LncRNA TYMSOS facilitates breast cancer metastasis and immune escape through downregulating ULBP3

Summary

The focus of the study is to examine the function of TYMSOS in immune escape of breast cancer, which is the most frequently diagnosed malignancy among women globally. Our study demonstrated that upregulated TYMSOS was associated with unfavorable prognosis and immune escape in breast cancer. TYMSOS promoted the malignant phenotypes of breast cancer cells, and reduced the cytotoxicity of NK92 cells on these cells. CBX3 was a downstream effector in TYMSOS-induced malignant phenotypes in breast cancer cells. Mechanistic studies showed that TYMSOS facilitated CBX3-mediated transcriptional repression of ULBP3, and it also promoted SYVN1-mediated ubiquitin-proteasomal degradation of ULBP3. TYMSOS promoted cell growth, metastasis, and immune escape via CBX3/ULBP3 or SYVN1/ULBP3 axis. The in vivo studies further showed that silencing of TYMSOS repressed tumor growth and boosted NK cell cytotoxicity. In sum, TYMSOS boosted breast cancer metastasis and immune escape via CBX3/ULBP3 or SYVN1/ULBP3 axis.

Graphical abstract

Highlights

• •TYMSOS acted as an oncogene and repressed the cytotoxicity of NK cells
• •CBX3 was a downstream effector of TYMSOS and a transcriptional repressor of ULBP3
• •TYMSOS promoted SYVN1-mediated ubiquitin-proteasomal degradation of ULBP3
• •TYMSOS promoted metastasis and immune escape via CBX3/ULBP3 or SYVN1/ULBP3 axis

Biological sciences; Molecular biology; Immunology; Cancer

Introduction

Breast cancer remains the frequently diagnosed malignancy among women globally.¹ As a highly heterogeneous disease, breast cancer has been classified into four intrinsic subtypes: triple-negative breast cancer, human epidermal growth factor receptor type 2-positive, luminal A, and luminal B subtypes.² Recently, immunotherapy has emerged as a novel therapeutic approach for breast cancer.³ Therefore, immune escape, one of the determinants of immunotherapy, has gained increasing attentions. In the early stage of breast cancer, the cellular effectors of the immune system, in particular natural killer (NK) cells, play a critical role in eliminating nascent transformed cells. Decreased NK cell cytotoxicity results in metastatic growth of cancer cells, leading to the unfavorable clinical outcomes.⁴ The underlying mechanisms associated with immune escape and metastasis in breast cancer merit in-depth investigation.

Long non-coding RNAs (lncRNAs) consisting of >200 nt in length have emerged as therapeutic targets, diagnostic, and/or prognostic biomarkers in different cancers.⁵ Aberrant expression of thymidylate synthetase opposite strand (TYMSOS) has been observed in different cancers.^6,7,8,9,10 For instance, TYMSOS promotes cell proliferation, migration, and invasion in gastric cancer (GC) through miR-4739/ZNF703 axis.⁶ It is also identified as an oncogene in non-small-cell lung cancer (NSCLC), thyroid carcinoma, and osteosarcoma.^7,8,9,10 Notably, The Cancer Genome Atlas (TCGA) data indicate that TYMSOS is elevated in breast cancer. However, the biological role of TYMSOS on immune escape in breast cancer remains ambiguous.

Chromobox protein homolog 3 (CBX3) belongs to CBX family which is involved in the transcriptional regulation of targets through chromatin modification.^11,12 Several studies have illustrated the upregulation of CBX3 in different cancers, and it is associated with cancer progression and the infiltration of immune cells.^12,13,14,15 RPISeq (http://pridb.gdcb.iastate.edu/RPISeq/) has predicted that CBX3 potentially interacts with TYMSOS. CBX3 attracts our attention due to its critical role in transcriptional regulation, as well as its elevated expression in breast cancer.^12,16 Additionally, a putative interaction between CBX3 and the NK group 2D (NKG2D) ligands ULBP3 was predicted by AnimalTFDB (http://bioinfo.life.hust.edu.cn/AnimalTFDB/#!/), indicating the potential role of TYMSOS/CBX3/ULBP3 axis in the regulation of NK cell cytotoxicity. In addition to the transcriptional regulation, ubibrowser (http://ubibrowser.bio-it.cn/ubibrowser/home/index) and RPISeq predicted the ubiquitination of ULBP3 and the association between TYMSOS and synoviolin 1 (SYVN1), respectively. Based on TCGA data, SYVN1 was increased in breast tumors. We thus hypothesized that TYMSOS possibly regulated ULBP3 turnover via SYVN1-mediated ubiquitination and degradation.

In this study, we demonstrated that TYMSOS was increased in breast tumors and cancer cells, and its expression was associated with unfavorable prognosis and immune escape. TYMSOS functioned as an oncogene in breast cancer cells, and lack of TYMSOS boosted NK cell cytotoxicity. TYMSOS facilitated CBX3-mediated transcriptional repression of ULBP3, and it also promoted SYVN1-mediated degradation of ULBP3 via ubiquitin-proteasome pathway. Functional studies have showed that TYMSOS promoted cell growth, metastasis, and immune escape via CBX3/ULBP3 or SYVN1/ULBP3 axis. The in vivo studies further showed that silencing of TYMSOS suppressed tumor growth and enhanced NK cell cytotoxicity.

Results

Upregulated TYMSOS indicated the poor prognosis and was correlated with cytokine levels related to immune escape

We first examined the level of TYMSOS in breast tumor (n = 58) and their normal counterparts (n = 58). As presented in Figures 1A and 1B, TYMSOS was elevated in breast tumors. The level of TYMSOS was positively correlated with tumor size, TNM stage, and lymph node metastasis (Table S1). Kaplan-Meier analysis showed that breast cancer patients with high TYMSOS level exhibited poor prognosis, compared to those with low TYMSOS level (Figure 1C). In comparison with non-metastatic breast tumors (n = 40), TYMSOS level was much higher in metastatic breast tumors (n = 18) (Figure 1D). Additionally, the immunosuppressive cytokines TGF-β, IL-10, and IL-1β were elevated in the breast tumors with high TYMSOS level, whereas the antitumor cytokines IFN-γ, IL-2, and TNF-α were decreased in these tumors (Figure 1E). Collectively, these findings indicate that high TYMSOS expression conferred unfavorable prognosis of breast cancer patients, and it was correlated to immune escape.

Figure 1.

