Resistance to chemotherapy represents a major limitation in the treatment of colorectal cancer. Novel strategies to circumvent resistance are critical to prolonging patient survival. Rac1b, a constitutively activated isoform of the small GTPase Rac1, is upregulated with disease progression and promotes cell proliferation and inhibits apoptosis by activation of NF-κB signaling. Here, we show that Rac1b overexpression correlates with cancer stage and confirmed Rac1b expression is associated with increased growth through enhancing NF-κB activity. Rac1b knockdown reduced cellular proliferation and reduced NF-κB activity. Surprisingly, Rac1b expression and NF-κB activity were upregulated in cells treated with chemotherapeutics, suggesting that Rac1b facilitates chemo-resistance through activation of NF-κB signaling. Knockdown of Rac1b or Rac inhibition increases the sensitivity of the cells to oxaliplatin. When used in combination, inhibition of Rac prevents the increase in NF-κB activity associated with chemotherapy treatment and increases the sensitivity of the cells to oxaliplatin. Although Rac inhibition or oxaliplatin treatment alone reduces the growth of colorectal cancer in vivo, combination therapy results in improved outcomes compared with single agents alone. We provide the first evidence that Rac1b expression confers resistance to chemotherapy in colorectal cancer. Additionally, we show that the use of a Rac inhibitor prevents chemoresistance by blocking activation of chemotherapy induced NF-κB signaling, providing a novel strategy to overcome resistance to chemotherapy in colorectal cancer.
Colorectal cancer is the third most common malignancy worldwide and continues to be a major health concern with over one million men and women being affected annually (1, 2). Fifty percent of patients progress to metastatic disease. The average 5-year survival rate for colorectal cancer is only 50% to 60%, resulting in close to 700,000 deaths worldwide in 2012 alone. Although colonoscopy has been effective in reducing mortality by detecting earlier stage disease, the identification of predictive biomarkers of disease progression remains limited.
Patients with colorectal cancer with late stage (stage III and IV) colon cancer are routinely treated with a chemotherapeutic regimen consisting of the combination of oxaliplatin (OXA) and 5-fluorouracil (5-FU). OXA, a third-generation platinum-based chemotherapy drug (3), has a major impact on the management and outcome of colorectal cancer patients in combination with 5-FU and leucovorin (FOLFOX regimen) or capecitabine (CapeOx regimen) for the adjuvant treatment of stage III (4). Other combined treatments are also in use in stage IV. However, the major obstacle in treating colorectal cancer is overcoming acquired resistance to chemotherapy. Targeted therapies have also been used to treat colorectal cancers. However, only a few of these targeted agents have shown activity in colorectal cancer (5, 6). Therefore, there remains an unmet need for additional treatment options for patients with colorectal cancer that become resistant to first-line therapy options (7). The discovery of novel drivers of colorectal cancer and the development of new compounds targeting these drivers may overcome mechanisms of resistance and improve upon the outcomes currently achieved.
The Rho family of GTPases has been widely studied in numerous types of cancer, including colorectal cancer. Identification of a novel splice variant of the Rho GTPase Rac1: Rac1b has piqued interest in Rac signaling in colorectal cancer. Rac1b is a naturally occurring splice variant, preferentially expressed in colon and breast tumors (8, 9). Rac1b is characterized by the insertion of 19 amino acid residues immediately C-terminal to the “switch II” domain. This insertion greatly reduces the intrinsic GTPase activity of Rac1b and impairs sequestration by RhoGDIs, resulting in Rac1b being preferentially found in a GTP-bound, active form (10). Although Rac1 activation strongly activates conical downstream signaling through p21-activated kinases (PAK) or Jun N-terminal kinase (JNK), Rac1b has been shown to poorly activate PAK and JNK (11).Instead, Rac1b has been shown to facilitate the proliferation of lung and colon cancer cells, via NF-κB pathway, and inducing Cyclin D1 expression (12, 13). The expression of Rac1b in fibroblasts stimulated cell-cycle progression and survival under conditions of serum starvation (14). It has been recently shown that Rac1b mediates an MMP-3-epithelial to mesenchymal transition (EMT) in cultured cells through the induction of reactive oxygen species (ROS; ref. 15). It has also been recently reported that Rac1b promotes canonical Wnt signaling, a pathway often deregulated in colorectal cancer (14).
