Oncogenic KRAS Mutations Confer a Unique Mechanotransduction Response to Peristalsis in Colorectal Cancer Cells

Abstract

Colorectal cancer (CRC) tumors start as polyps on the inner lining of the colorectum, where they are exposed to the mechanics of peristalsis. Our previous work leveraged a custom-built peristalsis bioreactor to demonstrate that colonic peristalsis led to cancer stem cell enrichment in CRC cells. However, this malignant mechanotransductive response was confined to select CRC lines that harbored an oncogenic mutation in the KRAS gene. Here, we explored the involvement of activating KRAS mutations on peristalsis-associated mechanotransduction in CRC. Peristalsis enriched cancer stem cell marker LGR5 in KRAS mutant lines, in a Wnt-ligand-independent manner. Conversely, LGR5 enrichment in wild type KRAS lines exposed to peristalsis were minimal. LGR5 enrichment downstream of peristalsis translated to increased tumorigenicity in vivo. Differences in mechanotransduction was apparent via unbiased gene set enrichment analysis, where many unique pathways were enriched in wild type vs. mutant lines. Peristalsis also triggered β-catenin nuclear localization independent of Wnt-ligands, particularly in KRAS mutant lines. The involvement of KRAS was validated via gain and loss of function strategies. Peristalsis induced β-catenin activation and LGR5 enrichment depended on the activation of the MEK/ERK cascade. Taken together, our results demonstrated that oncogenic KRAS mutations conferred a unique peristalsis-associated mechanotransduction response to colorectal cancer cells, resulting in cancer stem cell enrichment and increased tumorigenicity. These mechanosensory connections can be leveraged in improving the sensitivity of emerging therapies that target oncogenic KRAS.

Introduction

Colorectal cancer (CRC) is the second most common cause of cancer related deaths in the United States^{1, 2}. To best improve patient outcomes with targeted therapeutics, it is important to first understand mechanistic dependencies during the initial stages of CRC development. CRC tumors begin as a precancerous polyp on the inner lining of the colon or rectum^{3, 4}. Precancerous polyps can then switch to progressive and malignant carcinomas (CRC). One early contributing factor in the malignant progression of CRC polyps is oncogenic mutations⁵. In CRC, mutations in the Kirsten rat sarcoma virus (KRAS) gene typically occur in the early stages of the polyp to malignant carcinoma transition^6–9. Importantly, clinical pathology of CRC tumors indicates that 35–40% of CRC patients harbor a KRAS mutation, making it an appealing starting point to identify new targets to improve patient prognosis^10–12.

In homeostasis, the KRAS gene relays signals from upstream growth factor receptors to downstream effector pathways such as RAF/MEK/ERK, MAPK, and PI3K, among others. These pathways control important processes such as cell cycle progression, cell survival, cell polarity and movement, actin cytoskeletal organization, and extracellular signal transduction^{13, 14}. Oncogenic activating mutations, such as G13D, G12D, and G12C, in the KRAS oncogene drive cell transformation and uncontrolled cell division^{15, 16} by locking KRAS in an active GTP-bound conformation, thereby driving constant downstream pathway activation^{17, 18}. Clinical analysis of CRC tumor samples indicates that activating KRAS mutations are associated with increased metastasis, CRC recurrence, and overall poorer patient survival^{7, 19–21}.

As a key hallmark of cancer progression, activating oncogenic mutations continue to be influenced by the tumor microenvironment. Mechanical forces in the tumor microenvironment are an important contributor to malignant progression, modifying behaviors such as cellular migration, and stemness^{22, 23}. Oncogenic activation of KRAS alters cellular mechanics via cytoskeletal organization and actomyosin contractility²⁴, thereby altering how cells sense their physical microenvironment. For example, when breast epithelial lines are transformed with activating KRAS mutations, increasing substrate stiffness resulted in proliferative and self-renewing colony formation²⁵, indicative of malignant progression.

In the CRC tumor microenvironment, cells in the intestinal lumen, including cells in precancerous polyps, are continuously exposed to mechanical forces associated with colonic peristalsis^{26, 27}. Given the prevalence of KRAS mutations in early CRC tumorigenesis, and the propensity of KRAS mutations to drive disproportionate mechano-responses, we investigated the hypothesis that KRAS mutations may drive a unique mechano-response to peristalsis in CRC.

In previous studies, we developed and validated a peristalsis bioreactor that mimics the forces associated with colonic peristalsis²⁸, a nuanced combination of both shear stress and cyclic multi-axial strain. During the initial validation of the peristalsis bioreactor, we demonstrated how peristalsis forces elicited a unique mechanoresponse from mesenchymal stem cells compared to static, shear, or strain forces alone²⁸. Further, using the peristalsis bioreactor, we determined that HCT116 KRAS^MUT cells exposed to peristalsis increased cancer stem cell markers relative to static cultures, while untransformed intestinal epithelial cells did not²⁹. HCT116 cells also developed invasive signatures downstream of peristalsis, indicating malignant progression. Motivated by our previous studies and the amplified mechanosensing reported in other tumors harboring KRAS mutations, we hypothesized that activating KRAS mutations will alter peristalsis-associated mechanotransduction in CRC.

Leveraging the peristalsis bioreactor, multiple cell lines and patient-derived xenograft lines were employed to test the hypothesis that activating KRAS mutations change mechanotransduction via cancer stem cell enrichment. Our focus was on the Leucine rich repeat containing G protein couple receptor, LGR5, a cancer stem cell marker which our previous work strongly demonstrated was downstream of Wnt activation²⁹. Here, we characterize changes in the Wnt pathway as a function of KRAS mutation status in response to peristalsis. To confirm the involvement of mutant KRAS in the altered mechanotransduction response to peristalsis, we employed gain and loss of function techniques for KRAS and its downstream effectors. This work identifies a connection between activating KRAS mutations and altered cellular response to peristalsis in CRC that to our knowledge, has not been demonstrated before. Results from this work will be vital to improving our understanding of how activating KRAS mutations impact CRC malignant progression via mechanotransduction.

2. Materials and Methods

2.1. Materials

Cell culture reagents were purchased from ThermoFisher Scientific (Waltham, MA) unless otherwise specified. Cell lines were purchased from American Type Culture Collection (ATCC; Manassas, VA; RRID:SCR_001672) unless otherwise specified. Polydimethylsiloxane (PDMS) was purchased from DOW Chemical (Midland, MI). All other chemical reagents were purchased from Sigma Aldrich (St. Louis, MO; RRID:SCR_008988), unless otherwise indicated. Antibodies used for cellular staining were purchased from Santa Cruz Biotechnology (Dallas, TX; RRID:SCR_008987), unless otherwise indicated. Custom-made oligos, including CRISPR reagents, were purchased from Integrated DNA Technologies (Coralville, IA; RRID:SCR_025813). All other molecular biology-grade reagents were purchased from ThermoFisher Scientific (Waltham, MA; RRID:SCR_008452).

