The novel small molecule SR18662 efficiently inhibits the growth of colorectal cancer in vitro and in vivo.

Abstract

Krüppel-like factor 5 (KLF5), a member of the SP/KLF family of zinc finger transcription factors, is overexpressed in human colorectal cancer specimens and this overabundance is associated with aggressive cancer development and progression. We demonstrated that mice haploinsufficient for Klf5 had reduced intestinal tumor burden in the background of germline mutation in Apc, a gate keeper of intestinal tumorigenesis. Based on a high-throughput screening strategy, we developed ML264, a small molecule compound that inhibits KLF5 , and showed that it inhibits growth of colorectal cancer (CRC) in vitro and in vivo. Through optimization efforts based on structure of ML264, we have now identified a new lead compound, SR18662. We find that treatment with SR18662 significantly reduces growth and proliferation of CRC cells as compared to treatment with vehicle control, ML264, or SR15006 (a less optimized analogue from SAR efforts leading to SR18662). SR18662 showed improved efficacy in reducing viability of multiple CRC cell lines. Flow cytometry analysis following SR18662 treatment showed an increase in cells captured in either S or G2/M phases of the cell cycle and a significant increase in the number of apoptotic cells, the latter a unique property compared to ML264 or SR15006. SR18662 treatment also reduces the expression of cyclins and components of MAPK and WNT signaling pathways. Importantly, we observed a significant dose-dependent inhibition of xenograft growth in mice following SR18662 treatment that exceeded the effect of ML264 at equivalent doses. These findings support further development of SR18662 and its analogues for colorectal cancer therapy.

Introduction

Colorectal cancer (CRC) is the third most commonly diagnosed cancer and the third leading cause of cancer-related death in men and women in the US (1). Current approaches for treatment of CRC include surgery, radiation-, chemo-, immuno-, and targeted therapy (2). Standard chemotherapy treatments lead to the inhibition of DNA synthesis and transcription, available targeted therapies block the function of VEGF, EGFR or other kinases (3), and immunotherapy targets PLD-1 and CTLA-4 on T cells (4). These treatments are often used in combination for advanced stages of CRC, produce multiple side effects, and in some cases are ineffective due to specific mutations acquired during cancer progression. It has been shown that colorectal cancer development and progression results from impairments of function of signaling pathways (5), particularly the WNT and KRAS signaling pathways, which are altered even in the early stages (5). Specifically targeting these pathways would be an intriguing approach, though no compounds known to act in this manner have successfully progressed through (or even into) clinical trials (6).

Krüppel-like factor 5 (KLF5) is zinc finger transcription factor that has been shown to play a pro-proliferative role in many tissues (7-10). Studies have demonstrated that KLF5 is highly expressed in the intestinal epithelium within the transit-amplifying zone of crypts and in the active intestinal stem cells (11-13). We and others have shown that KLF5 modulates downstream signaling from the MAPK, WNT, and PI3K signaling pathways. Moreover, KLF5 also mediates the function of many components of these pathways by unidentified feedback mechanisms (8, 14, 15). KLF5 plays a very important role during colorectal cancer development and progression. Studies showed that the deletion of Klf5 from active intestinal cancer stem cells expressing the Leucine Rich Repeat Containing G Protein-Coupled Receptor 5 (Lgr5) and bearing an activating Ctnnb1 mutation prevents formation of colorectal tumors (13). KLF5 in conjunction with YAP1 also ensures the renewal of colon cancer progenitor cells (16). Moreover, data from our laboratory demonstrated that the deletion of one copy of the Klf5 gene in the context of an Apc^Min/+ mutation or with combination of Apc^Min/+ and KRAS^V12G mutations in mice significantly reduces development of intestinal polyps (17-19). A recent publication by Zhang and colleagues demonstrated that KLF5 undergoes diverse types of genomic alterations during cancer development that activate its oncogenic properties (20). Taken together, these lines of evidence suggest that the development of small molecules capable of inhibiting the expression level and/or activity of KLF5 may be beneficial for colorectal cancer treatment.

Using a high-throughput screening approach, we previously developed a luciferase assay to screen for compounds that inhibit the activity of the human KLF5 promoter (21, 22). Following structure-activity relationship studies we identified a small molecule lead, ML264, and showed that it is efficacious in CRC cell growth inhibition assays, both in vitro and in vivo (23, 24). Here, we show that further structural refinements have led to a new lead compound, SR18662, with an even higher efficacy for CRC cell growth inhibition in vitro. Moreover, we show that SR18662 significantly reduces the growth of tumors in a mouse xenograft model.

