3. Results
ddPCR detected BPV DNA in 95% (38/40) of the blood samples tested, whereas qPCR detected the virus in 57.5% (23/40) of the same samples, with the percentage difference being of statistical significance ( P ≤ 0.05). Specifically, the BPV genotypes were identified in 24 of the 25 (96%) healthy cows without any bladder lesions and in 14 of the 15 (∼ 93%) cows affected by naturally occurring PV-associated bladder tumors leading to chronic hematuria (P = 0.68). Furthermore, ddPCR revealed a single genotype in ∼ 37% (14/38), and multiple genotypes in ∼ 63% (24/38), of the cows harboring BPV DNA. qPCR failed to detect the BPV DNA in samples tested negative by ddPCR. In fact, qPCR detected BPV DNA in only 6 (∼ 16%) of the 14 samples detected by ddPCR to be infected with a single BPV genotype. Of the BPV co-infections detected by ddPCR, 13 (∼ 54%) were by two genotypes, 8 (∼33%) were by three genotypes and 3 (∼ 12.5%) were by four genotypes. By contrast, qPCR revealed BPV co-infections by multiple genotypes in only six samples; namely, four samples with co-infection caused by two genotypes, and two samples with co-infection caused by three and four genotypes, respectively. In particular, qPCR failed to detect co-infection by two genotypes in 10 of the 13 samples revealed by ddPCR. Furthermore, qPCR found only two BPV genotypes in a sample in which ddPCR detected four genotypes. Neither ddPCR nor qPCR were able to detect a single infection by BPV-1. Tables 2 summarizes these results, whereas Table 3 displays the BPV genotypic frequencies.
With regard to single and multiple BPV infections, BPV-13 was the most prevalent genotype as it was detected in ∼ 82% (31/38) of the samples by ddPCR and in ∼ 37% (14/38) of the samples by qPCR, with the percentage difference between the two assays being statistically significant (P = 0.005). BPV-2 was the second most-common genotype, being detected in ∼ 53% (20/38) of the samples by ddPCR and in ∼24% (9/38) of the samples by qPCR (P = 0.001). BPV-14 was detected in ∼ 40% (16/38) of the samples by ddPCR and in ∼13% (5/38) of the samples by qPCR (P = 0.005). Finally, BPV-1 was the least common of the four genotypes being found in a limited number of samples. Indeed, BPV-1 was detected in ∼24% (9/38) by ddPCR and 18% (7/38) by qPCR (P = 0.52).
ddPCR was also used for quantification of BPVs, with the results showing clear distinctions between positive FAM (blue), and VIC (green) and negative (gray) droplets (Figures 1-4). There were also differences in the fluorescence amplitude range of the background (negative) droplets among the δPV samples; that is 3000-8000 for BPV-1, 4000-9000 for BPV-2, 2500-5000 for BPV-13, 600-2000 for BPV-14. The numbers of BPV DNA copies per microliter of blood in the samples from the 24 healthy cows and the 14 cows suffering from PV-associated bladder tumors are summarized in Table 4.
4. Discussion
To the best of our knowledge, this is the first study that has performed both a qualitative and a quantitative estimation of circulating BPV DNA using the sensitive ddPCR approach. The BPV E5 DNA was detected in 95% of the blood samples tested (38/40); in particular, single or multiple BPV genotypes were detected and quantified in 14 of the 15 cows suffering from naturally occurring PV-associated, noninvasive, papillary urothelial tumors and in 24 of the 25 healthy cows belonging to the same cattle herd as the tumor-carrying animals. Both groups of cows shared the same grazing lands. The ddPCR assay quantified the number of E5 DNA copies per microliter of blood samples of all four bovine δPV genotypes. Our study showed that, as for human samples (Cheung et al., 2019), ddPCR could be a diagnostic procedure capable of detecting otherwise undetectable BPVs by liquid biopsy. To date, the prevalence and burden of bovine δPV have been investigated by PCR and qPCR only. A BPV-2 prevalence of 50% and 10% in blood from cattle with and without BPV-associated neoplastic bladder lesions, respectively, has been reported using semi-nested PCR (Wosiaki et al., 2005). In addition, Roperto et al. (2008) reported a 78% and 39% prevalence of BPV-2 in blood from cattle with urinary bladder tumors and from healthy cattle, respectively. High percentages of BPV-1 and -2 in the blood of cattle suffering from cutaneous papillomatosis have also been revealed by PCR (Diniz et al., 2009; Santos et al., 2016). Furthermore, in veterinary medicine, a few BPV detection and quantification studies have been performed using qPCR. For example, Yuan et al. (2007) reported quantitative data about BPV-1 and -2 in equine sarcoids and bovine bladder tumors, respectively. Pathania et al. (2012) studied the BPV-2 load in urinary bladder tumors from cows suffering from enzootic bovine hematuria (EBH), and Cota et al. (2015) reported quantitative data on BPV-2 infection in bovine urinary bladder lesions. In addition, Pangty et al. (2010) and Kumar et al. (2015) quantified the BPV-1, -2 and -5 DNA copy numbers in some cutaneous warts and wart-like lesions of the upper gastrointestinal tract in cattle and buffaloes.
