Irradiation suppresses STAT3-mediated MCL1 expression to augment CD8+ T cells cytotoxicity against EGFR-positive lung cancer

Tumor cells progress to evade immunological attacks and prohibit activity of CD8 + T cells. Irradiation damages tumor cells and augments tumor immunotherapy in clinical application. However, the detail mechanism remains elusive. We aimed to uncover the mechanism of irradiation augmenting cytotoxic CD8 + T cells to suppress tumor progression in non-small-cell lung cancer (NSCLC). NSCLC cell lines were co-cultured with isolated PBMCs from healthy volunteers, cell viability and apoptosis were RNAseq was to screen the IFNγ-mediated gene expression in A549 cells. was to augment PBMCs-mediated anti-tumor gene expression was qPCR and Western blots. to in the of and 5 and overnight. The medium for and of the in a The steps, including and are


Abstract Background
Tumor cells progress to evade immunological attacks and prohibit activity of CD8 + T cells. Irradiation damages tumor cells and augments tumor immunotherapy in clinical application. However, the detail mechanism remains elusive. We aimed to uncover the mechanism of irradiation augmenting cytotoxic CD8 + T cells to suppress tumor progression in non-small-cell lung cancer (NSCLC).

Methods
EGFR-positive NSCLC cell lines were co-cultured with isolated PBMCs from healthy volunteers, cell viability and apoptosis were measured. RNAseq was used to screen the IFNγ-mediated gene expression in A549 cells. Irradiation was used to augment PBMCs-mediated anti-tumor effect and the irradiation effect to IFNγmediated gene expression was investigated using qPCR and Western blots.

Results
Co-culture of tumor cells stimulates increase of granzyme B and IFNγ in CD8 + T, but A549 exhibits resistance against CD8 + T cytotoxicity. Irradiation inhibits A549 proliferation and enhances apoptosis, augmenting PBMCs-mediated cytotoxicity against A549. IFNγ simultaneously increased phosphorylation on STAT1 and STAT3 in EGFR-positive lung cancer, resulting in overexpression of PD-L1. In RNAseq analysis, MCL1 was identi ed and increased by IFNγ-STAT3 axis in A549 cells, we found that irradiation speci cally inhibits phosphorylation on STAT1 and STAT3 in IFNr-treated A549, resulting in reductions of PD-L1 and MCL1.

Conclusion
This study demonstrated that A549 expressed MCL1 against CD8 + T cell-mediated apoptosis. In addition, we found that irradiation suppressed STAT3 phosphorylation and IFNγ-mediated PD-L1 and MCL1 expression, revealing a potential mechanism of irradiation augmenting immune surveillance.