In this study we show that Rac1b is overexpressed at both the RNA and protein level in patients with colorectal cancer compared with normal colon tissue. Similar to previous findings, we observe that the overexpression of Rac1b in colon cancer cells enhanced the proliferation of cancer cells via NF-κB pathway activation. Additionally, the overexpression of Rac1b in HCT116 cells confers resistance to 5-FU and OXA, whereas the knockdown of Rac1b in HT29 cells sensitizes the cells to 5-FU or OXA. Moreover, we observe a significant upregulation of Rac1b protein upon treatment of either 5-FU or OXA, suggesting that Rac1b upregulation is part of a stress response of colorectal cancer cells to the treatment with these DNA-damaging drugs facilitating resistance through NF-κB signaling. Targeting Rac1b with a small molecule Rac inhibitor reduced the growth of colon cancer cells in vitro and in vivo. Importantly, the addition of a Rac inhibitor in combination with 5-FU or OXA enhanced the antitumoral effects of either agent alone in vitro and in vivo. Our combinatorial approach introduces a strategy to enhance the efficacy of DNA-damaging therapies for the treatment of cancer. Therefore, increased Rac1b levels contribute to the development of chemoresistance to these drugs in colorectal cancer.
Materials and Methods
Cell lines and cell culture
Colon cancer cells HCT116, HT29 cells were cultured in McCoy's 5A and SW48 and SW480 cells were cultured in Leibovitz's L-15 medium. Colo320DM cells were cultured in RPMI media. All cell lines were purchased from ATCC. Cells were tested for Mycoplasma (Lonza) every 10 passages. Media were supplemented with 5% (v/v) FBS (Invitrogen) and cultured at 37°C in a humidified atmosphere of 5% CO2. OXA, 5-FU, and JSH-23 were purchased from Sigma-Aldrich.
For stable knockdown of Rac1b the siRNA sequences, 5-GAAACGUACGGUAAGGAUA-3 and 5-GGCAAAGACAAGCCGAUUG-3 were cloned in a psi-LVRH1GH lentiviral vector. The vectors were cotransfected with plasmid pCMV-dR8.91 and plasmid encoding vesicular stomatitis virus G protein into 293T cells using Lipofectamine 2000. Viral supernatant was collected 48 hours posttransfection, filtered (0.45-μm pore size), and added to cells in the presence of 8 μg/mL polybrene (Sigma-Aldrich). Puromycin (2 μg/mL) was used for selection. For Rac1b overexpression, the Rac1b cDNA was cloned in pLv105 lentiviral vector. The transfection was done as mentioned earlier. Viral supernatant was added to cells and selected for with puromycin.
The Cancer Genome Atlas data mining
The colorectal cancer dataset was accessed and mined through The Cancer Genome Atlas (TCGA) Research Network (Provisional 2017; http://cancergenome.nih.gov/). Tumor samples with corresponding RNAseq (475 patients) were used for analysis. TCGA data were also obtained using the Oncomine platform (ThermoFisher Scientific).
Cell lysates were prepared in RIPA lysis buffer. Blots were probed with Rac1b (1:1,000; Millipore, #09-27 Rac1 (1:1,000; Millipore, clone 23A8), IκBα, phospho IκBα, and Phospho P65 (1:1,000; Cell Signaling Technology) Cyclin D1 (1:5,000; Abcam, EPR2241) antibodies. Fluorescent tag anti-rabbit (Licor) and anti-mouse (Licor) secondary antibodies were used. Signal was detected using Licor. Blots were reprobed with an anti-actin (Abcam) antibody.
NF-κB luciferase assay
Cells were transduced with NF-κB luciferase reporter lentiviral particles (Qiagen) per manufacture's protocol. Constructs were added as indicated in figures. Assays of luciferase activity were done as described previously (16). HCT116 and HT29 cells expressing NF-κB reporter were grown to a subconfluent density and treated either with OXA, GYS32661, or with the combination of both. Cell lysates were prepared and reporter activity was measured using the luciferase reporter assay system (Promega).
Rac pull-down experiments
GTP-bound Rac1 and Rac1b levels were determined using a Rac1 Activation Assay Kit according to the manufacturer's protocol (Pierce Biotechnology). Cells were washed in cold PBS and lysed on ice in 600 μL lysis buffer. Total lysates were cleared by centrifugation. Equal amount of lysate was incubated for 30 minutes at 4°C with PAK binding domain (PBD) agarose beads. Precipitated complexes were washed thrice with lysis buffer. After the final wash, the supernatant was discarded and 40 μL of 2× Laemmli buffer were added to the beads. Total lysates and precipitates were then analyzed by the Western blot analysis. In vitro pull-down experiments were conducted as previously described (17).
A total of 1 × 103 cells were cultured in six-well cell culture plates in triplicate. The cells were grown in McCoy's media supplemented either with 2.5% FBS or 5% FBS. The cells were trypsinized and counted with hemocytometer for 5 consecutive days.
Cells were seeded in the upper compartment of a 24-well Boyden chamber (8-mm pore size; Costar) and allowed to migrate for 16 hours in response to CM in the lower compartment. The cells that migrated to the underside of the filter were stained with crystal violet and counted under bright-field microscopy.
