Phosphatidylcholines (PCs) serve as biomarkers for prognosis of esophageal squamous cell carcinoma (ESCC) and compromise antitumor efficacy of focal adhesion kinase (FAK) inhibitor

doi:10.21203/rs.3.rs-1556187/v1

PDF

Research Article

Phosphatidylcholines (PCs) serve as biomarkers for prognosis of esophageal squamous cell carcinoma (ESCC) and compromise antitumor efficacy of focal adhesion kinase (FAK) inhibitor

https://doi.org/10.21203/rs.3.rs-1556187/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted 19 Apr, 2022

You are reading this latest preprint version

Abnormal metabolism is recognized as an oncogenic hallmark that is correlated with tumor progression and therapeutic resistance. Here, we used multi-omics, including phosphoproteomics, untargeted metabolomics and lipidomics to demonstrate that the cancer-associated fibroblasts (CAFs)-derived AKT2/choline-phosphate cytidylytransferase α (CCTα)/phosphatidylcholines (PCs) axis mediated the resistance of focal adhesion kinase (FAK) inhibitor-defactinib in esophageal squamous cell carcinoma (ESCC) treatment. We first identified extremely low levels of FAK Tyr³⁹⁷ expression in CAFs, which might possibly provide no available target for defactinib exerting its anti-growth effect in CAFs. Consequently, defactinib upregulated the intracellular concentration of Ca²⁺ in CAFs, and facilitated the formation of AKT2/CCTα protein complex to phosphorylate CCTα Ser^315/319/323 sites, and finally enhanced the production of stromal PCs, which activated intratumoral Janus kinase 2 (JAK2)/Signal transducer and activator of transcription 3 (STAT3) pathway to induce resistance of FAK inhibition. Analysis of clinical samples showed that stromal pAKT2 Ser¹²⁸ and pCCTα Ser^315/319/323 were associated with the tumor malignancy and reduced patient survival. Pseuo-targeted lipidomics and further validation cohort quantitatively demonstrated that plasma PCs allow to distinguish the malignant extent of ESCC patients. Overall, inhibition of stroma-derived PCs and related pathway might be used as potential therapeutic strategies for tumor treatment.

The prognosis for patients with esophageal squamous cell carcinoma (ESCC) is dismal, with the 5-year survival rate less than 15% [1, 2, 3]. This poor survival rate is driven by the lack of therapeutic efficacy from cytotoxic, targeted and immune-based therapeutics [4, 5]. The critical mechanism of therapeutic resistances is the ESCC cells surrounding tumor microenvironment (TME), especially its major component-cancer-associated fibroblasts (CAFs) [6, 7, 8]. The crosstalk between ESCC and their surrounding CAFs has important effect on the biological behavior of tumor cells. CAFs tightly interact with tumor cells via cell-cell contact, cytokine release and exosomal transmission [9, 10, 11]. However, CAFs-derived metabolites, the important signaling mediators, have ramifications for the function of tumor cells [12, 13, 14]. To manage the metabolic challenges imposed by the TME, tumor cells and CAFs cooperatively interact to support tumor malignancy. It is not clear what are the metabolic profiles of CAFs and how CAFs-derived metabolites act on tumor malignancy and the response of tumor cells toward therapeutic agents.

Focal adhesion kinase (FAK) is a cytoplasmic non-receptor protein tyrosine kinase and is ubiquitously expressed [15, 16]. Many reports indicate that FAK overexpression in several types of solid tumors is associated with tumor malignancy and functions as the nexus to transit the TME-derived signaling into tumor cells [16, 17, 18, 19, 20]. These findings have encouraged the development of FAK inhibitors for tumor treatment. However, even though some FAK inhibitors have obtained excellent antitumor effect in preclinical studies only using tumor cell lines, the clinical effect of FAK inhibitors is still controversial [15, 21]. We hypothesize that this discrepancy is at least in part induced by CAFs, which secrete some substances to facilitate the dysregulation of intratumoral signaling pathways, and resultantly impair the antitumor efficacy of chemotherapies.

While several studies have investigated cytokines or chemokines secreted by tumor cells or CAFs mediate the crosstalk between these two types of cells. A thorough understanding of TME-derived metabolites regulating tumor and CAFs communications in therapeutic resistance of tumor cells is still underexplored. In the present study, we aim to investigate whether CAFs-derived metabolites can be used as biomarkers to determine tumor malignancy and how these metabolites lead to alter the antitumor effect of FAK inhibitor through regulating the intercellular signaling crosstalk between tumor cell and CAFs.

1. CAFs impair the antitumor effect of FAK inhibitor in ESCC treatment

We first evaluated whether CAFs can affect the tumor inhibitory effect of FAK inhibitor-defactinib using ESCC cell lines/CAFs coculture system in transwell apparatus with 0.4 μm pore size, and then tumor cells (in the lower chamber of transwell plates) were subjected to MTS assay (Figure 1A). The IC₅₀ values of defactinib (0-10 μM) in KYSE410 and KYSE510 cells alone were 3.87 ± 0.06 and 4.33 ± 0.3 μM. However, in ESCC cells/CAFs coculture system, the IC₅₀ values of defactinib in these two ESCC cell lines were substantially upregulated (9.39 ± 0.37 μM in KYSE410 cells and 10.87 ± 0.42 μM in KYSE510 cells) (Figure 1B). Furthermore, the IC₅₀ value of defactinib (0-10 μM) in CAFs was 119.4 ± 16.78 μM (Figure 1C).

To assess whether CAFs contribute to the resistance of defactinib in ESCC treatment in vivo, KYSE410 or KYSE510 cells and CAFs were coinjected into BALB/c-nu mice. After tumor volume reached to approximately 100 mm³, animals were treated with defactinib. Defactinib effectively decreased the tumor volumes of KYSE410 or KYSE510 cells alone, whereas could not inhibited the tumor growth of these two ESCC tumors in the presence of CAFs (Figure 1D). To comprehensively evaluate the antitumor effect of defactinib in ESCC/CAFs coinjection xenografted model, the expression of Ki-67, CD31, and LYVE-1 in tumor tissues was measured using quantitative ELISA assays. As the results of Figure 1E-1G shown, defactinib could not exert its inhibitory effect on the expression of Ki-67, CD31 and LYVE-1 in KYSE410 and KYSE510 tumors in the presence of CAFs.

We then examined whether CAFs impaired defactinib-mediated inhibition of ESCC cells’ invasion using transwell apparatus with 8 μm pore size (Figure 1H). Defactinib (10 μM) significantly inhibited the invasion of KYSE410 and KYSE510 cells cultured alone (Figure 1I). However, the anti-invasive ability of 10 μM defactinib was weakened in the ESCC/CAFs coculture system (Figure 1I). The CAFs-impaired anti-invasive ability of FAK inhibition was also investigated in vivo using a popliteal lymph node metastasis model. Defactinib could not effectively inhibit the CAFs-facilitated formation of larger lymph nodes of KYSE410 and KYSE510 tumors (Figure 1J).

2. Untargeted metabolomics and lipidomics identify that FAK inhibitor induces the secretion of phosphocholines (PCs) from CAFs

Several studies have demonstrated that tumor microenvironment-derived metabolites can induce the resistance of tumor cells towards targeted agents [22,23,24]. To explore how FAK inhibition might change the metabolic profile of CAFs, we performed an untargeted LC-MS-based metabolomic analysis in CAFs treated with defactinib (10 μM). Several CAFs-derived metabolic pathways, such as protein digestion and absorption (KEGG ID: hsa04974), choline metabolism in cancer (KEGG ID: hsa05231), central carbon metabolism in cancer (KEGG ID: hsa05230), glutathione metabolism (KEGG ID: hsa00480), ABC transporters (KEGG ID: hsa02010), or glycerophospholipid metabolism (KEGG ID: hsa00564), have been upregulated upon defactinib treatment (Figure 2A and 2B). Among these pathways, we have focused on choline and glycerophospholipid metabolisms, due to the upregulation of several glycerophospholipids, such as PC (16:0/20:4), PC (20:5/20:4), PC (14:0/20:2), PC (16:0/20:3), glycerophosphocholine, or PC (22:6/18:1) in defectinib treatment (Figure 2C). The exact information of metabolomics was listed in Supplementary Table 1.

