Humans are natural hosts for two polyomaviruses, BK virus (BKV) and JC virus (JCV). BKV has been associated with several clinical syndromes, including interstitial nephritis, ureteritis, and cystitis, which occur chiefly in immunosuppressed humans (reviewed in reference 16
). Polyomavirus nephropathy, most commonly associated with BKV infection, is particularly prevalent among renal transplant recipients, where it affects 6 to 7% of patients (8
). Reduction of immunosuppressive medication may reduce BKV replication, but it is not always clear that a modest reduction in therapy has a beneficial effect on viral titers and a complete cessation of therapy is often not feasible. Few studies have examined the efficacy of antiviral agents against BKV infection. In vitro studies suggested that cidofivir has activity against murine polyomavirus and simian virus 40 (SV40), but the drug has significant renal toxicity which limits its utility in human patients (1
). To date, no animal model demonstrating BKV replication has been described.
species, including Rhesus macaque
and less commonly Rhesus cynomolgus
, are the natural hosts for SV40. SV40 was first identified as a contaminant of polio vaccine (20
). SV40 may have become prevalent in the human population, although considerable controversy surrounds this issue, and SV40 has been implicated as a cause of cancer and other diseases (reviewed in reference 5
). SV40 infection apparently causes little or no tissue injury in rhesus monkeys under normal circumstances, but it is associated with interstitial nephritis and encephalitis in immunosuppressed rhesus monkeys (9
). Other monkeys that are susceptible to experimental infection with SV40 include African green monkeys (Cercopithecus aethiops
). Cynomolgus monkeys have been found to be infected with a distinct polyomavirus that causes interstitial nephritis and enteritis (22
Polyomaviruses manifest considerable host restriction. BKV can induce tumors in rodents (e.g., hamsters), but rodent cells do not support viral replication. Inoculation of rodents, such as mice and hamsters, with SV40 results in characteristic tumors but does not lead to viral replication (6
). London et al. inoculated owl monkeys (Aotus trivirgatus
) with JCV, BKV, and SV40; JCV, but not the other viruses, was associated with brain tumors after a 1- to 3-year incubation period (14
). Similarly, London et al. found that JCV induced brain tumors in squirrel monkeys (Saimiri sciureus
We wished to determine whether BKV or SV40 could replicate in a primate that was not the natural host. In particular, we sought an animal model of BKV infection that might facilitate studies of pathogenesis and that might aid in the development of more effective antiviral therapies. Further, we sought to determine whether SV40 could replicate in a primate other than the rhesus monkey and whether it could induce pathology in this nonnatural host that was distinct from that seen in the rhesus monkey. We selected two standard viral strains, BKV Gardner strain and SV40 777 strain. The latter strain has an archetypal regulatory region, with two promoter elements and a single enhancer element, and was selected in view of the recovery of archetypal SV40 strains from tissues of rhesus monkeys with naturally acquired infection (10
). We selected New World squirrel monkeys for these experiments because of their availability and suitability as experimental animals.
Ten squirrel monkeys, aged 10 to 20 years (three males and seven females) were obtained from commercial breeders. Monkeys were maintained at a National Institutes of Health (NIH) animal care facility. The monkeys were inoculated with virus. The monkeys were anesthesized with ketamine given by intramuscular injection, and blood samples were obtained from the monkeys. Random urine samples were collected from the monkeys in standard cages. At the conclusion of the experiment, monkeys were euthanized using Beuthanasia (Schering Plough Animal Health, Union, N.J.) administered by intravenous injection, and selected tissues were obtained at autopsy. Monkeys were cared for according to NIH Animal Care Guidelines, and experiments were approved by the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) Animal Care and Use Committee.
In order to test the role of immunosuppression in promoting viral replication, we administered cyclophosphamide to eight monkeys, with the timing of immunosuppressive therapy as described below (see Table , footnote a). Cyclophosphamide (Bristol-Myers Squibb, Princeton, N.J.) was reconstituted in sterile water and administered by subcutaneous injection to conscious monkeys twice weekly. The initial cyclophosphamide doses were 30 mg/kg of body weight. Blood samples were obtained twice weekly for complete blood count with differential, and the cyclophosphamide doses were adjusted to maintain absolute neutrophil count under 1,000 cells/μl. Maintenance cyclophosphamide doses ranged from 15 to 35 mg/kg per day.
