Detection of XMRV in normal and tumor tissue of prostate cancer patients from the Southern United States is dependent on specific PCR conditions

Abstract

Background

There are questions regarding the prevalence of xenotropic murine leukemia virus-related virus (XMRV) in PCA patients and its association with the RNASEL R462Q polymorphism. We therefore investigated whether XMRV infection could be found in PCA patients from the southern United States (U.S.), and sought to verify the association with the R462Q.

Methods

Prostatic tissue specimens of 144 PCA patients from the southern U.S. were genotyped for R462Q by real time PCR allelic discrimination and screened for XMRV proviral DNA by nested PCR specific for the env gene.

Results

The R462Q polymorphism was found at an allelic frequency of 33%. XMRV was detected in 32 (22%) of the 144 patients. Patients were significantly more likely to test positive for XMRV in both tumor and normal tissue rather than either alone (κ = 0.64). Positivity for XMRV was not significantly correlated with the R462Q polymorphism (p = 0.82) or clinical pathological parameters of PCA, including Gleason score (p = 0.29).

Conclusions

XMRV is detectable in normal and tumor prostatic tissue from PCA patients independent of R462Q. The presence of XMRV in normal tissue suggests that infection may precede cancer onset.

INTRODUCTION

Prostate cancer (PCA) is a leading cancer in men in Western countries, accounting for 25% of incident cancers in American men in 2009 [1, 2]. Despite the high prevalence and gravity of this disease, there are currently few suitable biomarkers to distinguish between cancers with high and low recurrence potentials, and to determine whether patients require immediate therapeutic intervention or should only be under periodic observation [3]. Such biomarkers for classifying prostate cancers into different treatment categories may depend on the underlying etiology of each case. Epidemiological evidence suggests that environmental factors, such as diet and infectious agents may contribute to chronic inflammation of the prostate, and tumorigenesis [2]. An infectious etiology for PCA is supported by the linkage of hereditary PCA to the common R462Q polymorphism in the RNASEL gene. The polymorphism, which has been reported to be elevated among familial PCA patients, results in a reduced-activity variant of the innate antiviral factor, ribonuclease L [4, 5]. In one study, the R462Q polymorphism was implicated in up to 13% of PCA cases [4]. Correspondingly, xenotropic murine leukemia virus-related virus (XMRV) was discovered by searching for viruses in PCA patients homozygous for R462Q with a microarray (Virochip) designed to detect the most conserved regions of all viral families [6]. An expanded screen with PCA patients harboring wild type RNASEL alleles indicated a strong correlation with the R462Q variant, thus establishing a connection between infection and the disease [6].

The linkage of XMRV to PCA through the RNASEL R462Q polymorphism has become the subject of controversy as recent reports indicate that infection occurs independent of R462Q [7, 8]. Additional studies are needed to determine whether RNASEL genotype is a reliable indicator of susceptibility to XMRV infection. Furthermore, there is no agreement about the cell types infected in the prostate. XMRV was originally discovered exclusively in the non-malignant stromal and hematopoietic cells adjacent to the carcinoma [6]. By contrast, another study found XMRV primarily in prostate carcinoma cells [7]. Additional studies are therefore needed to determine whether non-malignant cells are susceptible to infection by XMRV in order to address whether infection may precede tumor initiation. Questions regarding the association of XMRV with PCA have also arisen in light of recent studies that detect little to no presence of the virus in PCA patients [9, 10]. Interestingly, studies that detect XMRV in PCA patients were conducted in the United States (U.S.), whereas those that do not detect the virus were conducted in Germany. These conflicting reports emphasize the need to confirm the presence of XMRV in PCA and to define the geographic distribution of the virus.

Here, we conducted a retrospective study in which we screened a cohort of PCA patients that is unique from those of previous studies with respect to its location within the U.S. Additionally, we selected for patients with a family history of PCA to enrich for carriers of the R462Q polymorphism. The goals of the study were to confirm the presence of XMRV in PCA patients, and to investigate the linkage of XMRV to the R462Q polymorphism. Here, we demonstrate that XMRV is present in 22.2% of PCA patients from the southern U.S., that infection does not correlate with R462Q, and that reliable detection of viral DNA was dependent on particular conditions of PCR. Additionally, we show that XMRV is detectable in both normal and cancer tissue in the prostate, suggesting that the virus does not specifically target transformed cells and that infection may therefore precede cancer onset. If XMRV is shown to promote PCA, it may prove to be a valuable biomarker for clinicians when considering treatment for patients.