Clonogenic assay (three-dimensional colony formation assay)
A total of 500 cells per well were seeded in six-well plates, and treated with or without the indicated drugs. At 2 to 3 weeks after seeding, the colonies were stained with crystal violet.
The soft-agar colony-forming assay was performed in six-well plates with a base of 2 mL of medium containing 5% FBS with 0.6% agar (Amresco). Cells were seeded in 2 mL of medium containing 5% FBS with 0.35% agar at 1 × 103 (for HCT116 and HT29 cells) cells/well and layered onto the base. Two milliliters of DMEM with 10% FBS was covered on the top of agar gel. The photographs of colonies growing in the plates were taken and scored using ImageJ.
Cells were plated in 96-well plates (5,000 cells per well), and 24 hours later, various concentrations of chemotherapeutic drugs or GYS32661 were added and incubated for 3 days. Cytotoxicity was measured using a standard PrestoBlue dye (Invitrogen).
Adult (8–10 weeks of age) SCID mice were used for xenograft studies. HT29 cells or HCT116 (1 × 106 cells) were injected subcutaneously into the left and right flank of the mice. When the tumors reached approximately 100 mm3, mice were divided into five groups (N = 5) for treatment with saline, GYS32661, OXA, and the appropriate combinations of GYS32661 and OXA. The OXA treatment group was injected intraperitoneally with 10 mg/kg OXA twice per week for 3 weeks. The GYS32661 treatment group was injected intraperitoneally 30 mg/kg for 5 consecutive days for 3 weeks. The combination treatment groups received OXA (twice per week, 10 mg/kg) and GYS32661 (5 days, 30 mg/kg). The control group received saline intraperitoneally. Tumor volume and body weight were measured every 3 days. Tumor volume was calculated using the formula V = (AB2)/2, where A is the largest diameter and B is the smallest diameter. These experiments have been conducted in accordance with and approved by the University of Miami Institutional Animal Care and Use Committee.
IHC staining for Rac1b was performed on colorectal cancer TMA (US Biomax). Slides were dewaxed with xylene and then hydrated in an ethanol series. Antigen retrieval was done by using citrate buffer solution (pH 6.0) for 30 minutes at 95°C. The endogenous peroxidase activity was blocked using hydrogen peroxide. The tissues were blocked with Avidin-biotin blocking system (Vector Laboratories, Burlingame, CA). After blocking, slides were incubated with rabbit anti-Rac1b antibody (Millipore, diluted 1:1,000) for 1 hour at room temperature. After incubation with secondary antibody (LSAB Kit, DakoCytomation), the slides were developed with diaminobenzidine tetrahydrochloride (DAB; Sigma-Aldrich). Slides were counterstained with hematoxylin. The IHC slides were scanned into high-resolution images using the Olympus VS120. All digital images obtained in .svs format were visualized with Olympus software. Each TMA spot was examined by two independent reviewers who assigned a score of 0 (no staining), 1 (<10% of malignant cells staining), 2 (10%–50% of malignant cells staining), or 3 (>50% of malignant cells staining) within carcinomatous areas.
Data are expressed as mean ± SEM. Differences between two groups and multiple groups were analyzed by two-tailed Student t test and one-way ANOVA, respectively. P-value less than 0.05 was considered significant.
Rac1b is overexpressed in colorectal adenocarcinoma
Rac1b has been shown to be upregulated in colon and breast cancers, suggesting an oncogenic role for Rac1b (8, 18). To confirm Rac1b overexpression occurs in clinical samples, Rac1b transcripts were queried from TCGA colorectal cancer (19) patient dataset containing data from 475 patients with colorectal cancer and 41 normal samples. RNAseq analysis from patients identified all Rac1 transcripts including the transcripts for Rac1 and Rac1b. The level of Rac1b transcript was analyzed from the total Rac1 levels (20) and the ratio of Rac1b over all Rac1 transcripts was calculated (20). Rac1b transcripts were significantly higher (P < 0.0001) in patients with colorectal cancer as compared with normal colon (Fig. 1A). To determine if Rac1 played a role in colorectal cancer progression, Rac1b transcript levels were analyzed in the different stages of colorectal cancer. Although we observed a trend that higher Rac1b expression correlated with higher colorectal cancer staging, only stage IV (metastatic) colorectal cancer was significantly elevated (Fig. 1B). Although Rac1 has been observed to be overexpressed in many cancer types (21), transcript levels of Rac1 were not significantly elevated compared with normal colon (Supplementary Fig. S1), indicating that Rac1b, and not Rac1, drives colorectal cancer progression. To confirm that Rac1b was elevated at the protein level in colorectal cancer patient samples, IHC for Rac1b was performed on tissue microarrays containing samples from stage IV metastatic colorectal cancer and normal tissue (Fig. 1C). Rac1b staining was observed in normal colon epithelial cells. However, Rac1b staining was significantly higher in colorectal cancer metastasis confirming that Rac1b is elevated upon progression to carcinoma and metastatic disease. IHC analysis of Rac1b also confirmed a correlation between Rac1b and colorectal cancer stage (Fig. 1D; Supplementary Table S1). Multiple colorectal cancer cell lines obtained from metastatic sites also express high levels of Rac1b protein (Fig. 1E). Therefore, an oncogenic splice variant of Rac1, Rac1b, is overexpressed upon progression to colon adenocarcinoma and metastatic disease.
