Metabolic stress conditions dictate MAPKAPK2-dependent efficiency of MEK1/2 inhibition in colorectal carcinoma

Significance

Reduced efficacy of anticancer agents is a major challenge for cancer treatment and is a critical priority in cancer research. MAPKAPK2, an effector molecule of the p38 signaling pathway, is a stress responsive kinase and regulates colorectal carcinoma (CRC) growth in nutrient abundant conditions. On the contrary, glucose deprivation (a common feature of solid tumors) downregulates MAPKAPK2 expression in cancer cells to promote cell survival. In addition, CRC tumors with low expression of MAPKAPK2 are less responsive to trametinib (an FDA-approved MEK1/2 inhibitor) and could compromise glucose analog based therapies in patients. Taken together, these results indicate that therapeutic strategies for cancer patients could be stratified based on MAPKAPK2 expression.

Abstract

The kinase MAPKAPK2 regulates cell survival, proliferation, and death, and is upregulated in colorectal carcinoma (CRC) where it is associated with tumor growth and progression. However, how it regulates tumor progression in conjunction with other signaling pathways, such as MEK/ERK, remains elusive. Solid tumors are often subjected to metabolic stress, notably glucose deprivation. Here, we demonstrate that MAPKAPK2 protein levels in CRC regulate cell fate decision during stress conditions, such as glucose deprivation and therapeutic treatment. While MAPKAPK2 expression is a limiting factor for CRC growth in vitro, depleting MAPKAPK2 or inhibiting its activity pharmacologically provides a survival advantage to CRC cells under glucose limiting conditions. Subjecting CRC cells to low glucose resulted in an ERK1/2-mediated decline in MAPKAPK2 to promote survival. Additionally, cells with reduced MAPKAPK2 activity were less sensitive to trametinib under glucose limiting conditions. Utilizing transcriptomic profiling, we found that glucose deprivation and MAPKAPK2 depletion activate pathways associated with survival during metabolic stress. This relationship was also observed in CRC patients (TCGA), where tumors with low MAPKAPK2 expression had higher ERK1/2 activation and upregulated stress-induced pathways, leading to poor survival. Finally, MAPKAPK2 modulated growth of CRC organoids, subcutaneous tumors, and patient-derived xenografts (PDX), and reduced MAPKAPK2 levels decreased efficacy of trametinib, in vitro and in vivo. Overall, this study identifies an interrelationship between MEK/ERK and p38/MAPKAPK2 signaling pathways during glucose deprivation to support cell survival and features MAPKAPK2 loss as a possible mechanism leading to reduced efficacy of trametinib-based anticancer therapy and poor patient outcomes in CRC.

Colorectal cancer (CRC) accounts for 9.6% of all cancer cases worldwide and 9.3% of cancer-related deaths, making it the third most frequent and second most lethal cancer (1). CRC commonly exhibits metabolic reprogramming, shifting toward aerobic glycolysis, also known as the Warburg effect (2). The underlying high glycolytic rate supports biomass production, invasion, metastasis, angiogenesis, chemoresistance, and changes in the tumor microenvironment, and is correlated with poor prognosis in CRC. Genetic perturbations in major CRC drivers, including KRAS, TP53, PTEN, and PI3K, function to sustain the high rate of glycolysis via various molecular mechanisms (2). Nevertheless, targeting glucose metabolism in CRC as monotherapy has not been successful despite promising results in preclinical studies (3, 4). A major reason underlying this limited effect is that metabolically CRC is highly heterogeneous and displays a high degree of plasticity. Instead of reducing cancer cell fitness, inhibition of glycolysis, and promoting nutrient stress such as glucose deprivation, induces various survival pathways in CRC. For example, activation of HIF-1α signaling under glucose deprivation mounts an antiapoptosis response in colon cancer cells (5). Similarly, glucose deprivation increases phosphorylation of the extracellular signal-regulated kinase (ERK) in CRC cells, stabilizing JMJD2B expression, thereby upregulating GLUT1 expression via epigenetic modifications (6). These pathways are also known to impart CRC with therapy resistance (3). This raises the possibility that differences in the glucose availability within the tumor allows individual tumor cells to activate various survival pathways and respond differently to therapeutic agents. Therefore, mechanisms by which CRC cells survive under glycolysis inhibition and glucose limitation warrants further investigation to devise effective ways to harness their underlying metabolic vulnerabilities.

Mitogen-activated protein kinase (MAPK) signaling regulates various processes, including cell proliferation, cell survival, metabolic reprogramming, inflammatory responses, and stress response in cancer. It has three major pathways: ERK-MAPK, c-Jun N-terminal kinase/stress-activated protein kinase (JNK/SAPK), and p38. Each of these pathways perform distinct functions: the ERK-MAPK signaling cascade strongly promotes cell proliferation, cell survival, and their underlying metabolic changes; the JNK signaling pathway has been implicated in both cell survival and apoptotic processes; whereas the p38-MAPK plays a major role in apoptosis, differentiation, survival, proliferation, development, inflammation, and other stress responses (7–9). However, in a broader context, JNK/p38 MAPK signaling pathways show antagonistic effects toward ERK-MAPK signaling (10–12). Therefore, while these three main branches of MAPK signaling may have overlapping functions, each may have unique functions which may also function in a largely context-dependent manner (7, 13).

MAPK signaling modulators are actively being developed as anticancer agents. Targeting ERK signaling in CRC is a promising clinical intervention. Trametinib, a MEK1/2 inhibitor FDA approved for unresectable or metastatic solid tumors with BRAF V600E mutation in combination with Dabrafenib (14), is under investigation in several clinical trials in combination with other agents for CRC (15–17). However, these combination therapies have shown either minimal efficacy or no improvement in patient survival (18, 19). Therefore, understanding critical factor(s) regulating the downstream effects of MAPK and how metabolic plasticity in therapy responses could be a contributing factor may be necessary to develop more efficacious therapeutic regimens (20). The p38-MAPK signaling pathway, including downstream MAPKAPK2 (MK2), has emerged as a potential drug target. Recent studies suggest that the p38/MAPKAPK2 signaling pathway regulates glucose metabolism in cancer (21, 22). MAPKAPK2 is upregulated in CRC and plays an important role in the development of colitis-associated colon cancer and improves tumor growth and invasive potential in the Apc^min/+ mouse model (23, 24). Targeting the activated form of MAPKAPK2 by sophoridine inhibited growth and invasion in CRC (25). Therefore, cotargeting ERK/MAPK and p38/MAPKAPK2 pathways could be sensible for CRC. However, while these two pathways may play an antagonistic role in certain cancer contexts, it is important to define the crosstalk between p38-MAPKAPK2 and ERK-MAPK signaling pathways to design improved targeting strategies.

Here, we identified that MAPKAPK2 is a glucose-sensitive protein, and that it may work together with ERK to regulate CRC, driving survival under glucose-restrictive conditions while supporting cell proliferation when glucose is abundant. Furthermore, we also identified that MAPKAPK2 dictates the efficacy of trametinib-mediated growth inhibition in CRC both in vitro and in vivo.

