INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Human immunodeficiency virus type 1 (HIV-1) infection of people results in a gradual loss of CD4⁺ T lymphocytes and immunological competence (52) well as other specific systemic complications that include encephalopathy (14, 35, 41, 55), interstitial pneumonia (36, 47, 60), and nephropathies. The most common nephropathy is known as HIV-associated nephropathy (HIVAN) (3, 4, 6, 9, 20, 43, 44, 53).

Renal failure in HIVAN is associated with enlargement of the kidneys and is characterized by focal segmental glomerulosclerosis (FSGS) with proliferation of mesangial cells, increased mesangial matrix, mesangial hyperplasia, vacuolation of glomerular epithelial cells, and collapse of the glomerular capillary system (9). Associated with this glomerular pathology are the deposition of immunoglobulin G (IgG), IgM, and C3 (9). Histologic changes are observed in the tubulointerstitium and include dilation of renal tubules, cast formation, tubular necrosis, and interstitial nephritis (9). The interstitial nephritis is characterized by fibrosis and infiltration of mononuclear cells.

Similar to HIV-1 infection of humans, inoculation of simian immunodeficiency virus (SIV) into rhesus macaques results in AIDS, encephalopathy, and interstitial pneumonia (11, 31, 33, 39). While the development of neurological disease and interstitial pneumonia in HIV-1-infected patients is associated with the selection of macrophagetropic (M-tropic) variants (5), it is unclear whether development of HIVAN is associated with altered cell tropism of the virus. In a previous study, we showed that inoculation of rhesus macaques with the molecularly cloned, lymphocytetropic (L-tropic) SIV_mac239 resulted in renal pathology that was characterized by focal segmental and global glomerulosclerosis, increased immunoglobulin and collagen (both type I and type IV) deposition in the glomerulus, and mild azotemia in some macaques (16). The glomerular pathology correlated with the generation of M-tropic variants in these animals (16). In this study, we have examined whether rhesus macaques inoculated with pathogenic M-tropic SIV_macR71/17E, recovered from the brains of macaques with fulminant SIV-induced encephalitis (54, 56), would develop more severe renal disease than macaques inoculated with L-tropic SIV_mac239. Our results indicate that of the seven macaques inoculated with SIV_macR71/17E, six had significant renal pathology, five developed focal segmental and global glomerulosclerosis, and four exhibited moderate to severe azotemia. These results further extend the association of M-tropic variants of SIV_mac with the glomerular pathology and indicate that SIV_macR71/17E infection of rhesus macaques is a useful animal model for HIVAN in humans.

(This work was presented at the 30th Annual Meeting of the American Society of Nephrology, 2 to 5 November 1997, San Antonio, Tex.).

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Viruses and inoculation of animals. SIV_macR71/17E was prepared from pooled brain homogenates prepared from macaques R71 and 17E, both of which developed SIV-induced encephalitis (54). The M-tropic and neurovirulent properties of this virus stock have been previously described (54, 56). L-tropic SIV_mac239 was obtained from R. C. Desrosiers, New England Primate Center, Harvard University. The CEMx174 cell line (50) was used to prepare stocks of virus as described previously. CEMx174 cells were maintained in RPMI 1640 supplemented with 10 mM HEPES buffer (pH 7.3), 2 mM glutamine, 50 µg of gentamicin per ml, and 10% fetal bovine serum. One milliliter of the SIV_macR71/17E virus stock, with a titer of approximately 10⁴ 50% tissue culture infective doses (TCID₅₀) per ml, was used to inoculate rhesus macaques (Macaca mulatta) AQ12, AQ20, AQ38, AQ43, AQ47, AQ69, and AQ70 via the intravenous route. Macaques W, X, and Y were uninfected control macaques. The time of euthanasia was dictated by the severity of their disease to reduce unnecessary pain and suffering.

Virus burdens in macaques at necropsy and CD4 assays. For evaluation of the virus burdens at necropsy, heparinized blood was centrifuged to separate the plasma from the buffy coat. The cells were centrifuged through Ficoll-Hypaque density gradients to isolate peripheral blood mononuclear cells (PBMC). The isolated cells were used in infectious center assays (ICA) to determine the number of PBMC per 10⁶ that were producing infectious, cytopathic virus as previously described (21). The level of p27 antigen in the plasma was also evaluated at necropsy by antigen capture assays (Coulter Corp., Hialeah, Fla.). For enumeration of the CD4⁺ T cells at necropsy, isolated PBMC were incubated with a monoclonal antibody to CD4 (SIM.4; NIH AIDS Research and Reference Reagent Program). Cells were washed and stained with a fluorescein isothiocyanate (FITC)-conjugated goat anti-mouse IgG (DAKO Corp., Carpinteria, Calif.), fixed in 1% buffered formalin, and analyzed on an EPICS fluorescence-activated cell sorter.

Fractionation of glomerular and tubulointerstitial fractions. Freshly harvested kidney tissue was sieved for isolation of glomeruli and cortical tubules by the method of Savin and Terreros (51). Cortical tissue was minced into 1- to 3-mm tissue fragments with scissors and then with a scalpel blade while being maintained in ice-cold RPMI 1640. Glomeruli were first isolated by pressing the tissue through a 40-mesh stainless steel screen with the plunger from a 20-ml syringe. The screen was washed with 50 ml of RPMI 1640. This material, rich in both tubules and glomeruli, was sieved through a series of 100-, 150-, and 200-mesh screens, and the retained material was washed with 25 ml of RPMI 1640. In our hands, the material retained by the 150-mesh screens was rich in glomeruli, while the material passing through the 200-mesh screen contained mainly tubules and mononuclear cells. The glomeruli retained by the 150-mesh sieve were further purified by preparing a series of twofold dilutions in Dulbecco modified Eagle medium supplemented with 10% fetal bovine serum. These dilutions were plated into 24-well plates, and those wells containing only glomeruli and no tubules (determined by microscopy) were pooled and concentrated by centrifugation. The purity of these glomerular preparations was greater than 99%. The glomerular and tubulointerstitial fractions were retained for DNA isolation.

