|Home | About | Journals | Submit | Contact Us | Français|
A minor fraction of simian immunodeficiency virus (SIV)-infected macaques progress rapidly to AIDS in the absence of SIV-specific immune responses. Common mutations in conserved residues of env in three SIVsmE543-3-infected rapid-progressor (RP) macaques suggest the evolution of a common viral variant in RP macaques. The goal of the present study was to analyze the biological properties of these variants in vitro and in vivo through the derivation of infectious molecular clones. Virus isolated from a SIVsmE543-3-infected RP macaque, H445 was used to inoculate six naive rhesus macaques. Although RP-specific mutations dominated in H445 tissues, they represented only 10% of the population of the virus stock, suggesting a selective disadvantage in vitro. Only one of these macaques (H635) progressed rapidly to AIDS. Plasma virus during primary infection of H635 was similar to the inoculum. However, RP-specific mutations were apparently rapidly reselected by 4 to 9 weeks postinfection. Terminal plasma from H635 was used as a source of viral RNA to generate seven full-length, infectious molecular clones. With the exception of one clone, which was similar to SIVsmE543-3, clones with RP-specific mutations replicated with delayed kinetics in rhesus peripheral blood mononuclear cells and human T-cell lines. None of the clones replicated in monocyte-derived or alveolar macrophages, and all used CCR5 as their major coreceptor. RP variants appear to be well adapted to replicate in vivo in RP macaques but are at a disadvantage in tissue culture compared to their parent, SIVsmE543-3. Therefore, tissue culture may not provide a good surrogate for replication of RP variants in macaques. These infectious clones will provide a valuable reagent to study the roles of specific viral variants in rapid progression in vivo.
Simian immunodeficiency virus (SIV) infection of macaques provides a model of the variable disease course of human immunodeficiency virus (HIV) infection in humans, ranging from long-term nonprogression to rapid progression (3, 13, 18-20, 22-24, 33, 46, 48, 55, 58). The level at which the plasma viremia stabilizes after primary infection is a strong predictor of the rate of disease progression in both SIV and HIV infection, with high viremia being associated with rapid disease (19, 33, 37, 55). The majority of SIV-infected macaques undergo a slow progressive disease with gradual loss of CD4+ T cells, ending with the development of opportunistic infections and a median survival of 1 to 2 years. Approximately 10 to 20% of SIV-infected macaques progress rapidly to disease within 3 to 6 months of inoculation (16, 18, 22-24, 46, 48, 58). While long-term-HIV-infected nonprogressors have been studied extensively, less attention has been paid to the study of rapid progression in humans. Individuals that progress to AIDS in a period of 1 to 2 years from the time of infection have been identified among adult and infant populations, with a higher frequency in infants (9, 14, 38-40). These individuals demonstrate rapid loss of CD4+ T cells and lack strong cellular and humoral immune responses. However, it is not clear why such patients develop AIDS so rapidly, and the relative contributions of host and viral factors remain undefined. Thus, the study of rapid progressors (RP) in an SIV model will be useful to understand the mechanism of rapid progression in HIV infection and to provide insight into the pathogenesis of classical AIDS development.
SIV-infected RP macaques are distinguished from conventional or slow progressors by transient antibody responses, a persistent high viral load, and uniform development of SIV encephalitis and pneumonia. Clinical disease in these animals is characterized by weight loss, chronic diarrhea, and wasting. Although the mechanism of the immune failure in RP macaques is not entirely clear, recent studies show that RP macaques mount an initial humoral and cellular immune response at the appropriate time following infection, but these responses wane rapidly within the first 3 to 4 weeks of infection (22). The subsequent immune defect in these animals affects both cellular and humoral immunity to SIV, as well as immune responses against unrelated antigens. Since these animals fail to respond to either recall antigens (tetanus toxoid) or new antigen exposure (hepatitis A), the immunologic defect is profound and global. The preservation of CD4+ T cells and loss of humoral and cellular immune responses are consistent with an early loss in T-helper function. Recent studies have now shown that SIV-infected RP macaques suffer an early massive loss of memory CD4+ T cells, presumably due to virus-induced destruction (32, 36, 44, 48, 54). While significant loss of memory CD4+ T cells occurs in all macaques during the primary phase of SIV infection (32, 36, 44, 48, 54), the loss is profound and irreversible in RP macaques, perhaps due to the lack of tissue delivery of new memory CD4+ T cells (44, 48). These data suggest a strong correlation between the maintenance of tissue CD4+ memory T cells and rapid progression to disease (44).
Not surprisingly, considering the lack of memory CD4+ T cells in RP macaques, another characteristic feature of RP macaques is the presence of multinucleated giant cells and the predominance of SIV-expressing macrophages in the lungs and brains of such animals (18, 22). These histopathologic features of RP macaques, as well as macrophage-associated disorders such as encephalitis and pneumonia in RP macaques, suggest that cell tropism of viruses in RP macaques may switch from CD4+ memory T cells to other type of cells, such as macrophages and brain cells. Indeed, despite the apparent lack of immune pressure, the envelope of SIV undergoes unique molecular evolution in RP macaques. Characteristic evolution of SIV in RP macaques has been previously reported for macaques infected with SIVsmE660 or SIVmac239 (6, 29, 51). Sequence analysis of viruses in three SIVsmE543-3-infected RP macaques (8) revealed a unique convergent pattern of substitutions in env, including the loss of a highly conserved potential glycosylation site in the V1/V2 region (N158D/S or S160N/G), substitutions in the V3 analog (P337T/S/H/L and R348W), and substitutions in the highly conserved GDPE motif (G386R and D388N/V). These envelope substitutions are associated with the acquisition of CD4-independent usage of CCR5 in cell fusion assays (8, 12, 51). This property may be associated with substitutions in the GDPE motif and/or V3 loop analog. Substitutions in the V3 loop analog of SIV are associated with changes in tropism in vitro and are associated with CD4-independent usage of CCR5 (12, 21). Previous studies of HIV and SIV have also shown that the GDPE motif is critical for interaction with the CD4 molecule, apparently by direct binding to the CD4 receptor molecule (31); mutations in this region significantly reduce the ability of virus to bind CD4 and impair infectivity (47, 51). Substitutions in the V3 analog and the GDPE motif, which decrease the ability to bind CD4, suggest that RP variants may target cells expressing low levels of CD4, such as macrophages and brain cells (12, 50). However, few analyses have been done to clarify the biological nature of RP variants.
