LATS1/2–Integrin Axis Confers Tumor‐Generated Forces That Activate Neighboring Fibroblasts

ABSTRACT

Tumors generate various forces during growth and progression, which in turn promote tumor development. Although fibroblasts are considered the primary force generators in the tumor microenvironment, recent studies have shown that cancer cells also generate considerable tensile forces. However, the roles that these forces play in the tumor microenvironment and the pathways regulating this process remain largely unknown. Here, we demonstrated that the Hippo pathway‐associated kinases, LATS1/2, in cancer cells are essential for collective force generation and fibroblast activation via extracellular matrix‐mediated cell–cell interactions. In murine breast cancer 4 T1 spheroids, the deletion of LATS1/2 dampened force generation and disrupted reorganization of the surrounding collagen matrix. LATS1/2‐mediated mechanical forces of tumors are required for fibroblast activation and differentiation into mechanoresponsive fibroblasts. Mechanistically, LATS1/2 regulate tumor force generation through the expression of collagen receptor integrins. Our findings not only identify the Hippo pathway as a critical regulator of tumor force generation but also suggest potential strategies for targeting it in cancer therapy from a mechanobiological perspective, offering new avenues in the fight against cancer.

This study shows that LATS1/2 kinases in cancer cells regulate mechanical forces and fibroblast activation in tumors. Deleting LATS1/2 disrupts force generation and collagen matrix organization, highlighting the Hippo pathway as a potential target for cancer therapy.

Abbreviations

CAF

cancer‐associated fibroblast

DIC

digital image correlation

ECM

extracellular matrix

ITGA2

integrin subunit alpha 2

LATS1/2

large tumor suppressor 1 and 2 kinases

TACS

tumor‐associated collagen signature

TAZ

transcriptional coactivator with PDZ‐binding motif

TGF‐β

transforming growth factor‐β

TME

tumor microenvironment

TNBC

triple‐negative breast cancer

YAP

yes‐associated protein

1. Introduction

Tumors generate mechanical forces during their growth and progression. These forces, which result from the abnormal behavior of residents in the tumor microenvironment (TME), play crucial roles in various pathological processes, including tumor invasion of neighboring tissues [1]. During the invasion of certain types of cancer, such as breast cancer, cancer cells align with and exert a traction force on the surrounding extracellular matrix (ECM). Each cell, not only the cells at the leading edge, coordinates its movement by actively interacting with its neighbors to generate a collective force [2]. Such forces can be potent enough to break down the basement membrane and pave the way for metastasis [3]. In addition to cell–ECM interactions, mechanical forces have been implicated in mediating cell–ECM–cell communication. For instance, force‐mediated myofibroblast–fibroblast interactions via the fibrous matrix enable long‐range mechanotransduction during fibrosis expansion [4]. However, the effects of tumor‐generated forces on other cell types in the TME remain largely unexplored. Unlike the force generated by a single cell, collective force generation requires a more delicate arrangement of multiple biological processes, including cell–cell and cell–ECM interactions, as well as force transmission [5]. The pathway that coordinates these processes has yet to be elucidated.

The Hippo pathway has been implicated as a regulator of cellular mechano‐sensing, ‐transmission, and ‐response [6]. Its molecular functions involve a kinase cascade primarily centered around the large tumor suppressor 1 and 2 kinases (LATS1/2). The main downstream effectors of the Hippo pathway are the yes‐associated protein (YAP) and transcriptional coactivator with PDZ‐binding motif (TAZ; also known as WWTR1). YAP and TAZ act as transcriptional coactivators that shuttle between the cytoplasm and nucleus. In the nucleus, they interact with TEA domain transcription factors 1–4 (TEAD1–4) to induce the expression of a broad range of genes. When the Hippo pathway is activated, LATS1/2 phosphorylate YAP and TAZ, promoting their cytoplasmic retention and inhibiting their transcriptional activities [7]. The Hippo pathway serves as a signaling hub that integrates information from various environmental cues and plays critical roles in numerous biological processes, including development, regeneration, and cancer biology [8, 9, 10]. Although the mechanisms by which cells sense and respond to biochemical and mechanical cues through the Hippo pathway have been extensively studied, how the Hippo pathway mediates environmental remodeling remains poorly understood, particularly under three‐dimensional (3D) multicellular conditions. Recent studies have suggested that the Hippo pathway may be involved in organizing multicellular behaviors because its disruption alters the expression patterns of proteins involved in focal adhesions and cell–cell junctions [11, 12]. However, whether the Hippo pathway plays a role in the tumor‐generated collective force requires further investigation.