Renal histopathology. For light microscopy, slices of kidney tissue were fixed in 10% buffered formalin, embedded in paraffin, and sectioned (6 µm). Separate sections were stained with hematoxylin and eosin and periodic acid-Schiff reagent (PAS). Morphological alterations (described below) in the kidneys were semiquantitatively evaluated (scored 0 to 3) from coded slides by two investigators, and the mean value of the scores was used in statistical analyses. Mesangial cellularity was quantified per mesangial stalk (0 = <3 cells, 1 = 4 to 6 cells, 2 = 7 to 9 cells, 3 = >10 cells), examining 25 glomeruli per kidney. Mesangial matrix expansion was defined as a significant increase in matrix (0 = <5% of the glomeruli involved, 1 = 5 to 25% involved, 2 = 26 to 49% involved, 3 = 50% involved). Glomerular sclerosis was defined as glomerular stalks being mainly matrix, with few, if any, open capillaries; this could be segmental (not all stalks involved) or global (involving essentially all of the glomerular tuft). The percentage of sclerotic glomeruli was determined. The percentage of normal glomeruli (0 for mesangial matrix expansion and 0 for mesangial cellularity) was also determined. Tubulointerstitial mononuclear leukocyte infiltration was qualitatively assessed (0 = no infiltration, 1 = focal infiltration, 2 = multifocal infiltration, 3 = diffuse infiltration). A Nikon Optiphot microscope equipped with bright-field and epifluorescence capabilities was used for all microscopy. Photographs were taken with a Nikon Microflex HFM photomicroscopy system on Kodak Technical Pan 2415 or TMAX 400 film. Data were grouped into three categories: SIV-infected macaques that were PCR positive in the glomerular fraction (group 1; AQ12, AQ20, AQ47, AQ69, and AQ70), SIV-infected macaques that were negative for SIV in the glomerular fraction (group 2; AQ38 and AQ43) and healthy, uninfected macaques (group 3; W, X, and Y). Data were compared by analysis of variance (ANOVA).

Immunohistochemistry. Frozen sections of unfixed renal tissues were cut and fixed in acetone for immunohistochemistry studies. To identify mononuclear cell types (CD4⁺ T lymphocytes, CD8⁺ T lymphocytes, and CD68⁺ macrophages), sections were incubated with mouse monoclonal antibodies directed against CD4 (SIM.4; NIH AIDS Research and Reference Reagent Program) and human CD8 and CD68 (DK25 and EMB11, respectively; DAKO). Sections were incubated with diluted primary antibodies (1:50) overnight at 4°C. A Vectastain Elite ABC kit including a biotinylated goat anti-mouse antibody and diaminobenzidine with nickel intensification (Vector Laboratories, Burlingame, Calif.) was used to visualize binding of the primary antibody. Positive cells were counted by light microscopy using a 40× objective lens (0.044 mm² per field). Ten random fields were evaluated and quantitated to indicate the average number of positive cells per high-power field from coded slides. Control sections were incubated with normal mouse sera rather than primary antibody.

Immunoprecipitation studies. To demonstrate the presence or absence of antibodies directed against SIV_mac proteins, 10⁶ CEMx174 cells were inoculated with 10⁴ TCID₅₀ of SIV_macR71/17E. At 3 days postinoculation, cells were starved for methionine and cysteine and then radiolabeled with 1 mCi of [³⁵S]methionine and cysteine for 18 h. The culture medium was retained, and SIV proteins were immunoprecipitated by using serum samples (20 µl) obtained at necropsy and protein A-Sepharose as previously described (58). In additional experiments, KappaLock covalently bound to Sepharose 4B (Zymed Laboratories) was used in place of protein A-Sepharose to bring down immunoprecipitates. The advantage of KappaLock is that it will react strongly with the kappa light chain from all subclasses of IgG as well as from IgA and IgM (2). To rule out the possibility that antibodies directed against SIV_mac proteins existed as an immune complex and could not be detected by conventional immunoprecipitation assays, serum samples were treated at low pH (0.33 M citrate buffer [pH 2.0]) for 1 h, neutralized, and used immediately in immunoprecipitation assays. Immunoprecipitates were recovered by using KappaLock-Sepharose 4B as described above. All immunoprecipitates were washed three times in radioimmunoprecipitation assay buffer, and samples were denatured by boiling in sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) sample reducing buffer (32). Proteins were separated by SDS-PAGE (10% gel) and visualized by standard autoradiography techniques. Control antisera included a serum sample from an uninfected macaque and a serum sample from a SIV_mac239-infected macaque that had generated antibodies against the virus as previously described (62).

Renal function studies. Sera were collected at the time of inoculation and necropsy. Sera were assessed for serum urea nitrogen concentration (SUN) by using a colorimetric assay (kit 640, Sigma Chemical Co., St. Louis, Mo.). Urine collected at necropsy (when available) was assessed for protein by the Bio-Rad protein assay (Bio-Rad Laboratories, Richmond, Calif.).