In order to study the biologic effects of these mutations, virus was isolated from the mesenteric lymph node (LN) of a SIVsmE543-3-inoculated RP macaque, H445, at euthanasia and used as the inoculum for naive rhesus macaques. Only one of six rhesus macaques inoculated with this stock developed rapid progression (22). This animal, H635, exhibited all the characteristics of RP macaques, including high persistent viremia, transient antibody responses, and the widespread presence of SIV-expressing macrophages in lymphoid tissues. The goal of the present study was to study the molecular composition of virus in the SIVsmH445 stock as it compared with tissues of the same macaque, to derive full-length infectious molecular clones of SIV directly from the plasma of H635, and to characterize these viruses in vitro.
Rhesus macaque H445 was inoculated intravenously with 2 × 103 50% tissue culture infectious doses of SIVsmE543-3, as previously reported (16). Six naive rhesus macaques were inoculated intravenously with 2 × 103 50% tissue culture infectious doses of SIVsmH445. H445 and one of the SIVsmH445-infected macaques, H635, rapidly progressed to disease and were euthanatized at 16 and 9 weeks postinoculation (wpi), respectively, as previously reported (16, 22). SIVsmE543-3 is a pathogenic molecular clone which was derived from a rhesus macaque (E543) with SIV-induced encephalitis and AIDS (8, 15, 16, 18). SIVsmH445 is a virus stock which was isolated from the mesenteric LN of H445 at necropsy by coculture with naive rhesus peripheral blood mononuclear cells (PBMC).
The viral RNA loads in plasma and cerebrospinal fluid (CSF) were determined by quantitative reverse transcriptase PCR (RT-PCR) using a Prism 7700 sequence detector (Applied Biosystems, Foster City, CA) as described previously (19). Briefly, viral RNA was isolated from plasma or CSF samples by using the QIAamp viral RNA kit (QIAGEN, Hilden, Germany). RT-PCRs were performed using primers S-GAG03 and S-GAG04 and probe P-SUS-05.
Absolute CD4+ or CD8+ T-cell counts in the blood were monitored using an EPICS XL-MCL flow cytometer (Beckman Coulter, Fullerton, CA) by staining cells with the following combination of antibodies: CD3-fluorescein isothiocyanate and CD4-phycoerythrin (BD PharMingen, San Diego, CA) and CD8-phycoerythrin-Cy5 (Beckman Coulter). Analysis of memory/naive subsets in CD4+ T cells was performed as described before (36). Briefly, cells were labeled with the following combination of monoclonal antibodies from BD PharMingen: CD3-Cy7-allophycocyanin, CD4-cascade blue, CD45RA-TRPE, and CD95-allophycocyanin. Labeled cells were fixed with 0.5% paraformaldehyde and analyzed using a modified Becton Dickinson Digital Vantage. Naive CD4+ T cells (CD95− CD45RA+) and memory CD4+ T cells (CD95+ CD45RA− or CD95+ CD45RA+) were shown as a percentage of CD3+ T cells.
Formalin-fixed, paraffin-embedded tissues were stained for SIV viral RNA by in situ hybridization (ISH), utilizing a modification of a method previously described (18). Briefly the sections were deparaffinized; rehydrated; subjected to high-temperature, high-pressure unmasking; treated with methanol-hydrogen peroxide, 0.2 N HCl, and proteinase K; and hybridized overnight at 50°C with either sense or antisense SIVmac239 digoxigenin-UTP labeled riboprobe. The hybridized sections were blocked with 3% normal sheep and horse serum in 0.1 M Tris, pH 7.4, and then incubated with sheep antidigoxigenin-horseradish peroxidase (SAD-HRP) (Roche Molecular Biochemicals). The SAD-HRP was detected with a tyramide signal amplification technique (TSA Plus FITC; Perkin-Elmer, NEL741) and rinsed with Tris buffer. After completion of the ISH assay, the sections were incubated in either rabbit anti-human CD3 (T-cell marker; DAKO A0452) or mouse anti-human macrophage (HAM56; DAKO M0632), rinsed in Tris buffer, and then stained with conjugated secondary antibodies (Invitrogen) goat anti-rabbit immunoglobulin G-Alexa633 or goat anti-mouse immunoglobulin M-Alexa633, respectively. The stained sections were then coverslipped in Vectashield Hardset mounting medium (Vector Laboratories, Burlingame, CA) and photographed with a Lecia confocal scanning microscope (Owen Schwartz, Juraj Kabat, and Meggan Czapiga, NIAID, NIH). Negative controls included antisense probe with uninfected tissues, sense probe with infected tissues, antisense probe with infected tissues without SAD-HRP, anti-human CD3, and anti-human HAM56.
