Targeting focal adhesion kinase boosts immune response in KRAS/LKB1 co-mutated lung adenocarcinoma via remodeling the tumor microenvironment

The OAK and POPLAR trials were available as supplementary information in Gandara DR et al. The MSKCC studies are available at www.​cbioportal.​org. The patient survival data and genetic aberrations were extracted for subgroup analysis.

Mouse KL cell line was a gift of Ji et al. Somatic KRAS activation and Lkb1 loss were induced in the lungs of genetically engineered male mice (Kras^G12D/+/Lkb1^L/L with the C57BL6/J genetic background) by nasal inhalation of Adeno-Cre (2 × 10⁶ PFU). After 12 weeks Adeno-Cre treatment, tumors from mice of defined genotypes were dissected and cut into small pieces and cultured in RPMI-1640 + 10% FBS + 1% penicillin–streptomycin. The medium was changed every other day until cells outgrew and stable KL cell lines formed. Mouse KP cell line was provided by InxMed. A549 and H1299 cell lines were purchased from American Type Culture Collection (ATCC). L929 cell line was kindly provided by Stem Cell Bank, Chinese Academy of Sciences. Cells were maintained in DMEM or RPMI1640 (HyClone) supplemented with 10% fetal bovine serum (FBS) (Gibco) and 1% penicillin–streptomycin (Gibco). The small molecule FAK inhibitor IN10018 was provided by InxMed. Anti-PD-1 mAb (#BE0146, clone: RMP1-14), IgG control (#BE0089, clone: 2A3), InVivoPure™ pH 7.0 Dilution Buffer (#IP0070) and InVivoPure™ pH 6.5 Dilution Buffer (#IP0065) were purchase from BioXcell (West Lebanon, New Hampshire, USA).

Human-STK11 and mouse-STK11 were generated by PCR, cloned into pHBLV-vector and transfected into cells using lipofectamine3000. siRNAs were commercially synthesized (Ribo-Bio Co) and transfected into cells using riboFECT CP Transfection Kit (Ribo-Bio Co). siRNA sequences targeting human genes were as below: siSTK11-1, TGAGGAGGTTACGGCACAA.

Total RNAs were isolated from cells using RNAiso Plus (TAKARA) and reverse-transcribed to cDNA using RevertAid First Strand cDNA Synthesis Kit (Thermo Scientific). Diluted cDNA was subject to qRT-PCR analysis using TB Green™ Premix Ex Taq™ II (TAKARA). Relative expression levels were calculated by the comparative Ct approach. qRT-PCR primers were listed as follows: m-FAP-Forward, 5′-GTCACCTGATCGGCAATTTGT-3′ and m-FAP-Reverse, 5′- CCCCATTCTGAAGGTCGTAGAT-3′; m-actin-Forward, 5′- TCCTGTGGCATCCACGAAACTACA-3′ and m-actin-Reverse, 5′- ACCAGACAGCACTGTGTTGGCATA-3′.

Protein samples were extracted from cells using RIPA lysis buffer (Beyotime). The quantification of protein level was measured with BCA kit (Beyotime). The samples were boiled in 5X loading buffer, separated by SDS-PAGE, transferred to PVDF membranes (Millipore). After blocking, the membrane was incubated overnight at 4 °C with primary antibodies as followings: LKB1 (#3047, Cell Signaling Technology; 1:1000 dilution), FAK (#610,087, BD Pharmingen; 1:1000 dilution), p-FAK (#81,298, Abcam; 1:1000 dilution), GAPDH (#5174, Cell Signaling Technology; 1:1000 dilution). The membrane was then washed with TBST and incubated with secondary antibody for 1 h at room temperature in the dark. The targeted proteins were detected by ECL reagent (Tanon) and the exposure was performed by ChemiDoc XRS (Bio-Rad).

