Breast cancer relapse can develop over the course of years as a result of dormant cancer cells that disseminate to secondary sites. These dormant cells are often resistant to conventional hormone and chemotherapy. Although recurrence is the main cause of death from cancer, microenvironmental factors that may influence resistance to therapy and duration of dormancy are largely unknown. Breast cancer relapse is often detected in tissues that are softer than the normal mammary gland or the primary breast tumor, such as bone marrow, brain, and lung. We therefore explored how stiffness of the microenvironment at secondary sites regulates tumor dormancy and the response of breast cancer cells to hormone and chemotherapy. In soft microenvironments reminiscent of metastatic sites, breast cancer cells were more resistant to the estrogen receptor modulator tamoxifen as a result of increased autophagy and decreased expression of estrogen receptor-α. Consistently, pharmacologic inhibition or genetic downregulation of autophagy increased the response of breast cancer cells to tamoxifen on soft substrata. In addition, autophagy was decreased downstream of integrin-linked kinase on stiff substrata. Altogether, our data show that tissue mechanics regulates therapeutic outcome and long-term survival of breast cancer cells by influencing autophagy.

Significance:

These findings characterize the persistence of dormant cells at metastatic sites, where soft microenvironments downregulate estrogen receptor expression and upregulate autophagy, thereby promoting therapy resistance in breast cancer cells.

Approximately 90% of cancer-related deaths are due to malignant growths that appear at sites far removed from that of the primary tumor (1). Although the mechanisms that drive metastatic growth in secondary organs are still unclear, it is well appreciated that disseminated cancer cells can remain clinically asymptomatic for as long as decades. The time from the formation of the first micrometastasis to the appearance of a clinically detectable macrometastasis is known as “tumor dormancy,” the rate-limiting step in metastasis (2). Although tumor dormancy was initially attributed to temporary mitotic arrest or quiescence (3–5), alternative hypotheses have suggested that a micrometastasis can fail to increase in size due to poor vascularization or cytotoxic activity of the immune system (6–8). This makes it difficult to detect dormant cancer cells using current diagnostic tools or to target them using conventional chemo-, hormone-, or radiotherapy approaches. Therefore, clinical practices rely on estimating the probability of recurrence from prognostic factors such as the grade and stage of the disease, and limited understanding of the biology of dormancy prevents its effective treatment (9). To improve treatment of metastatic disease, it is necessary to elucidate the pathways that promote dormancy and to uncover how dormant cancer cells interact with their surrounding niche.

Cancer cell behavior is influenced by the tumor microenvironment, which is comprised of the extracellular matrix (ECM), neighboring host cells, the immune system, and soluble factors including hormones and cytokines. The ECM is a major regulator of cell function, as cancer cells receive biochemical and mechanical signals from the host tissue, interpret these signals, and tailor their microenvironment into a hospitable niche (10). Breast cancer metastases are most frequently detected in the bone, brain, liver, and lung (11), and it is likely that the distinct microenvironments offered by these sites affect the temporal and spatial progression of metastatic disease. Consistently, microenvironmental signals appear to regulate tumor dormancy. For example, stable microvasculature promotes quiescence in breast cancer cells, which switch to a proliferative phenotype when they encounter newly sprouted blood vessels (12). This transition from quiescence to proliferation depends on the production of fibronectin and signaling through β1-integrin (13). Similarly, microenvironmental stresses such as hypoxia prime dormancy of head-and-neck squamous cell carcinoma and breast cancer cells, and posthypoxic disseminated tumor cells evade chemotherapy (14).

The mechanical signals that cancer cells receive from their microenvironment, including matrix stiffness as well as solid and fluid stress also influence cell proliferation, motility, differentiation, and resistance to apoptosis (15, 16). It is unclear, however, whether the mechanical properties of the microenvironment also regulate tumor dormancy. In hepatocellular carcinoma, matrix stiffness has been found to regulate cellular quiescence, response to chemotherapeutics, and clonogenic capacity following chemotherapy (17). Because breast cancer metastases are detected in tissues that are softer than a breast tumor or a healthy mammary gland (18–20), soft microenvironments could promote the survival of disseminated breast cancer cells at secondary sites.

Little is known about how micrometastases remain viable over prolonged periods of time. Recently, autophagy has emerged as a self-degradative process that has the potential to promote the survival of stem-like subpopulations of tumor cells (21–23) and that could contribute to the persistence of chemoresistant tumor cells at secondary sites (24). Monoallelic loss of beclin1, an essential autophagy gene, is common in breast cancer and is a precursor to tumorigenesis in mouse models of mammary carcinoma (25). These observations suggest that autophagy could play a tumor-suppressive role. Autophagy has also been demonstrated to promote the survival and chemoresistance of dormant ovarian and gastrointestinal cancer cells (26, 27). Overexpression of the tumor suppressor aplasia Ras homolog member I (ARHI) has been shown to induce dormancy and autophagy. In ovarian tumor xenografts, increased formation of autophagosomes is associated with reduced proliferation, and loss of ARHI expression correlates with a regain of proliferative potential (26). The fact that autophagy is a dynamic process that can be tumor-suppressive or tumor-promoting implies that its role in cancer progression might be influenced by the tumor microenvironment (28).