Background
Lung cancer is the most common type and the leading cause of cancer-related deaths worldwide [7]. The most common type is non-small cell lung cancer (NSCLC) which comprises 80% of lung cancers [56].
Selection of treatment for NSCLC mainly depends on the diagnostic stage of this disease and patient's immune diversity. Basic treatments included surgical resection and radiotherapy (RT) that are often preferred and e cient for the patients with early-stage of NSCLC, whereas RT is used at least once in over 60% of lung cancer patients until cure or palliation of this disease [35]. Moreover, sequential chemotherapy with RT or concurrent chemoradiotherapy is also suggested for disease remission in patients with locally advanced stage of lung cancer [5]. It indicates that RT plays critical for NSCLC treatment, such as external beam radiation therapy and brachytherapy.
Personal medicine targeting speci c gene characteristics currently applied in lung cancers, that results in better outcome for this disease [18]. The main characteristic of NSCLC is EGFR overexpression: targeted therapies against EGFR such as tyrosine kinases inhibitors (TKIs), including the rst-generation TKIs ge tinib and erlotinib, the second-generation TKI afatinib for EGFR and HER2 blockades, and the third-generation TKI osimertinib for EGFR T790M mutation have been demonstrated to improve clinical progression-free survival (PFS) compared with the conventional chemotherapies in patients with advanced NSCLC [17,21]. For tumor recurrence and resistance to TKIs, other treatments such as platinum-based chemotherapy combined with an agent directed to either ALK rearrangements or ROS1 rearrangements are recommended for patients with metastatic stage of NSCLC [16,51].
Besides, immunotherapy reactivating CD8 + T cells such as anti-PD-1 antibodies exhibits tumor therapeutic promise against NSCLC and results in 20% tumor remission in the patients fail to treatments of chemotherapies or targeted therapies [28,52,55]. To date, the monoclonal antibodies targeting PD-1 and PD-L1 interaction, pembrolizumab and nivolumab for PD-1 and atezolizumab for PD-L1, are approved by FDA for the treatment of metastatic NSCLC [34]. The immune surveillance is responsible for outside infections and currently found to potentially eradicate tumors: CD8 + T cells recognize tumor cells and secrete ganzymen B (GZMB) and perforin (Prf1) to elicit tumor cell apoptosis and death [14]. Immune checkpoints expressing in tumor cells, including PD-L1, galactine 9, HVEM, are able to exhaust CD8 + T cells and reduce CD8 + T-derived cytotoxicity [1,43,53]. In tumor microenvironment, IFNγ is secreted by T lymphocytes to reactivate macrophages and CD8 + T cells, but it also stimulates PD-L1 overexpression in tumor cells through JAKs-STAT1 signaling pathway [24]. Meanwhile, EGFR-STAT3 axis also mediates PD-L1 expression in lung cancer [12]. A case report has revealed a successful tumor therapy by Pembrolizumab immunotherapy in TKIs refractory NSCLC [57]. But EGFR-resistance and downstream signaling gene STAT3 activated by other stimulations contribute to cell proliferation, anti-apoptosis, and resistance to CD8 + T cells in EGFR-positive lung cancer [46,61].
Irradiation (IR) causes tumor DNA damage and results in consequent tumor apoptosis. In IR treatment, the accumulation of cytosolic DNA activates cGAS-STING signaling pathway [42], inducing type I IFN secretion and potentially reactivates immune surveillance against tumors. According to the preclinical and early clinical data, it demonstrates that IR enhances expression of MHC class I [19] and secretion of cytokines such as type I IFN [39], resulting in increase of homing rate of immune cells to tumor microenvironment [22,41]. Therefore, IR-mediated tumor therapies are considered for improving anti-tumor e cacy of clinical immunotherapies [27,37,50]. In addition, type I IFN is able to induce PD-L1 expression that PD-L1 expression detection plays a critical role for anti-PD-1 therapies [2,15]. Previous studies have demonstrated that low doses of fractionated IR signi cantly improves CD8 + T cell-mediated tumor remission in combination with anti-PD-1 or anti-PD-L1 therapies [15,25,26]. However, IFNγ from T lymphocytes and type I IFN from IRtreated tumors are able to activate JAKs-STAT1 axis [24,49], that induces expressions of not only PD-L1 but also cell proliferation and anti-apoptosis genes [31,54]. Since a previous study has indicated that UV light down-regulates IFNγ-mediated STAT1 phosphorylation [3], we, therefore, proposed a hypothesis that IR may augment CD8 + T cells-mediated cytotoxicity through blocking IFNs-mediated activations of JAKs-STATs to against immunotherapies-resistant EGFR-positive NSCLC.

Healthy volunteers
Clinical blood was acquired in Mackay Memorial Hospital, Taipei, Taiwan. The study protocol was approved by regulatory authorities and Institutional Review Boards in Mackay Memorial Hospital (20MMHIS018e).
Signed and informed written consent were obtained from all participants, and all research was performed in accordance with the relevant guidelines and regulations.

Cell culture
All the lung cancer cell lines used in this study were free of mycoplasma and reauthenticated through short tandem repeat pro ling (Applied Biosystems, Massachusetts, USA). A549, PC9, and H1650 were cultured in Dulbecco's Modi ed Eagle's Medium (DMEM). HCC827 was cultured in Roswell Park Memorial Institute (RPMI) 1640 medium. The tumor cells cultured in the medium were supplied with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin (P/S). All cells were incubated at 37°C with 5% CO 2 .

