Abstract
ABSTRACT. Simian virus 40 (SV40), a monkey polyomavirus that is believed to have entered the human population through contaminated vaccines, is known to be renotropic in simians. If indeed SV40 is endemic within the human population, the route of transmission is unknown. It was therefore hypothesized that SV40 might be renotropic in humans and be detected more frequently in samples obtained from patients with kidney diseases. This study found that typical polyomavirus cytopathic effects (CPE) were present and SV40 T antigen was detected in CV-1 cells cultured with peripheral blood mononuclear cells (PBMC) or urinary cells obtained from patients with kidney disease and healthy volunteers. DNA sequences homologous to the SV40 viral regulatory genome were detected by PCR in urinary cells from 15 (41%) of 36 patients with focal segmental glomerulosclerosis (FSGS), 2 (10%) of 20 patients with other kidney diseases, and 1 (4%) of 22 healthy volunteers (FSGS compared with other glomerular disease, P < 0.02; FSGS compared with healthy volunteers, P = 0.003). SV40 viral regulatory region genome was detected from PBMC at similar frequencies in patients with FSGS (35%), other glomerular diseases (20%), and healthy volunteers (22%). SV40 genome was detected by PCR in kidney tissues from 17 (56%) of 30 of patients with FSGS and 4 (20%) of 20 patients with minimal change disease and membranous nephropathy (P < 0.01). Considerable genetic heterogeneity of the viral regulatory region was detected, which argues against laboratory contamination. SV40 genome was localized to renal tubular epithelial cell nuclei in renal biopsies of patients with FSGS by in situ hybridization. This study demonstrates for the first time that human kidney can serve as a reservoir for SV40 replication and that SV40 may contribute to the pathogenesis of kidney disease, particularly FSGS.
SV40, a monkey polyomavirus, was present in poliovaccines and adenovirus vaccines used between 1955 and 1963, and it is believed to have been introduced into the human population at that time (1,2). Several lines of evidence suggest that new SV40 infections may be occurring in the human population. David et al. (3) found that 16% peripheral blood lymphocytes from non-cancer patients were positive for SV40 genome, and Martini et al. (4) detected SV40 DNA sequences in PBMC from 25% of healthy volunteers and in sperm from 45% of healthy volunteers. On the other hand, Shah et al. (5) was unable to identify SV40 genome in the urine of homosexual men. Recovery of infectious SV40 from human patients has been reported infrequently (reviewed in reference 6). The frequency of detection of SV40 DNA sequences suggests that most SV40 infections do not have long-term health consequences. Nevertheless, SV40 infection has been proposed as a possible cause of mesotheliomas (7), ependymomas and other brain tumors (8), non-Hodgkin lymphomas (9–11), and bone tumors (12). Consequently, considerable attention has been focused on the possibility of human SV40 infections and the implications for public health (13). Like other polyoma viruses, SV40 is renotropic and is believed to establish a latent infection in the kidney after primary infection (6). Rhesus monkeys coinfected with simian immunodeficiency virus and SV40 develop acute tubulointerstitial nephritis (14). Butel et al.(15) found SV40 seropositivity in children in association with renal transplantation and amplified SV40 genome from renal transplant biopsies, suggesting for the first time that SV40 might replicate in human kidney (16). The route of transmission SV40 and its possible role in human kidney disease remain to be defined. Interestingly, expression of SV40 large T antigen in transgenic mouse kidney results in focal segmental glomerulosclerosis (FSGS) (17). FSGS is a renal pathologic syndrome that shares a common histopathologic appearance but likely has diverse etiologies (18). The pathogenesis of FSGS is unclear, although the incidence of FSGS has risen considerably during the past 20 yr, and a new form, collapsing FSGS, emerged about 1980 (19). FSGS affects African Americans to a disproportionate extent. FSGS may be associated with viral infections, including HIV-1 (20) and possibly parvovirus B19 (21,22).
We now report the further evidence that human kidney can harbor SV40 and that patients with kidney disease shed the virus into urine more frequently that do controls. These data suggest that SV40 infections are occurring in the human population and suggest a possible link of SV40 with human kidney disease.
