5 months for the rhesus macaques. Evidence for the differential levels of development of disease was observed by monitoring the CD4⁺ cell levels in these animals (Fig. 2). By comparing levels between rhesus (Fig. 2A) and pig-tailed (Fig. 2B) macaques, it can be seen that the depletion of CD4⁺ cells occurred much more rapidly in the pig-tailed macaques. This was similar to the results observed with the initial transfusion cohort (data not shown), reaffirming that the pig-tailed macaques appear to be more sensitive to disease development with these viruses.

FIG. 2.

Longitudinal analysis of Circulating CD4⁺ cells in SIV-infected macaques. Blood samples from rhesus macaques (A) and pig-tailed macaques (B) infected with the FGb-derived viruses ROn2/CSF and PGm/MLN were used for the enumeration of absolute circulating CD4⁺ cells at the indicated time points by fluorescence-activated cell sorter analysis.

Neurologic involvement.

Clinical evidence of neurologic disease was observed in the last month of life for three of the six pig-tailed macaques in this second cohort. However, neurologic signs were not observed in any of the six rhesus macaques. Neurologic signs included head tilt, tremors, incoordination, and behavioral abnormalities such as cowering and unresponsiveness.

To investigate the extent of SIV infection in the lymphoid tissues and in the CNSs of these macaques, samples of lymphoid and brain tissue were used for histopathological studies. In situ hybridization was used to determine the presence of viral RNA in the lymph nodes and brains of all infected macaques. Results of these analyses are depicted in Fig. 3, which shows representative sections of lymph node and brain following hybridization with antisense SIV riboprobe. All macaques were examined in this manner; however, due to space constraints, the results from two rhesus and two pig-tailed macaques are presented. Moderate numbers of SIV-infected cells were found in the paracortices and medullas of lymph nodes from all infected rhesus macaques (Fig. 3A and C) and one pig-tailed macaque (PBt). Profuse viral replication was evident in lymph nodes from the five remaining pig-tailed macaques (Fig. 3E and G). SIV-positive cells were found in the brain tissue of only one of the six rhesus macaques, RLt2 (Fig. 3D). In marked contrast, SIV-infected cells were present throughout the CNS tissues of all pig-tailed macaques (Fig. 3F and H). The CNS virus loads in pig-tailed macaques ranged from moderate levels in two animals (PGt and PBt) to high levels in the remaining four animals. Virus was localized to cells which possessed the cytomorphologic characteristics of macrophages and microglial cells as well as to giant cells (Fig. 3). Infected cells were found in the meninges and were distributed in a multifocal-to-diffuse manner throughout both white matter and gray matter of the brain and spinal cord parenchyma. Infected cells were unassociated in the parenchyma and also occurred around blood vessels as perivascular macrophages (Fig. 3F). Thus, both the ROn2/CSF isolate and the PGm/MLN isolate induced significant neuropathologic lesions. These results suggest that both viruses are neurovirulent, regardless of the tissue of origin. Based on the results generated from these experiments, the focus of research was directed towards the use of pig-tailed macaques and the pig-tailed macaque virus isolate PGm/MLN.

FIG. 3.

In situ hybridization for SIV RNA in rhesus and pig-tailed macaques infected with PGm/MLN or RON2/CSF virus isolates. Representative lymph node (A, C, E, and G) and brain (B, D, F, and H) tissues from rhesus macaques RHd1 (A and B) and RLt2 (C and D) and from pig-tailed macaques PEs (E and F) and 4290 (G and H) were examined for the presence of productive SIV infection by in situ hybridization. Moderate numbers of SIV-positive cells (indicated with the bluish-purple NBT stain) are found within the paracortices of lymph nodes from infected rhesus macaques (A and C). However, infected cells were found only in the brain of one rhesus macaque, RLt2 (D). In contrast, extremely high numbers of SIV-positive cells are found within the lymph nodes (E and G) and brains (F and H) of infected pig-tailed macaques. Note the presence of SIV-positive giant cells in the meninges of pig-tailed macaque 4290 (H, top of photograph). Arrowheads indicate cells which possess the cytomorphological characteristics of macrophages and giant cells in lymph node samples or of macrophages and microglial cells in brain samples. The arrow (Fig. 3F) indicates an SIV-infected perivascular macrophage.