In this study, using an in vitro 3D culture system and an animal model, we revealed that the Hippo pathway is necessary for the tumor‐generated collective force, which promotes fibroblast activation and immunosuppression. Mechanistically, disruption of the Hippo pathway dampens the tumor‐generated collective force, partially by decreasing the expression of integrins responsible for collagen‐ and laminin‐binding. Our data implicate a critical role for the tumor‐generated collective force in modulating the TME and could inform strategies for targeting mechanical forces in cancer therapeutics.

2. Material and Methods

Wild‐type (WT) or LATS1/2 double‐knockout (dKO) 4 T1 breast cancer spheroids generated using the hanging drop method were embedded into the collagen type I gel mixed with fluorescent microbeads. Deformation of the collagen matrix was tracked by time‐lapse imaging and quantified using digital image correlation (DIC) analysis. NIH3T3 a‐SMA reporter cells were generated and co‐cultured with 4 T1 spheroids or conditioned medium to investigate the effects of cancer spheroid‐generated force on fibroblast activation in 3D collagen gels. Then, 4 T1 spheroids were subjected to RNA sequencing and analysis to find the potential molecules related to the mechanism of LATS1/2 in tumor‐generated force. ITGA2, identified in data analysis, was overexpressed in 4 T1 cells. The western blotting analysis, DIC analysis, and orthotopic transplantation mouse model were then conducted to validate the findings in RNA sequencing analysis. Bioinformatics analysis was performed on single‐cell RNA sequencing from 4 T1 tumors and human breast cancer to reveal the in vivo and clinical relevance of our findings. Detailed information can be found in Data S1. Antibodies, reagents, plasmids, and oligos used in this paper are listed in Tables S1–, S4.

3. Results

3.1. Loss of LATS1/2 Compromises the Tumor‐Generated Collective Force

To elucidate the role of the Hippo pathway in modulating tumor‐generated force under 3D multicellular conditions, we utilized an in vitro 3D culture system; 4 T1 triple‐negative breast cancer (TNBC) spheroids were generated using the hanging drop method [13]. After 96 h of culture, the spheroids were collected and embedded in a collagen type I gel mixed with fluorescent microbeads to facilitate tracking (Figure 1A). Ongoing spheroid force‐induced deformation of the collagen matrix was tracked using digital image correlation analysis [14], which compares digital images of the fluorescent microbeads before and after deformation. The radial displacement (U _r) and circumferential displacement (U _θ) vectors were quantified and then combined into vector U to better represent the general effects that the spheroid‐generated force has on the surrounding collagen matrix (Figure 1B, Movie S1). WT 4 T1 spheroids readily aligned with the surrounding collagen matrix and induced obvious collagen deformation before invading the gel matrix, as evidenced by the microbead displacements observed (Movie S2). Thus, multicellular tumor spheroids can serve as model systems for investigating the mechanics of cancer invasion, including collective cellular force generation and ECM remodeling. Previous studies have shown that deletion of LATS1/2 nearly abolishes YAP/TAZ regulation via the Hippo pathway, leading to the hyperactivation of YAP/TAZ [15]. Therefore, we used LATS1/2 dKO 4 T1 cells to investigate the role that the Hippo pathway plays in modulating the tumor‐generated force. Deletion of LATS1/2 was confirmed by the loss of their protein expression and impaired YAP/TAZ phosphorylation, as evidenced in our previous study [16]. Consistent with previous findings that deletion of LATS1/2 promotes anchorage‐independent tumor growth [16], LATS1/2 dKO 4 T1 spheroids were larger in size compared to WT spheroids (Figure S1A,B). While the migration abilities of WT and LATS1/2 dKO 4 T1 cells were comparable under two‐dimensional (2D) culture conditions (Figure S2A), LATS1/2 dKO 4 T1 spheroids (Movie S3) were less invasive and exhibited limited collagen deformation under 3D culture conditions compared with that of WT spheroids (Movie S2). To compare the influence of WT and LATS1/2 dKO spheroids on ECM deformation, microbead displacement (U) was quantified at various time points after embedding spheroids into the collagen gel and normalized by dividing by the radius of the spheroids (R), due to the size differences observed between the WT and LATS1/2 dKO spheroids (Figure 1C,D, Movies S4 and S5). To statistically compare collagen deformation over time, the normalized displacement (U/R) was fitted into the formula, U/R = A(1 − exp(−t/T)), where constants A and T were calculated. Constant A represents the magnitude of the deformation and was used to assess the force generated by the spheroids. While the tensile forces induced by LATS1/2 dKO spheroids developed more slowly and were weaker, those induced by WT spheroids were rapid and strong. In both cases, the tensile forces reached equilibrium at approximately 8 h. After achieving equilibrium, the spheroids maintained the tensile force, with WT spheroids generating significantly higher tensile force compared to LATS1/2 dKO spheroids (Figure 1E,F). These data indicate that the Hippo pathway is necessary for collective tumor force generation in vitro.

