Suppression of tumor immune microenvironment via microRNA‐1 after epidermal growth factor receptor‐tyrosine kinase inhibitor resistance acquirement in lung adenocarcinoma

Abstract Immunotherapy is considered one of the most important therapeutic strategies for patients with lung adenocarcinoma after the development of epidermal growth factor receptor tyrosine kinase inhibitor (EGFR‐TKI) resistance. However, useful predictors of immunotherapy for these patients has not been examined well, although the status of the tumor immune microenvironment (TIME), including programmed death‐ligand 1 expression and lymphocyte infiltration, has been generally known to provide predictive markers for the efficacy of immunotherapy. This study aimed to clarify novel predictors of immunotherapy following EGFR‐TKI resistance in lung adenocarcinoma, especially regarding micro RNA (miRNA). We evaluated the correlation between EGFR‐TKI resistance and lymphocyte infiltration, before and after acquiring EGFR‐TKI resistance, in 21 cases of lung adenocarcinoma, and further explored this by in vitro studies, using miRNA PCR arrays. Subsequently, we transfected miRNA‐1 (miR‐1), the most variable miRNA in this array, into three kinds of lung cancer cells, and examined the effects of miR‐1 on EGFR‐TKI sensitivity, cytokine expression and lymphocyte migration. Histopathological examination demonstrated that infiltration levels of CD8‐positive T cells were significantly decreased after development of EGFR‐TKI resistance. In vitro studies revealed that miR‐1 significantly inhibited EGFR‐TKI effect and induction of cytokines, such as C‐C motif chemokine ligand 5 and C‐X‐C motif chemokine ligand 10, causing inhibition of monocyte migration. These results indicate that the upregulated miR‐1 might suppress the TIME, following development of EGFR‐TKI resistance. Therefore, miR‐1 could be a clinically useful marker to predict therapeutic efficacy of immunotherapy in lung adenocarcinoma patients with EGFR‐TKI resistance.


| BACKGROUND
Lung cancer is a leading cause of cancer-related deaths, worldwide. 1,2 In non-small cell lung cancer (NSCLC), especially in adenocarcinoma, "driver mutations" have been discovered in genes such as EGFR (epidermal growth factor receptor), which are important to the processes of cancerization and proliferation. [3][4][5] In East Asian studies, approximately half of the adenocarcinoma patients were found to harbor EGFR mutations. 6 In addition, according to current guidelines for clinical practice, EGFR-tyrosine kinase inhibitor (TKI) is recommended as the first choice of therapy in the patients with advanced-stage NSCLC, harboring EGFR mutations. [7][8][9] EGFR-TKI exerts reasonable therapeutic effects upon EGFR-mutation-positive NSCLC patients [10][11][12][13][14][15][16] but, in many cases, the development of therapeutic resistance is inevitable within several months, due to the newly acquired T790M mutation, and others. 17,18 This is presently conceived as one of the most serious clinical problems of this particular therapy. On the other hand, an immune checkpoint inhibitor (ICI) has attracted great interest as a new treatment strategy for lung cancer patients and has been approved in EGFR-TKI-resistant NSCLC patients. In addition, immunohistochemical evaluation of programmed death-ligand 1 (PD-L1) in tumor cells is currently approved as a potential biomarker for predicting the clinical efficacy of ICI, by such as anti-programmed death 1 (PD-1) antibodies. 19 However, it is also true that the overall response rate to ICI in NSCLC is only about 20%, at most, 20,21 and 40%-50% in those with relatively abundant expression of PD-L1 in carcinoma cells. 22 In working toward solving these clinical problems, novel predictors of ICI therapy, other than the immunohistochemical status of PD-L1, have been explored. Among these, ICIs were reported to be more effective in the patients with abundant intra-tumoral infiltration of CD8+ T lymphocytes in lung cancer tissues, and these patients generally had good prognoses. 23,24 In addition, therapeutic effects of ICI have been reported to be poor in EGFR mutation-positive cases, 21,25 and there are some previous reports showing a correlation between EGFR-TKI resistance and poor benefit of ICI. 26 The administration of EGFR-TKI itself has been reported to change the tumor immune microenvironment (TIME), including reduced PD-L1 expression and forkhead box P3 (Foxp3)-positive T cell reduction, [26][27][28][29] and this change in PD-L1 expression has been cited as a reason for the poor effectiveness of ICI after the acquirement of EGFR-TKI resistance. 27,30 However, the relationship between changes in TIME and the mechanism of EGFR-TKI resistance in detail has not been examined yet.
Therefore, we hypothesized that the acquisition of EGFR-TKI resistance in carcinoma cells could alter the therapeutic efficacy of ICI through changing aspects of the TIME, such as the level of lymphocyte infiltration in NSCLC patients, based on the studies reported above. We also explored the status of microRNAs (miRNAs) in order to further elucidate its mechanisms, because miRNAs are well known to be involved in tumor development and progression [31][32][33] and in the EGFR signaling pathway and cytokine expression in lung carcinoma. 34 However, to the best of our knowledge, there have been few reports of a possible correlation between miR-NAs and ICI in lung cancer. This is the first study to examine the correlation between EGFR-TKI resistance and TIME, with particular emphasis on changes in miRNAs. Recently, ICI has become one of the main therapeutic methods after acquiring EGFR-TKI resistance, and it is considered to be important to study the relationship between them. In particular, we aimed to explore new predictors of therapeutic efficacy and the therapeutic targets of ICI in patients with EGFR-TKI-resistant lung adenocarcinoma.