Rac1b controls cell proliferation and cell viability through NF-κB
Rac1 is canonically involved in cytoskeletal rearrangements, cell-cycle progression, cell survival, migration, and invasion (22, 23). Rac1b however, has been shown to preferentially activate AKT signaling and NF-κB resulting in increased Cyclin D1 and progression through the cell cycle (24). To further investigate the role of Rac1b in colorectal cancer, Rac1b was stably overexpressed in HCT116 colon adenocarcinoma cells. Rac1b overexpression resulted in increased Cyclin D1 expression (Fig. 2A). To confirm the activity of exogenous expression of Rac1b in the HCT116 cells, cellular proliferation assays were carried out on HC116-EV and HCT116-Rac1b overexpressing cells in growth media containing 2.5% and 5% concentrations of serum (Fig. 2B and C). Rac1b overexpression resulted in a significant increase in cellular proliferation in media containing both 2.5% and 5% serum conditions (Fig. 2B and C). To test the effects of Rac1b overexpression on growth in three-dimensions, HCT116-EV and HCT116-Rac1b cells were seeded in soft agar (Fig. 2D). Rac1b overexpression resulted in a greater than two-fold increase in colony formation in soft agar. Together, these data support the role of Rac1b in enhancing the proliferative potential of colorectal cancer. To confirm that increased proliferation in Rac1b, overexpressing cells was attributed to Rac1b overexpression, HCT116-EV and HCT116-Rac1b cells were engineered to stably express NF-κB luciferase reporter. The overexpression of Rac1b in HCT116 cells resulted in a greater than three-fold increase in NF-κB reporter activity compared with the empty vector control (Fig. 2E). To confirm that Rac1b overexpression was elevating NF-κB activity, the addition of JSH-23, a NF-κB translocation inhibitor (25, 26), was used to block the activation of NF-κB activity caused by Rac1b overexpression (Fig. 2E). As expected, JSH-23 attenuated NF-κB signaling at previously reported doses (Supplementary Fig. S2; ref. 25). Therefore, Rac1b overexpression is sufficient to enhance NF-κB activity in colorectal cancer cells resulting in enhanced cellular proliferation.
To study the dependency of Rac1b in colorectal cells, Rac1b was stably knocked down using a Rac1b-specific shRNA. Rac1b shRNA expressing HT29 cells showed ∼80% reduction in Rac1b levels compared with a nonsilencing shRNA control as determined by the Western blot analysis (Fig. 2F). Rac1 levels were not altered, validating the specificity of the shRNA for Rac1b (Fig. 2F). Knockdown of Rac1b in HT29 cells reduced CyclinD1 levels further supporting the role of Rac1 as a determinate of colorectal cancer cell proliferation (Fig. 2A and F). To test if Rac1b knockdown effects cell growth, proliferation assays were conducted on shRac1b and shNSC cells. Rac1b knockdown resulted in a significant reduction in proliferation of HT29 cells grown in both 2.5% serum (Fig. 2G) and 5% serum concentrations (Fig. 2H). Knockdown of Rac1b resulted in significant suppression of anchorage-independent growth in soft agar (Fig. 2I). To confirm that the molecular knockdown of Rac1b resulted in reduced NF-κB activity, shRac1b and shNSC HT29 cells were engineered to express an NF-κB luciferase reporter. Similar to previous studies, the knockdown of Rac1b resulted in suppression of NF-κB activity (Fig. 2J; ref. 16). Together, these data showed that Rac1b enhances colorectal cancer growth and survival by activating NF-κB signaling resulting in enhanced CyclinD1 and cellular proliferation.