FIGURE 1.

Loss of LATS1/2 abolishes the tumor‐generated collective force. (A) 3D culture system used for tracking cancer spheroid‐induced deformation of the collagen matrix. Fluorescent microbeads were mixed into the collagen gel to facilitate tracking. Time‐lapse imaging using confocal microscopy was performed to monitor ECM displacement. (B) Representative time‐lapse images obtained using confocal microscopy. Scale bar = 100 μm. U is the combined vector of ECM displacement in the radial (U _r) and circumferential (U _θ) directions. (C) Color map of ECM displacement induced by WT or LATS1/2 dKO 4 T1 spheroids in the 3D culture system. Colors represent the level of displacement (green: Low; red: High). The white area represents the cancer spheroid. (D) Quantification of ECM displacement induced by WT or LATS1/2 dKO 4 T1 spheroids at each time point. The displacement (U) was normalized by dividing the radius (R) of the spheroids. All ECM displacements quantified at the same time point were averaged and plotted. U: Displacement. R: Radius of spheroids. U/R: Normalized displacement. (E) The normalized displacement U/R was fitted to time t using the equation U/R = A(1 − exp.(−t/T)), and constant A was calculated. A represents the magnitude of displacement and was used to assess the force generated by the spheroids. (F) Comparison of constant A between the WT and LATS1/2 dKO samples. Data are presented as means ± SD (n = 3 independent experiments). **p < 0.01; two‐tailed unpaired t‐test.

3.2. LATS1/2 in Cancer Cells Promote Neighboring Fibroblast Activation Through Cell–Extracellular Matrix–Cell Communication

Next, we investigate the effects of the tumor‐generated force on neighboring cells. The TME is a complex ecosystem comprising cancer and various noncancerous cells, including diverse immune cell types, endothelial cells, and fibroblasts [17]. Fibroblasts in tumors, termed cancer‐associated fibroblasts (CAFs), are perpetually activated and acquire features similar to those of myofibroblasts, which are temporarily activated during wound healing. CAFs play a primary role in the remodeling of the TME through ECM deposition and contraction, altering the mechanical properties of the ECM and affecting the behavior of both cancer and immune cells [18]. Although CAFs are critical to the TME, the mechanisms underlying fibroblast recruitment and activation in this environment remain poorly understood. Previous studies have reported that mechanical stimuli are required for complete fibroblast activation [19]. Therefore, we focused on the effects of tumor‐generated force on fibroblast activation.

To investigate the interactions between cancer cells and fibroblasts, we co‐cultured NIH3T3 fibroblasts with 4 T1 spheroids in 3D collagen gels. NIH3T3 cells co‐cultured with WT spheroids, but not with LATS1/2 dKO spheroids, acquired an elongated morphology shortly after co‐culture initiation (Figure 2A–C, Movies S6–, S8), indicative of their activation under mechanical stress. To better monitor fibroblast activation in the 3D collagen gel, we generated NIH3T3 reporter cells in which the mNeonGreen fluorescent protein is expressed under control of the Acta2 (commonly referred to as alpha‐smooth muscle actin [α‐SMA]) promoter by inserting a 3 × mNeonGreen cassette into one allele of the endogenous Acta2 locus using CRISPR (clustered regularly interspaced short palindromic repeats)/Cas9 genome‐editing technology (Figure 3A). Expression of α‐SMA is one of the most widely used markers for fibroblast activation. We confirmed the insertion of the 3 × mNeonGreen cassette into the endogenous Acta2 locus using genomic PCR analysis (Figure S3A). We also verified that this reporter system accurately represents fibroblast activation both in 2D and 3D culture conditions, as stimulation with transforming growth factor‐β (TGF‐β), which stimulates α‐SMA expression, induced the expression of mNeonGreen in NIH3T3 reporter cells (Figure S3B, Figure 3B). Notably, co‐culture of these reporter cells with WT spheroids, but not with LATS1/2 dKO spheroids, promoted mNeonGreen expression, indicating that the WT tumor spheroids induced fibroblast activation (Figure 3C). In contrast, treatment with conditioned medium derived from WT spheroids failed to induce mNeonGreen expression in NIH3T3 reporter cells (Figure S4A,B), suggesting that physical cues, rather than soluble factors, from 4 T1 spheroids may play key roles in driving the activation of co‐cultured NIH3T3 cells. Together, our data suggest that the LATS1/2‐dependent tumor‐generated force promotes fibroblast activation through ECM‐mediated tumor cell–fibroblast interactions.

FIGURE 2.