Viral RNA was isolated from SIVsmH445 viral stock and sequential plasma samples of macaque H635 by using the QiaAmp viral RNA kit (QIAGEN, Hilden, Germany). RT-PCR was performed using SuperScript III (Invitrogen, Carlsbad, CA) and Platinum Taq Hi Fidelity (Invitrogen) according to the manufacture's instructions. Primers SMR22 (5′-GAC GGT CAG TCG CAA CAA TTC-3′; nucleotides [nt] 8354 to 8374) and R-R (5′-TGC TTA CTT CTA AAT GGC AGC TTT-3′; nt 10170 to 10193) were used for reverse transcription of the gp120 region and the entire genome, respectively. For analysis of gp120, the region of gp120 from V1 to V5 was amplified using primers SME1 (5′-GCC CTG TGT AAA ACT TAC CCC A-3′; nt 6898 to 6919) and SME2 (5′-CTT GAG GCA CCA GTT GTG GT-3′; nt 8171 to 8190). The PCR product was cloned into plasmids using the TOPO TA cloning kit (Invitrogen), and the resulting clones were sequenced. To determine the consensus sequence of viruses in H635 at necropsy, three regions were amplified using the primers Nar-F (5′-GGTTGGCGC CCG AAC AGG GAC TT-3′; nt 833 to 855) and 3U3-R (5′-CAATTC TGG ATC AAA CTT CCA TGC-3′; nt 9732 to 9755), R-F (5′-CAGTCG CTC TGC GGA GAG GCT GG-3′; nt 531 to 553) and Gag-R (5′-TGA CGC AGA CAG TAT TAT AAA GGC TC-3′; nt 1292 to 1317), and Bgl-F (5′-GGC AGC AGA TCT TGG CCT TGG C-3′; nt 8897 to 8918) and R-R. These PCR products were directly sequenced. Sequence data were aligned by GeneWorks (Intelligenetics, Campbell, CA) and compared with those for SIVsmE543-3. All of the nucleotide and amino acid numbers used in this study are based on the sequence of SIVsmE543-3 (GenBank accession number U72748). Sequence analysis of envelope genes in H445 tissues was previously reported (8), and sequence data for clones from spleen, thymus, and mesenteric LN are used as sequences of H445 tissues at 16 wpi.
The RT-PCR products used to determine the consensus sequence of H635 at necropsy were used for construction of a cloning vector which has both the 5′ and 3′ ends of the SIV genome. Two RT-PCR products were amplified using R-F and Gag-R (nt 531 to 1317) and Bgl-F and R-R (nt 8897 to 10193) and used as templates for PCR amplification using primers Bgl-F and Gag-R. Since these two PCR products shared the R region of the long terminal repeat (LTR), the resultant PCR product had a complete LTR with the part of the env gene on the 5′ side and the primer binding site on the 3′ side. The PCR product was cloned into plasmid by using the TOPO TA cloning kit and sequenced. The clone with the consensus nucleotide sequence of H635 was obtained by ligation of two clones at the SstI site. The vector, designated pNNBN, was constructed from this consensus clone to allow insertion of a NarI-BglII fragment (nt 837 to 8903) amplified from plasma of H635. This vector has a 5′ fragment from the 5′ end of the SIV genome to the NarI site and a 3′ fragment from the BglII site to the 3′ end of the SIV genome. The 5′ fragment is identical to SIVsmE543-3, but the 3′ fragment contains three substitutions (see Fig. Fig.4).4). The RT-PCR product from plasma of H635 at necropsy was obtained using primers Nar-F and Bgl-R (5′-GGC CAA GAT CTG CTG CCA CCT CTG TC-3′; nt 8888 to 8913). This PCR product was digested with NarI and BglII, and directly cloned into pNNBN to obtain full-length SIV clones. Alternatively, the PCR product was cloned using the TOPO TA cloning kit before insertion into pNNBN because of the high efficiency of cloning PCR products. One infectious clone, SIVsmH635F-L3, which was obtained by the method described above, was used to obtain additional clones by replacing the SfuI-BglII region (nt 5878 to 8903) with other plasmid clones containing the NarI-BglII fragment (nt 837 to 8903). One clone, SIVsmH635FC, was constructed by recombination of three clones by using BsmI (nt 6709) and ClaI (nt 8094) sites to obtain a clone with the consensus nucleotide sequence in the SfuI-BglII region. The nucleotide sequence of SIVsmH635FC is identical to the consensus of H635 at necropsy except for two substitutions in pol and gag.
Human CD4+ cell lines CEMx174 (52) and PM-1 (34) were maintained in RPMI 1640 medium containing 10% heat-inactivated fetal calf serum (FCS), 2 mM l-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin. Human embryonic kidney cells (293 cells) (17) were maintained in Dulbecco's modified Eagle medium (DMEM) containing 10% FCS, 2 mM l-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin. PBMC from SIV-naive, healthy rhesus macaques were separated from whole blood, stimulated with 5 μg of phytohemagglutinin (PHA) per ml and 10% interleukin-2 (IL-2) (Advanced Biotechnologies, Columbia, MD) for 3 days, and maintained in RPMI 1640 medium containing 10% FCS and 10% IL-2. Rhesus monocyte-derived macrophages (MDM) were obtained from rhesus PBMC as previously described (30). In brief, 6 × 105 PBMC per well in a 96-well plastic plate were cultured for 5 days in RPMI 1640 containing 15% FCS, 10% human serum type AB (Sigma, St. Louis, MO), and 500 units/ml macrophage colony-stimulating factor (R&D Systems, Minneapolis, MN); washed four times with RPMI; and cultured in fresh medium for 2 days more. After being washed with RPMI, MDM were maintained in RPMI containing 15% FCS, 5% human serum type AB, and 200 units/ml macrophage colony-stimulating factor. Rhesus alveolar macrophages (AM) were prepared from bronchoalveolar lavage fluid as previously described (26). Briefly, the lavage fluid was filtered and centrifuged, and the cell pellet was washed four times with phosphate-buffered saline (PBS) supplemented with 1% bovine serum albumin, 100 U/ml penicillin, 100 μg/ml streptomycin, and 50 μg/ml gentamicin. AM were maintained in DMEM containing 10% human serum type AB, 5% FCS, 2 mM l-glutamine, 1 mM sodium pyruvate, and 50 μg/ml gentamicin at 3× 105 cells/well in a 48-well plate.