We further investigated whether the secretion of choline-related metabolites from CAFs could been stimulated by FAK inhibition using lipidomics. As shown in Figure 2D and Supplementary Figure 1A and 1B, glycerophospholipid (KEGG ID: hsa00564) and choline (KEGG ID: hsa05231) metabolisms have effectively been increased by defactinib (10 μM) treatment. The exact information of lipidomics was listed in Supplementary Table 2. Specifically, defactinib (10 μM) treatment effectively stimulated the secretion of PCs from CAFs (Figure 2E), whereas not from KYSE410 and KYSE510 cells (Figure 2F).

We chose two PCs-PC (16:0/20:4) and glycerophosphocholine for further functional assays to evaluate whether PCs induce the malignant progression of tumor cells, and found that these two PCs (10 μM) effectively stimulated the growth and invasion of KYSE410 and KYSE510 cells (Figure 2G and 2H). CTP-phosphocholine cytidyltransferase (CCT) enzymes (including CCTα and β) catalyze the key rate-limiting step in choline pathway for phosphatidylcholine (PC) biosynthesis (Glunde et al., 2011). We then knocked down the expression of CCTα and β in CAFs using siRNA, and further observed whether CCTα or β-depleted CAFs can induce the malignancy of ESCC cells. Depletion of CCTα, whereas not CCTβ in CAFs effectively blocked the secretion of PCs from CAFs (Supplementary Figure 2A and 2B). Correspondingly, CCTα siRNAs impaired CAFs-induced the growth and invasion of KYSE410 and KYSE510 cells in ESCC cells/CAFs coculture system (Supplementary Figure 2C and 2D). Furthermore, CCT inhibitor-miltefosine (25 μM) effectively inhibited the CAFs-induced ESCC malignancy (Supplementary Figure 2E and 2F).

3. CAFs-released PCs impair the antitumor effect of defactinib on ESCC cells

We further evaluated CAFs-released PCs induced the resistance of FAK inhibition in ESCC treatment. PC (16:0/20:4) and glycerophosphocholine (10 μM) induced the IC₅₀ value of defactinib (0-10 μM) to 9.14 ± 0.17, or 8.82 ± 0.26 μM in KYSE410 cells (Figure 3A). The IC₅₀ value of defactinib in KYSE410 cells/CAFs CCTα siRNA1/2 cocultured system was 4.58 ± 0.24, or 4.24 ± 0.07 μM, which was evidently lower than that in KYSE410 cells/CAFs control siRNA cocultured system (IC₅₀ value of defactinib was 9.3 ± 0.15 μM) (Figure 3B). However, knockdown of CCTβ in CAFs could not decrease the IC₅₀ value of defactinib in cocultured system (IC₅₀ value of defactinib in CCTβ siRNA1 group was 8.85 ± 0.52 μM, and in CCTβ siRNA2 group was 9.12 ± 0.26 μM) (Figure 3B). CCT inhibitor-miltefosine (25 μM) effectively reduced the IC₅₀ value of defactinib in KYSE410 cells/CAFs coculture system to 4.22 ± 0.15 μM (Figure 3C). Similar results were also obtained in KYSE510 cells (Figure 3A-3C).

PC (16:0/20:4) and glycerophosphocholine (10 μM) blocked the defactinib (10 μM)-mediated anti-invasive effect on KYSE410 and KYSE510 cells (Figure 3D). CCTα siRNAs, but not CCTβ siRNAs, impaired CAFs-mediated the anti-invasive resistance of defactinib (10 μM) in ESCC treatment (Figure 3E). Furthermore, miltefosine (25 μM) enhanced the anti-invasive ability of defactinib in ESCC cells in the presence of CAFs (Figure 3F).

4. Phosphoproteomics identifies that the activation of AKT2/CCTα axis in CAFs contributes to the secretion ofPCs

We further explored the molecular mechanism of the secretion of PCs from CAFs using phosphoproteomics. In an analysis of the proteomes, we identified 613 non-redundant, unique phospho-peptides, representing 1029 phospho-peptides in CAFs treated with defactinib (10 μM). Among these changed phospho-proteins, 565 phosphosites, representing 327 unique proteins, were found to be significantly upregulated in CAFs treated with defactinib (10 μM). Pathway enrichment analysis identified the statistically significant different pathways between control and defactinib (10 μM) treatment, such as VEGF signaling pathway (KEGG ID: hsa04370), Toll-like receptor signaling pathway (KEGG ID: hsa04620), TNF signaling pathway (KEGG ID: hsa04668), Rap1 signaling pathway (KEGG ID: hsa04015), Platinum drug resistance (KEGG ID: hsa01524), PD-L1 expression and PD-1 checkpoint pathway in cancer (KEGG ID: hsa05235), mTOR signaling pathway (KEGG ID: hsa04150), MAPK signaling pathway (KEGG ID: hsa04010), Longevity regulating pathway (KEGG ID: hsa04211), HIF-1 signaling pathway (KEGG ID: hsa04066), FoxO signaling pathway (KEGG ID: hsa04068), Focal adhesion pathway (KEGG ID: hsa04510), ErbB signaling pathway (KEGG ID: hsa04012), EGFR tyrosine kinase inhibitor resistance (KEGG ID: hsa01521), Choline metabolism in cancer (KEGG ID: hsa05231), cGMP-PKG signaling pathway (KEGG ID: hsa04022), Central carbon metabolism in cancer (KEGG ID: hsa05230), and Cellular senescence pathway (KEGG ID: hsa04218) (Figure 4A). The exact molecular information of these pathways was shown in Supplementary Table 3. Interestingly, AKT2 Ser¹²⁸ was appeared in almost all of these identified pathways, and the upregulated CCTα Ser^315/319/323 sites existed in choline metabolism in cancer (KEGG ID: hsa05231) (Figure 4A).

We assessed whether FAK inhibition affected the phosphorylation of AKT2 Ser¹²⁸ and CCTα Ser^315/319/323 sites in CAFs using coculture system (transwell apparatus with 0.4 μm pore size). The upper chamber of transwell apparatus was plated with KYSE410 or KYSE510 cells, and the lower chamber was cultured with CAFs. After 24 hours defactinib treatment, the lysates of CAFs were collected for evaluation of the phosphorylation status of AKT2 Ser¹²⁸ and CCTα Ser^315/319/323 sites. Defactinib (10 μM) effectively stimulated the phosphorylation of AKT2 Ser¹²⁸ and CCTα Ser^315/319/323 sites in CAFs alone or in the presence of KYSE410 or KYSE510 cells (Figure 4B). The phosphorylation extents of AKT2 Ser¹²⁸ and CCTα Ser^315/319/323 sites in defactinib treatment were no significant difference between CAFs alone and CAFs cocultured with KYSE410 or KYSE510 cells groups (Figure 4B). Function-loss AKT2 (S128A) was stably transfected into CAFs, and then the effect of FAK inhibition on the phosphorylation of CCTα Ser^315/319/323 was measured. Defactinib (10 μM) could not stimulate the phosphorylation of CCTα Ser^315/319/323 sites in CAFs harbored AKT2 S128A mutant (Figure 4C and 4D). Correspondingly, AKT2 S128A and CCTα S315/319/323A mutants effectively decreased the production of PCs from CAFs, and defactinib (10 μM) could not induce the secretion of PCs from CAFs stably transfected with AKT2 S128A or CCTα S315/319/323A mutant (Figure 4E).

We examined the physical association between AKT2 and CCTα using immunoprecipitation assays, and found that defactinib (10 μM) facilitated the formation of AKT2/CCTα complex, and the phosphorylation of CCTα Ser^315/319/323 in AKT2/CCTα complex in CAFs alone or in the presence of indicated ESCC cells. The interaction extent of AKT2 and CCTα and the phosphorylation level of CCTα Ser^315/319/323 in AKT2/CCTα complex were no difference between CAFs cultured alone and they cocultured with KYSE410 or KYSE510 cells (Figure 4F).

