CDK4/6 inhibitors sensitize gammaherpesvirus-infected tumor cells to T-cell killing by enhancing expression of immune surface molecules

BJAB cells, PEL cell lines JSC-1, BCBL-1, BC-1, and BC-2; and EBV cell lines Akata, Raji, and Daudi, were obtained and maintained as described previously [25, 26]. JSC-1, BC-1, and BC-2 are co-infected with EBV, while BCBL-1 is not. The KSHV-infected iSLK cell line (BAC16 strain) was kindly provided by Rolf Renne from the University of Florida and maintained in DMEM (Gibco) supplemented with 10% fetal bovine serum (HyClone), 1.0 μg/mL puromycin, 1.2 mg/mL hygromycin, and 250 μg/mL G418. HUVECs were purchased from Lonza and maintained in EGM-2 Bulletkit (Lonza) for up to 5 passages, with passages 3 to 5 used for experiments. All cells were maintained at 37 ℃ and 5% CO₂ in Falcon cell culture flasks. For the treatment of cells with inhibitors, floating cells were seeded at 2 × 10⁵/mL and treated with CDK4/6 inhibitors at indicated concentrations for up to 3 days, while adherent cells were treated for up to 7 days.

Abemaciclib (LY2835219) and palbociclib (PD-0332991) were purchased from Selleck Chemicals and dissolved in ethanol and water respectively at a stock concentration of 10 mM/mL. Ribociclib (LEE011, 10 mM/mL in DMSO) was purchased from MedChemExpress. All stocks were aliquoted and stored at − 20 °C.

For preparation of virus stock of KSHV-BAC16, which latently expresses green fluorescent protein (GFP), iSLK cells were treated with doxycycline (1 µg/mL; Thermo Fisher Scientific) and sodium butyrate (1 mM; Sigma Aldrich) for 3 d to induce lytic replication. Collected supernatants were cleared of debris with centrifugation at 500×g for 5 min, filtered (Rapid-Flow 0.45-µm filter; Thermo Fisher Scientific) and pelleted with ultracentrifugation (2.5 h at 4 °C at 50,000×g with SW 32 Ti; Beckman Coulter). Pellets were then resuspended with Dulbecco’s Modified Eagle Medium (DMEM) (Thermo Fisher Scientific). For de novo infection of HUVEC, cells were infected with diluted virus at a multiplicity of infection (MOI) of 15, as determined by LANA copy number, with 8 μg/ml polybrene. Virus supernatants were washed off after 8 h and replaced with fresh Endothelial Growth Media-2 (EGM2, Lonza Catalog #: CC-3162). HUVEC were refed every 2 days until harvested. Cells were visualized using ZOE fluorescent microscope (BioRad) and the percentage of infected cells were assessed as the percent of GFP+ cells using ImageJ software.

The relative number of viable cells was analyzed using CellTiter-Glo® Luminescent Cell Viability Assay kit (Promega). Briefly, 50 μL reagent was added to 50 μL cells in 96-well plate. Contents were mixed for 2 min on an orbital shaker and then incubated at room temperature for 10 min. Luminescence was recorded using a plate reader VICTOR™ X3 (PerkinElmer). The percentage of alive vs dead cells was assessed using trypan blue staining.

Analysis of cells for surface marker expression was carried out as described previously [25]. Briefly, control and treated cells were incubated with 1:50 diluted PerCP/Cyanine5.5-labeled isotype antibodies or antibodies against HLA class I (A, B and C) (Cat. #:311420), ICAM-1 (Cat. #:353120), B7-2 (Cat. #:374216), or PD-L1 (Cat. #:329738) from BioLegend, for 40 min, and then were washed three times with FACS buffer (2% FBS, 1 mM EDTA, Ca/Mg²⁺-free phosphate buffered saline [PBS]), suspended in FACS buffer, and then analyzed with a flow cytometrycalibur™ Flow Cytometry system (BD Biosciences). Data were analyzed and visualized using FlowJo software (flowjo.com).