Generation and analysis of a molecular clone of SIVsmmFGb.

An infectious molecular clone of SIVsmmFGb was generated with DNA isolated from PGm/MLN-infected PBMC as the template. The DNA was used as a template in experiments to amplify and clone the 5′- and 3′-half subgenomic fragments of SIV by PCR, as we have done in the past (39, 40). Generation of a full-length molecular clone was accomplished by joining the 5′ and 3′ halves at the BclI restriction enzyme site. By this methodology, several molecular clones were generated.

The results of transfection of these clones into CEMx174 cells revealed that only one clone, PGm5.3, was biologically active. Cytopathic effects (syncytium formation) in the cell culture were observed by 2 days posttransfection, and RT activity, indicating the presence of biologically active virus, was detected by 4 days posttransfection. A cell-free stock of virus was prepared from the transfected CEMx174 cells for additional characterization, described below.

The complete DNA sequence of PGm5.3 was determined, and the deduced sequences of protein products were compared to those of other SIV isolates (Table 3). Overall, amino acid homology studies indicated that PGm5.3 is closely related to other SIVsmm isolates. As expected, the proteins most conserved in amino acid homology were Gag and Pol; however, a high degree of conservation was also observed in three accessory proteins, Vif, Vpx, and Vpr. Markedly lower amino acid homologies were observed for Tat, Rev, Env, and Nef.

TABLE 3.

Comparison of PGm5.3 with other SIV isolates

Isolate	% Amino acid homologya of indicated protein of PGm5.3 to corresponding protein of isolate
	Gag	Pol	Vif	Vpx	Vpr	Tat	Rev	Env	Nef
SIVsmmPBj6.6	94.7	98.1	93.0	92.0	95.0	81.2	77.0	85.5	75.6
SIVsmH4	96.3	97.6	92.5	96.4	93.3	81.2	76.2	86.0	78.4
SIVstm37.16	87.6	91.6	79.0	87.5	86.1	69.8	68.0	81.5	75.0
SIVmac239	90.3	91.8	80.8	91.1	86.1	77.1	75.0	82.9	76.5

^aSequences of other viruses were obtained from GenBank. 

Upon closer examination of Env, it was found that amino acid differences between PGm5.3 and other SIV isolates were concentrated mainly within the variable domains of gp120 (Fig. 4A), with additional variation occurring in the C-terminal region of gp41 (3′ to the V5 region). Significant differences obvious from this alignment include (i) an insertion in the V1 region, similar to that present in the acutely pathogenic PBj6.6 isolate (40), and (ii) a 3-amino-acid deletion in gp41. The cysteine residues and N-linked glycosylation sites were found to be well conserved between PGm5.3 and other viruses, with the exception of an additional N-linked glycosylation site in the V3 region. Additionally, PGm5.3 contains an N-linked glycosylation site at the end of V2 that is not observed in other SIVsmm isolates but that is observed in the SIVmac239 isolate. The functional CD4 binding domain of PGm5.3 is highly conserved except for a K-to-R change, which does not alter the charge in this area.

FIG. 4.

Alignment of protein sequences of PGm5.3 and other SIV isolates. Deduced amino acid sequences of Env (A) and Nef (B) were aligned with amino acid sequences of the following isolates: SIVsmmPBj14 (PBj6.6), SIVsmH4, SIVstm (clone 37.16), and SIVmac239. For Env alignments, boxed areas indicate predicted N-linked glycosylation sites, asterisks indicate N-linked glycosylation sites in PGm5.3 not observed in other viruses, and the oval indicates a 3-amino-acid deletion in PGm5.3. For Nef alignments, boxes with solid lines indicate SH2 binding sites, boxes with dotted lines indicate conserved amino acids not present in PGm5.3, and boxes with dashed lines indicate SH3 binding sites.