2.2. Cell Culture

Three characterized colorectal cancer cell lines HCT116 (RRID:CVCL_0291), LS174T (RRID:CVCL_1384), and RKO (RRID:CVCL_0504) were used in this work. Additionally, three patient derived xenograft lines, PDX1, PDX2, and PDX3, established from primary patient samples³⁰, were utilized to assess the effects of peristalsis (PDX lines were gifts from the Kopetz lab at the University of Texas MD Anderson Cancer Center). All cells in this work were used before passage 15. Dulbecco’s Modified Eagle Medium (DMEM), Eagle’s Minimum Essential Media (EMEM, ATCC), and Roswell Park Memorial Institute (RPMI) 1640 Medium supplemented with 10% heat-inactivated fetal bovine serum (Peak Serum, Inc., Wellington, CO) and 1X Antibiotic-Antimycotic solution were used as the primary growth media for HCT116, LS174T and RKO, and PDX cells, respectively. All cells were cultured in standard 2D tissue culture flasks and treated with 0.25% trypsin to dissociate adherent cells. All cultures were routinely screened for mycoplasma contamination. Molecular characterization of the cell and PDX lines are presented in Supp. Table 1.

2.3. Preparation of Cell Seeded Membranes and Bioreactor Assembly

Polydimethylsiloxane (PDMS) membranes were prepared at a 10:1 ratio using previously established protocols^{28, 29}. To maximize seeding of cells on PDMS, the cell seeding area of the membranes were coated with 200 μg/mL of Collagen I. Cells were seeded onto PDMS at monolayer confluency of 500,000 cells/mL and allowed to adhere at 37°C for 4 hours. Cell attachment was confirmed via visual inspection with a cell culture microscope.

Operation of the peristalsis bioreactor was followed as previously reported^{28, 29}. Briefly, media was removed from each cell-seeded PDMS membrane. Then, cell-seeded peristalsis membranes were placed into the peristalsis bioreactor bottom while static membranes were maintained in cell culture dishes with 1 mL of fresh media. The bioreactor was assembled using commercially available zip ties and connected to the peristalsis pump. The assembled set up was placed into the 37°C incubator and connected to an Arduino that ran a pre-programmed code with the following parameters: 0.4 Pa shear and 15% cyclic strain at 12 rpm^{28, 29}. Cells were incubated, either as static controls or stimulated in the peristalsis bioreactor at 37°C for 24 hours and then collected for downstream analysis.

Control untreated conditions were maintained in static membranes, or peristalsis bioreactors in their respective growth medium. To determine the effects of inhibiting Wnt in some experimental conditions, a porcupine-selective pan-Wnt inhibitor (LGK974, 1 μM; Med Chem Express, Monmouth Junction, NJ) was incorporated into the growth media. To study the effect of MEK inhibition, after cells adhered to PDMS for 4 hours, media was removed from cell-seeded PDMS and growth media supplemented with Selumetinib (50 nM; Selleck Chemicals, Houston, TX) was added. Following overnight incubation, Selumetinib (50 nM) was incorporated into the growth media for appropriate peristalsis conditions.

2.4. Flow Cytometry Analysis of LGR5 Cancer Stem Cell Phenotype

Following 24-hour maintenance in static or exposure to peristalsis conditions, cells were detached using 0.25% trypsin and collected from PDMS into single cell suspensions in FACS buffer (Phosphate Buffered Saline (PBS) supplemented with 2% fetal bovine serum). Flow cytometry methods were performed using protocols optimized previously for cancer cells^{29, 33, 34}. Briefly, cells were incubated with AlexaFluor488-LGR5 antibody (R&D Systems, Minneapolis, MN; RRID:AB_3652857), or an isotype matched AlexaFluor-488 antibody (R&D Systems; RRID:AB_10718683) for 30 minutes at 37°C. Unbound antibody was removed by washing and resuspending in fresh FACS buffer and cells were analyzed on the Attune NxT flow cytometer (ThermoFisher Scientific). Isotype controls were used to establish a gating strategy by cutting off a background gate at 0.5% (gating strategy is demonstrated in Supp. Fig. 1–4). Based on the background gate, the percentage of cells expressing LGR5 was determined. Comparisons were drawn between static and peristalsis conditions.

2.5. Sample Collection and Processing of RNA-Seq Data

Cells were exposed to static or peristalsis conditions for 24 hours; RNA was extracted directly from the bioreactor membranes following manufacturer’s instructions from the RNeasy Mini Kit (Qiagen, Hilden, Germany; RRID:SCR_008539). RNA concentration was assessed using the Qubit Fluorometer (Life Technologies; Carlsbad, CA) and RNA quality was evaluating used the Tape Station (Agilent; Santa Clara, CA). Total RNA samples with adequate quality (RIN ≥6.0) and amount (≥5 ng/μl with >500ng) were sent to Azenta Life Sciences (Burlington, MA) for RNA-Sequencing.

Raw data files in FASTQ format were generated from the Illumina sequencer obtained from Azenta Life Sciences. Initial preprocessing involved quality control using FastQC software (RRID:SCR_014583) to trim adapters and low-quality bases. Clean reads were then mapped to the Human genes GRCh38.p13 reference genome available on ENSEMBL (RRID:SCR_002344) using the STAR aligner v.2.5.2b (RRID:SCR_004463). Differential expression analysis was conducted using DESeq2 (R package; RRID:SCR_000154), accounting for biological replicates, to identify genes with a fold change greater than 2 and an adjusted p-value (FDR) less than 0.1.

2.5.1. Weighted Gene Correlation Network Analysis (WGCNA)

K-means clustering was performed to visualize the expressed genes using iDEP tool suit³⁵. Functional enrichment analysis utilized EnRichGO (R library), on select genes grouped through WGCNA with minimum module size of 20 and a soft threshold of 15 to best fit the network structure using pickSoftThreshold function, considering the gene’s association with biological processes. All statistical analyses were conducted in R studio (v2023.12.1+402) and iDEP with significance thresholds adjusted based on the study design and corrected for multiple testing where applicable.

2.5.2. Gene Set Enrichment Analysis (GSEA)

Normalized count files from DESeq2 analysis performed by Azenta Life Sciences were populated into the Gene Set Enrichment Analysis Software (https://www.gsea-msigdb.org/gsea/downloads.jsp; RRID:SCR_003199)^{36, 37}. Analysis was run with two different gene sets: i) Gene Ontology Biological Process (GOBP)^{38, 39} and ii) Hallmark⁴⁰ MSigDB gene sets. For each analysis, gene set files, gene count files, and phenotype label files were loaded in GSEA software⁴¹. To reach a normalized enrichment score (NES), gene set permutations were conducted 1000 times⁴². The minimum and maximum criteria for gene sets selection were 1 and 1000 genes, respectively. The NES normalizes the enrichment score to the size of the gene set. To account for false positives in multiple testing, an NES was considered significant with a false discovery rate (FDR) less than 0.25⁴¹. A higher NES indicates more positively enriched genes in the gene set.

2.6. Assessment of Tumorigenicity in NSG Mice

HCT116 KRAS^MUT and RKO KRAS^WT cells were either maintained as static controls or stimulated with peristalsis in the bioreactor for 24 hours. Cells were harvested using 0.25% trypsin and prepped for subcutaneous injection into NOD.Cg-Prkdc^scid Il2rg^tm1Wjl/SzJ (NSG; RRID: BCBC_4142) mice. These methods were adapted from previous protocols established by Raghavan, et al.³⁴. Animal experiments were performed following approval by the Texas A&M University Institutional Animal Care and Use Committee under protocol 2020–0040 D. Three male mice and three female mice were used per test condition and randomly assigned to static and peristalsis groups to maintain minimal statistical power and illustrate reproducibility. All mice were 8–10 weeks of age and male mice were 27–30lb while female mice were 20–25lb.