| Patient information
A total of 119 cases were examined in this study in order to explore changes in the TIME before and after developing EGFR-TKI resistance. NSCLC specimens were obtained from Japanese patients who developed therapeutic resistance to EGFR-TKI therapy from 2010 to 2017 at Sendai Kosei Hospital, Tohoku University Hospital, and Miyagi Cancer Center. Of 119 cases, paired histological specimens of pre-and post-EGFR-TKI treatment were retrieved from 21 cases (totaling 42 samples; 36 samples: biopsy, six samples: surgical resection). All of these cases were histologically diagnosed as adenocarcinoma, seven male and 14 female, median age 73 years (range 56−87 years, standard deviation 7.2 years). Our cases included all major types of mutation of EGFR, including exon 19 deletion in nine cases, and L858R. Ninety-eight cases were not examined because either or both paired samples corresponded to lymph nodes or liquid samples such as pleural effusion, in which we could not evaluate the status of TIME. We could only evaluate the status of EGFR mutation in 10 of 21 specimens from the samples after acquiring EGFR-TKI resistance. As a result, five out of 10 cases were found to be positive for T790M (Table 1). The primary treatment was EGFR-TKI alone in 15 cases, chemotherapy (Carboplatin, Pemetrexed) combined with EGFR-TKI in three cases, where EGFR-TKI was used as the secondline treatment alone, or in combination with chemotherapy. Chemotherapy (Cisplatin, Carboplatin, Pemetrexed, Vinorelbine, Paclitaxel) was used alone in three cases. Carboplatin and Pemetrexed were the main drugs used for chemotherapy. All treatment was performed according to current clinical practice guidelines for treating lung cancer.
All specimens had been fixed in 10% formalin and embedded in paraffin. The study protocol was approved by the Ethics Committee at the Tohoku University School of Medicine (2018-1-612), Sendai Kousei Hospital  and Miyagi Cancer Center (2018-25-1).

| Histopathological evaluation
The stained slides were evaluated independently by two of the authors (SK and RS). We examined a whole slide for each individual case. The cells demonstrating higher immunointensity than the background were tentatively defined as "positive" in this study. Immunoreactivity of CD3, CD4, and CD8 was all detected in the cell membrane, that of Foxp3 in the nucleus of lymphocytes. The immunopositive rate for CD3 was used as a measure of infiltration level when we evaluated CD4, CD8 and Foxp3.