We show here that the mechanical properties of the surrounding microenvironment regulate tumor dormancy and autophagy in human breast cancer cells. Using synthetic substrata of tunable stiffness, we found that soft microenvironments reminiscent of metastatic sites harbor dormant estrogen receptor-alpha (ERα)-positive cancer cells that are resistant to treatment with the antiestrogen, tamoxifen. Soft microenvironments also cause an upregulation of autophagy, indicating that autophagy could serve as a survival mechanism for dormant cancer cells. In contrast, while breast cancer cells proliferate more on stiff microenvironments, they are also more prone to elimination by tamoxifen as well as common chemotherapeutics. Altogether, these results suggest that tissue mechanics regulate chemotherapeutic outcome and the long-term survival of breast cancer cells, in part, by influencing autophagy.

Cell culture and reagents

MCF7 and ZR-75-1 ERα-positive human breast cancer cells were obtained from the ATCC (in 2008 and 2018, respectively) and cultured in DMEM/F12 or RPMI medium respectively, supplemented with 10% heat-inactivated FBS (Atlanta Biologicals), 10 μg/mL insulin (Sigma), and 50 μg/mL gentamicin (Gibco) for 3 days unless otherwise specified. 5-fluorouracil (5 μmol/L, Sigma) or tamoxifen (2 μmol/L, Sigma) were added to fresh culture medium 24 hours after plating cells. To block autophagosome degradation, chloroquine (CQ; 0.1 μmol/L, Sigma) was added to the culture medium 24 hours before fixing cells. Cell lines were authenticated by short tandem repeat genotyping (ATCC) and used before passage 20. MCF7 cells were tested for Mycoplasma (2017; Lonza); experiments with ZR-75-1 cells were completed within 6 months of thawing the purchased vial and therefore not tested for Mycoplasma.

Viral transductions

A bicistronic recombinant adenovirus encoding ILK and GFP (AdILK) was obtained from Vector Biolabs (29). An adenovirus encoding GFP alone (AdGFP) was used as control. Each adenovirus was added to the cell culture medium at a multiplicity of infection of 100.

Transient RNA interference

SMARTpool containing 4 siRNA duplexes per gene were used to knockdown the expression of ILK1, Beclin1, and Atg7 (L-004499-00-0005; L-010552-00-0005; L-020112-00-0005, Dharmacon). A nontargeting SMARTpool was used as a negative control (D-001810-10-05, Dharmacon). Cells were plated at a density of 4 × 105 cells per mL and transfected using RNAiMAX transfection reagent (Invitrogen) to reach a final siRNA concentration of 20 nmol/L.

Synthetic substrata

Shear moduli have been measured in live, normal human breast tissue, solid human breast tumors, and common metastatic sites (30–33). To mimic the mechanical microenvironments of these tissues, we used the shear moduli values to calculate elastic moduli, generated polyacrylamide (PA) substrata with these elastic moduli, and adapted them for cell culture as described previously (34). Glass coverslips (31-mm diameter) were washed with 0.1 N NaOH for 30 minutes and then washed three times with MilliQ water (Millipore) and dried. Coverslips were then immersed in 2% (vol/vol) aminopropyltrimethoxysilane (Sigma Aldrich) in acetone for 30 minutes, washed three times with acetone, and dried. Finally, coverslips were immersed in 1% (vol/vol) glutaraldehyde (Sigma Aldrich) in PBS for 30 minutes, washed three times with MilliQ water, and dried. PA gels were synthesized on the treated coverslips as follows: 12.5% (vol/vol) acrylamide was mixed with either 0.5% (vol/vol) or 17.5% (vol/vol) bis-acrylamide in water. Polymerization was initiated by adding 10% ammonium persulfate (Bio-Rad) and N,N,N',N'-tetramethylethylenediamine (Sigma Aldrich). Thirty-six microliters of the mixture was placed onto treated coverslips and each sandwiched under an untreated glass coverslip. After drying for 1 hour at room temperature, PA gels were stored in PBS at 4°C. PA gels were characterized as described previously (35). The top coverslips were removed from the PA gels and functionalized with fibronectin using the heterobifunctional crosslinker, Sulfo-SANPAH (Thermo Fisher Scientific). Before plating cells, gels were washed three times with PBS and incubated with culture medium for 30 minutes at 37°C.

Immunofluorescence analysis

Samples were fixed with 4% paraformaldehyde in PBS for 15 minutes at room temperature, washed three times with PBS, blocked with 10% (vol/vol) goat serum (Sigma Aldrich) in PBS for 2 hours, and incubated overnight at 4°C with rabbit anti-LC3B (1:200; Sigma) antibody. Samples were washed four times for 15 minutes each with 0.5% Triton X-100 in PBS (PBST), and then incubated with Alexa 488 goat anti-rabbit secondary antibody (Invitrogen) overnight at 4°C. To label nuclei, samples were incubated with Hoechst 33342 (1:1,000; Invitrogen) for 15 minutes at room temperature and washed three times with PBS for 10 minutes each. Apoptosis analysis was conducted using a terminal deoxylnucleotidyl transferase dUTP nick-end labeling (TUNEL) kit (Thermo Fisher Scientific).

Images were acquired with a Hamamatsu Orca CCD camera attached to a Nikon Ti-U inverted fluorescence microscope using a ×20 magnification air objective (or ×60 magnification oil objective for autophagosome immunofluorescence). Phase-contrast and fluorescence images were merged using ImageJ. Proliferation analysis was based on EdU incorporation (Thermo Fisher Scientific). Briefly, the number of cells with EdU-positive nuclei was counted and recorded as the percentage of the total number of cells in each frame. The rolling average of the percentage of proliferating cells was calculated for each frame and for 400–500 cells for each condition. The size and number of autophagosomes (LC3B-labeled cytoplasmic vesicles) were quantified using a MATLAB code that employs Gaussian filtering to eliminate background noise.