Results

CRC Cells Are Less Sensitive to MEK1/2 Inhibition by Trametinib Treatment in Low Glucose.

Glucose availability to cancer cells supports a high proliferation rate, an important feature for cancer development and progression. After 72 h in culture with low glucose, CRC cell lines, HCT116 and DLD-1, showed a nearly 50% decrease in cell growth without any observable cell death (SI Appendix, Fig. S1 A–F). Similarly, 2-deoxyglucose (2-DG) treatment resulted in a dose-dependent reduction in cell viability (SI Appendix, Fig. S1 G–J). With the anticipation that extended culture of either cell line without glucose would abolish cell growth, we carried out colony formation assays. On the contrary, no significant change in colony number was observed (SI Appendix, Fig. S1 K–N). However, both colony size and perimeter were reduced by 45 to 55% and 64 to 71%, respectively, in the absence of glucose (SI Appendix, Fig. S1 K, L, and O–R). These data suggest that CRC cells can tolerate glucose deprivation by maintaining a reduced proliferation rate.

To understand molecular changes associated with CRC cell survival in low glucose, HCT116 cells grown with or without glucose for 24 h were subjected to RNA-sequencing. Differential gene expression analysis indicated a significant downregulation of 658 genes and upregulation of 693 genes upon glucose restriction (Fig. 1A). KEGG pathway analysis associated downregulated genes with cell cycle, meiosis, DNA replication, mismatch repair, spliceosome, and RNA transport (SI Appendix Fig. S2A). Similarly, Hallmark pathway analysis categorized these genes as E2F targets, G2M checkpoints, and Myc-regulated genes (SI Appendix, Fig. S2 B and C). A majority of these genes showed 20 to 40% reduction in expression and encode mitotic checkpoint proteins (BUB1B, BUB3, BUB1, MAD2L1), cyclin proteins (CCNA2, CCNB2, CCNE1, CCNE2), cell division cycle proteins (CDC25A, CDC20, CDC23, CDC25C), cyclin dependent kinases (CDK1, CDK2), a component of the origin recognition complex (ORC1), MCM complex (MCM2-7), clamp (PCNA), Replication protein A (RPA1, RPA2), and the helicase (FEN1) (SI Appendix, Fig. S2D). Reduction in mRNA expression of a subset of genes (FEN1, MCM2, MCM4, MCM6, MCM7, PCNA, CDC25A, ORC1) were also observed when DLD-1 cells were glucose-deprived for 24 h (SI Appendix, Fig. S2E). Top upregulated KEGG pathways included focal adhesion, microRNAs in cancer, MAPK signaling, circadian rhythm, p53 signaling pathway, ECM–receptor interaction, and proteoglycans in cancer (Fig. 1B). The Oncogenic Molecular Signatures Database (MSigDB) indicated upregulation of gene signatures associated with overexpressed EGFR, ERBB2, mTOR, and MEK (Fig. 1C). Upregulated Hallmark signatures included TNF-α signaling by NF-κB, hypoxia, estrogen response early, p53 pathway, unfolded protein response, cholesterol homeostasis, IL-2 STAT5 signaling, mTORC1, glycolysis, and apoptosis (Fig. 1D). Furthermore, we found that pathways we observed to be upregulated in our RNA-seq data were consistent with a prior analysis of glucose deprivation in HCT116 cells (GEO accession: GSE159351) (SI Appendix, Fig. S2 F–H). Focusing on the MAPK signaling pathway, we observed activation of ERK1/2 when CRC cells (HCT116, DLD-1, SW480, SW620, Caco-2, and HT-29) were maintained in media with 0.5 mM glucose for 24 h as compared to 20 mM glucose (Fig. 1 E–J). Connecting ERK1/2 activation to survival in low glucose, CRC cells were subjected to trametinib (a MEK1/2 inhibitor which inhibits phosphorylation of the downstream target ERK1/2) treatment. However, we observed that CRC cells were less responsive to trametinib treatment in 0.5 mM glucose compared to 20 mM glucose despite activation of ERK (Fig. 1 K–N and SI Appendix, Fig. S3 A and B). Taken together, these data indicate that while a lack of glucose supply leads to reduction in CRC cell proliferation, there is a simultaneous induction of cell survival pathway(s) which also provide an additional benefit to the CRC cells by promoting resistance to trametinib.

Fig. 1.

Colorectal carcinoma cells exhibit low sensitivity to trametinib treatment in low glucose conditions. (A) Volcano plot indicating differentially expressed genes (Log₂FoldChange > 0, −log₁₀ P-value ≥ 1.3) in HCT116 cells subjected to 0 mM glucose for 24 h as compared to 20 mM glucose. (B) Enrichment of KEGG pathways associated with upregulated genes in HCT116 cells subjected to 0 mM glucose for 24 h as compared to 20 mM glucose. (C) Enrichment of oncogenic molecular signatures (MS) pathways associated with upregulated genes in HCT116 cells subjected to 0 mM glucose for 24 h as compared to 20 mM glucose. (D) Enrichment of Hallmark pathways associated with upregulated genes in HCT116 cells subjected to 0 mM glucose for 24 h as compared to 20 mM glucose. (E–J) Immunoblots indicating increase in ERK1/2 phosphorylation in HCT116 (E), DLD-1 (F), SW480 (G), SW620 (H), Caco-2 (I), and HT-29 (J) cells subjected to 0.5 mM glucose-containing media for 16 to 18 h. (K and L) Cell viability (MTT assay) of HCT116 (K) and DLD-1 (L) cells subjected to indicated concentrations of trametinib in 20 mM and 0.5 mM glucose. (M and N) Cell viability of SW480 (M) and SW620 (N) cells subjected to indicated concentrations of trametinib in 20 mM and 0.5 mM glucose. The two-way ANOVA multiple comparison test was performed in K–N. Black arrows indicate first point of significant difference (***) in cell viability in K and L. *P < 0.05, **P < 0.01, ***P < 0.001.

MAPKAPK2 Is a Glucose-sensitive Kinase Associated with CRC Cell Survival in Low Glucose and Reduced Sensitivity to Trametinib.