| Cell culture
The following cell lines were used in this study: PC9 (Riken Cell Bank, Tsukuba, Japan) and NCI-H1975 (H1975) (American Type Cell Culture Collection (ATCC), Manassas, VA, USA) for lung adenocarcinoma; peripheral blood mononuclear cells (PBMCs) from a healthy donor (Precision Bioservices, Frederick, MD, USA). PC9/6M and PC9/ER (erlotinib-resistant cell line) were established in our laboratory by methods reported previously. 35 PC9 (Immunobiological Laboratories) was exposed to erlotinib and the concentration was gradually increased at 2-month intervals as follows: 10 nM, 1 μM and 5 μM. PC9 cells were also cultured for 6 months (PC9/6M) without erlotinib, to eliminate the effects of long-term cell culture. The EGFR mutation profiles of these cell lines were as follows: PC9 and PC9/6M; exon 19 deletion, PC9/ER; exon 19 deletion, L858R mutation, and T790M mutation, H1975; L858R mutation and T790M mutation.

| miRNA polymerase chain reaction (PCR) array
Total RNA was extracted from PC9/6M and PC9/ER using the TRI reagent (Molecular Research Center, Cincinnati, OH, USA). The cDNA synthesis was performed using 1000 ng of total RNA, based on the recommended protocol, using the miScript HiSpec Buffer (Qiagen, Hilden, Germany) (37°C for 60 min and then at 95°C for 5 min). After the dilution of cDNA, PCR array was performed based on a SybrGreen protocol and using human tumor suppressor gene miScript miRNA PCR array (MIHS-119Z) (Qiagen) following the manufacturer's protocol. These arrays contain the probes for 92 miRNAs whose expression has been known to be altered

| Cytokine antibody array
In this study, we treated PC9 (2.0 × 10 5 cells/mL) with 50 nM miR-1 for 24 h in the RPMI 1640 medium containing 10% FBS in order to further assess the potential effects of miR-1 on secretion levels of cytokines. The medium was subsequently removed. The cells were then washed with phosphate-buffered saline (PBS) and incubated in phenol red-and FBS-free RPMI 1640 medium for 24 h. The conditioned medium was used in samples. We used the Human Cytokine Antibody Array 5 (RayBiotech, Peachtree Corners, GA, USA), which can detect 80 cytokines. The membranes were then spotted with cytokine-specific antibodies and were analyzed. The signal was detected using the Image Lab TM software (BIO-RAD Laboratories, Inc., Hercules, CA, USA).

| Cell viability assays
To assess the effect of miR-1 on EGFR-TKI sensitivity, PC9 (5000 cells/well) was seeded into 96-well plates and incubated in RPMI-1640 medium containing 10% FBS for several hours before transfection. After preincubation, the cells were transfected with miR-1 or miRNA negative control. After 24 h of transfection, the medium was removed and each cell was treated with 5 µM gefitinib (Biaffin, Kassel, Germany) or 5 µM erlotinib (Roche Diagnostics, Mannheim, Germany). At 0, 24, 48, and 72 h after the treatment, the cell numbers were evaluated using a Cell Counting Kit-8 (Dojindo, Kumamoto, Japan).

| Real-time RT-PCR
Total RNA was extracted carefully from the cells, using the TRI reagent (Molecular Research Center) and was reverse transcribed to cDNA using a QuantiTect Reverse Transcription Kit (Qiagen). The levels of mRNA expression were semiquantified using a real-time RT-PCR in a LightCycler System (Roche Diagnostics). The PCR mixture (20 μL) included 1.0-μM primer and 2× QuantiTect SYBR Green PCR Master Mix (Qiagen). The following PCR protocol was used in this study: initial denaturation at 95°C for 5 min, followed by 40 amplification cycles of 95°C for 10 s and annealing at 60°C for 30 s.

| PBMC migration assay
We performed migration assays using a Chemotaxicell chamber containing membranes with 5-μm pore size (Kurabo, Osaka, Japan) and 24-well plates. After treating PC9 (2.0 × 10 5 cells/ml) with 50-nM miR-1 for 24 h, we washed the cells with PBS and exchanged the medium to phenol red-and FBS-free RPMI 1640 medium. After 24 h of incubation, we used this as a conditioned medium in the lower chambers. The PBMCs were then plated in the upper chambers (4.5 × 10 4 cells/well) in phenol red-and FBS-free RPMI 1640 medium. After a further 24 h of incubation, the number of migrated cells in the lower chamber was counted using a TC 20TM Automated Cell Counter (Bio-Rad Laboratories, Inc.).