Upregulated TYMSOS indicated the poor prognosis and was correlated with cytokine levels related to immune escape

(A and B) The TYMSOS level in breast tumors (n = 58) and their normal counterparts (n = 58) were detected by RT-qPCR.

(C) Kaplan-Meier survival analysis of breast cancer patients.

(D) The TYMSOS level in non-metastatic (n = 40) and metastatic breast tumors (n = 18) were detected by RT-qPCR.

(E) The cytokine levels (TGF-β, IL-10, IL-1β, IFN-γ, IL-2, and TNF-α) in breast tumors with high (n = 29) and low TYMSOS level (n = 29) were examined by RT-qPCR. Data were presented as mean ± S.D. ∗, p < 0.05, ∗∗, p < 0.01 by Student’s t test.

TYMSOS promoted the malignant phenotypes of breast cancer cells, and repressed the cytotoxicity of NK cells on breast cancer cells

To further delineate the biological function of TYMSOS in breast cancer, the TYMSOS level was examined in breast cancer cells. TYMSOS was remarkably increased in breast cancer cell lines, including Hs578T, T47D, MCF-7, MDA-MB-231, and MDA-MB-468 cells, in comparison with normal mammary epithelial cell line MCF10A cells (Figure 2A). Among these cells, MCF-7 and MDA-MB-231 cells expressed relatively low and high TYMSOS levels, respectively. These two cell lines were thus selected for subsequent gain- and loss-of-function studies. As expected, transfection of shTYMSOS-1 or shTYMSOS-2 led to a dramatic downregulation of TYMSOS in MDA-MB-231 cells, and transfection of TYMSOS overexpression plasmid resulted in upregulation of TYMSOS in MCF-7 cells (Figure 2B). Silencing of TYMSOS suppressed cell proliferation of MDA-MB-231 cells as detected by CCK-8 and colony formation assays (Figures 2C and 2D). By contrast, overexpression of TYMSOS promoted cell proliferation in MCF-7 cells (Figures 2C and 2D). Transwell assays further showed that knockdown of TYMSOS impaired the migratory and invasive capacities of MDA-MB-231 cells, whereas overexpression of TYMSOS exerted opposite effects on cell migration and invasion in MCF-7 cells (Figure 2E). We next investigated the role of TYMSOS on the killing effect of NK cells. NK92 cells were stimulated with IL-2, and breast cancer cells were transfected with short hairpin RNA or overexpression construct. NK92 cells were then co-cultured with MDA-MB-231 or MCF-7 cells. As presented in Figures 2F and 2G, silencing of TYMSOS in MDA-MB-231 cells enhanced the cytotoxicity of NK92 cells, along with the increased secretion of IFN-γ and TNF-α. Overexpression of TYMSOS in MCF-7 cells exerted opposite effects on NK cell cytotoxicity and the production of IFN-γ and TNF-α. These data were consistent with previous reports which have demonstrated that NK cells exhibited direct cytotoxicity toward breast cancer cells via IFN-γ production.^17,18 These findings indicate that TYMSOS served as an oncogene and repressed the cytotoxicity of NK92 cells on breast cancer cells.

Figure 2.

TYMSOS promoted the malignant phenotypes of breast cancer cells

(A) The TYMSOS level in MCF10A and breast cancer cells (Hs578T, T47D, MCF-7, MDA-MB-231, and MDA-MB-468 cells) was detected by RT-qPCR.

(B) The TYMSOS level in TYMSOS-knockdown MDA-MB-231 or TYMSOS-overexpressing MCF-7 cells were detected by RT-qPCR.

(C) Cell proliferation of transfected breast cancer cells was assessed by CCK-8 assay.

(D) The colony-forming ability of transfected breast cancer cells was monitored by colony formation assay.

(E) Cell migration and invasion of transfected breast cancer cells were detected by Transwell assays. Scale bar, 100 μm. Activated NK92 cells were co-cultured with transfected breast cancer cells.

(F) Cytotoxicity of NK92 cells was detected by CytoTox96 cytotoxicity assay.

(G) The secreted IFN-γ and TNF-α in cell culture supernatant of co-culture system were detected by ELISA assay. Data were analyzed from at least three independent experiments. Data were presented as mean ± S.D. ∗, p < 0.05, ∗∗, p < 0.01, ∗∗∗, p < 0.001 by Student’s t test or One-way ANOVA.

CBX3 was identified as a downstream effector in TYMSOS-regulated breast cancer progression

Bioinformatics analysis based on RPISeq predicted the potential interaction between TYMSOS and CBX3 (Figure S1A). We thus sought to unravel the role of CBX3 in breast cancer tumors and cells. As presented in Figures 3A and 3B, CBX3 was increased in breast tumors (n = 58), compared with the paired counterparts (n = 58), and it positively correlated with TYMSOS in breast tumors (Figure 3C). Consistent with our findings, analysis from TCGA data showed the upregulation of CBX3 in breast cancer (Figure 3D). In addition, CBX3 was elevated in breast cancer cells at both mRNA and protein levels (Figure 3E). As anticipated, silencing of CBX3 caused a reduction of CBX3 in MDA-MB-231 cells, while CBX3 overexpression induced CBX3 expression in MCF-7 cells (Figure 3F). We next investigated the role of CBX3 in cell proliferation, migration, and invasion. As presented in Figures 3G and 3H, silencing of CBX3 suppressed the growth of MDA-MB-231 cells. By contrast, CBX3 overexpression promoted cell proliferation in MCF-7 cells. Similarly, shCBX3 inhibited cell migration and invasion in MDA-MB-231 cells, while CBX3 overexpression exerted opposite effects in MCF-7 cells (Figure 3I). Moreover, CBX3 knockdown boosted the cytotoxicity of NK92 cells toward MDA-MB-231 cells, whereas the killing effects of NK92 cells toward CBX3-overexpressing MCF-7 cells were impaired (Figure 3J). This was accompanied with the similar tendency of IFN-γ and TNF-α secretion (Figure 3K). Together, these findings suggest that CBX3 was a downstream effector in TYMSOS-regulated breast cancer progression.

Figure 3.

CBX3 was identified as a downstream effector to be involved in TYMSOS-regulated breast cancer progression

(A and B) The mRNA level of CBX3 in breast tumors (n = 58) and their normal counterparts (n = 58) were detected by RT-qPCR.

(C) Pearson correlation analysis between TYMSOS and CBX3 in breast tumors.