The virus stocks were prepared from supernatant of 293 cells which was transfected by the plasmid DNA using FuGENE 6 transfection reagent (Roche Diagnostics, Indianapolis, IN) and were adjusted to contain equal virion-associated reverse transcriptase units (56) for infection experiments. Cells were infected by incubation of cell-free virus for 3 h, or for 16 h in the case of AM. After incubation, the cells were washed with PBS and the virus replication was monitored by RT activity of culture supernatants collected at 2- to 4-day intervals. RT activities were quantified with a Typhoon phosphorimager (Amersham Pharmacia Biotech, Piscataway, NJ). Virus replication in rhesus PBMC was assessed in duplicate using several donor macaques. Infection of MDM and AM was performed in triplicate, and the average RT activity and standard deviation were employed to compare infectivities of viruses. Macrophage-tropic SIVmac316 and T-cell-tropic SIVmac239 (41) were used as positive and negative controls of infection in AM, respectively.
Coreceptor usage was determined by infection of GHOST cells expressing CCR1, CCR2b, CCR3, CCR4, CCR5, CCR8, CXCR4, V28/CX3CR1, GPR15 (Bob), or CXCR6 (Bonzo/STRL33). These cells were maintained in DMEM containing 10% FCS, 500 μg/ml G418, 100 μg/ml hygromycin, and 1 μg/ml puromycin. GHOST parental cells, in which no vector expressing a coreceptor gene was introduced, were used as a control and were cultured in the same medium except that puromycin was omitted. GHOST cells were obtained through the AIDS Research and Reference Reagent Program, Division of AIDS, NIAID, NIH, from Vineet N. KewalRamani and Dan R. Littman (42). For infection experiments, 1 × 105 cells/well were seeded in 24-well plates the day before infection. RT-adjusted virus stocks were added in the presence of 20 μg/ml Polybrene and incubated for 3 h. Cells were washed with PBS twice, and fresh medium was added. Green fluorescent protein (GFP) expression was measured by flow cytometry using a FACSCalibur (Becton Dickinson, Franklin Lakes, NJ) at 3 days postinfection to determine coreceptor usage. To examine spreading replication in GHOST parental cells, the cells were infected and split 1:6 every 3 days, and GFP expression or RT activity was monitored. The CXCR4-specific inhibitor AMD3100 (10) was used to block CXCR4, which was expressed at a low level in GHOST parental cells. AMD3100 was obtained from the AIDS Research and Reference Reagent Program, Division of AIDS, National Institute of Allergy and Infectious Diseases, National Institutes of Health. Various concentrations of AMD3100 were added and incubated for 1 h at 37°C. Then, viruses were added, and infection experiments were performed as described. HIV-1NL432 (1) which uses CXCR4, was used as a control.
The sequences of representative infectious clones SIVsmH635F-L3, SIVsmH635SB10, and SIVsmH635FC have been submitted to GenBank under accession numbers DQ201172, DQ201173, and DQ201174, respectively.
In order to study the role of RP-specific viruses in disease, we isolated SIV from a SIVsmE543-3-inoculated RP macaque, H445, at necropsy. As previously described, rhesus H445 was typical of RP macaques in terms of transient immune responses, high plasma and tissue viral load (16, 22), and representative RP-specific mutations in the envelope gene (8). SIVsmH445 was isolated by coculture of 5 × 106 PHA-stimulated viably frozen, mesenteric lymph node mononuclear cells with naive rhesus PBMC. The cultures were monitored for RT activity every 3 days, and cell-free supernatants collected at 21 days were cryopreserved in liquid nitrogen. Six rhesus macaques were inoculated intravenously with uncloned SIVsmH445 (Table (Table1).1). Despite the isolation of this virus from an RP macaque, survival of SIVsmH445-infected macaques was similar to that of macaques inoculated with the parental SIVsmE543-3 (16, 18). Only one macaque (H635) developed high viremia in plasma and CSF (Fig. (Fig.1A)1A) and transient immune responses characteristic of rapidly progressive disease (22); this incidence of rapid disease is not significantly different from our experience with SIVsmE543-3. As shown in Table Table1,1, macaques inoculated with SIVsmH445 exhibited a range in characteristic SIV-induced lesions and opportunistic infections (four out of six), with a high frequency of meningoencephalitis.
The RP macaque, H635, was euthanatized at 9 weeks postinfection due to wasting and diarrhea, with pathological lesions characteristic of RP disease progression, SIV-induced encephalitis, and pneumonia (Table (Table1).1). The plasma virus load of the RP macaque H635 reached 109 copies/ml at the peak and was maintained at 108 copies/ml thereafter (Fig. (Fig.1A).1A). The virus load in CSF was also high in this animal, exceeding that in plasma after 4 wpi (Fig. (Fig.1A),1A), consistent with the development of encephalitis. Peripheral CD4+ T cells decreased at 1 wpi but subsequently rebounded to within the normal range for the remainder of infection (Fig. (Fig.1B),1B), similar to observations for other RP macaques (22). Despite relative preservation of CD4+ T cells, analysis of memory/naive subsets showed early, almost complete depletion of memory CD4+ T cells from the peripheral blood (Fig. (Fig.1C).1C). The depletion of the memory subset and predominance of naive CD4+ T cells were also observed in tissues such as lymph node mononuclear cells and spleen that were collected at necropsy (Fig. (Fig.1D).1D). Naive CD4+ T cells ranged between 77.1 and 90.3% of CD4+ T cells in the tissues of H635. This frequency was significantly higher than those in normal macaques, which range from 30 to 70% in PBMC and lymph node mononuclear cells and 10 to 50% of spleen lymphocytes (49). These results indicated that the loss of memory CD4+ T cells was severe and global in this animal, as observed in other RP macaques (44, 48). Consistent with the lack of memory CD4+ T cells, macrophages were predominant SIV-producing cells in samples from H635, similar to our observations for other RP macaques (18, 22). More than 90% of ISH-positive cells were macrophages in mesenteric lymph node (Fig. (Fig.2),2), intestine, and brain in H635, and CD3+ T cells were rarely ISH positive (data not shown).