| Statistical analysis
Statistical analysis was performed using JMP Pro 14.0.0 (SAS Institute, Cary, NC, USA). Statistical differences between the two groups and paired analyses were evaluated by the Student t-test. Statistical significance was defined as p < 0.05 in this study.

| The changes to CD8-positive T lymphocyte infiltration after EGFR-TKI resistance acquirement
We compared the changes in tumor infiltration levels of CD4, CD8, or Foxp3-positive T cells in tumor tissues before and after EGFR-TKI treatment, in the patients who developed EGFR-TKI resistance. Representative findings are summarized in Figure 1A. There were no significant differences in the overall CD8-positive rate (p = 0.669) but, when limited to the cases harboring more than 30% or 40% of CD8-positive rate before EGFR-TKI treatment, the rate significantly decreased following the EGFR-TKI treatment (30%: p = 0.009, 40%: p = 0.0407) ( Figure 1B). There were no significant changes in CD4, FoxP3-positive, and CD4/8 ratio (p = 0.5998, p = 0.5471, p = 0.5383, respectively) before and after the therapy.

| Comprehensive evaluation of miRNA expression before and after EGFR-TKI resistance acquirement
We subsequently performed the miRNA PCR array to explore the change of miRNA expression following acquired EGFR-TKI resistance. Fold change of expression levels for each miRNA in PC9/ER (an EGFR-TKI resistant cell line) from those in PC9/6M (an EGFR-TKI sensitive cell line) was evaluated in this study (n = 3). Seven miRNAs, whose expression levels had more than twofold changes, were summarized in Table 2. Among these, miR-1 demonstrated the largest increase (fold change: 8.42, p = 0.00014).

| MiR-1 inhibited cytokine expressions and monocytes migration
We performed a cytokine array analysis using miR-1 transfected PC9 for 24 h, against the control, in order to examine the effects of miR-1 on the change of cytokine expression. Both IL-6 and IL-8 were decreased in the miR-1 transfected cells ( Figure 3A). CCL5, CXCL10, and TNF-α, all of which were well known to be involved in immune cell migration, demonstrated no significant differences. Therefore, we subsequently validated the mRNA expression level of these cytokines (n = 3) and confirmed that those of CCL5 and CXCL10 were significantly inhibited by miR-1 transfection | 723 for 24 h in these three cell lines above, PC9 (p = 0.0394, p = 0.0101, respectively), PC9/ER (p = 0.0015, p = 0.0072, respectively), H1975 (p = 0.0013, p = 0.0096, respectively), and those of IL-6, IL-8 and TNF-α in two cell lines, PC9 (p = 0.0134, p = 0.0117, and p = 0.0047, respectively), H1975 (p = 0.0019, p < 0.001, and p = 0.0015, respectively) ( Figure 3B). IL-6, IL-8 and TNF-α were significantly decreased (48 h: p = 0.0008, p = 0.0195, p = 0.0006, 72 h: p = 0.0004, p = 0.0015, p = 0.0008, respectively) when the transfection time was extended to 48 and 72 h in PC9/ER cell lines ( Figure 3C). miRNAs have been reported to reduce the expression of cytokines involved in monocyte responses, including lymphocyte migration. 36 Therefore, in this study, we performed a monocyte migration test in order to determine whether miR-1 could inhibit the migration in human PBMC (n = 3). Representative findings are illustrated in Figure 3D. Monocyte migration was significantly inhibited in the miR-1 transfected group (p = 0.0018) ( Figure 3E). The migrated mononuclear cells were also confirmed under a light microscope and these cells were mainly composed of lymphocytes or macrophages.