Following harvest, a total cell count was obtained and then adjusted to a concentration of 140,000 cells per 50 μL of media. Ice cold Matrigel was added to cell concentrations at a ratio of 1:1 Matrigel:media. Each mouse received two subcutaneous flank injections (one on either side of the body) of 100 uL each (i.e. 140,000 cells per injection).

For two weeks following injections, mouse weight was monitored twice a week and injection sites were palpated for any tumor presence. Once tumors were present, measurements were obtained three times a week using a caliper to record tumor volumes (Eq. 1) and mouse weight was continually monitored.

$$\mathrm{V}\mathrm{o}\mathrm{l}\mathrm{u}\mathrm{m}\mathrm{e}=\left(\mathrm{m}\mathrm{i}\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{a}\mathrm{x}\mathrm{i}\mathrm{s}\mathrm{*}{\mathrm{m}\mathrm{a}\mathrm{j}\mathrm{o}\mathrm{r}\mathrm{a}\mathrm{x}\mathrm{i}\mathrm{s}}^2\right)/2$$ Equation 1:

Tumor volumes were recorded until tumors reached an end point of 2,000 to 2,500 mm³ and then mice were euthanized. Tumors were dissected, cleaned, and placed in biopsy cassettes for processing for histology and hematoxylin and eosin (H&E) staining. Histology and H&E slides were prepared by the Texas A and M University Veterinary Medicine and Biomedical Sciences Research Histology Unit Core Facility (RRID:SCR_022201). H&E slides were imaged with a Leica DM 6B upright microscope (Wetzlar, Germany) at the Texas A and M University Microscopy and Imaging Center Core Facility (RRID:SCR_022128).

Cells were also isolated from the tumors using previously established protocols^{30, 34}. Isolated cells were evaluated for LGR5⁺ cancer stem cell expression using flow cytometry following protocols in section 2.4. Cells were also collected for culture until 1 passage was reached whereby flow cytometry of the LGR5⁺ cancer stem cell phenotype was reassessed.

2.7. KRAS^MUT CRISPR Mutation in RKO Cells

All components for CRISPR procedures were purchased from Integrated DNA Technologies (IDT; Coralville, IA). The HDR donor template and associated Cas9 guide RNA (crRNA) was designed using the Alt-R^™ HDR Design Tool with IDT to create a KRAS^G13D mutation⁴³. Potential off-target effects of crRNA candidates were analyzed using the Alt-R^™ HDR Design Tool⁴³, and the crRNA sequences with the fewest off-target sites in the human genome were selected for further analysis. The target sequences of the crRNA and HDR Oligos used in this study are shown in Supp. Table 2. Additional components purchased included Alt-R CRISPR-Cas9 tracrRNA, Alt-R S.p. Cas9-GFP V3, Alt-R^® Cas9 Electroporation Enhancer, and Alt-R HDR Enhancer V2. Following manufacturer protocols, the guide RNA (gRNA) complex was created by a 1:1 combination of the designed crRNA:tracrRNA. Then, the ribonucleoprotein complex was created with a 1.2:1.0 molar ratio of gRNA:Cas9-GFP supplemented with PBS, per manufacturer’s protocols⁴³.

One day prior to electroporation, media was changed in RKO cells under passage number three in cell culture flasks. On the day of electroporation, RKO cells were detached using 0.25% trypsin and 1 million cells were spun down and resuspended in PBS. The ribonucleoprotein complex, electroporation enhancer, and HDR donor oligos (+ and −) were added to the suspended cells. The mixture was transferred to a pre-cooled (4°C) 0.4mm electroporator cuvette (BioRad, Hercules, CA). Electroporation was performed on the BioRad GenePulser XCell with the following protocol parameters: square, 110V, 25ms, 1 pulse. The protocol was run two times and then cells were transferred to the pre-warmed cell culture flask, with culture medium supplemented with HDR Enhancer V2. Cells were cultured in growth media for 48 hours and examined with a Leica DMi8 microscope (Wetzlar, Germany) to verify GFP expression. Cells were allowed to grow for an additional 48 hours prior to experimental use or collection for knock-in validation. To validate the KRAS mutation, a sample of cells was lysed in RIPA buffer, for subsequent western blot analysis of phosphorylated ERK and total ERK compared to non-electroporated RKO cells (see Section 2.8 for Western Blot procedure).

2.8. Western Blot Analysis of ERK Activity

Phosphorylation of ERK was used to determine the activation status of KRAS, and to verify CRIPSR mutation methods employed in RKO cells. Cells exposed to static or peristalsis conditions for 24 hours were detached using 0.25% trypsin, collected from PDMS and lysed in 100 μL of Radio-immunoprecipitation assay (RIPA) Buffer supplemented with 1 μL Halt Protease Inhibitor Cocktail. Extracted protein concentration was measured using the Pierce^™ BCA Protein Assay Reagent following manufacturer’s protocol for a 96 well format. Following protein quantification, 10 μg of protein from each sample were loaded onto 4–20% gradient polyacrylamide gels (Novex^™, ThermoFisher), and separated via electrophoresis. Protein was transferred to a PVDF membrane and blocked with 2.5% bovine serum albumin (BSA). Membranes were probed with primary antibody (pERK; RRID:AB_2533719) overnight at 4°C, washed with TBST buffer, and probed with an appropriate HRP-conjugated secondary antibody (RRID:AB_2533967). α-tubulin (RRID:AB_1965960) was used as a loading control to determine changes in phosphorylated ERK (pERK) among samples. KwikQUANT Imager (Kindle Biosciences) was used to visualize bands on the blots with ultra digital ECL substrate solution (Kindle Biosciences). Densitometry was performed using NIH Image J. Band intensities were normalized by dividing the intensity obtained from each protein band to their corresponding loading control band intensity.

2.9. WNT Pathway Gene Expression Analysis

Following exposure to static or peristalsis conditions for 24 hours, RNA was lysed directly from the PDMS membrane using the RNeasy Mini Kit (Qiagen, Hilden, Germany). RNA concentration and purity were evaluated using a NanoDrop OneC (ThermoFisher Scientific) and stored at −80°C until ready to use. Reverse transcription was performed following manufacturer’s protocols using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems). qPCR was performed with a QuantStudio5 (Applied Biosystems) using the Applied Biosystems PowerUp SYBR Green PCR Mastermix (Thermofisher Scientific) for detection. Genes that were investigated included canonical Wnt ligands: WNT1 (Wnt family member 1), WNT7b (Wnt family member 7b), WNT8a (Wnt family member 8a), and noncanonical Wnt ligands: WNT4 (Wnt family member 4), WNT5a (Wnt family member 5a), WNT5b (Wnt family member 5b). The primer sequences used for each gene are shown in Supp. Table 3. Changes in gene expression were calculated using the 2ΔΔC_t method, with GAPDH as the housekeeping control⁴⁴. qPCR experiments were run in triplicates, with 3–4 independent biological replicates. Comparisons were drawn between static controls and peristalsis.