BKV and SV40 infection of squirrel monkeysa
The BKV Gardner strain was isolated from a renal transplant recipient and was originally a gift of Sylvia Gardner. The virus was grown on primary human fetal glial cells, and the virus titer was estimated by hemagglutinin assay by Duard Walker. The viral stock was a generous gift of Richard Frisque. The SV40 777 strain was grown on CV1 cells, and virus titer was estimated by infection of primary African green monkey cells as previously described (12
). For both viruses, inoculation was accomplished by intravenous injection of virus diluted in 1 ml of sterile saline into the saphenous vein of a monkey.
DNA from blood, urine, and tissue samples was subjected to quantitative PCR analysis in a laboratory distant from the laboratory in which sample preparation took place. For whole-blood and urine samples, a 200-μl aliquot was transferred to a microcentrifuge tube containing 0.9 ml of lysis buffer included in the NucliSens isolation kit (bioMerieux, Durham, N.C.). The microcentrifuge tube was vortexed and stored at −70°C. The NucliSens isolation kit was used according to the manufacturer's instructions. Briefly, cells in the sample were lysed, and silica particles were added to bind the nucleic acid. After several washes, the nucleic acid was eluted from the silica using 50 μl of elution buffer.
Nucleic acid was extracted from pieces of tissue weighing approximately 10 mg for each spleen specimen and approximately 25 mg for all other tissue types using the DNeasy tissue kit (QIAGEN, Valencia, Calif.) following the manufacturer's instructions. Briefly, tissue samples were lysed in a buffer solution containing proteinase K and then loaded into the silica gel membrane columns, which bind the nucleic acid. Nucleic acid was eluted from the membrane using 100 μl of elution buffer. Eluate DNA concentration was measured on a UV-visible spectrophotometer (Shimadzu Scientific Instruments, Columbia, Md.).
Quantitative, real-time PCR was performed using the LightCycler (Roche Molecular Biochemicals, Indianapolis, Ind.). This instrument provides a platform for the amplification and continuous monitoring of the accumulation of target amplicons through application of fluorescence resonance energy transfer (FRET) technology using two fluorophore-labeled hybridization probes (23
BKV primers (Table ) (Midland Certified Reagent, Midland, Tex.) were used to amplify a 215-bp region of BKV (bases 1502 to 1716, Gardner strain, GenBank accession number V01108) that includes partial sequences for the VP1, VP2, and VP3 genes. SV40 primers (Table ) (Midland) were used to amplify a 476-bp region of SV40 (bases 363 to 838, MC-028846B strain, GenBank accession number AF180737) that includes partial sequences for the LP1 and VP2 genes. All reactions were performed in glass capillaries (Roche) using 1× LightCycler-FastStart DNA Master Hybridization Probes reaction mixture (Roche) containing FastStart Taq polymerase, reaction buffer (containing deoxynucleoside triphosphates, with dUTP substituted for dTTP), primers at concentrations of 1.0 μM, FRET probes at concentrations of 0.2 μM, 4.0 mM MgCl2, and 1 U of heat-labile uracil-DNA glycosylase (UNG) (Roche). Separate reactions were performed for the detection of BKV and SV40 using the appropriate primer and FRET probe sets (Table ) (Idaho Technologies, Salt Lake City, Utah.). Each reaction tube contained 10 μl of master mix and a 10-μl aliquot of extracted DNA or control material.