Since ERK1/2 inactivation was ineffective in promoting cell death in low glucose conditions, we further investigated the role of another component of MAPK signaling, p38/MAPKAPK2. Contrary to ERK1/2 activation, we observed that the expression of MAPKAPK2, an effector molecule of p38 signaling, declines in low glucose (Fig. 2 A–F). Loss of MAPKAPK2 in response to low glucose had not been observed previously which led us to explore the consequences of MAPKAPK2 decline on the cellular response to nutrient stress. Interestingly, inhibition of MAPKAPK2 with PF-3644022 in HCT116, HT-29, SW480, SW620, and Caco-2 cells reduced sensitivity to glucose restriction (Fig. 2 G–J and SI Appendix, Fig. S4 A–C). Similar to inhibition by PF-3644022, HCT116, and DLD-1 cells depleted of MAPKAPK2 expression with shRNA showed enhanced survival when subjected to low glucose conditions for 72 h compared to control cells targeted with scrambled shRNA (Fig. 2 K–N and SI Appendix, Fig. S4 D and E). MAPKAPK2-depleted cells also showed enhanced survival in the presence of 2-DG (SI Appendix, Fig. S4 F and G). On the contrary, MAPKAPK2 overexpressing cells had reduced survival in low glucose as compared to control cells (SI Appendix, Fig. S4 H and I). Moreover, MAPKAPK2-depleted cells were less sensitive to trametinib inhibition compared to control cells in both low and normal glucose conditions (Fig. 2 O and P). Reconstitution of exogenous MAPKAPK2 partially reversed the response of MAPKAPK2-depleted cells to trametinib (Fig. 2 J and K). Collectively, these results indicated that the loss of MAPKAPK2 in response to low glucose is associated with CRC cell survival as well as following exposure to trametinib.

Fig. 2.

MAPKAPK2 (MK2) expression regulates survival in low glucose and trametinib treatment. (A–F) Immunoblots indicating decline in MAPKAPK2 expression in HCT116 (A), DLD-1 (B), HT-29 (C), SW480 (D), SW620 (E), and Caco-2 (F) cells following growth in 0.5 mM glucose-containing media for 16 to 18 h. (G–J) Cell viability (expressed as fold change) of HCT116 (G), HT-29 (H), SW480 (I), and SW620 (J) cells grown in 20 and 0-0.5 mM glucose-containing media and treated with MAPKAPK2 inhibitor (PF-3644022) for 72 h. (K and L) Immunoblots representing shRNA-mediated depletion of MAPKAPK2 in HCT116 (K) and DLD-1 (L). (M and N) Cell viability (measured as cell number and expressed as fold change) in MAPKAPK2-depleted HCT116 (M) and DLD-1 (N) cells grown in 0.5 mM glucose-containing media for 72 h. (O and P) Cell viability (measured as cell number and expressed as fold change) in MAPKAPK2-depleted HCT116 (O) and DLD-1 (P) cells grown in 0.5 mM glucose-containing media or in the presence of trametinib at indicated concentration or both for 72 h. The two-way ANOVA multiple comparison test was performed in G–J and M–P. *P < 0.05, **P < 0.01, ***P < 0.001.

ERK1/2 and MAPKAPK2 Regulate Each Other in a Feedback Loop.

Next, we set out to determine how MAPKAPK2 expression declines during low glucose conditions to support cell survival. First, we observed a dose-dependent decrease in MAPKAPK2 expression with decreasing glucose concentrations (Fig. 3 A and B). Furthermore, MAPKAPK2 protein expression declined in a time-dependent manner following transfer to low glucose media (Fig. 3C). We also tested if the upstream regulator of MAPKAPK2, p38, undergoes a similar fate upon glucose deprivation. However, p38 protein levels were unaltered in response to low glucose suggesting that this effect was specific to the MAPKAPK2 component of this pathway (Fig. 3C). Furthermore, supplementing glucose to overnight glucose-starved cells resulted in a time-dependent rescue in MAPKAPK2 levels in both HCT116 and DLD-1 cells (Fig. 3 D and E). qPCR results indicated that the decline in MAPKAPK2 protein expression upon glucose starvation was not due to changes in MAPKAPK2 mRNA levels, suggesting a posttranslational mechanism of regulating MAPKAPK2 protein levels (SI Appendix, Fig. S5 A–D).

Fig. 3.

MAPKAPK2 expression declines in low glucose. (A and B) Immunoblots indicating change in MAPKAPK2 expression in HCT116 (A) and DLD-1 (B) cells subjected to different concentrations of glucose for 16 to 18 h. (C) Immunoblots indicating change in MAPKAPK2 expression in HCT116 cells subjected to 20 mM and 0.5 mM glucose-containing media for indicated times. (D) Immunoblots indicating restoration of MAPKAPK2 expression in 18 h glucose-starved HCT116 cells subjected to 20 mM glucose-containing media for indicated times. (E) Immunoblots indicating restoration of MAPKAPK2 expression in 18 h glucose-starved DLD-1 cells subjected to 20 mM glucose-containing media for 18 h. (F) Immunoblots indicating rescue of MAPKAPK2 expression in HCT116 cells subjected to 0 mM glucose-containing media and treated with 10 μM and 20 μM MG132, MLN4924, and Heclin for 16 h. (G) Immunoblots indicating increased ubiquitination of MAPKAPK2 in low glucose (0.5 mM) condition in the presence of MG132 (2 μM). (H–J) Immunoblots indicating rescue of MAPKAPK2 levels in low glucose upon treatment with p38 inhibitor (SB 203580) in HCT116 (H), DLD-1 (I), and HT-29 (J). (K) Immunoblots indicating change in ERK1/2 activation in HCT116 cells subjected to 20 mM and 0.5 mM glucose-containing media for indicated times. (L and M) Immunoblots indicating rescue of MAPKAPK2 expression in HCT116 (L) and DLD-1 (M) cells subjected to 0.5 mM glucose-containing media and treated with trametinib for 24 h. (N and O) Immunoblots indicating increase in ERK1/2 activation in HCT116 (N) and DLD-1 (O) cells depleted of MAPKAPK2 (MK2) compared to shControl and nontransduced (NT) cells.

We next investigated if downregulation of MAPKAPK2 is mediated by the ubiquitin–proteasome system. A dose-dependent increase in MAPKAPK2 level was observed in HCT116 cells in low glucose condition upon treatment with the proteasome inhibitor MG132 (Fig. 3F). To assess targeting of MAPKAPK2 by various classes of E3 ubiquitin ligases, we treated with MLN4924 which inhibits neddylation of Cullin scaffold proteins thereby inhibiting the activity of Cullin-dependent E3 ubiquitin ligases (26), as well as an inhibitor of the Heclin class of E3 ubiquitin ligase (27). We observed MLN4924 showed modest restoration of MAPKAPK2 expression whereas Heclin had no effect (Fig. 3F). Similarly, treatment of HCT116 and HT-29 cells with the proteasome inhibitor, epoxomicin, rescued MAPKAPK2 expression in low glucose (SI Appendix, Fig. S5 E and F). As further confirmation, we observed a moderate increase in ubiquitination of MAPKAPK2 in low glucose (Fig. 3G and SI Appendix, Fig. S5 I and J). However, we did not observe a similar effect in DLD-1 cells treated with MG132, MLN4924, or epoxomicin in low glucose conditions (SI Appendix, Fig. S5 G and H), suggesting that cancer cells may evolve to reduce MAPKAPK2 under stress conditions through multiple mechanisms.