Comparison of PGm5.3 Nef to the Nef’s of other SIVs reveals several differences (Fig. 4B). Perhaps the most significant mutation is the absence of an SH2 binding motif that is present in all other SIV isolates (amino acids 28 to 31; YXXL). However, the SH3 binding motif (PXXP) is conserved in this virus. Additionally, there are five sites that have conserved amino acid motifs in the other SIV isolates which are changed in PGm5.3. These are (i) the KGL (lysine, glycine, leucine) motif at amino acids 48 to 50, (ii) the S (serine) at amino acid 52, (iii) the C (cysteine) at amino acid 55, (iv) the T (threonine) at amino acid 200, and (v) the P (proline) at amino acid 217. Two of these changes are significant, namely, the change of the cysteine at amino acid 55 to a phenylalanine (C to F) and that of the proline at amino acid 217 to a serine (P to S). These amino acid differences may cause significant changes in the secondary and tertiary structures of Nef.

Other notable differences in the genome of PGm5.3 include (i) the insertion of two amino acids (AG; alanine, glycine) near the middle of Tat, relative to the sequence of Tat in SIVsmm and SIVstm (SIV from stump-tailed macaques) isolates, which is similar to Tat in the SIVmac239 isolate; (ii) an additional two arginine residues inserted near the middle of Tat, generating an R₅ motif, also similar to Tat in SIVmac239; and (iii) a predicted N-linked glycosylation site in the p27 region of the Gag polyprotein, which is similar to Gag in the PBj6.6 isolate. Other proteins contain various point mutations resulting in amino acid substitutions.

Growth characteristics of PGm, ROn2, and PGm5.3 in PBMC and macrophages.

To assess the biological activities of FGb-derived viruses and to compare their replicative abilities in cells of rhesus and pig-tailed macaques, we analyzed the growth of these viruses in bulk PBMC and primary macrophage cultures. In stimulated PBMC populations, all viruses exhibited vigorous growth in both pig-tailed and rhesus macaque PBMC (Fig. 5A and B, respectively) but grew to higher titers in pig-tailed macaque PBMC than in rhesus macaque PBMC. Virus derived from ROn2/CSF was able to replicate to higher titers than virus derived from either the uncloned PGm/MLN stock or the molecular clone PGm5.3. The ROn2/CSF virus also reached peak titers more rapidly (day 7) than the other viruses (day 10 postinfection for PGm/MLN and day 14 postinfection for PGm5.3). Virus derived from the molecular clone PGm5.3 was the most poorly replicating virus in both PBMC samples. However, replication levels of PGm5.3-derived virus still reached 10⁶ cpm/ml in pig-tailed macaque PBMC (twofold less than the peak in ROn2/CSF) and 5 × 10⁵ cpm/ml in rhesus macaque PBMC (fourfold less than the peak in ROn2/CSF).

FIG. 5.

Replication of SIVsmmFGb-derived viruses in PBMC. ConA-stimulated PBMC (10⁷) obtained from pig-tailed macaques (A) or rhesus macaques (B) were infected with virus isolated from either ROn2/CSF, PGm/MLN, or virus derived from the molecular clone PGm5.3. Cell-free supernatants were harvested at the indicated time points and used for quantitation of RT activity as described in Materials and Methods. 32-P, ³²P.

Because the majority of viruses in tissues of the SIVsmmFGb-infected macaques appeared to be localized in macrophages, we chose to assess the replicative abilities of the PGm, PGm5.3, and ROn2 viruses in macaque macrophage populations. These results (Table 4) show that FGb-derived viruses are highly macrophage tropic, more so than SIVsmmPBj, a macrophage-tropic virus (51) that is found mainly in macrophages in tissues of animals dying of acute disease. Similar to the results observed in PBMC cultures, FGb-derived viruses were able to replicate to higher levels in pig-tailed macaque macrophages than in rhesus macaque macrophages. In pig-tailed macaque macrophages, low levels of virus replication were seen as early as 7 days postinfection with FGb-based viruses and with the macrophage-tropic PBj isolate. Replication of the FGb-based viruses quickly accelerated, and high levels of RT activity could be detected in the supernatants by day 14 postinfection. In rhesus macaque macrophages, the outcome was more virus specific. While virus derived from ROn2 was able to replicate very well in rhesus macaque macrophages, virus derived from PGm showed only low-level replication at 14 days postinfection. SIVmac239 was used as a negative control and did not replicate to detectable levels in either pig-tailed or rhesus macaque macrophages. Additionally, SIVsmH4, a minimally pathogenic isolate, was not able to replicate in rhesus macaque macrophages.