Cell viability
The WST-1 assay (Sigma, Munich, Germany) was used to measure the cell viability according to the manufacturer's protocol. In brief, 1 x 10 3 cells were seeded in a 96-well plate, whereas 4-times repeat was necessary for measuring the cell viability in a time-dependent manner. For evaluating the cytotoxic effect of irradiation, A549, HCC827, PC9, and H1650 were exposed to 0, 10, 20 Gy of irradiation and incubated for 48 h.
For evaluating the cytotoxic effect of PBMC and CD8 + T cells, 1 x 10 5 PBMCs or 1 x 10 4 CD8 + T cells were cocultured with 1 x 10 3 A549, HCC827, PC9, and H1650 cells individually and incubated for 48 h. The same methodology was also used to determine the cytotoxic effect of PBMCs against A549shSTAT1, A549shSTAT3, A549shMCL1, A549shPDL1 compared to A549shLuc.

Western blots
The procedure of Western blots was described previously [10]. The speci c antibodies against STAT1, pSTAT1(Tyr701), STAT3, pSTAT3(Tyr705), ERK, pERK, and GAPDH were purchased from Cell Signaling Quantitative polymerase chain reaction (qPCR) The procedures for mRNA extraction and complementary DNA preparation were the same as described previously [10,13]. Quantitative polymerase chain reaction (qPCR) was performed using an SYBR Green system (Applied Biosystems, Foster City, CA, USA) according to the manufacturer's instruction. The primers are shown in Table 1.

Gene knockdown
Gene knockdown was conducted using a short-hairpin RNA (shRNA)-expression lentivirus system that contains the speci c shRNA (The target sequences are shown in Table 2) in the pLKO.1-puro vector generated by 293T cells. The plasmids were purchased from the National RNAi Core Facility of Academia Sinica, Taipei, Taiwan. The procedure was the same described in our previous study [13]. In brief, 293T cells (70% con uence) cultured in DMEM culture medium were transfected with 4 µg of pLKO.1 vector, 1 µg of the envelope plasmid pVSV-G, and 3.6 µg of the packaging plasmid pCMVΔR8.91. The plasmids were preincubated with 6 µL of JetPRIME (Polyplus-transfection, New York, NY, USA) for 20 min at room temperature and consequently added to 293T cells. The cultured medium was substituted with a fresh culture medium after 24 h and further incubated for 48 h. The virus solution was collected and stored at -80°C. A549 cells cultured in 80% con uence were infected with the prepared lentivirus for 24 h. The cells were then changed with DMEM medium containing 4 µg/mL of puromycin, which harvested after obtaining stable cells.

Isolation of peripheral blood mononuclear cells (PBMCs) and CD8 + T cells
The procedure for PBMCs and CD8 + T cell isolation was the same described previously [9]. In brief, blood was collected and analyzed from the individual healthy volunteer (n = 16, Table S1) in Mackay Memorial Hospital, Taipei, Taiwan. 6 mL of buffy coats were isolated from 20 mL of whole blood by an initial 1200 rpm centrifugation for 30 min. The collected 1 mL of buffy coats was mixed with 7 mL of PBS buffer and loaded onto the 4 mL of coll solution and consequently for 2000 rpm gradient centrifugation for 20 min. PBMCs on the interface between the plasma and coll media were collected for further analysis. The CD8 + T cell isolation kit (Miltenyi Biotec, North Rhine-Westphalia, Germany) was used to isolate CD8 + T cells from the PBMCs. The isolation steps were followed based on the manufacturer's manual.
All analytical data with more than two groups were evaluated using analysis of variance, followed by post hoc analysis with Bonferroni's test. The student's t-test was used to compare two groups. Moreover, p < 0.05 was considered to indicate a statistically signi cant difference.