Interestingly, our phosphoproteomics data showed that defactinib (10 μM) could stimulate the phosphorylation of several Ca²⁺-related proteins in CAFs, such as Ca²⁺-regulated kinases (CARHSP1 S30/S32/S41/S52/T54, CAMK2G T287, CAMK2D T287, CAMKK2 S495, MYLK S299/S305), heat shock proteins (HSPB1 S15/S65/S78/S82/S83, HSPB6 S16/S23, HSP90AA1 S231/S263, HSP90AB1 S226/S452), mitogen-activated protein kinases (MAP2K3 S218, MAP2K4 S394, MAPK1 T185/Y187, MAPK3 T202/Y204, MAPK14 T180/Y182), cytoskeleton (VIM S5/S29/T37/S39/S51/S72/S73/S325, AKAP12 S598/S612/S696/S1331), or Ca²⁺channel receptor (ITPR3 S1832) (Figure 4G, and Supplementary Table 4). Correspondingly, we detected the level of intracellular Ca²⁺ upon defactinib treatment (10 μM) using calcium detection assay. Defactinib (10 μM) effectively induced intracellular Ca²⁺ levels in CAFs (Figure 4H). Importantly, CAFs were pretreated with Ca²⁺ chelator-BAPTA-AM (10 μM), which effectively blocked defactinib-mediated the production of PCs, phosphorylation of AKT2 Ser¹²⁸ and CCTα Ser^315/319/323 sites, the formation of AKT2/CCTα complex and the activation of CCTα in this complex (Figure 4I-4K).

We then analyzed whether defectinib affect the phosphorylation of AKT2 and CCTα in ESCC cells. CAFs were cultured in the upper chamber of transwell apparatus (0.4 μm pore size), and the KYSE410 or KYSE510 cells were respectively cultured in the lower chamber of transwell apparatus. After 24 hours defactinib treatment, the lysates of indicated ESCC cells were collected for assessing the phosphorylation status of AKT2 Ser¹²⁸ and CCTα Ser^315/319/323 sites. Defactinib (10 μM) inhibited the phosphorylation of AKT2 Ser¹²⁸ and CCTα Ser^315/319/323 sites in KYSE410 or KYSE510 cells (Figure 4L). Co-cultured CAFs produced no effect on the phosphorylation of AKT2 Ser¹²⁸ and CCTα Ser^315/319/323 sites in KYSE410 or KYSE510 cells treated with defactinib (Figure 4L).

5. CAFs-derived AKT2/CCTα axis impairs the antitumor effect of defactinib on ESCC cells

The IC₅₀ value of defactinib (0-10 μM) in KYSE410 cells cocultured with CAFs harbored with AKT2 S128A or CCTα S315/319/323A mutant was 4.15 ± 0.42, or 4.51 ± 0.25 μM, respectively (Figure 5A). These IC₅₀ values were significantly lower than KYSE410 cells in the presence of control vector CAFs (IC₅₀ value of defactinib was 9.25 ± 0.46 μM) (Figure 5A). Similar results were also obtained in KYSE510 cells (Figure 5A). Correspondingly, AKT2 S128A or CCTα S315/319/323A mutant effectively blocked CAFs-mediated the anti-invasive resistance of defactinib (10 μM) in ESCC treatment (Figure 5B). The results of in vivo assays, including subcutaneous co-transplantation of ESCC cells and CAFs, or the popliteal lymph node metastasis model, confirmed those results obtained from in vitro assays (Figure 5C-5G). Taken together, these results indicated that CAFs-derived AKT2/CCTα axis critically contributes to the resistance of FAK inhibition in ESCC treatment.

6. CAFs-released PCs activate intratumoral STAT3 to mediate the resistance of defactinib in ESCC treatment

Because intratumoral STAT3 is contributed to the resistance of FAK inhibitor in tumor treatment [25]. We evaluated whether FAK inhibition stimulated the activation of STAT3 in ESCC cells, and found that defactinib (10 μM) significantly upregulated the phosphorylation of STAT3 Tyr⁷⁰⁵ in KYSE410 and KYSE510 cells in the presence of CAFs (Figure 6A). We then determined whether PCs induce the activation of intratumoral STAT3. PC (16:0/20:4) and glycerophosphocholine (10 μM) increased the phosphorylation of STAT3 Tyr⁷⁰⁵ in indicated ESCC cells (Figure 6B).

Furthermore, coinhibition of FAK and JAK2/STAT3 pathway by defactinib (10 μM) combined with JAK2 inhibitors-ruxolitinib, or fedratinib (10 μM), or STAT3 inhibitor-S3I-201 (20 μM) effectively inhibit the growth and invasion of indicated ESCC cells in the presence of CAFs (Figure 6C and 6D). The Results of in vivo xenografted models were similar with those of in vitro assays (Figure 6E, and Supplementary Figure 3A-3D).

7. Stroma-derived AKT2/CCTα axis determines the ESCC progression and the survival of ESCC patients

We further determined the clinical expression of pAKT2 Ser¹²⁸ and pCCTα Ser^315/319/323 in ESCC stroma using immunohistochemistry (IHC) assay, and found that the expression of pAKT2 Ser¹²⁸ (68.5%; 74/108) or pCCTα Ser^315/319/323(71.3%; 77/108) was high in tumor stroma (Figure 7A). The expression of stromal pAKT2 Ser¹²⁸ or pCCTα Ser^315/319/323 was positively correlated with advanced-stage, higher-grade tumor status and lymph node status of ESCC tumors (Figure 7B and 7C), and negatively correlated with the survival time of ESCC patients (Figure 7D and 7E). Critically, stromal pAKT2 Ser¹²⁸ and pCCTα Ser^315/319/323were coexisted with the biomarker of CAFs-αSMA (Figure 7A). The expression of CAFs-derived pAKT2 Ser¹²⁸ or pCCTα Ser^315/319/323 was positively correlated with the expression of intratumoral pSTAT3 Tyr⁷⁰⁵ (Figure 7F).

Our previous study had demonstrated that the expression of pFAK Tyr³⁹⁷ in ESCC tissues had positively correlated with tumor malignancy [26]. However, the staining intensity of pFAK Tyr³⁹⁷ was low in stroma (22.2%; 24/108) (Figure 7G).

8. Prognostic performance of plasma concentrations of PCs for ESCC patients

To evaluate the correlation between concentrations of plasma lipids and ESCC malignancy, pseuo-targeted lipidomics was applied to quantitatively and comprehensively screen 1, 000 lipids in plasma from 30 cases ESCC patients with 4 cases stage Ⅰ and 26 cases stage Ⅱ and Ⅲ. 559 lipids were detected in the plasma of ESCC patients. Among these, plasma concentrations of 31 lipids were statistically higher in stage Ⅱ and Ⅲ group than stage Ⅰ group (Figure 8A-8N, Supplementary Figure 4A-4Q, and Supplementary Table 7). Importantly, several PCs, PC (18:1/22:1) (stage Ⅱ and Ⅲ group: median: 16.733 ng/mL, IQR: 14.0-21.9; stage Ⅰ group: median: 12.414 ng/mL, IQR: 8.7-15.2), PC (18:1/20:5) (stage Ⅱ and Ⅲ group: median: 159.995 ng/mL, IQR: 122.1-263.9; stage Ⅰ group: median: 87.203 ng/mL, IQR: 69.4-156.9), PC (18:1/20:1) (stage Ⅱ and Ⅲ group: median: 195.706 ng/mL, IQR: 173.1-237.1; stage Ⅰ group: median: 146.432 ng/mL, IQR: 104.4-161.9), PC (18:0/22:1) (stage Ⅱ and Ⅲ group: median: 3.745 ng/mL, IQR: 3.3-5.4; stage Ⅰ group: median: 2.816 ng/mL, IQR: 2.5-3.2), PC (18:0/20:1) (stage Ⅱ and Ⅲ group: median: 92.182 ng/mL, IQR: 72.6-112.2; stage Ⅰ group: median: 66.871 ng/mL, IQR: 56.9-67.7), PC (18:0/18:1) (stage Ⅱ and Ⅲ group: median: 4424.936 ng/mL, IQR: 3774.9-5087.3; stage Ⅰ group: median: 3552.909 ng/mL, IQR: 3172.6-3766.6), PC (16:0/22:1) (stage Ⅱ and Ⅲ group: median: 10.267 ng/mL, IQR: 8.4-12.5; stage Ⅰ group: median: 7.392 ng/mL, IQR: 5.9-8.5), PC (16:0/20:1) (stage Ⅱ and Ⅲ group: median: 272.847 ng/mL, IQR: 222.5-292.9; stage Ⅰ group: median: 204.171 ng/mL, IQR: 179.4-227.2), and metabolites of PCs-LPCs, including LPC (18:1) (stage Ⅱ and Ⅲ group: median: 3531.124 ng/mL, IQR: 2733.6-4025.1; stage Ⅰ group: median: 2305.244 ng/mL, IQR: 2093.6-2943.7), LPC (20:1) (stage Ⅱ and Ⅲ group: median: 177.858 ng/mL, IQR: 138.8-211.0; stage Ⅰ group: median: 115.585 ng/mL, IQR: 96.7-150.2), LPC (22:1) (stage Ⅱ and Ⅲ group: median: 16.167 ng/mL, IQR: 11.4-21.8; stage Ⅰ group: median: 10.547 ng/mL, IQR: 9.1-12.9), LPC (22:2) (stage Ⅱ and Ⅲ group: median: 6.571 ng/mL, IQR: 5.7-7.8; stage Ⅰ group: median: 5.101 ng/mL, IQR: 4.7-6.9), LPC (22:5) (stage Ⅱ and Ⅲ group: median: 1367.435 ng/mL, IQR: 1044.5-1668.4; stage Ⅰ group: median: 922.245 ng/mL, IQR: 818.1-1042.9), LPC (24:1) (stage Ⅱ and Ⅲ group: median: 19.745 ng/mL, IQR: 16.2-23.4; stage Ⅰ group: median: 12.639 ng/mL, IQR: 11.8-17.5), were statistically high in stage Ⅱ and Ⅲ group (P< 0.05, Mann-Whitney U test) (Figure 8A-8N and Supplementary Table 5). Furthermore, some other lipids (among 31 lipids), such as PS (18:0/22:5), PI (18:1/22:6), PE (18:1p/20:5), PE (18:1/22:4), PE (18:1/22:1), PE (18:0p/20:3), PE (18:0/20:1), PE (16:0p/20:5), PE (16:0p/20:3), LPE (24:1), LPE (22:1), LPE (20:1), LPE (19:0), LPA (20:0), FA (16:0), FA (22:6), or Hex2Cer (d18:2/24:1), were also statistically high in stage Ⅱ and Ⅲ group (Supplementary Figure 4A-4Q and Supplementary Table 5). In the validation set, data from 89 ESCC patients showed that plasma PCs were positively correlated with the advanced-stage, higher-TN stages of ESCC patients (Figure 8O). Therefore, plasma PCs concentrations had a positive diagnostic performance and allowed evaluation of ESCC patients’ malignancy.