METHODS

Prostate cancer cohort and tissue preparation

Frozen prostate tissue cores were obtained from a PCA tissue bank at Baylor College of Medicine. Details of the donor patients have been previously described [3]. All prostatic tissues used in the study were derived from patients who underwent radical prostatectomy and had provided consent in accordance with the Baylor College of Medicine Institutional Review Board. No patients underwent preoperative treatment for their cancer. To enrich for carriers of the R462Q polymorphism in RNASEL, only tissues from patients having at least one first or second degree relative diagnosed with PCA were selected for XMRV screening and RNASEL genotyping. In total, 144 patients were screened for XMRV and the R462Q polymorphism in RNASEL. For 57 of the 144 patients, both normal and tumor tissue were available for screening.

All prostate tissues were prepared for DNA extraction in a separate laboratory from the laboratory in which the infectious XMRV clone VP62 (NC_007815.1) was handled. DNA was extracted from sections of prostate biopsies using the QIAamp DNA Mini kit (Qiagen). All prostatic DNA samples were stored at −20°C immediately following extraction in a laboratory free of amplified or cloned DNA.

RNASEL genotyping

All patients were genotyped for RNASEL G1385A (R462Q) using the Applied Biosystems real time PCR TaqMan SNP assay (Assay ID: C____935391_1_) with TaqMan Universal PCR Master Mix (Applied Biosystems). A 7500 real time PCR system (Applied Biosystems) was used for amplification and analysis of RNASEL genotyping reactions containing 20ng of prostatic DNA. Specimens of predetermined genotypes (homozygous wild type, heterozygous, and homozygous variant) were used as controls for genotyping reactions. All tissues were tested in duplicate.

Provirus screen

A nested PCR assay was developed to screen prostatic DNA for XMRV provirus. First-round Primers, 5′-ACCAGACTAAGAACTTAGAACCTCG-3′ and 5′-AGCTGTTCAGTGATCACGGGATTAG-3′ amplify an 888 bp region containing the 5′ terminus of envelope (env). The nested, second-round primers, 5′-GAACAGCATGGAAAGTCCAGCGTTC-3′ and 5′-CAGTGGATCGATACAGTCTTAGTCC-3′ amplify a 653 bp region encompassing the three variable regions (VR) of env, VRA, VRB, and VRC. First-round reactions contained 650ng of prostatic DNA, 2.5mM MgCl₂, 800μM dNTPs, 100ng of each primer, and 1.5 units of AmpliTaq Gold DNA polymerase (Applied Biosystems) in 50 μl total volume. Two microliters of first-round reactions were transferred to 48μl of a PCR master mix containing 100ng of each second-round primer and the same concentrations of each of the components of the first-round reactions. Thermocycling conditions were as follows: 94°C for 5 minutes; 35 cycles of 94°C for 30 seconds, 56°C for 1.5 minutes, 72°C for 1 minute; and ending with 72°C for 10 minutes. The master mixes for each set of PCR reactions were tested for sensitivity and nucleic acid contamination by incorporation of positive and negative controls, respectively. The positive control consisted of three separate reactions; each with 100ng of DNA isolated from XMRV-infected LNCaP cells diluted 1 to 10³ in uninfected LNCaP cells. The master mix was considered adequately sensitive only if all three positive control reactions tested positive. Negative controls consisted of three separate reactions with H₂O in place of DNA template and three separate reactions of 650ng of uninfected LNCaP DNA. After thermocycling, second round reactions were electrophoresed on agarose gels containing ethidium bromide, and visualized under UV light. All tissues were screened in triplicate, and patients/tissues were considered positive if one or more PCR reactions tested positive.