KL cells were split into 96-well plate (3 × 10³ cells, 100 μL per well, six replicates). The culture medium was changed into medium with different dose of FAK inhibitor on the other day. After incubating for 24 h, 48 h and 72 h, 10 μL of the CCK-8 (Dojindo) was added to each well of the plate and continued to incubate for 2 h. Absorbance was measured at a wavelength of 450 nm with microplate reader (Bio-Tek).

The rates of cell apoptosis were measured by an Annexin V-FITC/PI Apoptosis Detection Kit (Yeasen biotech Co., Ltd.). Briefly, cells were harvested and washed with cold PBS twice. After being resuspended in 100 μL of binding buffer, cells were incubated with Annexin V and PI for 15 min in the dark. Then the apoptosis data was acquired on Beckman Cytoflex (Beckman Coulter). The results were analyzed with FlowJo 10.6.1 software (TreeStar).

The mouse tumor tissue was harvested for total RNA extraction with TRIzol (Invitrogen). RNA integrity was assessed using the RNA Nano 6000 Assay Kit of the Bioanalyzer 2100 system (Agilent Technologies). A total amount of 1 μg RNA per sample was used as input material for the RNA sample preparations. The cDNA libraries were generated with NEBNext Ultra RNA Library Prep Kit for Illumina (New England Biolabs Inc). The clustering of the index-coded samples was performed on a cBot Cluster Generation System using TruSeq PE Cluster Kit v3-cBot-HS (Illumia) according to the manufacturer’s instructions. After cluster generation, the library preparations were sequenced on an Illumina Novaseq platform and 150 bp paired-end reads were generated. Gene expression levels were quantified as normalized FPKM (fragments per kilobase of exon per million mapped fragments). GSEA was performed using local version of the GSEA analysis tool (http://​www.​broadinstitute.​org/​gsea/​index.​jsp). KEGG pathways and GO enrichment analysis of differentially expressed genes were implemented by clusterProfiler R package. The Reactome database brings together the various reactions and biological pathways of human model species. The corrected P value less than 0.05 were considered significantly enriched by differential expressed genes in Reactome database. Heatmaps were produced by Novomagic (https://​magic.​novogene.​com). FPKM RNA sequencing data was uploaded to GEO (GSE244452).

A total of 2 × 10⁶ KP or KL cells were suspended in 200 μL phosphate-buffered saline (PBS) and then injected subcutaneously into 6 week old male C57BL/6 mice in the right flank. The tumor volumes were measured by caliper and calculated with a formula of π/6 × long diameter × short diameter × short diameter. Mice were sacrificed when tumor volume reached 700–800 mm³. Tumors from KP or KL cell-bearing mice were excised for further analysis.

To explore the efficacy of PD-1 inhibitor in KP and KL mouse model, 2 × 10⁶ KP or KL cells were suspended in 200 μL PBS and were injected subcutaneously in the right flank of syngeneic recipient 6–8 week old male mice. Mice bearing tumors between 150 and 200 mm³ were randomly assigned to intraperitoneal treatment with six doses of 200 μg anti-PD-1 (clone RMPI-14; BioXCell) or isotype control antibody (clone 2A3; BioXCell) administered every 3 days (N = 5–8 mice per group). The body weights and tumor volumes were monitored and recorded twice a week.

To explore the impact of FAK inhibitor and combination therapy in KL mouse models, 2 × 10⁶ KL cells were suspended in 200 μL PBS and were injected subcutaneously in the right flank of syngeneic recipient male mice. FAK inhibitor treatment or vehicle control reagent (0.5% Natrosol 250 HX dissolved in distilled water) was initiated at the first day when tumor volume reached at indicated size and administered at 25 mg/kg by oral gavage every day. Anti-PD-1 (clone RMPI-14; BioXCell) or isotype control antibody (clone 2A3; BioXCell) were administered at 200μg/mouse in total of six doses intraperitoneally. Tumor volume was measured with a formula of 0.5 × long diameter × short diameter × short diameter. At indicated day, tumor tissues or spleen were harvested for analysis.