In this study, we demonstrate that FAK inhibitor leads to altered choline metabolism in CAFs, and this alteration mediates resistance to FAK-targeted therapy. Mechanistically, our findings indicate that FAK inhibitor regulates the CAFs-derived AKT2/CCTα axis and the secretion of lipid metabolites-PCs, which induce malignant cells STAT3 activation to facilitate the therapeutic resistance of tumor cells. Present data establish a concept in CAFs-FAK-regulated and metabolites-mediated control of tumor malignancy with relevance to human ESCC with low stromal FAK expression, and identify potential novel actionable targets for anticancer therapy. Although metabolites have been the focus on tumor cells themselves, our study demonstrate that stroma-derived PCs can play critical role to support tumor malignancy and chemoresistance. Importantly, we found that concentrations of plasma PCs were increased in ESCC patients with stage Ⅱ and Ⅲ, compared with those of stage Ⅰ ESCC patients, indicating plasma PCs can be served as biomarkers for ESCC diagnosis.

Our data show that FAK suppression increases the stromal level of PCs and their metabolites-LPCs, the major membrane structural phospholipids, and the stromal levels of other types of phospholipids, such as PE, PS, PI, or LPS. Moreover, FAK inhibition caused upregulation of unusual lipid subclass-the (O-acyl)-ω-hydroxy FAs (OAHFAs), such as OAHFA(36:1), OAHFA(38:4) from CAFs, suggesting that inhibiting FAK activity results in disruption of stromal choline and its related glycerophospholipid homeostasis, which could contribute to the resistance of FAK inhibition in tumor treatment. Furthermore, our data show that FAK inhibition increases CAFs-released ceramide (CER) and sphingomyelin (SM) levels, whose productions are induced by cellular stress response [27, 28]. It is possible that FAK inhibitor may function as an exogenous stress to dysregulate the choline homeostasis in stromal cells, due to the low expression of stromal FAK, which mediates no available target for FAK inhibitor exerting its anti-signaling function and subsequent anti-growth effect [29]. Overall, dysregulated choline homeostasis and enhanced cellular stress work together to mediate FAK inhibition-induced secretion of PCs from CAFs to mediate FAK inhibitor resistance.

Our MS-based phosphoproteomics indicated that AKT2, the stress-induced protein kinase [30, 31], had effectively stimulated the production of PCs from CAFs after FAK inhibitor treatment. AKT2 is an important signaling regulator of metabolism and can be stimulated to counteract stress-induced apoptosis [32]. Previous study has indicated that stress-responsive FKBP51 activated AKT2 signaling to enhance glucose uptake in skeletal myotubes [33]. Importantly, we identified that FAK inhibition facilitated the phosphorylation of the key rate-limiting step enzyme of PC biosysthesis-CCTα at Ser^315/319/323 sites, and subsequently induced the overproduction of PCs from CAFs. It is clear that the control of CCTs activity is complex and that is involves multiple oncogenic signaling pathways-related factors that modulate expression and function of CCTs [34, 35]. With a combination of phosphoproteomics and functional assays, we found that AKT2 interacted with CCTα and induced the phosphorylation of CCTα to enhance its activity in stromal cells. When stroma-derived metabolites harshly elevated, tumor cells could quickly utilize these metabolites to boost their own growth and resistance to the cytotoxic effect of chemotherapies [34, 36]. Thus, targeting stromal AKT2/CCTα axis and their-derived PCs has been suggested to a effective strategy for enhancing the antitumor effect of FAK inhibitor. Specifically, our data also indicate that FAK inhibition-stimulated the activation of AKT2/CCTα axis and PC production uniquely occurred in CAFs, while not in tumor cells. Combination with our previous report that FAK inhibition could effectively inhibit the expression of several metabolism-related molecules and the malignancy of ESCC cells cultured alone. We suggested that identifying the metabolic processes specifically operating in TME could provide an opportunity for developing the antitumor strategies.

Hostile microenvironmental conditions within tumors, such as nutrient deprivation, oxygen limitation, high metabolic demand, oxidative stress, and drug stimulation, provoke persistent stress to endow malignant cells with greater tumorigenic, metastatic, and drug-resistant capacity [37]. Specifically, calcium signaling pathways have been identified as playing important roles in the establishment and maintenance of drug resistance [38, 39]. Combined these believes with our data, we suggest that rapidly rising Ca²⁺ concentration in stroma and increasing the production of stromal metabolites conferred increased resistance to cell death from stress or apoptotic stimuli, providing a drug-resistant stromal niche for TME-mediated tumor malignancy. Thus, the balance between signaling alterations due to direct effects of FAK inhibition in cancer cells and stromal fibroblasts could potentially be important to determine the overall treatment outcome.

Drug resistance is a common issue when targeted therapy is deployed in both preclinical and clinical settings. Constitutive activation of STAT3 has frequently been observed in several types of solid tumors, and mediates resistance to conventional chemotherapy and targeted therapy [40, 41, 42]. In the present study, we have uncovered a TME-derived metabolites-induced STAT3 activation in response to FAK inhibition. We provide novel evidence that the crosstalk between TME and tumors is critical for driving the targeted therapy response. Correspondingly, ruxolitinib, baricitinib and S3I-201, targeting different levels of the STAT3 signaling cascade, showed a strong synergism with defactinib both in vitro and in vivo.

Our study highlights a previously unclear role of high plasma PCs in facilitating tumor progression and may be exploited as targets for therapy development against solid tumors. Accumulating reports have indicated the relationship between metabolites and the development of tumors [43, 44, 45]. In light of our findings, we speculated that high concentration of plasma PCs in ESCC patients playing a critical role in ESCC malignancy, irrespective of the origin of plasma PCs. Inhibition the effect of PCs on tumor cells can effectively block tumor malignant progression. Taken our data together, plasma PCs levels can not only be used as biomarker to discriminate tumor stages but also be utilized as a potential target for tumor treatment or enhancement the antitumor efficacy of targeted therapies.