A prior study reported that MAPKAPK2 is activated and subsequently degraded due to its interaction with p38 in response to high intensity ultra-violet light (UV-C) treatment (28). Furthermore, MDM2, an E3 ubiquitin ligase, targeted MAPKAPK2 for degradation in response to UV-C treatment, while depleting MDM2 via shRNA protected MAPKAPK2 from poly-ubiquitination and subsequent degradation by the 26S proteasome (28). We observed a similar consequence in our study where low glucose was associated with MAPKAPK2 downregulation and an increase in phosphorylation of HSP27, a downstream substrate of MAPKAPK2, while inhibition of p38 by SB 203580 reversed HSP27 phosphorylation and protected MAPKAPK2 from degradation (Fig. 3 H–J). Consequently, inhibiting p38 resulted in reduction in ubiquitination of MAPKAPK2 (SI Appendix, Fig. S5 I and J). However, we did not observe any reversal of MAPKAPK2 protein decline in 0.5 mM glucose when MDM2 was inhibited (SI Appendix, Fig. S5 K–N). Therefore, we suspect that a yet to be identified E3 ligase is responsible, in part, for MAPKAPK2 degradation in glucose restriction stress conditions.

We next assessed if activation of ERK1/2 regulates expression of MAPKAPK2 in low glucose conditions as we observed that ERK1/2 activation occurred prior to MAPKAPK2 decline following glucose deprivation (Fig. 3K). We found that inhibition of ERK1/2 activation by trametinib maintained MAPKAPK2 expression at its basal level when HCT116 and DLD-1 cells were subjected to low glucose (Fig. 3 L and M). Similar results were obtained when another structurally unrelated ERK1/2 inhibitor, temuterkib, was used in low glucose conditions (SI Appendix, Fig. S5O). However, ERK1/2 mediated regulation of MAPKAPK2 expression may not occur via direct interaction (SI Appendix, Fig. S5 P and Q). Rather, it was observed that inhibiting ERK1/2 in low glucose reduced ubiquitination of MAPKAPK2, indicating it promotes MAPKAPK2 ubiquitination in low glucose through a currently unknown mechanism (SI Appendix, Fig. S5 I and J). Furthermore, we also observed an increase in ERK1/2 activation in MAPKAPK2-depleted cells (Fig. 3 N and O), suggesting that upon glucose deprivation, ERK1/2 activation leads to MAPKAPK2 degradation and its absence in turn promote activation of ERK1/2 leading to a positive feedback loop maintaining activated ERK1/2 and low MAPKAPK2 protein levels.

MAPKAPK2 Loss Promotes Reduced Cell Proliferation but an Increase in Stress-Mediated Cell Survival.

As MAPKAPK2 loss in CRC cells was associated with enhanced cell survival following nutrient stress as well as reduced sensitivity for trametinib, and MAPKAPK2 was observed to decline in low glucose conditions, we next set out to determine how MAPKAPK2 affects CRC at the molecular level. HCT116 cells expressing control or MAPKAPK2 shRNA were subjected to RNA-sequencing. Differential gene expression analysis indicated that 2,051 genes were significantly upregulated in MAPKAPK2-depleted cells, while 1,268 were downregulated (Fig. 4A). Pathway analysis of downregulated genes using the KEGG dataset indicated that these genes were associated with DNA replication, mismatch repair, cell cycle, and hippo signaling pathways (SI Appendix, Fig. S6A). Similarly, the Hallmark dataset revealed downregulation of gene signatures such as E2F transcription factor targets, G2M cell cycle checkpoints, cholesterol homeostasis, and mTORC1 signaling (SI Appendix, Fig. S6B). Consistently, depleting MAPKAPK2 expression in HCT116 and DLD-1 cells or inhibiting MAPKAPK2 activity by PF-3644022 in HCT116 and HT-29 cells reduced cell growth (SI Appendix, Fig. S6 C–H). Genes impacted by MAPKAPK2-depletion were associated with DNA replication process, such as components of DNA polymerase α-primase complex (PRIM1; Pri1), DNA polymerase δ-complex (POLD1; δ1), MCM complex (MCM2, MCM4, MCM5, MCM6, and MCM7), replication protein A (RPA2), clamp (PCNA), clamp loader (RFC2/4), and helicases (DNA2 and FEN1). Similarly, cell cycle-related genes including cell division cycle proteins (CDC6, CDC25A, CDC45), cyclin-dependent kinase (CDK6), cyclin-dependent kinase inhibitor (CDKN2D), transcription factors (E2F1 and E2F2), spindle assembly checkpoint component (MAD1L1), DNA metabolic enzymes (TK1, RRM2, TYMS), and a component of the origin recognition complex (ORC1) were also downregulated (SI Appendix, Fig. S6 I and J). Interestingly, we observed that mRNA levels of two glucose transporters, SLC2A1 and SLC2A3, were consistently downregulated in MAPKAPK2-depleted HCT116 cells in the RNA-seq dataset (SI Appendix, Fig. S6 K and L) and loss of MAPKAPK2 resulted in downregulation of SLC2A1 and SLC2A3 mRNA as well as corresponding proteins, GLUT1 and to a greater extent GLUT3, in both cell lines (SI Appendix, Fig. S6 M–P). While glucose transporters, such as GLUT1 and GLUT3, are upregulated in response to low glucose conditions (29, 30), we observed that MAPKAPK2-depleted cells were unable to efficiently upregulate the expression of GLUT1 and GLUT3 when cells were grown in 0.5 mM glucose for 24 h (SI Appendix, Fig. S6 O and P). These data indicate that MAPKAPK2 is one of the prominent regulators of glucose metabolism to promote cell growth.

Fig. 4.

MAPKAPK2 depletion leads to upregulation of stress-induced survival pathways. (A) Volcano plot indicating differentially expressed genes (Log₂FoldChange > 0, −log₁₀ P-value ≥ 1.3) in HCT116 cells depleted of MAPKAPK2 compared to shControl. (B) Graph representing enrichment of Hallmark pathways associated with upregulated genes in HCT116 cells depleted of MAPKAPK2 compared to shControl. (C) Graph representing enrichment of oncogenic MS pathways associated with upregulated genes in HCT116 cells depleted of MAPKAPK2 compared to shControl. (D) Circle plot indicating common upregulated genes between 24 h glucose-deprived or MAPKAPK2-depleted HCT116 cells. (E) Hallmark pathways commonly upregulated in 24 h glucose-deprived or MAPKAPK2-depleted HCT116 cells. (F) Oncogenic MS pathways commonly upregulated in 24 h glucose-deprived or MAPKAPK2-depleted HCT116 cells. (G). Heatmap indicating expression changes in genes commonly upregulated or downregulated in 24 h glucose-deprived or MAPKAPK2-depleted or a combination of both conditions in HCT116 cells.