Rac1b expression confers resistance to chemotherapy
5-FU in combination with OXA are components of the standard of care regimen for colorectal cancer (27, 28). Previous studies have shown that the NF-κB transcription factor becomes activated upon resistance to OXA (29). Given that Rac1b expression activates NF-κB activity, we queried whether Rac1b modulation in colorectal cells changes the sensitivity to chemotherapy. To do so, cellular viability assays were conducted on HCT116 EV and Rac1b overexpressing cells exposed to varying concentrations of OXA (Fig. 3A). Rac1b overexpression in HCT116 shifted the sensitivity of cells to OXA compared with control cells with IC50 values of 1.1 μmol/L and 570 nmol/L, respectively (Fig. 3A). Similar results were observed when the cells were treated with 5-FU (Supplementary Fig. S3). To further observe the effects of Rac1b overexpression on cells treated with clinically relevant concentrations of OXA, a three-dimensional (2D) colony formation assay was conducted on HCT116-EV and Rac1b cells treated with 500 nmol/L and 1 μmol/L OXA (Fig. 3B). To confirm that the growth and survival benefit was attributed to the enhanced NF-κB activity, cells were treated with 500 nmol/L OXA along with the NF-κB inhibitor JSH-23 (Fig. 3C). As expected, cells that overexpressed Rac1b were resistant to the OXA. However, the addition of the NF-κB inhibitor resensitized the cells to the OXA. These findings were recapitulated in a three-dimensional soft agar colony formation assay, where Rac1b overexpression resulted in anchorage independent growth in the presence of OXA whereas growth in control cells was diminished (Fig. 3D). As observed in the 2D colony formation assay, the treatment of cells with JSH-23 rescued the enhanced growth and survival phenotype observed by Rac1b overexpression (Fig. 3E). Therefore, cells that overexpress Rac1b are capable of sustained growth in the continual presence of clinically relevant levels of OXA due to enhanced activation of NF-κB signaling.
To determine if the knockdown of Rac1b could sensitize colorectal cells to OXA, shNSC and shRac1b HT29 cells were treated with varying concentrations of OXA and a cellular viability assay was performed. Knockdown of Rac1b resulted in a 2.5-fold increase in sensitivity to OXA shifting the IC50 from 3.4 μmol/L in the control cells down to 1.2 μmol/L in the shRac1b cells (Fig. 3F). Similar data were observed with 5-FU (Supplementary Fig. S4). The Rac1b knockdown cells grown under anchorage-independent conditions showed significantly increased sensitivity to OXA compared with control cells (Fig. 3G). These results suggest that expression of Rac1b in colon cancer confers resistance to OXA, whereas suppression of Rac1b by molecular knockdown can shift the sensitivity of colorectal cancer cells to OXA. Together, these data strongly support that Rac1b expression confers resistance to OXA.
Although many patients with colorectal cancer respond initially to treatment, a major reason for the failure of advanced colorectal cancer treatment is the occurrence of chemoresistance to OXA-based chemotherapy (7). It has become clear that a well-established mechanism of chemoresistance to OXA-based chemotherapy is the hyper-activation of NF-κB pathway (7, 30). While treating HCT116 colorectal cancer cells with OXA, we observed a dramatic increase of Rac1b protein expression in HCT116 EV control cells (Fig. 3H). Moreover, a pronounced increase in Rac1b protein was also observed in the HCT116-Rac1b cells that were treated with OXA for the same time period (Fig. 3H). Therefore, OXA treatment resulted in an increase in both endogenous and exogenous levels of Rac1b. Interestingly, Rac1 levels did not seem to fluctuate upon OXA treatment, suggesting that Rac1b is selectively upregulated in response to OXA treatment. Because we have previously shown that increased Rac1b expression enhances NF-κB signaling, we performed NF-κB luciferase assays on HCT116-control and HCT116-Rac1b cells treated with OXA for 72 hours (Fig. 3I). Consistent with previous observations, OXA increased Rac1b protein expression in both the control HCT116 cells and Rac1b overexpressing cells.
As expected, the HCT116-Rac1b cells had higher basal NF-κB activity as compared with HCT116-EV cells. Upon OXA treatment, the HCT116-control cells significantly enhanced NF-κB activity as well as the HCT116-Rac1b overexpressing cells. Thus, upregulation of Rac1b activates NF-κB signaling and occurs when HCT116 colorectal cancer cells are treated with OXA. Upregulation of Rac1b upon OXA was also observed in HT29 cells, indicating that this is not a cell type–specific phenomenon (Fig. 3J). Rac1b was also upregulated in a similar manner upon treatment of 5-FU, indicating that upregulation of Rac1b may be a common mechanism for overcoming resistance to chemotherapeutics (Supplementary Fig. S5).