Materials and Methods
Subject Groups
Patient characteristics are summarized in Table 2. We studied 40 patients with FSGS, including 29 patients with idiopathic FSGS, 6 patients with idiopathic collapsing FSGS and 5 patients with HIV-associated FSGS. We also studied 20 patients with other kidney disease, including 8 patients with proliferative lupus nephritis, 3 patient with lupus membranous nephropathy, 3 patients with idiopathic membranous nephropathy, 5 patients with diabetic nephropathy, and 1 patient with minimal change disease. In general, patients were studied many months to a few years after the onset of kidney disease. As a control group, we studied 22 healthy volunteers. As expected, the proportion of African Americans was higher in the FSGS group than in the general population. The severity of renal disease was similar in patients with FSGS and patients with other kidney disease. Immunosuppression was defined as the presence of HIV-1 infection or the use of immunosuppressive medication within a period of 3 mo preceding collection of samples for viral studies. For longitudinal studies, five FSGS patients were studied twice at intervals of 6 mo to 1 yr. All renal diagnoses were made on renal biopsies carried out for clinical indications, except for two patients with diabetic nephropathy. Information was recorded on the use of immunosuppressive medication at the time of sample collection or during the preceding 3 mo. Most of the subjects resided in the metropolitan Washington DC area. All patients gave informed consent and participated in clinical trials approved by an Institutional Review Board.
Subject characteristics and SV40 detection: Individual dataa
Continued
Renal Biopsies
For studies of kidney tissue requiring DNA extraction, frozen kidney tissue was not available on the subjects described above, as the patients were referred to NIH and had undergone renal biopsy at other institutions. Therefore, PCR was performed on DNA extracted from a separate set of 50 frozen tissue blocks, including ten biopsies each from patients with idiopathic FSGS, collapsing FSGS, HIV-associated FSGS, membranous nephropathy, and minimal change nephropathy. Blood and urine samples were not obtained on these patients. All renal biopsies were performed solely for clinical indications and were submitted to the Department of Pathology at the University of North Carolina, Chapel Hill.
DNA was extracted from frozen kidney tissue, embedded in OCT (Miles, Elkhart, IN). Tissue was dissected free, and DNA was purified using a commercial kit (Qiagen, Valencia, CA). Considerable care was taken to avoid crosscontamination of DNA between samples; this included dissecting each tissue on fresh aluminum foil and using a fresh disposable scalpel blade.
Collection of Samples and Detection of SV40 Replication in CV-1 Cell Cultures
PBMC were isolated from 10 ml of blood, and urinary cells were isolated from 50 ml of freshly voided urine. PBMC or urinary cells were inoculated into 60% confluent CV-1 cells (American Type Culture Collection, Manassas, VA). These cells were cultured in 25-cm2 tissue culture flasks in DMEM supplemented with 10% fetal bovine serum (FBS) and antibiotics and then maintained culture with DMEM supplemented with 2% FBS for 10 to 15 d for observation. Second passage cultures were made by inoculating 1 ml of first culture supernatant to fresh CV-1 cell cultures in a 75-cm2 flask, and these were maintained for an additional 10 to 15 d. Cultured cells were scraped from the culture flasks for DNA extraction. Stringent precautions to prevent the spread of infectious SV40 between cultures were taken. These included UV light sterilization within the culture hood when handling patient samples and separation of stock CV-1 cell culture and infected CV-1 cell cultures in different rooms. No SV40, BKV, or JCV contamination was detected by nested PCR using SV40, BKV, and JCV virus-specific regulatory region primers, respectively, in CV-1 cell stocks tested on a monthly basis over a period of 2 yr.
Immunofluorescent staining was performed to identify SV40 T antigen in infected cells. CV-1 cells in first culture were passaged into 8-well chamber slides, maintained overnight, and fixed with 4% buffered formaldehyde. Cell layers were permeablized with 0.05% saponin, blocked with 5% normal goat serum, and incubated with a 1:150 dilution of monoclonal antibody directed against the C terminus of SV40 T antigen for 1 h at 37° (Mab 990; Chemicon, Temecula, CA [23]). With indirect immunofluorescence technique, we used hamster cells expressing BK T antigen (gift of Dr. Andrew Lewis, Food and Drug Administration) and hamster cells expressing JCV T antigen (gift of Dr. Eugene Major, National Institute of Neurological Disorders and Stroke, NIH) to confirm that this SV40 antibody has no crossreaction with BKV T antigen and no crossreaction with JCV T antigen (24). The sections were washed in phosphate-buffered saline (PBS) and incubated with FITC-conjugated goat anti-mouse IgG (Sigma, St. Louis, MO) for 1 h. Mouse IgG-2a was used as an isotype control antibody. Human glomerular epithelial cells transformed with adenovirus bearing an origin-defective SV40 genome (25) served as positive cell control.