Materials and Methods

Cell lines and reagents

DLD-1 (CCL-221), HCT116 (CCL-247), HT29 (HTB-38), and SW620 (CCL-227) colorectal cancer cell lines were purchased from the American Type Culture Collection (ATCC) in 2015, 2016, and 2018. DLD-1 and SW620 cells were maintained in RPMI1640 medium supplemented with 10% FBS and 1% penicillin/streptomycin, while HCT116 and HT29 in McCoy’s medium supplemented with 10% FBS and 1% penicillin/streptomycin. The DLD-1/pGL4.18hKLF5p cell line was maintained in RPMI1640 with 10% FBS and 1% penicillin/streptomycin supplemented with 800 μg/mL of geneticin (21). The cell lines were passaged for three months while used for experiments. We tested all cell lines for Mycoplasma contamination upon thawing and routinely performed morphology checks on all tested cell lines. Furthermore, each experiment had appropriate controls to assure the behavior of tested cell lines and was performed with an appropriate number of biological replicates. The previously-described lead compound ML264 and subsequent new leads SR15006 and SR18662 were synthesized at The Scripps Research Institute in the laboratory of Dr. Thomas Bannister. For in vitro experiments the lead compounds were dissolved in dimethyl sulfoxide (DMSO, Fisher Scientific).

ML264, SR15006, and SR18662 synthesis

The synthesis of ML264 has been previously described (23). SR15006 and SR18662 were prepared in 3 steps as shown in Supplementary Figure 1 with the following protocol. Step 1: A solution of 1-methylsulfonyl-piperazine (1.2g, 7.3 mmol, 1.0 eq) and Cbz-Glycine (1.5 g, 7.3 mmol, 1 eq.) in DMF was treated with diisopropylethylamine (1.27 mL, 21.9 mmol, 3.0 eq.) at room temperature, stirred 10 min, then the coupling agent EDCI (1.25 g, 8.03 mmol, 1.1 eq) was added along with HOBt (1.09 g, 8.03 mmol, 1.1eq). After 8 h aq. sat NH₄Cl solution was added. Extraction with ethyl acetate, drying, and concentration of the extracts gave a crude colorless solid (>95% pure by LCMS, 2.2g, 85%) which was used in the next step without further purification. Step 2: The product of step 1 (2.2 g, 6.2 mmol) was dissolved in a minimal amount of a 1:1 mixture of MeOH and THF in a pressure vessel. 10 mol% of Pd/C catalyst (20% Pd by weight) was added and the flask was pressurized with H₂ to 55 psi. This mixture was shaken for 12 h. The vessel was depressurized, the solution was bubbled with nitrogen, the mixture was filtered through Celite, and the solution was concentrated. A colorless solid (>95% pure by LCMS, 1.0g, 74%) was obtained and was again used in the next step without further purification. Step 3: The product of step 2 (30 mg, 0.14 mmol) and 3-chlorocinnamic acid (25 mg, 0.14 mmol, 1.0 eq.) were together dissolved in a minimal amount of dry DMF. Diisopropylethylamine (71 uL, 0.42 mmol, 3.0 eq.) was added at room temperature, as was the coupling agent HATU (54 mg, 0.143 mmol, 1.05 eq) in small portions. After 8 h the yellow solution was quenched with aq. sat. NH₄Cl solution. Extraction with ethyl acetate, drying, and concentration of the extracts gave a colorless solid. This material was further purified by silica gel chromatography, giving pure SR15006 (35 mg, 67%). Alternatively, this third step was run instead using 3,4-dichlorocinnamic acid, giving pure SR18662 in 69% yield. Supporting data for (E)-3-(3-chlorophenyl)-N-(2-(4-(methylsulfonyl)piperazin-1-yl)-2-oxoethyl)acrylamide (aka SR15006): MS(ESI): m/z 386 [M+1]⁺; ¹HNMR (400Hz, CDCl₃) δ (ppm) 2.82 (s, 3H), 3.26-3.31 (dt, J = 4.8 Hz, 4H), 3.59 (t, J = 4.8 Hz, J = 5.2 Hz, 2H), 3.80 (t, J = 4.8 Hz, J = 5.2 Hz, 2H), 4.20 (d, J = 4.0 Hz, 2H), 6.50 (d, J = 15.6 Hz, 1H), 6.73 (t, J = 4.0 Hz, 1H), 7.29-7.40 (m, 3H), 7.51 (s, 1H), 7.57 (d, J = 15.6 Hz, 1H). HPLC purity >95%. Supporting data for (E)-3-(3,4-dichlorophenyl)-N-(2-(4-(methylsulfonyl)piperazin-1-yl)-2-oxoethyl)acrylamide (aka SR18662): MS(ESI): m/z 420 [M+1]⁺; ¹H NMR (400Hz, DMSO) δ (ppm) 2.90 (s, 3H), 3.11 (t, J = 6.0 Hz, 2H), 3.16 (t, J = 6.0 Hz, 2H), 3.57 (d, J = 4.8 Hz, 4H), 4.13 (d, J = 5.2 Hz, 2H), 6.94 (d, J = 16.0 Hz, 1H), 7.42 (d, J = 16.0 Hz, 1H), 7.59 (dt, J = 2.0 Hz, J = 8.4 Hz, 1H), 7.68 (d, J = 8.4 Hz, 1H), 7.88 (d, J = 2.0 Hz, 1H), 8.19 (t, J = 5.2 Hz, 1H). HPLC purity >95%.