All the animal studies were approved by the Institutional Committee for Animal Care and Use, Tongji University School of Medicine, and were performed in accordance with the institutional guidelines.

4 μm paraffin-embedded tissue were deparaffinized and rehydrated, stained with hematoxyline and eosin (H&E), IHC, IF, Masson's Trichrome Kit (SenBeiJia Biological Technology Co., Ltd.) and Picro-Sirius Red (SenBeiJia Biological Technology Co., Ltd.) following manufacturer protocol. For IHC, slides were incubated overnight in a humidified chamber at 4 °C with primary antibodies as followings: LKB1 (#13,031, Cell Signaling Technology; 1:250 dilution), Collagen I (#BA0325, Boster; 1:250 dilution), Collagen III (#ab7778, Abcam; 1:200 dilution), CD8 (#98,941, Cell Signaling Technology; 1:400 dilution), HIF-1α (#ab114977, Abcam; 1:250 dilution), p-FAK (Tyr397) (#44624G, Invitrogen; 1:200 dilution). The whole slides were scanned with Pannoramic 250 (3D HISTECH). The intensities of brown-colored precipitate stained with DAB were measured and quantified with IOD or cell intensity by Image Pro Plus 6 (Media Cybernetics). For IF, the FAP/α-SMA was used for activated CAF analysis (FAP: #ab53066, Abcam, 1:200 dilution; α-SMA: #ab7817, Abcam, 1:1000 dilution). CD31 (#77,699, Cell Signaling Technology; 1:100 dilution) was used for vascular analysis. The whole slides were scanned with Pannoramic MIDI (3D HISTECH).

The single-cell suspensions of tumor tissue were generated with tumor dissociation kit, mouse (Miltenyi Biotec) according to manufacturer’s instructions and mouse spleen were mechanically dissociated. After re-suspended in FACS buffer (PBS, 2%FBS), the cells were blocked with anti-mouse CD16/CD32 (eBioscience) 10 min prior to staining with surface antibodies at 4 °C. Surface markers were then added along with 10 μL Brilliant Stain Buffer (BD Bioscience) in a final staining volume to 100 μL. For intracellular staining, cells were fixed and permeabilized with Transcription Factor Buffer Set (BD Bioscience) followed by staining with intracellular antibodies. Fluorochrome-conjugated antibodies for flow cytometry analysis were listed in Additional file 1: Table S1. The flow cytometry data was acquired on Beckman Cytoflex (Beckman Coulter) and the results were analyzed with FlowJo 10.6.1 software (TreeStar).

The intratumoral CCL2, CCL5, CCL7, CXCL1, CXCL2 and VEGF were measured using Luminex liquid suspension chip. Mouse Premixed Multi-Analyte Kit (#LXSAMSM-06, R&D systems) was used in accordance with manufacturer’s instructions. Briefly, the lysate of tumor tissue was embedded with microbeads for 2 h, and then incubated with detection antibody for 1 h. Subsequently, streptavidin-PE was added into each well for 30 min, and values were read using the Bio-Plex MAGPIX System (Bio-Rad). The intratumoral CXCL9 (#MCX900, R&D systems) and CXCL5 (#ab100719, Abcam) were analyzed with commercial ELISA Kit according to the manufacturer's instructions.

Results were presented as mean ± SEM. The log-rank test was used for comparison of survival outcomes downloaded from OAK, POPLAR and MSKCC with Kaplan–Meier method. Unpaired Student’s t-test was used to compare the statistical significance between tested and control groups. For comparison of tumor growth between two groups in vivo, two-way analysis of variance (ANOVA) with Tukey’s post hoc test was performed (control versus FAK inhibitor group in Fig. 6A; control versus anti-PD1 and FAK inhibitor versus FAK inhibitor + anti-PD1 group in Fig. 7C) [13].Survival time of mice was defined from the day of tumor cell inoculation until the tumor volume reached 1000 mm³.Statistical analyses were performed using GraphPad Prism 8 (GraphPad Software). Statistical significance was determined as indicated in the figure legends. Significance was set to P < 0.05 and represented as *P < 0.05, **P < 0.01, ***P < 0.001.

