The cohort study consisted of 101 patients who received kidney (n = 83) or kidney–pancreas (n = 18) transplants during 2002–2004 at Leiden University Medical Center (LUMC), Leiden, the Netherlands (Table 1). This study population is part of a larger prospective European multicenter study designed to investigate the role of human papillomavirus infection in the development of skin cancer in solid-organ transplant patients (27). The study adhered to the Declaration of Helsinki Principles, and the medical ethical committee of the LUMC approved of the study design (Medical Ethical Committee no. P02.111). Participants gave written informed consent.

All patients received induction with interleukin-2 receptor blocker daclizumab (100 mg/d) on the day of transplantation and 10 days after transplantation or basiliximab (40 mg at days 0 and 4), followed by triple therapy with prednisone, tacrolimus, or cyclosporine and mycophenolate mofetil. For kidney transplant patients, the dose of the calcineurin inhibitor (tacrolimus or cyclosporine) was tapered at 6 weeks after transplantation, whereas for kidney–pancreas transplant patients, the calcineurin inhibitor was reduced at 3 months after transplantation.

The time points of serum sample collection and the number of samples collected per time point are summarized in Table 2 and shown in relation to the date of transplantation. The baseline samples were obtained in the days immediately after transplantation (time point 0, T0). To collect the subsequent samples, patients were asked to visit the LUMC outpatient clinic for follow-up sample collection at the preferred time points of 3, 6, 9, 12, and 18 months after transplantation (T3–T18). A total of 58 patients provided a sample at all 6 time points: 31, 6, 4, and 2 patients provided 5, 4, 3, or 1 samples, respectively.

Stored pretransplantation serum samples, if available, were retrieved and tested for HPyV9 DNA (n = 65; Table 2) and antibodies (n = 45, 40 of which were also included in pretransplantation DNA testing). The average dates of obtaining samples for DNA and antibody testing were 8 and 2 days before transplantation, respectively.

To obtain a healthy control population, we analyzed anonymized samples from 87 random unpaid blood donors (Table 2). For each donor, 2 follow-up serum samples were studied, collected 1 year apart (T0 and T12). The donors were matched for age and sex with the transplant patient study population (Table 1).

Total nucleic acids were extracted from 200 µL serum by using the MagNA Pure LC Total Nucleic Acid Isolation Kit–High Performance and MagNA Pure LC Instrument (Roche Diagnostics, Indianapolis, IN, USA). To monitor the quality of DNA extraction and potential PCR inhibition, we added low concentrations of phocine herpesvirus (28) to the lysis buffer. DNA was eluted in a final volume of 100 μL elution buffer, of which 10 μL was used as input for real-time quantitative PCR (qPCR).

Primers and Taqman probes were designed by using Beacon Designer software (Premier Biosoft, Palo Alto, CA, USA). For HPyV9, we used the following primers and probe, located in the viral protein (VP) 1 gene amplifying a product of 109 nt: sense primer 5′- CCTGTAAGCTCTCTCCTTA-3′, antisense primer 5′- CCTGATAAATTCTGACTTCTTC-3′, and probe FAM-5′- CTTGTTCTCTGGTCTTATGCCTCA-3′-BHQ-1. For BKPyV, we used the following primers and probe, located in the VP1 gene amplifying a product of 90 nt: sense primer 5′-GAAAAGGAGAGAGTGTCCAGGG-3′, antisense primer 5′-GAACTTCTACTCCTCCTTTTATTAGT-3′, and probe FAM-5′-CCAAAAAGCCAAAGGAACCC-3′-BHQ1.

The BKPyV qPCR and phocine herpesvirus PCR were duplexed for DNA quality and potential PCR inhibition monitoring. Furthermore, the BKPyV qPCR was validated to detect BKPyV genotypes I–IV. qPCR reactions were performed in a total volume of 50 μL, containing 25 μL HotStart Taq mastermix (QIAGEN, Hilden, Germany), 0.5 µmol/L of each primer, 0.35 µmol/L BKPyV probe or 0.4 µmol/L HPyV9 probe, and 3.5 mmol/L MgCl₂. Reactions were performed by using a CFX96 real-time detection system (Bio-Rad, Hercules, CA, USA) with the following cycle conditions: 15 min at 95°C followed by 45 cycles of amplification (30 s at 95°C; 30 s at 60°C for HPyV9 qPCR and 55°C for BKPyV qPCR; 30 s at 72°C). For quantification, a standard of pGEX 5×3 HPyV9 VP1 plasmid (Genscript, Piscataway, NJ, USA) and of a quantified BKPyV-positive urine sample were used. Analytical sensitivity of the HPyV9 and BKPyV qPCRs was ≈10 copies/mL. CMV load was measured as described (29) with minor adjustments: 0.5 µmol/L of each primer, 0.2 µmol/L probe, and HotStart Taq mastermix (QIAGEN) was used with the following cycle conditions: 15 min at 95°C, followed by 45 cycles of amplification (5 s at 95°C, 15 s at 63°C, 15 s at 72°C). Analytical sensitivity of the CMV qPCR was ≈100 copies/mL. On each plate, 3 negative controls were included; these controls tested negative in all PCR assays. PCR results with a cycle threshold ≥40 were considered negative.