Primer and probe sequences used for the detection of BKV and SV40 genomes by quantitative PCR
To verify that impurities that may inhibit PCR amplification were removed during the extraction procedure, separate plasmid internal controls (IC), one of which is amplifiable by the BKV primers and the other of which is amplifiable by the SV40 primers, were constructed as previously described (4
). Since the internal sequences of both the BKV IC and the SV40 IC are identical, one set of FRET probes (Table ) (Idaho Technologies) was designed for detection of the IC products. The concentration (number of copies per microliter) of each IC was calculated using the A260
optical density value and the plasmid molecular weight. Each IC plasmid was linearized and then diluted in Tris-EDTA buffer (pH 8.0) supplemented with glycogen (33.3 μg/ml) to obtain the working stock concentration and stored at −70°C. In parallel reactions used to evaluate the presence of amplification inhibition, 1,000 copies of the BKV or SV40 IC was added to the reaction mixture, and the target FRET probes were replaced with 0.2 μM IC FRET probes.
Standards for the quantification of BKV and SV40 genomes were generated by cloning the target amplicon into the pCR2.1 vector (InVitrogen, Carlsbad, Calif.). After propagation and purification of each plasmid, the plasmid concentration was calculated as described above. The plasmids were linearized and then diluted in Tris-EDTA buffer (pH 8.0) supplemented with glycogen (33.3 μg/ml) to obtain concentrations of 5, 50, 500, and 5,000 copies per reaction mixture. Amplification of a BKV stock with 107 hemagglutinin units (HAU) yielded a measurement of 13 × 107 genomes, suggesting that the two methods differ by approximately 1 log unit and that the PCR appears to be more sensitive than the hemagglutination assay.
The thermocycling protocol consisted of an initial 10-min incubation at 30°C for UNG activity, followed by an 8-min incubation at 95°C to activate the DNA polymerase, inactivate the UNG, and melt double-stranded DNA. For amplification of BKV and for the IC reactions, this was followed by 50 cycles, with 1 cycle consisting of denaturation (5 s at 95°C), annealing (10 s at 62°C), and extension (20 s at 72°C). For amplification of SV40, the annealing temperature was 70°C for 10 s. In each experiment, the standards (5, 50, 500, and 5,000 copies/reaction mixture) were included to generate a standard curve for quantification of positive samples. In addition, each experiment also included positive controls (extracted DNA from BKV [ATCC VR-837] or SV40 [ATCC VR-820]) and negative controls (sterile water). The minimum level of detection for both BKV and SV40 was five plasmid copies per reaction mixture. All specimens were tested in duplicate with one reaction mixture containing the IC. A negative BKV or SV40 result had to have a positive result for the corresponding IC to be considered valid.
Tissues were embedded in zinc formalin (Anatech, Battle Creek, Mich.), and paraffin sections were cut at 5 μm. Sections were stained with hematoxylin and eosin and examined by a pathologist (M. Raffeld). Also, brain sections were stained with Luxol Fast Blue to detect demyelination and examined by a neuropathologist. Sections from all 10 tissues from all 10 monkeys were deparaffinized and incubated with a mouse monoclonal antibody directed against SV40 T antigen (Ab2, clone Pab 416, diluted 1:200; Oncogene Science, Cambridge, Mass.) that cross-reacts with BKV and JCV T antigens. The primary antibody was detected by the streptavidin-biotin method and developed using diaminobenzidine as the chromogen.
We inoculated four squirrel monkeys by intravenous administration with BKV Gardner strain, using 105 HAU in two monkeys and 5 × 105 HAU in two monkeys. Using genome estimates from PCR, we estimate that these two HAU values represent approximately 1.3 × 106 and 6.5 × 106 genome equivalents, respectively. We inoculated six squirrel monkeys with SV40 777 strain intravenously, using 107 PFU in four monkeys and 106 PFU in two monkeys. Cyclophosphamide was administered to eight monkeys as shown in Table , and the resulting immunosuppression medication was generally well tolerated. Lymphocyte counts were generally maintained below 2,000 cells/μl. One monkey developed a vesicular lesion on the lips consistent with acute herpetic infection, which improved when cyclophosphamide was temporarily withheld.
Four weeks after BKV inoculation, BKV was detected by quantitative PCR in blood from all four monkeys that had been inoculated (Table ). Viral copy number was higher in monkeys given the higher inoculating dose but was unaffected by immunosuppression therapy. BKV was detected in only one monkey which had been inoculated with the higher dose and exposed to immunosuppression therapy; the virus was detected in urine at low copy number. Using quantitative PCR, we found that SV40 was absent in blood from all six monkeys inoculated with SV40 and was present at low copy number in urine obtained from only one monkey (Table ).