In order to understand evolution of viral genotype in a passage from one RP macaque to another RP macaque and to clarify the potential in vivo importance of the RP genotype, we evaluated the genetic composition of the virus stock used to inoculate H635, as well as sequential samples from this animal.
As shown in Table Table2,2, the Env gp120 region cloned from the SIVsmH445 virus stock was compared with that observed in macaque H445 tissues. Substitutions specific to RP macaques, such as loss of a potential glycosylation site in the V1/V2 region (N158D/S or S160N/G), P337T/S/H/L and R348W in the V3 analog, G386R and D388N/V in the GDPE motif, and E477K in V5, were frequently observed in H445 tissues, as previously described (8). Other substitutions, such as E340K in the V3 analog, P430T/S/L/R in V4, and D519N in C terminus of gp120, were also frequent in H445, and these substitutions were also observed in multiple RP macaques, although the frequency was not high enough to note as an RP mutation (8). However, these mutations were observed only infrequently among clones derived from the SIVsmH445 stock (Table (Table22 and Fig. Fig.3).3). Almost all of the major mutations in H445 tissues were represented at less than 10% of clones derived from the SIVsmH445 stock virus. A combination of all these characteristic RP-specific mutations was not observed in any of the clones. The virus genotype was similar to that of the parental SIVsmE543-3, consistent with a selective advantage of this virus genotype in vitro. Therefore, the RP-specific mutations appeared to be at a selective disadvantage in rhesus PBMC cultures in vitro.
We then evaluated the viral populations in plasma samples collected sequentially from the RP H635. As shown in Fig. Fig.33 and Table Table2,2, RP-specific mutations were seldom observed in the plasma of H635 during primary viremia (1 and 2 wpi), consistent with the establishment of infection with the predominant SIVsmE543-3-like virus sequence from the SIVsmH445 stock. However, RP-specific mutations identical to those observed in tissues of H445 first appeared as minor populations in clones derived from plasma collected at 4 wpi and became the dominant virus genotype by 9 wpi (Table (Table22 and Fig. Fig.3).3). All the dominant mutations in H445 tissues except R477K were observed in more than 90% of clones at 9 wpi, an even higher proportion than that in H445 tissues. The average number of these mutations in each clone was 5.1 mutations per clone in H445 tissues but 6.7 in H635 at 9 wpi, although those in SIVsmH445 and H635 at 1 and 2 wpi were <1.0 mutation/clone (Fig. (Fig.3).3). In addition to the high frequency of RP mutations, the viral population in the plasma of H635 at 9 wpi was extremely homogeneous, suggesting the selection of a specific genotype of SIV in this animal. Since all the mutations in H635 resulted from the same nucleotide substitutions as those in H445, it is probable that these variants were present at a low frequency in the SIVsmH445 stock and were reselected and expanded in H635. Reselection of RP-specific mutations did not occur in at least two other macaques with conventional disease courses following inoculation with the same virus stock (Q. Dang, unpublished data). As we observed in three SIVsmE543-3-infected RP macaques, this unique convergent pattern of substitutions in env was limited to RP macaques and was not observed in animals that progressed more slowly to AIDS (8). These data suggest that viruses with RP-specific mutations efficiently replicate only in RP macaques and that the dominance of the RP-specific genotype may be an adaptation to a particular microenvironment in an RP macaque.
As shown by the comparison of genotypes in the SIVsmH445 stock with those in tissues of H445, the RP-specific genotypes were at a selective disadvantage in rhesus PBMC cultures in vitro (Table (Table22 and Fig. Fig.3).3). Thus, it was expected that a swarm virus isolated from an RP macaque would not be suitable for in vitro biological analyses of RP-specific genotypes because of an outgrowth of the SIVsmE543-3-like genotype. For this reason, molecular clones were constructed to analyze the biologic properties of viruses with RP-specific mutations. The viral RNA isolated from plasma of H635 at 9 wpi was used to construct molecular clones, since the RP-specific genotype was present at a high frequency and homogeneity in this sample (Fig. (Fig.3).3). This high homogeneity would ensure that the consensus sequence of viruses would be close to the representative genotype in this plasma. Thus, we determined the consensus sequence of a whole viral genome before the construction of clones and used it as a reference to obtain representative clones.
The entire viral genome was amplified by RT-PCR and directly sequenced to determine the consensus sequence of viruses in H635 at 9 wpi. In addition to the seven mutations observed in gp120 (Fig. (Fig.3),3), six additional mutations were identified in the gag, vif, tat, env, and nef genes (Fig. (Fig.4).4). Three mutations, containing one synonymous mutation, were observed in nef, and two of them overlapped env. Based on the consensus sequence, an LTR vector, pNNBN, was constructed to use as a vector to reconstruct a full-length SIV genome by insertion of a NarI-to-BglII fragment amplified from plasma viral RNA. Since pNNBN contained the three consensus mutations in nef, all full-length clones shared a common sequence with these three mutations in nef but differed in the representation of substitutions in viral genes as shown in Fig. Fig.44.