Detection of Polyomavirus Regulatory Region by Nested PCR
We selected a 315-bp region of the SV40 regulatory region for amplification, as this region shows considerable diversity among polyomaviruses, with this region of SV40 showing 31% identity with BKV and 28% identity with JCV, and would facilitate distinction among members of this virus family. To screen for SV40, we used a sensitive culture system, as CV-1 cells are known to support SV40 replication. Total DNA from CV-1 cell cultures inoculated with PBMC or urinary cell material as described above was extracted using QIAamp DNA blood kit (Qiagen) according to the manufacturer’s instructions with one modification. Before elution of DNA from the column, we performed an additional centrifugation to eliminate washing buffer that might inhibit the subsequent PCR reaction.
Nested PCR primer sets for the SV40 regulatory region (external primers RA3/RA4; internal primers RA1/RA2; Table 1) were used and described by Lednicky et al. (26). When used under high stringency conditions as in the present study, these PCR primers have high specificity for SV40 and do not amplify BKV or JCV genomes (26). Briefly, first reaction cycle parameters using the RA3/RA4 primer pairs for SV40 were 2 min at 94°C followed by 35 cycles at 94°C for 30 s, 65°C for 30 s, and 72°C for 30 s and then followed by a single 7-min extension at 72°C. An aliquot (1 μl) from the first amplification reaction served as a template for second-round PCR. Second-reaction parameters using the RA1/RA2 primer pair were 2 min at 94°C followed by 30 cycles at 94°C for 30 s, 62°C for 30 s, and 30 s at 72°C and then followed by a single 15-min extension at 72°C.
Regulatory region primers for SV40, BKV, and JCVa
Plasmid pBRSV (ATCC), containing the entire genome of reference strain SV40–776, was used as positive control in PCR amplification. DNA extracted from uninoculated CV-1 cells served as negative control. The sensitivity of nested PCR for SV40 genome was found to be ten copies, using serial dilutions of plasmid pBRSV in buffer. Stringent precautions were taken to avoid PCR contamination, including performing cell culture, DNA extraction, and PCR in separate rooms by different technicians.
To screen for BKV and JCV, DNA was extracted directly from 63 PBMC and 61 urine cells and subjected to nested PCR to amplify BKV and JCV regulatory regions (Table 2). BKV and JCV were not cultured in vitro because CV-1 cells are not permissive for their replication. The annealing temperature was 55°C for both BKV and JCV detection, and the other PCR parameters were the same as for the SV40 PCR reaction.
PCR was performed with a Perkin-Elmer GeneAmp thermocycler and a mixture of Taq DNA polymerase and the proofreading GB-D polymerase, with reaction constituents as recommended by the manufacturer (PCR SuperMix High Fidelity kit; Life Technologies, Rockville, MD). The amount of primers used per 25-μl reaction volume was 10 pm in the first round PCR and 20 pm in the second round PCR.
DNA Cloning and Sequencing
Nested PCR products were cloned into the pCR2.1 plasmid using the TOPO TA cloning kit (Invitrogen, Carlsbad, CA), and DNA sequencing was performed using automated methods. Two to four independent clones were sequenced for each patient sample showing an amplicon of the appropriate size. Sequences were compared with reference strains using MacVector (Oxford Molecular, Oxford, UK) and BLASTN algorithms.