2.10. Immunofluorescent Staining and Fluorometry of β-Catenin

Immunofluorescence of PDMS membranes exposed to peristalsis or maintained as static controls was performed using previously established protocols^{28, 29}. Following formalin fixation, and blocking and permeabilization, cell-seeded membranes were incubated with the fluorescently tagged primary antibody βcatenin-AlexaFluor647 (RRID:AB_626807) and a nuclear counterstain (DAPI) for 1 hour at room temperature. Unbound antibodies were rinsed using PBS and DI water and mounted using the ProLong^™ Diamond antifade mounting reagent (ThermoFisher). Fluorescence was observed using an Olympus Fluoview FV3000 Confocal Laser Scanning Microscope (Tokyo, Japan) with 5 independent, non-overlapping regions for analysis. NIH Image J (RRID:SCR_003070) was employed to perform fluorometry. β-catenin activation was quantified as the mean fluorescent intensity of β-catenin (pink) located within the nucleus (blue) (Supp. Fig. 5). Mean intensity values from each test condition were compared to their respective static controls to produce a fold change. Microscopy was performed at the Texas A&M Health Science Center Integrated Microscopy and Imaging Laboratory Core Facility (RRID:SCR_021637).

2.11. Statistical Analysis

Statistical analysis was performed on GraphPad Prism 10 (RRID:SCR_002798). All reported values result from 3–5 independent biological replicates. In vivo tumor and xenograft harvest analyses result from 10–12 independent biological replicates. All qPCR data was normalized to static conditions within each experimental set and performed in triplicates over at least 3 biological replicates. Image analysis and morphometry included 5 non-overlapping fields of view from 3 biological replicates. Two-way ANOVA, t-test, and one-way ANOVA analyses were performed as appropriate, and statistical significance is indicated within each experimental data set with associated p-values. A p-value of less than 0.05 was considered significant. Box and whiskers plots were used to represent collected data. Box values range from the 25^th to 75^th percentiles with a line at the median and whiskers extend from the smallest value to the largest value.

2.12. Data Availability

The data generated in this work are publicly available in a Texas Data Repository with the following doi: https://doi.org/10.18738/T8/RXKKND. RNA-Seq data is deposited in the Gene Expression Omnibus under “GSE270759” (RRID:SCR_005012). RNA-Seq codes are deposited in a Code Ocean compute capsule found at https://codeocean.com/capsule/5943906.

Results

Peristalsis enhances LGR5 expression in CRC cells harboring activating KRAS mutations

To establish a link between LGR5 expression and KRAS activating mutations in CRC, cell and PDX lines were exposed to peristalsis or maintained as static controls for 24 hours prior to flow cytometry analysis of LGR5, a CRC stem cell marker. The gating strategy for all flow analysis is provided in Supp. Fig. 1–4. In HCT116 KRAS^MUT, peristalsis significantly increased LGR5⁺ cells by 1.7-fold compared to static controls (**p<0.01, two-way ANOVA; Fig. 1A–B; representative flow analysis plots and quantification). Similar to HCT116 KRAS^MUT, exposure to peristalsis increased LGR5⁺ expression 2.5–3.8-fold in LS174T KRAS^MUT and PDX1 KRAS^MUT cells (****p<0.0001, **p<0.01; two-way ANOVA; Fig. 1B). In contrast, exposure of RKO KRAS^WT cells to peristalsis did not increase LGR5 expression compared to static controls (ns, two-way ANOVA; Fig. 1C–D; representative flow analysis plots and quantification). Similarly, KRAS^WT lines, PDX2 (*p<0.05) and PDX3 (ns) resulted in no LGR5 enrichment compared to static controls (two-way ANOVA; Fig. 1D).

Figure 1: LGR5⁺ cancer stem cell enrichment in KRAS mutant CRC cells exposed to peristalsis.

(A) Representative LGR5 flow cytometry plots of HCT116 KRAS^MUT cells maintained in static controls or exposed to peristalsis in the bioreactor. (B) Box and whiskers plots summarizing flow analysis of KRAS mutant LGR5⁺ expression (%) after 24hr exposure to peristalsis bioreactor or maintenance in static controls. Significant increase in LGR5⁺ expression was noted in all cell types exposed to peristalsis compared to static controls (**p<0.01, ****p<0.0001, two-way ANOVA). (C) Representative LGR5 flow cytometry plots of RKO KRAS^WT cells maintained in static controls or exposed to peristalsis in the bioreactor. (D) Box and whiskers plots summarizing flow analysis of KRAS wild type LGR5⁺ expression (%) after 24hr exposure to peristalsis bioreactor or maintenance in static controls (*p<0.05, ns, two-way ANOVA). (E) Gene ontology pathway analysis using Weighted Gene Correlation Network Analysis in HCT116 KRAS^MUT and RKO KRAS^WT cells exposed to peristalsis. The top 9 GO terms from each population are displayed.

To evaluate peristalsis-associated responses in HCT116 KRAS^MUT cells compared to RKO KRAS^WT, global RNA-Seq gene expression analysis methods were employed. Gene ontology pathway analysis using Weighted Gene Correlation Network Analysis (WGCNA) between HCT116 KRAS^MUT and RKO KRAS^WT cells identified 6 unique pathways when assessing the top 9 GO terms for each population (Fig. 1E). Unbiased bioinformatics analysis demonstrated that HCT116 KRAS^MUT cells showed enrichment in response to stimulus, adhesion, and Wnt signaling gene sets, while RKO KRAS^WT cells showed enrichment for developmental processes, cell differentiation, and proliferation gene sets. Importantly, many of the genes involved in the GOBP Wnt signaling pathways are associated with increased cell stemness and cancer progression^{45, 46}. Specifically, FZD7 activates both the canonical and noncanonical Wnt signaling pathway in colon cancer leading to downstream effects in stemness and even cell migration^{47, 48}. Further, TCF7 is a canonical Wnt signaling gene shown to increase migration and invasion as well as adhesion in HCT116 cells⁴⁹. The RKO KRAS^WT gene enrichment sets demonstrate a focus on basic developmental processes related to maintenance of cell behavior with genes such as PGDFA and BMPR2^{50, 51}.

Peristalsis drives increased tumorigenicity in HCT116 KRAS^MUT tumors

We next assessed the role peristalsis plays in driving tumor growth in response to the KRAS activating mutation in vivo. HCT116 KRAS^MUT and RKO KRAS^WT cells were pre-exposed to peristalsis or static control conditions for 24 hours, then injected subcutaneously into NOD.Cg-Prkdc^scid Il2rg^tm1Wjl/SzJ (NSG) mice (Fig. 2A). Representative images of harvested tumors showed overall differences in tumor sizes across conditions (Fig. 2B). HCT116 KRAS^MUT tumors initiated with peristalsis-conditioned cells were larger compared to static tumors (Fig. 2B). Interestingly, the size differences in peristalsis versus static conditions was not observed in RKO KRAS^WT tumors (Fig. 2B).

Figure 2: Increased tumor growth in HCT116 KRAS^MUT mutant tumors in peristalsis-conditioned cells.

(A) Depiction of in vivo subcutaneous xenograft procedure from cell collection to analysis method. Cell lines used for in vivo studies and their corresponding KRAS status information is included in the table. Cells were cultured statically or exposed to peristalsis for 24 hours prior to collection and subsequent subcutaneous injection into the left and right flank of NSG mice. Mouse weight and tumor volume measurements were continuously monitored. (B) Representative images of harvested tumors from static- and peristalsis-conditioned HCT116 KRAS^MUT and RKO KRAS^WT cells. Scale bar 1cm. (C) Tumor initiation and growth kinetics are shown for both HCT116 KRAS^MUT and RKO KRAS^WT. In HCT116 KRAS^MUT, peristalsis tumors (pink curve) demonstrated elevated tumorigenicity, with faster tumor initiation and tumor burden development (gray shaded area) compared to static tumors (black curve; ns, **p<0.01, ***p<0.001, ****p<0.0001, two-way ANOVA). RKO KRAS^WT tumors demonstrated overlap in growth kinetics with no significant differences in tumor initiation nor development (ns, two-way ANOVA). (D) Histological examination of H&E stains demonstrates that both static and peristalsis tumors in HCT116 KRAS^MUT cells revealed highly invasive cell morphology. Harvested RKO KRAS^WT tumors in peristalsis and static conditions demonstrated less invasive cell morphology with necrosis present. Scale bar 100μm.