LATS1/2 in cancer cells promote fibroblast morphological change in 3D collagen gel. (A) NIH3T3 cells co‐cultured with WT 4 T1 spheroids in 3D collagen gel. (B) NIH3T3 cells co‐cultured with LATS1/2 dKO 4 T1 spheroids in 3D collagen gel. (C) NIH3T3 cells embedded in 3D collagen gel without 4 T1 spheroids.

FIGURE 3.

LATS1/2 in cancer cells promote fibroblast activation in vitro. (A) Knock‐in strategy used for generating NIH3T3 α‐SMA (ACTA2) reporter cells. CRISPR/Cas9 was used to create a double‐stranded DNA break near the start codon of the endogenous Acta2 locus. A targeting vector encoding the mNeonGreen reporter and puromycin‐resistant (PuroR) proteins was used as the template for homologous recombination (HR). pA, Poly A. (B) NIH3T3 reporter cells in the 3D collagen gel were treated with TGF‐β1 (20 ng/mL) for 24 h. Representative confocal microscopy images (left). Quantification data (right). Data are presented as means ± SD (n = 30 cells per group). ****p < 0.0001; two‐tailed unpaired t‐test. (C) NIH3T3 reporter cells (green) were co‐cultured with 4 T1 spheroids (red) in the collagen gel for 24 h. Representative images (left). Quantification data (right). Data are presented as means ± SD (n = 54 cells per group). ****p < 0.0001; two‐tailed unpaired t‐test.

3.3. LATS1/2 Deletion in Cancer Cells Hinders Mechanoresponsive Cancer‐Associated Fibroblast Differentiation In Vivo

It is generally accepted that the Hippo pathway inhibits cellular growth, serving as a tumor suppressor. However, emerging evidence suggests that it plays a more complex role in tumor progression [20]. Our previous study revealed that LATS1/2 deletion in cancer cells increased cancer cell proliferation while simultaneously enhancing tumor immunogenicity, leading to tumor destruction by boosting anticancer immunity in vivo [16]. A subsequent study using a single‐cell RNA sequencing (scRNA‐seq) approach revealed that LATS1/2 deletion in 4 T1 breast tumors altered the proportion of CAFs in the TME, contributing to attenuation of the immunosuppressive environment [21]. However, the mechanism by which LATS1/2 deletion in cancer cells reshapes CAF composition remains unknown. Therefore, we reanalyzed the scRNA‐seq data [21] to investigate the link between the tumor‐generated force and CAF proportions in vivo.

To this end, either WT or LATS1/2 dKO 4 T1 breast tumors were collected for tumor dissociation, followed by stromal cell enrichment using a cell sorter with selection for CD45⁻EpCAM⁻CD31⁻ stromal cells [21]. The resulting stromal cell‐enriched samples were processed using 10× Chromium to profile single‐cell transcriptomes and subjected to data analysis. As we had observed a certain amount of cancer cell contamination in the stromal cluster, CAFs were subclustered based on the specific expression of Lumican (Lum) and Decorin (Dcn) (Figure S5A). Six clusters with distinct gene signatures were identified in our CAF population. Clusters 1 and 2, characterized by high expression of activated myofibroblast markers, such as α‐SMA, decreased markedly in LATS1/2‐deficient tumors (Figure 4A,B). To better understand the characteristics of clusters 1 and 2, we reanalyzed a published breast cancer dataset as a reference to annotate our CAF population. The CAFs were manually selected from the total cell population and annotated according to the criteria described in the previous study [22] as three super‐clusters: Mechanoresponsive (MR), immunomodulatory (IM), and steady‐state‐like (SSL) CAFs (Figure S6A). Integration of our data with reference data using Seurat's label transfer algorithm [23] revealed that clusters 1 and 2 in our dataset corresponded to the MR CAF subset, which demonstrates elevated expression of mechanosensitive signaling mediators and ECM components [22] (Figure 4C). This suggests that the reduction of clusters 1 and 2 CAFs in LATS1/2‐deficient tumors may be associated with the reduced tumor‐generated force in LATS1/2 dKO cancer cells.

FIGURE 4.

LATS1/2 deletion in cancer cells reshapes the composition of the cancer‐associated fibroblast population in vivo. (A) WT or LATS1/2 dKO 4 T1 breast tumors collected were subjected to scRNA‐seq and data analyses. A uniform manifold approximation and projection (UMAP) plot of fibroblasts colored by cluster membership (n = 18 tumors per group) (left). The ratio of each cluster in WT or LATS1/2 dKO 4 T1 breast tumors (right). (B) Relative average expression of indicated marker genes in clusters from the UMAP shown in (A). (C) Label transfer projection of the CAF clusters obtained from Foster et al. [22] (left) onto CAF clusters from 4 T1 tumors (right). Clusters were annotated as three superclusters [mechanoresponsive (MR), immunomodulatory (IM), and steady‐state‐like (SSL)] as indicated in the figure panel. (D) UMAP showing the global relationship between CAF clusters in 4 T1 tumors inferred by Slingshot. (E) UMAP showing that the lineage started in cluster 0 and ended in cluster 1. The gray color indicates that the cells were not related to this lineage. The initial cluster was determined using Slingshot without prior knowledge. (F) Relative average expression of the indicated marker genes in clusters from the lineage shown in (E). (G) Heatmap showing the average expression of the top 30 genes identified using Tradeseq whose changes in expression were associated with pseudotime. (H) KEGG enrichment analysis was performed on the top 300 genes identified using Tradeseq, with expression changes associated with pseudotime.

