Inhibition of ATM or ATR in combination with hypo-fractionated radiotherapy leads to a different immunophenotype on transcript and protein level in HNSCC

Abstract

Background: The treatment of head and neck tumors remains a challenge due to their reduced radiosensitivity. Small molecule kinase inhibitors (smKI) that inhibit the DNA damage response, may increase the radiosensitivity of tumor cells. However, little is known about how the immunophenotype of the tumor cells is modulated thereby. Therefore, we investigated whether the combination of ATM or ATR inhibitors with hypo-fractionated radiotherapy (RT) has a different impact on the expression of immune checkpoint markers (extrinsic), the release of cytokines or the transcriptome (intrinsic) of head and neck squamous cell carcinoma (HNSCC) cells.

Methods: The toxic and immunogenic effects of the smKI AZD0156 (ATMi) and VE-822 (ATRi) in combination with a hypo-fractionated scheme of 2x5Gy RT on HPV-negative (HSC4, Cal-33) and HPV-positive (UM-SCC-47, UD-SCC-2) HNSCC cell lines were analyzed as follows: cell death (necrosis, apoptosis; detected by AnxV/PI), expression of immunostimulatory (ICOS-L, OX40-L, TNFSFR9, CD70) and immunosuppressive (PD-L1, PD-L2, HVEM) checkpoint marker using flow cytometry; the release of cytokines using multiplex ELISA and the gene expression of Cal-33 on mRNA level 48 h post-RT.

Results: Cell death was mainly induced by the combination of RT with both inhibitors, but stronger with ATRi. Further, the immune phenotype of cancer cells, not dying from combination therapy itself, is altered predominantly by RT+ATRi in an immune-stimulatory manner by the up-regulation of ICOS-L. However, the analysis of secreted cytokines after treatment of HNSCC cell lines revealed an ambivalent influence of both inhibitors, as we observed the intensified secretion of IL-6 and IL-8 after RT+ATRi. These findings were confirmed by RNAseq analysis and further the stronger immune-suppressive character of RT+ATMi was enlightened. We detected the down-regulation of a central protein of cytoplasmatic sensing pathways of nucleic acids, RIG-1, and found one immune-suppressive target, EDIL3, strongly up-regulated by RT+ATMi.

Conclusion: Independent of a restrictive toxicity, the combination of RT + either ATMi or ATRi leads to comprehensive and immune-modulating alterations in HNSCC. This includes pro-inflammatory signaling induced by RT + ATRi but also anti-inflammatory signals. These findings were confirmed by RNAseq analysis, which further highlighted the immune-suppressive nature of RT + ATMi.

Keywords: ATM inhibition; ATR inhibition; DNA damage repair; HNSCC; immunomodulation; kinase inhibitors.

Figure 1

Treatment scheme of HNSCC cell lines. HPV-negative (HSC4, Cal-33) and HPV-positive (UM-SCC-47, UD-SCC-2) HNSCC cells were seeded on day 1 and treated with 1µM AZD0156 or 0,1 µM VE-822 24 hours later. Irradiation with 5 Gy was performed 3 hours after treatment with smKI and repeated 24 hours later. Cells were harvested and analyzed 24 hours after the last treatment.

Figure 2

Cell death analysis of HNSCC treated with ATMi or ATRi w/o 2x5Gy RT. RT combined with ATRi shows the highest toxicity in HNSCC cells, regardless of the HPV status. 48 hours post-RT, HPV-negative HSC4 (A) and Cal-33 (B) cells as well as HPV-positive UM-SCC-47 (C) and UD-SCC-2 (D) cells were identified as viable (AnnexinV-, PI-), apoptotic (AnnexinV+, PI-) or necrotic (AnnexinV+, PI+). Percentages of viable, apoptotic, and necrotic cells are shown as stacked bars representing the mean ± SD. A one-tailed Mann-Whitney U test was performed to compare the different treatment approaches within one cell line: *p ≤ 0.05, n=4.

Figure 3

Flow cytometric analysis of immune checkpoint surface marker on HNSCC after ATMi or ATRi w/o 2x5Gy RT. Combining RT with ATRi leads to the upregulation of immunostimulatory ICOS-L and CD137-L on HNSCC cells, irrespective of the HPV status. 48 hours after the last treatment, the HNSCC cells were harvested and the expression of immunostimulatory ICMs ICOS-L and CD137-L was examined by flow cytometry. The gating strategy is presented in (A) ICM expression is presented as ΔMFI (delta of mean fluorescence intensity of stained samples – background fluorescence) (B, C). The combined treatment of RT and ATRi increased ICOS-L expression in all four cell lines compared to ATRi or RT alone, as well as in comparison to RT combined with ATMi (B). Dual treatment with RT and ATMi tended to even decrease ICOS-L surface expression compared to RT or RT with ATRi (B). RT and ATRi led to enhanced expression of immunostimulatory CD137-L in HPV-negative HSC4 cells compared to ATRi or RT alone, or the combined treatment of RT and ATMi (C). In HPV-positive UM-SCC-47 and UD-SCC-2 cells, the combination of RT and ATRi also demonstrated higher surface expression of CD137-L compared to RT alone or treatment with RT and ATMi, respectively (C). A one-tailed Mann-Whitney U test was performed to compare the different treatment approaches within one cell line: *p ≤ 0.05, n=4.