Cell culture and generation of PCR sensitivity controls

The LNCaP human prostate carcinoma cell line was used to test the sensitivity of the PCR assay and to generate XMRV stocks. XMRV has previously been shown to infect and replicate within this cell line [11]. LNCaP cells were cultured in RPMI 1640 (Invitrogen), 10% heat-inactivated fetal bovine serum (Sigma), glutamine, and penicillin and streptomycin (Invitrogen), and were incubated at 37°C with 5% CO₂.

To generate PCR sensitivity controls, LNCaP cells were transfected with an infectious XMRV clone (VP62), a generous gift from Robert Silverman (Cleveland Clinic) [11]. One day following transfection, the cells were washed with PBS and supplied with fresh media. Two days post-transfection, the conditioned media was passed through a 0.45μm syringe filter and used to infect a new stock of LNCaP cells. The infected cells were cultured for 40 days, splitting them 1:10 every five to seven days. The infected cells and a separate stock of uninfected LNCaP cells were washed with PBS, trypsinized, and mixed together at ratios of 1:100, 1:10³, 1:10⁴, and 1:10⁵ (infected cells : uninfected cells). Without further culturing, total cellular DNA was extracted from the cell mixtures using the QIAamp DNA Mini kit (Qiagen). Extracted DNA was used as template to test the sensitivity of the XMRV env nested PCR assay.

To test for VP62 plasmid contamination in prostate specimens, a set of four primers were designed for nested PCR. The two forward primers, 5′-TCTGGCTAACTAGAGAACCCACTG −3′ and 5′-AATACGACTCACTATAGGGAGACC −3′, were specific to the multiple cloning site of pCDA3.1(−) (Invitrogen). The two reverse primers, 5′-AAGGTAACCCAGCGCCTCTTC −3′ and 5′-GTTACGGTCTGTCCCATGATCTC −3′, were specific to the 5′ terminus of VP62 gag. The VP62 nested PCR assay was found to be capable of detecting 10 plasmids diluted in 600ng of uninfected LNCaP DNA in three of three samples, and 1 plasmid in 600ng of uninfected LNCaP DNA in one of three samples (data not shown).

Cloning and sequencing of patient-derived PCR products

Positive PCR reactions were electrophoresed on agarose gels and extracted using the Qiaex II Gel Extraction Kit (Qiagen) following the manufacturer’s instructions. Extracted PCR products were cloned into pCR2.1-TOPO using the TOPO TA Cloning Kit (Invitrogen) according to the manufacturer’s protocol. The cloned PCR sequences were propagated in NEB 10-beta (New England BioLabs Inc.) E. coli, isolated with the QIAprep Spin Miniprep Kit (Qiagen), and the sequence of the DNA inserts were determined.

Phylogenetic analysis

Env sequences were aligned using Clustal and were trimmed to the same length with gaps. The maximum likelihood tree of env sequences was generated using PhyML [12].

Statistical analysis

Statistical analyses were performed using Stata 10 software. Correlation between XMRV positivity and tissue type was analyzed by measuring the simple kappa coefficient. Correlations between XMRV positivity and Gleason score or seminal vesicle invasion were analyzed using Fischer’s exact test. Correlation between XMRV positivity and extracapsular extension or surgical margin invasion was assessed using the Chi-Square test.

Accession numbers

Sequences of cloned XMRV env genes were deposited in Genbank under numbers GU812341-GU812357. Accession numbers from GenBank (http://www.ncbi.nlm.nih.gov/Genbank/) for other viral sequences are DG-75, af221065; Raucher MuLV, NC_001819; Friend MuLV, M93134; Moloney MuLV, NC_001501.1; MERV Chr12, ac153658; MTCR, NC_001702; and MERV Chr7, ac127565.

RESULTS

Distribution and frequency of RNASEL R462Q

In order to investigate the linkage of XMRV infection to RNASEL R462Q, we obtained prostatic tissue specimens from PCA patients to screen for the virus and the R462Q polymorphism in RNASEL. Since XMRV was originally found to be strongly associated with R462Q, we screened PCA patients with a family history of PCA, which are reported to have an elevated R462Q allelic frequency [4–6]. In total, 144 PCA cases were screened by a real-time PCR-based allelic discrimination assay for the R462Q polymorphism in RNASEL. We found there to be 66 (45.8%) wild type (RR) individuals, 61 (42.4%) heterozygotes (RQ), and 17 (11.8%) individuals homozygous for the Q variant (Table 1). The allelic frequency for R462Q was determined to be 33%, which is intermediate in comparison to previously described PCA cohorts of unselected or sporadic cases (ranging from 38% to 25%) [4, 7, 10].