KLF5 promoter activity assay

DLD-1/pGL4.18hKLF5p cells were seeded in 96 well plate format and treated with DMSO or with test compounds dissolved in DMSO in the range of 0.001 to 20 μM final concentration for 24 h and the human KLF5 promoter activity was determined with the ONE-Glo luciferase assay system (Promega) using a SpectramMax M3 (Molecular Devices) plate reader. The IC₅₀ values were calculated using GraphPad Prism version 5.00 for Windows (GraphPad Software) (22).

Cell viability

DLD-1, HCT116, HT29, and SW620 cells were seeded in 96 well plate format and treated with DMSO or with compounds dissolved in DMSO in the range of 0.001 to 20 μM final concentration for 24 h and analyzed with the Cell Titer-Glo luciferase assay system (Promega) using a SpectraMax M3 (Molecular Devices) plate reader. The IC₅₀ values were calculated using GraphPad Prism version 5.00 for Windows (GraphPad Software).

Cell proliferation, cell cycle and apoptosis assays

For cell proliferation experiments, cell cycle, and apoptosis determination, DLD-1 and HCT116 cells were treated with 1 or 10 μM of each compound or with vehicle (DMSO). The cells were collected at 24, 48, and 72 h post-treatment and analyzed as previously described (24). Each experiment was performed in triplicate.

Western blot analysis

Total protein was extracted from cells with Laemmli buffer and the analysis was performed as described previously (24).

Immunofluorescence and immunohistochemistry

Tumors dissected from mice were first fixed in Bouin’s fixative (50% ethanol + 5% acetic acid in water) for 1 h, then fixed overnight in 10% buffered formalin (Fisher Scientific). The tissues were then paraffin-embedded using an automated processor, sectioned at 5μm, collected onto charged slides and baked in a 65°C oven overnight, and were subsequently deparaffinized in xylene. Sections were incubated in a 2% hydrogen peroxide in methanol bath to block endogenous tissue peroxidases and were then rehydrated by incubation in a decreasing ethanol bath series (100%, 95%, 70%) followed by antigen retrieval in citrate buffer solution (10 mM sodium citrate, 0.05% Tween-20, pH 6.0) at 120°C for 10 min using a decloaking chamber (Biocare Medical). Tissue sections were first incubated with blocking buffer (5% BSA in TBS-Tween) for 30 min at 37°C and then with primary antibody at 4°C overnight in a humidified chamber with gentle shaking (24). Sections were washed and incubated with secondary antibodies (HRP-conjugated or fluorescent-tagged) at the appropriate concentration for 30 min at 37°C. Betazoid DAB (Biocare Medical) was used to reveal IHC staining in tissues. For fluorescent sections, slides were washed after secondary antibody treatment and then stained with Hoechst (AnaSpec Inc.) and mounted with Prolong gold antifade (Life Technologies). Slides were analyzed under a Nikon Eclipse 90i microscope (Nikon) and representative photomicrographs were taken.

H&E stain

Five μm sections were fixed, paraffin-embedded, deparaffinized and rehydrated as discussed above. Then the sections were stained with Hematoxylin Stain Solution, Gill 3 (Ricca Chemical Company) and Eosin Y (Sigma-Aldrich), dehydrated in an increasing series of ethanol bath (70%, 95%, 100%), cleared in xylene and mounted with Cytoseal XYL xylene-based mounting media (Thermo Scientific). Images were taken using a Nikon Eclipse 90i microscope (Nikon).