Irradiation and PBMCs synergistically inhibited tumor cell viability and induced apoptosis in lung cancer cells
To investigate whether irradiation augments immunological surveillance against NSCLC cells, the EGFRpositive tumor cell lines, including A549 (wild-type EGFR; KRAS mutation), HCC827, PC9, and H1650 (EGFR E746-A750 deletion: autophosphorylation), were selected for treatments with irradiation and the isolated healthy PBMCs. Cell viability and apoptosis were then measured. We found that irradiation signi cantly suppressed cell viability on A549, HCC827, PC9, and H1650 (Fig. 1A). In addition, 20 Gy of irradiation increased late apoptosis rates on A549, HCC827, and PC9 (Fig. 1B), and early apoptosis rates on HCC827 and PC9 (Fig. 1B). By contrast, CD8 + T cells signi cantly suppressed HCC827, PC9, and H1650 cell viabilities but did not on A549 (Fig. 1C). We noticed that HCC827 pre-incubated CD8 + T cells possessed higher cytotoxic e cacy than the parental CD8 + T cells to suppressing cell viability on HCC827, but the phenomena did not observe on A549 cells (Fig. 1C). Furthermore, A549 cell line was selected as CD8 + T-resistant model to investigate the synergistic effect of irradiation on immunological surveillance. We found that combined irradiation and PBMCs treatment signi cantly led to remarkable inhibition on A549 (Fig. 1D), also increased late apoptosis and dead cell rates compared to individual irradiation or PBMCs treatment (Fig. 1E).

Reactivation of healthy CD8 + T cells in encountering HCC827 and A549 in vitro
To further investigate the mechanism on A549 cells resisting cytotoxicity of the healthy CD8 + T cells, the antitumor status of reactivation on immune surveillance was measured, particularly T lymphocytes. We rst measured the percentages of Naïve, central memory (CM), effector memory (EM), and terminally differentiated effector memory (TDEM) in the isolated healthy PBMCs through detecting CD45RA and CCR7 using ow cytometry ( Fig. 2A). The percentage of the reactivation markers, including CD16, CD25 (IL2RA), CD28, CD122 (IL2RB), CD178 (FASL), and inhibitory marker CD279 (PD-1) were measured in the isolated healthy PBMCs depending on CD4 + or CD8 + lymphocytes (Fig. 2B). We noticed and validated that CD16 and CD122 were positively correlated in CD4 + and CD8 + T cells (Fig. 2C). The personal characteristics of the enrolled volunteers were shown in Table S1 and Table S2. Moreover, we found that CD16-high expression CD8 + T cells were able to signi cantly inhibited HCC827 cell viability compared to CD16-low expression CD8 + T cells (Fig. 2D). To make sure CD8 + T cells were reactivated by tumor cells, the PBMCs were co-incubated with HCC827 and A549 cells for 24 h and CD8 + T cells were isolated and analyzed consequently. The GZMB as the reactivation marker for CD8 + T cells and the qPCR results revealed that HCC827 and A549 both activated CD8 + T cells (Fig. 2E). Meanwhile, A549 pre-treated with IFNγ for 2 h caused a remarkable increase of GZMB in isolated CD8 + T cells (Fig. 2E). Moreover, CD16 and CD122 both increased in HCC827-incubated CD8 + T cells and only CD122 increased in A549-incubated CD8 + T cells (Fig. 2F). Since we demonstrated that irradiation improved anti-tumor cytotoxicity of PBMCs ( Fig. 1D and 1E), the irradiation marker ISG15 was measured and it increased in irradiated A549 in a dose-dependent manner (Fig. 2G). ISG15 is a secreted protein playing as a cytokine to activate CD8 + T cells through LFA-1 receptor [47]. Therefore, the cultured medium was collected and added to PBMCs, and GZMB and IFNγ expressions were measured in CD8 + T cells. We found that A549 cells stimulated CD8 + T cells to express GZMB (Fig. 2H) but supernatant medium stimulated CD8 + T cells to express IFNγ (Fig. 2I). In addition, A549 and irradiated A549 both stimulated PBMCs to secret IFNγ in 24 h of incubation, rather than HCC827 (Fig. 2J).