(D) Data analyses of CBX3 expression in breast cancer based on UALCAN database (https://ualcan.path.uab.edu/index.html).

(E) The expression of CBX3 in MCF10A and breast cancer cells (Hs578T, T47D, MCF-7, MDA-MB-231, and MDA-MB-468 cells) were detected by RT-qPCR and Western blot, respectively.

(F) The protein level of CBX3 in CBX3-knockdown MDA-MB-231 or CBX3-overexpressing MCF-7 cells was detected by Western blot.

(G) Cell proliferation of transfected breast cancer cells was assessed by CCK-8 assay.

(H) The colony-forming ability of transfected breast cancer cells was monitored by colony formation assay.

(I) Cell migration and invasion of transfected breast cancer cells were detected by Transwell assays. Scale bar, 100 μm. Activated NK92 cells were co-cultured with transfected breast cancer cells.

(J) Cytotoxicity of NK92 cells was detected by CytoTox96 cytotoxicity assay.

(K) The secreted IFN-γ and TNF-α in cell culture supernatant of co-culture system were detected by ELISA assay. Data were analyzed from at least three independent experiments. Data were presented as mean ± S.D. ∗, p < 0.05, ∗∗, p < 0.01, ∗∗∗, p < 0.001 by Student’s t test or One-way ANOVA.

CBX3 facilitated the transcriptional repression of ULBP3

TCGA data and RT-qPCR unequivocally revealed that ULBP3 was remarkably decreased in breast tumors, compared with their normal counterparts (Figures 4A–4C). Intriguingly, silencing of CBX3 upregulated ULBP3 in MDA-MB-231 cells, while CBX3 overexpression downregulated ULBP3 expression in MCF-7 cells (Figure 4D). Bioinformatics analysis based on AnimalTFDB predicted three putative binding sites (BSs) between CBX3 and ULBP3 promoter, including BS1 p(-1677/-1665), BS2 p(-1245/-1232), and BS3 p(-530/-515) (Figure 4E). Chromatin immunoprecipitation (ChIP) assay further showed that the chromatin fragment containing BS3 was successfully enriched by anti-CBX3 antibody in MDA-MB-231 and MCF-7 cells (Figure 4F). In line with this finding, luciferase assay revealed that co-transfection of ULBP3-WT luciferase construct and CBX3 overexpression plasmid resulted in a dramatic reduction of luciferase activity, whereas ULBP3-MUT abolished these effects (Figure 4G). Moreover, silencing of CBX3 impaired the association between CBX3 and ULBP3 promoter, as well as between H3K9me3 and ULBP3 promoter. CBX3 overexpression exerted an opposite effect as detected by ChIP assay (Figure 4H). These findings suggest that CBX3 served as a transcriptional repressor of ULBP3.

Figure 4.

CBX3 mediated the transcriptional repression of ULBP3

(A) Data analyses of ULBP3 expression in breast cancer based on UALCAN database (https://ualcan.path.uab.edu/index.html).

(B and C) The mRNA level of ULPB3 in breast tumors (n = 58) and their normal counterparts (n = 58) were detected by RT-qPCR.

(D) The expression of ULPB3 in CBX3-knockdown MDA-MB-231 or CBX3-overexpressing MCF-7 cells was detected by Western blot.

(E) The putative binding sites between CBX3 and ULBP3 promoter.

(F) The interaction between CBX3 and ULBP3 promoter in breast cancer cells was detected by ChIP assay. Normal IgG served as a negative control.

(G) The interaction between CBX3 and ULBP3 promoter in transfected breast cancer cells was detected by dual-luciferase assays. WT, wild-type; MUT, mutant.

(H) The interaction between CBX3 or H3K9me3 and ULBP3 promoter in transfected breast cancer cells was detected by ChIP assay. Data were analyzed from at least three independent experiments. Data were presented as mean ± S.D. ∗, p < 0.05, ∗∗, p < 0.01, ∗∗∗, p < 0.001 by Student’s t test or One-way ANOVA.

TYMSOS facilitated CBX3-mediated transcriptional repression of ULBP3

We next investigated whether TYMSOS served as an upstream molecule of CBX3/ULBP3 axis. RNA fluorescence in situ hybridization and subcellular fractionation showed that TYMSOS was predominantly expressed in the both cytosol and nucleus of breast cancer cells (Figures S1B-C). Western blot revealed that TYMSOS downregulated CBX3 expression, but upregulated ULPB3 protein level in breast cancer cells (Figure 5A). RNA pull-down and RNA immunoprecipitation (RIP) assays unequivocally showed that TYMSOS bound to CBX3 (Figures 5B and 5C). Gain- and loss-of-function assays further showed that lack of TYMSOS impaired the association between CBX3 and ULBP3 promoter, whereas TYMSOS overexpression potentiated this interaction (Figure 5D). These findings were validated by dual-luciferase assay in which TYMSOS knockdown partially rescued CBX3-decreased luciferase activity, while TYMSOS overexpression exerted an opposite effect (Figure 5E). These data suggest that TYMSOS facilitated CBX3-suppressed ULBP3 in breast cancer cells.

Figure 5.

TYMSOS facilitated CBX3-mediated transcriptional repression of ULBP3

(A–C) The protein levels of CBX3 and ULBP3 in transfected breast cancer cells were detected by Western blot with quantitative analysis. The interaction between TYMSOS and CBX3 in breast cancer cells was detected by RNA pull-down assay (B) and RIP assay (C). Anti-sense TYMSOS (As-TYMSOS) acted as a negative control for RNA pull-down assay. Normal IgG served as a negative control for RIP assay.

(D and E) The interaction between CBX3 and ULBP3 promoter in transfected breast cancer cells was detected by (D) ChIP and (E) luciferase assays. Normal IgG served as a negative control for ChIP assay. WT, wild-type; MUT, mutant. Data were analyzed from at least three independent experiments. Data were presented as mean ± S.D. ∗, p < 0.05, ∗∗, p < 0.01, ∗∗∗, p < 0.001 by Student’s t test or One-way ANOVA.