TUNEL stain

Five μm sections of formalin-fixed, paraffin-embedded tissues were processed according to the manufacturer’s protocol and stained with In Situ Cell Death Detection Kit, TMR Red (Sigma Aldrich).

Xenografts

All mice studies were approved by the Stony Brook University Institutional Animal Care and Use Committee (IACUC protocol number: 354445) and performed in accordance with institutional policies,NIH and ARRIVE guidelines. Nude mice were purchased from Jackson Laboratories (NuJ/Foxn1^nu, Stock number: 002019, Bar Harbor,). Animals were housed under specific pathogen-free conditions in ventilated and filtered cages under positive pressure. DLD-1 human colorectal cells at concentration of 5×10⁶ were subcutaneously injected into the right flank of 7 week old male nude mice. Tumor volume was determined by caliper measurement and calculated by the established method (25). When tumors reached a volume of about 100 mm³, mice were treated intraperitoneally (i.p.) with varying doses of SR18662: 5mg/kg daily, 5mg/kg twice a day,10 mg/kg daily, 10 mg/kg twice per day, 25mg/kg daily, and 25 mg/kg twice per day. Each treatment regimen was as follows: 5 days of injections, 2 days break, and 5 days of injections, and mice were collected 24 h after the last injection. The vehicle solution (30% 2-hydroxypropyl-beta-cyclodextrin) was used as the control treatment. Mice were monitored and weighed every two days. Tumors were excised and retained for further analyses.

Statistical analysis

The analysis of significance of in vitro experiments was performed using a student’s t-test, with a value of p < 0.05 considered significant. This analysis was performed using GraphPad Prism version 5.00 for Windows (GraphPad Software). The analysis of tumor growth was performed using a linear mixed effect model for longitudinal data which was used to compare volume difference of each dose treatment versus vehicle after 5 days of injections and after 10 days of injections. Mice were selected at random for the various treatment schedules and treated as a random effect in the linear mixed effect model. Such analysis was performed in SAS 9.4 (SAS Institute Inc., Cary, NC) and significance level was set at 0.05.

Results

The design of SR18662

We previously described ML264 as a potent inhibitor of colorectal cancer growth in vitro and in vivo (24), showing that the growth of the CRCs and xenografts in nude mice were arrested upon treatment with ML264 in comparison to vehicle. We sought to perform structure-activity relationship studies to further optimize ML264 with respect to anti-tumor activities and pharmacokinetic properties (Fig. 1A), focusing upon chemistry that would permit greater structural diversification. These efforts led to the discovery of SR15006 and SR18662 (Fig. 1A), which, while structurally related to ML264, are markedly different in the glycine amide region.

Figure 1. SR18662 induces anti-tumor activity in colorectal cancer cell lines.

(A) Chemical structures of ML264, SR15006, and SR18662. (B) Percentage of the human KLF5 promoter activity as measured by a luciferase assay in the DLD-1/pGL4.18hKLF5p cell line upon ML264, SR15006, and SR18662 24 h treatment. ---- lines represent ML264, •••• lines – SR15006, and -••-•• lines – SR18662, (C-D) Proliferation assays of DLD-1 (C) and HCT116 (D) cells treated with DMSO or with 10μM ML264, SR15006, or SR18662 at 24, 48 and 72 h. The solid lines represent control (DMSO- treated), ---- lines - ML264, •••• lines – SR15006, and -••-•• lines – SR18662. Data represent mean ± SEM (n=3). *p < 0.05, ***p < 0.001, (E-F) Percentage cell growth of DLD-1 (E) and HCT116 (F) cells treated with DMSO or 10μM ML264 or SR15006 or SR18662 at 24, 48 and 72 h. Data represent mean ± SEM (n=3). *p < 0.05, **p < 0.01, ***p < 0.001.