Immunoblotting analysis

Samples were lysed in RIPA lysis buffer (Thermo Fisher Scientific) supplemented with a protease inhibitor (Roche). Protein concentrations were measured using the Bradford Assay (Bio-Rad). Equal amounts of total protein were separated by SDS-PAGE and transferred to polyvinylidene difluoride membranes, which were then blocked in 5% nonfat milk and incubated overnight at 4°C in blocking buffer containing primary antibodies. Antibodies used for immunoblotting were: rabbit anti-LC3B (1:1,000; Sigma), rabbit anti-p21 (1:1,000; Cell Signaling Technology), rabbit anti-p27 (1:1,000; Cell Signaling Technology), rabbit anti-cyclin D1 (1:1,000; Cell Signaling Technology), mouse anti-p62 (1:1,000; Abcam), rabbit anti-ERα (1:1,000; Cell Signaling Technology), rabbit anti-Atg7 (1:1,000; Cell Signaling Technology), rabbit anti-Beclin1 (1:1,000; Cell Signaling Technology), rabbit anti-ILK1 (1:100; Cell Signaling Technology), and rabbit anti-GAPDH (1:2,000; Cell Signaling Technology). After three washes with TBST, blots were probed with horseradish peroxidase-conjugated anti-rabbit secondary antibodies for 45 minutes (1:5,000; Cell Signaling Technology). Following incubation with ECL Plus Western Blotting Detection System (GE Healthcare), blots were imaged using a FluorChemE Imager (Cell Biosciences). For each protein, densitometry analysis was carried out in ImageJ by dividing the total intensity of each band by that of GAPDH in the same sample.

Cell-cycle analysis

To assay for cell-cycle stage, the FUCCI cell-cycle sensor (Invitrogen) was added to culture medium at 50 particles per cell and the cells were cultured for another 24 hours. The cells were then fixed with 4% PFA, washed three times with PBS, and counterstained with Hoechst 33342 to visualize the nuclei. Images were acquired as described above using a ×10 magnification air objective. Nuclei with red, yellow, or green labeling were recorded as G0–G1, G1–S transition, or S–G2–M phases, respectively.

Soft microenvironments harbor growth-arrested breast cancer cells

Breast cancer metastases are most often detected in tissues that are inherently softer than a breast tumor or a healthy mammary gland (18, 20, 30–32). This observation suggests a role for tissue stiffness in the regulation of dormancy, as well as in the response of disseminated breast cancer cells to commonly used cancer treatments. MCF7 or ZR-75-1 ERα+ human breast cancer cells cultured on “soft” substrata (E∼0.1, 0.9, 2, 4 kPa) that mimic metastatic sites including the bone marrow, the liver, and the brain (Fig. 1A; ref. 30) exhibit a rounded morphology (Fig. 1B; Supplementary Fig. S1A), and are significantly less proliferative than their counterparts cultured on “stiff” substrata (E∼45, 100 kPa) that mimic the human mammary gland and breast tumors (Fig. 1C and D; Supplementary Fig. S1B and S1C), as indicated by EdU-incorporation assays. Cyclin D1, which regulates progression through the G1 phase of the cell cycle, is downregulated, while the cyclin-dependent kinase inhibitors p21 and p27 are upregulated (Fig. 1E) in MCF7 cells cultured on soft substrata. Consistently, cell-cycle analysis using the FUCCI reporter system revealed that the percentage of cells in the G0–G1 phase of the cell cycle is increased on soft substrata (Fig. 1F). These data suggest that culture in soft microenvironments might induce quiescence in breast cancer cells.

Given that dormant phenotypes can either result from cellular quiescence or from a balance between proliferation and cell death within micrometastases (8), we also investigated the effect of substratum stiffness on apoptosis. We cultured breast cancer cells on soft or stiff substrata and used the TUNEL assay to detect apoptosis (Fig. 1G; Supplementary Fig. S1D). TUNEL staining revealed that apoptosis is increased in both MCF7 and ZR-75-1 cells on soft microenvironments (Fig. 1H; Supplementary Fig. S1E).

Therefore, both proliferation and apoptosis are influenced by substratum stiffness in these ER+ breast cancer cells. Prolonged periods of dormancy in breast cancer metastases could therefore result from the fact that soft microenvironments promote a balance between proliferation and apoptosis, or because they decrease cell-cycle progression.

Substratum stiffness regulates the response of ER+ breast cancer cells to tamoxifen, in part by modulating ERα expression

One of the biggest challenges in cancer treatment is resistance to conventional therapy. Because proliferation is altered by substratum stiffness, we explored how this microenvironmental parameter regulates the response of ERα+ cells to the ER modulator tamoxifen (36). Breast cancer cells were cultured on soft or stiff substrata for 24 hours and treated with tamoxifen at a concentration that matches that found in the serum of breast cancer patients (2 μmol/L for 72 hours; refs. 37, 38). We found that whereas tamoxifen inhibits the proliferation of MCF7 cells cultured on stiff microenvironments (Fig. 2A), those cultured on soft substrata are impervious to treatment with tamoxifen (Fig. 2B; Supplementary Fig. S2A). Given that breast cancer metastases are often detected in tissues softer than the primary tumor, our results suggest that conventional therapies may be ineffective at eliminating cells that have disseminated to softer secondary sites. We also assessed apoptosis in tamoxifen-treated MCF7 cells (Fig. 2C) and found that this treatment specifically increases apoptosis in the cells cultured on stiff substrata, but does not affect those on soft substrata (Fig. 2D). These data show that tamoxifen decreases proliferation and enhances death in cells cultured on substrata that mimic the stiffness of primary tumors, but has no effect in cells on substrata that mimic soft metastatic sites.