Quantitative PCR revealed BKV DNA in all tissues harvested from all four monkeys (Table ). Viral copy numbers were higher in the two monkeys that received higher inoculating dose and were not increased appreciably by immunosuppression.
By contrast, SV40 DNA was detected at low levels and was detected in relatively few tissues, primarily lymphoid tissues and kidney. Among monkeys that received delayed immunosuppression therapy, SV40 DNA was at low levels in lymph node tissue from three of four monkeys (<1, 2, and 2 copies/μg of DNA, respectively) and was present in the lung of one monkey (9 copies/μg of DNA) and in the spleen of another monkey (<1 copy/μg of DNA). Among monkeys who received continuous immunosuppression therapy, SV40 DNA was detected in one monkey only in the renal medulla (0.03 copy/μg of DNA) and in another monkey in lymph node (21 copies/μg of DNA), spleen (1 copy/μg of DNA), renal cortex (<1 copy/μg of DNA), and renal medulla (<1 copy/μg of DNA).
Several negative results add credence to the specificity of the PCR amplification reactions. PCR analysis for BKV and SV40 using baseline blood and urine samples from all monkeys obtained prior to inoculation showed no amplification of viral sequences in any sample. Amplification of blood, urine, and tissue samples from SV40-inoculated animals using BKV primers yielded no amplicons. Similarly, amplification of blood, urine, and tissue samples from BKV-inoculated animals using SV40 primers yielded no amplicons.
At autopsy, tissues, including cerebral cortex, cerebellum, ependyma, lung, liver, spleen, lymph node, tonsil, thymus, heart, and renal cortex and medulla, appeared normal in all SV40-infected monkeys. Three of four BKV-infected monkeys exhibited focal interstitial nephritis, largely confined to the renal cortex. Focal mononuclear cell infiltrates were present surrounding up to 10% of glomeruli and occasional cortical venules, with the infiltrates largest and most numerous in the monkey with the highest copy number of BKV detected in kidney tissue (Fig. ). This infiltrate was composed primarily of lymphocytes and plasma cells, suggesting a chronic process. Tubular cell nuclear morphology was normal, and the enlarged nuclei with a ground-glass morphology that are typical of human BKV nephropathy were absent.
FIG. 1. Kidney histology and immunostaining for BKV T antigen. The kidney tissue sections shown are from the monkey that was inoculated with a high dose of BKV (5 × 105 HAU) and did not receive immunosuppression therapy. At low power (A), the cortex, (more ...)
Immunostaining of all 10 tissues from all monkeys using an antibody cross-reactive with all polyomavirus T antigens was negative with the exception of the monkey with the highest renal BKV copy number. In this monkey, BKV staining was confined to occasional cell nuclei within the renal medulla, particularly within the papilla (Fig. ). These nuclei were round or spindle-shaped and were generally located within peritubular capillaries, presumably within endothelial cells.
These results suggest that BKV establishes a productive infection in squirrel monkeys. To our knowledge, this represents the first time that BKV has been shown to replicate in a nonhuman species. The monkeys remained healthy and did not exhibit any apparent symptoms over the period of infection. Analysis of blood and tissue showed high viral genome numbers, consistent with active viral replication. BKV genome was detected in plasma at levels (104
copies/ml) that are similar to those seen in human patients with BKV nephropathy. High levels of BKV genome were detected consistently in the brain, particularly in the cortex and ependyma, with quantitatively lower levels in the kidney and spleen. In humans with BKV nephropathy, BKV has been located in the brain and kidney (2
). In the monkey that received high inoculation doses and continuous immunosuppression therapy, the BKV genome was detected at high levels in all tissues examined, with the exception of low levels in lymph nodes. Somewhat surprisingly, the BKV genome was detected in urine in only one monkey, although the BKV genome was detected in kidney tissue in all four monkeys inoculated with BKV.