PCR amplification of gp120 sequences from renal tissues. Total cellular genomic DNA was extracted from tubulointerstitial and glomerular fractions of kidneys as well as from lymph node tissue and then used as a template in nested PCR (48, 49) to amplify SIV gp120 sequences. The oligonucleotide primers used in the first round were 5'-GGCTAAGGCTAATACATCTTCTGCATC-3' and 5'-ACCCAAGAACCCTAGCACAAAGACCCC-3', which are complementary to bases 6565 to 6591 and 8179 and 8205, respectively, of SIV_mac239 (46). One microgram of genomic DNA was used in the PCR mixture, which contained 4.0 mM MgCl₂, 200 µM each of the four deoxynucleoside triphosphates, 100 pM each oligonucleotide primer, and 2.5 U of Taq polymerase (Perkin-Elmer Cetus, Norwalk, Conn.). The template was denatured at 92°C for 3 min, and PCR amplification was performed with an automated DNA Thermal Cycler (Perkin-Elmer Cetus) for 35 cycles of denaturation at 92°C for 1 min, annealing at 55°C for 1 min, and primer extension at 72°C for 3 min. Amplification was completed by incubation of the PCR mixture for 10 min at 72°C. One microliter from the 100-µl PCR mixture was used in a nested PCR performed under the reaction conditions described above. For the second round of amplification, the nested set of primers consisted of 5'-GTAAGTATGGGATGTCTTGGGAATCAG-3' and 5'-GACCCCTCTTTTATTTCTTGAGGTGCC-3', which are complementary to bases 6598 to 6624 and 8158 to 8184, respectively, of the SIV_mac239 genome (46). To confirm the specificity of the PCR products, the DNA in the gel was transferred onto nitrocellulose by the Southern technique and then hybridized with a ³²P-labeled gp120 probe generated by the random primer labeling method (15). Blots were washed for 30 min at 65°C with 2× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate)-0.5% SDS, 0.2× SSC-0.5% SDS, and finally 0.1× SSC-0.5% SDS. Blots were then exposed to Kodak X-Omat film and processed by standard autoradiographic procedures. As an internal control for the amplification process, DNA samples from all tissues were used in PCR with oligonucleotides that amplified the -actin gene as previously described (21, 38). The -actin gene was amplified from DNA samples from all tissues analyzed in this report (data not shown). Other controls done with each PCR included (i) a reaction control with no template, (ii) a reaction control with no oligonucleotide primers, (iii) a reaction control with a known template lacking SIV sequences (normal rhesus macaque kidney DNA), and (iv) a positive DNA control (a plasmid with the gp120 sequence) whose amplified products have been molecularly cloned and sequenced.

Construction of chimeric viruses. To determine if the gp120 sequences isolated from the cortical fractions of kidney from the macaques with the most severe glomerular disease would confer macrophage tropism onto L-tropic SIV_mac239, we constructed chimeric viruses in which the majority of the SIV_mac239 gp120 sequence (amino acids 107 to 490) was replaced with the corresponding region from gp120 sequences from AQ20 and AQ47. DNA sequences corresponding to the gp120 region of Env was amplified as described above and digested with NsiI and ClaI to release a 1,148-bp fragment encoding the variable (V1 to V5) regions of gp120. The DNA was gel purified and ligated into a pBS-derived plasmid containing the entire SIV_mac239 genome (Stratagene, La Jolla, Calif.) that was also digested with NsiI and ClaI to release a 1,148-bp fragment encoding the V1 to V5 regions of gp120. The plasmid was gel purified and ligated with the NsiI/ClaI fragment isolated from the pGEM3Zf() vector (Promega, Madison, Wis.) containing the appropriate V1 to V5 sequences. The resulting ligated DNA was used to transform Escherichia coli JM109, and the resulting transformants were screened for the presence of an NsiI/ClaI insert. To ensure that the insert corresponded to the appropriate sequence and that no premature termination codons were present in the DNA, the entire gp120 region was sequenced. Plasmids with the correct NsiI/ClaI fragments were used to transfect CEMx174 cells. Syncytial cytopathic effect was observed within 48 h of transfection and increased with continued incubation of cultures. Stock viruses were prepared at 7 days posttransfection and titrated in CEMx174 cells.

Assessment of macrophage tropism of chimeric viruses. The M-tropic nature of the various viral constructs was assessed by the criteria established for the L-tropic SIV_mac239 and a M-tropic SIV_macLG1 as reported previously (57). Monolayers of macrophage cultures in 35-mm-diameter dishes were washed three times with RPMI 1640, inoculated with 0.1 ml of undiluted virus stocks in 0.5 ml of macrophage differentiation medium (MDM), incubated for 2 h at 37°C, and then supplemented with 2 ml of fresh MDM and reincubated for up to 10 days. Macrophage tropism was assessed by assaying culture medium at 0, 2, 4, 6, 8, and 10 days for the presence of p27 core antigen. Controls included SIV_mac239 as an L-tropic virus and SIV_macLG1 as an M-tropic virus (57).

Statistical evaluation of data. ANOVA (one-way) was used to determine the statistical significance for data on histopathology (mesangial hyperplasia and expansion), immunohistochemical staining (p27), immunoglobulin and matrix staining, the density of cellular infiltrates, and SUN. Table 3 provides means and standard errors of the means for groups 1 to 3 for several parameters; Tables 2, 4, 5, and 6 provide the individual data derived from each macaque.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Virus infection in the macaques and kidney. The macaques used in this study, the lengths of time for which they were infected with SIV_macR71/17E, CD4⁺ T-cell counts, and virus burdens in the blood and other histopathological observations made at necropsy are listed in Table 1. The length of infection varied between 42 to 97 days, and all macaques were actively producing virus at the time of necropsy. All macaques developed typical SIV_mac disease, including lymphoid depletion of lymph nodes (in six of seven macaques), thymic atrophy (four of seven macaques), and gastrointestinal disease (seven of seven macaques). In addition, six macaques also developed interstitial pneumonia and/or encephalitis, characteristic of this M-tropic strain of SIV_mac. As previously shown (16), the number of circulating CD4⁺ T cells in the blood did not correlate with when the macaques became moribund and developed terminal SIV_mac disease. All seven macaques were actively producing virus from lymph nodes at the time of euthanasia (Table 1).