HPyV9-positive PCR samples were confirmed by sequencing. PCR products were cloned by using the TOPO TA Cloning Kit (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instructions and subsequently sequenced. Sequence reactions were performed by using the BigDye Terminator Kit (Applied Biosystems, Foster City, CA, USA) and analyzed on an ABI Prism 3130 Genetic Analyzer (Applied Biosystems).

To detect IgG seroresponses against the major capsid protein VP1 of HPyV9, we performed an antibody-binding assay using Luminex xMAP technology (30), as described (26). Briefly, the assay is based on cross-linking of glutathione to casein, which is subsequently coupled to fluorescent polystyrene beads (Bio-Rad). Glutathione S-transferase HPyV9 VP1 fusion protein was affinity purified on the beads. Serum samples were tested in a 1:100 dilution, and VP1-bound antibodies were detected with biotinylated goat anti-human IgG (H+L; Jackson Immuno Research, West Grove, PA, USA), followed by streptavidine-R-phycoerythrin (Invitrogen). Finally, the beads and the phycoerythrin signal were analyzed in a Bio-Plex 100 Analyzer (Bio-Rad), which gave results in median fluorescent intensity (MFI). For background correction, MFI values of glutathione S-transferase alone were subtracted to obtain HPyV9 VP1–specific signals. Quality control was performed on each plate with a serum pool consisting of 4 serum samples that had been analyzed in a 1:4 serial dilution, starting with a dilution of 1:100 up to 1:409,600. Little interplate variance was observed.

The cutoff value of the antibody-binding assay was defined on the basis of a group of healthy children 0.5–2 years of age and determined as described by van der Meijden et al. (26). The transplant patients and the healthy blood donors were analyzed in 2 independent antibody-binding assays. HPyV9 cutoff values of 252 MFI and 311 MFI were determined for the transplant patient group and the blood donor group, respectively.

Differences between groups in terms of HPyV9 DNA or seroprevalence were assessed by using the Pearson χ² or Fisher exact test, as appropriate for population size. Independent Student t tests or analyses of variance were used for comparisons of mean values between groups. Occurrence of HPyV9 viremia at transplantation was calculated by using the Kaplan-Meier method, with the time from transplantation to the next detected HPyV9 DNA as the outcome variable for HPyV9-seronegative and -positive patients at baseline (T0). HPyV9 seroconversion at transplantation was calculated by using the Kaplan-Meier method with the time from transplantation to the next seropositive sample as the outcome variable for HPyV9 viremic and nonviremic patients during follow-up. For all tests, 2-tailed p values ≤0.05 were considered significant. The statistical analyses in this study were performed by using SPSS 20 (IBM, Armonk, NY, USA) and Prism 3 statistical software (GraphPad Software Inc., San Diego, CA, USA).

To determine HPyV9 viremia in the transplant patients, we assessed the presence of HPyV9 DNA in the complete sample set. During the 18 months after transplantation, HPyV9 DNA was detected at some point in 21 (20.8%) of the 101 patients (Table 2). No significant difference in the detection of HPyV9 DNA was observed between kidney and kidney–pancreas transplant patients (20.5% vs. 22.2%, respectively). For 3 (3.0%) patients, results were positive for consecutive serum samples; persistent HPyV9 DNA detection throughout the follow-up period was observed for 1 patient. Shortly after transplantation (T0, on average 11 days after transplantation), HPyV9 DNA was detected in 3.0% of patients. Detection of HPyV9 DNA peaked 3 months after transplantation (7.1% positivity) and gradually decreased to 2.5% 18 months after transplantation (Figure 1; Table 2).

To explore the possibility that the observed viremia was not related to the transplant and immunosuppression but to the underlying cause of the kidney disease (e.g., diabetes), we retrieved and analyzed pretransplantation serum samples for 65 (64%) of the 101 transplant patients. In addition, the group of 87 healthy blood donors was analyzed for the presence of HPyV9 DNA. No DNA was detected in either of these sample sets. To confirm the HPyV9-specificity of our PCR findings in the transplant patients, we cloned and sequenced 13 of the 27 HPyV9-positive PCR products, 109 nt in length. The results revealed a complete match with the described HPyV9 DNA sequence in GenBank (accession no. NC_015150) for 12 of 13 samples. In 1 sample, a single nonsynonymous nucleotide mismatch was observed (A→G at position 2403), resulting in an I321V amino acid mutation in the VP1 capsid protein.