In conclusion, combining multi-omics, we systematically investigated PCs-based paracrine communication between CAFs and tumor cells to limit the antitumor efficacy of FAK inhibitors. Mechanistically, the alteration of CAFs-derived AKT2/CCTα axis and its-activated intratumoral JAK2/STAT3 pathway induces the resistance of FAK inhibitor in tumor treatment. Importantly, PCs can potentially be used as new biomarkers for ESCC diagnosis. These data provide a new strategy for targeting metabolites-related pathway for ESCC treatment (Supplementary Fig. 5).

Antibodies and reagents

All information of antibodies and reagents were listed in Supplementary Table 6.

Cell culture and transfection

ESCC cell lines-KYSE410 and KYSE510 were provided by Dr. Yutaka Shimada (Kyoto University). Primary CAFs were purchased from CHI Scientific Inc. All cells were cultured in RPMI-1640 medium contained with 10% heat-inactivated FBS (Gibco) and 1% penicillin/streptomycin in a 37 °C humidified incubator under 5% CO₂.

The siRNA-based approach was applied to generate CCTα or CCTβ-knockdown cells. CCTα or CCTβ siRNAs were transfected into primary CAFs using Lipofectamine 2000 reagent. For plasmid stable transfection, pcDNA 3.1-Flag plasmid contained AKT2 S128A or CCTα S315/319/323A mutant was transfected into CAFs. Subsequently, positive clones were selected for further experiments. Tansfection efficacy was evaluated using immunoblotting. Sequences of siRNAs and mutant plasmids were listed in Supplementary Table 7.

Proliferation assay

For evaluation of CAFs-mediated FAK inhibitor resistance in ESCC treatment, CAFs (3 × 10⁵/well) were plated in the upper chamber of transwell apparatus (0.4 μm insert, 24-well), and ESCC cells (1 × 10⁵/well) were cultured in the lower chamber. Defactinib (0-10 μM) or/and other inhibitors, including miltefosine (25 μM), ruxolitinib, fedratinib (10 μM), or S3I-201 (20 μM), were treated for 4 days, and then the upper chamber was discarded. The growth rate of ESCC cells was assessed using MTS assay (Progema). The IC₅₀ value of defactinib was listed in presented Figures. Other MTS assays of indicated ESCC cells and CAFs were performed in 96-well plates. The exact protocols of MTS assay was according to our previous studies.

Invasion assay

EZCell^TM cell invasion assay kit (Biovision) was applied to evaluate CAFs-mediated defactinib resistance in the anti-invasion of ESCC cells. Briefly, indicated ESCC cells (2 × 10⁴/well) were plated in the upper chamber of transwell apparatus (8 μm insert), and indicated CAFs (6 × 10⁴/well), PC (16:0/20:4) and glycerophosphocholine (10 μM) were added to the lower chamber. Defactinib (10 μM) or/and other inhibitors, including miltefosine (25 μM), ruxolitinib, fedratinib (10 μM), or S3I-201 (20 μM), were treated for 24 hours, and then remove the non-invasive cells from the top chamber using a cotton swab. The invasive cells in top chamber were incubated with cell dye and dissociation solutions for 1 hour at 37 °C in CO₂ incubator. Read the plate at Ex/Em (530/590 nm) to obtain the invasion rate of indicated ESCC cells.

Concentration of phosphoatidylcholine (PC) in condition media (CM) and plasma of ESCC patients

PC assay kit (Sigma-Aldrich) was used to measure the levels of PC in the CM from indicated ESCC cells and CAFs, or ESCC patients’ plasma. Supernatants of indicated cells were collected, and centrifuged at 3000 rpm for 10 minutes to obtain CM. Then, 50 μL sample buffer (44 μL CM, 2 μL PC hydrolysis enzyme, 2 μL PC development mix, and 2 μL fluorescent peroxidase substrate) was combined with 50 μL reaction mix, and incubated at room temperature for 30 minutes. Then, OD value was measured spectrophotometrically at 570 nm. The plasma (50 μL/sample) was directly subjected to PCs assay.

Intracellular concentration of Ca²⁺

CAFs were collected and suspended in 500 μL calcium assay buffer and put on ice for quickly pipetting. Then, intracellular concentration of Ca²⁺was evaluated using Ca²⁺ detection kit (Nanjing Jiancheng Bioengineering Institute). Briefly, 50 μL sample were incubated with 500 μL MTB reagent, 1mL alkaline solution, and 50 μL protein clarifier for 5 minutes. OD value was measured spectrophotometrically at 610 nm.

STAT3 activity

Human phospho-STAT3 Tyr⁷⁰⁵ and total STAT3 ELISA kit (Raybiotech) was used to evaluate the activity of intratumoral STAT3. The protocols were according to manufacture’s instruction and our previous studies.

Immunoprecipitation (IP) and immunoblotting (IB) analysis

For IP assay, indicated cells were washed with PBS, lysed in NP40 buffer supplemented with protease and phosphatase inhibitors for 30 minutes, and then centrifuged at 12000 g for 20 min at 4℃. Supernatants were collected to incubated with indicated primary antibodies (approximately 10 μg antibody/sample) and protein A/G sepharose beads (Thermo Fisher) on a rotator at 4℃ overnight. Then, samples were centrifuged at 4°C for 5 minutes at 3000 g, supernatants were discarded, and pellets were washed with 800 μL cold NP40 buffer for 3 times. Finally, beads were collected, and 60 μL loading buffer was added to the beads. The beads were bathed in metal for 5 minutes, and supernatants were subjected to IB assay.

For IB assay, proteins were separated using sodium dodecyl sulfate (SDS)-PAGE, and transferred onto a nitrocellulose (NC) membrane. After blocking with PBS buffer solution containing 0.1% Tween-20 and 5% nonfat milk for 1 hour, the membranes were incubated with indicated primary antibodies at 4°C overnight. PBST was used to wash NC membranes for three times. The membranes were incubated with secondary antibodies for 1 hour, and then washed an additional three times with PBST and detected by chemiluminescence (Thermo Fisher).

Xenograft study

Female BALB/c-nu mice (purchased from Beijing Vital River Laboratory) with 3-4 weeks of age were used in present assay. All animal procedures were approved by Institutional Review Board of Peking University Cancer Hospital & Institute.

KYSE410 or KYSE510 cells were subcutaneously co-injected with indicated CAFs into the flank of mice. When the tumor reached around 100 mm³, defactinib (25mg /kg/day, p.o) alone or in the presence of ruxolitinib (10 mg/kg/day, p.o.), or fedratinib (10 mg/kg/day, p.o.), or S3I-201 (25 mg/kg/day, p.o.) for consecutive 3 weeks (n=5/group). Tumor volume was evaluated using our reported formula. Human Ki67, CD31, or LYVE-1 ELISA kits (Raybiotech) were applied to measure the proliferation, angiogenesis, or lymph-angiogenesis abilities of indicated ESCC tumors. The experimental protocols were according to manufacturer’s instructions.

For evaluation of lymph node metastasis of ESCC cells, KYSE410 or KYSE510 cells were subcutaneously co-injected with indicated CAFs into the footpads of mice (n=5/group). The agents used in this assay was consistent with the model that subcutaneous tumor cells inoculation. Treatment was started for week 2 and this experiment was sustained for 5 weeks. Lymph node volume was evaluated by our reported formula [12].

ESCC tissues and IHC staining

All procedures and experiments of ESCC tissues were approved by the institutional Review Board of Peking University Cancer Hospital. The protocols of IHC staining and the calculation of staining index were according to our previous studies [12,26]. The dilution of primary antibodies was as follow: FAK (1:100), pFAK Tyr³⁹⁷ (1:100), pAKT2 Ser¹²⁸ (1:100), pCCTα Ser^315/319/323 (1:500), pSTAT3 Tyr⁷⁰⁵ (1:4000), or αSMA (1:1500).

Phosphoproteomic analysis

Protein extraction and tryptic digestion

CAFs samples were washed with three times with phosphate buffer saline (PBS), and then lysed in SDS lysis buffer supplemented with PMSF and phosphatase inhibitors (Cat # 4906837001, Roche). Samples were sonicated with 3 min (1 s on and 1 s off, 80 W power) on ice and centrifuged at 12, 000 g for 10 min to remove insoluble debris. The supernatants were collected, and the protein concentration was measured using BCA protein assay (Cat # 23225, ThermoScientific). Extracted proteins were reduced in 5 mM DTT at 55 °C for 30 min and then alkylated in 10 mM iodoacetamide at room temperature for 15 min in darkness. Six times of the volume of precooled acetone was added to above system to precipitate the proteins, and place samples at -20 °C overnight, then centrifuged at 8, 000 g for 10 min to collect the precipitate. Samples underwent trypsin digestion (enzyme-to-substrate ratio of 1:50 at 37 °C overnight) and lyophilized.

