Human cytomegalovirus miR-UL112-1 promotes the down-regulation of viral immediate early-gene expression during latency to prevent T-cell recognition of latently infected cells

Abstract

Human cytomegalovirus, a member of the herpesvirus family, can cause significant morbidity and mortality in immune compromised patients resulting from either primary lytic infection or reactivation from latency. Latent infection is associated with a restricted viral transcription programme compared to lytic infection which consists of defined protein coding RNAs but also includes a number of virally encoded microRNAs (miRNAs). One of these, miR-UL112-1, is known to target the major lytic IE72 transcript but, to date, a functional role for miR-UL112-1 during latent infection has not been shown. To address this, we have analysed latent infection in myeloid cells using a virus in which the target site for miR-UL112-1 in the 3' UTR of IE72 was removed such that any IE72 RNA present during latent infection would no longer be subject to regulation by miR-UL112-1 through the RNAi pathway. Our data show that removal of the miR-UL112-1 target site in IE72 results in increased levels of IE72 RNA in experimentally latent primary monocytes. Furthermore, this resulted in induction of immediate early (IE) gene expression that is detectable by IE-specific cytotoxic T-cells (CTLs); no such CTL recognition of monocytes latently infected with wild-type virus was observed. We also recapitulated these findings in the more tractable THP-1 cell line model of latency. These observations argue that an important role for miR-UL112-1 during latency is to ensure tight control of lytic viral immediate early (IE) gene expression thereby preventing recognition of latently infected cells by the host's potent pre-existing anti-viral CTL response.

Fig. 1.

Removal of the miR-UL112-1 target site leads to induction of immediate early (IE) gene expression in the absence of virion production. (a) The target site of miR-UL112-1 in the 3′ UTR of IE72 was deleted to generate a recombinant virus in which IE72 is no longer targeted by miR-UL112-1. (b) Monocytes (mono) were either uninfected, infected with wild-type TB40E (WT) or the miR-UL112-1 target site mutant virus (∆112-1) and latency was established for 4 days before harvesting and analysing mRNA levels for viral products UL138, IE and pp28 relative to cellular gene glyceraldehyde 3-phosphate dehydrogenase (GAPDH). Alternatively, these gene products were analysed following reactivation from latency by differentiation to DCs (reactivation). (c) Both latent monocytes and reactivating DCs, described in (a), were also co-cultured with indicator fibroblasts and supernatants were harvested and seeded onto fresh fibroblasts for 24 h before fixing and immunofluorescence staining for IE protein, which was used to determined the IE foci forming units. Error bars shown in (b) and (c) denote the standard error of two independent experiments, with each containing three technical repeats.

Fig. 2.

Monocytes latently infected with HCMV ∆112-1 target site mutant are recognized specifically by IE-specific CD8⁺ T-cells. Monocytes (Mono) were latently infected with either wild-type HCMV (WT) or miR-UL112-1 target site mutant virus (∆112-1) and a proportion of cells from each population were also differentiated to DCs by IL-4 and granulocyte-macrophage colony-stimulating factor treatment to induce reactivation. Both the undifferentiated and the differentiated cells were then co-cultured with IE- (VLE) specific CD8 T-cells and then assayed for IFNγ secretion using Fluorospot assays. Mock-infected monocytes (Mono mock) and dendritic cells (DC mock) were also co-cultured with IE- (VLE) specific CD8 T-cells. The level of T-cell recognition is shown relative to the level of T-cell recognition of reactivating DCs. Error bars denote standard deviation of three biological replicates. T-cells alone showed no background spots and reactivating DCs routinely showed in the region of 40–50 spot-forming units (SFU).

Fig. 3.

Characterization of the THP-1 latency and reactivation model in low serum. (a) THP-1 cells were infected with TB40/E at an m.o.i. of 5 and left for 3 days to obtain a latently infected population. Phorbol 12-myristate 13-acetate (PMA) was added on day 3 and left on the cells for 24, 48 or 72 h; cDNA was produced and amplified using IE, UL138 and GAPDH primers. Reverse transcriptase (RT) minus control samples (no RT) were run in parallel to exclude genomic DNA contamination. A representative experiment of three independent replicates is shown. Lane 1: 100 bp DNA ladder (Life Technologies); Lane 2: mock infected, untreated; Lane 3: mock infected, 100 nM PMA; Lane 4: TB40/E infected, untreated; Lane 5: TB40/E infected, 100 nM PMA; Lane 6: TB40/E infected, untreated; Lane 7: TB40/E infected, 100 nM PMA; Lane 8: TB40/E infected, untreated; Lane 9: TB40/E infected, 100 nM PMA. (b) Micrographs of TB40/E-infected, undifferentiated THP-1 cells (latency) (left panel) and TB40/E latently infected THP-1 cells, differentiated on day 3 post-infection with 100 nM PMA (reactivation) (right panel). (c) Cells from (b), left panel or right panel, were co-cultured with fibroblasts for 5 days (left panel or right panel, respectively). For all panels in (b) and (c), IE expression was visualized using immunofluorescence staining with anti-IE72/86 antibody. Arrows indicate IE-expressing cells.

Fig. 4.

Latent infection with Δ112-1 target site mutant virus results in increased levels of IE72 RNA expression compared with WT virus. THP-1 cells were infected at an m.o.i. of 5 with the parental WT virus or the Δ112-1 target site mutant. Latently infected cells were sorted for GFP expression at 2 days post-infection, before mRNA expression was analysed by quantitative reverse transcriptase PCR at 3 days post-infection. The level of UL138 and IE72 normalized to the mRNA level of housekeeping gene GAPDH is shown in (a) and (b), respectively. Data shown is representative of three independent repeats, whilst error bars shown mark standard deviation of technical replicates.

Fig. 5.

IE- (VLE) specific CD8⁺ T-cells recognize lytically but not latently infected THP-1 cells. (a) THP-1 cells were stained with monoclonal antibody specific for MHC Class I HLA-A2 (filled plot) or an isotype-matched control antibody (open plot). (b) THP-1 cells were pulsed with IE (VLE) peptide (pulsed) and assayed with control THP-1 cells (unpulsed) for their ability to be recognized by increasing numbers of IE- (VLE) specific CD8⁺ T-cells, as measured by IFNγ production in ELISPOT assays. (c) Latently infected THP-1 cells (latent), lytically infected differentiated THP-1 cells (lytic) or latently infected THP-1 cells differentiated with PMA to induce HCMV reactivation (reactivated) were co-cultured with increasing numbers of IE- (VLE) specific CD8⁺ T-cells (as detailed in the figure) and assayed for IFNγ using ELISPOT assay. Data shown is representative of three independent experiments. (d) The ∆112-1 target mutant allows recognition and removal of latently infected cells. THP-1 cells were infected at an m.o.i. of 5 with the parental WT virus (WT) or the Δ112-1 target virus before latently infected cells were sorted for GFP expression at 2 days post-infection. At 3 days post-infection, the cells were co-cultured with CD8⁺ T-cells as indicated for a further 24 h before the numbers of GFP-expressing cells were enumerated. Changes in GFP⁺ cell numbers are shown as fold change over the number of GFP⁺ cells present in WT virus-infected THP-1 cells cultured in the absence of T-cells. Statistical analysis was performed using a t-test, with an asterisk indicating significant difference (P<0.05). Error bars denote standard deviation of technical replicates.