A clinical trial in patients with familial adenomatous polyposis (FAP) demonstrated that sulindac plus erlotinib (SUL+ERL) had good efficacy in the duodenum and colon, but toxicity issues raised concerns for long-term prevention. We performed a biomarker study in the polyposis in rat colon (Pirc) model, observing phosphorylated Erk inhibition in colon polyps for up to 10 days after discontinuing ERL+SUL administration. In a follow-up study lasting 16 weeks, significant reduction of colon and small intestine (SI) tumor burden was detected, especially in rats given 250 ppm SUL in the diet plus once-a-week intragastric dosing of ERL at 21 or 42 mg/kg body weight (BW). A long-term study further demonstrated antitumor efficacy in the colon and SI at 52 weeks, when 250 ppm SUL was combined with once-a-week intragastric administration of ERL at 10, 21, or 42 mg/kg BW. Tumor-associated matrix metalloproteinase-7 (Mmp7), tumor necrosis factor (Tnf), and early growth response 1 (Egr1) were decreased at 16 weeks by ERL+SUL, and this was sustained in the long-term study for Mmp7 and Tnf. Based on the collective results, the optimal dose combination of ERL 10 mg/kg BW plus 250 ppm SUL lacked toxicity, inhibited molecular biomarkers, and exhibited effective antitumor activity. We conclude that switching from continuous to once-per-week ERL, given at one-quarter of the current therapeutic dose, will exert good efficacy with standard-of-care SUL against adenomatous polyps in the colon and SI, with clinical relevance for patients with FAP before or after colectomy.

Prevention Relevance:

This investigation concludes that switching from continuous to once-per-week erlotinib, given at one-quarter of the current therapeutic dose, will exert good efficacy with standard-of-care sulindac against adenomatous polyps in the colon and small intestine, with clinical relevance for patients with FAP before or after colectomy.

Mortality rates for colorectal cancer vary according to geographic region, gender, age, ethnicity, and diet/lifestyle factors (1, 2). Genetic influences also contribute to the etiology of colorectal cancer, for both sporadic cases and hereditary syndromes, such as familial adenomatous polyposis (FAP) driven by germline mutation in Adenomatous Polyposis Coli (APC; refs. 2–4, 5). Patients with FAP develop hundreds to thousands of adenomatous polyps in the gut, necessitating risk‐reducing surgical intervention coupled to long-term chemoprevention strategies (6–9).

Sulindac (SUL) and other nonsteroidal anti-inflammatory agents are effective in the suppression of adenomatous colon polyps (6–9), but not without toxicity, and they have lower efficacy in the duodenum, where tumor formation also is of concern (10). In patients with FAP, a randomized clinical trial combined standard-of-care SUL (150 mg p.o. twice daily) with erlotinib (ERL, 75 mg p.o. daily) for 6 months, reducing the duodenal polyp burden markedly compared with historic controls (11). Secondary analysis identified a reduction in polyp burden in those with an entire colorectum, as well as those with ileal pouch or retained rectum after colectomy (12), suggesting broad-spectrum antitumor efficacy of the combination regimen in the gastrointestinal (GI) tract. However, Samadder and colleagues also noted that skin toxicity or other adverse events might limit the long-term use of ERL+SUL in combination at the doses tested in the clinical trial (11).

Efficacy has been reported for SUL in preclinical models of FAP, including the Apc‐mutant polyposis in rat colon (Pirc) model (13, 14). The Pirc model has a high colon tumor burden and a lifespan suitable for biomarker, prevention, toxicity, and resistance studies (13–20). Importantly, the Pirc model mimics the tumor distribution in patients with FAP; rats develop both duodenal polyps and adenomatous lesions in the colorectum (13–15, 21).