Detection of SV40 Genome in Kidney Biopsies by In Situ Hybridization
Tissue from eight renal biopsies was available for in situ hybridization. Paraffin sections, 3- to 5-μm-thick, were placed on precoated glass slides, deparaffinized using three changes of xylene for 10 min each, rinsed in 100% ethanol and rehydrated, and rinsed once in PBS. Sections were then pretreated with proteinase K (125 μg/ml) for 10 min at 37°C followed by incubation for 20 min with avidin D and biotin (Vector Laboratories, Burlingame, CA). The biotin-labeled SV40 DNA probe was prepared by nick translation of cloned full-length genomic SV40 DNA (Enzo Diagnostics, Farmingdale, NY) and has no crossreaction with BK virus and JC virus genome (Enzo, unpublished observations). In situ hybridization (ISH) was applied on sections incubated with SV40 probe (1 ng/μl in hybridization solution composed of 40% formamide and 40% 2× SSC) at 95°C for 10 min and then incubated at 37°C overnight. Detection was performed with horseradish-peroxidase-strepavidin conjugate, diaminobenzidine as substrate, and methyl green as counterstain. SV40-infected monkey renal tissue served as positive control (gift of Michael Eckhaus, Veterinary Resources Program, NIH). As a negative control for each sample, serial sections were processed identically using equivalent concentration of Bio-labeled BK probe and an irrelevant Bio-labeled DNA probe prepared from lambda phage DNA (Enzo Diagnostics).
Statistical Analyses
Data are presented as mean ± SD. ANOVA was used to test differences among three groups, with intergroup comparisons using the LSD test. An exact contingency table analysis was used to compare frequencies among three groups. P < 0.05 was considered to be significant.
Results
Detection of SV40 Replication in Cell Cultures
Cytopathic effect highly characteristic of SV40 infection was observed in CV-1 cells inoculated with material from PBMC or urinary cells derived from some patients. This cytopathic effect consisted of nonisometric cytoplasmic vacuolization of CV-1 cells, which initially affected foci of cells within the culture and with time became extensive across the cell monolayer with longer time in culture (Figure 1,A through C). Cytopathic effect was not always present in CV-1 cultures from patient samples; in some cases, it was absent from cultures from which SV40 DNA sequences were subsequently amplified by PCR.
Figure 1. CV-1 cultures with patient material: cytopathic effect and detection of SV40 T antigen. (A) CV-1 cells cultured with urinary cells from a healthy volunteer after 15 d. (B) CV-1 cells inoculated with urinary cells obtained from idiopathic focal segmental glomerulosclerosis (FSGS) after 7 d, showing cytopathic effect (CPE) with cytoplasmic vacuoles characteristic for polyoma virus infection. (C) CV-1 cells inoculated with urinary cells obtained from a patient with HIV-associated FSGS after 15 d, showing progressively more CPE. (D) CV-1 cells inoculated with urinary cells from a healthy volunteer showed no SV40 T antigen staining. (E) CV-1 cells inoculated with urinary cells obtained from a patient with idiopathic FSGS, showing strong nuclear staining for SV40 T antigen. (F) Human glomerular epithelial cells transformed with adenovirus bearing an origin-defective SV40 genome served as positive cell control for detection of SV40 T antigen. Magnifications: ×20 in A; ×32 in B and C; ×40 in D and E.
Uninoculated CV-1 cell cultures showed no staining for SV40 T antigen (Figure 1D). CV-1 cells inoculated with patient material showed nuclear staining for SV40 T antigen in some cells. SV40 T antigen was most prominently localized to CV-1 nuclei but was also present to a lesser extent within cytoplasm. The positive staining was present in 3 of 13 PBMC and 4 of 13 urinary cell inoculated cultures that were selected for staining (Figure 1, E and F). SV40 genome was confirmed by PCR in all of these seven immunofluorescence-positive samples. One sample immunofluorescence-positive for SV40 T antigen was also positive for BKV by PCR; none of the immunofluorescence-positive samples was positive for JCV by PCR.
Detection of SV40 Regulatory Region Sequence by Nested PCR
Nested PCR was used to amplify SV40 DNA extracted from CV-1 cultures (Figure 2). The frequency of SV40 genome detection in CV-1 cultures inoculated with urine cell material differed among patient groups (Tables 2 and 3). SV40 was present in the urinary cells obtained from 15 (41%) of 36 of all FSGS patients, including 8 of 26 idiopathic FSGS patients, 3 of 6 collapsing FSGS patients, and 4 of 4 HIV-associated FSGS patients. This rate of detection was significantly greater than the rate of detection in other kidney diseases (2 of 20; 10%; P < 0.02) and healthy volunteers (1 of 22; 5%; P = 0.003). In contrast to the finding with urine cells, the rate of detection of SV40 in PBMC was not significantly different among patients with FSGS (14 of 40; 35%), other kidney diseases (3 of 20; 15%), and healthy volunteers (5 of 22; 22%).