Firstly, we investigated the impact of KRAS/LKB1 co-mutations on survival outcomes in PD-1/PD-L1 inhibitor-treated lung adenocarcinoma patients from POPLAR and OAK trials and confirmed that KL subgroup had significantly shorter progression-free survival (PFS) compared with other KRAS-mutated patients (median PFS:1.3 vs. 4.1 months, p = 0.0025). Integrating the data from MSKCC cohort, we further demonstrated that overall survival (OS) also tended to be shorter in KL group (median OS: 9.0 vs. 15.0 months, p = 0.065) (Fig. 1A).

Previous studies reported that among all subtypes of KRAS-driven lung cancers, tumors with frequently inactivated TP53 were most sensitive to PD-1 blockade [4]. Therefore, to explore the underlying mechanisms for the distinct therapeutic response, we used KRAS^G12DLKB1^−/− (hereafter referred to “KL”, previously established from a spontaneously arising lung adenocarcinoma in the Kras^G12DLkb1^fl/fl mouse model, (Additional file 3: Fig. S1A) and KRAS^G12DTP53^−/− (hereafter referred to “KP”) murine lung adenocarcinoma cell lines [14, 15] for further analyses. Isogenic cell lines upon confirmation of STK11 knockout (by immunoblotting for the LKB1 protein) (Additional file 3: Fig. S1B) were implanted into the right flank of syngeneic recipient mice and cohorts of tumor-bearing mice were randomized to be treated with anti-PD-1 antibody or IgG control (Fig. 1B). The expression of LKB1 by IHC in KP or KL tumors further confirmed that STK11 was successfully knocked out (Fig. 1C and D). Tumor growth curve suggested that anti-PD-1 antibody significantly attenuated tumor growth in subcutaneous KP mouse model (Fig. 1C) but showed no efficacy in KL mouse model (Fig. 1D). No significant differences in body weight were found between anti-PD-1 antibody and control group in KP or KL mouse model. These results suggested that KL tumors were resistant to anti-PD1 antibody and our KL mouse model was reliable for subsequent research.

We next sought to elucidate how LKB1-deficiency impacted the TME. Firstly, comprehensive immune cell populations between KP and KL tumors via flow cytometry analysis were compared. In general, the infiltrated quantity of T cells (CD3) and B cells (CD19) were significantly decreased in KL tumors as compared with KP tumors (Fig. 2A and B) while no significant difference was observed regarding NK cells (CD3⁻NK1.1⁺) (Fig. 2C). Moreover, KL tumors had significantly fewer CD8⁺T cells, CD4⁺T cells and Foxp3⁺Tregs infiltration (Fig. 2D–F). In addition, DCs (CD45⁺CD11c⁺MHC-II⁺) and tumor-associated macrophage (TAMs)[16], regardless of M1-like macrophage (CD11b⁺F4/80⁺CD86⁺) or M2-like macrophage (CD11b⁺F4/80⁺CD206⁺) were also less infiltrated in KL tumors than KP tumors (Fig. 2G and I). Tumor-associated neutrophils (TANs) and polymorphonuclear myeloid-derived suppressor cells (PMN-MDSCs, CD45⁺CD11b⁺Ly6G⁺Ly6C^low), which suppressed T cells function to help form immunosuppressive TME [17], were significantly highly infiltrated in KL tumors (Fig. 2H and J). Taken together, our data demonstrated that KL tumors revealed immunosuppressive TME, with fewer immune-active cells infiltration but with more immunosuppressive cells infiltration, including PMN-MDSC and TANs.