TABLE 4.

Replication of FGb-derived viruses in macaque macrophages

Virus	Day postinfection	RT activity in macrophages ofa:
		Pig-tailed macaques	Rhesus macaques
SIVsmmPGm	7	15,496	6,539
	14	392,236	12,766
SIVsmmROn2	7	8,340	138,918
	14	225,117	165,273
PGm5.3	7	68,965	NDb
	14	265,636	ND
PBj6.6	7	16,198	ND
	14	77,708	ND
SIVmac239	7	7,488	5,187
	14	9,100	4,680
SIVsmH4	7	ND	5,217
	14	ND	7,809
Negative	7	6,643	4,530
	14	6,065	5,142

^aIn counts per minute per milliliter of culture supernatant. 

^bND, not done. 

DISCUSSION

The detrimental effects of HIV-1 infection in the brain have been well documented, but the pathogenesis of disease is poorly understood. An animal model system demonstrating the neuropathogenic effects of HIV or SIV on a consistent basis would be invaluable to elucidating the mechanisms of neurovirulence and development of clinical neurologic disease. Although the SIV-macaque system has been the model of choice for investigating the pathogenesis of AIDS, most non-brain-passaged SIV isolates induce neurovirulent infections in about 25 to 40% of infected animals (43, 50). The in vivo adaptation of viruses by brain passage has resulted in increased neurovirulence (12, 32, 55); however, the development of an easily reproducible SIV-macaque system that induces neurovirulent infection in 100% of animals is still highly desirable in order to perform meaningful studies with smaller numbers of animals.

In this paper, we have described a new SIV isolate, termed SIVsmmFGb, derived from a sooty mangabey (FGb), which induces neurovirulent infections in 100% of infected pig-tailed macaques. In the initial cohort, two pig-tailed macaques and one rhesus macaque transfused with blood from this animal developed clinical neurologic symptoms in addition to AIDS. Viruses derived from two of these animals, ROn2 and PGm (a rhesus and pig-tailed macaque, respectively), were used for subsequent inoculations into pig-tailed and rhesus macaques to investigate the basis for neurologic disease. The animals inoculated with these viruses displayed a differential development of disease with respect to time and with respect to neuropathogenesis. Pig-tailed macaques inoculated with these viruses were more susceptible to development of disease than were rhesus macaques. In addition to exhibiting a more rapid progression to disease, the pig-tailed macaques demonstrated a higher level of neuropathology (100%) and neurologic dysfunction (62.5%) than did rhesus macaques (25 and 12.5%, respectively). This finding is in agreement with the observation that pig-tailed macaques are more susceptible to infection and/or disease development with SIV from sooty mangabeys, SIV from African green monkeys, and HIV-1 (2, 19, 29). We believe that the susceptibility of pig-tailed macaques is a factor in this disease development; however, we also believe that the virus phenotype contributes significantly. For example, we have not observed an increased level of neurovirulence in pig-tailed macaques that have died of AIDS following infection with SIV isolates, including SIVsmm9 and various SIVsmmPBj isolates and clones which do not induce acutely lethal disease.

The finding that neurovirulent infection could be induced by virus derived from the brain or lymphoid tissue suggests that neurovirulence is an inherited trait and is not derived through selective adaptation and expansion of the virus in the brain. While these results contradict those of previously described studies, which found that HIV isolates from brain or CSF differed from viruses recovered from blood (7, 8), the results of more recent studies suggest that HIV isolates from CSF and blood are genetically similar (23). Still, this concept has not been thoroughly examined in the SIV arena, and more-focused work on the genetics of neurovirulent viruses is one of our future directions.