Figure 2. Nested PCR to detect SV40 regulatory region genome. (A) PCR of DNA from CV-1 cells cultured with peripheral blood mononuclear cells (PBMC; lanes 3 to 6) and urine cells (lanes 8 to 11) obtained from FSGS patients with idiopathic FSGS, collapsing FSGS, and HIV-associated FSGS. Molecular weight (MW) marker is shown (lane 1). Controls include SV40–776 plasmid (lane 2), CV-1 cells only (lane 7), and buffer (lane 12). All bands were sequenced; the 315-bp proximal size bands were shown to represent SV40 (GenBank Accession Nos. AF416272, AY054453, AF416243, AF416277, AF416273, and AY054452). The DNA in lanes 4 and 10, with sizes divergent from 315 bp, represent human genomic sequence. (B) PCR of DNA from CV-1 cells cultured with PBMC (lanes 4 to 7 and 9 to 11) and urinary cells obtained from healthy volunteers (lane 8). MW marker is shown (lane 1). Positive control is SV40–776 plasmid (lane 2); negative control is CV-1 cell DNA (lane 3). The 315-bp amplicons in lanes 7 and 10 were SV40 sequences with point mutations compared to SV40–776 (GenBank Accession Nos. AY052640 and AF416281). The 350-bp amplicon in lane 11 was found to be human genomic DNA. (C) PCR screening of SV40 clones obtained from PBMC from a patient with idiopathic FSGS. Lane 1, MW marker; lane 2, amplification of reference strain SV40–776 (315 bp); lane 3, buffer control (negative); lanes 4 to 8, amplicons from clones derived from one FSGS patient PBMC; lane 4, 315-bp amplicon identical to SV40–776; lane 5, 314 amplicon with 1-bp deletion; lanes 6 and 7; 293-bp amplicons with a 21-bp promoter deletion, sequence identical to each other; lane 8, 242-bp amplicon, representing an archetypal regulatory region, with one 72-bp enhancer deletion compared with SV40–776 (GenBank accession Nos. AY054454, AY052636, AY052635, and AY052634).
Subject characteristics and SV40 detection: Summarya
Among 40 FSGS patients, SV40 genome was detected by PCR in both PBMC and urine cells in eight patients, only in PBMC in six patients, and only in urine cells in seven patients. Among 20 patients with other glomerular disease, SV40 genome was detected in both PBMC and urine cells in one patient, only in PBMC in two patients, and only in urine cells in one patient. Among 22 healthy volunteers, SV40 genome was detected in both PBMC and urine cells in one patient, only in PBMC in four patients, and only in urine cells in no patients.
Among African American patients with FSGS, 11 of 29 PBMC and 13 of 26 urinary cells contained SV40. Among FSGS patients of other racial backgrounds, 3 of 11 PBMC and 2 of 10 urine cells contained SV40 (P = 0.72 and P = 0.14, respectively). These differences were NS between racial groups. Among subjects born before 1964 and therefore >35 yr at the time of sample collection, 13 (26%) of 50 of PBMC samples and 10 (21%) of 47 urine samples contained SV40 genome. Among subjects born after 1963, 9 (28%) of 32 of PBMC samples and 8 (26%) of 31 of urinary cell samples contained SV40 genome. These differences between the age groups were NS.
SV40 was recovered from PBMC in 8 (50%) of 16 of immunosuppressed FSGS patients and 6 (25%) of 24 of FSGS patients without immunosuppression (P = 0.18). Similarly, SV40 was recovered from urine cells in 7 of 16 immunosuppressed FSGS patients and 8 of 24 FSGS without immunosuppression (P = 0.53) There was also no association between immunosuppression and recovery of SV40 in other glomerular diseases or in all kidney patients taken as a single group. Thus, there was no convincing evidence that SV40 recovery was associated with immunosuppression, at least according to the definition used in the present study.