IFNγ dominantly phosphorylated STAT3 in the premise of phosphorylated EGFR
IFNγ is a cytokine secreted by T lymphocytes for further stimulating macrophages and CD8 + T cells. The increase of GZMB and IFNγ in CD8 + T cells and in the medium of the cell mixture, respectively, indicated that the CD8 + T cells were reactivated after A549 co-culture but we found A549 remained resistant to CD8 + T cells.
We found intrinsic STAT3 phosphorylation in HCC827 and PC9 cells but no STAT1 phosphorylation was detected in the selected cell lines (Fig. 3C). Further investigation revealed that IFNγ increased phosphorylation on STAT1 and STAT3 in HCC827 and A549 (Fig. 3D). Interestingly, consequent and synergic EGF treatment highly elicited IFNγ-mediated STAT3 phosphorylation with simultaneous inhibition of STAT1 phosphorylation (Fig. 3D). No effect of EGF was found on HCC827 because that HCC827 was an EGFR autophosphorylated cell line. The immune checkpoints, including PD-L1, galactin-9, and HVEM were measured using qPCR in the arranged treatments. We demonstrated that IFNγ signi cantly increased PD-L1 and consequent and synergic EGF treatment increased IFNγ-mediated PD-L1 expression in A549 cells (Fig. 3E).

IFNγ increased MCL1 expression in the premise of phosphorylated EGFR
Besides PD-L1 overexpression in A549 to exhaust CD8 + T cells, we further investigated the mechanism eliciting resistance against CD8 + T cells in A549 cells. RNAseq was used to investigate the differential gene expression derived by IFNγ treatment. The differential genes > 2 and <-2 fold changes with p-value < 0.001 were selected and analyzed ( Fig. 4A and Table S3). NetworkAnalst revealed the driver gene from the IFNγincreased differential genes, including PSMB8, PSMB10, NFKB2, SOCS3, UBE2L6, STAT1, EFNA1, PML, FASL1, MCL1, and STAT2 (Fig. 4B). After validation using qPCR, SOCS3, UBE2L6, STAT1, FASL1, MCL1 were signi cantly increased in IFNγ-treated A549 (Fig. 4C). The ve genes were associated with survival probability in patients with lung cancer using Kaplan-Meier plotter analysis, which demonstrated that increase of each gene was associated with poor survival rate (Fig. 4E). SOCS3 and MCL1 were particularly investigated since literature indicated that they were associated with apoptosis [38] and anti-apoptosis function [58], respectively. PBMCs co-culture signi cantly increased SOCS3 and MCL1 expression in A549 cells (Fig. 4D).

Irradiation speci cally blocked IFNγ-mediated phosphorylations on STAT1 and STAT3 in A549 cells
The IFNγ-mediated phosphorylations on STAT1 and STAT3, and gene expressions on PD-L1, SOCS3, and MCL1 were detected in irradiation-treated A549 cells. The results indicated that 20 Gy of irradiation speci cally inhibited IFNγ-mediated STAT1 phosphorylation and synergic effect of EGF with IFNγ on STAT3 phosphorylation (Fig. 5A). PD-L1, SOCS3, and MCL1 expressions were consequently reduced in 20 Gy of irradiation treatment compared to the corresponding arranged treatments (Fig. 5B).
We also demonstrated that IFNγ-mediated SOCS3 and MCL expressions were decreased in A549shSTAT3 cells but not in A549shSTAT1 cells when compared to A549shLuc cells (Fig. 6C). To clarify the potential regulatory mechanism, we used ChIP to investigate whether STAT3 and STAT1 directly bind to the promoters of individual PD-L1 and MCL1. The results revealed that STAT3 and STAT1 both were able to bind to the promoters of PD-L1 (Fig. 6D) and MCL1 (Fig. 6E) in IFNγ and IFNγ with consequent EGF treatments.
Furthermore, PD-L1 and MCL1 were knockdowned in A549 cells, the four stable A549 knockdown cell lines, including A549shPD-L1, A549shSTAT3, A549shSTAT1, and A549shMCL1 were further investigated compared to A549shLuc after co-cultured with individual healthy PBMCs. We found that cell viability was decreased in A549shSTAT3 and A549shMCL1 compared to A549shLuc in PBMCs treatment for 48 h (Fig. 6F). Knockdown of PD-L1 and STAT1 did not affect the PBMCs cytotoxicity against A549 and knockdown of STAT1 slightly increased cell viability (Fig. 6F). To address and validate that MCL1 elicited resistance to the cytotoxicity of PBMCs, apoptosis was measured in A549shMCL1 with irradiation and PBMCs treatments. The results demonstrated that knockdown of MCL1 increased A549 cell death rate in 20 Gy with PBMCs treatment compared to control A549shLuc cells (Fig. 6G).