Next, we explored gene signatures that are negatively regulated by MAPKAPK2. Pathway analysis using the Hallmark dataset of upregulated genes in MAPKAPK2-depleted HCT116 cells indicated an increase in p53-pathway, heme metabolism, estrogen response early, hypoxia, glycolysis, apical junction, UV response DN, EMT, TNF-α signaling via NF-κB, and IL-2 STAT5 related gene signatures (Fig. 4B). Similarly, upregulation of genes associated with overexpression of EGFR, MEK, LTE2, ERBB2, and RAF were observed (Fig. 4C). Interestingly, although MAPKAPK2 depletion/inhibition had a negative effect on cell growth (SI Appendix, Fig. S6 C–H), these cells also displayed upregulation of gene signatures usually associated with cell survival, in particular when cells are subjected to stress conditions, similar to glucose deprived cells. In fact, at least 146 genes (−log₁₀ P-value ≥ 1.3 and fold change > 1.2) that were common in glucose deprived and MAPKAPK2-depleted HCT116 cells datasets were associated with TNF-α signaling by NF-κB, hypoxia, p53 pathway, IL-2 STAT5 signaling, estrogen response early, as well as ERBB2, EGFR, MEK, and KRAS associated pathways (Fig. 4 D-F). Notably, while cell cycle, DNA replication, DNA repair, and genome stability related genes, such as TUBA1C, E2F1, E2F2, FEN1, MCM4, HELLS, BRIP, HASPIN, and MASTL are downregulated upon glucose deprivation or MAPKAPK2 depletion, several survival promoting or cell death inhibiting genes, such as BEX2, ZMIZ2, ITGB4, LGALS3, and NOXA1 were upregulated (Fig. 4G). Having demonstrated that ERK activation and MAPKAPK2 expression are inversely correlated, but are mutually inclusive events, similar activation of transcriptional pathways supports further the existence of a feedback loop between ERK1/2 activation and MAPKAPK2 downregulation to sustain survival pathway induction in cells.

MAPKAPK2 Expression Correlates Cell Survival Gene Signatures in CRC Patients and Trametinib Sensitivity In Vitro and In Vivo.

To expand our findings to CRC patients, we extracted MAPKAPK2 protein expression via Clinical Proteomic Tumor Analysis Consortium (CPTAC) (n = 84). Patients were divided into high MAPKAPK2 (MK2^HI, n = 40), and low MAPKAPK2 (MK2^LO, n = 40) groups (Fig. 5A). The mRNA level of MAPKAPK2 was unaltered between these two patient cohorts (Fig. 5B), indicating that MAPKAPK2 is regulated posttranscriptionally in CRC patients and is consistent with our findings that cells under stress downregulate MAPKAPK2 via posttranslational mechanisms. Additionally, MK2^LO patients had higher activation of ERK1/2 compared to the MK2^HI cohort (Fig. 5C), which further corroborated our in vitro findings. In addition, MK2^LO patients showed upregulated Hallmark gene signatures such as TNF-α signaling via NF-κB, KRAS signaling, estrogen response early, and hypoxia. Additional gene signatures were also observed, including xenobiotic metabolism and fatty acid metabolism (Fig. 5 D and E). Similarly, an increase in MEK, ERBB2, and RAF-related gene signatures were observed via oncogenic signaling dataset analysis (Fig. 5F). Stress-induced survival is often associated with poor survival in patients (31). A comparison of overall survival in MK2^LO CRC patients compared to the MK2^HI cohort, indicated a trend toward poor probability of progression free and overall survival in MK2^LO patients (P-value = 0.0710 and 0.0759, respectively) (SI Appendix, Fig. S7 A and B). Collectively, low expression of MAPKAPK2 in CRC patients is associated with upregulation of stress induced survival pathway expression patterns and is a clinical marker of poor overall survival.

Fig. 5.

MAPKAPK2 loss is associated with survival pathways in CRC patients and low trametinib response in vivo. (A) Violin plot indicating MAPKAPK2 protein expression in MAPKAPK2-high (MK2^HI) and -low (MK2^LO) cohorts of colorectal carcinoma patients. (B) Violin plot indicating MAPKAPK2 mRNA expression in MK2^HI and MK2^LO cohorts of colorectal carcinoma patients. (C) Graph indicating ERK1/2 activation (phosphorylation) in MK2^HI and MK2^LO cohorts of colorectal carcinoma patients. D. Volcano plot indicating differentially expressed genes (Log₂FoldChange > 0, −log₁₀P-value ≥ 1.3) in MK2^LO vs. MK2^HI cohorts. (E) Enrichment of Hallmark pathways associated with upregulated genes in MK2^LO vs. MK2^HI cohorts. (F) Enrichment of oncogenic MS pathways associated with upregulated genes in MK2^LO vs. MK2^HI cohorts. (G) Correlation between MAPKAPK2 protein (UniProtKB accession p49137) expression and trametinib response [area under curve (AUC)] in colorectal carcinoma cell lines. (H) Comparison of trametinib dose response in HCT116, DLD-1, HT-29, and Caco-2 cell lines and immunoblot indicating MAPKAPK2 expression in HCT116, DLD-1, HT-29, and Caco-2 cell lines. (I) Trametinib response between HCT116 (shControl) and DLD-1 (shControl) subcutaneous xenografts. (J) Graph comparing trametinib response between HCT116 control and MAPKAPK2-depleted cells in subcutaneous xenografts. Student’s t test was performed in A–C. Pearson’s correlation coefficient and Spearman rank correlation was performed in G. The two-way ANOVA multiple comparison test was performed in H–J. *P < 0.05, ***P < 0.001. Tram, Trametinib.

Importantly, our in vitro findings indicated that MAPKAPK2 expression is critical for the therapeutic response of trametinib. Using DepMap portal, we found that MAPKAPK2 expression also correlated with trametinib response in various CRC cell lines with an inverse correlation between trametinib AUC (area under curve) and MAPKAPK2 expression with a Pearson’s coefficient of −0.597, and a Spearman’s coefficient of −0.561 (P = 0.0187) (Fig. 5G). We further assessed the expression of MAPKAPK2 in four different CRC cell lines (HCT116, DLD-1, HT-29, and Caco-2) and their sensitivity for trametinib. Of the four CRC cell lines we tested, DLD-1 had the lowest basal MAPKAPK2 expression (Fig. 5H). Assessing cell viability in response to a dose-dependent treatment with trametinib, we found DLD-1 cells also showed the lowest sensitivity to trametinib compared to the other three cell lines with relatively higher basal MAPKAPK2 protein expression (Fig. 5H). The effect of MAPKAPK2 expression on trametinib sensitivity was also validated using an in vivo mouse model. Consistent with our in vitro findings, subcutaneous HCT116 tumors were more responsive to trametinib as compared to DLD-1 tumors in mice (Fig. 5I and SI Appendix, Fig. S7 C and D). Additionally, HCT116 subcutaneous tumors with MAPKAPK2 depleted (shMK2 #1 and shMK2 #2) exhibited reduced sensitivity to trametinib as compared to tumors expressing control shRNA (Fig. 5J and SI Appendix, Fig. S7 C and E). We further validated our findings in clinically relevant settings by employing use of two CRC organoids. While trametinib treatment resulted in reduction in organoid growth as compared to vehicle, MAPKAPK2 inhibition mitigated this response (Fig. 6 A–D). Similarly, while trametinib significantly inhibited growth of subcutaneous implanted patient-derived-xenograft (PDX) tumors, inhibition of MAPKAPK2 suppressed the trametinib response (Fig. 6 E–G). Overall, these data indicate that altered MAPKAPK2 levels due to inherent properties of the tumor microenvironment such as glucose restriction might play a crucial role in dictating trametinib response in CRC.