Renal pathology. In addition to typical SIV_macR71/17E-induced disease previously described, we report here on renal pathology associated with SIV_macR71/17E-infected macaques. The kidneys of macaques infected with SIV_macR71/17E exhibited both glomerular and tubulointerstitial lesions. For quantitative comparisons, macaques with glomeruli that were positive for SIV_mac sequences (group 1; AQ12, AQ20, AQ47, AQ69, and AQ70) were compared with the other macaques with glomeruli that were negative for SIV_mac sequences (group 2; AQ38 and AQ43) and healthy, uninfected macaques (group 3; W, X, and Y).

View larger version (203K): [in this window] [in a new window]   FIG. 1.   Renal histopathology in macaques infected with SIVmacR71/17E. (a) Light micrograph of AQ20 (a group 1 macaque) with a focus of inflammatory cell infiltrate (arrowhead), slightly dilated tubules with cast material (arrows), and glomeruli (g), many of which are sclerotic (magnification, ×25). (b) A globally sclerotic glomerulus from AQ20 in which mesangial matrix has completely replaced the glomerular capillaries (magnification, ×100). (c) A sclerotic glomerulus from AQ47 with peripheral glomerular capillary collapse and sclerosis associated with the core of each lobule (magnification, ×100). (d) A glomerulus from AQ70 (a group 1 macaque) exhibiting FSGS. The top lobe of the glomerulus has numerous capillaries, while much of the bottom lobe is composed of mesangial matrix (magnification, ×100). (e and f) Light micrographs of AQ43 (a group 2 macaque) with normal-appearing kidney parenchyma and glomeruli (magnifications, ×25 and ×100).

View larger version (164K): [in this window] [in a new window]   View larger version (193K): [in this window] [in a new window]   FIG. 2.   Electron microscopy of glomerular pathology in macaques infected with SIVmacR71/17E. (a) Transmission electron micrograph of a portion of a glomerulus from AQ20 (a group 1 macaque) with prominent mesangial cells (M) surrounded by excessive amounts of extracellular matrix including both fibrillar (large arrows) and fibrous (small arrows) collagenous material. Within some of the fibrillar matrix are foci of amorphous material which has an increased electron density consistent with deposits of immunoglobulin (arrowheads) (magnification, ×7,600). (b) Transmission electron micrograph of a glomerular capillary loop from AQ43 (a group 2 macaque) showing a paracrystalloid tubuloreticular inclusion (curved arrow) within the glomerular endothelial cells. These inclusions were evident in all SIV-infected macaques (magnification, ×3,800).

Collagen type I and IV reactivity in SIV_mac-infected macaques. Because an increased deposition of collagen is seen in sclerotic glomeruli in HIV-1 transgenic mice (30) and in the kidneys from macaques inoculated with L-tropic SIV_mac239 (16), the kidneys from SIV_macR71/17E-infected and control macaques were examined by immunohistochemistry for collagen types I and IV. As shown in Tables 3 and 4 and in Fig. 3, glomerular collagen type I and IV immunoreactivity was highest in those macaques exhibiting focal segmental and global glomerulosclerosis (group 1; AQ12, AQ20, AQ47, AQ69, and AQ70 [Fig. 1a and d]) but was also higher in the macaques without glomerular SIV (group 2; AQ38 and AQ43 [Fig. 1b and e]) than in the uninfected macaques (Fig. 1c and f). Similarly, tubulointerstitium staining of collagen I was highest in the group 1 macaques (Table 4), but staining for collagen I also appeared to be slightly higher in group 2 macaques than in the uninfected controls. Tubulointerstitium staining for collagen IV was also higher in group 1 macaques than in the uninfected control macaques (Table 4). These data indicate that interstitial fibrosis was occurring in the macaques with SIVAN, in a manner similar to that seen in humans with various glomerulopathies (59).

View larger version (176K): [in this window] [in a new window]   FIG. 3.   Increased type IV and I collagen deposition in macaques infected with SIVmacR71/17E. Acetone-fixed frozen sections from macaques AQ47 (group 1 macaque with severe glomerulosclerosis), AQ43 (group 2 macaque with minimal renal pathology), and Y (group 3, uninfected) were stained for collagen type IV (a to c) and type I (d to f) as described in Materials and Methods. (a) Immunofluorescence staining for collagen IV of a sclerotic glomerulus from AQ47 showing a relatively homogeneous, dense staining of collagen IV throughout the glomerulus; (b) immunofluorescence staining for collagen IV of a glomerulus from AQ43 showing a foci of dense staining of collagen IV within the glomerulus; (c) immunofluorescence staining for collagen IV of a normal glomerulus from uninfected macaque Y showing a relatively light homogeneous staining for collagen IV; (d) immunofluorescence staining for collagen I of a sclerotic glomerulus from AQ47 showing a relatively homogeneous, dense staining of collagen I throughout the glomerulus; (e) immunofluorescence staining for collagen IV of a glomerulus from AQ43 showing a foci of dense staining of collagen I within the glomerulus; (f) immunofluorescence staining for collagen I of a normal glomerulus from uninfected macaque Y to demonstrate the distribution of collagens from data in Tables 3 and 4. It can be seen that there was relatively little glomerular staining of collagen I compared to that evident in infected macaques. (Magnification of all panels, ×100.)