In situ hybridization studies revealed the presence of replicating SIV in the CNS (brain parenchyma, meninges, and spinal cord) of all pig-tailed macaques. To our knowledge, no other virus consistently induces this level of SIV replication in the brain of any macaque. Hybridization signals from replicating virus in tissues of pig-tailed macaques were primarily localized in cells that were morphologically similar to macrophages and glial cells. Only rarely did SIV-positive cells appear to be of lymphocyte origin; this may reflect the severe depletion of peripheral CD4⁺ cells observed in most animals at necropsy, or it may reflect the highly macrophage-tropic nature of this virus. Evaluation and quantitation of viral expression and dissemination in other cell types (neurons, astrocytes, and endothelial cells) and tissues are under way and would be best served in a separate publication focusing on pathology.

In these studies, clinical neurologic disease was observed in five of eight pig-tailed macaques but in only one of eight rhesus macaques. It must be noted, however, that the clinical neurological signs described in this study are entirely based upon observations of infected animals. Neurologic deficiencies in humans and domestic animals are usually documented by hands-on neurologic examination of an unanesthetized patient. The facts that these animals are infected with SIV and that they are prone to biting and scratching preclude the examination of an unanesthetized animal. Further documentation of neurologic involvement will need to be accomplished by behavioral and cognitive testing, as has been recently demonstrated (11, 38), and by molecular scanning techniques currently being used for HIV-1-infected persons (41).

In vitro, the viruses used for inoculation replicated well in PBMC derived from pig-tailed and rhesus macaques. Of note is the result that all FGb-derived viruses replicated to higher titers in pig-tailed than in rhesus macaque PBMC. A similar result was observed when virus replication was tested in monocyte-derived macrophages. Again, the viruses displayed differential replication patterns, with growth more vigorous in macrophages derived from pig-tailed macaques than in those derived from rhesus macaques. These results suggest that the combination of replicative ability in the host and macrophage tropism may play a significant role in determining neuroinvasion and development of disease, as both progression to AIDS and neurologic involvement were enhanced in pig-tailed macaques relative to those factors in rhesus macaques. Additional studies will need to be performed in order to determine the exact mechanism of differential levels of replication in these cells.

The isolation of an infectious molecular clone from the PGm-based virus provided an opportunity to examine the molecular characteristics of this virus. As expected, the virus was most closely related to other SIVsmm isolates. While significant changes relative to other SIV isolates were observed, the ability to associate pathogenesis with specific sequence variations was not possible. Most important will be the ability to define sequences that are important for macrophage tropism. SIVsmm isolates have been less well characterized at the genetic level than have SIVmac isolates, where determinants for pathogenesis and macrophage tropism have been identified (35). Recent studies of the SIVsmmPBj14 variant have shown that Vpx and Nef contribute significantly to the ability of this virus to replicate in macrophages (13, 51). The fact that viruses derived from FGb (including virus derived from the molecular clone PGm5.3) replicate much better in macrophages than the PBj isolate strongly suggests that other factors influencing macrophage tropism may be present. However, comparison of the PGm5.3 env gene sequence to those of the env genes of known macrophage-tropic SIVmac isolates (SIVmac316 and SIVmac17E-Fr) revealed no homologies with specific sequences associated with macrophage tropism (data not shown). Additional work will thus be necessary to define important phenotypic determinants of the FGb-derived viruses. Virus derived from PGm5.3 has recently been used to inoculate two animals by the oral route. Both animals have become infected, as determined by virus isolation. These animals are currently being observed for development of disease. Additionally, two macaques were inoculated orally with PGm5.3 virus and sacrificed at 5 days postinfection. Virus was easily isolated from the brains of both animals, showing that this virus has the ability to establish infection in the brain (data not shown). The availability of a molecular clone that reproduces neuropathogenic effects in animals will be of great value for defining determinants of neurovirulence.

In summary, we have isolated an SIV that is 100% neuropathogenic in pig-tailed macaques. This virus-animal system should provide an excellent model for investigating the basis of HIV-induced neurologic disease. Combining pathogenesis studies with behavioral and cognitive studies will be crucial in the elucidation of when neurologic disease occurs and will also be valuable in testing the effectiveness of antiretroviral therapies for AIDS-related CNS disease.