Fig. 6.

MAPKAPK2 inhibition attenuates trametinib response in patient-derived CRC organoids and xenograft mouse models. (A–D) Images (A and C) representing modulation of trametinib response by MAPKAPK2 inhibition in CRC organoid 1 and 2 and respective quantifications (B and D) of growth of organoids (measured as diameter) in 5 µM PF-3644022, 20 nM trametinib, or in combination. (E–G) Images representing modulation of trametinib response by MAPKAPK2 inhibition in patient-derived xenografts tumors (E), tumor volumes (F) and tumor weights (G). The two-way ANOVA multiple comparison test was performed in B, D, and F–G. **P < 0.01, ****P < 0.0001.

Discussion

The MAPK signaling pathway regulates several aspects of tumor growth and progression, including survival in unfavorable conditions and therapeutic resistance. MEK/ERK and p38/MAPKAPK2 are two major pathways dysregulated in CRC and are currently under investigation as targets for clinical intervention (23, 32). These pathways often integrate to reinforce or oppose cellular fate decisions (cell survival or cell death). However, their contextual relationship has remained unclear (33). Given that these pathways may modulate CRC therapeutic responses upon targeting, it is important to fully characterize the context-dependent nature of these pathway interactions in tumor growth, under nutrient stress, and in response to therapeutic treatment. In the present study, we found that glucose deprivation is a solid tumor-relevant context that impacts cellular response to survival mediated by loss or inhibition of MAPKAPK2. MEK/ERK and p38/MAPKAPK2 promote glycolytic metabolism in cancer by regulating glucose transporters and glycolytic enzymes (4, 21, 22). However, how these two signaling pathways integrate during low glucose stress in CRC has not been explored. Our results are in line with earlier observations indicating that these pathways play an important role in regulating cell growth and survival under glucose starvation (6, 21). Furthermore, our results define a unique interaction between the MEK/ERK and p38/MAPKAPK2 pathways by which low glucose conditions leading to activation of MEK/ERK downregulates MAPKAPK2 expression. Inhibition of MEK/ERK signaling was ineffective in low glucose to reverse cell survival, whereas MAPKAPK2 inhibition enhanced cell survival. Mechanistically, we identified that ERK mediated regulation of MAPKAPK2 is, in part, via promoting ubiquitin-proteasome-mediated degradation and is an adaptation response to support cell survival over proliferation in stress condition. However, we found that while DLD-1 cells downregulated MAPKAPK2, it was unlikely through the ubiquitin–proteasome pathway. This could be because DLD-1 cells already show relatively low levels of MAPKAPK2, or through tumor evolution DLD-1 developed a different mechanism for stress-induced MAPKAPK2 downregulation. The latter would further support the importance of cancer cells downregulating MAPKAPK2 to promote tumor cell survival under stress conditions, as different cancers may evolve different mechanisms to achieve MAPKAPK2 suppression. In essence, MAPKAPK2 could be a regulatory point which is modulated to control cell proliferation when conditions for cancer cells are not favorable for optimum growth and rather maintain cell survival. Together, these findings further support a context-dependent role of MEK/ERK signaling, such that in optimal growth conditions it promotes rapid proliferation, but under stress conditions, such as low glucose, the pathway is dedicated to support cell survival. However, ERK does not interact directly with MAPKAPK2 and therefore likely impacts ubiquitination of MAPKAPK2 through an unknown mechanism. ERK may either promote MAPKAPK2 phosphorylation indirectly through p38 as a downstream consequence of ERK1/2 activation leading to MAPKAPK2 recognition for degradation, or ERK may regulate an E3 ubiquitin ligase to promote MAPKAPK2 ubiquitination. However, these mechanisms need further investigation.

Given that glucose limitation could be an important context for diverse and opposing phenotypic regulation by MAPKAPK2, its inhibition under glucose-rich conditions impeded CRC cell growth. Previous studies have shown that its pharmacological inhibition, or genetic deletion in the Apc^min/+ mouse model, reduced tumor growth (23, 25). Our transcriptomic profiling analysis suggests that MAPKAPK2 controls expression of several DNA replication, cell cycle, and glucose metabolism genes, that could be the reason for reduced cell/tumor growth observed in prior studies as well as our own results. On the contrary, these gene signatures are negatively impacted by glucose deprivation and manifest as reduced cell growth, as also observed earlier in lung adenocarcinoma (34). Accordingly, in glucose abundant conditions, quintessential for accelerated proliferation, MAPKAPK2 loss is a limiting factor affecting cellular machinery responsible for rapid cell growth. Whereas loss of MAPKAPK2 expression provided survival benefits to CRC cells when exposed to nutrient stress such as low glucose or 2-DG. A similar observation was reported where MAPKAPK2 inhibition rescued U2OS cells from lethal effects of genetically induced sustained activation of the p38 signaling pathway in high glucose medium (21). These observations suggest an alternative role for the MAPKAPK2 pathway where it regulates various stress response and cell death pathways. CRC cells possibly downregulate the expression of MAPKAPK2 when glucose is restricted via activation of MEK/ERK, to slow down cell growth and its glucose prerequisite, to maintain cell survival. However, our observation contrasts with a prior study showing that prolonged stress to U2OS cells by UV-C irradiation leads to MAPKAPK2 degradation to induce cell death (28). These differing results suggest that MAPKAPK2 may have differential effects depending on the form of stress. UV-C induces DNA-damage and therefore may require a different set of pathways to sense and respond to, in which p38/MAPKAPK2 may promote cell death. This may contrast with nutrient stress, such as glucose deprivation, in which p38/MAPKAPK2 may suppress cell cycle progression but support cell survival. Together, these results further emphasize the context-dependent role of the p38/MAPKAPK2 pathway. In our RNA-seq data, MAPKAPK2 depletion vs control cells and glucose starved vs supplemented cells exhibited upregulation of common gene signatures, including p53 pathway, estrogen response early, hypoxia, glycolysis, TNF-α signaling via NF-kB, and IL-2 STAT5 signaling. Upregulation of these gene signatures have been implicated in stress conditions and associated with cancer cell survival in a context-dependent manner (35–40). Additionally, we observed that MAPKAPK2 depletion in CRC cells leads to high ERK1/2 activation. This is particularly interesting as it suggests a positive feedback loop, such that ERK1/2 activation and MAPKAPK2 loss coexist and regulates each other leading to sustained signaling and consequently desensitizes cancer cells to trametinib. It is a possibility that ERK1/2 itself promotes survival pathways when activated under stress, such as glucose deprivation, and simultaneously degrades MAPKAPK2 which helps to sustain its activated form as well as downregulate glucose metabolism and proliferation. The prospect of targeting these pathways in combination should be explored further to enhance trametinib efficacy in patients.