To support endpoint visual analysis, growth rates were tracked throughout tumor growth. Overall growth kinetics in the HCT116 KRAS^MUT mice were significantly different between the static and peristalsis conditions (Fig. 2C). HCT116 peristalsis tumors developed at a faster rate of 177.1 mm³/day while static tumors were 0.58-fold slower at a rate of 103.1 mm³/day in the first week of growth (days 14–21). Peristalsis tumors reached the maximum tumor burden window much earlier (Day 24), compared to static tumors in HCT116 (Day 28; Fig. 2C). These differences in growth kinetics of the tumors appear as divergent curves between the static and peristalsis conditions in HCT116 KRAS^MUT.

In the RKO KRAS^WT tumors, the growth kinetics were similar in both the static and peristalsis conditions (almost overlapping static and peristalsis growth curves in Fig. 2C). RKO peristalsis tumors grew at a rate of 60.9 mm³/day and static tumors were only 0.78-fold slower at 47.8 mm³/day in the first week of growth (days 14–21). Both peristalsis and static tumors reached the tumor burden window on Day 30 with little difference in overall growth patterns (Fig. 2C). Histologic examination of harvested HCT116 xenografts revealed highly invasive cell morphology in both static and peristalsis conditions (Fig. 2D). In harvested RKO xenografts, histological examination demonstrated necrosis and less invasive cell morphology across both static and peristalsis conditions (Fig. 2D).

Since there were divergent growth rates in peristalsis conditioned HCT116 KRAS^MUT tumors, we tested the hypothesis that LGR5 enrichment would be seen across the xenograft process. At the termination of the growth studies, harvested xenografts were dissociated and collected for flow LGR5 enrichment analysis on the day of harvest and following the first passage of isolated cells. HCT116 KRAS^MUT xenografts showed increased LGR5 expression in the peristalsis conditioned group, both on the day of harvest and at the time of the first passage (****p<0.0001, two-way ANOVA compared to xenografts from static groups, Fig. 3A). LGR5 expression analysis from RKO KRAS^WT xenografts demonstrated no significant differences between static and peristalsis conditions on the day of harvest or at the time of the first passage (ns, two-way ANOVA, Fig. 3B). Note that no additional peristaltic mechanical stimulation was provided following xenograft tumor harvests, yet LGR5 enrichment persisted in the HCT116 KRAS^MUT cell line, while it did not in the RKO KRAS^WT cell line.

Figure 3: Peristalsis-conditioned HCT116 KRAS^MUT mutant xenografts maintain enriched LGR5⁺ cancer stem cell expression upon harvest and passage.

(A) Representative LGR5⁺ flow analysis plots and quantification for harvested HCT116 KRAS^MUT xenografts at passage 0 (P0) and passage 1 (P1). Both P0 and P1 analysis demonstrated increased LGR5 enrichment in peristalsis compared to static-conditioned cells (****p<0.0001, two-way ANOVA). (B) Representative LGR5⁺ flow analysis plots and quantification for harvested RKO KRAS^WT xenografts at passage 0 (P0) and passage 1 (P1). P0 and P1 flow cytometry analysis resulted in no differences between peristalsis and static-conditioned cells (ns, two-way ANOVA).

The introduction of the KRAS mutation in WT RKO cells increased LGR5 expression

To test the sufficiency of the KRAS mutation in driving LGR5 expression, we generated a CRISPR mediated KRAS mutation in RKO cells. Pathway activation downstream of active KRAS mutation was functionally validated by western blot analysis of phosphorylated ERK (pERK). Band intensities were normalized to their corresponding α-tubulin loading control. The CRISPR edited RKO KRAS mutation resulted in a 17% increase in phosphorylated ERK expression compared to RKO KRAS^WT cells (*p<0.05, t-test, Fig. 4A). Once increased pERK activity was confirmed as a consequence of introducing the KRAS mutation into RKO cells, we tested the hypothesis that peristalsis would result in LGR5 enrichment, similar to that observed in the HCT116 KRAS^MUT cell line. As predicted, exposure of RKO KRAS^MUT cells to peristalsis resulted in a 3.17-fold increase in LGR5 expression compared to static controls (****p<0.0001, t-test, Fig. 4B).

Figure 4: Introduction of the KRAS mutation in WT RKO cells increases LGR5 expression and alters mechanoresponse.

(A) Densiometric analysis of western blots for phosphorylated ERK (pERK) activity. Band intensities were normalized to their corresponding α-tubulin loading control. Introduction of the KRAS mutation into RKO cells resulted in increased pERK expression (*p<0.05, t-test). (B) Representative analysis graphs and box and whiskers plot of flow cytometry LGR5 expression (%) after 24hr exposure to static or peristalsis in RKO KRAS^MUT cells (****p<0.0001, t-test). (C) Gene ontology pathway evaluation using Weighted Gene Correlation Network Analysis in RKO KRAS^MUT cells exposed to peristalsis. The top 9 GO terms from the tested population are displayed. (D) Gene set enrichment analysis plots for ERK1_and_ERK2_cascade (GO:0070371) and response_to_mechanical_stimulus (GO:0009612) selected from significantly enriched (FDR < 0.25) gene ontology terms for RKO KRAS^WT and RKO KRAS^MUT cells. Plots include the profile of the running enrichment score and positions of gene set members on the rank-ordered list. Normalized enrichment scores (NES) are also reported where a higher NES indicates more positively enriched genes in the gene set.

In tandem, RNA-seq analysis was employed to evaluate gene expression pathway involvement in peristalsis-associated mechanotransduction in RKO cells with the KRAS mutation (Fig. 4C). Unbiased analysis with WGCNA plots showed that 5 unique pathways were identified out of the top 9 GO terms between RKO KRAS^WT (Fig. 1E) and RKO KRAS^MUT cells (Fig. 4C). In the analysis of GO pathways with WGCNA, RKO KRAS^MUT cells showed altered responses from RKO KRAS^WT with significant enrichment in motility, migration, and mechanical response gene sets. With motility and migration gene sets in the RKO KRAS^MUT cells, SNAIL1, ANG, and SERPINE2 are key genes known to increase cancer progression through the induction of epithelial to mesenchymal transition and migration^52–54. Further, Gene Set Enrichment Analysis (GSEA) was performed against the entire gene ontology biological process gene set to look at overall enrichment differences. Closer examination of significant (FDR <0.25), relevant biological processes led to the selection of the ERK1_and_ERK2_cascade (GO:0070371) and response_to_mechanical_stimulus (GO:0009612) plots for comparison (Fig. 4D). Interestingly, RKO KRAS^MUT cells had a higher normalized enrichment score (NES) for both gene sets compared to RKO KRAS^WT cells, indicating global changes following the introduction of the KRAS mutation.