TYMSOS promoted ubiquitin-proteasomal degradation of ULBP3

Interestingly, the ubiquitination of ULBP3 was predicted by ubibrowser (Figure S1D), and a direct association between TYMSOS and SYVN1 was predicted by RPISeq (Figure S1E). We next sought to test if TYMSOS was implicated in the posttranslational regulation of ULBP3. As presented in Figure 6A, TYMSOS knockdown induced ULBP3 expression, and this induction was potentiated in the presence of proteasome inhibitor MG132. In MCF-7 cells, TYMSOS overexpression decreased ULBP3 expression, while the degradation of ULBP3 was blocked by MG132, indicating that TYMSOS regulated ULBP3 expression via ubiquitin-proteasome pathway. Consistently, TYMSOS promoted ULBP3 turnover in the presence of protein synthesis inhibitor CHX (Figure 6B), confirming that posttranslational modification was responsible for TYMSOS-mediated degradation of ULBP3. MDA-MB-231 or MCF-7 cells were then co-transfected with HA-tagged ubiquitin and shTYMSOS or TYMSOS overexpression plasmid. The immunoprecipitated ubiquitin was detected by anti-ULBP3 antibody for immunoprecipitation (IP) and anti-HA tag antibody for immunoblotting. Co-IP further revealed that silencing of TYMSOS decreased the ubiquitination of ULBP3, whereas TYMSOS overexpression facilitated ULBP3 ubiquitination in the presence of proteasome inhibitor MG132 (Figure 6C). In line with the bioinformatics analysis, RNA pull-down and RIP assays revealed the direct association between TYMSOS and SYVN1 (Figures 6D and 6E). As anticipated, Co-IP showed the direct interaction between SYVN1 and ULBP3 in MDA-MB-231 and MCF-7 cells (Figure 6F), and co-localization of SYVN1 and ULBP3 was also detected by immunofluorescence (Figure 6G). Silencing of TYMSOS decreased SYVN1 expression, but increased ULBP3 expression in MDA-MB-231 cells. Co-IP further showed that the interaction between SYVN1 and ULBP3 was impaired in TYMSOS-knockdown cells. By contrast, overexpression of TYMSOS upregulated SYVN1, alone with the downregulation of ULBP3. A markedly increase of SYVN1, as well as a slight decrease of ULBP3, was observed in the SYVN1-ULBP3 protein complex in MCF-7 cells. The association between SYVN1 and ULBP3 might be enhanced in TYMSOS-overexpressing MCF-7 cells. In addition, the slight downregulation of ULBP3 might be attributed to SYVN1-catalyzed degradation of ULBP3 in the presence of TYMSOS (Figure 6H). Collectively, these findings suggest TYMSOS promoted ULBP3 turnover via ubiquitin-proteasome pathway, and SYVN1 was implicated in this process.

Figure 6.

TYMSOS promoted ubiquitin-proteasomal degradation of ULBP3

(A and B) The protein level of ULBP3 in transfected breast cancer cells was detected by Western blot.

(C) HA-tagged ubiquitin and shTYMSOS/TYMSOS were co-transfected into breast cancer cells. The ubiquitination of ULBP3 in transfected breast cancer cells was assessed by Co-IP.

(D and E)The interaction between TYMSOS and SYVN1 was detected by RNA pull-down assay (D) and RIP assay (E). Anti-sense TYMSOS (As-TYMSOS) and Poly(A) acted as negative controls for RNA pull-down assay. Normal IgG served as a negative control for RIP assay.

(F, H) The association between SYVN1 and ULBP3 was detected by Co-IP. Whole cell lysates and normal IgG served as an input or negative control, respectively.

(G) The co-localization of ULBP3 and SYVN1 in breast cancer cells was detected by IF. Scale bar, 50 μm. Red, SYVN1; Green, ULBP3; Blue, DAPI. Data were analyzed from at least three independent experiments. Data were presented as mean ± S.D. ∗, p < 0.05, ∗∗, p < 0.01, ∗∗∗, p < 0.001 by Student’s t test or One-way ANOVA.

TYMSOS promoted cell growth, migration, invasion, and the killing effect of NK cells via CBX3/ULBP3 or SYVN1/ULBP3 axis

In accordance with TCGA data, RT-qPCR showed elevated expression of SYVN1 in breast tumors (Figures 7A and 7B). SYVN1 was also upregulated in breast cancer cells with similar tendency to TYMSOS (Figure 7C). As expected, silencing or overexpression of SYVN1 decreased or increased its expression, respectively. SYVN1 negatively regulated ULBP3 expression in breast cancer cells (Figure 7D). Functional experiments were next conducted to study the biological functions of CBX3/ULBP3 or SYVN1/ULBP3 axis. CCK-8 and colony formation assays showed that TYMSOS-enhanced cell proliferation was abolished by shCBX3 or shSYVN1 in MCF-7 cells, while silencing of TYMSOS-decreased cell proliferation was reversed by CBX3 or SYVN1 overexpression in MDA-MB-231 (Figures 7E and 7F). Transwell assay further revealed that TYMSOS-facilitated cell migration and invasion were attenuated by CBX3 or SYVN1 knockdown in MCF-7 cells (Figure S2A). TYMSOS knockdown-repressed cell migration and invasion was counteracted by CBX3 or SYVN1 overexpression in MDB-MB-231 cells (Figure S2A). In co-culture system, cytotoxicity assay showed that TYMSOS knockdown enhanced NK cell cytotoxicity, while CBX3 or SYVN1 overexpression exerted a negative effect on NK cell cytotoxicity. Co-transfection of CBX3 or SYVN1 overexpression construct abrogated shTYMSOS-increased NK cell cytotoxicity toward MDA-MB-231 cells. By contrast, TYMSOS-impaired NK cell cytotoxicity was rescued by shCBX3 or shSYVN1 in MCF-7 cells (Figure S2B). In line with these findings, shTYMSOS-induced IFN-γ and TNF-α secretion were attenuated by CBX3 or SYVN1 overexpression in MDA-MB-231 cells. On the contrary, silencing of CBX3 or SYVN1 led to a rebound of TYMSOS-suppressed IFN-γ and TNF-α secretion in MCF-7 cells (Figure S2C). These findings suggest that TYMSOS promoted cell growth, metastasis, and NK cell cytotoxicity via CBX3/ULBP3 or SYVN1/ULBP3 axis.

Figure 7.

TYMSOS promoted cell growth, migration, invasion, and immune escape via CBX3/ULBP3 or SYVN1/ULBP3 axis

(A) Data analyses of SYVN1 expression in breast cancer based on UALCAN database (https://ualcan.path.uab.edu/index.html).

(B) The mRNA level of SYVN1 in breast tumors (n = 58) and their normal counterparts (n = 58) were detected by RT-qPCR.