Twenty-eight of a total of 36 clones, which were obtained by this cloning procedure produced RT activity upon transient transfection of 293 cells, but only two of these clones, SIVsmH635F13 and SIVsmH635F-L3, were confirmed to be infectious in screening of rhesus PBMC cultures. As shown in Fig. Fig.4,4, the envelope sequence of these clones was more similar to that of the parental SIVsmE543-3 than to the consensus sequence of env in H635 plasma. Since missense mutations in other regions of the genome might be responsible for the lack of replication of the other clones in rhesus PBMC, we used one of the SIVsmE543-3-like clones as a basis to generate full-length clones that were representative of the consensus sequence of H635. Additional clones were constructed by replacing the SfuI-BglII region (vpx-to-env region) of SIVsmH635F-L3 with this fragment from other subclones. In addition, a true consensus clone, SIVsmH635FC, was constructed using portions of three subclones. Five of the eight resultant clones, SIVsmH635SB2, -SB5, -SB10, -SB11, and -FC, replicated in rhesus PBMC. The sequences of these clones were much closer to the H635 consensus sequence than to those of SIVsmH635F13 and SIVsmH635F-L3 (Fig. (Fig.4).4). They had seven characteristic mutations in the env gene, although SIVsmH635SB11 had P430T rather than P430S. SIVsmH635FC had only two minor mutations in gag and pol, and other regions were completely identical to the consensus of H635 at 9 wpi. Western blot analysis confirmed that each of the H635 infectious clones was indistinguishable from the parental SIVsmE543-3 in terms of expression of viral proteins and processing of gp160 (data not shown). A total of seven infectious clones were obtained (two clones by the first screening and five clones by the second), but five clones were used for further analysis because SIVsmH635F13 and SIVsmH635SB11 replicated very inefficiently in rhesus PBMC (data not shown).
Replication of H635 clones was examined in various types of cells, including the human T-cell lines CEMx174 and PM-1. CEMx174 is a commonly used target cell for propagation of SIVsm/mac viruses but does not express CCR5, which is the main coreceptor for SIV, whereas PM-1 cells express CCR5 (11). The SIVsmE543-3-like clone SIVsmH635F-L3 replicated with similar or slightly delayed kinetics in CEMx174 and PM-1 cells compared to SIVsmE543-3 (Fig. (Fig.5).5). The other four H635 clones, which were closer to the consensus sequence in H635 plasma, did not replicate in CEMx174 cells over a 35-day period. However, as shown in Fig. Fig.5B,5B, these four clones replicated with extremely slow kinetics in PM-1 cells (peaking by25 to 30 days postinfection, versus 10 to 14 days for SIVsmE543-3 and SIVsmH635F-L3). Peak RT activity was also consistently lower in PM-1 cells infected with either of these four 635 clones.
Replication of H635 clones in primary rhesus PBMC and macrophages was then evaluated. As shown in Fig. Fig.6,6, SIVsmH635F-L3 replicated to a higher titer than SIVsmE543-3 in PHA-activated rhesus PBMC, although both reached peak RT levels by 3 to 6 days postinfection. In contrast, the other RP clones replicated with slower kinetics than SIVsmE543-3 and SIVsmH635F-L3, with RT activity peaking by 9 to 15 days (Fig. (Fig.6A).6A). While SIVsmE543-3 and SIVsmH635F-L3 replicated efficiently in rhesus PBMC, regardless of the donor, replication of SIVsmH635FC, which was used as a representative for RP clones, was variable (Fig. (Fig.6B),6B), depending upon the specific macaque donor. Even in susceptible donors, the replication was delayed compared with that of SIVsmE543-3 and SIVsmH635F-L3.
As shown in Fig. Fig.7,7, replication of H635 clones was assessed in both monocyte-derived macrophages and alveolar macrophages. SIVsmE543-3, the parental virus, replicated productively in MDM as previously reported (18) but was variable in AM, depending upon the donor (Fig. (Fig.7).7). We therefore used the macrophage-tropic SIVmac316 as a positive control in AM since it replicated more consistently in this cell type. Replication of clones from H635 was extremely inefficient in both MDM and AM (Fig. (Fig.7).7). Slight replication of H635 clones was observed frequently in MDM and rarely in AM, but the level ofreplication was much less than that of SIVsmE543-3 or SIVmac316.
In summary, viruses with RP-specific mutations appeared to replicate inefficiently in all cell lines and primary cell types tested, including rhesus PBMC and macrophages, compared with SIVsmE543-3. This result, although unexpected, is consistent with the disadvantage of RP-specific mutations in rhesus PBMC cultures in vitro, which was shown by the loss of RP-specific mutations during the isolation of viruses from RP macaque H445 (Table (Table22 and Fig. Fig.3).3). The disadvantage of RP-specific mutations was generally observed in all the cells tested, including macrophages and human cell lines. Clearly, the genotype similar to that of SIVsmE543-3 would have a selective advantage over clones with RP-specific envelope mutations in vitro.
The GHOST assay was performed to determine the coreceptor usage of clones from H635 (Fig. (Fig.8A).8A). As expected, the parental virus, SIVsmE543-3, used CCR5 and GPR15 as coreceptors (11). In contrast, the SIVsmE543-3-like clone, SIVsmH635F-L3, used only CCR5 as a coreceptor. This was also the case with the other H635 clones with RP mutations (5.6 to 6.3% GFP-positive cells). This may explain the lack of replication of these clones in CEMx174 cells, which do not express CCR5, though SIVsmH635F-L3 can replicate in CEMx174 despite the lack of GPR15 usage. In addition, RP clones showed a low level of infectivity (0.3 to 1.3%) in all the GHOST cell lines regardless of coreceptor expression. Significant viral replication was observed in RP clone-infected GHOST parental cells over a period of 12 days (Fig. (Fig.8B).8B). The replication of RP clones in GHOST parental cells was not affected by the CXCR4-specific inhibitor AMD3100, while the replication of HIV-1NL43, which uses CXCR4 as a coreceptor, was completely blocked by AMD3100 (Fig. (Fig.8C).8C). This result indicated that SIVsmH635FC did not use CXCR4 expressed on the surface of GHOST parental cells at a low level and suggested the use of an alternative coreceptor which was expressed on GHOST parental cells. The use of an alternative coreceptor was not advantageous in the replication of RP clones in human cell lines, rhesus PBMC, and macrophages (Fig. (Fig.55 to to7)7) but may broaden the in vivo cell tropism of RP variants.