Discussion
In this study, we used healthy PBMCs to investigate the mechanism of IR augmentation for clinical tumor immunotherapies but the status of immune surveillance is ultimately different between healthy volunteers and tumor patients, including macrophages and T lymphocytes. Literature has indicated that macrophages are usually classi ed as M1 (classical-activated macrophages) and M2 (tumor-associated macrophages, TAMs) phenotype [8], that M1 macrophages enhance in ammation response against initial tumor cells, but M2 macrophages suppresses immune system to trigger tumor progression. TAMs enriched in the tumor microenvironment. Moreover, CD8 + T cells are exhausted in patients with tumors. Exhausted CD8 + T cells are characterized by loss of cytotoxic functions with increased expression of multiple inhibitory receptors such as PD-1 and decreased secretion of effector cytokines such as IFNγ. Current immunotherapies are mainly designed to reactivate the immune system in tumor patients. For example, TAMs as a therapeutic target [20] that the TAM-targeted therapeutics are mainly focused on to eliminate M2 or educate M2 to transfer to M1 phenotype [45]. Reactivation of the exhausted CD8 + T cells with inhibitory receptor blockade promotes immunity and survival outcome in tumor patients. Besides, CD8 + T cells can also be activated by inserting a speci c TCR recognizing tumor antigen called chimeric antigen receptor T-cell therapy (CAR-T). Once activation in CD8 + T cells, IFNγ is secreted and the function of which is to bind and activates macrophages to induction of M1 differentiation [36,44]. However, IFNγ also induced JAKs-STATs signaling pathways in tumor cells. We clari ed the potential IFNγ-mediated mechanism resisting to cytotoxicity of CD8 + T cells through expressing MCL1 in EGFR-positive NSCLC in this study.
We have previously demonstrated that CD16 (FcγRIIIA) and CD122 (IL2/IL15Rβ) are correlated in the healthy CD8 + T cells [11]. This study also evaluated and demonstrated CD16-highly expressed CD8 + T cells possessed higher anti-HCC827 capacity compared to CD16-low expressed CD8 + T cells. To our knowledge, CD16 on CD8 + T cells acts as an activation site exerting antibody-dependent cellular cytotoxicity (ADCC) function, the levels of which are decreased in exhausted CD8 + T cells treated by nicotine treatment and in smokers [11]. CD16 + CD8 + T cells exhibit natural killer (NK)-like and are terminally differential memory effector T phenotype with high levels of granzyme B and perforin [6,40] [59]. Moreover, CD16 is frequently increased in the unconventional NK-like CD8 + T cells [4]. To our knowledge, Eomes determines the differentiation of CD8 + T cells into effector and memory phases and regulates CD122 expression [30]. CD122 is the receptor of IL2 and IL15 contributing to T cell proliferation and granzyme B expression through activating JAKs-STAT5. A literature has indicated that IL15 is able to substantially increase CD16 expression in CD8 + T cells [4]. In addition, IL2 is used in clinical applications to against tumors. Based on the evidence, CD16 + CD122 + CD8 + T cells may re ect the reactivation of CD8 + T cells as a biomarker of cancer immunotherapies since they are surface receptors to be easily detected.