Antitumor activity of ERL also has been reported in preclinical models of colorectal cancer (22–24). Consistent with Wingless-related integration site (WNT) plus ERK pathway interaction (25, 26), WNT, EGFR, and prostaglandin E2 (PGE2)–related genes were identified as mechanistic targets modulated in patients with FAP taking ERL+SUL (27). We utilized the Pirc model to define scheduling and dosing regimens for ERL+SUL that would maintain antitumor efficacy while limiting toxicity. A secondary objective was to validate the above-mentioned mechanistic targets in the Pirc colon and small intestine (SI), with a view to future clinical trials in patients with FAP using optimized combination regimens.

Preclinical studies

Experiments were approved by the Institutional Animal Care and Use Committee. Under a Taconic Breeding License, Pirc males were generated for the various preclinical studies and genotyped as reported (13), with AIN93 and custom diets from Research Diets, Inc. ERL (>99%) was supplied by LC laboratories, whereas SUL (>98%) was obtained from MilliporeSigma. For intragastric (i.g.) oral gavage, the test agent was administered in sterile 30% Captisol, a modified β-cyclodextrin facilitating drug solubilization (28).

Short-term biomarker study

As shown in Fig. 1, 6-month-old rats were provided with AIN93 basal diet, or switched to 250 parts per million (250 ppm) SUL in AIN93 diet, equivalent to 150 mg SUL twice daily in patients with FAP (11). Based on a prior report (29), ERL was given in the first week by one of four treatment schedules: i.g. oral gavage every day at 6 or 12 mg/kg body weight (BW), i.g. at 21 mg/kg BW twice per week, or i.g. at 42 mg/kg BW once per week. Oral dosing of 6 mg/kg BW ERL for 1 week equates to the daily intake of 70 ppm ERL in rodent diet, or 75 mg ERL q.d. in patients with FAP (11). ERL at 42 and 21 mg/kg BW (ERL42, ERL21) corresponds, respectively, to “loading” and “half-loading” bolus doses from 1-week daily intake at 6 mg/kg BW i.g. (29). After the final ERL treatment, rats were sacrificed at 1, 2, 7, 10, and 14 days.

Figure 1.

Dose-response and time-course outcomes for biomarker modulation in Pirc colon tumors. A, Biomarker study protocol, with n = 5 rats per group. B, Immunoblotting of pErk normalized to Erk in Pirc colon tumors after animals were fed with AIN or SUL diets, or treated with dietary SUL plus i.g. ERL at 6 mg/kg BW (ERL+SUL). Representative findings from two or more independent experiments. The corresponding densitometry data in B are shown as mean ± SE; data bars sharing the same superscript letter were not significantly different by one-way ANOVA. C–F, Densitometry data were further analyzed at 6, 12, 21, and 42 mg/kg ERL±SUL, to determine the kinetics of pErk inhibition for up to 14 days after drug dosing, plotted as a ratio of the AIN control group in the corresponding immunoblot. Dotted line, normalized pErk/Erk level in AIN controls, given an arbitrary value of 1.0. Original uncropped immunoblotting images are presented in Supplementary Fig. S1.

Figure 1.

Dose-response and time-course outcomes for biomarker modulation in Pirc colon tumors. A, Biomarker study protocol, with n = 5 rats per group. B, Immunoblotting of pErk normalized to Erk in Pirc colon tumors after animals were fed with AIN or SUL diets, or treated with dietary SUL plus i.g. ERL at 6 mg/kg BW (ERL+SUL). Representative findings from two or more independent experiments. The corresponding densitometry data in B are shown as mean ± SE; data bars sharing the same superscript letter were not significantly different by one-way ANOVA. C–F, Densitometry data were further analyzed at 6, 12, 21, and 42 mg/kg ERL±SUL, to determine the kinetics of pErk inhibition for up to 14 days after drug dosing, plotted as a ratio of the AIN control group in the corresponding immunoblot. Dotted line, normalized pErk/Erk level in AIN controls, given an arbitrary value of 1.0. Original uncropped immunoblotting images are presented in Supplementary Fig. S1.