To better understand this, we performed a trajectory inference on our CAF data using Slingshot [24] to reveal the global relationship between the CAF clusters (Figure 4D). Cluster 0, annotated as SSL CAFs, was identified as the origin of the other clusters. One lineage originating from cluster 0 CAFs and ending at cluster 1 (Figure 4E,F)—was identified within the CAF population. As cluster 1 CAFs were markedly decreased in LATS1/2 dKO tumors, we focused on their lineage. To identify the critical genes responsible for lineage induction, we used tradeSeq [25] to recognize genes differentially expressed along the cluster 1 lineage (Figure 4G, Table S5). KEGG pathway enrichment analysis showed that focal adhesion and ECM–receptor interaction pathways were the top two enriched pathways among the top genes correlated with pseudotime (Figure 4H, Table S6), suggesting that ECM–cell interactions are important for the induction of clusters 1 and 2 CAFs. Together, these data suggest that LATS1/2 deletion in cancer cells hinders fibroblast activation and reshapes CAF composition through mechanical cell–ECM–cell communication in vivo.

3.4. ITGA2 Overexpression in LATS1/2 Double‐Knockout Cancer Cells Partially Rescues the Tumor‐Generated Force In Vitro and Tumor Growth In Vivo

Next, we explored the underlying mechanism by which LATS1/2 organizes the tumor‐generated force in cancer cells. To this end, we extracted RNA from WT and LATS1/2 dKO 4 T1 spheroids and conducted RNA‐seq analysis. We identified 1246 upregulated and 1302 downregulated genes as differentially expressed genes (DEGs) in LATS1/2‐deficient 4 T1 spheroids compared with those in the WT (Figure 5A,B, Table S7). Enrichment analysis of the DEGs revealed that focal adhesion and ECM‐receptor interaction pathways were among the most significantly enriched pathways in WT spheroids (Figure 5C, Table S8). Since these pathways were also enriched during CAF activation in vivo (Figure 4H), we focused on gene expression related to these two pathways.

FIGURE 5.

LATS1/2 promote expression of collagen‐binding integrins in 4 T1 cells. (A) Volcano plot of the DEGs between WT and LATS1/2 dKO 4 T1 spheroids identified through RNA‐seq data analysis. (B) Heatmap of the DEGs between WT and LATS1/2 dKO 4 T1 spheroids. (C) DEGs between the WT and LATS1/2 dKO spheroids were subjected to KEGG pathway analysis. The top KEGG pathways enriched in WT 4 T1 spheroids are displayed. (D) Expression of the collagen‐ and laminin‐binding integrin subunits in WT or LATS1/2 dKO spheroids. (E) Schematic representation of the integrin pairs involved in collagen‐ and laminin‐binding.

Interestingly, the overall expression of integrins responsible for collagen‐ and laminin‐binding was decreased in LATS1/2‐deficient spheroids (Figure 5D,E). Given that breast tissues are rich in collagen matrix, we further focused on collagen‐binding integrins with decreased expression in LATS1/2 dKO spheroids, including integrin α1 and α2 (Figure 5D). Among these, integrin subunit alpha 2 (ITGA2) is the most abundant collagen‐binding integrin subtype found in 4 T1 cells. Moreover, ITGA2 is consistently downregulated upon YAP activation, either by LATS1/2 inhibition through the small molecule inhibitor TRULI treatment (Figure S7A) or by overexpression of the constitutive active mutant YAP(5SA) (Figure S7B,C) in another breast cancer cell line, 67NR. Therefore, we selected ITGA2 for further investigation. We generated LATS1/2 dKO cells overexpressing ITGA2 (Figure 6A) and found that its overexpression partially rescued the tumor‐generated force in LATS1/2 dKO spheroids, as evidenced by the deformation field observed (Figure 6B,C, Movie S9). These data suggest that the loss of LATS1/2 in 4 T1 spheroids diminishes the tumor‐generated force, partially due to the reduction of ITGA2 expression.

FIGURE 6.