We compared the diagnostic sensitivity of ddPCR with that of qPCR in evaluating the same liquid biopsy. In our study, ddPCR showed superior sensitivity compared with qPCR, the latter so far believed to be the gold standard for measuring PV DNA (Isaac et al., 2017). All percentage differences in the detection of circulating BPV E5 DNA between these two diagnostic procedures were of statistical significance. ddPCR appeared to be an assay that would allow for the remarkably improved diagnosis of BPV infections in comparison with both PCR and qPCR. Therefore, our results suggest that ddPCR is by far the most sensitive and accurate assay for BPV detection, in accordance with its recently shown performance for oncogenic human papillomavirus (HPV) detection (Biron et al., 2016).
No statistically significant differences in the viral load estimations were found between the blood samples from healthy and bladder tumor-affected cows, suggesting that circulating BPVs represent a primary blood infection rather than being the consequence of virus release from bladder tumors. Therefore, it is conceivable that BPVs can spread through the blood to infect all organs of the cow, including the bladder, thereby representing the major pathway responsible for a latent bladder infection by the virus. However, it is also possible that other minor pathways may occur, such as a secondary infection deriving from the paragenital area harboring the BPV infection (Campo, 2006). Although PVs are strictly species-specific, experimental data showing that these viruses can produce infection via the bloodstream in some animal models appear to strengthen our suggestions (Cladel et al., 2019). Furthermore, it is believed that blood could be an actual route of HPV dissemination as the virus has been found in peripheral blood mononuclear cells from healthy blood donors (Bodaghi et al., 2005; Chen et al., 2009; Vergara et al., 2019). A further biological feature characterizing BPV-associated bladder carcinoma would seem to corroborate our suggestion. It has been suggested that the circulating viral DNA detectable in human patients reflects the aggressive biological behavior of HPV-associated carcinoma thus representing a prognostic tumor biomarker (Jeannot et al., 2016; Damerla et al., 2019; Veyer et al., 2019). BPV-associated bladder carcinomas of cattle have a relatively low metastatic potential, as only 8-10% of them migrate to distant organs (Pamukcu, 1974; Roperto et al., 2010). We detected circulating BPV DNA in cows affected by noninvasive urothelial tumors, including papilloma and papillary urothelial neoplasms of low malignant potential, which showed that the circulating BPVs appeared to be independent of the status of urothelial neoplasms. Therefore, circulating BPVs cannot be used as an actual prognostic biomarker of urothelial tumors in cattle.
ddPCR may be an essential tool for improving diagnostic procedures, helping us to accurately identify the genotypic distribution of BPV and to better understand the territorial divergence, if any, of the BPV prevalence in different areas and allowing us to gain insights into the molecular and ecological epidemiology of infectious diseases, including viral ones. In particular, ddPCR may be a better choice than qPCR for clinical application being able to detect otherwise undetectable PV infections and co-infections. This baseline information is necessary before any decision about livestock health management can be made. ddPCR can provide great advantages toward informing on the necessary measures to adopt in order to reduce the risk of BPV infection and/or co-infection and improving our knowledges about the molecular mechanisms of the disease. Of note, whether PV co-infections increase the risk of disease is not clear, as controversial results exist (Chen et al., 2019). Indeed, it has been suggested that PV co-infections are prone to increase the risk of disease, as multiple PV genotypes might act synergistically (Trottier et al., 2006). Recently, it was shown that superinfection exclusion may occur during PV co-infections, whereby a PV genotype would be able to block or exclude other PV genotypes, as was described for many other viruses (Biryukov and Meyers, 2018).
In conclusion, ddPCR provides new perspectives for the clinical diagnosis of BPV infections. Observational studies using diagnostic tools, such as ddPCR, are required to gain further insights into the biology of BPVs. Moreover, these studies could verify the role of BPVs also as non-oncogenic infectious agents. In this context, there is an urgent need to study whether there are relationships between BPV infections and novel molecular mechanisms that underlie emerging non-oncogenic diseases, including the reproductive disorders that are a major problem in farm animals. The accurate interpretation of such association will have significant implications for the therapeutic, prognostic, and management considerations in cattle breeding and general animal husbandry.
Acknowledgments: The authors wish to thank Dr G. Borzacchiello, Department of Veterinary Medicine and Animal Productions, Naples University, and Dr A. Venuti, IRCSS Regina Elena National Cancer Institute, Rome, for providing BPV-1 and BPV-2 positive samples as a kind gift; Dr G. Salvatore of the Regione Basilicata, Dr P. Sarnelli of the Regione Campania, and Dr S. Morace of the University of Catanzaro “Magna Graecia” for their technical help.
Funding: The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This project was partly supported by grants from the Regione Basilicata and the Regione Campania.
Declaration of Conflicting Interests; The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data Availability Statement: All data supporting this manuscript are reported and can be found in our paper.