CRC patients often harbor EGFR hyperactivation or KRAS and BRAF mutations, both of which feed through MEK1/2 and ERK1/2, which makes MEK1/2 inhibitors a preferred choice to treat these cancers. Various clinical trials investigating the potential of MEK1/2 inhibition by trametinib in combination with experimental and approved therapeutic regimens are being conducted. However, a heterogenous response toward MEK1/2 inhibitor-based therapies is a major challenge to their clinical use (18, 19). We observed that one such factor could be the expression of MAPKAPK2, with its loss leading to a decrease in trametinib efficacy. A negative correlation of MAPKAPK2 protein expression and trametinib sensitivity in CRC cell lines was observed, suggesting that apart from glucose induced changes in MAPKAPK2 expression, other intrinsic and extrinsic mechanisms that downregulate its expression may impact trametinib efficacy in CRC. Clinically, these results could be relevant in CRC tumors that express low levels of MAPKAPK2 and might therefore be refractory to trametinib, or other MEK1/2 inhibitors. As loss of MAPKAPK2 is vital for stress-induced survival of CRC cells, low expression of these similar pathways (Figs. 4 and 5) in tumors might indicate the existence of stress conditions and adaptation responses. Segregating the heterogeneity in CRC tumors based on MAPKAPK2 protein expression might be useful when considering trametinib treatment in CRC patients. Moreover, CRC tumors are prone to heterogeneity in glucose distribution and MAPKAPK2 being a glucose-sensitive protein might lead to intratumor MAPKAPK2 heterogeneity, ultimately affecting therapy response and recurrence (41). For instance, the possibility of combining trametinib with 2-DG has been explored previously for breast cancer (42). However, the combination was effective for MDA-MB-231 but not in MDA-MB-468 xenografts (42). In support of this notion, utilizing data from the DepMap portal, we observed that MAPKAPK2 expression of MDA-MB-231 is 3.7 times higher than that of MDA-MB-468 (SI Appendix, Fig. S7F). Additionally, MAPKAPK2 was among the top five targets in a CRISPR knockout screen carried out in the lung cancer cell line NCI-H1299 treated with trametinib, further supporting the view that MAPKAPK2 modulates trametinib response (43). Given our findings highlight the importance of MAPKAPK2 stability for the efficient response of trametinib (or ERK inhibitors), maintaining CRC patients on high glucose while trametinib administration might improve the therapy response, as was observed for pancreatic cancer (44).

In conclusion, we show that MAPKAPK2 could be an inherent phenotypic switch in CRC that is turned on to support rapid cell growth when glucose is present but is turned off in the absence of glucose to support cell survival. Additionally, such tumors expressing low or no MAPKAPK2 might be refractory to trametinib regimen in CRC.

Materials and Methods

Please see SI Appendix, Materials and Methods for additional details related to this study.

Cell Lines.

HCT116 (ATCC, CCL-247, RRID: CVCL_0291), DLD-1 (ATCC, CCL-221, RRID: CVCL_0248), HT-29 (ATCC, HTB-38, RRID: CVCL_0320), Caco-2 (ATCC, HTB-37, RRID: CVCL_0025), SW480 (ATCC, CCL_228, RRID: CVCL_0546), SW620 (ATCC, CCL_227, RRID: CVCL_0547), and HEK293T (ATCC, CRL-3216, RRID: CVCL_0063)

Cell Culture Reagents.

Dulbecco’s Modified Eagle Medium, DMEM (Gibco, 11965092), Dulbecco’s Modified Eagle Medium, DMEM (with glucose, without glutamine) (Gibco, 31053028), Dulbecco’s Modified Eagle Medium, DMEM (without glucose, without glutamine) (Gibco, A1443001), L-Glutamine (Gibco, 25030081), Fetal Bovine Serum (HyClone, SH30910.03), and Penicillin/streptomycin (Gibco, 15140122)

Other Reagents.

MTT (3-(4,5-Dimethylthiazol-2-yl)-2,5-Diphenyltetrazolium Bromide) (Invitrogen, M6494), DMSO (Fisher Bioreagents, BP231-100), Hoechst stain (Invitrogen, H3570), crystal violet (Thermo Scientific, B21932.36), polybrene (Santa Cruz Biotechnology, sc-134220), puromycin (Thermo Scientific, AC227420100), 0.45 μM syringe filter (Whatman, 99142504), 100X HALT (Thermo scientific, 1861281), nitrocellulose blotting membrane (Amersham Protran, 10600002), chemiluminescent HRP substrate (Millipore, WBKLS0500), methylcellulose (Thermo Scientific, 258115000), Tween-80 (Fisher Bioreagents, BP338), Trametinib (MedChemExpress, HY-10999), Matrigel (Corning, 354234), DNA oligos (Integrated DNA Technologies), p38-inhibitor, SB 203580 (MedChemExpress, HY-10256), Ni-NTA-agarose beads (Qiagen, 30210) MK2-TRAP agarose beads (Chromotek, mta-20), control agarose beads (Chromotek, bab-20), pLenti-III-Blank vector (abm, LV587) MAPKAPK2 lentiviral vector (abm, 280780610295), and Lipofectamine 2000 (Invitrogen, 11668019). pCMV-8xHis-Ub was a gift from William Kaelin (Addgene plasmid #107392; RRID:Addgene_107392) (45).

Antibodies.

MAPKAPK2 (Invitrogen, PA5-17729, RRID: AB_10979499), alpha-tubulin (Santa Cruz Biotechnology, sc-23948, RRID: AB_628410), p-ERK1/2 (T202/Y204) (Cell Signaling Technology, 4370T, RRID: AB_2315112), t-ERK1/2 (Cell Signaling Technology, 4695, RRID: AB_390779), p38 (Cell Signaling Technology, 9218), GLUT1 (Cell Signaling Technology, 12939, RRID: AB_2687899), GLUT3 (Proteintech, 20403-1-AP, RRID: AB_10694437), HSP27 (Cell Signaling Technology, 2402, RRID: AB_331761), p-HSP27 (Cell Signaling Technology, 2401, RRID: AB_331644), Anti-HA (Sigma, H6908, RRID: AB_260070), and Anti-His (Santa Cruz Biotechnology, sc-57598).

Kits.

MycoAlert PLUS Mycoplasma Detection Kit (Lonza LT07-710), RNeasy Kit (Qiagen, 74104), qScript cDNA Synthesis Kit (Quanta bio, 95047), Powerup SYBR Green Master Mix for qPCR (Applied biosystems, A25742), DC Protein Assay Kit (Bio-Rad, 5000112), and CellTiter-Glo® 2.0 Luminescent Cell Viability Assay (Promega, G9241)

Methods

RNA-Sequencing.