Figure 4

Flow cytometric analysis of immune checkpoint surface marker on HNSCC after ATMi or ATRi w/o 2x5Gy RT. Combining RT with ATRi leads to decreased expression of PD-L2 in HPV-positive cell lines UM-SCC-47 and UD-SCC-2. 48 hours after the last treatment, the HNSCC cells were harvested and the expression of the immunosuppressive ICMs PD-L1 (A), PD-L2 (B), HVEM (C), OX40-L (D) and CD70 (E) was examined by flow cytometry. ICM expression is presented as ΔMFI (delta of mean fluorescence intensity of stained samples – background fluorescence) (A–E). Combined treatment with RT and ATRi had no significant impact on the expression of PD-L1 compared to RT alone, except for the Cal-33 cell line, where it resulted in upregulation. However, it led to decreased expression of PD-L2 (B) in both HPV-positive cell lines and decreased expression of HVEM (C) in one HPV-positive cell line. The impact of the treatment approaches on the expression of OX40-L (D) and CD70 (E) showed cell line-specificity, regardless of the HPV status. A one-tailed Mann-Whitney U test was performed to compare the different treatment approaches within one cell line: *p ≤ 0.05, n=4.

Figure 5

Multiplex ELISA analysis (MSD) of seven cytokines released by HNSCC after ATMi or ATRi w/o 2x5Gy RT. Combining RT with ATRi enhances the secretion of proinflammatory cytokines IL-8, IL-6, TNF-α, IL-13, IL-12, IL-1β, IL-10 by HNSCC cells, independent of their HPV status. Cytokine secretion was quantified 48 hours after the last treatment. Samples were collected in four independent experiments (n = 4) and the mean values + standard deviation (SD) were used for statistical analysis. The combined treatment of RT and ATRi led to an increased secretion of immunostimulatory cytokines IL-8 (A), IL-6 (B), TNF-α (C), IL-13 (D) compared to the combination of RT and ATMi in all four cell lines, along with increased levels of IL-12p70 (E) and IL-1β (F) in two or three out of four cell lines, respectively. Regarding immunoinhibitory cytokines, combined treatment with RT and ATRi led to increased levels of IL-10 in three of the four cell lines (G). A one-tailed Mann-Whitney U test was performed to compare the different treatment approaches within one cell line: *p ≤ 0.05, n=4.

Figure 6

RNAseq analysis of treated Cal-33. Modulation of specific transcripts based on a RNAseq analysis of treated HPV-negative Cal-33 cells (RT vs. RT+ATMi vs. RT+ATRi). Cells were treated according to the standard scheme in this study (see Figure 1 ). Immune-modulating cell surface marker such as CD70 (A), TNFRSF9 known as CD137-L (B), TNFRSF8 known as CD30 (C) and LGALS9 known as Galectin-9 (D) were significantly downregulated by the combination of RT and ATMi AZD0156. Further, soluble factors such as IDO1 (E) and immune-related cytokines such as IL-24 (F), IL-1b (G), IL-1a (H) and IL-6 (I) were also downregulated by the combination of RT + ATMi. Data representing 4 independent replicates. All transcripts were declared as significant when log-(foldchange) > ± 1 was determined and the adjusted p-value was ≤ 0.050. All transcripts were declared as significant (*) when log-(foldchange) > ±1 was determined and the adjusted p-value was ≤0.050.

Figure 7

Heatmap including subsets of shared DEGs of Cal-33 treated with RT, RT+ATMi or RT+ATRi for 48h. First subset shows all genes down-regulated only in RT+ATMi and includes 690 genes. Second subset shows all genes down-regulated only by RT (20 genes). Third subset shows all down-regulated genes by RT and RT+ATRi (188 genes). Forth subset shows all genes down-regulated only in RT+ATRi treated Cal-33 (50 genes). Fifth subset shows all genes down-regulated by both combinations RT+ATMi and RT+ATRi (34 genes). Data representing 4 independent replicates. All transcripts were declared as significant when log-(foldchange) > ± 1 was determined and the adjusted p-value was ≤ 0.050. Representative MA plots show comparison of RT vs. RT+ATMi treated Cal-33 mRNA expression.

Figure 8

Treatment of Cal-33 with a combination of hypo-fractionated RT and ATM inhibitor AZD0156 leads to significant downregulation of nucleic acid sensor RIG-1 and significant up-regulation of immune-execution protein EDIL3. RIG1 plays a central role in recognition and processing of dsRNA (double-stranded RNA) in mammalian cells and is strongly down-regulated by the combination of ATM inhibitor AZD0156 with RT compared to RT or RT+ATRi treatment alone. In contrast to this, EDIL3, which is associated with immune exclusion and poor prognosis, is highly up-regulated by this combination, without differing between RT or RT+ATRi treatment in Cal-33. All transcripts were declared as significant (*) when log-(foldchange) > ±1 was determined and the adjusted p-value was ≤0.050.