Immunoglobulin deposition in the glomeruli. Since HIVAN (and HIV immune complex disease) and SIVAN are associated with glomerular deposits of immunoglobulins, we examined the kidneys from SIV_mac-infected and control macaques for the presence of immunoglobulin deposits. As shown in Table 5, IgM immunoreactivity was present in the glomeruli of all SIV_macR71/17E-infected macaques, whereas none of the uninfected macaques exhibited glomerular IgM. The IgM immunoreactivity was localized mainly to the mesangial compartment as described for HIVAN (Fig. 4a). IgG immunoreactivity in the infected macaques (Fig. 4b), although slightly weaker than IgM reactivity, was significantly higher than in the uninfected controls (Fig. 4d). In addition, six of seven macaques demonstrated significant glomerular IgA immunoreactivity (Fig. 4c). Control sections incubated with normal rabbit serum showed no specific immunoreactivity. Interestingly, the macaque that survived the shortest period of time following inoculation with SIV_macR71/17E (AQ43) had no IgA immunoreactivity and the weakest IgG and IgM immunoreactivity. Unlike previous studies showing polyclonal gammopathy in the sera of SIV_mac-infected macaques (16), the results for the SIV_macR71/17E-infected macaques indicate decreased serum levels of IgG, elevated levels of IgM, and normal levels of IgA (data not shown). These results suggest that the increased IgG and IgA glomerular immunoreactivity was not due to nonspecific trapping of serum immunoglobulins in the widened mesangial interstitium. By electron microscopy, discrete deposits of immunoglobulin were evident predominantly in mesangial regions, with some deposits in the glomerular capillary wall (Fig. 2a).

View larger version (60K): [in this window] [in a new window]   FIG. 4.   Deposition of immunoglobulin in the glomeruli of macaques infected with SIVmacR71/17E. Acetone-fixed frozen sections from representative macaques were stained for the presence of IgG, IgM, or IgA as described in Materials and Methods. (a) Micrograph showing that IgG is present largely within the mesangial region of this sclerotic glomerulus from macaque AQ20. (b) Micrograph showing that IgM is localized largely to the mesangial region of this sclerotic glomerulus from AQ20. (c) IgA is present within this glomerulus from AQ20. (d) Micrograph showing that there is no IgG (nor IgM or IgA [data not shown]) in the glomeruli from uninfected macaque Y. (Magnification of all panels, ×100.)

SIV_mac p27 antigen is detected in glomeruli from some macaques inoculated with SIV_macR71/17E. We determined if the SIV_mac core antigen (p27) could be localized to certain regions of the kidney. By immunohistochemistry, SIV_mac p27 antigen was detected in the glomeruli but not the tubulointerstitium from four of the five macaques which were positive for glomerular SIV by PCR (Fig. 5a and b). Further, the immunostaining for p27 was most intense in the glomeruli from macaques AQ20 and AQ47, which exhibited the most severe focal segmental and global glomerulosclerosis (Tables 2 and 5). Core antigen was not detected in those macaques whose glomeruli were negative for SIV by PCR (group 2) or in the uninfected macaques (group 3). The p27 antigen did not colocalize with cells that were positive for staining with anti-CD68 antibody (Fig. 6b) or with areas staining for IgG (Fig. 5c). Also shown in Table 5 are the levels of p27 antigenemia at necropsy. These results indicate that the detection of p27 antigen in the glomeruli was not the result of trapping of plasma p27, since macaques negative for glomerular p27 staining (AQ12, AQ38, and AQ47) also had extremely high levels of p27 antigen in the plasma. These results indicate a strong correlation between the presence of viral sequences in the glomeruli, the presence of viral p27 antigen, and glomerular pathology. These results also indicated that a productive viral infection was present in the glomeruli of at least four of the five macaques in group 1. 

View larger version (34K): [in this window] [in a new window]   View larger version (67K): [in this window] [in a new window]   FIG. 5.   Glomerular SIVmacp27 antigen in macaques infected with SIVmacR71/17E. Acetone-fixed frozen sections of kidney were prepared and stained for the presence of p27 antigen as described in Materials and Methods. (a) Low-magnification immunofluorescence micrograph of AQ47 showing that the glomeruli are the only structures stained (magnification, ×25). (b) Higher magnification of the same section showing that p27 staining is relatively uniform throughout the glomerulus. The oval dark regions appear to be the nuclei of glomerular cells. This pattern of staining is in sharp contract to the multifocal nature of the staining for macrophages (compare to the CD68+ cells in Fig. 6b). Therefore, the p27 staining appears to be in more cells than can be explained on the basis of resident glomerular macrophages (magnification, ×100). (c) The glomerulus shown in panel b was stained for IgG and shows the diffuse but localized deposits of IgG. Since the patterns of distribution of p27 and IgG do not appear to parallel each other, it is unlikely that the p27 staining can be explained solely on the basis of glomerular deposition of IgG-p27 immune complexes (magnification, ×100).

Renal function. SUNs were elevated in group 1 macaques with evidence of glomerular SIV by PCR and immunohistochemistry (Table 3). The average SUN in the group 1 macaques was 63 ± 14.6 mg/dl (range, 28.8 to 112.4 mg/dl), compared to 24.7 ± 0.8 and 26.8 ± 3.3 mg/dl for the group 2 and group 3 macaques, respectively. These results indicate that mild to severe azotemia was present in the macaques in group 1. Further, those macaques exhibiting the most extensive focal segmental and global glomerulosclerosis, AQ20 and AQ47, had the highest SUNs, 112.4 and 76 mg/dl, respectively. The macaques with the most severe glomerulosclerosis, AQ20 and AQ47, were also evaluated for proteinuria. Both AQ20 and AQ47 had proteinuria (2 and 1.5 mg/ml, respectively), while macaque AQ43, which had no glomerulosclerosis, had undetectable amounts of protein in the urine.