Table 1.

Summary of RNASEL genotyping and XMRV screen

Total patients	144
PCR-positive patientsa	32 (22.2%)
R462Q allelic frequency
Total patients	33.0%
PCR-positive patientsb	29.7%
Genotypic distributionc
Wild type, RR	66 (45.8%)
Heterozygous, RQ	61 (42.4%)
Homozygous variant, QQ	17 (11.8%)

^aPatients that tested positive by PCR for XMRV regardless of tissue type

^bR462Q allelic frequency for the 32 patients testing positive for XMRV DNA

^cPercentage of patients with respective genotype is indicated in parentheses

XMRV is detected in prostate cancer patients

A highly-sensitive, nested PCR assay for XMRV env was developed to screen patients for XMRV infection. The PCR assay was found to be capable of detecting a single copy of VP62 plasmid (data not shown). We also tested the sensitivity of the PCR assay in the context of integrated provirus. We found that XMRV provirus could be detected at a dilution of one infected human prostate carcinoma cell per 10⁴ uninfected cells in three of three samples using 600ng of DNA (approximately 10⁵ cells). Using the same quantity of DNA, the nested PCR assay was found to be capable of detecting XMRV provirus in one infected cell per 10⁵ uninfected cells in one of three samples (data not shown). Importantly, assuming XMRV provirus is present at a frequency of 0.15 to 1.5%, which has been estimated in previous reports, our nested PCR assay is greater than 15 to 150 times more sensitive than that which is minimally required to detect the virus [6, 7]. Thus, the nested PCR assay is a highly sensitive method to detect XMRV provirus. Each patient specimen was screened in triplicate using 650ng of prostatic DNA. In total, 32 of 144 (22%) patients were found to be PCR-positive for XMRV (Table 1). The majority of tissue specimens that were positive for XMRV tested positive in only one or two out of three replicates (data not shown).

To confirm that XMRV was detected, the 653 base pair env PCR products were sequenced from 17 patients that tested positive (nucleotide sequence alignment listed in the appendix, which appears only in the electronic edition of the Journal). In comparison to reference strain VP62, 3 of the 17 sequences encoded non-synonymous nucleotide differences representing a total of 5 amino acid differences (Figure 1A). With respect to all 17 predicted Env peptide sequences, differences from VP62 were observed at a rate of 0.14%, ranging from 0% to 1.4% (patient PCA1). The high degree of sequence identity with VP62 suggests that positive PCR results for the tissue specimens were due to the presence of XMRV DNA. This was confirmed by phylogenetic analysis of the sequences (Figure 1B). Additionally, we tested for the presence of contaminating VP62 plasmids in the DNA isolated from the patient tissue specimens using a nested PCR assay targeting the pcDNA3.1(−) multiple cloning site/XMRV genome junction. We found no evidence for contamination in specimens scoring positive by PCR for XMRV env (data not shown).

Figure 1.

Sequence analysis of patient-derived PCR products. A, Comparison of the predicted Env protein sequences from the PCR products of 17 patients with the Env sequence of the XMRV clone VP62; variable regions (VR) A, B, and C are indicated; dots indicate identical residues; and a stop codon is indicated by an asterisk. B, Phylogenetic tree of the XMRV patient clones compared with other murine retroviruses.

XMRV is present in cancer and normal tissues

For 57 of the 144 patients, both normal and tumor prostate tissues were available for screening, whereas only tumor tissue was available for the remaining 87 patients. In this subset of patients, XMRV was detected in 21 of 57 (36.8%) normal tissues and in 25 of 57 (43.9%) tumor tissues. The virus was detected exclusively in the normal tissue of three patients, and was detected exclusively in tumor tissue of seven patients, while 18 patients had provirus in both tissue types (Figure 2). Statistical analysis of these results indicate that patients were more likely to harbor provirus in both normal and tumor tissue rather than one or the other (kappa coefficient of agreement = 0.64), suggesting that XMRV does not specifically target tumor tissue in the prostate.