(C) The protein level of SYVN1 in MCF10A and breast cancer cells was detected by Western blot.

(D) The protein levels of SYVN1 and ULBP3 in transfected MDA-MB-231 and MCF-7 cells were detected by Western blot.

(E) Cell proliferation of transfected breast cancer cells was assessed by CCK-8 assay.

(F) The colony-forming ability of transfected breast cancer cells was monitored by colony formation assay with quantitative analysis. Data were analyzed from at least three independent experiments. Data were presented as mean ± S.D. ∗, p < 0.05, ∗∗, p < 0.01, ∗∗∗, p < 0.001 by Student’s t test or One-way ANOVA.

Depletion of TYMSOS repressed tumor growth and boosted NK cell sensitivity in vivo

To validate the in vitro findings, xenograft models were established. Transfected MDA-MB-231 cells were injected into NICD-SOD mice subcutaneously; activated NK92 cells were injected though tail vein on day 3 and 7 post inoculation. As shown in Figures 8A–8C, silencing of TYMSOS suppressed tumor growth, and injection of NK cells further potentiated this effect as detected by the measurement of tumor volume and weight. Ki-67-positive cells were reduced by TYMSOS knockdown in xenograft tumors, and the reduction was more prominent in shTYMSOS + NK groups as detected by Ki-67 staining (Figure 8D). Flow cytometry further showed that the percentage of CD107a⁺ NK cells was dramatically increased by shTYMSOS in xenograft tumors. As expected, the injection of NK cells further increased the population of NK cells (Figure 8E). Together, these findings indicate that silencing of TYMSOS inhibited tumor growth and boosted NK cell sensitivity in vivo.

Figure 8.

Silencing of TYMSOS repressed tumor growth and boosted NK cell sensitivity in vivo

(A) Photographs of xenograft tumors derived from transfected MDA-MB-231 cells.

(B) Quantitative analysis of tumor volume.

(C) Quantitative analysis of tumor weight.

(D) The immunoreactivity of Ki-67 in xenograft tumors was analyzed by IHC. Scale bar, 50 μm.

(E) Flow cytometry was performed to analyze the percentage of CD107a⁺ NK cells in xenograft tumors. Data were analyzed from at least three independent experiments. Data were presented as mean ± S.D. ∗, p < 0.05, ∗∗, p < 0.01, ∗∗∗, p < 0.001 by Student’s t test or One-way ANOVA.

Discussion

In recent years, immunotherapy with checkpoint inhibitors has emerged as a promising therapeutic strategy for breast cancer.^19,20 Accumulating evidence supports that immune escape is required for breast cancer progression.²⁰ It is of interest to delineate the mechanism underlying immune escape and provide novel insights into the immunotherapy for breast cancer. In this study, we reported that TYMSOS was upregulated in breast cancer tissues and cells. Two regulatory axes responsible for immune escape and metastasis, namely CBX3/ULBP3 and SYVN1/ULBP3, were identified in the current study. These findings shed light on the targeted therapeutic approaches for breast cancer.

TYMSOS, also known as C18orf56, is located on human chromosome 18. In the past decade, emerging evidence illustrates that TYMSOS functions as an oncogene in different cancers, including GC, NSCLC, thyroid carcinoma, and osteosarcoma.^6,7,8,9,10 Several TYMSOS-associated competing endogenous networks have been identified. For instance, FOXM1/TYMSOS/miR-214-3p/NCAPG axis facilitates NSCLC progression by modulating cell proliferation, stemness, migration, and immune cell infiltration.⁷ In addition, lack of TYMSOS suppresses cell growth, migration, invasion, epithelial-mesenchymal transition, or self-renewal capabilities of various cancer cells.^6,7,9,10 More importantly, TCGA data support that TYMSOS is upregulated in breast cancer, and it has been identified as a risk lncRNA in breast cancer.²¹ In accordance with these reports, we found that TYMSOS was markedly increased in breast tumors, especially metastatic breast tumors. High TYMSOS expression conferred unfavorable prognosis in breast cancer patients. Consistently, functional studies further revealed the oncogenic effects of TYMSOS on the proliferation and metastatic properties of breast cancer cells. In addition, it is worth noting that high TYMSOS was also associated with dysregulated immunosuppressive and antitumor cytokines, indicating its potential function in regulating breast cancer cell immune escape. Following studies showed that silencing of TYMSOS increased NK cell cytotoxicity. CD107a is a well-accepted marker for measuring NK cell cytotoxicity.²² Lack of TYMSOS increased the percentage of CD107a⁺ NK cells, suggesting that cytotoxic ability possessed by NK cells was enhanced by TYMSOS knockdown. The findings were further validated by in vivo experiments in which NK cell injection increased the silencing of TYMSOS-boosted NK cell cytotoxicity, thus leading to more prominent suppressive effects on tumor growth. We demonstrated for the first time that TYMSOS was implicated in breast cancer immune escape.

A number of studies have reported the oncogenic roles of CBX3 in various cancers. For instance, CBX3 contributes to cell proliferation and glycolysis through modulating FBP1 in pancreatic cancer.¹⁴ ONCOMINE analysis demonstrated that CBX3 is increased by 2.48-fold in breast tumors.¹⁶ In accordance with this finding, we found that CBX3 was elevated in breast tumors and cells, and it exhibited oncogenic roles in breast cancer cells. Recently, the immunological potential of CBX3 in cancer has gained increasing attentions. In GC, CBX3 expression is negatively associated with the immune checkpoint genes, such as PDCD1 and PDCD1LG2, as well as the responses of immunotherapy.¹⁵ Computational analysis has illustrated that CBX3 expression is negatively correlated with the infiltration of most types of immune cells in invasive breast cancer,¹² supporting the pivotal role of CBX3 in the immune responses in breast cancer. In addition, CBX3 is required for the recognition of H3K9me3, leading to transcriptional repression of CBX3 targets.²³ Consistently, our findings revealed that CBX3 negatively regulated ULBP3 expression at the transcriptional level, possibly in a H3K9me3-dependent manner, thereby modulating NK cell cytotoxicity on breast cancer cells. A recent study has reported that LINC00998 suppresses the progression of glioma via the direct interaction with CBX3.²⁴ Similarly, a direct association between TYMSOS and CBX3 was observed in this study, suggesting that CBX3 acts as an important binding partner of lncRNA in different cancers.