Peristalsis results in Wnt-independent β-catenin activation in KRAS mutant cells

The Wnt pathway is a known driver of cancer cell proliferation, stemness, and malignant transformation⁵⁵, including driving the enrichment of LGR5. Given the differences in LGR5 expression between KRAS WT and Mutant lines, we evaluated if there were similar differences in Wnt pathway involvement in KRAS^WT and KRAS^MUT cell populations (Fig. 5). We first assessed Wnt ligand expression via qPCR gene expression. In the HCT116 KRAS^MUT line, cells exposed to peristalsis significantly increased gene expression of many canonical and noncanonical Wnt ligands compared to static controls (WNT7b, WNT4, WNT5a, and WNT5b; Fig. 5A, Supp. Fig. 3). In contrast, only a mild increase in non-canonical Wnt ligands (WNT4, WNT5a, and WNT5b) were observed in RKO KRAS^WT cells as a result of peristalsis, differing quite significantly from the patterns observed in the HCT116 KRAS^MUT cell line (Fig. 5A, Supp. Fig. 3). Importantly, when the KRAS mutation was introduced into RKO, peristalsis again resulted in significant increases in Wnt ligand genes similar to the HCT116 KRAS^MUT condition (WNT1, WNT7b, WNT4, WNT5a, and WNT5b; Fig. 5A, Supp. Fig. 3).

Figure 5: Peristalsis drives Wnt-independent β-catenin activation in KRAS mutant cells.

(A) Heat map of Wnt ligand gene expression. Gene expression changes in peristalsis for HCT116 KRAS^MUT, RKO KRAS^WT and RKO KRAS^MUT cells are demonstrated as fold-change values compared to static controls (indicated by 1.0). Statistical differences between each cell type exposed to peristalsis and static controls individually can be found in Supplementary material (Supp. Fig. 6). (B) Representative flow cytometry analysis graphs and box and whisker plot of flow cytometry LGR5 expression (%) following 24 hour exposure to peristalsis or static conditions with Wnt inhibition (WNTi). HCT116 KRAS^MUT and RKO KRAS^MUT resulted in sustained increases in LGR5 expression in peristalsis compared to static controls (**p<0.01, *p<0.05, two-way ANOVA). (C) Representative micrographs of cells maintained as static controls or exposed to peristalsis with and without Wnt inhibition (WNTi) stained with β-catenin (magenta) and counterstained with DAPI (blue). Scale bar 10 μm. Bar graphs quantifying nuclear β-catenin localization relative to their respective static controls for all tested conditions. Peristalsis increased nuclear localization of β-catenin compared to static controls both with and without Wnt inhibition in HCT116 KRAS^MUT and RKO KRAS^MUT (****p<0.0001, t-test). RKO KRAS^WT cells exposed to peristalsis with and without Wnt inhibition did not observe increases in β-catenin nuclear localization. Detailed quantification methods and channel separated images are found in supplementary information (Supp. Fig. 2 and 4).

Wnt ligand gene expression trends implicated the involvement of the Wnt pathway in response to peristalsis. This led us to test the hypothesis that inhibiting pan-Wnt secretion with LGK974^{56, 57} would abrogate peristalsis-associated LGR5⁺ enrichment observed in the KRAS mutant cell lines. Interestingly, HCT116 KRAS^MUT cells still resulted in a significant 2.5-fold increase of LGR5 expression following peristalsis compared to static controls with Wnt inhibition (**p<0.01, two-way ANOVA, Fig. 5B). As expected, neither Wnt inhibition nor peristalsis made a difference in LGR5 levels in RKO KRAS^WT cells (ns, two-way ANOVA, Fig. 5B). Interestingly, despite Wnt inhibition in the KRAS mutated RKO cells, peristalsis continued to result in a significant 3.1-fold increase in LGR5 expression (*p<0.05, two-way ANOVA, Fig. 5B). Further, compared to uninhibited peristalsis, Wnt inhibition did not significantly decrease LGR5 expression in any of the tested cell lines (Supp. Fig. 9A).

Having demonstrated that LGR5 enrichment downstream of peristalsis was independent of Wnt ligand expression in KRAS mutant cells, we then evaluated the activation of the main Wnt pathway effector, β-catenin. We used immunofluorescence to localize β-catenin expression in cells exposed to peristalsis with or without the pan-Wnt secretion inhibitor (WNTi; Fig. 5C). Activated β-catenin localized to the nuclear region, indicated nuclear translocation of β-catenin (detailed methods for quantification provided in Supp. Fig. 5 and representative channel separated images are provided in Supp. Fig. 7). All test conditions were normalized to their respective static conditions to report as a fold-change. Exposure to peristalsis significantly increased β-catenin activation by 1.25-fold in HCT116 KRAS^MUT cells, relative to static controls (****p<0.0001, t-test, Fig. 5C). RKO KRAS^WT cells resulted in no significant change in peristalsis, relative to static controls (ns, t-test, Fig. 5C). Similar to HCT116 KRAS^MUT cells, RKO KRAS^MUT cells demonstrated a 1.60-fold increase in peristalsis exposed cells relative to static controls (****p<0.0001, t-test, Fig. 5C). Interestingly, even when Wnt secretion was inhibited, β-catenin activation was sustained by exposure to peristalsis in the mutant HCT116 KRAS^MUT and RKO KRAS^MUT cell lines, suggesting that the β-catenin activation is not being triggered by Wnt (****p<0.0001, t-test, Fig. 5C).

Peristalsis-associated LGR5 enrichment is sensitive to MEK inhibition

To directly evaluate the dependency of KRAS downstream effectors in regulating LGR5 expression in response to peristalsis, we assessed the effects of Selumetinib (50 nM), a highly selective inhibitor of mitogen-activated protein kinase kinase (MEK) and extracellular signal-regulated kinase (ERK). First, we verified that phosphorylated ERK was reduced with selumetinib, resulting in a 22.8% decrease of phosphorylated ERK in HCT116 KRAS^MUT cells compared to uninhibited controls (****p<0.0001, t-test, Fig. 6A). Although the ERK suppression was modest, we chose this low dose of Selumetinib to preserve cell viability^{58, 59}. It was included in the media circulating through the peristalsis bioreactor (MEKi), and cells were collected for analysis of LGR5 expression. Upon MEK inhibition via Selumetinib, exposure to peristalsis no longer resulted in LGR5 enrichment in the HCT116 KRAS^MUT cell line compared to inhibited static cells (****p<0.0001, two-way ANOVA, Fig. 6B). Similarly, both RKO KRAS^WT cells and RKO KRAS^MUT cells demonstrated a significant decrease in LGR5 expression in MEK inhibited peristalsis compared to inhibited static controls (***p<0.001, *p<0.05, two-way ANOVA, Fig. 6D). Further, compared to uninhibited peristalsis, MEK inhibition in peristalsis only demonstrated a significant decrease in LGR5 expression in the mutant cell lines (Supp. Fig. 9B). Together, these studies link MEK activation to peristalsis-induced enrichment of LGR5.

Figure 6: LGR5 enrichment downstream of peristalsis is sensitive to MEK inhibition.

(A) Densiometric analysis of western blots for phosphorylated ERK (pERK) activity analysis in HCT116 KRAS^MUT with and without MEK inhibition (Selumetinib, 50 nM). Band intensities were normalized to their corresponding α-tubulin loading control. MEK inhibition in HCT116 KRAS^MUT cells resulted in decreased pERK expression (****p<0.0001, t-test). (B) Representative flow cytometry analysis graphs and box and whisker plot of flow cytometry LGR5 expression (%) following 24-hour exposure to peristalsis or static conditions with MEK inhibition (MEKi). All tested cells demonstrated no enrichment in LGR5 expression in peristalsis compared to static controls with MEK inhibition (****p<0.0001, ***p<0.001, *p<0.05, two-way ANOVA).