To study the effect of MAPKAPK2 depletion, HCT116 cells expressing shRNA against MAPKAPK2 (shMK2 #1 and shMK2 #2), or control shRNA (shControl) were subjected to RNA-sequencing. To study the effect of glucose deprivation, HCT116 cells expressing control shRNA (shControl) were cultured in 20 mM or 0 mM glucose-containing media for 24 h. Total RNA was extracted using RNeasy Kit (Qiagen) as per the manufacturer’s instructions. Minimum two biological replicates were used for all treatments. After quality check, total RNA was sent to Novogene for RNA-sequencing. Briefly, messenger RNA was purified and fragmented followed by synthesis of first strand cDNA using random hexamer primers and second strand cDNA synthesis using either dUTP for directional library or dTTP for nondirectional library. Quantified library was then sequenced using Illumina NovaSeq X plus platform to generate paired end reads.

Bioinformatics Analysis of RNA-Sequencing.

Clean reads were mapped to reference genome (GRCh38) using Hisat2 v2.0.5. The number of mapped reads for each gene was then counted using featureCounts v1.5.0-p3 and FPKM of each gene was calculated. For differential expression analysis DESeq2 R package (1.20.0) was used. For enrichment analysis of differentially expressed genes between two groups, enrichr (RRID: SCR_001575) and Shiny go 0.82 (RRID: SCR_019213) platforms were used.

Ubiquitin Assay.

Ubiquitin assay was performed as described previously (46). Cells were cotransfected with His-Ub expression vector together with either empty or MAPKAPK2-HA expression vector using Lipofectamine 2000. After 36 h post transfection, cells were treated with glucose 20 or 0.5 mM and inhibitors (as indicated) for 12 to 16 h. Cells were collected in 15 mL tubes by scrapping, washed with ice-cold PBS, 10% cells were separated as inputs and remaining cells were lysed in denaturing Buffer A (6 M Guanidine-HCL, 0.1 M Na₂HPO₄/NaH₂PO₄, 10 mM Imidazole, pH 8.0) with sonication (8 pulses of 1 s each). Supernatants were incubated with 30 μL of equilibrated (50%) Ni-NTA-agarose for 3 h at room temperature. Beads were collected and washed sequentially with Buffer A/TI (1:3 Buffer A:Buffer TI) and Buffer TI (25 mM Tris-Cl, 20 mM Imidazole, pH-6.8). Finally, the collected beads were denatured in 2X Laemmli containing 200 mM Imidazole at 95 °C for 15 min and subjected to immunoblotting.

CRC Organoids Growth.

Patient-derived cancer organoids were cultured as previously described (47, 48). Cancer samples were collected following approval by the Ethics Committees of Daping Hospital [Ratification NO. 2022(223)], in accordance with the Declaration of Helsinki [2021(266-01)], and deidentified prior to use in the current study. CRC tissues were washed with PBS, minced, and subsequently digested by collagenase IV (300 U/mL, Sigma) and dispase (125 µg/mL, Sigma) for 1.5 h at 37 °C. The suspension was then filtered through 70 µm cell strainers and centrifuged at 300 g for 10 min. The pellet was washed by advanced DMEM/F12 medium (Gibco) and mixed with Matrigel (Corning). Approximately 3,000 cells were seeded in a 24-well plate. After the Matrigel was fixed, the human colorectal tumor organoids were cultured in medium containing Advanced DMEM/F12, Penicillin/Streptomycin (Gibco), Glutamax (Gibco), HEPES (10 mM, Gibco), B27 supplement (Gibco), N2 supplement (Gibco), N-acetylcysteine (1 mM, Sigma), Recombinant human EGF (50 ng/ml, PeproTech), A-83-01 (500 nM, Sigma), and SB202190 (10 µM, Sigma) and incubated at 37 °C with 5% CO₂.

In Vivo Xenografts.

Animal experiments were approved by Creighton University Institutional Animal Care and Use Committee (IACUC) Policies and Procedures. Female nude mice (002019, RRID: IMSR_JAX:002019) were purchased from The Jackson Laboratory and housed with 12 h light/12 h dark cycles. At 7 wk of age, mice were randomly divided into four groups (n = 8) and subcutaneously injected with 10⁶ HCT116 cells (1:1 ratio, cells in DMEM:Matrigel) expressing either control shRNA (HCT116 shControl) or shRNA against MAPKAPK2 (HCT116 shMK2 #1 or HCT116 shMK2 #2) or 10⁶ DLD-1 cells expressing control shRNA (DLD-1 shControl). When tumors reached 100 to 200 mm³, each group was further subdivided into two groups (n = 4) and administered either vehicle (0.5% methylcellulose and 0.25% Tween-80) or trametinib (0.1 mg/Kg/day) via oral gavage every day until day 18 after which the study was terminated. Tumor volumes were measured using calipers every 3 d and calculated [volume = (length*width²)/2] starting from the first day of drug treatment.

Patient-Derived Xenografts.

To establish patient-derived CRC tumor xenograft (PDX) models, NOD/SCID mice (6 wk old, female) were purchased from Gempharmatech China. PDX was established as previously described (47, 48), with mouse studies approved by the Institutional Animal Care and Use Committee of Jinfeng laboratory (IACUC-JFLAB-2022-003). When xenografted tumors reached a sufficient size, they were cut into approximately 2 mm³, and reimplanted into the right flank of NOD/SCID mice. Tumor volumes were calculated using the formula [tumor volume = 0.5 × length × width²] based on caliper measurements. Once tumor volume reaches 100 mm³, mice were randomly assigned into two groups and given peritumoral injection of either vehicle (n = 12) or MAPKAPK2i (PF-3644022) (n = 12). Treatment of MAPKAPK2i was given every other day at 50 µM. Three days post first MAPKAPK2i injection, each group was subdivided into two groups to receive vehicle (n = 6) or trametinib (0.2 mg/kg/day) (n = 6) daily via oral gavage until day 28 upon which the study was terminated.

Correlation Analysis of MAPKAPK2 Expression and Trametinib Response.

For correlation analysis, MAPKAPK2 protein expression and trametinib sensitivity (PRISM repurposing secondary screen) data for various CRC cell lines was downloaded from Cancer Dependency Map Portal (RRID: SCR_017655).

CRC Patient Data Analysis.

cBioPortal (RRID: SCR_014555) was used to access proteomic data for MAPKAPK2 expression in tumors from TCGA CRC patients. Based on MAPKAPK2 expression, patients were divided into either high MAPKAPK2 (n = 40) or low MAPKAPK2 (n = 40) groups. Both groups were then assessed for ERK1/2 phosphorylation, progression free survival, overall survival, and differential gene expression based on transcriptomic data available from TCGA. For enrichment analysis of differentially expressed genes between two cohorts, enrichr (RRID: SCR_001575) and Shiny go 0.82 (RRID: SCR_019213) platforms were used.

Data Representation and Statistical Analysis.

For data representation and statistical analysis GraphPad PRISM 10 (RRID: SCR_002798) was used. Data are represented as mean ± SEM. The statistical tests used are indicated in the figure legends.

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

Sequence data that support the findings of this study have been deposited in the National Center for Biotechnology Information (NCBI) Gene Expression Omnibus (GEO) with the primary accession number (GSE291324) (49).