Using BKV- and JCV-specific primers extracted directly from PBMC and urinary cells, BKV DNA was detected in 3 of 63 PBMC and 7 of 61 urine cell samples. Coinfection with SV40 and BKV was found in one FSGS PBMC sample and one urine sample. JCV DNA was detected in 2 of 63 PBMC and 9 of 61 urine cell samples. Coinfection with SV40 and JCV was found in one FSGS urine sample. Coinfection with BKV and JCV was found in one urine sample from a healthy volunteer. Detection of BKV and JCV was similar among the patient groups.
Among the archived renal biopsies, SV40 genome was more frequently detected from FSGS biopsies than from biopsies from patients with membranous nephropathy or minimal change disease (P < 0.02) (Table 4).
Detection of SV40 genome in renal biopsies
Genetic Analysis of SV40 Regulatory Region Sequence Variability
To verify the accuracy of sequence determination, we sequenced eight clones of PCR products derived from PCR amplification of plasmid containing SV40–776 in both forward and reverse direction. All 16 sequences were identical to the published sequence.
Seventy-nine clones containing SV40 regulatory region were obtained from study subjects. 24 (30%) of 79 clones were identical to the reference strain SV40–776 (non-archetypal with two 72-bp enhancer elements). The SV40–776 regulatory sequence was found in PBMC from 13 patients and in urine cells from 11 patients. The detection of SV40–776 was not limited to any particular time period of sample acquisition, arguing against laboratory contamination. Forty-seven (60%) of 79 sequences showed 1 to 6-bp mutations or deletions. Eight sequences (10%) showed large deletions that involved a functional unit such as a 21-bp promoter element or one 72-bp enhancer element.
SV40 sequences obtained from five subjects (four FSGS patients and one healthy volunteer) demonstrated both non-archetypal (containing two enhancer elements) and archetypal regulatory region (containing one enhancer element). No patient had exclusively archetypal regulatory region sequences. One FSGS patient had SV40 clones with three types of regulatory region (Figure 2C): archetypal, non-archetypal, and non-archetypal with a deletion of one of the three promoter elements, the last one representing a novel structure for the SV40 regulatory region (Figure 4).
Figure 4. Sequence variation in the SV40 virus regulatory region. The full sequence is shown of the 315-bp product of nested PCR obtained from the SV40–776 reference strain, which is non-archetypal, with three 21-bp promoter elements and two 72-bp enhancer elements (termed P3E2). Numbering of the bases is by convention, with the origin of replication shown as +1. The four T antigen binding sites (5′GAGGC3′ or its palindrome 5′CTCCG3′) are located within the core origin region and are shown in blue; the TATA box is underlined; the promoter elements are shown in green, with the initial base of each promoter element boxed; the enhancer elements are shown in red, with the initial base of each enhancer element boxed. Shown are eight representative amplicon sequences of SV40 strains which were isolated and cloned from human samples and which have mutations compared with SV40–776. The Genbank accession numbers are shown. Four sequences are shown with one or more point mutations and/or deletions, obtained as follows: from urine cells of FSGS patient G25, urine cells of lupus nephritis patient K3, PBMC of healthy volunteer N8, and PBMC from FSGS patient G37. Sequence G8M TagBS, obtained from PBMC of HIV-associated FSGS patient G8, demonstrates deletions within the T antigen binding site. One sequence, obtained from PBMC of FSGS patient G37, had only two promoter elements (the third promoter element was deleted) and two enhancer elements (termed P2E2). One sequence, also obtained from PBMC of FSGS patient G37, had three promoter elements and one enhancer element, equivalent to an archetypal region (termed P3E1). One sequence, obtained from PBMC of HIV-associated FSGS patient G8, had a combination of deletion of the second enhancer element and deletions within the T antigen binding site (termed P3E1+TagBS.1). Thus, regulatory regions of human SV40 strains in some instances show substantial divergence from SV40–776, and multiple strains were recovered from an individual patient.
Longitudinal studies in five FSGS patients with SV40 in urine cells showed that four patients had persistent shedding of infectious SV40 and one patient did not have SV40 recovered after 1 yr, coincident with resolution of proteinuria.