MEK inhibition in KRAS^MUT mutant cells abrogates peristalsis induced Wnt and β-catenin activation

Since MEK inhibition (MEKi) with Selumetinib abrogated peristalsis-associated LGR5 enrichment in KRAS mutant lines, we explored the effect of MEK inhibition on Wnt ligand expression as well as on β-catenin activation. Despite MEK inhibition, peristalsis continued to increase gene expression of many Wnt ligands in HCT116 KRAS^MUT cells (WNT1, WNT7b, WNT4, WNT5a, and WNT5b; Fig. 7A, Supp. Fig. 8). Consistent with previous observations, there were no significant Wnt ligand increases in the RKO KRAS^WT cells downstream of peristalsis, with MEK inhibition (Fig. 7A, Supp. Fig. 8). Introducing the KRAS mutation into the RKO WT line modestly increased Wnt ligand expression downstream of peristalsis with a significant increase in WNT1 and modest increase in WNT8a, WNT5a, and WNT5b in RKO KRAS^MUT, relative to static controls (Fig. 7A, Supp. Fig. 8). Ultimately, MEK inhibition impacted peristalsis induced Wnt ligand expression less significantly in KRAS mutant cells compared to RKO KRAS^WT cells.

Figure 7: MEK inhibition reduces peristalsis induced Wnt and β-catenin activation in KRAS mutant cells.

(A) Heat map of Wnt ligand gene expression. Gene expression changes in peristalsis with MEK inhibition (MEKi; Selumetinib 50 nM) for HCT116 KRAS^MUT, RKO KRAS^WT and RKO KRAS^MUT cells are demonstrated as fold-change values compared to static controls (indicated by 1.0). Statistical differences between each cell type exposed to peristalsis with MEK inhibition and static controls individually can be found in Supplementary material (Supp. Fig. 8). (B) Representative micrographs of cells exposed to peristalsis with and without MEK inhibition (MEKi) stained with β-catenin (magenta) and nuclear counterstained with DAPI (blue). Scale bar 10 μm. Bar graphs quantifying nuclear β-catenin localization relative to static untreated controls for all tested conditions. Peristalsis with MEK inhibition decreased nuclear localization compared to peristalsis alone in both HCT116 KRAS^MUT and RKO KRAS^MUT (*p<0.05, ****p<0.0001, two-way ANOVA). RKO KRAS^WT cells exposed to peristalsis with and without MEK inhibition did not observe differences in β-catenin nuclear localization. Detailed quantification methods and channel separated images are found in supplementary information (Supp. Fig. 5 and 7).

β-catenin activation in peristalsis was similarly evaluated for its response to MEK inhibition (Fig. 7B, Supp. Fig. 7). Compared to peristalsis alone, MEK inhibition resulted in a 12.3% decrease of β-catenin activation in KRAS^MUT HCT116 cells (*p<0.05, two-way ANOVA, Fig. 7B). RKO KRAS^MUT cells pheno-copied this sensitivity to MEK inhibition, with a 23.9% decrease in β-catenin activation observed during uninhibited peristalsis (****p<0.0001, two-way ANOVA, Fig. 7B). Our data strongly suggest the involvement of the activating KRAS mutation in sustaining β-catenin activation downstream of peristalsis since RKO KRAS^WT cells displayed no difference with or without MEK inhibition.

Discussion

Mechanics of the colorectal tumor microenvironment (peristalsis) are complex, with propagating waves of multi-axial strain and shear forces always present^{26, 27}. In order to study mechanotransduction in response to colonic peristalsis, we previously custom-built a peristalsis bioreactor that mimicked mechanical patterns of the colon in vitro²⁸. The use of mesenchymal stem cells in our previous work²⁸ and HCT116 CRC cells (Supp. Fig. 10) confirmed that peristalsis forces elicit a unique mechanoresponse compared to static, shear, or strain forces alone. Peristalsis is a unique combination of forces holistically represented in the peristalsis bioreactor, as confirmed by previous computational modeling²⁸. Leveraging the peristalsis bioreactor, we demonstrated that peristalsis resulted in malignant progression of colorectal cancer cells, enriching cancer stem cells and increasing invasive features in mechanically stimulated cells²⁹. Interestingly, the mechanotransductive response to peristalsis was specific to cancer cells that carried activating KRAS mutations, since LGR5⁺ enrichment was not observed in mechanically stimulated non-cancerous intestinal epithelial cells nor KRAS wild type cancer cells. Overall, this led us to hypothesize that activating KRAS mutations might alter mechano-responsiveness of colorectal cancer cells to peristalsis.

Oncogenic KRAS mutations (G13D, G12D, G12C, etc.) lead to malignant transformation via the constant activation of the mitogen-activated protein kinase (MAPK) cascade^{15, 16}. Activation of RAS (through oncogenic mutations or extracellular signals) results in sequential activation of Raf and MEK kinases, ultimately leading to the phosphorylation and activation of extracellular signal-regulated kinase (ERK)⁶⁰. In CRC, oncogenic KRAS mutations are associated with decreased patient survival^{7, 19–21}. Interestingly, in breast and pancreatic cancers, oncogenic KRAS alters cancer cell mechanosensing²⁵. Given the near constant exposure of colorectal cancer cells to colonic peristalsis, we hypothesized that a similar KRAS-specific mechanosensitive phenotype might exist in CRC.

First, we focused on Leucine-rich repeat-containing G-protein-coupled receptor 5 (LGR5), due to previous observations that peristalsis resulted in LGR5 enrichment in colorectal cancer cells²⁹. In this work, our data further demonstrated that LGR5 enrichment downstream of peristalsis was not universal and confined to oncogenic KRAS mutant cells (Fig. 1). To further support this result, we introduced the KRAS^MUT mutation into RKO KRAS^WT – to acknowledge that they are BRAF mutant cells but lack the KRAS mutation. Interestingly, when mutant KRAS^MUT was induced into the WT background cell line, exposure to peristalsis led to LGR5 enrichment (in RKO KRAS^WT vs. RKO KRAS^MUT cells; Fig. 1, 4). Despite the BRAF mutation in RKO cells, LGR5 changes were evident between RKO KRAS^WT and RKO KRAS^MUT. This data was also backed by observations from unbiased RNA-Seq analysis. Global gene expression analysis via WGCNA demonstrated there were 6 unique pathways between HCT116 KRAS^MUT and RKO KRAS^WT. With 67% of the top 9 GO pathways altered in the unmodified KRAS WT and Mutant cell lines, this demonstrated how KRAS mutation status may alter cellular response to peristalsis at the gene expression level. Specifically, enrichment of Wnt pathway effectors like FZD7 and TCF7 in the HCT116 KRAS^MUT cells was indicative of enhanced stemness and migration,^47–49. The highlights from RKO KRAS^WT instead focused on basic cellular processes via PGDFA and BMPR2^{50, 51}. Further, the introduction of the KRAS mutation to the RKO KRAS^WT cells resulted in 5 unique pathways between RKO KRAS^WT and RKO KRAS^MUT out of the top 9 GO terms in response to peristalsis, with the shifts primarily occurring in motility and migration-based gene sets. SNAIL1, ANG, and SERPINE2 were three key genes enriched in RKO KRAS^MUT cells that drive cancer progression via epithelial to mesenchymal transition and increased migration^52–54. WGCNA findings were also corroborated by altered gene set enrichment in the ERK1_and_ERK2_Cascade in RKO KRAS^WT versus RKO KRAS^MUT. Within the top 5 enriched genes of this pathway for RKO KRAS^MUT cells, the LIF gene negatively regulates the tumor suppressor gene p53⁶¹ and the GAREM1 gene regulates MAPK activation which controls cellular proliferation⁶². Thus, our results strongly suggest that oncogenic KRAS altered mechanosensing in response to peristalsis. These results are in line with evidence that oncogenic RAS signaling alters cytoskeletal organization and actomyosin contractility in epithelial cells²⁴. Both these avenues play important roles in cellular mechanotransduction, implicated in cellular ability to sense strain⁶³. In fact, emerging evidence suggests that oncogenic RAS increases the elastic modulus of many types of cancer cells, conferring growth advantages in confined spaces via altered mechanosensing^{64, 65}. To the best of our knowledge, our work is the first instance where we report that oncogenic RAS signaling alters mechanotransductive signaling in response to a tissue-level mechanical stimulus, like colonic peristalsis.