Given that tamoxifen targets ERα, we performed immunoblotting analysis for ERα in MCF7 or ZR-75-1 cells treated with or without tamoxifen on soft or stiff substrata. This analysis revealed that culture on soft substrata decreases the expression of ERα (Fig. 2E; Supplementary Fig. S3A). We also found that the levels of ERα are increased in cells treated with tamoxifen (Fig. 2E), consistent with previous reports (39). These data suggest that cells on soft microenvironments downregulate ERα expression, which might partially account for their resistance to tamoxifen.

Soft microenvironments induce chemoresistance by increasing autophagy

It is well appreciated that breast cancer cells can survive at secondary sites for prolonged periods of time, but how they do so remains a mystery. Autophagy was recently suggested as a mechanism to prevent slowly cycling ovarian or quiescent breast cancer cells from undergoing apoptosis (26, 40). To determine whether autophagy in breast cancer cells is affected by the mechanical properties of the microenvironment, we first examined how substratum stiffness regulates autophagosomes. LC3B-II, the lipidated version LC3B-I (cytosolic LC3B), is incorporated into autophagosomal membranes and serves as a marker for autophagosomes (41). We found that the expression of LC3B-II is elevated specifically in MCF7 or ZR-75-1 cells cultured on soft substrata (Fig. 3A; Supplementary Fig. S3B), suggesting that autophagy may be increased in these microenvironments. We also examined the expression of the autophagy substrate p62, which serves as a link between LC3 and ubiquitinated substrates and is degraded by autophagy. The levels of p62 are therefore inversely correlated with induction of autophagy (41). As predicted, we found that the levels of p62 are decreased in cells on soft substrata (Fig. 3B; Supplementary Fig. S3C), consistent with an increase in autophagy on these microenvironments. Immunofluorescence analysis for LC3B (Fig. 3C; Supplementary Fig. S3D) revealed that the number (Fig. 3D) and total volume (Fig. 3E) of autophagosomes per cell are highest in MCF7 or ZR-75-1 cells on soft substrata, and decrease as substratum stiffness increases.

Inhibiting autophagy sensitizes breast cancer cells on soft microenvironments to tamoxifen

To determine whether the increase in autophagosomes on soft substrata is due to increased autophagosome formation or impaired autophagosome degradation, we performed an autophagosome turnover assay. We cultured MCF7 cells on soft or stiff substrata for 24 hours and then treated them with CQ (0.1 μmol/L for 48 hours), which is a lysomotropic agent that increases lysosomal pH and inhibits autophagosome-lysosome fusion, consequently blocking autophagosome degradation (41). We found that treatment with CQ leads to an increase in the number and volume of autophagosomes (Fig. 3F), specifically in MCF7 cells cultured on soft substrata (Fig. 3G). These data suggest that autophagosome formation is enhanced in MCF7 cells on soft microenvironments. Altogether, our data show that autophagy is regulated by substratum stiffness in MCF7 cells, and reveal that autophagy may contribute to the survival of breast cancer cells on soft microenvironments.

To determine whether the increase in autophagy is responsible for resistance to tamoxifen on soft substrata, we treated MCF7 cells with tamoxifen and stained for LC3B as a marker of autophagosomes (Fig. 4A). We found that treatment with tamoxifen increases the number of autophagosomes on both soft and stiff substrata (Fig. 4B). However, the total volume of autophagosomes on soft substrata is significantly higher than that on stiff substrata upon exposure to tamoxifen (Fig. 4C). Immunoblotting analysis revealed that the expression of LC3B-II is increased while the expression of p62 is decreased upon tamoxifen treatment on soft substrata (Fig. 4D). To uncover whether autophagy promotes tamoxifen resistance on soft substrata, we treated MCF7 or ZR-75-1 cells with CQ and investigated their response to tamoxifen. While blocking autophagy does not affect proliferation (Fig. 4E; Supplementary Fig. S3E), our data suggest that apoptosis increases in autophagy-impaired cells in response to tamoxifen on soft substrata (Fig. 4F; Supplementary Fig. S3F). In contrast, blocking autophagy does not affect apoptosis in cells on stiff substrata (Fig. 4F; Supplementary Fig. S3G and S3H). Soft microenvironments therefore render ERα+ breast cancer cells resistant to tamoxifen by upregulating autophagy.

Formation of the autophagosome consists of initiation, nucleation, and elongation steps that are regulated by multiprotein complexes including those containing Beclin1 and Atg7 (Fig. 5A). Beclin1 interacts with class III type phosphoinositide 3-kinase/Vsp34 to coordinate membrane nucleation, whereas Atg7 catalyzes lipidation of LC3-I to LC3-II with phosphatidylethanolamine during elongation of the autophagosomal membrane (42). To define how the mechanical properties of the microenvironment affect autophagosome formation, we examined the expression levels of upstream regulators Beclin1 and Atg7 in cells cultured on soft or stiff substrata. Immunoblotting analysis revealed that the levels of Beclin1 are increased in cells on soft substrata (Fig. 5B), consistent with a possible increase in vesicle nucleation (43). To determine whether decreasing upstream initiation of autophagy alters response to tamoxifen, we used siRNAs to deplete Atg7 (Fig. 5C) or Beclin1 (Fig. 5D) in MCF7 cells cultured on soft or stiff substrata in the presence or absence of tamoxifen. We found that decreasing the levels of Atg7 or Beclin1 sensitizes MCF7 cells to tamoxifen treatment. Specifically, we observed that proliferation decreases in autophagy-deficient cells on soft substrata in response to tamoxifen (Fig. 5E). We also found that apoptosis increases dramatically on soft substrata in Beclin1- or Atg7-deficient cells treated with tamoxifen (Fig. 5F). These data suggest that soft microenvironments upregulate formation of autophagosomes in ERα+ breast cancer cells, which renders them resistant to tamoxifen. However, depleting Atg7 or Beclin1 does not alter the expression levels of ERα (Fig. 5G). Impairing autophagosome nucleation, therefore, enhances the response of MCF7 cells to tamoxifen without affecting ERα expression.