The enrichment of phosphorylated peptides

For iTRAQ labeling, the lyophilized samples were resuspended in 60 μL 100 mM TEAB solution and then transferred into 1.5 mL Eppendorf tube for labeling. 70 μL isopropanol was added to iTRAQ reagent vial (Cat # 4381663, ABSCIEX) at room temperature, and then vortexed for mixing, centrifuged. This step was repeated 1 time. 100 μL iTRAQ label reagent was added to samples for mixing. Tubes were incubated at room temperature for 2 h. Finally, 200 μL HPLC water was added to each sample and incubated for 30 min to stop reaction. The labeled peptides samples were lyophilized and stored at -80 °C.

For phosphopeptides enrichment, the high-select^TM Fe-NTA phosphopeptide enrichment kit was used (Cat # A32992, ThermoFisher). Briefly, the labeling peptides were incubated with 200 μL binding/washing buffer in spin columns and incubated for 30 min. Then, placed the columns into microcentrifuge tubes for centrifuging at 1000 g for 30 s, and discarded the flowthrough. Bound columns were washed by 200 μL binding/washing buffer, and then centrifuged at 1000 g for 30 s. This step was repeated for three times. Bound peptides were washed with 200 μL HPLC water to remove the unphosphorylated peptides. The remaining peptides were added with 100 μL elution buffer, resolved on columns, and then centrifuged at 1, 000 g for 30 s. The incubation reaction was repeated. The final enriched phosphopeptides were immediately vacuum concentrated for further use.

LC-MS/MS

Peptides samples were analyzed on an EASY-nLC 1200 system (ThermoFisher) coupled with a TimsTOF pro mass spectrometry (Bruker). Peptides were re-dissolved in mobile phase A (H₂O-FA; 99.9: 0.1, v/v) and immediately loaded onto an 25 cm long, C18 analytical column (75 μm inner diameter; Cat #AUR2-25075C18A; Ionopticks).

Peptides were separated at an fluent flow rate of 400 nL/min. Peptide separation was achieved with a 90 min gradient (0-66 min, 2 to 21% of buffer B; 66-73 min, 21 to 42% of buffer B; 73-82 min, 42 to 95 % of buffer B; 82-90 min, 95% of buffer B). The eluted phosphopeptides were ionized and detected by the TimsTOF pro mass spectrometry. Mass range was 100-1700 m/z. Ion mobility was 0.6-1.6. Collision energy was 20-59 eV. The temperature of drying gas was 180 °C. Capillary voltage was 1.5 KV.

MS database searching

MS raw files generated by LC-MS/MS were searched using Max Quant (version 1.6.17.0.). Protease was Trypsin/P. Up to 2 missed cleavages were allowed. iTRAQ (N-term, K) and carbamidomethyl (C) were considered as fixed modifications. Variable modifications were oxidation (M), acetyl (Protein N-term), and phospho (S/T/Y). The cutoff of false discovery rate (FDR) by using a target-decoy strategy was 1% for peptides.

Identification of variant peptides

MaxQuant search data were analyzed by “R” package (version 2.15) software. MaxQuant output files were log₂-transformed, normalized and adjusted in “R” package. Protein, peptide, and phosphorylation-site identifications were filtered by removing entries corresponding to reverse database identifications, potential contaminants and those identified by site alone. The screening conditions were as follows: the phosphorylation sites with expression value ≥ 50% in any groups were retained. Phosphorylation sites with missing value ≤ 50% were filled with the mean value of the same group, screen localization probability ≥ 0.75, Delta score ≥ 8, and converse data via median normalization and log₂-transformation to obtain highly reliable phosphorylation sites. Based on obtained highly reliable phosphorylation sites, the differences between two groups was calculated (Difference screening condition: fold change ≥ 2), and the MOMO (http://meme-suite.org/tools/momo) software was applied to analyze the motif of phosphorylation sites to obtain the specific substrates. Then, Uniprot, KEGG/GO databases were used to annotate the function of phospho-proteins. Method of KEGG/GO functional enrichment analysis was as follow: human proteins were used as the background, and the proteins corresponding to the differential phosphorylation sites were screened from background. The Fisher’s exact test (as implemented in R software) were used to evaluate the different pathways corresponding to phospho-proteins between control and treated groups. Pathways with P value ≤ 0.05 were considered to be statistically differential pathways.

Untargeted metabolomics

Sample preparation

1 mL of CM were transferred to a 5 mL Eppendorf tube, and then lyophilized. 400 μL mixture of methanol and water (1/4, vol/vol) were added to each sample. Samples were vortexed for 30 s, sonicated for 3 min, and then added with 10 μL internal standard (2-chloro-L-phenylalanine; 0.3 mg/mL, dissolved in methanol). Mixed samples were place at -20 °C for 2 h, and then centrifuged at 13, 000 g for 10 min at 4 °C. 150 μL supernatants from each sample were collected using crystal syringes (filtered through 0.22 μm microfilters) and transferred to LC vials, which were stored at -80 °C for further LC-MS assay.

LC-MS analysis

Metabolic analyses were performed on an UPLC Ultimate 3000 system (Dionex) coupled to a A-Exactive mass spectrometer (ThermoFisher). The chromatographic separation was conducted using ACQUITY UPLC HSS T3 column (100 mm × 2.1 mm, 1.8 μm; Waters) with temperature was set at 45 °C. The mobile phases A and B were consisted of water and acetonitrile both containing 0.1 % formic acid, and run at a flow rate of 0.35 mL/min. The linear gradient was as follows: 0.01 min, 5% of phase B; 2 min, 5% of phase B; 4 min, 30 % of phase B; 8 min, 50 % of phase B; 10 min, 80 % of phase B; 14-15 min, 100 % of phase B; 15.1-18 min, 5 % of phase B. The mass spectrometer was performed in a full-scan mode ranging from 100-1000 m/z, running at a 70, 000 resolution in both positive and negative modes simultaneously.

Data extraction and analysis

The acquired raw data were analyzed using the Progenesis QI (version 2.3; Nonlinear Dynamics), and the applied parameters were as follows: precursor tolerance was set 5 ppm, product tolerance was set 10 ppm, and product ion threshold was set 5 %. The metabolite identification was conducted using EMDB database (http://www.emdataresource.org/mission.html). The extracted data were further reduced by removing any peaks with missing value (ion intensity =0) in more than 50 % samples, and the selected score is >36 (full score is 60). Results of score lower than 36 were deleted. The positive and negative data were combined to obtain a combined data which were imported into “R” package (version 2.15) software. Unsupervised principle component analysis (PCA) and PLS-DA and OPLS-DA were combined to discriminate the metabolite profile between groups. The variable importance in projection (VIP) value (from OPLS-DA) > 1.00 and P value (Student’s test) < 0.05 of each metabolite were used as the combined cut-offs of the statistical significance.

Untargeted lipidomic analysis

Sample preparation

After lyophilized, CM lipids were extracted by 400 μL mixture of chloroform and methanol (2/1, vol/vol) in the presence of 20 μL internal standard (Lyso PC-17:0, 0.1 mg/mL), and then vortexed for 30 s, sonicated for 3 min, plated for 30 min at -20 °C. Subsequently, samples were centrifuged at 12, 000 g for 10 min at 4 °C, and volatilized. The residue of lipids were re-dissolved with 200 μL isopropanol and methanol (1/1, vol/vol), vortexed for 30 s, sonicated for 3 min, and then centrifuged at 12, 000 g for 10 min at 4 °C. 150 μL supernatants were collected for further study.

LC-MS analysis

Lipidomic analyses were conducted using an UPLC Ultimate 3000 system (Dionex) coupled to a A-Exactive mass spectrometer (ThermoFisher). The chromatographic separation was conducted using ACQUITY UPLC BEH C8 column (100 mm × 2.1 mm, 1.7 μm; Waters) with temperature was set at 55 °C. The mobile phase A was consisted of acetonitrile and water (6/4, vol/vol) containing 10 mM ammonium acetate. The mobile phase B was consisted of isopropanol and acetonitrile (9/1, vol/vol) containing 10 mM ammonium acetate. The flow rate was set at 0.26 mL/min. The linear gradient was as follows: 0.00 min, 32 % of phase B; 1.5 min, 32 % of phase B; 15.5 min, 85 % of phase B; 15.6 and 18.0 min, 97 % of phase B; 18.1 and 21.0 min, 32 % of phase B. The mass spectrometer was performed in a full-scan mode ranging from 100-1500 m/z, running at a 70, 000 resolution in both positive and negative modes simultaneously.