To uncover the mechanisms that contributed to the formation of immunosuppressive TME of KL tumors, RNA-seq-based transcriptome profiling of KL and KP subcutaneous tumors was analyzed. Briefly, 3497 genes were up-regulated and 3952 genes were down-regulated in KL tumors compared with KP tumors. Furthermore, GO, Reactome and gene set enrichment analysis (GSEA) were performed and showed that genes encoding for ECM-related components were significantly enriched (Fig. 3A–C). The ECM provided a physical scaffold for all cells in the TME and is the major component. Collagen has been implicated in cancer progression, metastasis, and drug resistance. As expected, the transcriptomic data showed that collagen-formation related genes were mostly upregulated in KL tumors. Among top 20 collagen-related genes, COL3A1 and COL1A1 were significantly upregulated in KL tumors (Fig. 3D, Additional file 2: Table S2). Consistently, Sirius Red and Masson staining confirmed the increased intra-tumoral collagen deposition and specifically, collagen I and III, which are encoded by COL1A1 and COL3A1, in KL tumors (Fig. 3E ). To elucidate the impact of collagen deposition on the distribution of immune cells, we further analyzed the spatial correlation between collagen deposition and CD8⁺ T cells infiltration. The results showed that abundant CD8⁺ T cells were present at sites of collagen deposition (Fig. 3F). Therefore, we analyzed the distribution of CD8⁺ T cells in two regions (tumor nest and stroma, including tumor margin and intratumor stroma). In tumor nest, CD8⁺ T cells were significantly less infiltrated in KL tumors, however, at the tumor margin where collagen deposited, CD8⁺ T cells infiltration were much higher in KL tumors (Fig. 3G).

The above results revealed that (i) compared with KP tumors, KL tumor had excessive ECM deposition; (ii) excessive collagen deposition blocked CD8⁺ T cells infiltrating into tumor nest to exert killing function; and (iii) CD8⁺ T cells could barely infiltrate into tumor nest owing to the collagen deposition and were accumulated in the margin of KL tumors.

CAFs are the predominant stromal cells in the TME [18]. Activated CAFs, which highly express α-SMA and fibroblast activating protein (FAP), usually accumulate at the edge of tumor nest, secret collagen fibers to surround the malignant cells with dense desmoplastic stroma, hence facilitate drug resistance via blocking anti-tumor drugs and immune cells into tumor nest [19, 20]. Therefore, we questioned whether CAFs were significantly activated in KL tumors. Firstly, using TIMER2.0 database (http://​timer.​cistrome.​org/​), we observed a positive correlation between the expression of dominant genes encoding collagen fibers and markers for activated CAFs (Fig. 4A and B). Consistently, our RNA-seq data also showed the more CAFs were infiltrated in KL tumors (Fig. 4C). To investigate whether KL tumor cells directly impacted the activation of fibroblasts, KL and KP tumor cells were co-cultured with mouse fibroblast (L929) in vitro. The results showed that co-culture with KL tumor cells significantly increased expression of FAP protein, suggesting that KL tumor cells could activate CAFs (Fig. 4D). Immunofluorescence further confirmed that activated CAFs were highly infiltrated in KL tumors (Fig. 4E). Next, we performed KEGG pathway analysis of dissected KL and KP tumors to explore the mechanisms of abundant activated CAFs. Interestingly, focal adhesions-related genes were enriched in KL tumors (Fig. 4F). Focal adhesions are integrated points of cellular membrane which link to extracellular matrix via integrin heterodimers [21, 22]. FAK is recruited to the sites of focal adhesions and get auto-phosphorylated upon ECM-induced integrin receptor activation [23]. Previous studies demonstrated that FAK activation was closely associated with stromal fibrosis and immunosuppressive TME [24, 25]. Further co-localization analysis in tumor specimens demonstrated that FAK was hyperactivated in CAFs (Fig. 4G). Given the fact that FAK activation within pancreatic cancer CAFs was associated with ECM synthesis and deposition [26], we studied whether fibroblastic FAK hyperactivation could lead to CAFs activation and collagen deposition. KL tumors-bearing mouse were treated with FAK inhibitor (IN10018, InxMed, 25 mg/kg) or natrosol as control and tumor components were assessed at different time points (Day 3, Day 7, Day 14) (Fig. 5A). As indicated, FAK inhibitor effectively inhibited FAK phosphorylation (Fig. 5B). Further GO analysis revealed that FAK inhibitor had significant impact on ECM-related genes (FAK inhibitor vs control at Day 14, FAK inhibitor at Day14 vs FAK inhibitor at Day 7) (Fig. 5C). Immunofluorescence and IHC results also demonstrated FAK inhibitor could significantly decrease fibrosis and CAFs activation (Fig. 5D-G).

