Serum samples (n = 446) were collected from healthy subjects (HSs), mean age = 49.73, at the Clinical Laboratory Analysis, University Hospital of Ferrara. Sera were taken from discarded laboratory analysis samples, after routine analyses at the University Hospital, Ferrara, Italy, before their destruction by incineration. The hospital records annotated these serum samples as belonging to HS. Indeed, parameters of blood analyses were all in the normal index range.

Sera were collected anonymously, coded with indications of age and gender only. The County Ethical Committee, Ferrara approved the project. Written informed consent was obtained from all participants at the time of the hospital admission. All serum samples were stored at −20°C until testing.

Computer-assisted analyses enabled two specific BKPyV peptides to be selected from the late viral region by comparing VP1 proteins from BKPyV, with amino acids from JCPyV and SV40 polyomaviruses, which are highly homologous to BKPyV, as well as with other less homologous polyomaviruses (http://blast.ncbi.nlm.nih.gov; Figure 1; Figure S1 in Supplementary Material). Preliminary ELISA results indicated that the two peptides did not cross-react with the SV40 and JCPyV hyperimmune sera employed (see data in the Section “Results”). The amino acid (a.a.) sequences of the two peptides, known as VP1-L and VP1-M, which start at the N-terminal with a.a. L and M, respectively, are as follows:

VP1-L: NH2—LKLSAENDFSSDSPERK—COOH

Amino acid sequences of BKPyV viral capsid protein 1 (VP1), named peptide L viral capsid protein 1 (VP1 L) and VP1 peptide M (VP1 M), were characterized regarding stable secondary structure formation. Analysis was carried out by PSIPRED server (21–23). Moreover, peptide sequences were mapped on native virion proteins to verify structural similarities. VP1-L and VP1-M, in their monomeric forms, were obtained from computational prediction carried out by I-TASSER server (24–26). Molecular visualizations were performed by PyMOL (PyMOL Molecular Graphics System, Version 1.3, Schrödinger, LLC). Computational tools were available through ExPASy server (27).

Indirect ELISA was developed and standardized to detect specific antibodies against BKPyV VP1 in human sera using VP1-L and VP1-M synthetic peptides.

A BKPyV viral working stock, Gardner strain (1, 19), for the HAI assay was obtained from infected Vero cells (African green monkey kidney cells, ATCC catalog number CRL-1587) as described before (19, 30, 31). The viral titer, determined by HA assay, was 1.6 × 10² hemagglutinating units, corresponding to 1.6 × 10⁶ plaque-forming units (PFUs)/ml (19). BKPyV was employed as antigen in both HA and HAI assays.

BK polyomavirus HA and HAI titrations were carried out as described in detail elsewhere (19, 30, 31). The HA titer was calculated based on the highest virus dilution that gave complete HA. The HAI assay allows the detection of serum antibodies against the polyomavirus BKPyV, which abolish its agglutination property. The HAI titer was defined as the highest dilution of each serum sample that inhibited viral HA completely.

Permissive Vero cells were used as described and modified slightly as follows: cells were grown and propagated in BioWhittaker^® RPMI 1640 with l-glutamine medium (RPMI), supplemented with 10% fetal bovine serum (FBS) at 37°C in a humidified atmosphere with 5% CO₂ (19). BKPyV infectivity neutralization was carried out by incubating each human serum, diluted at 1:20 in phosphate-buffered saline (PBS) with 5 × 10⁶ PFU of BKPyV at 37°C for 30 min. Then, the suspension was added to Vero cell monolayers for 2 h at 37°C. The inoculum was removed, cells were washed three times with RPMI, and overlaid with the medium containing 1% FBS. Infected Vero cell monolayers were maintained in RPMI + 1% FBS for 3 weeks, until cytopathic effect (CPE) appearance (19, 32). Each sample was tested in duplicate and with different sera with distinct levels of antibodies (OD reading). The neutralization assay included the following controls (i) BKPyV only in PBS and (ii) cells only in PBS. Cultures were observed using a light microscope for CPE presence for 3 weeks.

BK polyomavirus CPE was assessed by three cell biologists of our working group. CPE inhibition was graded as negative, grade 0 (complete CPE), grade 1–2 as weak inhibition, grade 3 as moderate inhibition, grade 4 as intermediate inhibition, and grade 5 as strong inhibition.