Thirty-nine clones containing SV40 regulatory region sequences were obtained from 50 renal biopsies. Among these, 25 (64%) were identical to SV40–776 and thus non-archetypal, 11 (28%) were non-archetypal SV40 sequences that showed one- to three-point mutations or 1- to 3-bp deletions, and 3 (12%) were archetypal SV40.
Detection of SV40 Genome in Kidney Biopsies by In Situ Hybridization
Eight patients had kidney tissue available for in situ hybridization. Seven patients had SV40 detected in urinary cells, and none had BKV or JCV detected by PCR; one patient was anuric. SV40 genome was localized within kidney in five of eight renal biopsies, including two of four idiopathic FSGS, two of three collapsing FSGS, and one of one HIV-associated FSGS. Normal kidney from nephrectomy performed for cancer showed no SV40 hybridization signal. When present, SV40 genome was localized to the nuclei of some interstitial inflammatory cells and tubular epithelial cells; multiple tubular cells within a tubular profile typically showed evidence of infection (Figure 3). Glomerular cells were uniformly negative for SV40 genome. BKV was not detected in the SV40-positive control samples, showing the specific detection of SV40 in situ.
Figure 3. In situ hybridization for SV40 genome in renal biopsies. (A) SV40 genome was detected in the nuclei of tubular epithelial cells of the renal biopsy from HIV-associated FSGS patient. (B and C) SV40 genome was detected in the infiltrating inflammatory cells and tubular epithelial cell nuclei from a patient with idiopathic FSGS. Glomeruli showed no hybridization signal. (D) Positive control is SV40-infected monkey kidney, with SV40 genome present in the nuclei of tubular epithelial cells and infiltrating inflammatory cells. (E) Negative control is SV40-infected monkey kidney hybridized with lambda phage DNA probe; no signal was observed. (F) Another negative control is normal human kidney obtained at cancer nephrectomy hybridized with SV40 probe. Magnifications, ×40 for all.
Discussion
In this study, we have found evidence suggesting that SV40 infects human subjects and that human kidney can serve as a viral reservoir, leading to viral shedding in urine. Viral shedding is largely restricted to patients with kidney disease and is particularly prominent in FSGS. This evidence for SV40 infection includes the following observations: typical polyomavirus cytopathic effect observed in cell cultures inoculated with patient material, the presence of SV40 T antigen in cultured cells, the identification of SV40 genome within kidney tissue by PCR and by in situ hybridization, the detection of SV40 regulatory region sequences by PCR and confirmation by sequence analysis, and the considerable heterogeneity of the regulatory region sequence. Our finding that SV40 infection is present in subjects born after 1964 suggests that at least some infections have been acquired by routes other than contaminated vaccines.
This is the first time that SV40 has been identified in human urine samples. Possible reasons for the increased rate of detection of SV40 in urinary cells in the present report include the use of a sensitive culture system coupled with nested PCR and the inclusion of patients with kidney disease. Using in situ hybridization, we localized SV40 genome to tubular epithelial cells and infiltrating cells. After initial dissemination, SV40 may establish latency in kidney, as does BKV. We can envision at least three scenarios for the relationship between SV40 recovery in urine and tubular cell injury. First, SV40 may reactivate as a consequence of tubular inflammation associated with kidney disease in a manner similar to activation of mouse polyomavirus replication in kidney after renal injury from toxins or ischemia (27). Second, increased shedding of virus in kidney could also result from the loss of renal epithelial cells or immune cells in the urine consequent to renal inflammation or from immunosuppression associated with kidney disease or its therapy. Although we could not demonstrate an association between immunosuppression and SV40 recovery in this study, the number of patients studied is relatively small and we feel that the relationship between immunosuppression and shedding of SV40 remains an open question. Third, it is possible that SV40 may cause tubulitis, as is the case of BKV, and thus might contribute to the progression of renal injury.