Data extraction and analysis

The raw data were extracted and analyzed using the Lipid Research (version 4.1.16，ThermoFisher). Following analysis methods and protocols were according to metabolomics analysis.

Pseudo-targeted Lipidomics

Sample preparation

The lipid was extracted from 100 μL plasma using 300 μL choroform-methanol (2:1, v/v, supplementing with 0.1 mM BHT) contained 20 μL known amounts of 74 isotope-labeled internal mix standards (identify 1, 000 lipids), vortex for 30 seconds and ultrasonic extraction for 10 minutes, and then hold 30 minutes at -20 °C, centrifuge 10 minutes (13, 000 rpm) at 4 °C, 200 μL chloroform layers were transferred into a centrifuge tube. Residues were reextracted according to above condition. Combine the chloroform layers together and lyophilize in a centrifugal vacuum evaporator at 4 °C. Samples were reconstituted in isopropanol-methanol (1:1, v/v), vortex for 30 seconds, ultrasonic extraction for 3 minutes, centrifuge 10 minutes (13, 000 rpm) at 4 °C, and then stored at -20 °C. Finally, 150 μL supernates were subjected to LC/MS analysis.

LC-MS analysis

LC-MS analysis was conducted using an ExionLC^TM system (ABSCIEX) coupled to Qtrap 6500 plus system (ABSCIEX) equipped with an IonDrive^TM Turbo V source. The chromatographic separation was conducted using ACQUITY UPLC BEH C8 column (100 mm × 2.1 mm, 1.7 μm; Waters) with temperature was set at 55 °C. The mobile phase A was consisted of acetonitrile and water (6/4, vol/vol) containing 0.1 % formic acid and 10 mM ammonium formate. The mobile phase B was consisted of isopropanol and acetonitrile (9/1, vol/vol) containing 0.1 % formic acid and 10 mM ammonium formate. The flow rate was set at 0.35 mL/min. The sample injection volume was 5 μL. The linear gradient was as follows: 0.00 min, 0 % of phase B; 1.5 min, 0 % of phase B; 5.0 min, 55 % of phase B; 15.0 min, 90 % of phase B; 16.0 and 18.0 min, 100 % of phase B; 18.1 and 20.0 min, 0 % of phase B. The MS analysis was conducted in the negative/positive-ion mode working in the time-scheduled MRM method to high-throughoutly screen more than 1000 lipids. The source condtion is as follow: curtain gas is 35 psi, the ion spray (IS) voltage is -4500 V/+5500 V, source temperature is 350 °C, and Gas 1 and Gas 2 is 40 psi and 45 psi.

Data extraction and analysis

The acquired raw data were analyzed using the MRMPROBS software [46] to perform automated batch processing, including peak extraction, alignment, identification, peak area integral. Relevant parameters are set as follows: smoothing level is 2, minimum peak width is 5, minimum peak height is 500, retention time tolerance is 0.2 min, and supplemented by manual correction. The obtained qualitative and quantitative tables were quantitatively analyzed using response factor method. Furthermore, integral peak area of metabolites was brought into following formula: Lipid concentration = Target lipid peak area (A1)/Peak area of lipid internal standard corresponding to target lipid (A2) × The internal standard concentration value of the corresponding lipid internal standard in the sample (C) × The constant volume (V)/Weighed sample mass or volume (N). The concentration of lipids in plasma from different tumor stages of ESCC patients (30 cases) were divided into two groups, including Stage Ⅰ and stage Ⅱ and Ⅲ, and the difference between this two groups was analyzed using Mann-Whitney U test. Variables are expressed as the median with interquartile range (IQR) in Supplementary Table 7.

Data availability

The phosphoproteomic data have been deposited in https://www.iprox.cn/page/home.html, and the accession number was: PXD032254. The metabolomics related data have been deposited in https://www.ebi.ac.uk/metabolights/, and the accession numbers were: MTBLS4489 and MTBLS4502.

Statistical analysis

All data are expressed as the mean ± SD, and statistical analyses are performed by Graphpad 7.0 software. Unpaired Student’s t-test (two-tailed) was applied to compare the difference between two groups. For analysis of clinical samples, Chi-square test was used to evaluate the correlation between two factor. Kaplan-Meier method was employed to establish the survival curves of ESCC patients. The statistical analysis of phosphoproteomics, metabolomics, lipidomics, or pseudo-targeted Lipidomics was listed in respective “Materials and methods” section. P-value < 0.05 was considered statistically significant.

Competing Interests

The authors have declared that no competing interest exists.

Author contributions

Qimin Zhan designed the experiments and wrote the paper. Jie Chen, Lingyuan Zhang, Di Zhao, Jing Zhang, Yuanfan Xiao, Qingnan Wu, and Yan Wang performed the experiments and analyzed the data.

Acknowledgement

This work was supported by the National Natural Science Foundation of China (81988101, 81830086, and 81972243), CAMS Innovation Fund for Medical Sciences (2019-I2M-5-081), Funding by Major Program of Shenzhen Bay Laboratory (S201101004), Guangdong Basic and Applied Basic Research Foundation (2019B030302012), the Fund of “San-ming” Project of Medicine in Shenzhen (No. SZSM201812088).