Angiogenesis is one of the hallmarks of cancer and blood vessels are also the important components of the TME [11, 27]. Notably, previous evidence showed that FAK pathway was associated with angiogenesis and tumor growth [28‐30]. Therefore, we questioned whether the use of FAK inhibitor has an impact on TME via inhibiting tumor angiogenesis in KL tumors. Indeed, GO analysis and immunofluorescence staining revealed that angiogenesis-associated genes were highly enriched in KL tumors (Additional file 4: Fig. S2A, B). Although FAK inhibitor slightly alleviated hypoxia at Day 7, FAK inhibitor had no significant impact on angiogenesis (Additional file 4: Fig. S2C, D).

In summary, these findings suggested that (i) abundant activated CAFs resulted in excessive collagen deposition in KL tumors; and (ii) fibroblastic FAK activity contributed CAFs activation and hence, induced collagen formation.

Therapeutic benefits of pharmacological FAK inhibition in previous preclinical models have been confirmed to directly inhibit FAK expressed by neoplastic cells and thus, reduced tumor growth, invasion and metastasis [31, 32]. To illustrate the cytotoxic effects of FAK inhibition on tumor cells, we utilized human NSCLC cell lines with LKB1-mutant (A549), LKB1-WT (H1299), mouse lung cancer cell line with LKB1-wild type (KP) and mouse lung cancer cell line with LKB1-loss (KL). However, knockdown of LKB1 in LKB1-wild type cell lines and upregulation of LKB1 in LKB1-mutant or LKB1-loss cell lines did not significantly influence the activation of FAK pathway (Additional file 5: Fig. S3A). Additionally, FAK inhibitor had minimal impact on cell viability and no obvious effects on apoptosis of KL cells in vitro (Additional file 5: Fig. S3B, C). These results provided the evidence that FAK inhibitor did not directly exert cytotoxic effects on lung tumor cells.

Next, we investigated the anti-tumor effect of FAK inhibition on tumor growth in KL mouse model and found that tumor growth was significantly attenuated (Fig. 6A). Therefore, we speculated FAK inhibitor might exert anti-tumor effect via remodeling the TME. Interestingly, FAK inhibitor significantly promoted proliferation of entire immune cells (CD45⁺Ki67⁺, Additional file 6: Fig. S4A). Moreover, FAK inhibitor significantly increased the proportions of T cells (on Day 7 and Day 14), B cells (on Day 3 and Day 7) and NK cells (on Day 3 and Day 14) (Fig. 6B–D). Specifically, a simultaneous increase in CD8⁺ T cells was observed with FAK inhibition on Day 3, Day 7 and Day 14 (Fig. 6E). Interestingly, on Day 14, more Foxp3⁺Tregs were also infiltrated in FAK inhibitor groups (Fig. 6F). Regarding myeloid cells, FAK inhibition facilitated more DCs infiltration on Day 7 and Day 14 (Fig. 6H). The quantity of M-MDSC were temporarily elevated with the treatment of FAK inhibitor on day 3 and day 7. In contrary, less PMN-MDSCs were infiltrated with FAK inhibition on day 14 (Fig. 6G). Although the total quantity of macrophages in two groups were not significantly altered, FAK inhibitor changed the proportions of macrophage phenotype, with more infiltration of M1-like macrophage and less infiltration of M2-like macrophage (Fig. 6I). The dynamic changes of immune cells in spleen were generally consistent with tumors upon FAK inhibition (Additional file 7: Fig. S5).