To test the effects of Rac1b expression on chemoresistance of colorectal cells in vivo, HCT116-EV control and HCT116-Rac1b overexpressing cells were inoculated subcutaneously into nude mice. Once tumors reached 100 mm3, mice were randomized to vehicle or treatment arms. The HCT116-EV tumors treated with the OXA grew significantly slower compared with vehicle-treated group (Fig. 3K). However, no difference in the tumor volume was observed in HCT116-Rac1b overexpressing cells treated with OXA as compared with vehicle-treated group (Fig. 3K). The average tumor weight was significantly smaller for HCT116-EV OXA treated as compared with vehicle, but no difference was seen between HCT116-Rac1b treated with OXA as compared with vehicle-treated group (Fig. 3L). From these findings we conclude that Rac1b is upregulated upon treatment of colorectal cancer cells with chemotherapeutic agents and provides resistance to chemotherapy. The Rac1b-mediated acquired resistance is mediated by activation of the NF-κB pathway.
GYS32661 inhibits both Rac1 and Rac1b in colorectal cells
Although Rac1 and Rac1b are established drivers of cancer in a number of different malignancies (8, 31, 24), clinically available compounds have yet to be developed. Rac1 inhibitors such as NSC23766 and EHT1864 have moderate effects in a number of different diseases, including cancer (32–35). However, these compounds lack efficacy in vivo. Therefore, there exists a need for inhibitors with improved pharmacologic properties. Modification of the central heterocyclic ring structure to an imidazole while maintaining the quinolone and morphaline moieties improved both the activity of the compound while driving down metabolic clearance as measured by human liver microsomes (hCLint) (Fig. 4A and B). To demonstrate that GYS32661 inhibits Rac1, in vitro pull-down assay using the PBD was conducted with recombinant Rac1 protein (Fig. 4C). Pull-down results confirmed that GYS32661 at IC50 1.18 μmol/L inhibited activated Rac1 (Fig. 4C).
To confirm that GYS32661 is capable of inhibiting endogenous Rac1 and Rac1b in colorectal cancer cells, Rac1 and Rac1b in vivo activation assays were conducted using GYS32661 in colorectal cancers cells (Fig. 4D). GYS32661 treatment showed significant inhibition of Rac1 and Rac1b activity at 5 μmol/L in HCT116 colorectal cancer cells (Fig. 4D). Therefore, compound GYS32661 is a potent Rac inhibitor capable of inhibiting both Rac1 and Rac1b in vivo. To investigate the effects of Rac1 and Rac1b inhibition on colorectal cancer cell lines, cellular viability assays were performed on a panel of colorectal cancer cell lines (Supplementary Fig. S6A). All of the colorectal cancer cell lines were sensitive to Rac inhibition with IC50 ranging from 2.9 to 8.2 μmol/L (Supplementary Fig. S6A). Canonical Rac signaling results in enhanced cell migration and is a prerequisite for tumor invasion and metastasis (36, 37). To determine if Rac inhibition would reduce cellular motility, Boyden chamber migration assays were conducted on HCT116 cells. Rac inhibition significantly reduced the migratory ability of HCT116 cells (Fig. 4E). Rac inhibition was able to significantly reduce the number of colonies in the soft agar (Fig. 4F). To confirm that GYS32661 was capable of attenuating NF-κB signaling, luciferase assays were conducted on HCT116 cells. Cells were stimulated with TNFα, a potent activator of NF-κB signaling in colorectal cancer (38), and incubated with increasing concentrations of GYS32661 (Fig. 4G). Rac1b overexpression in HCT116 cells showed enhanced NF-κB reporter activity (Fig. 2E). To study whether GYS32661 would overcome the Rac1b mediated enhanced NF-κB reporter activity, HCT116-EV and HCT116-Rac1b cells expressing the NF-κB reporter was treated with different doses of GYS32661. 10 μmol/L GYS32661 significantly blocked Rac1b-mediated NF-κB activity (Supplementary Fig. S6B). Rac inhibition with GYS32661 significantly attenuated TNFα-stimulated NF-κB activity. Under conditions when NF-κB signaling is inactive, IκBα sequesters NF-κB (RelA/p65) in the cytosol. Upon stimulation with TNFα, IKBα, becomes phoshorylated targeting it for proteasomal degradation (39, 40). Once released, RelA/p65 becomes phosphorylated and forms an active dimer with the p50 NF-κB subunit. Once the subunits dimerize, NF-κB translocates to the nucleus and turns on target gene transcription (39, 40). Additional confirmation of the ability of GYS32661 to inhibit NF-κB signaling was performed by observing activation of upstream IKBα and phosphorylation of the RelA/p65 subunit by the Western blot analysis (Fig. 4H). HCT116 cells stimulated with TNFα resulted in phosphorylation of IKBα and P65. However, treatment with GYS32661 attenuated phosphorylation of both IKBα and RelA/p65 (Fig. 4H). These data confirm that inhibition of Rac with GYS32661 inhibits downstream activation of NF-κB signaling in colorectal cancer.