SR18662 is potent inhibitor of CRC cells growth in vitro

Firstly, to assess the effect of the new compounds on the activity of the human KLF5 promoter, we used a DLD-1 CRC cell line stably transduced with the luciferase reporter gene under the control of the human KLF5 promoter (DLD-1/pGL4.18hKLF5p cell) (21). The results showed that SR18662 is much more potent than ML264 and SR15006 and has an IC₅₀ of 4.4 nM, comparing to ML264 and SR15006 whose IC₅₀ values are 43.9 nM and 41.6 nM, respectively (Fig. 1B). To examine the impact of these compounds on viability of CRC cell lines, we performed cell proliferation and cell growth assays using DLD-1 and HCT116 CRC cells. As shown in Figures 1C-F, the three compounds, each tested at 10 μM, significantly inhibited proliferation and growth of CRC cells over the course of three-day treatment in comparison to vehicle-treated cells. Additional analysis showed that SR18662 demonstrated a robust inhibitory effect, not only in comparison with vehicle but also in comparison to ML264 and SR15006 (Suppl. Table 1). To investigate if these compounds exhibit a similar effect on other CRCs, we performed comparative cell viability analysis of four CRC cell lines (DLD-1, HCT116, HT29, and SW620) (Figs. 2A-C). SR18662 had the highest inhibitory effect on HT29 and SW620 CRC cell types as well, with an IC₅₀ value about one log unit lower than those for ML264 and SR15006.

Figure 2. ML264, SR15006, and SR18662 inhibit viability of multiple colorectal cancer cell lines.

Percentage cell viability of colorectal cancer cell lines treated with ML264 (A), SR15006 (B), or SR18662 (C). Cells were treated with test compounds for 24 h and cell viability was measured using Cell Titer Glo. Each experiment was performed in triplicate and data is shown as mean ± SEM (n=3). • represents DLD-1, ■ – HCT116, ▲ – HT29, and ▼ – SW620 cell line.

SR18662 alters the cell cycle pattern of CRCs

Previously, we showed that ML264 can modify the cell cycle of CRC cells in comparison to vehicle (24). We now evaluated the impact of SR15006 and SR18662 on cell cycle progression. We treated DLD-1 and HCT116 CRC cells over the course of three days with vehicle or 1 μM or 10 μM of test compounds and then analyzed cell cycle using flow cytometry. As previously shown, treatment with 10 μM ML264 caused a significant decrease in the number of cells in the G0/G1 phase and an increase in cells in S- and G2/M phases of the cell cycle, in both DLD-1 and HCT116 cells over three days (Figs. 3A-F). SR15006 treatment affected the cell cycle in the similar way as did ML264. SR18662 treatment, however, showed a different pattern (Figs. 3A-F). There was a decrease in the number of cells in G0/G1 phase, however to a lesser degree than had been observed following treatment with either ML264 or SR15006. Furthermore, in the case of SR18662 there was a significant increase in the number of cells within the subG1 population, in comparison to not only vehicle-treated cells but also to cells treated with ML264 or SR15006. Similar patterns of changes in the cell cycle were observed upon treatment of DLD-1 and HCT116 cells with 1 μM of these compounds over three days (Suppl. Fig. 2 and Suppl. Table 2). The increase in the number of cells in subG1 suggests that SR18662 may not only inhibit growth of CRCs and have cytostatic activity (as does ML264), but that SR18662 may also cause cell death and have cytotoxic potential. To address this thesis, we performed apoptosis assays using DLD-1 and HCT116 cells treated with vehicle or 1 μM or 10 μM of test compounds over the course of three days with Annexin V/PI stain in combination with FACS analysis. Our data showed that treatment of DLD-1 cells with 10 μM SR18662 caused a significant decrease in the population of healthy cells at 48 and 72 h in comparison to control, and also in comparison to treatment with 10 μM ML264 or SR15006 (Figs. 4A-C and Suppl. Table 3). This is accompanied by an increase in the population of cells in early and late apoptosis, as well as an increase in the dead cell population over the three days of treatment. In addition, even treatment of DLD-1 cells with 1 μM SR18662 caused a decrease in healthy cell count at 48 h and 72 h and an increase in early and late apoptosis, as compared to vehicle- or ML264- or SR15006-treated cells (Suppl. Figs. 3A-C, and Suppl. Table 3). However, treatment of HCT116 with all three compounds decreased the population of healthy cells and gradually increased the population of cells in the early and late apoptosis (Figs. 4D-F). This effect was more pronounced upon treatment with SR18662 than ML264 or SR15006 (Figs. 4D-F). When a lower concentration (1 μM) of test compounds was used, only a small increase of the population of cells undergoing early apoptosis was seen after 72 h treatment with SR18662 (Suppl. Figs. 3D-F). These data suggest that treatment with all three compounds alters the cell cycle of CRCs and that treatment with SR18662 can additionally induce cell apoptosis and death.

Figure 3. SR18662 changes cell cycle profiles of colorectal cancer cell lines.