Autophagy is increased downstream of integrin-linked kinase

We then investigated the mechanisms by which substratum stiffness regulates autophagy. Mechanical signals received by β1-integrin are transduced through integrin-linked kinase (ILK) to the actin cytoskeleton and downstream pathways (44). Consistent with previously reported findings in other breast cancer cell lines (29), we found that MCF7 cells cultured on stiff substrata have elevated expression of ILK (Fig. 6A). To determine whether integrin signaling through ILK regulates autophagy, we used an adenoviral approach to ectopically express ILK in cells cultured on soft or stiff substrata (Fig. 6B). Immunofluorescence analysis for LC3B (Fig. 6B) revealed that ectopic expression of ILK downregulates the number (Fig. 6C; Supplementary Fig. S3G) and volume (Fig. 6D; Supplementary Fig. S3H) of autophagosomes in MCF7 or ZR-75-1 cells on soft substrata. Our data therefore suggest that substratum stiffness regulates autophagosome formation, in part, by signaling through ILK.

The tendency of a tumor to metastasize to different tissues reflects how well it can adapt to the microenvironment of a secondary site. More than a century ago, Paget's seed and soil hypothesis posited that disseminated tumor cells will grow only in hospitable niches (45). Consistently, breast cancer metastases are most often detected in tissues such as the brain, lung, liver, and bone marrow, which are softer than both the normal mammary gland as well as the primary breast tumor. Secondary sites harbor dormant tumors that have been implicated in breast cancer relapse (5, 8, 9); however, the survival mechanisms used by dormant cancer cells to perdure at soft metastatic sites have remained elusive. Here, we demonstrate that breast cancer cells cultured on soft microenvironments upregulate autophagy, increasing their ability to survive for prolonged periods of time. In addition, we show that breast cancer cells that are cultured on soft microenvironments downregulate expression of ERα and become impervious to treatment with tamoxifen (Fig. 7).

The majority of breast cancers are ER+ and can therefore be targeted with endocrine therapies such as tamoxifen. However, many patients eventually develop resistance to these therapies due to decreased ER expression (46). Our data suggest that the mechanical microenvironment regulates the expression of ERα, thereby modulating the response to tamoxifen treatment. Novaro and colleagues previously reported that decreasing cell–ECM attachments in normal mammary epithelial cells by blocking β1-integrin decreases ERα expression (47). Concordant with this finding, our data suggest that ERα expression is decreased in cells on soft microenvironments (Fig. 2E), which have weaker integrin-mediated cell–matrix adhesions (Fig. 6A). These observations could partially explain why secondary lesions are resistant to hormonal therapy and remain asymptomatic at soft metastatic sites.

Our data also demonstrate that on soft substrata, ERα+ breast cancer cells activate autophagy, which enhances their survival and promotes resistance to antiestrogen therapy (Figs. 4 and 5). Autophagy mobilizes intracellular energy stores to meet cellular metabolic demands (48) and can protect cancer cells from apoptosis (49, 50). Consequently, preclinical studies have investigated combining autophagy inhibitors with chemotherapeutics (51), and it has been suggested that inhibiting autophagy leads to apoptosis of gastrointestinal stromal tumor cells (27). Autophagy has also been shown to play a role in the maintenance of dormant tumors (40). In an ovarian carcinoma model, the formation of dormant tumors was found to correlate with increased formation of autophagosomes (26). It is therefore possible that breast cancer cells activate autophagy pathways to survive in soft microenvironments and that this contributes to resistance to tamoxifen. Consistently, we found that inhibiting autophagy pharmacologically or by knocking down the expression of Atg7 or Beclin1 sensitizes breast cancer cells to tamoxifen when they are on soft substrata (Figs. 4F and 5G). Conversely, we found a decrease in ERα expression under conditions that increase autophagy, but it is unclear whether these two processes are directly coupled. Cook and colleagues demonstrated that knockdown of ERα leads to resistance to antiestrogen therapy while causing an increase in autophagy (52); however, it remains unclear whether ERα regulates induction of autophagy or if the increase in autophagy is solely a stress response caused by a decrease in ERα levels.

The effects of the mechanical microenvironment on therapeutic resistance are not limited to hormone therapies. We found that the response of breast cancer cells to 5-fluorouracil (5-FU) is similarly regulated by substratum stiffness (Supplementary Fig. S4A–S4D), and that breast cancer cells residing on soft substrata are sensitized to treatment with 5-FU upon pharmacologic inhibition of autophagy (Supplementary Fig. S5A–S5E). It was recently reported that autophagy is critical for the maintenance of dormant breast cancer cells, and that once dormant cells have transitioned to proliferation, they no longer require autophagy for survival (40). Our data also indicate that autophagy is increased in breast cancer cells that become less proliferative in soft microenvironments, and that response to chemotherapeutic agents could be ameliorated by impairing autophagy.