The failure of immune surveillance in cancer leads to the evasion of cancer cells, as well as metastasis.²⁵ NKG2D is an activating receptor on NK cells which are the effectors of immune surveillance, and ULPB1-6, MICA, and MICB are well-known human ligands for NKG2D.²⁶ According to TCGA data, MICA and ULBP3 were downregulated, while MICB and the other ULBPs were upregulated in breast tumors. We focused on ULBP3 due to its remarkable reduction in breast cancer tissues. Bioinformatic analysis through TCGA database has showed that SYVN1 is upregulated in breast tumors, while the biological role of SYVN1 in breast cancer is uninvestigated. In this study, we demonstrated that SYVN1 interacted with ULBP3, and it was required for the TYMSOS-mediated ubiquitination of ULBP3. Our findings identified a posttranslational regulation of ULBP3 in breast cancer cells. Intriguingly, TYMSOS was implicated in the transcriptional and post-translational regulation of ULBP3. We also proved that TYMSOS localized in both cytoplasm and nucleus of breast cancer cells which was similar with its subcellular localization in GC cells.⁶ Future research will explore the specific sites or sequences where TYMSOS binds to CBX3 and SYVN1, while distinguishing the differences in the binding between the nucleus and cytoplasm.

In conclusion, TYMSOS facilitated breast cancer metastasis and immune escape through CBX3-mediated transcriptional regulation of ULBP3 and SYVN1-catalyzed degradation of ULBP3. The CBX3/ULBP3 and SYVN1/ULBP3 axes might be promising targets for breast cancer treatment.

Limitations of study

Two regulatory axes, namely CBX/ULBP3 and SYVN1/ULBP3, have been identified in this study, while the crosstalk between these two axes remains elusive. In the future study, the regulatory mechanisms merit in-depth investigation in the mice model.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
CBX3	Abcam	Cat# ab10480; RRID:AB_1545104
CBX3	Proteintech	Cat# 11650-2-AP; RRID:AB_790137
SYVN1	Abcam	Cat# ab170901; RRID:AB_2479092
ULBP3	Abcam	Cat# ab300102; RRID:AB_2288423
H3K9me3	Abcam	Cat# ab8898; RRID:AB_2927650
H3	Abcam	Cat# ab176842; RRID:AB_371489
Ago2	Abcam	Cat# ab32381; RRID:AB_2784637
IgG	Abcam	Cat# ab109489; RRID:AB_1068097
HA tag	Abcam	Cat# ab9110; RRID:AB_10887802
CD107a	Abcam	Cat# ab25630; RRID:AB_1543789
Ki67	Abcam	Cat# ab15580; RRID:AB_854581
β-actin	Abcam	Cat# ab8227; RRID:AB_2750915
Bacterial and virus strains
pCDH-CMV-MCS-EF1-Puro	YouBio	Cat# VT1480
Biological samples
breast tumors and matched normal counterparts from breast cancer patients	Xiangya Hospital of Central South University	N/A
Chemicals, peptides, and recombinant proteins
Cycloheximide	MedChemExpress	Cat# HY-12320
IL-2	Gibco	Cat# PHC0023
Lipofectamine 3000 reagent	Invitrogen	Cat# L3000150
ECL	Beyotime	Cat# P0018S
Critical commercial assays
IFN gamma Human ELISA kit	Invitrogen	Cat# BMS228
TNF alpha Human ELISA kit	Invitrogen	Cat# BMS223-4
PARIS kit	Invitrogen	Cat# AM1921
Magna RIP Kit	Millipore	Cat# 17-701
RNA Pull-down Kit	Pierce	Cat# 20164
ChIP Assay Kit	Pierce	Cat# 26159
Deposited data
UALCAN	This paper	Figures 3D/4A/7A; https://ualcan.path.uab.edu/index.html
AnimalTFDB	This paper	Figure 5E; http://bioinfo.life.hust.edu.cn/AnimalTFDB/#!/
ubibrowser	This paper	Figure S1D; http://ubibrowser.bio-it.cn/ubibrowser/home/index
RPISeq	This paper	Figures S1A/S1E; http://pridb.gdcb.iastate.edu/RPISeq/
Experimental models: Cell lines
MCF10A	ATCC	Cat# CRL-10317; RRID:CVCL_0598
MCF-7	ATCC	Cat# HTB-22; RRID:CVCL_0031
Hs578T	ATCC	Cat# HTB-126; RRID:CVCL_0332
T47D	ATCC	Cat# HTB-133; RRID:CVCL_0553
MDA-MB-231	ATCC	Cat# HTB-26; RRID:CVCL_0062
MDA-MB-468	ATCC	Cat# HTB-132; RRID:CVCL_0419
NK92	ATCC	Cat# CRL-2407; RRID:CVCL_2142
Experimental models: Organisms/strains
Female NOD-SCID mice	Hunan SJA laboratory animal Co., Ltd	Cat# hnslkjd008
Oligonucleotides
shRNA targeting sequence: TYMSOS#1: 5′-CACCGGAACACT GAGGCCTAGAATGCGAACATTCTAGGCCTCAGTGTTCC-3′	This paper	N/A
shRNA targeting sequence: TYMSOS#2: 5'- CACCGCCTAGA ATGGTTGAGGTACTCGAAAGTACCTCAACCATTCTAGGC-3'	This paper	N/A
shRNA targeting sequence: CBX3#1: 5'- CACCGCTGCTGAC AAACCAAGAGGACGAATCCTCTTGGTTTGTCAGCAGC-3'	This paper	N/A
shRNA targeting sequence: CBX3#2: 5′- CACCGCTGACAAA CCAAGAGGATTTCGAAAAATCCTCTTGGTTTGTCAGC-3′	This paper	N/A
shRNA targeting sequence: SYVN1#1: 5'- CACCGCTGCTGAC AAACCAAGAGGACGAATCCTCTTGGTTTGTCAGCAGC-3'	This paper	N/A
shRNA targeting sequence: SYVN1#2: 5′- CACCGCTGACAAA CCAAGAGGATTTCGAAAAATCCT CTTGGTTTGTCAGC -3′	This paper	N/A
BS1(ChIP): TCAGCTGACTTCC	This paper	N/A
BS2(ChIP): GCCCCTCCTGCACA	This paper	N/A
BS3(ChIP): AGTGGAGAGGCAAAGA	This paper	N/A
Primers for TYMSOS, see Table S2	This paper	N/A
Primers for TGF-β, see Table S2	This paper	N/A
Primers for IL-10, see Table S2	This paper	N/A
Primers for IL-1β, see Table S2	This paper	N/A
Primers for IFN-γ, see Table S2	This paper	N/A
Primers for TNF-α, see Table S2	This paper	N/A
Primers for ULBP3, see Table S2	This paper	N/A
Primers for GAPDH, see Table S2	This paper	N/A
Recombinant DNA
Plasmid: pGL3 vector	YouBio	Cat# VT1554; RRID:Addgene_48743
Plasmid: pcDNA3.1	YouBio	Cat# VT1001; RRID:Addgene_200458
Dual-Luciferase Assay System	Invitrogen	Cat# T1033
Software and algorithms
ImageJ	Schneider et al.7	https://imagej.nih.gov/ij/
microplate reader	Bio-Rad
confocal microscopy	Nikon