ITGA2 overexpression in LATS1/2 double‐knockout 4 T1 cells partially rescues the tumor‐generated force and tumor growth in vivo. (A) WT and LATS1/2 dKO 4 T1 cells, with or without ITGA2 overexpression, were subjected to immunoblotting analysis using antibodies against the indicated proteins. (B) Color map of ECM displacement induced by LATS1/2 dKO spheroids with or without ITGA2 overexpression in the 3D culture system. (C) Normalized ECM displacement (U/R average) induced by LATS1/2 dKO spheroids with or without ITGA2 overexpression over time. (D) WT or LATS1/2 dKO 4 T1 cells with or without ITGA2 overexpression were transplanted into BALB/c mice, and tumor weight was determined 16 days after transplantation. A representative tumor image (left). Quantitative data (right). Data are presented as means ± SD (n = 26 tumors per group). ****p < 0.0001, *p < 0.05; one‐way ANOVA.

To investigate the in vivo relevance of this finding, we transplanted equal numbers of WT or LATS1/2 dKO 4 T1 cells, with or without ITGA2 overexpression, into the mammary fat pads of syngeneic BALB/c mice to monitor their tumor growth in vivo. We previously found that the deletion of LATS1/2 in 4 T1 breast cancer cells induces a robust immune response [16], which is associated with reduced immunosuppressive CAFs in the TME [21], leading to tumor growth suppression. Therefore, we hypothesized that partial recovery of the tumor‐generated force in ITGA2‐overexpressing LATS1/2 dKO cells might induce fibroblast activation and immunosuppression to assist tumor growth. Indeed, ITGA2 overexpression in LATS1/2 dKO cells partially recovered the reduced tumor growth, whereas ITGA2 overexpression in WT cells did not promote further tumor growth (Figure 6D). These observations suggest that the reduction in ITGA2 expression was partially responsible for the decreased tumor contraction force and growth suppression observed in LATS1/2‐deficient breast cancer. Taken together, our data suggest that integrins play pivotal roles downstream of the Hippo pathway in mediating the tumor‐generated force and tumor growth.

3.5. Niche With Enhanced Focal Adhesion and Extracellular Matrix‐Receptor Signaling Is Associated With Fibroblast Activation in Human Cancer

Our data indicate that the LATS1/2–integrin axis in cancer cells modulates focal adhesion and ECM‐receptor interaction pathways, enabling cell–ECM–cell communication to activate neighboring fibroblasts. To investigate the clinical relevance of our findings, we reanalyzed the scRNA‐seq and Visium spatial transcriptomic datasets from human TNBC samples published in a previous study [26]. We first analyzed the Visium spatial transcriptomic data to determine whether local activities of the focal adhesion and ECM‐receptor interaction pathways are associated with fibroblast activation. The activities of these two pathways were scored at each spot using AUCell and projected onto tissue images (Figure 7A). We then calculated the correlation between activities of these pathways and expression levels of the markers for activated (ACTA2, LRRC15, and TAGLN) and steady‐state fibroblasts (DPT and PI16). High activities of the focal adhesion and ECM‐receptor interaction pathways were positively correlated with the expression levels of ACTA2, LRRC15, and TAGLN while being negatively correlated with those of DPT and PI16 (Figure 7B). These results suggest that a niche with high cell–ECM interaction activity is associated with local fibroblast activation. To further delineate cell–cell communication in TNBC, we subclustered TNBC scRNA‐seq datasets from the same study [26] (Figure 7C) and performed a CellChat [27] analysis to predict the cell–cell interactions. We found that fibroblasts and epithelial cells communicated most frequently in TNBC, especially through the collagen signaling pathway (Figure 7D,E). These data suggest that a niche with enhanced cell–ECM–cell communication, likely between cancerous epithelial cells and fibroblasts, is associated with fibroblast activation in human cancer.

FIGURE 7.

Niche with enhanced focal adhesion and extracellular matrix‐receptor signaling is associated with fibroblast activation in human cancer. (A) Heatmap showing the ACTA2 expression levels and activities (AUCell scores) of the following two pathways: Focal adhesion (FA) and ECM receptor interaction (ERI) pathways in representative TNBC Visium spatial transcriptomic data. (B) Heatmap of the Spearman correlation analysis values measured between ACTA2 expression levels and AUCell scores of the FA and ERI pathways (n = 6 samples). (C) UMAP of an integrated TNBC scRNA‐seq dataset from a previous study. (D) CellChat analysis was performed on scRNA‐seq data to infer the cell–cell interactions that likely occur in TNBC. Circle plots depict the interaction numbers and strength between different cell types in TNBC. Arrows: Direction of intercellular communication. Thickness of the lines: Number/strength of the intercellular communication. (E) Circle plots depicting intercellular communication through collagen signaling pathways between different cell types in TNBC.

