As a preliminary to in vivo
inoculation, the growth of XMRV strain VP62 (4
) was tested in rhesus macaque primary fibroblasts () using a reverse transcriptase assay (41
), and XMRV was shown to replicate and produce Gag protein (). Next, the presence of preexisting immunity to XMRV was tested in 25 plasma samples from rhesus macaques from the Yerkes colony, including the five animals used for the inoculation. As shown in , no XMRV-specific antibodies were found, in contrast to a human plasma sample from a prostate cancer case harboring antibody reactive to XMRV p30 capsid (lane 1). Next, an initial set of three monkeys were inoculated with XMRV VP62, a molecular clone isolated from the prostate of a human patient (4
), and followed for plasma viremia using quantitative RT-PCR (). Among the three animals, plasma viremia was detected in two: one male (RIl-10) with viremia starting on day 4 postinfection (p.i.), which reached peak at 7,500 viral copies on day 7 and decreased to below the detection levels by day 14 (); and a female (RYh-10), who showed a delayed and lower acute viral replication kinetics at days 14 to 20. The third animal (a male, RLq-10) never showed vRNA levels above the detection threshold. However, proviral DNA was detectable in all three animals during the initial 3 to 4 weeks p.i. in blood (), confirming the susceptibility of macaques to XMRV infection. Viral and proviral signals became consistently negative beyond 1 month p.i. until reinfection of two animals at day 158 (). At that time, plasma viral RNA remained below the detection threshold, but transient proviral signals were detected, suggesting reinfection (). More importantly, viremia was again detected in monkey RIl-10 16 days after the immunization, suggesting that XMRV may be reactivated by immune activation and confirming the presence of replication-competent XMRV in this animal 4 months after its last exposure to the virus ().
Fig. 2. (a) Outline of the experimental monkey XMRV infection protocol and sample collections. LN coll, lymph node collection. (b) Plasma viral loads measured by real-time RT-PCR during the acute infection of three rhesus macaques. (c) Detection of XMRV Env provirus (more ...)
The lineage of blood cells replicating XMRV during acute infection was determined by the sorting of monocytes, B cells, CD4+
T cells, and NK cells from pooled peripheral blood mononuclear cells (PBMC) collected on days 3, 5, and 7 p.i. from each of the three animals using magnetic beads coated with antibodies to CD14, CD20, CD3, and CD4 or CD8, with NK cells being the remaining cells. The purity of the enriched subsets was verified by flow cytometry and ranged from approximately 60% for monocytes, 75% for B and CD8+
T cells, 80% for NK cells, and >93% for CD4+
T cells. XMRV DNA isolated from the enriched populations was amplified in duplicates and analyzed by electrophoresis (). XMRV signal was consistently detected in CD4+
T cells and in NK cells. B cells were not consistently positive, and monocytes were mostly negative. These data mirrored the lymphotropism of XMRV reported in a cohort of chronic fatigue patients (17
). To demonstrate the presence of nearly full-length XMRV genomic DNA, a 7.2-kb amplicon was amplified from day 18 p.i. PBMC DNA (animals RIL-10 and RLq-10) ().
None of the infected animals showed any obvious clinical symptoms. Complete blood counts and sequential serum chemistry panels revealed only a mild elevation of creatine phosphokinase levels postinfection, which resolved rapidly and remained normal during the 9 months of monitoring. However, marked variations were observed in the phenotypic profiles of peripheral blood postinfection (). XMRV infection and reinfection were both accompanied by a marked increase in the frequency of circulating B and NK cells in blood. Even more remarkable was the increase in Ki67 expression by B cells and NK cells ( to l), as well as on memory CD4+ and CD8+ T cells ( to h), suggesting marked activation of lymphoid subsets following XMRV infection in vivo. Of note, immunization of the remaining two monkeys at day 275 with a cocktail of recombinant XMRV proteins in incomplete Freund's adjuvant also induced a recrudescence of circulating B cells but a concomitant decrease in circulating NK cells ( and k). It was assumed that infections would lead to detectable cytokine production, and, therefore, plasma samples corresponding to acute infection were subjected to multiplex flow-based analysis (data not shown). Out of 23 factors tested, transient changes were noted for only the chemokine IL-8 and soluble CD40L and only in two out of the three monkeys, with a peak on day 4 p.i. ().
Fig. 3. Phenotypic analyses of monkey PBMC post-XMRV infection (day 0) and reinfection (day 158) and postimmunization with XMRV proteins (day 275) of the three monkeys followed longitudinally. (a) Frequencies of total T cells (CD3+). (b) Frequencies of proliferating (more ...)
Quantitation of soluble CD40 ligand (a) and IL-8 in monkey plasma (b) using the BioPlex technology.
XMRV-specific cell-mediated responses measured both by short-term restimulation/flow analysis for gamma interferon (IFN-γ), interleukin-2 (IL-2), and tumor necrosis factor alpha (TNF-α) and by proliferation were essentially at background levels throughout the monitoring period, suggesting low levels of cell-mediated responses (data not shown). A caveat to these analyses was the limited availability of peptides spanning only the transmembrane Env protein and of whole virus as the stimulating antigens. More extensive analyses with a larger set of reagents will be needed to ascertain the magnitude of XMRV cell-mediated responses in this model. In contrast, antibody responses were clearly elicited after the initial infection () and were boosted following reinfection as well as after immunization. However, the titers to XMRV antibodies rapidly decreased to a low level even after the second infection, indicative of a lack of sustained antigen stimulation of humoral responses, due to either poor release of XMRV from infected cells or the presence of immunosuppressive mechanisms (35
). Of note, antibody responses to the envelope glycoprotein p70 and the transmembrane protein p15E dominated the early response although all structural proteins were recognized in Western blotting (28
). Even though the antibody titers were relatively modest, plasma samples tested at the chronic phase (day 114 p.i.) showed evidence of neutralizing activity ().