| DISCUSSION
In this study, we first examined the changes in the TIME before and after developing EGFR-TKI resistance in human lung cancer tissues. Our results demonstrated that miR-1 induction altered the sensitivity of EGFR-TKI and the features of the TIME by inhibiting migration of monocytes, including CD8-positive T cells. These results also provided new insights into the prediction of ICI efficacy and therapeutic strategy for lung cancer patients with EGFR-TKI resistance.
In our present study, we focused on tumor-infiltrating T lymphocytes as a factor influencing the TIME. This is because PD-L1 has been widely used as a marker to predict the therapeutic efficacy of ICI, but the effects of ICI are known to be weak when tumor-infiltrating T lymphocytes are relatively scarce, despite high levels of PD-L1 in carcinoma cells. 37,38 In addition, in the cases whose levels of CD8-positive T cell infiltration were high prior to EGFR-TKI therapy, their levels decreased after acquiring EGFR-TKI resistance. Although the number of cases was limited to only 21, the results of our immunohistochemical analysis was consistent with a previous study, in which 138 NSCLCs demonstrated that the status of CD8-and Foxp3-positive T cells significantly decreased after developing EGFR-TKI resistance. 26 However, the mechanisms of the association between these changes of the TIME and EGFR-TKI resistance has remained unknown. Therefore, we performed in vitro study to explore the rationale for reduction of lymphocytes infiltration after acquiring EGFR-TKI resistance.
We first revealed an association between miR-1 expression, TIME, and EGFR-TKI resistance in lung adenocarcinoma patients. In particular, we focused on miRNAs because they are known to exert effects on various biological activities of cancer, such as progression and metastasis, and have recently been proposed as a useful marker to detect cancer status. [39][40][41] We found that miR-1 expression increased after acquiring EGFR-TKI resistance in carcinoma cells although only one cell line model, revealed by comprehensive miRNA analyses. MiR-1 has been found to play pivotal roles in muscle formation and muscle growth in skeletal and myocardial myocytes. 42 It was subsequently reported to be involved in metastasis and suppression of progression in various human malignancies. [43][44][45] On the other hand, there have been no reports regarding the interrelationship between miR-1 and EGFR-TKI resistance, to the best of our knowledge. Our results firstly clarified the significantly inhibitory effect In addition, it is also true that miR-1 expression was reported to be influenced by cytokines or other miRNAs involved in lymphocytes migration. [46][47][48] However, it has been unknown whether miR-1 regulates status of cytokines or lymphocyte migration or not. Results of our current comprehensive study of cytokines did reveal that the transfection of miR-1 into three different human lung carcinoma cell lines harboring EGFR mutations significantly decreased the expression of CCL5, CXCL10, IL-6, IL-8, and TNF-α. Among those cytokines, both CCL5 and CXCL10 are reported to be involved in the migration of CD8-positive T cells and to induce migration of intratumoral CD8-positive T cells in various tumors, including NSCLC. 49,50 In addition, both IL-6 and IL-8 were previously reported to be involved in the differentiation of CD8-positive T cells and the migration of T cells. 51,52 Results of our present study reveal that miR-1 transfection into human lung carcinoma cell lines harboring EGFR mutations does inhibit the migration of human monocytes. Therefore, we are able to demonstrate that miR-1 reduced the expression of cytokines, resulting in an inhibition of monocytes, including lymphocyte migration. Unexpectedly, no significant reduction in cytokines was observed in PC9/ER compared with PC9/6M in this study (data not shown). However, as is well known, many factors have been known to regulate the expression of cytokines. Therefore, we thought that these unexpected results are because other factors affecting cytokines may have outweighed the effect of miR-1 in PC9/ER. For further investigation to support our hypothesis that miR-1 might suppress the TIME following development of EGFR-TKI resistance, examination using multiple EGFR-TKI resistant strains is required in the future.
In summary, the results of our present study indicate that intra-tumoral CD8-positive T cells are decreased in lung adenocarcinoma patients acquiring EGFR-TKI resistance when levels of miR-1 are raised. In addition, patients with high expression of miR-1 may be less likely to benefit from subsequent ICI therapy. Therefore, an evaluation of serum miR-1, before treatment and after acquiring EGFR-TKI resistance, could contribute to predicting therapeutic efficacy of subsequent ICI therapy for the patient. In addition, chemotherapy rather than ICI treatment may be recommended for these patients, although further evaluation is required.