We also examined the dynamic change levels of related cytokines in tumor tissues. Firstly, RNA-seq demonstrated the corresponding changed immune cell-related cytokines, including CXCL9, CCL5, CCL2, VEGFA, CXCL5, CXCL1, CXCL2, CCL7 et al., were also significantly altered (Additional file 8: Fig. S6A). Specifically, CXCL1 was significantly reduced on Day 3 and Day 7 and CXCL2 was reduced on Day 7 with FAK inhibition, Meanwhile, CXCL9 was upregulated on Day 7 (Additional file 8: Fig. S6B).

The above findings suggested that FAK inhibitor effectively suppressed tumor growth via boost immune response, converting “immune-cold” tumors into “immune-hot” tumors. Upon FAK inhibition, more lymphocytes were infiltrated, especially CD8⁺ T cells. Meanwhile, several immunosuppressive cells, including PMN-MDSC and M2-like macrophage, were less infiltrated. Interestingly, FAK inhibitor could also upregulate PD-L1 expression on non-immune cells (CD45⁻PD-L1⁺, Additional file 6: Fig. S4B), implying a potential synergetic anti-tumor effect with anti-PD-1 antibody.

Based on our previous findings, we further examined the anti-tumor effect of combining FAK inhibition and PD-1 blockade in KL mouse models. Tumor-bearing KL models and the following schedules were shown in Fig. 7A. PD-1 blockade had no impact on FAK phosphorylation (Fig. 7B). As expected, FAK inhibitor monotherapy or plus PD-1 blockade could both attenuate tumor growth compared with either PD-1 blockade group or control. Importantly, as compared with FAK inhibitor alone, the combination therapy resulted in a synergetic anti-tumor effect (Fig. 7C). Furthermore, the combination of FAK inhibitor and PD-1 blockade also led to prolonged time to reach endpoint (the time when tumor volume reached 1000 mm³) (Fig. 7D). Notably, on Day 30, histological analyses of lung tissues from subcutaneous mouse model showed that tumor cells had metastasized to lung, suggesting KL tumor cells were aggressive and invasive (Fig. 7E). Interestingly, FAK inhibitor with and without PD-1 blockade could also effectively decrease the metastatic area in lung tissue and inhibit lung metastases compared with control of anti-PD1 monotherapy group (Fig. 7F). As expected, FAK inhibitor and the combination therapy also significantly decreased collagen deposition and inhibited CAFs activation (Fig. 8A–C). Moreover, vessel density was similar among the four groups (Additional file 9: Fig. S7).

We then investigated the impact of different treatment regimens on the TME. The results showed that the proportions of T cells, specifically CD8⁺T cells were significantly increased in the combination group (Additional file 10: Fig. S8A and Fig. 8D). Additionally, CD8/Treg ratio was numerically elevated in the combination group compared with FAK inhibitor monotherapy group (p = 0.08) (Fig. 8D). Further IHC staining showed that FAK inhibitor in combination with PD-1 blockade could increase CD8⁺ T cells infiltration in both subcutaneous and lung metastatic tumor tissues (Fig. 8E and F). The proportions of B cells, NK cells and CD8⁺GZMB⁺ T cells were also significantly increased with FAK inhibition while there was no significant difference between FAK inhibitor monotherapy and the combination therapy (Additional file 10: Fig. S8B–D). In terms of Tregs, the combination therapy numerically reduced Tregs infiltration compared with FAK inhibitor monotherapy group (Additional file 10: Fig. S8E). Regarding myeloid cells, the combination group or FAK inhibitor significantly promoted DCs infiltration and reduce PMN-MDSC and macrophages infiltration (Additional file 10: Fig. S8F–H).

Collectively, these results demonstrated FAK inhibitor could overcome primary resistance of KL tumors to PD-1 blockade.