Collagen production and niche engineering: A novel strategy for cancer cells to survive acidosis in DCIS and evolve

Abstract

Growing tumors are dynamic and nonlinear ecosystems, wherein cancer cells adapt to their local microenvironment, and these adaptations further modify the environment, inducing more changes. From nascent intraductal neoplasms to disseminated metastatic disease, several levels of evolutionary adaptations and selections occur. Here, we focus on one example of such an adaptation mechanism, namely, "niche construction" promoted by adaptation to acidosis, which is a metabolic adaptation to the early harsh environment in intraductal neoplasms. The avascular characteristics of ductal carcinoma in situ (DCIS) make the periluminal volume profoundly acidic, and cancer cells must adapt to this to survive. Based on discovery proteomics, we hypothesized that a component of acid adaptation involves production of collagen by pre-cancer cells that remodels the extracellular matrix (ECM) and stabilizes cells under acid stress. The proteomic data were surprising as collagen production and deposition are commonly believed to be the responsibility of mesenchymally derived fibroblasts, and not cells of epithelial origin. Subsequent experiments in 3D culture, spinning disk and second harmonic generation microscopy of DCIS lesions in patients' samples are concordant. Collagen production assay by acid-adapted cells in vitro demonstrated that the mechanism of induction involves the RAS and SMAD pathways. Secretome analyses show upregulation of ECM remodeling enzymes such as TGM2 and LOXL2 that are collagen crosslinkers. These data strongly indicate that acidosis in incipient cancers induces collagen production by cancer cells and support the hypothesis that this adaptation initiates a tumor-permissive microenvironment promoting survival and growth of nascent cancers.

Keywords: K‐RAS; Niche construction and engineering; SMAD; anoikis; cancer evolution; collagen production; extracellular matrix remodeling enzymes; tumor microenvironment.

Figure 1

Acid adaptation prevents anoikis as a survival mechanism of early lesion cancer cells. (a) (b) Soft Agar clonogenic assay on MCF7 and AA MCF7 cells. AA MCF7 cells grow more and bigger clones than non‐adapted MCF7. (c) Spherogenicity assay on MCF7 and AA MCF7 cells. Acid‐adapted MCF7 cells grow sphere in ultra‐low adhesion plates contrary to the non‐adapted cells. (d) Proteomics analysis of AA MCF7 versus MCF7 cells. The analysis was applied on filtered proteins that have role in ECM and anoikis. More than 80% of the top ten pathways that were upregulated in acid‐adapted cells belong to cell adhesion, ECM–cell adhesion, and integrin‐mediated cell adhesion pathways implying the role of acid adaption in regulation of those pathways. (e) Spinning disk experiment to measure cell–matrix adhesion. AA MCF7 cells are more adhesive than MCF7 while the normal MCF10A cells have opposite behavior. (f) Western blot of FAK and pFAK on lysates from AA and NA MCF7 cells. (g) 3D viability assay of AA MCF7 and MCF7 spheroid treated with FAK inhibitors

Figure 2

Acid adaptation promotes collagen production of cancer cells. (a) Second harmonic generation (SHG) microscopy of breast tumors confirms the existence of collagen in the center of ducts in DCIS breast tumors. DCIS is the most acidic part of the tumor due to avascular nature of early carcinoma. (b) Hydroxyproline assay showed higher amount of hydroxyproline in acid‐adapted cancer cells compared to non‐adapted ones. Hydroxyproline is an essential precursor of the collagen fibers that is necessary for its stability. The data are repeated in three biological replicates, and data are presented as mean with SD as error bar. (c) Fluorescent confocal microscopy coupled with second harmonic generation operator confirmed the higher expression of collagen and fibrillary structures in acid‐adapted cancer cells. Collagen production was measured using SHG and confocal microscopy on paraformaldehyde fixed breast cancer cells. CNA‐35‐GFP marker was used to distinguish collagen from other repetitive structure detected by SHG. Acid‐adapted cells produce more collagen significantly. Data are shown as standard deviation with mean as error bars with three separated biological replicates. (d) Expression pattern of collagen genes in acid‐adapted MCF7 cells against non‐adapted ones. The expression response pattern is quite heterogeneous and varies from over expression to lower expression. The y axis scale is in logarithmic some genes such as Col17a1 are highly expressed in acid‐adapted cells while Col21a1 is overly downregulated. (e) SILAC proteomics analysis of AA MCF7 against NA MCF7 cells revealed increased expression of collagen production enzymes such as PLODs, and P4Has as well as some collagens in acid‐adapted MCF7 cells. Data are presented in two set that are two separate runs of mass spectrum. F) Western blot validation of PLOD higher expression in acid‐adapted cells compared to non‐adapted one