4. Discussion

In this study, we demonstrated that the Hippo pathway‐associated kinases, LATS1/2, are necessary for modulating collective tumor‐generated forces. LATS1/2 deletion in cancer cells hindered fibroblast activation induced by cancer cell–fibroblast interactions both in vitro and in vivo. We identified integrin pathways as pivotal mechanisms that drive the tumor force and subsequent induction of mechanoresponsive CAFs. Restoring ITGA2 expression in LATS1/2 dKO 4 T1 cancer cells partially rescued tumor force generation in vitro and promoted tumor growth in vivo. Moreover, a niche with enhanced focal adhesion and ECM‐receptor signaling was associated with fibroblast activation in human breast cancer.

4.1. Tumor‐Generated Force in Tumor Progression

During tumor progression, cancer cells hijack numerous noncancerous host cells to support their growth through cell–cell communication [28]. While biochemical factors have been extensively studied as important mediators of these types of communication, emerging evidence has shifted the focus to mechanical crosstalk between cells. In addition to well‐studied mechanical factors—stiffness, solid stress, and fluid stress [29]—several recent studies have described the tensile forces generated by cancer cells that radially deform the ECM, creating a preferential alignment of fibers perpendicular to the spheroid surface. This pattern resembles the well‐established breast tumor model known as the tumor‐associated collagen signature (TACS). TACS describes three distinct layers of collagen that radiate from the main body of a tumor. The outermost layer, termed TACS3, features bundles of straightened collagen fibers oriented perpendicular to the tumor surface [30]. TACS has been validated in human samples and is strongly correlated with a poor prognosis [31]. However, the mechanisms by which collagen is reorganized into a TACS and how TACS contributes to tumor progression remain poorly understood. Our study suggests that similar ECM remodeling induced by the tumor‐generated force may induce TACS formation and activate surrounding fibroblasts to construct a tumor‐promoting microenvironment. However, fibroblasts are only one of the many cell types that are regulated by mechanical cues. For instance, mechanical stretching can promote the activation of macrophages to create an inflammatory environment that favors tumor progression [32]. Although we did not directly demonstrate that tumor‐derived integrin signaling activates immunosuppressive fibroblasts in vivo, our previous study showed that LATS1/2 deletion enhances the secretion of extracellular vesicles (EVs), which activate dendritic cells to induce anti‐tumor immunity [16]. While the current study focuses on the mechanical aspects of CAF activation, the potential contribution of EVs warrants further investigation. Future studies are needed to dissect how tumor‐generated force influences various components of the TME.

4.2. Hippo Pathway and the Tumor‐Generated Force

The Hippo intracellular signaling pathway is well‐established as a key regulator of cellular responses to environmental cues and facilitates cellular adaptation to dynamic microenvironments. Recent studies have expanded our understanding of the Hippo pathway, suggesting that it plays an active role in modulating the extracellular environment. Notably, YAP and TAZ, two core effectors of the Hippo pathway, directly regulate the differentiation of fibroblasts into myofibroblasts, which in turn contribute to tissue remodeling through deposition and contraction of the ECM [33]. However, the role of the Hippo pathway in other contexts, particularly within the TME, remains unclear. Here, we propose that the Hippo pathway regulates tumor cell‐driven ECM remodeling by modulating mechanical forces within the TME. Our findings suggest that the mechanical tension exerted by cancer cells is a critical factor in ECM‐mediated cell–cell communication and the activation of adjacent fibroblasts, thereby influencing the composition and organization of the TME. ECM remodeling, particularly TACS, plays a pivotal role in tumor progression by facilitating the initiation of CAFs and other stromal cells within the TME [30]. Therefore, our results underscore the potential of targeting the Hippo pathway as a therapeutic strategy for disrupting ECM remodeling and the mechanical forces that drive tumor progression.

We observed that ITGA2 re‐expression in LATS1/2‐deficient 4 T1 cells partially restored tumor‐generated force. This incomplete rescue may result from the concurrent downregulation of multiple integrins in LATS1/2‐deficient cells, suggesting that full restoration of force generation may require upregulation of additional integrin subunits. While our data support the role of integrins as downstream effectors of the LATS1/2–YAP/TAZ axis, we cannot exclude the contribution of other YAP/TAZ targets, such as CDC42 [34], which may contribute to force generation. Additionally, LATS1/2 are known to have functions beyond the Hippo pathway, including roles in maintaining cell shape [35] and genomic integrity [36], which could influence force production through YAP/TAZ‐independent mechanisms. Prior work has shown that YAP regulates focal adhesion integrity and that its inactivation can downregulate certain integrins, while upregulating others in a context‐dependent manner [11]. Given the heterogeneity and complexity of the ECM across tissues and cancer types [37], the Hippo pathway's role in regulating tumor‐generated forces is likely to be highly context‐specific. Indeed, a previous study reported that LATS1/2 are required for sphere formation and self‐renewal of cancer stem cells in oral squamous cell carcinoma [35]. Although the Hippo pathway is generally considered as tumor‐suppressive, emerging evidence suggests that it may promote tumor growth in certain cancers, including colorectal [38], hematological malignancies [39] and ERα‐positive breast cancers [40]. In such cases, YAP/TAZ remain under the regulation of the Hippo pathway but inhibit tumor growth through alternative mechanisms. Our findings emphasize the context‐dependent duality of the Hippo pathway in cancer biology, wherein it can exert both tumor‐suppressive and tumor‐promoting functions [20].