To test the effects of Rac inhibition of colorectal cells in vivo, HT29 colorectal cancer cells were inoculated subcutaneously into the flanks of nude mice. Once tumors reached 100 mm3, mice were randomized to treatment groups. The tumors treated with the Rac inhibitor grew significantly slower and the average tumor weight was significantly smaller compared with the vehicle control group (Fig. 4I). Moreover, no change in the body weight was observed at the end of the experiment indicating that systemic treatment with a Rac inhibitor is tolerable (Supplementary Fig. S6C). Altogether, Rac inhibition resulted in reduced cellular viability, migration, and colorectal cancer cell growth both in vitro and in vivo.
Inhibition of Rac sensitizes colorectal cancer cells to chemotherapy
As we showed above that Rac1b is upregulated in response to OXA and 5-FU treatment resulting in elevated NF-κB signaling, we wanted to study whether treatment with the Rac inhibitor will resensitize colorectal cancer cells to OXA. To do so, a colony formation and soft agar assays were conducted on HCT116 cells in the presence of OXA, GYS32661, or the combination of both drugs (Fig. 5A and B). For colony formation assay, the single agent treatment of HCT116 cells with OXA or GYS32661 resulted in a 60% and 55% reduction, respectively. When the two drugs were added in combination, however cellular viability dropped to 95% (Fig. 5A). Similarly in soft agar HCT116 cells treated with OXA resulted in a 60% reduction and GYS32661 resulted in a 55% reduction in growth in soft agar. However, the combination of OXA and GYS32661 resulted in 85% reduction in growth in soft agar (Fig. 5B).
To study whether combination treatment overcomes the OXA-mediated increase in NF-κB signaling, HCT116 and HT29 cells expressing the NF-κB reporter were treated either with DMSO, GYS32661, OXA, or combination of both. As shown above, OXA treatment increased the NF-κB reporter activity, which was drastically reduced when the cells were treated with OXA and Rac inhibitor (Fig. 5C and D). These results were in concordance with activation of NF-κB signaling pathway, where treatment with OXA or 5-FU showed increased IκB and p65 phosphorylation and combining with GYS32661 results in downregulation of phospho-IκB and phospho-p65 in HT29 colon cancer cell (Fig. 5E). Similar results were observed in HCT116 cells were OXA-mediated IκB and p65 phosphorylation was attenuated by combining Rac inhibitor GYS32661 (Supplementary Fig. S7). To study the in vivo effect of the combination of Rac inhibitor GYS32661 with OXA, we subcutaneously inoculated 1 × 106 HCT116 cells in immunocompromised mice. When the tumors grew 100 mm3, the mice were separated into 4 groups—vehicle, OXA (10 mg/kg), GYS32661 (30 mg/kg), and OXA + GYS32661. The tumors treated with the combination grew significantly slower compared with OXA and GYS32661 alone treated tumors (Fig. 5F and G). Together, we show that combining Rac inhibition with OXA sensitizes colorectal cancer to OXA by inhibiting the NF-κB pathway.
In this study, we demonstrate for the first time that Rac1b plays a critical role in conferring chemoresistance in colorectal cancer. In colon cancer cells, expression of Rac1b is upregulated during treatment with the chemotherepeutic drugs 5-FU and OXA as part of a cellular stress response. The overexpression of Rac1b results in enhanced NF-κB, signaling driving both cellular growth and survival of colorectal cancer cells (Fig. 5H). These results implicate Rac1b as playing a significant role in the development of chemotherapeutic drug resistance in colorectal cancer.
The proto-oncogene Rac1 has been reported to drive progression and metastasis of numerous cancers, including colorectal cancer. Rac1b, an alternatively spliced variant of Rac1 that is constitutively active, has been previously reported to be upregulated in primary colorectal tumors (18). High Rac1b expression is also reported in breast and lung cancer (8, 10). In a large sample cohort of patient data, our present study showed a similar observation of high Rac1b expression in patients with colorectal adenocarcinoma. The expression of Rac1b protein was significantly high in patients with metastasis compared with nonmetastatic colorectal cancer patients having localized disease and healthy controls. Rac1b has been associated with poor prognosis in KRAS/BRAF WT metastatic colorectal cancer patients treated with first-line FOLFOX/XELOX therapy (41). Previous studies demonstrate that Rac1b stimulates cell proliferation by enhancing the NF-κB–mediated signaling in colorectal and thyroid cancer cells (12, 13, 24, 42). In our study, we also showed that Rac1b expression provides a proliferative advantage to colorectal cancer cells via activation of the NF-κB pathway.