Focal interstitial nephritis was present in three of four monkeys that were inoculated with BKV and was most striking in the monkey with the highest BKV copy number. The infiltrates were largely confined to the cortex, although the expression of BKV antigen was confined to the inner medulla. The reasons for this discrepancy are not entirely clear. The absence of inflammatory response to BKV in other tissues and in the other monkeys may be a consequence of lower viral copy number, which makes detection in random tissue sections more difficult. Since the monkeys were sacrificed only 4 weeks after inoculation, it is possible that allowing viral infection in the kidney to persist for a longer time would have been associated with more histological signs of disease.
It was interesting that despite high BKV genome copy numbers in certain tissues, viral protein in only one tissue in one monkey could be detected by immunostaining. It may be that immunostaining is relatively insensitive or that viral replication may be focally distributed, causing it to be missed when tissue is sampled. When BKV T antigen was detected in squirrel monkey kidney, it was present in the renal papilla and was localized to peritubular capillary endothelial cells. In humans with polyomavirus nephropathy, BKV-infected cells are generally tubular epithelial cells, particularly those within the collecting tubules of medulla and cortex (reviewed in reference 16
). Recently, a syndrome of BKV infection of endothelial cells, leading to fulminant vasculopathy (18
), was described. Little is known about determinants of cellular tropism by BKV.
One consideration is whether BKV genomes detected in blood and tissue could represent the inoculated virus, without viral replication in vivo. We consider this to be unlikely for two reasons. First, the spleen of a squirrel monkey weighs approximately 50 g and the BKV-inoculated monkey spleens had ≥2,000 genomes/mg of tissue weight, suggesting that each spleen contained ≥108 virions. This is approximately 2 log units higher than the total body inoculum (accounting for differences in quantification between hemagglutinin assay and PCR) and thus supports the interpretation that there is viral replication in vivo. Second, SV40 was not detected in blood and most tissues, suggested that circulating SV40 has been sequestered and cleared by 4 weeks after inoculation. If BKV were not replicating, it appears likely that it also would not be detected.
One striking feature in BKV infection in squirrel monkeys is that inoculum size appears to influence viral copy number in blood and tissue at sacrifice 4 weeks later. Thus, a fivefold increase in inoculum size was associated with approximately 10-fold-higher viral copy number in many tissues. As only four monkeys were tested in this experiment, these results must be interpreted with caution. Nevertheless, these results suggest that inoculum size might influence the initial level of viral infection and that in turn this might influence the viral set point for persistent infection.
This model may prove useful in studies to elucidate the pathogenesis of BKV infection and to test antiviral agents designed for human use. At this time, we are analyzing regulatory region sequences amplified from different tissues to see whether there is selection pressure on that region of the genome and whether this pressure differs in different tissues. Although BKV DNA was detected in the kidney, viral DNA was detected in urine in only one monkey. It is possible that alternate immunosuppressive regimens or more prolonged use of immunosuppressive medication might facilitate BKV replication in kidney and increase the severity of renal disease.
In contrast to robust replication of BKV, squirrel monkeys infected with SV40 exhibited very low viral copy numbers, largely limited to lymphoid tissue and kidney. These results most likely suggest that squirrel monkey cells have limited ability to support SV40 replication. Squirrel monkeys, as New World primates, diverged evolutionarily from Old World primates, such as rhesus monkeys, approximately 35 to 40 million years ago. While this long period of divergence might explain the relative resistance of squirrel monkeys, which are New World primates, to infection with SV40 (whose natural hosts are within the Rhesus genus and are Old World primates), it does not explain why squirrel monkeys are susceptible to infection with BKV (whose natural host is Homo sapiens, an Old World primate). The differences in susceptibility to BKV and SV40 may be a fruitful area for future investigation.
In conclusion, we have demonstrated that BKV can infect a nonhuman species and we have shown that squirrel monkeys may provide a useful model to study BKV replication in vivo. The ability of a polyomavirus to infect more than one primate host suggests the potential for monkey polyomaviruses, notably SV40, to infect humans.