Figure 3

Acid‐induced collagen production is controlled by SMADs and K‐Ras. (a) Immunocytochemistry (ICC) of acid‐adapted MCF7 and non‐adapted MCF7 for SMAD4. SMAD4 is a transcription factor that controls the production of collagen when it is located in nucleus. (b) Nuclear localization of SMAD4 versus its cytoplasmic localization revealed the higher nuclear localization of this protein in acid‐adapted cells. (c) pSMAD3 staining of acid‐adapted and non‐adapted MCF7 cells showed upregulation of pSMAD3 in acid‐adapted cells. (d) Nuclear and cytoplasmic localization analysis of pSMAD3. SMAD3 is a cytoplasmic protein that binds to SMAD4 in its phosphorylated form and translocates SMAD4 into the nucleus. pSMAD3 has higher nuclear expression and lower cytoplasmic expression in acid‐adapted MCF7 cells compared to non‐adapted MCF7. (e) Western blot analysis shows Ras activation and downstream protein regulations. K‐Ras is activated in AA MCF7 and AA MCF10AT cancer cells. Downstream of K‐Ras is ERK that gets activated through phosphorylation. Western blot shows higher p‐ERK in acid‐adapted cells compared to controls. Higher level of LAMP2b is a marker of acid adaptation that shows the cells are completely acid adapted. (f) Tumor growth of acid‐adapted MCF10AT and non‐adapted ones injected to nude mice. AA MCF10AT cell frequency of forming tumor is higher, and the tumors that are formed grow faster. (g) Schematic of Ras activation role in collagen production in cancer cells. We proposed Ras activation phosphorylate ERK that can activate SMAD3 through phosphorylation. pSMAD3 will translocate SMAD4 to nucleus to activate collagen production genes

Figure 4

Acid‐adapted cells use collagen remodeling enzymes to engineer their niche. (a) TGM2 validation by (b) Western blot and (c) Immunocytochemistry (ICC). TGM2 expression in acid‐adapted cancer cells is significantly higher compared to non‐adapted cells. (d) Bicarbonate buffer therapy in animal reduced the expression of TGM2 compared to control group, indicating the effect of acidosis on expression of TGM2. (e, f) Translation of TGM2 expression to clinic and patient samples. LAMP2b is a marker of acidosis that we reported in our previous work. The TMA from same patients was stained for TGM2, and expression level of TGM2 and LAMP2b was compared for each sample. There is a correlation between LAMP2b and TGM2 expression in patients. (g) Acid‐adapted MCF‐7 cells not only have higher amount of TGM2 and LOXL2, but also secrete more of these enzymes to the environment. Both enzymes have been shown to play role in collagen crosslinking and stability. This strengthen our acid‐induced niche engineering phenotype of cancer cells that they build the niche they need it to survive the harsh environment such as acidosis

Figure 5

Acid‐induced collagen production promotes tumors early stage development and progress in different cancer models. Schematic of breast cancer tumors growth regarding the acidic microenvironment and collagen morphology and structure change. Solid tumors are profoundly and continuously acidic due to Pasteur and Warburg effect. The collagen morphology variation has been reported previously with mostly unknown trigger and mechanism. Here, we show how microenvironmental factors such as acidosis can promote the collagen production and its role in tumor evolution. At the early stages, acid‐adapted cancer cells inside the DCIS gain the ability of producing the extracellular matrix component such as collagen. This will enable them and probably their neighbors surviving the anoikis and progress toward next stage