T-cell activation assays were performed using the T-cell activation bioassay IL-2 promoter kit (Promega, cat# J1651) as described previously [27]. Briefly, PEL cells were treated with the indicated concentrations of CDK4/6 inhibitors for 3 days, after which the cells were washed with PBS and mixed with T cell receptor (TCR)/CD3 Effector Cells (Jurkat T-cells expressing a luciferase reporter gene under IL-2 promoter) at a 2:1 ratio for BCBL-1 to Jurkat, and 1: 2 for Akata to Jurkat, and stimulated with various concentrations of anti-human CD3 monoclonal antibody (OKT3) from ThermoFisher Scientific (cat# 16-0037-81) in a 37 °C incubator for 6 h. The mix and incubation were done in triplicate in 96 well-plates containing 25 μL of target cells, 25 μL of Jurkat cells, and 25 μL of anti-CD3 antibody per well. Bio-Glo reagent was then added for a 10 min incubation at room temperature. The signal was captured using a Victor X3 multilabel plate reader (PerkinElmer). Wells containing cells but no Bio-Glo reagent served as background luminescence control. Luminescence data was plotted as a 4PL regression graph using GraphPad Prism software. Fold change in activation was calculated after subtracting signal obtained from Jurkat cells without co-stimulation by target cells from that obtained with co-stimulation by target cells.

mRNA was extracted using the RNeasy kit (Qiagen). cDNA synthesis was performed using random primers with the High-Capacity cDNA Reverse Transcription Kit (ThermoFisher Scientific) on a T100 Thermal Cycler (Bio-Rad). To evaluate the mRNA level of host genes including MHC-I, ICAM-1, B7-2,PD-L1, ACTB, DDX58, DNMT1, ERV3-1, IFIT1, interferon(IFN)-λ2 (IFNL2), IFN-α, IFN-β, IFN-γ, NLRC5, OAS2, and STAT1, and viral genes including KSHV-encoded open reading frame (ORF)73 (LANA), ORF50 (regulator of transcription activator [RTA]), K2 (viral interleukin-6 [vIL6)], ORF45, and ORF57, and EBV-encoded EBER2 and BMRF1, SYBR green qPCR assays were performed using the Applied Biosystems® SYBR® Green PCR Master Mix (ThermoFisher Scientific) on an ABI StepOnePlus real-time PCR system (ThermoFisher Scientific). Primers used are listed in Additional file 8. Relative mRNA expression levels were analyzed using the ΔΔCt method with genes coding for β-actin as the reference gene.

Statistical analysis was performed using two-tailed student’s t-test (paired or unpaired where indicated) on experiments with 3 or more biological replicates. P-values less or equal to 0.05 were considered statistically significant. Asterisks indicate p values: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

We first investigated the impact of CDK4/6 inhibitors on cell growth of various KSHV-infected cells including the PEL cell lines JSC-1 (Fig. 1a), BCBL-1 (Fig. 1b), BC-1 (Additional file 1a), and BC-2 (Additional file 1b), EBV-infected BL lines including Akata (Fig. 1c), Raji (Fig. 1d), and Daudi (Additional file 1c), and two EBV-negative BL lines including BJAB (Additional file 1d) and CA46 (Additional file 1e). All three CDK4/6 inhibitors exhibited dose-dependent inhibitory effect on cell growth for all the tested cell lines at day 3 (72 h) with about 30 to 70% inhibition observed at the highest doses tested (1 µM of Abe and Pal, or 5 µM Rib). We also tested the effects of Abe on the growth of uninfected HUVEC and KSHV-infected HUVEC cells. To produce KSHV-infected HUVEC, cells were exposed to concentrated viral stocks of KSHV.BAC16 using 15 viral DNA copy numbers per cell to reach around 85% GFP-positive rate by 24 h (Additional file 2). GFP is constitutively expressed in KSHV.BAC16-infected cells and GFP expression can be used to identify infected cells. Abe inhibited cell growth of KSHV-infected HUVEC starting at day 4 at concentrations of 0.02 µM or higher; by day 7, cell growth was inhibited about 50% at the highest dose tested (0.5 µM). In addition, Abe similarly inhibited the growth of uninfected HUVEC. The decrease in viable cell numbers after culture with CDK4/6 inhibitors was caused by a decrease in proliferation rather than an increase in cell death. After 3 days (B cells) or 4 days (uninfected or infected HUVEC cells) culture with Abe at doses that reduced cell numbers, there was no decrease in the percentage of viable cells as assessed by trypan blue exclusion (Additional file 3).