We did not localize SV40 genome to the glomerulus, although with the limited number of glomeruli present in the tissues we examined, we cannot exclude that glomerular cell infection occurs. In one patient after renal transplantation, polyoma virus T antigen staining was positive in glomerular parietal epithelial cells, and SV40 genome was detected in the patient biopsy DNA (Li and Kopp, unpublished observation). In this regard, it may be relevant that SV40 T antigen transgenic mice develop FSGS, but in these mice T antigen expression is confined to the tubular epithelium (17). It is possible that SV40 might play a role in human kidney diseases such as FSGS, either by initiating renal injury or by acting as a progression factor. The hypothesis that SV40 infection might be specifically associated with FSGS will require more detailed epidemiologic investigation.
The detection rate of SV40 genome in PBMC of healthy volunteers in our study was similar to that found using direct PCR methodology by David et al. (3) and Martini et al. (4), but such findings remain controversial. Though most SV40 infections in humans appear to be asymptomatic, the presence of SV40 in PBMC suggests that hematogenous spread of viral infection may occur,
The rate of SV40 detection in PBMC and urinary cells was similar when we compared older subjects (born before 1963) who might have received SV40 contaminated vaccines and younger subjects who did not receive contaminated vaccine. This suggests that SV40 may be transmitted through other routes than contaminated vaccines. Animal studies have demonstrated that SV40 can be vertically transmitted (28). It is likely that polyoma viruses are transmitted in other ways as well, such as via respiratory secretions or fecal or urine shedding, as the frequency of seropositivity rises during childhood. Our data are congruent with previous studies, which showed that antibodies against SV40 were present in children too young to have received the contaminated vaccine (15).
The genetic heterogeneity of the SV40 genomes described here has several implications. First, it argues that the detection of SV40 genomes in human patients is unlikely to be due to laboratory contamination. Second, analysis of the regulatory region sequences shows that the dominant strain of virus is non-archetypal, containing two enhancer elements, with either no mutations or few mutations in the regulatory region. Only 10% of SV40 regulatory region sequences were archetypal, containing a single enhancer element. A novel sequence with a promoter element deletion with two enhancer elements was present in the PBMC of an FSGS patient. Our findings are also similar to those of other investigators, who have demonstrated SV40 regulatory region sequence heterogeneity in renal transplant biopsies (15), peripheral blood lymphocytes (3), and brain tissue obtained from SIV-infection monkeys (28). These findings differ, however, from those of Arringon et al. and Tognon et al., whose work on SV40 regulatory region and T antigen sequences obtained from patients with mesothelioma and neurologic disease suggests that such tissue possess only a single viral variant (29,30). These differences may reflect the differing tissue origins of samples. Furthermore, our findings of the diverse regulatory region of SV40 demonstrate considerable heterogeneity, both within and between patients, indicating polymorphism among SV40 viruses that are present in the human population.
Importantly, kidney appears to be a site from which both archetypal and non-archetypal SV40 strains can be recovered. Our findings suggest that human kidney can sustain replication of both strains of SV40. In this regard, mesothelial cells are unusual in that they can be infected with non-archetypal SV40 (31), and the present findings suggest that kidney cells may share this susceptibility. It has been difficult to understand why different tumors are associated with different SV40 strains and what the source of those strains might be. Thus archetypal SV40 is typically found in brain tumors (8), both archetypal and non-archetypal SV40 are found in bone tumors (32), and non-archetypal SV40 is found in mesotheliomas (33). A renal reservoir of SV40 would help explain how multiple strains may be present in the human population, and as part of the premalignant process particular variants may preferentially infect certain cell type or accomplish malignant transformation more efficiently.
In conclusion, we have detected SV40 genome from PBMC and urinary cells and renal biopsies from patients with kidney disease and healthy volunteers. We have found considerable genetic diversity in the regulatory region of human SV40 isolates. Human kidney is a reservoir for SV40 viruses showing both archetypal and non-archetypal regulatory regions. Inflammatory renal disease or immunosuppression may promote the excretion of SV40 virus into urine, or alternatively SV40 may be responsible for renal epithelial injury.
Acknowledgments
We wish to acknowledge the assistance provided by physicians who referred patients and of the patients who participated in this study, to Jorge Briones for technical assistance, to Dr. Eugene Major, NIH, for the hamster JC cells, and to Drs. Andrew Lewis, Keith Peden, and James Balow for critical review of the manuscript.
- © 2002 American Society of Nephrology