The ability of a cell to proliferate depends on cell-ECM adhesion as well as exposure to growth factors and cytokines (53). A lack of ECM attachment in anchorage-dependent cells, as well as other microenvironmental stress factors, activates autophagy. Autophagy has been implicated in developmental processes that involve luminal clearing such as mammary morphogenesis (54), as well as in disease, where its role in cancer has been shown to be context-dependent. Although previous studies have not explored the regulation of autophagy by the mechanical microenvironment, several of them have emphasized the role of integrin–ECM interactions in autophagy regulation. During dissemination, detachment from the ECM can induce autophagy to give cancer cells time to establish cell–ECM contacts necessary to survive at secondary sites (53). The travel of cancer cells from a stiff tumor microenvironment to a soft secondary site could similarly result in a decrease in integrin–ECM interactions. Considering its role in the transmission of mechanical signals from β1-integrin to downstream signaling pathways, we predicted that ILK could be involved in the regulation of autophagy. Consistently, we found that ectopically expressing ILK, which has previously been shown to strengthen cell–ECM adhesion (55), leads to a decrease in autophagy (Fig. 6; Supplementary Fig. S3G and S3H). Weaker cell–ECM attachments induced by soft substrata also lead to decreased proliferation and growth arrest, which could enable the long-term survival of breast cancer cells in soft microenvironments. These findings provide a potential connection between increased autophagy and tumor dormancy, as suggested recently (40), and implicate a role for the mechanical microenvironment in regulating cell-survival mechanisms in breast cancer. The role of ILK and its downstream cytoskeletal and signaling targets in autophagosome assembly will be the subject of a future study.

No potential conflicts of interest were disclosed.

A.A. Anlaş: Conceptualization, investigation, writing-original draft. C.M. Nelson: Conceptualization, supervision, funding acquisition, project administration, writing-review and editing.