The mean HPyV9 DNA load after transplantation was 137 copies/mL (range 25–530 copies/mL). On average, the highest viral loads were observed immediately after transplantation (Figure 1). The kidney–pancreas transplant patients tended to show higher HPyV9 DNA loads than did kidney transplant patients (mean values of 250 and 135 copies/mL, respectively; Table 2), but this difference was not statistically significant (p = 0.123 by Student t test).

HPyV9 seroresponses were analyzed for the complete sample set. At baseline, just after transplantation (T0), 33% of transplant patients were HPyV9 seropositive. This percentage corresponds to the percentages that we measured in healthy blood donors (29%) and in 45 pretransplantation serum samples (31%). However, at 1 year after transplantation (T12), the seropositivity rate for transplant patients rose to 46%. This percentage differed significantly from the rate measured for blood donors, which remained stable at ≈30% during 1 year of follow-up (p = 0.029 by χ² test) (Figure 2, panel A). In total, 15 (15%) of 101 transplant patients seroconverted during follow-up (Technical Appendix Figure 1); these patients represent 23% (15/66) of the patients who were seronegative at baseline.

The intensity of measured HPyV9 serologic responses also increased after transplantation, whereas HPyV9 seroreactivity in blood donors was lower at baseline and remained low within a comparable follow-up period of 1 year (Figure 2, panel B). Mean HPyV9 seroreactivity in the 1-year follow-up samples was significantly higher for the transplant group than for blood donors (p = 0.008 by Student t test). Further analysis of the transplant population revealed that kidney–pancreas transplant patients in particular were responsible for the observed increase in HPyV9 seroreactivity after transplantation (Figure 2, panels C, D). Kidney–pancreas transplant patients had lower mean seroreactivity at baseline than did kidney transplant patients (Figure 2, panel D); the relative increase of seroreactivity in kidney–pancreas transplant patients was confirmed by analyzing the complete dataset with a mixed model analysis (p = 0.003; data not shown).

Because HPyV9 DNA detection and seroresponses increased after transplantation, we investigated the correlation between these parameters. Comparable proportions of patients who were HPyV9-seropositive and -seronegative at baseline became HPyV9 DNA-positive during follow-up (6/35 [17%] and 15/66 [23%], respectively; p = 0.510 by χ² test) (Figure 3, panel A), and the measured mean viral loads were comparable for the 2 groups (169 and 152 copies/mL, respectively; p = 0.798 by Student t test). Furthermore, we analyzed whether the presence of HPyV9 DNA influenced HPyV9 seroreactivity during follow up and found no association (Figure 3, panel B); we also found no association when we compared HPyV9 DNA positivity between high and low seroresponders (above and below median MFI) and seroconverters (data not shown). Stratified analyses for kidney–pancreas and kidney transplant patients did not alter the lack of association. HPyV9 DNA and seroreactivity profiles for patients who seroconverted and/or became viremic during follow-up are shown in the online Technical Appendix.

Of 541 samples tested, 225 (42%) were BKPyV DNA positive; these samples came from 86 (85%) of the 101 transplant patients. HPyV9 DNA was detected significantly more frequently in BKPyV DNA–positive samples than in BKPyV DNA–negative samples (9.8% vs. 1.6%, respectively; p<0.001 by χ² test) (Figure 4, panel A). During follow-up, HPyV9 DNA was more often detected in BKPyV DNA–positive patients than in BKPyV-negative patients (23.3% vs. 6.7%%, respectively; p = 0.185 by Fisher exact test) (Figure 4, panel B). Furthermore, we divided BKPyV viremic patients into 2 groups, those with high (>10³ copies/mL) and low (<10³ copies/mL) BKPyV DNA loads, and found HPyV9 DNA–positive patients were overrepresented among patients with high BKPyV loads (p = 0.001 by Fisher exact test; Figure 4, panel C). For 11 (55%) of 20 co-infected patients, BKPyV viremia coincided with HPyV9 viremia; for 8 (40%), BKPyV viremia preceded HPyV9 viremia.

We additionally assessed the presence of viremia caused by CMV, a herpes virus that is not phylogenetically related to HPyV9 and not particularly related to urinary tract infections but that frequently reactivates during immunosuppressive drug use after transplantation. CMV DNA was detected at some point after transplantation in 27 (27%) of the 101 patients (43 [8%] of 541 tested samples). The proportion of HPyV9 DNA–positive samples was similar among CMV DNA positive and negative samples (7% and 5%, respectively; Figure 4, panel A), and no associations were found when comparing HPyV9 DNA positivity and seropositivity among patients who were negative or positive for CMV DNA (Figure 4, panel B; data not shown).