Resource availability

Lead contact

Further information and request for resources and reagents should be directed to and will be fulfilled by the lead contact, Lei Guo (guolei436@163.com).

Materials availability

This study did not generate new unique reagents.

Experimental model and study participant details

This study was approved by the Xiangya Hospital of Central South University (ethics number: 202109924). Female patients with breast cancer (n = 58) were recruited in this study. Written consents were obtained from all participants. The clinicopathological characteristics of breast cancer patients, including age, menopause, tumor size, lymph node metastasis and TNM stage, were listed in Table S1.

All methods and protocols were performed in accordance with Practical Guide for Animal Experiment Management and Operation. Female NOD-SCID mice (4 ∼ 6-wk-old) were used in this study. All mice have free access to standard chow and water under controlled temperature and constant 12h/12h light/dark cycles.

Human mammary epithelial cell line MCF10A and human breast cancer cell lines MCF-7, Hs578T, T47D, MDA-MB-231, MDA-MB-468 cells and human NK cell line NK92 cells were from ATCC (Manassas, VA, USA). MCF10A cells were cultured in F12/DMEM (Gibco, Grand Island, NY, USA). The breast cancer cells were grown in DMEM (Gibco) containing 10% FBS. NK92 cells were cultured in MEMα containing 12.5% FBS, 2 mM L-glutamine and 12.5% horse serum (Gibco). For NK92 cell activation, cells were stimulated with 100 U/mL IL-2 (Gibco, PHC0023) for 24 h. For co-culture, the activated NK-92 cells were co-cultured with MCF-7 or MDA-MB-231 cells at a ratio of 10:1 for 4 h. All cells were maintained at 37°C/5% CO₂.

Method details

Clinical specimens

The breast tumors (n = 58) and matched normal counterparts (n = 58) (female: 58, male: 0; Age(<50 years): 33, Age(≥50 years): 25) were collected from breast cancer patients at the Xiangya Hospital of Central South University. This study was approved by the Xiangya Hospital of Central South University (ethics number: 202109924). Written consents were obtained from all participants. The clinicopathological characteristics of breast cancer patients were listed in Table S1. Patients met the diagnostic criteria of breast cancer and the clinical TNM staging criteria, and there was no surgery history of the patients within the latest half-year. Patients who had taken drugs affecting the results or had severe dysfunction in lung, heart and kidney were excluded from this study.

Cell culture and transfection

Human mammary epithelial cell line MCF10A and human breast cancer cell lines MCF-7, Hs578T, T47D, MDA-MB-231, MDA-MB-468 cells and human NK cell line NK92 cells were from ATCC (Manassas, VA, USA). MCF10A cells were cultured in F12/DMEM (Gibco, Grand Island, NY, USA). The breast cancer cells were grown in DMEM (Gibco) containing 10% FBS. NK92 cells were cultured in MEMα containing 12.5% FBS, 2 mM L-glutamine and 12.5% horse serum (Gibco). For NK92 cell activation, cells were stimulated with 100 U/mL IL-2 (Gibco, PHC0023) for 24 h. For co-culture, the activated NK-92 cells were co-cultured with MCF-7 or MDA-MB-231 cells at a ratio of 10:1 for 4 h. All cells were maintained at 37°C/5% CO₂. shTYMSOS-1(5′-CACCGGAACACTGAGGCCTAGAATGCGAACATTCTAGGCCTCAGTGTTCC -3'), shTYMSOS-2(5'- CACCGCCTAGAATGGTTGAGGTACTCGAAAGTACCTCAACCATTCTAGGC-3'); shCBX3-1 (5'- CACCGCTGCTGACAAACCAAG

AGGACGAATCCTCTTGGTTTGTCAGCAGC -3'), shCBX3-2 (5′- CACCGCTGA

CAAACCAAGAGGATTTCGAAAAATCCTCTTGGTTTGTCAGC-3'), shSYVN1-1 (5′- CACCGCTGCTGACAAACCAAGAGGACGAATCCTCTTGGTTTGTCAGCA

GC -3', shSYVN1-2(5′- CACCGCTGACAAACCAAGAGGATTTCGAAAAATCCT

CTTGGTTTGTCAGC -3′) were purchased from Invitrogen (Carlsbad, CA, USA). TYMSOS, CBX3 or SYVN1 were constructed into pcDNA3.1 (YouBio, VT1001). Breast cancer cells were transfected using Lipofectamine 3000 reagent (Invitrogen).

RT-qPCR

Total RNA was isolated using Trizol (Invitrogen), and cDNA was synthesized using Advantage RT-PCR Kit (TaKaRa, Dalian, China). The target gene expression was analyzed with iQTM SYBR Green Supermix (Bio-Rad, Hercules, CA, USA) and calculated using 2 ^–ΔΔCt method. The primer sequences were listed in the Table S2.

Cell counting Kit-8 (CCK-8) assay

Cells (2×10³ cells/mL) were plated into 96-well plates. At the designated time points, CCK-8 solution (20 μL, Beyotime) was added into each well, and incubated at 37°C for 1 h. A450 was determined using a microplate reader (Bio-Rad).

Colony formation assay

Cells (5×10² cells/well) were plated into 6-well plate. After 14 d, fixed cells were stained with crystal violet. The colonies (> 50 cells) were photographed and counted.

Transwell assay

Transfected cells (2×10⁴ cells/well) were plated in the Matrigel (Corning, Corning, NY, USA)-coated top chamber and cultured in serum-free medium, while the complete medium was filled in the bottom chamber. After 1 d, the invaded cells were fixed and stained with crystal violet. The migration assay was performed using similar approach without Matrigel coating.