This work was supported in part by grants from the NIH (CA187692 and CA214292), the NSF (CMMI-1435853), the David & Lucile Packard Foundation, and the Camille & Henry Dreyfus Foundation. A.A. Anlaş was supported, in part, by a predoctoral fellowship from the New Jersey Commission on Cancer Research. C.M. Nelson was supported in part by a Faculty Scholars Award from the HHMI.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Chaffer
CL
,
Weinberg
RA
. 
A perspective on cancer cell metastasis
.
Science
2011
;
331
:
1559
64
.
2.
Giancotti Filippo
G
. 
Mechanisms governing metastatic dormancy and reactivation
.
Cell
2013
;
155
:
750
64
.
3.
Willis
R
. 
The spread of tumors in the human body
.
London, England
:
Butterworth & Co. Ltd.
, 
1952
, p.
1
18
.
4.
Hadfield
G
. 
The dormant cancer cell
.
BMJ
1954
;
2
:
607
10
.
5.
Naumov
GN
,
MacDonald
I
,
Weinmeister
P
,
Chambers
A
. 
Persistence of solitary mammary carcinoma cells in a secondary site
.
Cancer Res
2002
;
62
:
2162
8
.
6.
Holmgren
L
,
O'Reilly
MS
,
Folkman
J
. 
Dormancy of micrometastases: balanced proliferation and apoptosis in the presence of angiogenesis suppression
.
Nat Med
1995
;
1
:
149
53
.
7.
O'Reilly
MS
,
Holmgren
L
,
Chen
C
,
Folkman
J
. 
Angiostatin induces and sustains dormancy of human primary tumors in mice
.
Nat Med
1996
;
2
:
689
92
.
8.
Sosa
MS
,
Bragado
P
,
Aguirre-Ghiso
JA
. 
Mechanisms of disseminated cancer cell dormancy: an awakening field
.
Nat Rev Cancer
2014
;
14
:
611
22
.
9.
Aguirre-Ghiso
JA
. 
Models, mechanisms and clinical evidence for cancer dormancy
.
Nat Rev Cancer
2007
;
7
:
834
46
.
10.
Bissell
MJ
,
Hall
HG
,
Parry
G
. 
How does the extracellular matrix direct gene expression?
J Theor Biol
1982
;
99
:
31
68
.
11.
Chambers
AF
,
Groom
AC
,
MacDonald
IC
. 
Dissemination and growth of cancer cells in metastatic sites
.
Nat Rev Cancer
2002
;
2
:
563
72
.
12.
Ghajar
CM
,
Peinado
H
,
Mori
H
,
Matei
IR
,
Evason
KJ
,
Brazier
H
, et al
The perivascular niche regulates breast tumour dormancy
.
Nat Cell Biol
2013
;
15
:
807
17
.
13.
Barkan
D
,
Kleinman
H
,
Simmons
JL
,
Asmussen
H
,
Kamaraju
AK
,
Hoenorhoff
MJ
, et al
Inhibition of metastatic outgrowth from single dormant tumor cells by targeting the cytoskeleton
.
Cancer Res
2008
;
68
:
6241
50
.
14.
Fluegen
G
,
Avivar-Valderas
A
,
Wang
Y
,
Padgen
MR
,
Williams
JK
,
Nobre Ana
R
, et al
Phenotypic heterogeneity of disseminated tumour cells is preset by primary tumour hypoxic microenvironments
.
Nat Cell Biol
2017
;
19
:
120
32
.
15.
Paszek
MJ
,
Zahir
N
,
Johnson
KR
,
Lakins
JN
,
Rozenberg
GI
,
Gefen
A
, et al
Tensional homeostasis and the malignant phenotype
.
Cancer Cell
2005
;
8
:
241
54
.
16.
Discher
DE
. 
Tissue cells feel and respond to the stiffness of their substrate
.
Science
2005
;
310
:
1139
43
.
17.
Schrader
J
,
Gordon-Walker
TT
,
Aucott
RL
,
van Deemter
M
,
Quaas
A
,
Walsh
S
, et al
Matrix stiffness modulates proliferation, chemotherapeutic response, and dormancy in hepatocellular carcinoma cells
.
Hepatology
2011
;
53
:
1192
205
.
18.
Budday
S
,
Nay
R
,
de Rooij
R
,
Steinmann
P
,
Wyrobek
T
,
Ovaert
TC
, et al
Mechanical properties of gray and white matter brain tissue by indentation
.
J Mech Behav Biomed Mater
2015
;
46
:
318
30
.
19.
Mueller
S
,
Sandrin
L
. 
Liver stiffness: a novel parameter for the diagnosis of liver disease
.
Hepat Med
2010
;
2
:
49
67
.
20.
Evans
A
,
Whelehan
P
,
Thomson
K
,
McLean
D
,
Brauer
K
,
Purdie
C
, et al
Quantitative shear wave ultrasound elastography: initial experience in solid breast masses
.
Breast Cancer Res
2010
;
12
:
R104
.
DOI: 10.1186/bcr2787
.
21.
Espina
V
,
Mariani
BD
,
Gallagher
RI
,
Tran
K
,
Banks
S
,
Wiedemann
J
, et al
Malignant precursor cells pre-exist in human breast dcis and require autophagy for survival
.
PLoS One
2010
;
5
:
e10240
.
22.
Cufí
S
,
Vazquez-Martin
A
,
Oliveras-Ferraros
C
,
Martin-Castillo
B
,
Vellon
L
,
Menendez
JA
. 
Autophagy positively regulates the CD44 <sup>+</sup>CD24 <sup>-/low</sup>breast cancer stem-like phenotype
.
Cell Cycle
2011
;
10
:
3871
85
.
23.
Wolf
J
,
Dewi
DL
,
Fredebohm
J
,
Müller-Decker
K
,
Flechtenmacher
C
,
Hoheisel
JD
, et al
A mammosphere formation RNAi screen reveals that ATG4A promotes a breast cancer stem-like phenotype
.
Breast Cancer Res
2013
;
15
:
R109
.
DOI: 10.1186/bcr3576
.
24.
Schoenlein
PV
,
Periyasamy-Thandavan
S
,
Samaddar
JS
,
Jackson
WH
,
Barrett
JT
. 
Autophagy facilitates the progression of ERalpha-positive breast cancer cells to antiestrogen resistance
.
Autophagy
2009
;
5
:
400
3
.
25.
Liang
XH
,
Jackson
S
,
Seaman
M
,
Brown
K
,
Kempkes
B
,
Hibshoosh
H
, et al
Induction of autophagy and inhibition of tumorigenesis by beclin 1
.
Nature
1999
;
402
:
672
6
.
26.
Lu
Z
,
Luo
RZ
,
Lu
Y
,
Zhang
X
,
Yu
Q
,
Khare
S
, et al
The tumor suppressor gene ARHI regulates autophagy and tumor dormancy in human ovarian cancer cells
.
J Clin Invest
2008
.
27.
Gupta
A
,
Roy
S
,
Lazar
AJF
,
Wang
WL
,
McAuliffe
JC
,
Reynoso
D
, et al
Autophagy inhibition and antimalarials promote cell death in gastrointestinal stromal tumor (GIST)
.
Proc Natl Acad Sci
2010
;
107
:
14333
8
.
28.
White
E
. 
Deconvoluting the context-dependent role for autophagy in cancer