DLD-1 and HCT116 cells were treated with DMSO or 10 μM ML264, SR15006, or SR18662 for 24 (A and D), 48 (B and E) and 72 h (C and F), respectively, stained with propidium iodide, and analyzed by flow cytometry. Each experiment was performed in triplicate and data is shown as mean ± SEM (n=3). *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 4. SR18662 increases apoptosis of colorectal cancer cell lines.

DLD-1 and HCT116 cells were treated with DMSO or 10 μM ML264, SR15006, or SR18662 for 24 (A and D), 48 (B and E) and 72h (C and F), respectively, and the apoptosis rate was determined using the Alexa Fluor 488 Annexin V/Dead Cell Apoptosis Kit analyzed by flow cytometry. Each experiment was performed in triplicate and data is shown as mean ± SEM (n=3). *p < 0.05, **p < 0.01, ***p < 0.001.

SR18662 negatively regulates MAPK and WNT signaling pathways and cyclin levels

Myriad publications have shown that in both in vitro and in vivo contexts KLF5 is regulated by MAPK and WNT signaling pathways and that KLF5, in turn, has the ability to modulate the activity of these pathways (14, 18, 26-32). We have previously shown that treatment of CRCs with ML264 results in decreased activities of both the MAPK and WNT signaling pathways (24). Because SR15006 and SR18662 are structurally related to ML264, we decided to investigate their ability to alter the expression levels of the components of these pathways. Thus, we treated DLD-1 and HCT116 CRCs with DMSO (vehicle) and tested compounds at 1 μM or 10 μM, collecting the cells for western blot analysis at 24, 48, and 72 h. As shown on Figure 5A, after 72 h of treatment, ML264, SR15006 and SR18662 all negatively affected the levels of the components of the MAPK pathway. Treatment with these compounds decreased the basal levels of EGFR and ERK proteins and additionally altered their phosphorylation status. Importantly, KLF5 and its direct transcriptional activator EGR1 are significantly downregulated upon treatment with each of these compounds as compared to vehicle (14, 33). Similar effects were seen when both cell lines were treated with these compounds at 1 μM (Suppl. Fig. 4A). Furthermore, the inhibitory effects of SR18662 on the components of the MAPK signaling pathway were observed at earlier time points than were pronunced following treatment with ML264 or SR15006 (Suppl. Fig. 5). It has been previously shown that WNT signaling regulates the expression of KLF5 (28, 29, 31, 34) and in turn KLF5 can modulate the levels of β-catenin, altering its nuclear accumulation and perturbing its transcriptional activity (30, 35). Thus, we examined the impact of tested compounds on the components of the WNT signaling pathway and observed a substantial reduction upon treatment (Fig. 5B). The treatment most notably results in a significant downregulation of AKT and its active form, phosphorylated at Serine 473 (36), which has the ability to phosphorylate and activate β-catenin (37). As shown in Figure 5B, we observed not only a reduction in the expression levels of AKT but also of total β-catenin and active β-catenin, which is phosphorylated at Serine 552 (38). A downregulation of the expression levels of the components of the WNT signaling pathway was also observed at 24 and 48 h (Suppl. Fig. 6). Moreover, the treatment with the compounds at 1 μM demonstrated a very similar pattern in inhibition of the WNT signaling pathway, with the effects most pronounced at 72 h (Suppl. Fig. 4B). Cyclins and their appropriate cyclin-dependent kinases form a regulatory network that allows cells to progress through the cell cycle (39, 40). The results presented in Figure 3 demonstrate substantial changes in the cell cycle progression upon treatment with ML264, SR15006, or SR18662, as compared to vehicle-treated cells. Thus, we investigated if these changes are accompanied by differences in the expression levels of cyclins. We focused our attention on cyclin D1, E, A2, and B1, as we have previously shown that ML264 inhibits their levels (24). The treatment of DLD-1 CRC cells with 1 μM or 10 μM of ML264, SR15006, or SR18862 caused a significant reduction in the levels of tested cyclins over the course of 72 h (Fig. 5C and Suppl. Fig. 4C). The treatment of HCT116 CRC cells with 10 μM concentration of these compounds showed significant reduction in the levels of cyclin E, A2 and B1. However, the treatment with 1 μM concentration of these compounds showed minimal or no changes in the levels of tested cyclins. The reduced levels of cyclins E, A2, and B1 upon treatment with these compounds correlate with the reduction of the cells in G0/G1 and the accumulation of the cells within S and G2/M phases, as the lack of these cyclins does not allow cells to enter or exit these specific stages of the cell cycle. Thus far, using an in vitro model, we demonstrated that SR18662 more effectively inhibits the proliferation of CRC cells. Importantly, in contrast to ML264 or SR15006, SR18662 induced apoptosis and cell death. Hence, we decided to investigate the ability of SR18662 to affect the growth of a tumor in a xenograft model.