Immunofluorescence (IF)

Cells grown on the coverslips were fixed and permeabilized. After blocking with 1% BSA, cells were incubated with primary antibody and Alexa Fluor-conjugated secondary antibody. Nucleus was detected by DAPI. Images were acquired using confocal microscopy (Nikon, Tokyo, Japan).

Western blot and protein stability assay

Protein extracts were prepared in RIPA lysis buffer (Pierce, Rockford, IL, USA). After SDS-PAGE separation, proteins were transferred onto NC membrane, and blocked with 5% BSA. The blot was then incubated with primary antibody (Table S3), and followed by the incubation with corresponding secondary antibody. The signal was detected using ECL substrates (Beyotime). For protein stability assay, cells were incubated with Cycloheximide (CHX, 2 μg/mL, MedChemExpress) for 0, 2, 4 and 8 h prior to Western blot analysis.

Co-immunoprecipitation (Co-IP)

Protein lysates were extracted using IP lysis buffer (Pierce). Cell lysates (1000 μg) were incubated with primary antibody or normal IgG (Table S3) at 4°C overnight. The protein complexes were then enriched using Protein A/G agarose (Pierce) at 4°C for 4 h, and analyzed by western blot. Whole cell lysates were used as an input control, and normal IgG acted as a negative control.

Cytotoxicity assay

NK cell cytotoxicity was assessed to detect the level of lactate dehydrogenase (LDH) using CytoTox96 Non-Radioactive Cytotoxicity Assay Kit (Promega, Madison, WI, USA). Briefly, cells were (2×10⁵ cells/mL) plated into 96-well plate. 50 μL CytoTox96 Reagent was added into each well, and incubated for 30 min at room temperature, followed by the addition of Stop Solution (50 μL). A490 was determined using a microplate reader (Bio-Rad). Percent cytotoxicity was calculated as instructed.

ELISA assay

The IFN-γ and TNF-α levels in cell culture medium were quantified using ELISA kits (BMS228 and BMS223-4, Invitrogen). Briefly, the culture medium was collected and centrifugated at 1400 rpm for 1 min. ELISA assay was performed as instructed, and A450 was determined using a microplate reader (Bio-Rad).

Flow cytometry

NK92 cells were stained with anti-CD107a antibody-PE (Abcam) at 4°C for 30 min. For cell apoptosis assay, transfected cells were incubated with Annexin V-FITC and PI (Invitrogen) for 15 min. The stained cells were analyzed using flow cytometer (BD Biosciences, San Jose, CA, USA).

Fluorescence in situ hybridization (FISH)

PKH67-labeled TYMSOS was designed and synthesized by Ribobio (Guangzhou, China). Cells were fixed and permeabilized, and incubated with FISH probe, followed by DAPI staining. Images were acquired using confocal microscopy (Nikon).

Subcellular fractionation

The nuclear and cytoplasmic RNAs were isolated using PARIS kit (Invitrogen). Briefly, Cells were lysed with cell fractionation buffer. After centrifugation, the supernatants (cytoplasmic lysates) were collected, and the pellets (nuclear lysates) were then lysed with cell disruption buffer. RT-qPCR was then conducted to detect the TYMSOS level. GAPDH was used as the cytoplasmic or nuclear marker, respectively.

RNA immunoprecipitation (RIP) assay

RIP assay was conducted using Magna RIP Kit (Millipore, Billerica, MA, USA). In brief, MDA-MB-231 and MCF-7 cells were lysed using RIP lysis buffer. Anti-CBX3 (ab10480, Abcam) antibody or normal IgG conjugated beads were incubated with cell lysates. The immunoprecipitated TYMSOS was detected by RT-qPCR.

RNA pull-down assay

RNA pull-down assay was conducted using RNA Pull-down Kit (Pierce, #20164). Briefly biotin-labeled TYMSOS or antisense TYMSOS (As-TYMSOS) was conjugated to streptavidin beads, followed by the incubation with cell lysates. The complexes were then eluted and subjected to western blot analysis.

Chromatin immunoprecipitation (ChIP) assay

ChIP assay was carried out using Pierce ChIP Assay Kit (Pierce, #26159). In brief, cells were crosslinked with 1% formaldehyde. The lysed cells were then digested by MNase. CBX3, H3K9me3 antibody (Table S3) or normal IgG conjugated beads were incubated with the chromatin fractions. After purification, the immunoprecipitated DNA was analyzed by RT-qPCR.

Dual-luciferase reporter assay

The wild-type or mutated promoter regions BS1 (-1677/-1665), BS2 (-1245/-1232) and BS3 (-530/-515) of ULBP3 were constructed into pGL3 vector (Youbio, VT1554). Cells were co-transfected with shTYMSOS or TYMSOS overexpression construct, CBX3 and luciferase constructs. Luciferase activity was measured using Dual-Luciferase Assay System (Invitrogen, T1033).

Animal study

Female NOD-SCID mice (4 ∼ 6-wk-old, n=6 per group) were purchased from Hunan SJA laboratory animal Co., Ltd. This study was approved by the Xiangya Hospital of Central South University (ethics number: 202109924). The transfected MDA-MB-231 cells were subcutaneously injected into the flank of mice. Tumor size was monitored every 5 days, and calculated with the formula: Volume= 1/2 × length ×width². IL-2 activated NK92 cells were injected though tail vein on day 3 and 7 post-inoculation. On day 28, the xenograft tumors were harvested and weighed.

Immunohistochemistry (IHC)

The deparaffinized sections were subjected to antigen retrieval. After blocking with 1% BSA, the slides were incubated with Ki67 antibody (ab15580), followed the incubation with HRP-conjugated secondary antibody. Signal was detected using the DAB color development kit (Beyotime).

Quantification and statistical analysis

Statistics

Data were presented as mean ± S.D. and analyzed using GraphPad Prism software 8.0 (GraphPad, La Jolla, CA, USA). One-way ANOVA or Student’s t-test was performed to assess the differences. Kaplan-Meier plot was used for survival analysis. The correlation was performed by Pearson correlation analysis. P<0.05 was considered statistically significant. ∗, P < 0.05, ∗∗, P < 0.01, ∗∗∗, P < 0.001. The statistical details can be found in figure legends.

Published: August 7, 2023

Data and code availability

• •All data reported in this paper will be shared by the lead contact upon request.
• •This paper does not report original code.
• •Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.