.
Nat Rev Cancer
2012
;
12
:
401
10
.
29.
Pang
M-F
,
Siedlik
MJ
,
Han
S
,
Stallings-Mann
M
,
Radisky
DC
,
Nelson
CM
. 
Tissue stiffness and hypoxia modulate the integrin-linked kinase ILK to control breast cancer stem-like cells
.
Cancer Res
2016
;
76
:
5277
87
.
30.
McKnight
AL
,
Kugel
JL
,
Rossman
PJ
,
Manduca
A
,
Hartmann
LC
,
Ehman
RL
. 
MR elastography of breast cancer: preliminary results
.
Am J Roentgenol
2002
;
178
:
1411
7
.
31.
Sack
I
,
Beierbach
B
,
Hamhaber
U
,
Klatt
D
,
Braun
J
. 
Non-invasive measurement of brain viscoelasticity using magnetic resonance elastography
.
NMR Biomed
2008
;
21
:
265
71
.
32.
Mariappan
YK
,
Glaser
KJ
,
Levin
DL
,
Vassallo
R
,
Hubmayr
RD
,
Mottram
C
, et al
Estimation of the absolute shear stiffness of human lung parenchyma using <sup>1</sup>H spin echo, echo planar MR elastography: lung EpiMRE
.
J Magn Reson Imaging
2014
;
40
:
1230
7
.
33.
Yin
M
,
Talwalkar
JA
,
Glaser
KJ
,
Manduca
A
,
Grimm
RC
,
Rossman
PJ
, et al
Assessment of hepatic fibrosis with magnetic resonance elastography
.
Clin Gastroenterol Hepatol
2007
;
5
:
1207
13
e2.
34.
Simi
AK
,
Anlaş
AA
,
Stallings-Mann
M
,
Zhang
S
,
Hsia
T
,
Cichon
M
, et al
A soft microenvironment protects from failure of midbody abscission and multinucleation downstream of the EMT-promoting transcription factor snail
.
Cancer Res
2018
;
78
:
2277
89
.
35.
Lee
K
,
Chen
QK
,
Lui
C
,
Cichon
MA
,
Radisky
DC
,
Nelson
CM
. 
Matrix compliance regulates Rac1b localization, NADPH oxidase assembly, and epithelial-mesenchymal transition
.
Mol Biol Cell
2012
;
23
:
4097
108
.
36.
Jordan
VC
. 
Tamoxifen: a most unlikely pioneering medicine
.
Nat Rev Drug Discovery
2003
;
2
:
205
13
.
37.
Kisanga
ER
,
Gjerde
J
,
Guerrieri-Gonzaga
A
,
Pigatto
F
,
Pesci-Feltri
A
,
Robertson
C
, et al
Tamoxifen and metabolite concentrations in serum and breast cancer tissue during three dose regimens in a randomized preoperative trial
.
Clin Cancer Res
2004
;
10
:
2336
43
.
38.
Nakamura
T
,
Hashimoto
I
,
Sawada
Y
,
Mikami
J
,
Yoshimoto
M
,
Nishidai
H
, et al
[5-Fluorouracil concentration in various tissues from cancer patients after oral administration of 5-fluorouracil]
.
Gan To Kagaku Ryoho
1984
;
11
:
1037
48
.
39.
Fan
P
,
Wang
J
,
Santen
RJ
,
Yue
W
. 
Long-term treatment with tamoxifen facilitates translocation of estrogen receptor α out of the nucleus and enhances its interaction with EGFR in MCF-7 breast cancer cells
.
Cancer Res
2007
;
67
:
1352
60
.
40.
Vera-Ramirez
L
,
Vodnala
SK
,
Nini
R
,
Hunter
KW
,
Green
JE
. 
Autophagy promotes the survival of dormant breast cancer cells and metastatic tumour recurrence
.
Nat Commun
2018
;
9
:
1944
.
41.
Klionsky
DJ
,
Abdelmohsen
K
,
Abe
A
,
Abedin
MJ
,
Abeliovich
H
,
Acevedo Arozena
A
, et al
Guidelines for the use and interpretation of assays for monitoring autophagy (3rd edition)
.
Autophagy
2016
;
12
:
1
222
.
42.
Xie
Z
,
Klionsky
DJ
. 
Autophagosome formation: core machinery and adaptations
.
Nat Cell Biol
2007
;
9
:
1102
9
.
43.
Furuya
N
,
Yu
J
,
Byfield
M
,
Pattingre
S
,
Levine
B
. 
The evolutionarily conserved domain of beclin 1 is required for Vps34 binding, autophagy and tumor suppressor function
.
Autophagy
2005
;
1
:
46
52
.
44.
Hannigan
G
,
Troussard
AA
,
Dedhar
S
. 
Integrin-linked kinase: a cancer therapeutic target unique among its ILK
.
Nat Rev Cancer
2005
;
5
:
51
63
.
45.
Paget
S
. 
The distribution of secondary growths in cancer of the breast, 1889
.
Cancer Metastasis Rev
1989
;
8
:
98
101
.
46.
Waks
AG
,
Winer
EP
. 
Breast cancer treatment: a review
.
JAMA
2019
;
321
:
288
.
47.
Novaro
V
. 
Collagen-IV and laminin-1 regulate estrogen receptor expression and function in mouse mammary epithelial cells
.
J Cell Sci
2003
;
116
:
2975
86
.
48.
Levine
B
,
Kroemer
G
. 
Autophagy in the pathogenesis of disease
.
Cell
2008
;
132
:
27
42
.
49.
Thorburn
J
,
Andrysik
Z
,
Staskiewicz
L
,
Gump
J
,
Maycotte
P
,
Oberst
A
, et al
Autophagy controls the kinetics and extent of mitochondrial apoptosis by regulating PUMA levels
.
Cell Rep
2014
;
7
:
45
52
.
50.
Fitzwalter
BE
,
Towers
CG
,
Sullivan
KD
,
Andrysik
Z
,
Hoh
M
,
Ludwig
M
, et al
Autophagy inhibition mediates apoptosis sensitization in cancer therapy by relieving FOXO3a turnover
.
Dev Cell
2018
;
44
:
555
65.e3
.
51.
Levy
JMM
,
Towers
CG
,
Thorburn
A
. 
Targeting autophagy in cancer
.
Nat Rev Cancer
2017
;
17
:
528
42
.
52.
Cook
KL
,
Clarke
PAG
,
Parmar
J
,
Hu
R
,
Schwartz-Roberts
JL
,
Abu-Asab
M
, et al
Knockdown of estrogen receptor-α induces autophagy and inhibits antiestrogen-mediated unfolded protein response activation, promoting ROS-induced breast cancer cell death
.
FASEB J
2014
;
28
:
3891
905
.
53.
Guadamillas
MC
,
Cerezo
A
,
del Pozo
MA
. 
Overcoming anoikis - pathways to anchorage-independent growth in cancer
.
J Cell Sci
2011
;
124
:
3189
97
.
54.
Fung
C
,
Lock
R
,
Gao
S
,
Salas
E
,
Debnath
J
. 
Induction of autophagy during extracellular matrix detachment promotes cell survival
.
Mol Biol Cell
2008
;
19
:
797
806
.
55.
Han
S
,
Pang
M-F
,
Nelson
CM
. 
Substratum stiffness tunes proliferation downstream of Wnt3a in part by regulating integrin-linked kinase and frizzled-1
.
J Cell Sci
2018
;
131
:
jcs210476
.