Fig. 5. Time course of antibody response to the XMRV transmembrane protein p15E (anti-p15E) (a) and capsid protein p30 (anti-p30) (b) in three monkeys post-XMRV infection and immunization. Antibody responses were determined by recombinant protein (p15E and p30)-based (more ...)
Two additional animals were infected with XMRV and sacrificed during the acute infection period to obtain tissues representative of early virus dissemination. Serial sacrifice of XMRV-infected monkeys failed to demonstrate signs of pathogenicity in any of the five animals, two of which were followed for 9 months p.i. Immune activation was evident during acute infection, based on the formation of germinal centers in spleen and lymphoid organs ( and b), although compared to other lymphotropic retrovirus infections such as simian immunodeficiency virus (SIV), this activation was modest. Since XMRV was discovered initially linked with prostate cancer, the prostate was of particular interest as a potential site for pathogenicity. Histologic analyses showed mild to multifocal interstitial infiltrates of lymphocytes, neutrophils, and a few plasma cells in the prostate during acute infection ( and d), which appeared less pronounced in the two animals sacrificed at the chronic stage ( and f). However, prostates from age-matched control rhesus macaques also exhibited such findings without other evidence of prostate infection, arguing against a viral specificity for these observations ( and h).
Fig. 6. Histological analysis of XMRV-infected monkeys in spleen and prostate using hematoxylin and eosin staining. (a and b) Splenic lymphoid architecture during acute XMRV infection on days 6 and 7 postinfection. Other images show prostate histology in monkeys (more ...)
In stark contrast to the absence of detectable XMRV genomic RNA in blood, detection of XMRV in situ was surprisingly abundant during acute infection as well as at the chronic stage. Lymphoid organs such as spleen, lymph nodes, and the gastrointestinal (GI) lamina propria all contained single round cells positive for XMRV by both immunohistochemistry (IHC) and fluorescence in situ hybridization (FISH) (). Moreover, the frequency of positive cells did not appear to vary dramatically between acute and chronic infection, except perhaps in the colon (), where increasing numbers of XMRV-positive cells were observed with time p.i. Using fluorescent IHC for XMRV and CD3 or CD4, the only cells identified to be positive for the presence of XMRV protein in the GI tract were CD3+ and CD4+ T cells ( and n), which was also the case for other lymphoid tissues such as spleen and lymph nodes, corresponding to the cell types infected in the blood during acute infection (). The exception among these tissues was lung, which exhibited large numbers of XMRV-positive macrophages () throughout infection () and few or no XMRV-positive T cells. There was an absence of signal by both IHC and FISH in the organs of control rhesus macaques that were not infected with XMRV ( and l and 8v and w).
Fig. 7. Detection of XMRV infection in lymphoid organs in situ using IHC and FISH. (a to e) Spleen samples from acutely (a and b) and chronically (c to e) infected monkeys analyzed by IHC (a and d) and FISH (e). (f) Noninfected control spleen tested by IHC. (g (more ...)
Semiquantification of XMRV signal in relation to tissue and time postinfection
To our surprise, when other tissues were investigated, completely distinct host cell lineages were observed to be XMRV positive. In the prostate, large foci of XMRV-positive acinar epithelial cells were detected by IHC during acute infection, suggesting rapid cell-to-cell transmission in glandular acini, while the surrounding stroma was uniformly negative ( and b). In contrast to acute infection, XMRV was no longer detected by IHC in prostate during chronic infection (); although the virus was not eliminated, low-level nucleic acid signals were still observed using FISH (). While not providing an etiological link to prostate carcinoma, these findings clearly indicate that the prostate constitutes an early target for XMRV. In contrast to findings in the prostate, XMRV-positive cells in the testes were far less abundant at any time and were exclusively interstitial, showing short strings of positive cells likely indicative of cell-to-cell transmission throughout infection ( to h). Other tissues of the reproductive tract showed evidence of replicating XMRV, such as the epididymis and seminal vesicle in male monkeys ( and k) and cervix and vagina in the female monkey ( and ), suggesting the potential for sexual transmission. XMRV infection of interstitial cells was also detected in the pancreas during acute but not chronic infection. Other organs scored negative for XMRV by IHC (, brain, heart, kidney, bladder, gallbladder, etc.). These tissues were, however, not clear of XMRV since low-frequency nucleic acid signals were detected by FISH, suggesting generalized dissemination of XMRV ( to u) but lack of replication due perhaps to a paucity of suitable target cells, an unsuitable environment, or some form of active control of viral replication which remains to be determined.
Fig. 8. Detection of XMRV infection in nonlymphoid organs in situ using IHC and FISH. (a to d) Prostate samples from acutely (a and b) and chronically infected monkeys (c and d, showing IHC and FISH, respectively). (e to h) Testis samples from acutely (e and (more ...)