Composition of mononuclear infiltrates in the glomeruli and interstitium. Glomerular and interstitial inflammatory cell infiltrates typically occurs with viral infections of the kidney. Glomerular and tubulointerstitial infiltrates were quantitated in SIV_mac-infected and noninfected macaques. Very few CD4⁺ lymphocytes were evident in glomeruli and interstitium from infected or noninfected kidney tissue (Table 6) despite the finding that all macaques except AQ20 had CD4⁺ T-cell levels above 500 cells/µl at necropsy (Table 1). There were variable numbers of CD8⁺ lymphocytes in infected and noninfected kidneys (in glomeruli or interstitium), and the numbers of cells were not statistically different between the three groups of macaques (Table 6). The numbers of CD68⁺ cells (monocytes/macrophages) were statistically increased in the glomerular compartment of macaques in group 1 compared to group 2 and 3 macaques (Table 6; Fig. 6). There were increased tubulointerstitial CD68⁺ cells in the two macaques (AQ20 and AQ47) which had the most severe renal disease.

View larger version (106K): [in this window] [in a new window]   FIG. 6.   Macrophage (CD68+ cells) are present in the kidney of macaques infected with SIVmacR71/17E. Acetone-fixed frozen sections from the kidneys of macaques AQ20 (with severe glomerulosclerosis) and AQ43 (with minimal renal pathology) were stained for the presence of CD68+ cells as described in Materials and Methods. (a) Low-magnification micrograph of kidney tissue from AQ20 stained for CD68+ cells (monocytes/macrophages) in which a few glomeruli (arrowheads) are evident. There are a number of CD68+ cells (black foci) scattered throughout the parenchyma (magnification, ×25). (b) Higher magnification of a glomerulus from AQ20 showing several CD68+ cells (magnification, ×100). (c) Low-magnification micrograph of kidney from AQ43 stained for CD68+ cells (monocytes/macrophages). A few glomeruli (arrowheads) are evident, as are a number of CD68+ cells (black foci) scattered throughout the parenchyma (magnification, ×25). (d) Higher magnification of a glomerulus from AQ43 with only a few CD68+ cells (magnification, ×100).

Immune complexes are not responsible for the observed renal pathology. We determined whether SIV_macR71/17E-infected macaques generated antibodies during the course of infection. Serum samples obtained at necropsy were used in immunoprecipitation studies with radiolabeled SIV_macR71/17E (Fig. 7). The results indicate that of the seven macaques inoculated with SIV_macR71/17E, only two (macaques AQ69 and AQ70) developed antibodies against SIV_mac proteins. Because protein A-Sepharose preferentially binds to IgG subclasses 1, 2, and 4 but not to IgG subclass 3 and only weakly to IgA and IgM, the same experiment was performed with a KappaLock bound to Sepharose 4B, which binds to the kappa light chains from IgG, IgA, and IgM. The results obtained were identical to those seen in with protein A-Sepharose (data not shown). Because all seven macaques had high levels of plasma p27, we determined if circulating immune complexes were preventing the detection of antibodies in our immunoprecipitation assays. Serum samples were pretreated at low pH to disrupt antigen-antibody complexes, neutralized, and immediately used in immunoprecipitation assays. Again, the results were identical to those shown in Fig. 7 (data not shown). Taken together, the results from these three experiments indicate that five of the seven macaques did not develop antibodies against SIV_mac proteins.

View larger version (62K): [in this window] [in a new window]   FIG. 7.   Immunoprecipitation of SIVmac proteins with serum samples taken from macaques at necropsy. SIVmacR71/17E (104 TCID50) was used to inoculate 2 × 106 CEM174 cells. At 4 days postinoculation, cells were starved for methionine and cysteine and then radiolabeled with 1,000 µCi of [35S]methionine and cysteine for 18 h. The culture supernatant was retained, and SIV proteins were immunoprecipitated with 10 µl of each serum sample and protein A-Sepharose as described in the text. Immunoprecipitates were washed three times in radioimmunoprecipitation assay buffer, samples were denatured by boiling in SDS-PAGE sample reducing buffer, and proteins were separated by SDS-PAGE (10% gel). Proteins were visualized by standard autoradiographic techniques. Lanes: 1, SIV proteins immunoprecipitated from a macaque infected with SIVmac239; 2, SIV proteins immunoprecipitated with a serum from an uninfected macaque; 3, SIV proteins immunoprecipitated with serum from macaque AQ70; 4, SIV proteins immunoprecipitated with serum from macaque AQ69; 5, SIV proteins immunoprecipitated with serum from macaque AQ47; 6, SIV proteins immunoprecipitated with serum from macaque AQ43; 7, SIV proteins immunoprecipitated with serum from macaque AQ38; 8, SIV proteins immunoprecipitated with serum from macaque AQ20; 9, SIV proteins immunoprecipitated with serum from macaque AQ12.

Chimeric viruses constructed with the V1 to V5 gp120 sequences isolated from glomerular fractions of AQ20 and AQ47 are M-tropic. To confirm the M-tropic phenotype of the viruses in the glomeruli, we constructed chimeric viruses in which the variable (V1 to V5) region of SIV_mac239 gp120 was replaced with the corresponding regions from gp120 amplified from the glomerular fractions isolated from macaques AQ20 and AQ47, the two macaques with the most severe glomerular disease. Five chimeric viruses were prepared from the gp120 region amplified from the glomerular fractions of AQ20 and AQ47; the results from representative clones from each glomerular fraction are shown. The chimeric viral genomes SIV_macAQ20GLO and SIV_macAQ47GLO were transfected into cultures of CEMx174 cells to prepare stock viruses. Both chimeric viruses replicated efficiently in CEMx174 cells and caused syncytial cytopathology in cultures (Fig. 8A). Viruses were then used to infect rhesus macrophage cultures, which were assayed for the presence of p27 at several times after inoculation. The results shown in Fig. 8B indicate that L-tropic SIV_mac239 infected rhesus macrophage cultures very poorly, whereas cultures infected with SIV_macLG1 (a molecularly cloned M-tropic virus) released large amounts of p27 into the culture medium over the 10-day period. Cultures inoculated with SIV_macAQ20GLO and SIV_macAQ47GLO exhibited a growth curve similar to that of M-tropic SIV_macLG1 (Fig. 8B). Similar results were obtained for the other four chimeric viruses prepared from the glomerular fraction from each macaque. These data indicate that the predominant viral species in the glomerular fractions from AQ20 and AQ47 had an M-tropic phenotype.