Differences among age groups were statistically evaluated. Statistical analyses were performed using Prism 4.0 software (GraphPad, San Diego, CA, USA). Data are presented as a percentage of the positive samples. The 95% confidence intervals of the percentage of positive samples are also reported. Differences among proportions were calculated by χ² test for independence in the contingency tables. The small sample size was statistically analyzed using χ² with Yates’ correction. The correlation between the OD value and BKPyV CPE was evaluated by non-parametric Spearman analysis, which indicates Spearman r and p values.

The serologic profile of serum antibody reactivity were analyzed with one way Anova analysis, and Newman–Keuls Multiple Comparison Test (OD mean, 95% CI).

Computational analysis showed that the two linear peptides (http://blast.ncbi.nlm.nih.gov; Figure 1; Figure S1 in Supplementary Material) are characterized by their secondary structure, as follows (PSIPRED server). Peptide VP1-L has a random coil secondary structure, whereas it does not contain any alpha helix or beta sheet domain (Figure 2A, on the left). Peptide VP1-M forms a stable secondary structure from a.a. 9 to a.a. 13, i.e., ₉QKVHE₁₃, where an alpha helix domain is present (Figure 2A, on the right).

Tertiary structures of BKPyV viral capsid protein VP1 (Figures 2B–D), selected among computationally determined structures, presented C-scores of 0.51, with an estimated model accuracy: 0.78 ± 0.10 (TM-score), 5.5 ± 3.5Å (RMSD). For details, see Ref. (24–26). Mapping two linear peptides of the inferred protein structures showed that VP1-L does not fold in any structure type, whereas VP1-M folds in a short alpha helix domain (Figures 2B,C). The three-dimensional graph from L and M peptide structures, shows the specific amino acids exposed to the protein surface (Figure 2D), which may interact with the environment.

An indirect ELISA was setup using synthetic peptides in order to check whether human sera contain IgG antibodies, which react to BKPyV peptides and to determine the spread of BKPyV infection in humans. These synthetic peptides correspond to BKPyV VP1 mimotopes. The unrelated human peptide hNPS was employed as a negative control (19).

In the first step of this investigation, indirect ELISA was used to test serum samples taken from HS ranging from 18 to 90 years old. The sera were diluted at 1/20 in order to react to BKPyV VP1 epitopes. Serum samples, which reacted to BKPyV VP1-L mimotope reached an overall prevalence of 80%. The same assay was then performed to detect IgG class serum antibodies against BKPyV VP1 epitopes, which are known as VP1-M. It turned out that serum samples reacted with the VP1 M peptide with the same prevalence, 80%, as had been detected previously for the VP1 L peptide. Conversely, seronegative samples for the BKPyV VP1 L peptide failed to react with BKPyV VP1 M epitopes. The exceptions were negligible and were represented by a few serum samples, which were found to be negative for VP1 L, while testing positive for VP1 M peptide, and vice-versa. The difference was not statistically significant (P > 0.05) (Table 1).

Combining the BKPyV-positive sera, both for the VP1 L and VP1 M peptides, the overall prevalence was 72%, with no difference between male and female (Table 1). The two indirect ELISA tests, with two distinct VP peptides gave overlapping results, thus confirming the presence of anti-BKPyV VP antibodies in human sera from HS (Table 1).

A prevalence selection corresponding to 76, 81, 65, and 65%, within the cohorts ranging from 18 to 90 years old, was observed in subjects 18–39, 40–50, 51–61, and 62–90 years old, respectively. Interestingly, serum antibodies prevalence against BKPyV VP declined in cohorts of individuals ranging from 51 to 61 and 62 to 90 years old, with a percentage of 65% (Table 1).

BK polyomavirus-positive sera tested by indirect ELISA diluted at 1/20 had a general cutoff, range of 0.19 OD, using a spectrophotometric reading. This cutoff represents the discriminatory value between BKPyV-negative samples (below OD 0.19) and BKPyV-positive samples (above OD 0.19). The positive control, represented by BKPyV hyperimmune rabbit serum, had an OD of up to 0.3, while the two JCPyV and SV40 hyperimmune sera, employed as negative controls, had an OD of less than 0.01–0.02. The OD value was usually in the 0.01–0.02 range, which is consistent with the OD for BKPyV-negative sera.

The human neuropeptide hNPS (33), which is unrelated to BKPyV, was employed as a negative control peptide in the ELISA experiments. Data indicate that this negative control peptide did not react with 322 BKPyV-positive sera, nor with 124 BKPyV-negative samples. The OD value was usually in the 0.01–0.02 range, which is consistent with the OD for BKPyV-negative sera.