EGFR overexpressed in the selected cell lines, including A549, HCC827, PC9, and H1650, and intrinsic PD-L1 expression was detected. The EGFR autophosphorylation has been reported in HCC827, PC9, and H1650, which may lead to highly phosphorylation of STAT3 (Fig. 3C). Since the characteristic, EGF had no effect on STAT3 phosphorylation but IFNγ caused highly STAT3 phosphorylation in HCC827. Same phenomena were observed in the sequential treatment of EGF and IFNγ in A549 (Fig. 3D). It also validated and revealed that STAT3 phosphorylation suppressed STAT1 phosphorylation, indicating the possibility the lower levels of PD-L1 in HCC827 compared to that in A549. We expected that low levels of STAT1 and pSTAT1 in HCC827 resulted in susceptibility to the cytotoxicity of CD8 + T cells since we demonstrated that STAT1 bound to the promoter of PD-L1 and MCL1. Compared to HCC827, there was an instant increase of MCL1 in the sequential IFNγ with EGF-treated A549 (Fig. 4F). Meanwhile, JAKs-STATs signaling axis also regulated MHCI expression [32,60]. We found that IFNγ speci cally increased non-classical MHCI HLA-E molecules which is the ligand of CD94/NKG2 receptors expressed in NK cells and a subset of CD8 + T cells [48]. Therefore, IFNγ-pretreated A549 exhibited higher immunogenic capacity to immunity since we uncovered the isolated CD8 + T cells with highly granzyme B expression (Fig. 2E). However, IFNγ-pretreated A549 also expressed highly MCL1 expression responding to PBMCs (Fig. 4D). This study provides evidence suggesting that IFNγ-mediated JAKs-STATs axis plays a double-edged sword in NSCLC stimulating reactivation of CD8 + T cells but increasing anti-apoptosis MCL1 expression in A549.
This study provides evidence that IR enhances the reactivation of CD8 + T cells against NSCLC through speci cally inhibiting phosphorylation of STAT1 and STAT3 and reducing the IFNγ-mediated anti-apoptosis MCL1 expression. IR is considered to break DNA and leads to consequent tumor apoptosis as reliable tumor therapy. Although previous studies have indicated that tumor cells under IR-mediated stress trends to stimulate STAT3 activation for spontaneous defends to DNA damage and cellular apoptosis [23]. With su cient IR strength causing irreversible progress such as 20 Gy in this study, cellular cytosolic DNA activates cGAS-STING signaling pathway and results in type 1 IFN-mediated ISG15 expression. ISG15, therefore, secreted by IR-treated tumors play as a cytokine that activates CD8 + T cells through binding to LFA1 receptor, resulting in IFNγ expression in CD8 + T cells [47]. Interestingly, this study demonstrated that IR suppressed phosphorylation on STAT1 and STAT3 speci cally. We speculated that SHP1/2 may be activated by IR [29,33], which dephosphorylated STAT1 and STAT3. Based on the ndings, we demonstrated a new mechanism of radiotherapy to augment immunity, that may indicate that radiotherapy enhances anti-tumor e cacy of immunotherapies in clinical practice.

Conclusions
In this study, we demonstrated that IR elicited A549 cell apoptosis and augmented PBMCs-mediated cell death through prohibiting IFNγ and EGF-mediated phosphorylations on STAT1 and STAT3 and gene overexpression on PD-L1 and MCL1 (Fig. 7). Healthy CD8 + T cells were stimulated and reactivated in cocultured with HCC827 and A549 since granzyme B (GZMB) and IFNγ were increased. Granzyme B was considered a hydrolase enzyme leading to tumor cell apoptosis but CD8 + T also secreted IFNγ and resulted in STAT3-mediated MCL1 expression against cell apoptosis. We gured out the mechanism of IR augmented immune surveillance against EGFR-positive NSCLC in this study.

CONSENT FOR PUBLICATION
Not applicable.

CONFLICT OF INTEREST
We declare no con ict of interest for this article.
ANIMAL SOURCE Not applicable.

CELL LINE AUTHENTICATION
The lung cancer cell lines A549, HCC827 and PC9 and H1650were purchased from the American Type Culture Collection (ATCC, Manassas, VA, USA); HCC827 and A549 were reauthenticated through short tandem repeat    Figure 1