View larger version (17K): [in this window] [in a new window]   FIG. 8.   Interviral recombinants constructed with the gp120 regions isolated from the glomerular fractions of macaques AQ20 and AQ47 are M-tropic. Interviral recombinants were constructed as described in Materials and Methods and used to inoculate either CEMx174 or rhesus macaque macrophage cultures. (A) Growth curves of SIVmac239, SIVmacR71/17E, SIVmacAQ20GLO, and SIVmacAQ47GLO in CEMx174 cultures. CEMx174 cells (106 cells/culture) were inoculated with 1,000 TCID50 of each virus (multiplicity of infection of approximately 0.001) for 24 h, washed three times to remove the virus inoculum, and maintained in the appropriate medium for the course of the infection. Culture medium was harvested at the time points indicated and assayed for the presence of p27 antigen as described in Materials and Methods. , SIVmac239-inoculated CEMx174 cultures; , SIVmacR71/17E-inoculated CEMx174 cultures; , SIVmacAQ20GLO-inoculated CEMx174 cultures; , SIVmacAQ47GLO-inoculated CEMx174 cultures. (B) Growth curves of SIVmac239, SIVmacLG1, SIVmacAQ20GLO, and SIVmacAQ47GLO in macrophage cultures. Rhesus macrophages in 35-mm-diameter dishes were prepared as described earlier (58). All cultures were inoculated, washed, and maintained as described above. Culture medium was harvested at the time points indicated and assayed for the presence of p27 antigen by using antigen capture assays (Coulter Corp.). , SIVmac239-inoculated rhesus macrophage cultures; , SIVmacR71/17E-inoculated rhesus macrophage cultures; , SIVmacAQ20GLO-inoculated rhesus macrophage cultures; , SIVmacAQ47GLO-inoculated rhesus macrophage cultures.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

In this study, we showed that five of seven macaques infected with M-tropic SIV_macR71/17E rapidly developed glomerulosclerotic lesions similar to those in HIVAN (reviewed by Pardo et al. [40]). The lesions found in SIVAN were also very similar to glomerular lesions found in mice expressing an HIV transgene (12, 28, 29). The glomerular changes seen in both HIVAN and SIVAN described herein include glomerular capillary collapse, mesangial cell hyperplasia, focal and global glomerulosclerotic lesions with increased extracellular matrix and immunoglobulin deposits within the mesangial (and to a lesser extent capillary) matrix, and endothelial tubuloreticular inclusions. However, there are some dissimilarities between the SIVAN induced by SIV_macR71/17E and HIVAN in humans, including the paucity of SIV_macR71/17E-induced tubulointerstitial changes (with very few small foci of interstitial inflammatory cells, sparse evidence of microcystic change, and limited evidence of interstitial fibrosis). Previously, we described more prominent tubulointerstitial pathology in the kidneys of macaques infected with L-tropic SIV_mac239 (16). However, SIV_macR71/17E appeared to cause more glomerulosclerosis and less tubulointerstitial pathology, suggesting that the tubulointerstitial changes are not critical to the development of the glomerulopathy and renal failure in SIVAN. These results also suggest that M-tropic SIV_macR71/17E is a more glomerulopathic virus whereas L-tropic SIV_mac239 induces more tubulointerstitial pathology. The glomerulopathic nature of this virus correlated with renal function data, which indicated that four of seven macaques inoculated with SIV_macR71/17E developed moderate to severe azotemia. Furthermore, the two macaques with the most severe focal segmental and global glomerulosclerosis, AQ20 and AQ47, developed proteinuria, a further indication of renal dysfunction. These results contrast with those obtained for macaques inoculated with SIV_mac239, which developed only mild azotemia.

We examined the potential relationship(s) between the presence of viral p27 and the numbers of CD68⁺ cells in the glomeruli from infected macaques. While all SIV_macR71/17E-infected macaques had high levels of plasma p27 antigen, not all were positive for glomerular p27 antigen (Table 5). We also found that those macaques with glomerulosclerosis also had the highest numbers of CD68⁺ cells in the glomerulus (Table 3). We found a statistically significant relationship (by regression analysis) between the number of CD68⁺ cells in the glomerulus and the relative staining for p27 but not between the localization of p27 and CD68⁺ cells (Fig. 5b and 6b). Since immunohistochemistry revealed viral p27 throughout the glomerulus (Fig. 5b), the focal nature of the CD68⁺ cells (Fig. 6b) could not account for all of the viral p27 antigen. The finding of increased numbers of CD68⁺ cells in the glomeruli with sclerosis suggests that these cells may be responsible for ferrying virus into the glomerulus and/or for exacerbating the viral infection within the glomerulus. Similarly a recent study has also implicated macrophage infiltration in the pathogenesis of HIV focal segmental glomerulosclerosis (1). However, the present study indicates that CD68⁺ cells are not the only cell type in the glomerulus infected with the virus. The likelihood that SIV_macR71/17E infects other cell types in the glomerulus (renal endothelial, epithelial, and mesangial cells) is supported by studies showing that HIV-1 can infect human renal endothelial, epithelial, and mesangial cells (7, 18, 25, 45) and by our recent observations that primary cultures of glomerular and tubular epithelial cells from rhesus macaques can be infected with SIV_mac (55a).