Our recent studies have been directed at defining the roles of host and virus factors in the development of rapid progression. In the present study, we generated full-length infectious molecular clones that were representative of the spectrum of genetic RP-specific variants in RP macaques. Despite their obvious growth advantage in vivo, as evidenced by high plasma viremia and high levels of viral RNA expression in tissues, these mutations were disadvantageous for growth of virus in tissue culture. Thus, propagation of uncloned virus from an RP macaque resulted in the loss of RP-specific mutations. However, sequence analysis of sequential variants in an RP macaque inoculated with this isolate demonstrated reselection of RP mutations in vivo. This result indicates that RP mutations are a characteristic of end-stage viruses in RP macaques, but the dominance of the RP genotype is not correlated with the ability to replicate in human cell lines or primary macaque cells in vitro. The reselection of RP mutations, which were a minor population within the inoculum, suggests a strong positive selection pressure for this genotype in RP macaques.
One of the potent selective pressures on viruses is immune responses, but immune responses in RP macaques have generally waned by 4 weeks postinfection and thus are unlikely to be responsible for the selection of RP-specific mutations (22). Actually, the pattern of sequence variation of gp120 in RP macaques differed from that previously observed in conventional SIV-infected macaques (8). Variation was concentrated to variable regions, designated V1, V2, V3, V4, and V5, in conventional SIV infection that induces immune responses (5, 27), but these variable regions were conserved in viruses from RP macaques. In addition, some of the RP mutations were found in highly conserved regions, such as the GDPE motif and the V3 loop analog, which is the region homologous to the V3 loop in HIV-1 (7, 57). Taken together, these RP mutations do not appear to be the consequence of immune pressure.
A few prior studies have characterized infectious clones with RP-specific mutations that were derived from SIV-infected macaques (2, 21, 41), with a specific focus on analyzing the viral determinants of macrophage tropism and neurotropism. Three such infectious clones are SIVmac316 (41), SIVmac17E-Fr (2), and SIVsm62d (21). Due to the propensity for macrophage-associated disease in RP macaques, these clones were derived from RP macaques. However, although RP mutations were dominant in the tissues of these macaques, many of the RP mutations we have observed were not found in these infectious clones. Thus, SIVmac316 did not show the loss of the glycosylation site in V1/V2, P334L/R in the V3 analog, and D385N in the GDPE motif, although these mutations were highly dominant in macaque 316 (29, 41). Anderson et al. constructed molecular clones from two SIVmac239-infected macaques with SIV encephalitis in which loss of the glycosylation site in V1/V2 and D385N in the GDPE motif were dominant. However, all the clones with the D385N mutation were not infectious, and only SIVmac17E was shown to be infectious (2). Finally, eight infectious clones were obtained from a SIVsm-infected RP macaque and designated as SIVsm62A through -J (21). Although scattered mutations in the V1/V2 glycosylation site, V3 analog, and GDPE motif were observed, these specific clones showed low infectivity and narrow in vitro tropism. These previous studies indicated that RP mutations, especially D388N (D385N in SIVmac239) in the GDPE motif, were selectively excluded, because of extremely low in vitro infectivity. In the present study, we successfully constructed infectious clones with the constellation of characteristic RP-specific mutations, i.e., loss of the potential glycosylation site in the V1/V2 region (N158D), substitutions in the V3 analog (P337T, E340K, and R348W), and the D388N substitution in the GDPE motif. Although these viruses replicated inefficiently in vitro, they were still infectious in primary rhesus PBMC and will allow us to evaluate their properties in vivo.
Two of the motifs that are altered in RP viruses are highly conserved in SIV and/or HIV. Although the V3 loop is variable in HIV-1, it is highly conserved among SIVmac/sm viruses; it is involved in coreceptor recognition (7, 57), and changes in this region alter the tropism of both of these viruses (21, 25, 28, 53). The GDPE motif is highly conserved in both HIV-1 and SIV. Previous studies have shown that the GDPE motif is important for CD4 binding (45), appearing to bind to the CD4 receptor molecule directly (31). Mutational analyses revealed that D368N in the HIV-1 gp120 (GGN368PE) and the analogous D385N substitutions in SIVmac impaired CD4 binding and viral infectivity (43, 45, 47). Another related mutation in this motif, G383R (R383GDPE) in SIVmac, also reduced CD4 binding and viral infectivity (47). These two mutations were also suggested to be critical for CD4-independent use of CCR5 as a coreceptor for SIV (47), which is a characteristic of RP variants (8, 12, 51). RP clones constructed in this study had the D388N substitution in the GDPE motif (GGN388PE) but retained infectivity in rhesus PBMC, although the replication was low and variable depending on the donor macaque. Therefore, the loss of infectivity of the GGN385PE mutant of SIVmac239 suggests that other compensatory mutations may be required for replication of viruses with the GGN385PE mutation.