Figure 5. SR18662 inhibits activity of MAPK, WNT/β-catenin signaling pathways and decreases the levels of cyclins.

DLD-1 and HCT116 cells were treated with DMSO or 10 μM ML264, SR15006, or SR18662 for 72 h. Representative western blots of selected components of MAPK (A), WNT/β-catenin (B) signaling pathways and cyclins (C).

SR18662 inhibits the growth of CRC xenografts in vivo

To evaluate SR18662 in an in vivo model we used CRC cells for a mouse xenograft study. As described in the Material and Methods section, nude mice were injected with DLD-1 CRC cells. Tumors were allowed to grow to an approximate size of 100 mm³, and then mice were injected with vehicle or with varied concentrations of SR18662, given once a day or twice a day for five days, followed by two days without injection, and then followed by five days with injections. The experiment was completed the day after the last injection. As shown in Figure 6A, we observed a dose-dependent reduction in tumor growth upon injection with SR18662, compared to vehicle-treated mice. Furthermore, tumor growth was significantly reduced after five days of SR18662 treatment (Suppl. Table 4). Significant differences in tumor size are apparent at the end of the treatment schedule (Fig. 6B). Over the course of the study we observed a loss of the weight of the animals treated with increased concentration of SR18662 (Suppl. Fig. 7). To further characterize the impact of SR18662 on the features of isolated tumors, we compared vehicle-treated mice with mice that had been treated with SR18662 twice a day. The initial histological analysis by H&E stain showed an increase in the level of inflammation and fibrosis in xenografts that had been treated with SR18662, in comparison to vehicle-treated xenografts (Fig. 6C). The PanCK/Vimentin stain confirmed a higher level of fibrosis in mice that had been treated with SR18662 (Figs. 6D and 6H). Because SR18662 had induced cell apoptosis and death in the in vitro setting, we stained the xenografts with PanCK and TUNEL to estimate the levels of apoptosis in the in vivo context. As shown in Figure 6E there is an increase of TUNEL staining in SR18662-treated xenografts in comparison to those that had been vehicle-treated.

Figure 6. SR18662 inhibits the growth of DLD-1-derived tumor xenografts in a nude mice model.

DLD-1 cells were subcutaneously injected into nude mice to develop xenograft tumors. Mice were then treated with vehicle with various doses of SR18662. (A) Growth of the tumors after injection with vehicle (•), or with SR18662 at 5 (■), 2×5 (▲), 10 (▼), 2×10 (♦), 25 (○), and 2×25 (□) mg/kg. The arrows indicate the days of injections. Data represent mean ± SEM (n=5). (B) Tumors sizes at the end of the treatment. Data represent mean ± SEM (n=5). *p < 0.05, ***p < 0.001. (C - G) Representative images of DLD-1–derived xenografts treated with vehicle (left) and treated with SR18662 twice per day at 25 mg/kg (right). (C) – H&E, (D) – PanCK and Vimentin, (E) – PanCK and TUNEL, (F) KLF5–, and (G) –EGR1. Quantification of Vimentin positive staining (H), KLF5 positive staining (I), and EGR1 positive staining (J) of DLD-1–derived xenografts treated with vehicle (left) and treated with SR18662 twice per day at 25 mg/kg (right). Data represent mean ± SEM (n=5). *p < 0.05, **p < 0.01.

ML264 and related compounds have been shown to inhibit the growth of CRCs and to decrease the levels of KLF5, a strong pro-proliferative transcription factor. We previously showed that ML264 inhibits the growth of CRC cells in vivo and reduces the levels of KLF5 and its direct transcriptional regulator EGR1. Hence, we decided to examine the levels of these proteins in xenografts treated with SR18662. Our immunohistochemistry analysis demonstrates that twice a day treatment with 25mg/kg of SR18662 significantly decreased the percentage of the cells that are positive for KLF5 and EGR1, as compared to vehicle-treated mice (Figs. 6F-G and 6I-J).

To summarize, we have generated a new small molecule lead, designated SR18662, that efficiently inhibits the growth of CRC cells both in vitro and in vivo and that reduces the expression levels of both KLF5 and EGR1.