In this investigation, BKPyV-positive sera (65%), which was detected both in 51- to 61 and 62- to 90-year-old subjects, was lower than that detected in the other cohorts of healthy individuals (Table 1). These differences are statistically significant when compared to other cohorts from different age groups (18–39; 40–50 years old) (Table 1). The reduced prevalence of BKPyV-positive sera in individuals 51–90 years old could be ascribed to natural age-dependent partial immune decline. Indeed, it is well established that the immune system physiologically declines with age, rendering individuals less responsive to infection. An alternative interpretation of the result, as suggested by other studies (34, 35) is that the low prevalence of BKPyV antibodies in sera from the elderly could depend on low BKPyV antibody stability overtime. Serologic profiles of serum antibody reactivity to BKPyV mimotopes are presented in Figure 3. Serologic profile of serum antibody reactivity to BKPyV mimotopes VP1 L (A) and VP1 M (B) and VPs L + M (C). Data are OD values at 405 nm of serum samples diluted 1:20, detected in indirect ELISA. In scatter dot plotting, each plot represents the dispersion of OD values to a mean level indicated by the line inside the scatter with SEM for each age group of subjects analyzed (mean OD + SEM). The difference in OD mean values were not statistically significant in sera from HSs (P > 0.05; Anova analysis, and Newman–Keuls Multiple Comparison Test).

A fraction of randomly chosen human sera (n = 87), which had been investigated by indirect ELISA were also analyzed using the well-established technique, known as HAI assay (19, 20, 30). Serum samples, serially twofold diluted from 1:16, 1:32, 1:64, to 1:128, were analyzed by HAI to evaluate both the presence of antibodies against BKPyV and their titer.

The 1:128 dilution was chosen since other lower dilutions, such as 1:32 or 1:64 contain antibodies at higher concentrations, which may give rise to false positive reaction; on the other hand, higher dilutions could carry antibodies at much lower concentration giving false negative reaction (19, 20, 30). The seroprevalence of BKPyV-positive samples, diluted at 1:128, was 63% (55/87). Seropositivity obtained with indirect ELISA on the same sera was 62% (54/87) (Table 2). The seroprevalence of BKPyV-positive in sera from HS, studied by HAI, did not differ statistically from that determined by indirect ELISA (P > 0.05). Interestingly, BKPyV-positive samples concordance from the two techniques was 98%.

An inhibition test was performed in order to check whether immune sera, containing antibodies against BKPyV, neutralized its infectivity. To this end, 15 immune sera, which ranged between a 0.19 and 0.65 OD reading, out of 322 BKPyV-positive samples, were selected to test their ability to inhibit the CPE in permissive BKPyV-infected Vero cells, along with 4 out 124 BKPyV-negative sera, with an OD of below 0.19, employed as control. In this context, it is worth recalling that BKPyV CPE generates many cytoplasmic vacuoles in the first phase of the infection and then cell lysis of Vero monolayers (19). To avoid any possible cross-reactivity, the 15 BKPyV-positive sera analyzed for their neutralization activity, were selected among SV40- and JCPyV-negative sera, as determined by indirect ELISA for SV40 VP (19) and HAI assay for JCPyV (unpublished results). The data are in agreement with previous reports. JCPyV- and SV40-positive sera showed no higher OD values toward VP1 L and VP1 M peptides than those, which tested negative. The neutralization effect of BKPyV-immune sera is apparent in infected cells when BKPyV CPE is abolished or hampered. The CPE inhibition test was performed by mixing the serum sample with BKPyV virions. Then, the mixture was employed as a viral inoculum to infect Vero cells. In inhibition experiments on BKPyV CPE, the positive and negative controls were (i) cells cultured without human serum and without BKPyV and (ii) cells infected with BKPyV, cultured without human serum, respectively. CPE inhibition in BKPyV-infected cells indicated that tested sera were able to neutralize BKPyV infectivity. Indeed, three sera with a mean OD = 0.65, taken from BKPyV-positive samples, completely inhibited BKPyV CPE (Figure 4A), while 12 sera with a mean OD = 0.25–0.55, only partially inhibited BKPyV CPE, with a different degree (Figures 4B–E). Four samples with an OD < 0.19 did not inhibit BKPyV CPE (Figure 4F). There was a correlation between the OD reading and the CPE inhibition activity of the immune sera. Indeed, the correlation between OD and CPE inhibition is statistically significant (Spearman r = 0.988 and P < 0.0001) (Figure 5).

This assay detects functional antibodies to VP1 L that are capable of neutralizing BKPyV infectivity in tissue culture.