Development of a Primary Human Cell Model for the Study of Human Cytomegalovirus Replication and Spread within Salivary Epithelium

Abstract

Various aspects of human cytomegalovirus (HCMV) pathogenesis, including its ability to replicate in specific cells and tissues and the mechanism(s) of horizontal transmission, are not well understood, predominantly because of the strict species specificity exhibited by HCMV. Murine CMV (MCMV), which contains numerous gene segments highly similar to those of HCMV, has been useful for modeling some aspects of CMV pathogenesis; however, it remains essential to build relevant human cell-based systems to investigate how the HCMV counterparts function. The salivary gland epithelium is a site of persistence for both human and murine cytomegaloviruses, and salivary secretions appear to play an important role in horizontal transmission. Therefore, it is important to understand how HCMV is replicating within the glandular epithelial cells so that it might be possible to therapeutically prevent transmission. In the present study, we describe the development of a salivary epithelial model derived from primary human "salispheres." Initial infection of these primary salivary cells with HCMV occurs in a manner similar to that reported for established epithelial lines, in that gH/gL/UL128/UL130/UL131A (pentamer)-positive strains can infect and replicate, while laboratory-adapted pentamer-null strains do not. However, while HCMV enters the lytic phase and produces virus in salivary epithelial cells, it fails to exhibit robust spread throughout the culture and persists in a low percentage of salivary cells. The present study demonstrates the utility of these primary tissue-derived cells for studying HCMV replication in salivary epithelial cells in vitroIMPORTANCE Human cytomegalovirus (HCMV) infects the majority of the world's population, and although it typically establishes a quiescent infection with little to no disease in most individuals, the virus is responsible for a variety of devastating sequelae in immunocompromised adults and in developing fetuses. Therefore, identifying the viral properties essential for replication, spread, and horizontal transmission is an important area of medical science. Our studies use novel human salivary gland-derived cellular models to investigate the molecular details by which HCMV replicates in salivary epithelial cells and provide insight into the mechanisms by which the virus persists in the salivary epithelium, where it gains access to fluids centrally important for horizontal transmission.

Keywords: HCMV; cytomegalovirus; horizontal dissemination; latency; lytic replication; organoid; salisphere; salivary gland.

FIG 1

Establishment of salispheres and primary epithelial lines from human salivary gland. (A) Freshly isolated salivary gland tissues (top) were digested with collagenase-dispase, and single-cell suspensions were then cultured on Cultrex PathClear basement membrane extract (BME) gels, where multicellular salispheres developed within 72 to 96 h (middle). Salispheres were then dissociated, counted, and replated onto Cultrex PathClear BME gels as in the middle panel or plated onto plastic culture dishes to facilitate HCMV infection experiments (bottom). (B) Salispheres were used for mRNA preparation and amylase 1 and hypoxanthine-guanine phosphotransferase (HPRT) expression. (C) Salispheres were then used for more-detailed analyses of salivary gland-specific gene expression. Taken together, the results indicate that the salispheres grown in vitro exhibit gene expression profiles similar to those of salivary tissue in vivo.

FIG 2

Low-passage-number clinical HCMV isolates and TB40E efficiently infect primary salivary cells and initiate IE gene expression. (A) Primary salivary epithelial cells (Par1) or HS68 fibroblasts were infected at an MOI of 1 with various HCMV strains, including the low-passage-number clinical isolate MOLD and the laboratory strains TB40E and AD169. Cells were analyzed for IE expression by flow cytometry to determine the percentage of cells that become infected and enter the immediate early phase. FSC, forward scatter. (B) Experiments performed as described above for panel A are presented quantitatively as the relative ability of each strain to infect parotid gland-derived salivary cells compared to its ability to infect HS68 fibroblasts. (C) Comparison of AD169 and MOLD infections in two additional salivary lines (one parotid and one submandibular) demonstrates that similar results are obtained in multiple lines. (D) Determination of relative infectivity, calculated as described above for panel B, was repeated using the ARPE19 epithelial line to allow comparisons between primary salivary lines and an immortalized epithelial line often used in HCMV studies. Values above the error bars in panels B to D represent the ability of each strain to infect epithelial cells relative to fibroblasts. *, P < 0.05 compared to AD169, as determined by unpaired Student’s t tests.