Treatment of unresectable colorectal cancer is empirical and most patients are candidates for FOLFOX therapy as a first-line treatment (43, 44). However, approximately 40% of colorectal cancer become refractory to FOLFOX. The development of resistance occurs by several processes, like enhanced DNA damage repair, altered expression of oncogenes, reduced apoptosis, decreased drug import, and enhanced tolerance to platinum adduct accumulation (3, 45). One of the major clinical challenges in treating colorectal cancer is identifying the subset of patients who will respond to chemotherapy or alternatively, identifying patients that will be resistant to chemotherapy. Studies looking at gene expression signatures of patients with colorectal cancer have suggested that biomarkers can identify patients that may respond poorly to conventional therapy (46). However, biomarkers of resistance that also guide subsequent treatment options for clinicians are still needed. Overexpression of Rac1b has been associated with poor prognosis in KRAS/BRAF WT metastatic colorectal cancer patients treated with FOLFOX therapy (41). In our studies, we observed that Rac1b expression was upregulated upon 5-FU and OXA treatment of colorectal cancer cells. Expression of Rac1 has been linked to chemo and radioresistance in head and neck squamous cell carcinomas (HNSCC), where Rac1 expression was increased after treatment with cisplatin or ionizing radiation (47). Inhibition of Rac1 in HNSCC cells restores anoiksis, decreases cell motility, and enhances cell sensitivity to cisplatin and ionizing radiation (47). Similar results were shown in doxorubicin-resistant squamous cell carcinoma (SCC), were pharmacological inhibition of Rac1 restores doxorubicin sensitivity (48). Rac1 expression is also implicated in radioresistance in different cancer cell types such as breast cancer, glioblastoma, pancreatic, and cervical cancers (49, 50, 51, 52).
It has now been well established that the sensitivity of colorectal cancer to OXA-induced death is adversely affected by elevated NF-κB activity (53, 54). Colorectal cancer cell lines with acquired resistance to OXA showed increased activation of NF-κB in comparison to their matched sensitive parental cells, implicating NF-κB as a significant mediator of OXA resistance in these cells (53, 54). Our data and others have shown that Rac1b expression activates the NF-κB signaling pathway in colon, lung, and thyroid cancer (12, 13, 24). Interestingly, we showed that OXA treatment results in upregulation of Rac1b, which leads to further enhanced NF-κB–mediated signaling and a significant shift in the sensitivity to chemotherapeutic agents. This increase in Rac1b was observed in both endogenous and exogenous Rac1b, suggesting that the increase in Rac1b may be regulated in part at the protein level and merits furthers investigation. Nevertheless, the increase in NF-κB signaling and subsequent chemoresistance caused by Rac1b upregulation can be reversed through the use of a small molecule inhibitor targeting Rac1 and Rac1b both in vitro and in vivo. Importantly, combining the Rac inhibitor with OXA resensitizes colorectal cancer cells to OXA and was capable of significantly reducing tumor growth compared with either agent alone. The results suggest the combination of OXA and a Rac inhibitor may translate into improved outcomes in the clinic.
Overall, in this study we showed that Rac1b overexpression plays an important role in colon cancer progression from normal colon epithelium to colorectal adenocarcinoma and eventually metastatic disease. As Rac1b activates prosurvival pathways such as AKT and NF-κB, it is conceivable that Rac1b may be a critical component in the establishment of distal metastases allowing cancer cells to proliferate in a foreign environment and resist the effect of chemotherapeutic drugs. Therefore, Rac1b may be a useful clinical prognostic molecular biomarker for both disease progression as well as a marker for chemo resistance. Finally, we show evidence that the use of a small molecule inhibitor is both safe and effective in treating colorectal cancer alone or in combination with standard of care therapy.
Disclosure of Potential Conflicts of Interest
E.T. Goka has ownership interest (including stock, patents, etc.) in patents, stock, and employee. D.T.M. Lopez is an employee. M.E. Lippman is a board member at GENEYUS; and has Ownership Interest (including stock, patents, etc.) in GENEYUS. No potential conflicts of interest were disclosed by the other authors.
Conception and design: E.T. Goka, P. Chaturvedi, M.E. Lippman
Development of methodology: E.T. Goka, P. Chaturvedi, A. De La Garza, M.E. Lippman
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): E.T. Goka, P. Chaturvedi, D.T.M. Lopez, M.E. Lippman
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): E.T. Goka, D.T.M. Lopez, M.E. Lippman
Writing, review, and/or revision of the manuscript: E.T. Goka, P. Chaturvedi, D.T.M. Lopez, A. De La Garza, M.E. Lippman
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): D.T.M. Lopez, E.T. Goka, M.E. Lippman
Study supervision: E.T. Goka
We would like to thank Lucas Outcault for his technical assistance.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.