Despite significant efforts aimed at targeting the proteins involved in mechanotransduction, such as integrins, clinical trials evaluating integrin‐targeting therapeutics for cancer have largely been unsuccessful [41]. This may be partly attributed to the incomplete understanding of cancer mechanobiology. Further investigation of the mechanistic links between the Hippo pathway and cancer mechanobiology is crucial for advancing therapeutic strategies targeting the biomechanical aspects of tumor progression.

Author Contributions

Yanliang Liu: conceptualization, data curation, formal analysis, investigation, methodology, visualization, writing – original draft, writing – review and editing. Saisai Liu: formal analysis, investigation, writing – review and editing. Yasuhisa Sakamoto: formal analysis, project administration, writing – original draft, writing – review and editing. Paweenapon Chunthaboon: investigation, validation, writing – review and editing. Chanida Thinyakul: investigation, writing – review and editing. Ryunosuke Mori: data curation, formal analysis, investigation, writing – review and editing. Shuran Li: investigation, writing – review and editing. Takahiro Watanabe‐Nakayama: investigation, writing – review and editing. Seiji Omata: formal analysis, methodology, project administration, writing – review and editing. Yasuyuki Morita: formal analysis, methodology, supervision, visualization, writing – review and editing. Toshiro Moroishi: conceptualization, formal analysis, funding acquisition, methodology, project administration, resources, supervision, visualization, writing – original draft, writing – review and editing.

Disclosure

Animal Studies: All animal experiments were approved by the Kumamoto University Animal Experiment Committee and conducted following the laws and regulations concerning animal experiments and animal care.

Ethics Statement

The authors have nothing to report.

Consent

The authors have nothing to report.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1. LATS1/2 dKO 4 T1 cells form larger spheroids than WT.

Figure S2. WT and LATS1/2 dKO 4 T1 cells have no significant differences in migration capacity under 2D culture conditions.

Figure S3. Validation of the NIH3T3 reporter cells.

Figure S4. Conditioned medium derived from 4 T1 spheroids is not sufficient to induce mNeonGreen expression in NIH3T3 reporter cells.

Figure S5. Clustering CAF subset from the total cell population in scRNA‐seq data.

Figure S6. Clustering CAF subset from the reference dataset.

Figure S7. YAP activation suppresses ITGA2 expression in 67NR cells.

Movie S1. Representative movie of the 3D culture system used for tracking spheroid‐induced deformation of the collagen matrix.

Movie S2. Collagen matrix deformation induced by WT 4 T1 spheroids in the 3D culture system.

Movie S3. Collagen matrix deformation induced by LATS1/2 dKO 4 T1 spheroids in the 3D culture system.

Movie S4. Color map showing the change in ECM displacement over time induced by WT 4 T1 spheroids in the 3D culture system.

Movie S5. Color map showing the change in ECM displacement over time induced by LATS/2 dKO 4 T1 spheroids in the 3D culture system.

Movie S6. Morphological changes in NIH3T3 cells co‐cultured with WT 4 T1 spheroids in a collagen gel.

Movie S7. Morphological changes in NIH3T3 cells co‐cultured with LATS1/2 dKO 4 T1 spheroids in collagen gel.

Movie S8. Morphological changes in NIH3T3 cells cultured in collagen gel.

Movie S9. Collagen matrix deformation induced by ITGA2‐overexpressing LATS1/2 dKO 4 T1 spheroids in the 3D culture system.

Table S1. The antibodies used in this study are listed as follows.

Table S2. The reagents used in this study are listed as follows.

Table S3. The plasmids used in this study are listed as follows.

Table S4. The oligos used in this study are listed as follows.

Table S5. Genes differentially expressed along the cluster 1 lineage identified by Tradeseq.

Table S6. Result of the KEGG enrichment analysis on genes in Table S5.

Table S7. DEG between WT and LATS1/2 dKO 4 T1 spheroids.

Table S8. Result of the KEGG enrichment analysis on DEG between WT and LATS1/2 dKO 4 T1 spheroids.

Data S1.