Discussion

Many recent studies have demonstrated that KLF5 plays an important pro-proliferative role during tissue homeostasis and that under specific conditions KLF5 can become oncogenic and drive tumorigenesis (10-12, 18, 30, 41). We had previously shown that ML264, a small molecule compound discovered by an HTS campaign and follow-up SAR studies, has potent anti-colorectal cancer activity (23, 24). In the present studies we demonstrated that a new compound, SR18662, shows superior efficacy relative to ML264 in the inhibition of CRC cell proliferation in vitro and in vivo. We demonstrated that SR18662 more efficiently reduces proliferation of CRC cells in comparison to vehicle-treated cells but also showed significant improvement over treatment with either ML264 or SR15006 (Figs. 1 and 2, Suppl. Table 1). This enhanced ability of SR18662 in preventing CRC cell proliferation was observed in both microsatellite instable (DLD-1 and HCT116) and microsatellite stable (HT29 and SW620) cell types (42). Moreover, these cell lines have different genomic alterations of KRAS, BRAF, PIK3CA or TP53, suggesting that SR18662 may be effective against a broad spectrum of CRC cells (42). However, the strongest effect of SR18662 is more pronounced in the cell lines with relatively high levels of the KLF5 expression (DLD-1 and HCT116) as shown in Fig. 2. Furthermore, SR18662’s increased ability to inhibit growth of CRC cells, as compared to ML264 and SR15006, was accompanied by more potent inhibition of the activity of the human KLF5 promoter. Importantly, and in contrast to ML264 and SR15006, SR18662 not only modified the progression of the cell cycle but also induced apoptosis and death in CRC cells (Figs. 3 and 4, Suppl. Tables 2 and 3). Treatment of CRC cells with all three compounds led to reduced expression of the components of MAPK and WNT signaling pathways and to a reduction in cyclin levels. We have observed this previously while treating CRC cells with ML264. However, the effect of SR18662 is more pronounced at 1 μM and 10 μM over the three day time course. Furthermore, levels of EGR1, a direct transcriptional activator of KLF5, were significantly reduced, with notable more notable effect even at early time points following treatment with SR18662 compared with treatment with either ML264 or SR15006. This more rapid onset of activity may be an important advantage of SR18662 over its predecessors. Moreover, the potent activity of SR18662 from in vitro studies translated to an in vivo colorectal cancer xenograft model. While ML264 significantly reduced the growth of the CRC xenografts in the mouse model with treatment twice a day at a concentration of 10 mg/kg and 25 mg/kg, SR18662 showed strong inhibitory effects with treatment of twice a day at 5mg/kg and once a day at 10 mg/kg, and significant decrease in tumor growth with treatment once a day at 5mg/kg. The treatment of mice with ML264 and SR18662 resulted in the loss of body weight. The treatment with 10 mg/kg of ML264 resulted on average in the loss of 4.8%±3 of the starting body weight, with 2 × 10 mg/kg of 0.3%±2, with 2 × 25 mg/kg of 5%±3 while the treatment with 10 mg/kg of SR18662 with 1.5%±3, with 2 × 10 mg/kg with 5.4%±3, and with 2 × 25 mg/kg with 14.2%±4, respectively. The loss of body weight after the treatment with SR18662 was more pronounced than after the treatment with ML264 probably due to the cytotoxic feature of the new compound. Currently, the exact molecular mechanism of action of SR18662 is unknown and is under investigation. SR18662’s chemical structure suggests that it is a covalent and irreversible modifier of its target protein(s), likely acting upon an active site cysteine residue in its native target(s). It may directly modify a protein that associates with EGR1, since downregulation of EGR1 was seen within 24 h of treatment with SR18662, with a later decrease in KLF5 levels. Testing of SR18662 in patient-derived organoids (PDO) is planned, to assess its potency in a context more relevant to human CRC (43). We also plan to more fully describe, in a chemistry-focused publication, the molecular design and SAR studies leading to SR18662. Because ML264 has been shown to inhibit KLF5 expression and led to the repression of matrix degradation in the nucleus pulposus (NP), an effect that was enhanced by treatment with TGF-β (44), similar work with SR18662 or further-improved analogues are warranted. The finding that ML264 reduces the expression level of KLF5 in a breast cancer cell line in vitro (45) further suggests that targeting the transcription factor KLF5 may give benefits not only in CRC treatment but also in the treatment of other type of tumors where KLF5 plays a pro-proliferative role.