Having established the role of oncogenic KRAS-specific mechanotransduction leading to LGR5 enrichment, we then evaluated if a functional advantage was also present. Since increased LGR5 expression is correlated with poorer prognosis in vivo^66–68, we hypothesized that the differences in LGR5 expression observed in the HCT116 KRAS^MUT and RKO KRAS^WT cells would translate to altered tumor growth kinetics. As we expected, xenografts generated from peristalsis-exposed HCT116 KRAS^MUT cells grew at a significantly faster rate compared to RKO KRAS^WT cells, likely due to the significant LGR5 enrichment due to peristalsis (Fig. 2, 3). This was unsurprising, since the LGR5⁺ population is associated with increased tumor growth, invasion, and therapy resistance in CRC clinical samples^69–72. KRAS mutations conferred a functional tumor growth advantage, evident when growth rates of xenografted tumors from HCT116 KRASMUT cells were compared to RKO KRAS^WT cells, maintained as static controls or exposed to peristalsis. We undertook this functional follow-up in these two cell lines motivated by the stark differences and divergence in peristalsis driving increases in LGR5 expression in vitro in both these cell types. The caveat in this experimental methodology however is that ultimately, these are two non-isogenic cell lines. A significant future scientific opportunity is to explore the functional consequences of different activating oncogenic KRAS mutations in otherwise isogenic WT and oncogenically activated lines.

In many studies of oncogenic KRAS related mechanosensing, the focus has typically stayed on cell response to environmental stiffness⁶⁴. Since our readout of malignant progression was focused on quantifying increases in LGR5 expression, we focused on the intersection of the Wnt pathway and oncogenic KRAS. Our motivations were two-fold: i) our previous work demonstrated a connection between Wnt activation and subsequent LGR5 expression²⁹; and ii) LGR5 is a target gene of Wnt signaling^73–75. As expected, Wnt ligand gene expression was highly increased in response to peristalsis in HCT116 KRAS^MUT (Fig. 5A). These results were somewhat unsurprising as HCT116 cells harbor a CTNNB1 mutation that prevents the degradation of β-catenin leading to enhanced Wnt pathway signaling^76–78. While RKO KRAS^WT cells demonstrated mild changes in Wnt ligands, the introduction of the KRAS^MUT into RKO cells (despite being CTNNB1^WT) resulted in a similar genotype to HCT116 cells in response to peristalsis. Wnt/β-catenin signaling is activated in response to mechanics in many non-epithelial cell types⁷⁹. For example, mechanical stretch in skeletal osteoblasts and oscillatory shear stress in lymphatic endothelial cells have been cited to increase Wnt/β-catenin signaling and their downstream targets^{80, 81}. Our results demonstrate that Wnt ligand increases occur downstream of peristalsis mechanics in epithelial cancer cells, but only in those that harbor oncogenic KRAS mutations. Surprisingly, inhibiting all Wnt ligand secretion via a porcupine-inhibitor^{56, 57} still demonstrated significant increases between static and peristalsis conditions in cases where oncogenic KRAS is active (Fig. 5B). Further, when compared to peristalsis alone, Wnt inhibition with peristalsis was insufficient in decreasing LGR5⁺ expression in the HCT116 KRAS^MUT cell line (Supp. Fig. 9).

Importantly, peristalsis resulted in β-catenin nuclear translocation independent of Wnt-ligand inhibition, but only in cells that harbored oncogenic KRAS mutations (Fig. 5C). This phenomenon was independently tested with both a gain of function mutation in the RKO KRAS^WT cell line, and with inhibition of MEK/ERK in the HCT116 KRAS^MUT, RKO KRAS^WT, and RKO KRAS^MUT cell lines. Our observations are in line with reported evidence that cytoskeletal stiffness-mediated mechanics can trigger mechanical activation of β-catenin in cancer cells independent of Wnt ligands⁷⁹. This Wnt-ligand-independent mechanical activation of β-catenin subsequently led to cancer stem cell enrichment. Similarly, another report demonstrated that external solid mechanical stresses from the growing colorectal tumor microenvironment also activate β-catenin, leading to enhanced tumor growth⁸². Our data corroborates that β-catenin activation occurs in response to the mechanics of peristalsis, but only in cells harboring oncogenic KRAS mutations. The intersection of oncogenic RAS and β-catenin activation has also previously been reported in many different cancers either via PI3K/AKT or MEK/ERK activation^83–87. Hyper-activation of ERK1/2 or PI3K/AKT observed due to oncogenic KRAS requires β-catenin stabilization and activation in melanoma⁸⁴. Interestingly, RAS-driven hyper-activation and malignant progression in thyroid cancers depend on Wnt-independent β-catenin activation^{87, 88}. These reports reinforce our findings that peristalsis-driven, Wnt-ligand-independent β-catenin activation is unique to cells that harbor activating oncogenic KRAS mutations, where we see hyperactivation of ERK.

Lastly, the LGR5 enrichment and β-catenin activation phenotype downstream of peristalsis were sensitive to MEK inhibition in KRAS mutant cells (Fig. 6–7, Supp. Fig. 9). Importantly, while this confirms the role of oncogenic KRAS itself, it also opens out the possibility of MEK/ERK as potential druggable mechano-sensors in colorectal cancer. Evidence of their mechanosensory activity is found in non-epithelial cell literature, where cardiomyocytes and chondrocytes independently exposed to strain phosphorylate MEK/ERK^89–91. To the best of our knowledge, this is the first study that links MEK to a mechano-sensory response in KRAS mutant colorectal cancer, especially in response to the native mechanics of colonic peristalsis.

Collectively, this work supports the hypothesis that oncogenic KRAS mutations dictate a unique mechanotransduction response to peristalsis in colorectal cancer (CRC). With the application of peristalsis forces, CRC cells harboring a KRAS mutation demonstrated an altered functional phenotype and genotype that portends to increased tumor growth in vivo. Importantly, KRAS targeted therapeutic options are growing in the treatment of KRAS mutant cancers and can be used uniquely with other pharmacologic modulators that regulate colonic peristalsis to achieve favorable therapeutic outcomes. Our work supports the idea that the combination of peristalsis modulation and targeted KRAS therapeutics can be employed to modulate CRC progression and thereby increase patient survival.

Implications.

Oncogenic KRAS empowers colorectal cancer cells to harness the mechanics of colonic peristalsis for malignant gain, independent of other cooperating signals.