We also performed studies examining the potential relationship between the presence of viral p27, immunoglobulin deposition, and the histopathologic changes in the glomeruli from SIV_macR71/17E-infected macaques. Our results indicate that the patterns of cellular distribution of p27 and IgG were not identical, suggesting that the presence of viral p27 was not due to the trapping of circulating immune complexes (Fig. 5). These observations suggest that the p27 present in the kidney was not the result of nonspecific deposition of immune complexes from the circulation and cannot be entirely accounted for by infected macrophages within the glomerulus. Our results indicated neither a relationship between the relative amount of p27 and glomerular immunoglobulin staining nor a relationship between the relative amount of glomerular immunoglobulin staining and glomerulosclerotic changes (Table 5). However, we found a statistically significant relationship (by regression analysis) between the amount of immunoglobulin deposition in the glomerulus and the degree of mesangial cell hyperplasia in the SIV_macR71/17E-infected macaques. Mesangioproliferative glomerulonephritis is known to occur secondary to anti-glomerular basement membrane disease (Goodpasture syndrome) and other conditions associated with antibody deposition within the glomerulus (17). Because mesangial hyperplasia seen in SIV_macR71/17E-infected macaques was seen only in the context of glomerulosclerosis, it may represent an intermediate histopathologic state in the progression toward glomerulosclerotic lesions, which are the lesions correlated with and responsible for the development of the renal dysfunction. However, the immunoglobulin deposition may represent the superimposition of immune complex glomerulonephritis on the glomerulosclerosis. Support for the relationship between mesangial hyperplasia and FSGS comes from one study that examined sequential biopsy samples from a HIV-1-infected patient and showed a diffuse mesangial hypercellularity that evolved to FSGS (10).

Polyclonal gammopathy and glomerular deposits of IgM are relatively common in HIV-infected patients, with and without renal disease (26, 27). Some studies have suggested that immune complex glomerulonephritis (HIV immune complex disease) is a common nephropathy reported for HIV-infected patients (27). In a previous study of four HIV-1 patients with renal disease, the antigen specificity of these renal antibodies was identified (27). Kidney-eluted IgA and/or IgG reacted with HIV p24 or gp120 in three patients, while in one patient no HIV protein-antibody complexes were identified. In acid-treated biopsy sections, p24 was identified by immunohistochemistry in the same way that we identified SIV p27 core antigen in glomeruli from SIV_mac-infected macaques (Fig. 5). Similar to what was found for HIV-1-infected patients, we demonstrated significant deposition of not only IgG and IgM but also IgA in the glomeruli from macaques infected with SIV_macR71/17E. This finding raised the question as to whether the deposition of immunoglobulin in the glomeruli was due to immune complexes formed between antiviral antibodies and viral antigen or due to nonspecific deposition. Previous studies have shown that following inoculation with certain strains of SIV_mac, macaques that develop an acute disease characterized by high virus burdens as measured by ICA and plasma p27 levels rarely develop antibodies to the virus (62). Thus, it was of interest to determine if there was an association between severe glomerular disease and the presence or absence of antiviral antibodies. Immunoprecipitation studies performed with serum samples obtained at necropsy indicated that only two of the seven macaques inoculated with SIV_macR71/17E had developed antibodies against the virus (Fig. 7). In addition, pretreatment of serum samples at low pH to disrupt antibody-antigen complexes prior to immunoprecipitation analysis did not identify anti-SIV antibodies, indicating that the inability to detect antibodies was not due to the presence of circulating antibody-antigen complexes. The lack of antiviral antibodies in five of seven macaques including AQ20 and AQ47, which had the most severe focal segmental and global glomerulosclerosis, indicates that the severe glomerular disease observed in these macaques was not associated with an immune complex-mediated disease involving antiviral antibodies. These results confirm our observations that the immunoglobulin deposition and presence of p27 in the glomerulus were not due to the trapping of antiviral immune complexes. Additionally, the detection of significant amounts of immunoglobulin in the glomeruli from infected macaques suggests that the deposition of immunoglobulin in the glomeruli was either nonspecific in nature or possibly against clinical or subclinical opportunistic infectious agents. The antigens to which these glomerular antibodies are directed have yet to be determined. The similarity between the glomerular pathologic effects observed in HIVAN and SIVAN described herein suggests that immune complexes probably do not play a significant role in the pathogenesis of HIVAN.

The results presented in this report both confirm and extend our findings that suggest that pathogenic M-tropic strains of SIV_mac reproducibly cause a glomerulosclerosis which is similar to the HIVAN observed in humans. In support of our studies presented here, a recent study that used a human kidney organ culture system showed productive replication by M-tropic HIV-1 (BaL strain) and not L-tropic HIV-1 (IIIB strain) (22). Similarly, the HIV MB strain (which has the characteristics of a M-tropic virus) infected glomerular endothelial and mesangial cells, whereas T-cell-tropic HIV (IIIB strain) did not infect mesangial cells (18). Numerous studies have mapped macrophage tropism of HIV-1 isolates to amino acid substitutions within the envelope glycoprotein of the virus (8, 19, 24, 34, 37, 61) and differential coreceptor usage by the virus (13). Recently, chimeric primate lentiviruses containing the tat, rev, vpu, and env of HIV-1 (strain HXB2) in a genetic background of L-tropic SIV_mac239 that cause severe CD4⁺ T-cell loss and AIDS, are M-tropic, and cause encephalopathy in rhesus macaques have been described (42). It will be of interest to determine if pathogenic molecular clones derived from these virus stocks are capable of inducing renal disease. Such molecular clones may permit the identification of HIV-1 genes and amino acid substitutions that are important in the genesis of primate lentivirus-associated nephropathy.