FIG 3

Human salivary cells infected with HCMV-TB40E are able to support all phases of HCMV lytic gene expression and release nascent virus into the culture medium. (A) Human parotid gland-derived (left) or submandibular gland-derived (right) salivary epithelial cells were infected with TB40E at an MOI of 1, and cell extracts were prepared at the indicated time points. Replicate Western blots prepared with these infected cell extracts were then probed with anti-CMV IE1/IE2 (Chemicon), anti-UL44 (12G5), anti-pp65 (Virusys), and β-actin antibody (Bethyl Laboratories). (B) Human parotid gland-derived (left) or submandibular gland-derived (right) salivary epithelial cells were infected with TB40E at an MOI of 0.1, the culture supernatant was isolated at the indicated times postinfection, and the titers were subsequently determined on HS68 fibroblasts. The growth curves depicted in panel B are representative of data from multiple independent experiments, each performed in duplicate.

FIG 4

Primary salivary epithelial cells do not facilitate efficient HCMV-TB40E spread. (A) HS68 fibroblasts were infected with TB40E at an MOI of 0.02, and the percentages of IE⁺ cells were quantified at the indicated time points postinfection by IE staining and flow cytometry. (B) ARPE19 cells were similarly infected but at an MOI of 0.2 and analyzed as described above for panel A. (C) Parotid gland-derived epithelial cells were infected as described above for panel A with TB40E at an MOI of 0.2 and similarly analyzed by flow cytometry for IE expression at various times postinfection. (D) Submandibular gland-derived epithelial cells were infected as described above for panel A with TB40E at an MOI of 0.2 and similarly analyzed by flow cytometry for IE expression at various times postinfection. (E) Salivary cell salispheres grown as clusters on BME gels were infected as described above for panel A with TB40E at an MOI of 0.2 and similarly analyzed by flow cytometry for IE expression at various times postinfection.

FIG 5

HCMV-MOLD and HCMV-FIX similarly fail to spread throughout salivary epithelial cells. (A and C) The MOLD (A) or FIX (C) strain of HCMV was used to infect HS68 fibroblasts, and spread was measured by flow cytometry using IE expression (MOLD) or GFP expression (FIX) as a marker. Using a starting MOI of 0.02, the virus quickly and efficiently spreads throughout the culture, reaching a CPE level of close to 100% by day 12 postinfection. (B and D) Human parotid salivary epithelial cells were similarly infected with HCMV-MOLD at an MOI of 0.1 or HCMV-FIX-GFP at an MOI of 20. A high MOI for FIX was used as this strain infects Par1 cells relatively poorly. Spread was measured by flow cytometric analyses of IE or GFP expression as described above for panels A and C.

FIG 6

The failure of HCMV to robustly and efficiently spread in salivary epithelial cell cultures results in a low-level persistent infection. Parotid primary epithelial cells were infected with the TB40E-GFP virus at an MOI of 0.1 to 0.5 (A) or with the FIX-GFP virus at an MOI of 1 to 5 (B) and harvested at the indicated time points. At each time point, the percentage of infected cells was tabulated by calculating the percentage of cells expressing GFP. Analysis of GFP expression enables the determination of the number of infected cells remaining in this “persistent” state. Cells were fed with fresh medium every 3 to 4 days. The overall viability of the cultures remained high based on basic analyses of flow cytometric analysis of forward scatter (FSC) and side scatter (SSC). The data depicted in panels A and B were derived from 4 or more independent experiments.

FIG 7

HCMV-infected primary salivary cells exhibit extensive cell death following infection. (A) HS68 fibroblasts were infected with the FIX-GFP virus at an MOI of 0.02, and the percentage of infected cells was tracked by flow cytometry using GFP expression as a marker of infected cells (left). At each time point, cells were then stained with SYTOX orange for analyses of cell viability. Cells from each time point were gated such that cellular viability could be examined in both uninfected (middle) and infected (right) cell populations. For example, at the day 9 time point, when gating on uninfected GFP-negative cells, 96.7% of the cells were viable and SYTOX negative, while 3.1% of the cells were dead and SYTOX positive. At the day 9 time point, when gating on infected GFP⁺ cells, 85.2% of the cells were viable and SYTOX negative, while 14.4% of the cells were dead and SYTOX positive. (B) Primary salivary epithelial cells were infected with the FIX-GFP virus at an MOI of 2.0, and the percentage of infected cells and viability of uninfected and infected cell populations were assessed as described above for panel A. (C) Data from three independent experiments performed in duplicate are shown in graphical form. Cell death in Par7 cells was compared to cell death in HS68 cells at each time point for statistical analyses. **, P < 0.01 for comparison of Par7 cells to HS68 cells, as determined by unpaired Student’s t tests. ns, nonsignificant.