Figure 2.

Distribution of XMRV between normal and cancer tissue of 57 PCA patients. The white circle represents patients from whom XMRV DNA was detected by PCR in normal tissue, the dark gray circle represents patients from whom XMRV DNA was detected by PCR in tumor tissue, and the light gray overlap represents patients from whom both tissue types tested positive for XMRV DNA. Twenty-nine of the 57 patients tested negative for XMRV DNA. Patients were found to be more likely to test positive for both tissue types (simple kappa coefficient = 0.64).

XMRV infection does not correlate with R462Q, Gleason score, or other pathological parameters of prostate cancer

We investigated whether XMRV infection is enriched among carriers of the R462Q polymorphism of RNASEL in this cohort. XMRV was detected in 24.2%, 21.3%, and 17.6% of wild type (RR), heterozygous (RQ), and homozygous variant (QQ) patients, respectively (Table 2). However, infection was not found to be significantly associated with the R462Q polymorphism of RNASEL (Chi-Square, p = 0.82).

Table 2.

XMRV screening by nested PCR for env

PCR	RNASEL genotypea			Total
	RR	RQ	QQ
PCR +	16	13	3	32
PCR −	50	48	14	112
Totalb	66 (24.2%)	61 (21.3%)	17 (17.6%)	144 (22.2%)

^aRNASEL genotypes are: RR, homozygous wild-type, RQ, heterozygous, QQ, homozygous R462Q variant.

^bThe percentages in parenthesis indicate the proportion of XMRV PCR positive specimens.

XMRV infection is reportedly associated with higher Gleason score prostate cancers [7]. We therefore examined whether XMRV infection correlates with tumor grade. The patients in our study consisted of 12, 35, 82, 9, and 6 patients with Gleason scores of 5, 6, 7, 8, and 9, respectively. Although there appears to be a trend between XMRV infection and increasing Gleason score in Figure 3, no statistically significant association was found (Fischer’s exact test, p = 0.29). Furthermore, we examined whether XMRV infection correlates with seminal vesicle invasion (SVI), extracapsular extension (ECE), and surgical margin invasion (SMI), which are indicators of spreading prostate cancer (Table 3). However, we found no significant correlation between XMRV infection and SVI (Fischer’s exact test p = 0.33), ECE (Chi-Square p = 0.59), or SMI (Chi-Square p=0.89).

Figure 3.

XMRV infection is not significantly correlated with Gleason score. Numbers of infected patients (light gray) and uninfected patients (dark gray) are graphed according to Gleason score. No association between provirus positivity and Gleason score was found (Fisher’s exact test, p = 0.29; two sample t-test, p = 0.30).

Table 3.

Statistical analysis of XMRV positivity versus clinical pathological parameters of spreading prostate cancer

Parametera	Pos.		Neg.		Test	p-value
	PCR+	PCR−	PCR+	PCR−
ECE	11(25.6)	32	21(21.4)	77	Chi-Square	0.59
SVI	5(33.3)	10	27(21.4)	99	Fischer’s exact	0.33
SMI	8(23.5)	26	24(22.4)	83	Chi-Square	0.89

Note. Patients either scored positive (Pos.) or negative (Neg.) for clinical pathological parameters of prostate cancer, and were either positive (PCR+) or negative (PCR−) by PCR for XMRV. Numbers in parentheses indicate percentage of patients among parameter and score.

^aClinical parameters of prostate cancer include extracapsular extension (ECE), seminal vesicle invasion (SVI), and surgical margin invasion (SMI).

DISCUSSION

Our screen of PCA patients confirms the presence of XMRV among PCA patients in the U.S. We detected XMRV DNA in normal and tumor tissue, indicating that non-malignant cells may be susceptible to infection. In agreement with recent studies, we find no correlation between the presence of XMRV infection and the R462Q polymorphism of RNASEL, confirming that the population at risk for infection is not confined to homozygous carriers of the Q variant [7, 8].