Despite changes in the V3 loop analog, all of our clones retained the ability to use CCR5 as their major coreceptor. This was not unexpected, since envelopes of SIVsm62 clones with similar substitutions in the V3 loop retained the ability to use CCR5 in cell fusion assays but demonstrated relative CD4 independence (12). CD4 independence of the clones from H635 was not examined, but since the substitutions are similar to those observed in previous studies (8, 12, 51), we assume that this is a conserved property of RP clones. The RP clones in the present study lost the ability to use GPR15 as a coreceptor, a conserved property of many SIVsm and SIVmac viruses, and had gained the ability to replicate (albeit to low levels) in parental GHOST cells that expressed CD4 but lacked other known coreceptors of HIV and SIV. This finding suggested the use of an alternative coreceptor in addition to CCR5. In terms of in vitro properties, the ability to use an alternative coreceptor was the only potential advantageous feature of these RP clones that was observed in this study. The use of an alternative coreceptor may be related to the donor-dependent replication in PBMC, which was observed in this study, and may be advantageous for in vivo replication after depletion of memory CD4+ cell subsets, in which most of CCR5+ CD4+ cells are included. However, the significance of this finding will be unclear until the actual molecule is identified. Further analysis is required to identify the cell types which express this unknown coreceptor and whether cells infected in vivo express this coreceptor. The implication of this finding is that RP variants may be capable of infecting CD4+ T cells that do not coexpress CCR5, in addition to their ability to infect cells that only express CCR5 in the absence of CD4, thus potentially broadening the range of susceptible cells in macaques.
In the present study, full-length infectious clones of SIV that are representative of the consensus sequence of SIV in the plasma of an RP macaque were generated. These clones are thus likely to be characteristic of replicating virus in this macaque. We then evaluated the replication properties in primary macaque cells in vitro as an indirect measure of their in vivo tropism and replicative efficiency. In the absence of immune pressure, the efficiency of replication and the breadth of cellular tropism may be critical for the success of the virus. Thus, we expected an advantage of RP variants in the replication efficacy or cell tropism that would explain the high viral load in plasma and tissues of RP macaques (22). In terms of tropism, macrophage tropism would be highly advantageous in a host after depletion of the classic target cells, CD4+ memory T cells. Consistent with this theory, SIV-producing macrophages were predominant in H635 and other RP macaques (18, 22). SIV-producing macrophages and multinucleated giant cells were commonly observed in the lungs and brains of RP macaques by immunohistochemistry and in situ hybridization analyses (18, 22; C. R. Brown, unpublished data). Accordingly, we predicted that RP-specific envelope mutations might be responsible for the enhanced replication in tissue macrophages at the expense of replication in CD4+ T cells. However, the cloned viruses from H635 replicated less efficiently in macaque macrophages (MDM and AM) in vitro than the parental strain. Thus, the selection of RP mutations in vivo did not correlate with enhanced replication of these variants in macrophages in vitro. This lack of correlation between in vitro and in vivo tropism was not entirely unanticipated, since studies with SIVmac239 and SIVmac316 have demonstrated that viral tropism for macrophages in vitro was not reflective of the subsequent replication in macrophages in vivo (4). However, it is clear that the dominance of the RP genotype in vivo is not correlated with the ability to replicate in macrophages in vitro.
RP clones replicated less efficiently in primary macaque PBMC and macrophages than their parent, SIVsmE543-3. The low in vitro infectivity of RP clones is an enigma, since these viruses replicate to high levels in the RP macaques from which they were derived, as evidenced by high plasma viral RNA loads and large numbers of SIV RNA-expressing cells in tissues (22). One possible key to this discrepancy between in vitro and in vivo infectivity may be the profound early immune disorder in RP macaques (22), which may create a specific microenvironment that promotes the growth of RP variants. Supporting this hypothesis, high levels of proinflammatory cytokines such as IL-6 or chemokines such as monocyte chemoattractant protein-1 have been observed in RP macaques with SIV encephalitis (35). Although these may be the product of infected macrophages, alternatively they may be driving the activation of macrophages and potentially promoting the growth of RP variants. The in vivo target cells for SIV infection, including macrophages, may exhibit altered susceptibility compared with that observed in culture systems in vitro. We consider it likely that tissue culture systems may not be reflective of the in vivo replication potential of RP variants.
The predominance of the RP-specific genotype in different RP macaques suggests a direct role of these variants in the pathogenesis of rapid progression. However, the poor replicative ability of these variants in primary cells in vitro was unexpected. If growth in primary macrophages and PBMC in vitro is reflective of in vivo replication potential, these variants may not play a direct role in inducing the early immune failure that characterizes RP macaques. Instead, these variants may have evolved specifically to replicate efficiently in the absence of immune pressure in a host with a paucity of memory CD4+ T cells. We consider it likely that the early immune damage observed in RP macaques may be the direct result of the parental virus, which efficiently replicates in memory CD4+ T cells. Such early immune dysfunction may be critical for the subsequent evolution and development of RP-specific variants. This theory is consistent with the result of genotype analysis of sequential plasma samples in H635, since the parental virus predominated early in infection coincident with the loss of memory cells and RP variants did not appear until 4 weeks. RP-specific variants may be linked with the development of macrophage-specific disorders such as SIV encephalitis but could play little direct role in the early immune events. This two-step process is reminiscent of the rapid, consistent SIV encephalitis model system using SIVsmB670 and the neurotropic SIVmac17E strain (59), which essentially produces a high frequency of rapid progressors. Future studies will address the role of RP variants cloned from H635 in vivo in naive macaques, alone or in combination with the parental strain.
In summary, RP variants appear to be differentially selected in vivo in RP macaques specifically but are relatively unfit for replication in primary macaque cells in vitro. The infectious molecular clones obtained in the present study will be useful for in vivo studies with macaques to investigate the contribution of the RP variants to rapid progression and to understand the mechanisms underlying SIV encephalitis and other SIV-associated macrophage disorders.
We thank Russell Byrum (Bioqual Inc.) for animal studies, Que Dang (LMM, NIAID) for discussions, and Sonya Whitted and Robert Goeken (LMM, NIAID) for technical assistance.
This work was supported by the intramural program of the National Institute of Allergy and Infectious Diseases.