We next explored the effect of CDK4/6 inhibitors on the surface expression of MHC-I on KSHV- and EBV-infected cells. Several KSHV-encoded lytic genes including K3 and K5 can ubiquitinate MHC-I and reduce its surface expression. Consistent with this, MHC-I expression decreased to 37.5% of control for JSC-1 (Fig. 2a and g) and 31% for BCBL-1 (Fig. 2b and g) after lytic induction by butyrate. However, pretreatment of these cells with 1 μM Abe substantially prevented downregulation of MHC-I by butyrate (Fig. 2a, b, and g). Even without lytic induction, treatment of JSC-1 and BCBL-1 PEL lines with 1 µM Abe increased surface MHC-I expression (Fig. 2a, b). Treatment of two other PEL cell lines, BC-1 and BC-2, with 1 µM Abe, exhibited similar upregulation of MHC-I on the cell surface (Additional file 4, a-d). Abe also increased surface expression of MHC-I in the EBV-infected Akata and Raji cell lines (Fig. 2c, d). Figure 2g shows the mean and standard deviation from 3 experiments of the expression of MHC-I on JSC-1 and BCBL-1 PEL lines and on Akata and Raji EBV-infected BL lines. We further tested the effects of Pal and Rib, in addition to Abe, on MHC-I expression on JSC-1 and BCBL-1 PEL lines and on Akata and Raji EBV-infected BL lines. As can be seen in Fig. 2g, all three CDK4/6 inhibitors increased MHC-I expression on PEL lines induced to lytic replication as well as uninduced lines. Moreover, all three drugs enhanced MHC-I expression on Akata and Raji BL lines. EBV-uninfected BJAB a Burkitt’s lymphoma B cell line, also exhibited an increased MHC-I when treated with 1 µM Abe, although it was less than that of the virus-infected lines(Additional file 4e, f).

We also tested Abe’s effect on MHC-I expression in KSHV-infected and uninfected HUVEC (Fig. 2e, f, and h). Cells were infected with KSHV.BAC16 to obtain about 85% cells expressing GFP. As seen in Fig. 2h, while both 0.1 and 0.5 µM Abe induced significant and dose-dependent increases in MHC-I on KSHV-infected HUVEC, only 0.5 µM Abe induced a small increase of MHC-I in uninfected HUVEC.

KSHV and EBV also downregulate surface expression of ICAM-1 and B7-2 [6, 13], which are important for T-cell and NK-cell activation and effector function. We assessed the effects of CDK4/6 inhibitors on surface expression of ICAM-1 and B7-2, as well as PD-L1 on virus-uninfected BJAB cells and the same set of virus-infected PEL and BL cell lines (Fig. 3; Additional file 5). BJAB, PEL cells and EBV-infected BL cells were treated with 1 μM Abe, 1 μM Pal, or 5 μM Rib for 3 days, while uninfected and KSHV-infected HUVEC were treated with 0.5 μM Abe, 0.5 μM Pal, or 2.5 μM Rib for 4 days. Cells were then analyzed using flow cytometry. All the tested cell lines exhibited significant increases in ICAM-1 and B7-2 surface expression in response to all 3 CDK4/6 inhibitors although the virus-infected lines showed bigger increases compared to the virus-negative BJAB line. The average fold increase for ICAM-1 in KSHV and EBV-infected cells ranged from 2.5 to 4.5 fold (Fig. 3a), and for B7-2 from 3.2 to 6.0 fold (Fig. 3b). In addition, all 3 CDK4/6 inhibitors increased expression of PD-L1 from 4.2 to 8.3 fold (Fig. 3c) in the virus-infected lines. Virus-uninfected BJAB cells exposed to the drugs had a substantially smaller increase in the surface markers. We also tested the effects of the drugs in uninfected and infected HUVEC cells. While there was a relatively small increase (1.5 for ICAM-1, 2.0 for B7-2 and PD-L1) in uninfected HUVEC, there was a more substantial increase (2.7 to 4.2 for ICAM-1, 25.4 to 5.8 for B7-2, and 5.2 to 7.2 for PD-L1) in expression of all 3 markers in KSHV-infected HUVEC (Fig. 3d–f, Additional file 6).

We further evaluated the impact of CDK4/6 inhibitors on the expression of mRNA for these surface markers in JSC-1 (Fig. 4a), BCBL-1 (Fig. 4b), Akata (Fig. 4c), and Raji (Fig. 4d) cells. Cells were cultivated with 1 μM Abe or solvent control (diluted ethanol in medium) for 24 h or 48 h and the total RNA was then extracted for qPCR analysis. As the result shows, mRNA expression of all these 4 genes were significantly increased at both time points after treatment, with an average fold increase of 1.6–1.8 for MHC-I, 1.4–1.5 for ICAM-1, and 2.8–3.4 for B7-2 at 24 h, and 2.0–2.4 for MHC-I, 1.9–2.2 for ICAM-1, and 2.8–3.3 for B7-2 at 48 h. Also, mRNA for PD-L1 increased significantly from 1.5 to 1.8 fold at 24 h, and 2.4 to threefold at 48 h for all cells except for BCBL-1 cells at 24 h. These results provide evidence that the increased expression of the surface markers by Abe was at least in part from increased transcription of the genes.