Interestingly, three independent studies, including two surveys of German prostatic tissue specimens and a screen of English chronic fatigue syndrome (CFS) patients, found little to no evidence of XMRV infection [9, 10, 13]. However, in agreement with studies performed in the U.S., we found the presence of XMRV in PCA tissues [6, 7, 14]. It is possible that XMRV is mostly absent from the European population. If so, it would be interesting to uncover the reason for this geographic distribution. Alternatively, the inability to detect XMRV in Europe may possibly reflect genetic differences between American and European strains. However, this seems unlikely considering the high degree of sequence conservation among XMRV isolates and the variety of primer target sequences used for detection among the studies in Europe [6, 8–10, 13]. Additionally, the failure to detect XMRV may be due to differences in the detection techniques employed. We have found that detection of XMRV required rather specific conditions. For instance, at least 600ng of prostatic DNA was necessary for reliable detection with our PCR assay. XMRV was detected in 3.2% of the patients when we initially used 100–140ng of prostatic DNA compared to 22.2% of the patients when we used 650ng. Additionally, we found that detection of XMRV from patient specimens, but not from LNCaP cells infected in vitro, depended on the gene targeted in the PCR assay. We were unable to detect XMRV in the patient tissue samples by nested PCR with primers specific for the gag and pol genes regardless of whether 100 or 650ng of DNA was used as template. We found the gag primers to be at least 10-fold less sensitive than the env primers, and the pol primers tended to amplify a competing region from the human genome (data not shown). It is unclear whether these deficiencies account for the inability to detect XMRV in patient samples, or if XMRV is mainly present as an incomplete provirus in the cells of these patients. Nonetheless, the difficulty associated with detecting XMRV in patient samples may perhaps explain studies that do not detect the virus among large cohorts.

We found our nested PCR assay for XMRV env capable of detecting one infected cell per 10⁵ uninfected LNCaP cells in one of three samples using 600ng of DNA. The fact that the PCR-positive tissue specimens tested positive in only one or two out of three replicates may indicate that XMRV provirus is present at a very low copy number. This interpretation would be consistent with an earlier report [7]. Alternatively, it is possible that the quality of the tissue specimens was low due to preservation, handling and the duration of storage prior to DNA isolation. However, we were able to genotype the patients for R462Q using 20ng of DNA without difficulty.

Our finding that XMRV can be detected in the normal tissue of PCA patients suggests that non-malignant cells may also be susceptible to XMRV infection. If this is correct, XMRV infection may precede, and possibly partake in the process of tumorigenesis. There is currently little evidence to suggest that XMRV employs any traditional mechanisms for transforming cells. The virus harbors no known oncogenes, and a clonal integration pattern indicative of insertional mutagenesis has not been observed in prostate cancer specimens. In accord with previous studies, we predict a proviral copy number of far less than one per cell, arguing against insertional mutagenesis as a mechanism of transformation [6, 7].

A limitation of our PCR-based screen is that it does not identify the infected cell types. It is possible that the XMRV we detected was exclusively from non-malignant cells since tumor tissue consists of both malignant and non-malignant cells types. It is important to note that XMRV may promote tumorigenesis through paracrine and cell-cell interactions. PCA has been shown to depend on the biology of the surrounding stromal microenvironment, and a reactive stromal phenotype has been shown to promote cancer progression [3, 15–18]. It would be interesting to determine whether XMRV elicits the conversion of prostate stromal cells to a reactive phenotype, regardless of the cell type infected.

We did not find a correlation between XMRV infection and various clinical pathological parameters of PCA, including seminal vesicle invasion, extracapsular extension, and surgical margin invasion. Similar to a previous report, which found a correlation with higher Gleason scores, we observed a slight trend in favor of increasing Gleason score [7]. However, our results were not statistically significant. Additional studies with a greater number of patients will be required to evaluate a correlation between XMRV infection and Gleason score.

In conclusion, our data support a hypothesis that XMRV is endemic to North America. However, further investigation into the association of XMRV with PCA and other human diseases is needed. If established as an agent of human disease, XMRV may prove to be an important biomarker for selecting a suitable course of treatment.