Isolation of a Human Betaretrovirus from Patients with Primary Biliary Cholangitis

Abstract

A human betaretrovirus (HBRV) has been linked with the autoimmune liver disease, primary biliary cholangitis (PBC), and various cancers, including breast cancer and lymphoma. HBRV is closely related to the mouse mammary tumor virus, and represents the only exogenous betaretrovirus characterized in humans to date. Evidence of infection in patients with PBC has been demonstrated through the identification of proviral integration sites in lymphoid tissue, the major reservoir of infection, as well as biliary epithelium, which is the site of the disease process. Accordingly, we tested the hypothesis that patients with PBC harbor a transmissible betaretrovirus by co-cultivation of PBC patients' lymph node homogenates with the HS578T breast cancer line. Because of the low level of HBRV replication, betaretrovirus producing cells were subcloned to optimize viral isolation and production. Evidence of infection was provided by electron microscopy, RT-PCR, in situ hybridization, cloning of the HBRV proviral genome and demonstration of more than 3400 integration sites. Further evidence of viral transmissibility was demonstrated by infection of biliary epithelial cells. While HBRV did not show a preference for integration proximal to specific genomic features, analyses of common insertion sites revealed evidence of integration proximal to cancer associated genes. These studies demonstrate the isolation of HBRV with features similar to mouse mammary tumor virus and confirm that patients with PBC display evidence of a transmissible viral infection.

Keywords: biliary epithelial cells (BEC); common insertion sites (CIS); human betaretrovirus (HBRV); mouse mammary tumor virus (MMTV); primary biliary cholangitis (PBC).

Figure 1

Isolation of HBRV by repeat subcloning of Hs578T cells infected with PBC lymph node homogenate. Co-cultures growing to 70% confluency with PBC lymph node homogenates were maintained in culture for 7 days. Then, three rounds of subcloning by limiting dilution of single cells exposed to PBC conditioned media, supernatants were assayed for RT activity and HBRV RT-PCR positive cells were subcloned for a total of three rounds [Pink and red circles represent increasing HBRV load with subcloning infected Hs578T cells].

Figure 2

Dark field microscopy showing betaretrovirus RNA by in situ hybridization (green signal) (A) MMTV RNA in the MM5MT murine breast cancer cells (control), (B) HBRV in the cytoplasm of Hs578T following co-culture with PBC conditioned media and (C) HBRV signal in biliary epithelial cells infected with lymph node conditioned Hs578T supernatants [white arrows indicative of HBRV RNA, DAPI blue stain for nuclei, magnification ×200].

Figure 3

Negatively stained supernatant from infected Hs578T cells was analyzed by transmission electron microscopy to identify 85–95 nm virus-like particles with acentric cores and membrane spikes, comparable with MMTV and prior betaretrovirus-like particles derived from patients with PBC.

Figure 4

(A) HBRV integrations in vitro demonstrate a variability in the density of sites with areas of clustering within the genome [with red/orange signal demonstrating 5–10 frequency of integrations per Mb on Chr 3q, Chr 5q, Chr 8q, Chr 9q, Chr 11q, and the centromere on X Chr.] (B) Compared to a computer-generated set of random integrations, HBRV showed little difference from the random plot of insertions within the genome with regard to integration proximal to TSS. (C) HBRV avoided insertion within 2 kb proximal to CpG islands in the Hs578T breast cancer cell line but aggregated proximal to CpG in the samples derived from patients with liver disease. (D) In the Hs578T breast cancer cell line, MMTV integrations did not aggregate around CpG and TSS as previously reported [36]. [The three black lines represent the median and range [5–95%] of insertions randomly generated with 1000 iterations throughout the human genome].