Bray J, Ferlay J, Soerjomataram I, et al. Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2018; 68(6): 394-424.
Chen W, Zheng R, Baade PD, et al. Cancer statistics in China, 2015. CA Cancer J Clin. 2016; 66(2): 115-132.
Lin Y, Totsuka Y, Shan B, et al. Esophageal cancer in high-risk areas of China: research progress and challenges. Ann Epidemiol. 2017; 27(3): 215-221.
Hirano H, Kato K. Systemic treatment of advanced esophageal squamous cell carcinoma: chemotherapy, molecular-targeting therapy and immunotherapy. Jpn J Clin Oncol. 2019; 49(5): 412-420.
Reichenbach ZW, Murray MG, Saxena R, et al. Clinical and translational advances in esophageal squamous cell carcinoma. Adv Cancer Res. 2019; 144: 95-135.
Zhang H, Hua Y, Jiang Z, et al. Cancer-associated fibroblast-promoted LncRNA DNM3OS confers radioresistance by regulating DNA damage response in esophageal squamous cell carcinoma. Clin Cancer Res. 2019; 25(6): 1989-2000.
Zhang H, Xie C, Yue J, et al. Cancer-associated fibroblasts mediated chemoresistance by a FOXO1/TGFbeta1 signaling loop in esophageal squamous cell carcinoma. Mol Carcinog. 2017; 56(3): 1150-1163.
Zhao Q, Huang L, Qin G, et al. Cancer-associated fibroblasts induce monocytic myeloid-derived suppressor cell generation via IL-6/exosomal miR-21-activated STAT3 signaling to promote cisplatin resistance in esophageal squamous cell carcinoma. Cancer Lett. 2021; 518: 35-48.
Chen Y, McAndrews KM, Kalluri R. Clinical and therapeutic relevance of cancer-associated fibroblasts. Nat Rev Clin Oncol. 2021. [Online ahead of print].
Higashino N, Koma YI, Hosono M, et al. Fibroblast activation protein-positive fibroblasts promote tumor progression through secretion of CCL2 and interleukin-6 in esophageal squamous cell carcinoma. Lab invest. 2019; 99(6): 777-792.
Yang X, Li Y, Zou L, et al. Role of exosomes in crosstalk between cancer-associated fibroblasts and cancer cells. Front Oncol. 2019; 9: 356.
Chen J, Wang Y, Zhang W, et al. NOX5 mediates the crosstalk between tumor cells and cancer-associated fibroblasts via regulating cytokine network. Clin Transl Med. 2021; 11(8): e472.
Wu D, Zhuo L, Wang X. Metabolic reprogramming of carcinoma-associated fibroblasts and its impact on metabolic heterogeneity of tumors. Semin Cell Dev Biol. 2017; 64: 125-131.
Li Z, Sun C, Qin Z. Metabolic reprogramming of cancer-associated fibroblasts and its effect on cancer cell reprogramming. Theranostics. 2021; 11(17): 8322-8336.
Dawson JC, Serrels A, Stupack DG, et al. Targeting FAK in anticancer combination therapies. Nat Rev Cancer. 2021; 21(5): 313-324.
Sulzmaier FJ, Hean C, Schlaepfer DD. FAK in cancer: mechanistic findings and clinical applications. Nat Rev Cancer. 2014; 14((): 598-610.
Zhou J, Yi Q, Tang L. The roles of nuclear focal adhesion kinase (FAK) on cancer: a focused review. J Exp Clin Cancer Res. 2019; 38(1): 250.
Mitra SK, Schlaepfer DD. Integrin-regulated FAK-Src signaling in normal and cancer cells. Curr Opin Cell Biol. 2006; 18(5): 516-523.
Osipov A, Saung MT, Zheng L, et al. Small molecule immunomodulation: the tumor microenvironment and overcoming immune escape. J Immunother Cancer. 2019; 7(1): 224.
Murphy JM, Rodriguez YAR, Jeong K, et al. Targeting focal adhesion kinase in cancer cells and the tumor microenvironment. Exp Mol Med. 2020; 52(6): 877-886.
Lv PC, Jiang AQ, Zhang WM, et al. FAK inhibitors in cancer, a patent review. Expert Opin Ther Pat. 2018; 28(2): 139-145.
Ahmed N, Escalona R, Leung D, et al. Tumour mciroenvironment and metabolic plasticity in cancer and cancer stem cells: perspectives on metabolic and immune regulatory signatures in chemoresistant ovarian cancer stem cells. Semin Cancer Biol. 2018; 53: 265-281.
Bu L, Baba H, Yasuda T, et al. Functional diversity of cancer-associated fibroblasts in modulating drug resistance. Cancer Sci. 2020; 111(10): 3468-3477.
Ramapriyan R, Caetano MS, Barsoumian HB, et al. Altered cancer metabolism in mechanisms of immunotherapy resistance. Pharmacol Ther. 2019; 195: 162-171.
Jiang H, Liu X, Knolhoff BL, et al. Development of resistance to FAK inhibition in pancreatic cancer is linked to stromal depletion. Gut. 2020; 69(1): 122-132.
Chen J, Zhang W, Wang Y, et al. The diacylglycerol kinase alpha (DGKalpha)/Akt/NF-kappaB feedforward loop promotes esophageal squamous cell carcinoma (ESCC) progression via FAK-dependent and FAK-independent manner. Oncogene. 2019; 38(14): 2533-2550.
Nikolova-Karakashian MN, Rozenova KA. Ceramide in stress response. Adv Exp Med Biol. 2010; 688: 86-108.
Herr I, Debatin KM. Cellular stress response and apoptosis in cancer therapy. Blood. 2001; 98(9): 2603-2614.
Demircioglu F, Wang J, Candido J, et al. Cancer associated fibroblast FAK regulates malignant cell metabolism. Nat Commun. 2020; 11(1): 1290.
Li G, Anderson RE, Tomita H, et al. Nonredundant role of Akt2 for neuroprotection of rod photoreceptor cells from light-induced cell death. J Neurosci. 2007; 27(1): 203-211.
Swaminathan S, Rosner MH. The podocyte under stress: AKT2 to the rescue. Am J Kidney Dis. 2014; 63(4): 555-557.
Muslin AJ. Akt2: a critical regulator of cardiomyocyte survival and metabolism. Pediatr Cardiol. 2011; 32(3): 317-322.
Balsevich G, Häusl AS, Meyer CW, et al. Stress-responsive FKBP51 regulates AKT2-AS160 signaling and metabolic function. Nat Commun. 2017; 8(1): 1725.
Glunde K, Bhujwalla ZM, Ronen SM. Choline metabolism in malignant transformation. Nat Rev Cancer. 2011; 11(12): 835-848.
Cornell RB, Ridgway ND. CTP: phosphocholine cytidylytransferase: function, regulation, and structure of an amphitropic enzyme required for membrane biogenesis. Prog Lipid Res. 2015; 59: 147-171.
Ackerman D, Simon MC. Hypoxia, lipids, and cancer: surviving the harsh tumor microenvironment. Trends Cell Biol. 2014; 24(8): 472-478.
Cubillos-Ruiz JR, Bettigole SE, Glimcher LH. Tumorigenic and immunosuppressive effects of endoplasmic reticulum stress in cancer. Cell. 2017; 168(4): 692-706.
Santoni G, Morelli MB, Marinelli O, et al. Calcium signaling and the regulation of chemosensitivity in cancer cells; role of the transient receptor potential channels. Adv Exp Med Biol. 2020; 1131: 505-517.
Kucukkaya B, Basoglu H, Erdag D, et al. Calcium homeostasis in cisplatin resistance epithelial ovarian cancer. Gen Physiol Biophys. 2019; 38(4): 353-363.
Shih PC, Mei KC. Role of STAT3 signaling transduction pathways in cancer stem cell-associated chemoresistance. Drug Discov Today. 2021; 26(2): 1450-1458.
Yang PL, Liu LX, Li EM, et al. STAT3, the challenge for chemotherapeutic and radiotherapeutic efficacy. Cancers (Basel). 2020; 12(9): 2459.
Huynh J, Chand A, Gough D, et al. Therapeutically exploiting STAT3 activity in cancer-using tissues repair as a road map. Nat Rev Cancer. 2019; 19(2): 82-96.
Martínez-Reyes I, Chandel NS. Cancer metabolism: looking forward. Nat Rev Cancer. 2021; 21(10): 669-680.
Sullivan LB, Gui DY, Vander Heiden MG. Altered metabolite levels in cancer: implications for tumour biology and cancer therapy. Nat Rev Cancer. 2016; 16(11): 680-693.
Elia I, Haigis MC. Metabolites and the tumour microenvironment: from cellular mechanisms to systemic metabolism. Nat Metab. 2021; 3(1): 21-32.
Tsugawa H, Kanazawa M, Ogiwara A, et al. MRMPROBS suite for metabolomics using large-scale MRM assays. Bioinformatics. 2014; 30(16): 2379-2380.

No competing interests reported.

PDF

Version 1

posted 19 Apr, 2022

You are reading this latest preprint version

Phosphatidylcholines (PCs) serve as biomarkers for prognosis of esophageal squamous cell carcinoma (ESCC) and compromise antitumor efficacy of focal adhesion kinase (FAK) inhibitor

Status:

Version 1

Abstract

Figures

Introduction

Results

1. CAFs impair the antitumor effect of FAK inhibitor in ESCC treatment

2. Untargeted metabolomics and lipidomics identify that FAK inhibitor induces the secretion of phosphocholines (PCs) from CAFs

3. CAFs-released PCs impair the antitumor effect of defactinib on ESCC cells

4. Phosphoproteomics identifies that the activation of AKT2/CCTα axis in CAFs contributes to the secretion ofPCs

5. CAFs-derived AKT2/CCTα axis impairs the antitumor effect of defactinib on ESCC cells

6. CAFs-released PCs activate intratumoral STAT3 to mediate the resistance of defactinib in ESCC treatment

7. Stroma-derived AKT2/CCTα axis determines the ESCC progression and the survival of ESCC patients

8. Prognostic performance of plasma concentrations of PCs for ESCC patients

Discussion

Methods And Materials

Antibodies and reagents

Cell culture and transfection

Proliferation assay

Invasion assay

Concentration of phosphoatidylcholine (PC) in condition media (CM) and plasma of ESCC patients

Intracellular concentration of Ca2+

STAT3 activity

Immunoprecipitation (IP) and immunoblotting (IB) analysis

Xenograft study

ESCC tissues and IHC staining

Phosphoproteomic analysis

Protein extraction and tryptic digestion

The enrichment of phosphorylated peptides

LC-MS/MS

MS database searching

Identification of variant peptides

Untargeted metabolomics

Sample preparation

LC-MS analysis

Data extraction and analysis

Untargeted lipidomic analysis

Sample preparation

LC-MS analysis

Data extraction and analysis

Pseudo-targeted Lipidomics

Sample preparation

LC-MS analysis

Data extraction and analysis

Data availability

Statistical analysis

Declarations

References

Competing Interests

Supplementary Files

Status:

Version 1

Intracellular concentration of Ca²⁺