We next sought to assess whether these CDK4/6 inhibitor-treated cells would enhance T-cell activation through increased expression of co-stimulatory molecules. As in previous experiments, 3-days of treatment of BCBL-1 PEL cells and Akata BL cells by Abe upregulated surface expression of ICAM-1 and B7-2 (Fig. 5a, b). T-cell activation induced by Abe was assessed in aliquots of the same cell culture using Jurkat cells expressing a luciferase reporter gene under control of the IL-2 promoter as the effector cells, and anti-CD3 antibody was used to activate these cells. Both BCBL-1 and Akata cells cultured in the absence of CDK4/6 inhibitors increased Jurkat T-cell activation above the baseline (Fig. 5c, e). However, Abe-treated BCBL-1 and Akata cells further stimulated the Jurkat T-cell activation compared to the control untreated cells in a dose-dependent manner (Fig. 5c, e). When co-stimulated with 0.6 μg/mL anti-CD3 antibody (red arrow in Fig. 5c, d), the mean increases in T-cell activations induced by 0.3 μM Abe-treated BCBL-1 cells and 1 μM Abe-treated Akata cells were 2.6 fold and fivefold over control-treated cells, respectively.

Abe has been shown to suppress expression of DNA methyltransferases which in turn leads to increased expression of the endogenous retrovirus ERV3-1, and it has been suggested that this may be a mechanism for the increased expression of MHC-I [23, 28]. In particular, it has been suggested that the ERV3-1 stimulates interferon type III which in turn leads to increased expression of interferon-induced genes including MHC-I. We were interested to see if such a mechanism might apply in the cell lines studied here, which were infected with an exogenous virus and might thus have high basal levels of interferon expression that would not be increased by ERV3-1. We were also interested to explore the possibility that Abe-induced expression of KSHV and/or EBV genes might contribute to the effect [29‐31].

To this end, we evaluated the mRNA expression levels of selected latent and lytic gammaherpesvirus genes, DNA methyltransferase 1 (DNMT1), the endogenous retrovirus ERV3-1, IFN-α, IFN-β, IFN-γ, IFN-λ2, and selected interferon-induced genes after treatment of KSHV- and EBV-infected tumor lines with 1 µM Abe for 24 h. After Abe treatment, the KSHV lytic genes ORF45 and ORF57 were significantly increased in both BCBL-1 and JSC-1 PEL lines (Fig. 6a). In addition, the lytic genes K2 (vIL-6), ORF50 (encoding (RTA)), and ORF73 (encoding latency-associated nuclear antigen (LANA)) were upregulated in BCBL-1 cells, but not JSC-1 cells. Also, treatment with 1 µM Abe for 24 h upregulated expression of EBER2, the most abundant EBV latent transcript, and BMRF1, a lytic gene, in Akata cells and JSC-1 cells, but not in Raji cells (Fig. 6b). Thus, Abe treatment variably increased expression of gammaherpesvirus latent and lytic genes in these cell lines.

DNMT1, which suppresses expression of endogenous retroviruses and has been reported to suppress of certain gammaherpesvirus genes [32], was substantially downregulated in both JSC-1 and BCBL-1 PEL cells and in Raji EBV-infected BL cells after treatment with 1 µM Abe; there was also a trend towards a decrease in Akata cells although the change was not significant (Fig. 6c). Expression of the endogenous retroviral gene ERV3-1, which is suppressed by DNMT1, was significantly upregulated in all the tested lines except Raji (Fig. 6c). In addition, the dsRNA sensor DDX58, which senses endogenous retroviruses and stimulates IFN signaling, was also upregulated in all four cell lines (Fig. 6c). We next looked specifically at expression of interferons. We found that IFNL2, IFN-α, and IFN-β, but not IFN-γ mRNA expression was increased in three of the lines tested (JSC-1, BCBL-1, and Akata), but not in Raji cells (Fig. 6c). Moreover, using an ELISA kit that measures IL28B plus IL29, two other members of the interferon family, we found an increase in the supernatants of JSC-1 and BCBL-1 cells after treatment with 1 μM Abe for 3 days (Additional file 7).

To evaluate the downstream effects of enhanced interferon activity, we measured the IFN-sensitive transcription factors STAT1 and NLRC5, as well as two IFN-stimulated genes, IFIT1 and OAS2 in these cells. All these 4 genes exhibited significantly enhanced expression in JSC-1, BCBL-1, and Akata cells after exposure to Abe (Fig. 6c). However, Raji was again somewhat of an outlier in that IFIT1 was not significantly changed and NLRC5 was decreased (Fig. 6c).