Close modal

Efficacy study

As shown in Fig. 2, at 6 weeks of age, rats were given AIN basal diet or switched to AIN diet containing 250 ppm SUL. In the presence or absence of SUL, other groups comprised of ERL42 and ERL21 qw (i.g. once/wk), ERL 70 or 140 ppm in the diet on continuous weeks (ERL70c, ERL140c), and ERL 70 ppm in the diet on alternating weeks (ERL70a). Adenomatous polyps were tracked via colonoscopy (14–16) after 6, 10, and 14 weeks of ERL±SUL treatment, and 2 weeks thereafter rats were sacrificed, i.e., at 16 weeks of drug treatment in total, coinciding with the approximate age of the rats in the Biomarker study.

Figure 2.

Short-term assessment of antitumor efficacy of ERL±SUL in the Pirc model. A, Efficacy study protocol, with n = 10 rats per group. B, Temporal tracking of antitumor response by colonoscopy (mean ± SE). C–F, Tumor outcomes determined at 16 weeks (mean ± SD). G, Immunoblotting of pErk/Erk in colon tumors. Representative findings from two or more independent experiments, assaying the same set of negative and positive controls (AIN n = 4, SUL n = 4) within each panel. The corresponding densitometry data indicate mean ± SE. In C–G, bars sharing the same superscript letter were not significantly different by one-way ANOVA.

Figure 2.

Short-term assessment of antitumor efficacy of ERL±SUL in the Pirc model. A, Efficacy study protocol, with n = 10 rats per group. B, Temporal tracking of antitumor response by colonoscopy (mean ± SE). C–F, Tumor outcomes determined at 16 weeks (mean ± SD). G, Immunoblotting of pErk/Erk in colon tumors. Representative findings from two or more independent experiments, assaying the same set of negative and positive controls (AIN n = 4, SUL n = 4) within each panel. The corresponding densitometry data indicate mean ± SE. In C–G, bars sharing the same superscript letter were not significantly different by one-way ANOVA.

Close modal

Long-term efficacy/resistance/toxicity study

As shown in Fig. 3, at 6 weeks of age, rats were maintained on AIN basal diet or switched to AIN diet containing 250 ppm SUL. Other groups included ERL10 qw, ERL21 qw, and ERL42 qw, with or without SUL. The study was terminated at 52 weeks, facilitated by occasional polypectomy (16), mainly in the AIN controls (n = 5 rats).

Figure 3.

Long-term assessment of antitumor efficacy in the Pirc model. A, Study protocol, with n = 22 rats per group. B, Temporal tracking of antitumor response by colonoscopy (mean ± SE). C–F, Tumor outcomes determined at 52 weeks (mean ± SD). G, Immunoblotting of pErk/Erk in colon tumors. Representative findings from two or more independent experiments. The corresponding densitometry data indicate mean ± SE. In C–G, bars sharing the same superscript letter were not significantly different by one-way ANOVA.

Figure 3.

Long-term assessment of antitumor efficacy in the Pirc model. A, Study protocol, with n = 22 rats per group. B, Temporal tracking of antitumor response by colonoscopy (mean ± SE). C–F, Tumor outcomes determined at 52 weeks (mean ± SD). G, Immunoblotting of pErk/Erk in colon tumors. Representative findings from two or more independent experiments. The corresponding densitometry data indicate mean ± SE. In C–G, bars sharing the same superscript letter were not significantly different by one-way ANOVA.

Close modal

At necropsy, the entire GI tract was opened, cleaned of its contents, and subjected to gross examination, including for possible ulceration (see below). Colon and SI polyps were recorded for location, multiplicity, tumor volume, and total tumor burden. Tumor dimensions were measured using Vernier calipers. Tumor volume was calculated using the formula: volume = length × width2 × 0.5, and the total tumor burden was calculated as the sum of the individual tumor volumes of each polyp, expressed as mm3. Tumors and adjacent normal-looking tissues were frozen for molecular analyses of protein and RNA expression changes, or fixed in 10% buffered formalin for histopathologic diagnosis. Other tissues were examined for signs of toxicity or gross pathology, and organ weights were recorded. Blood samples at termination were collected for biomarker assessment. For the biochemistry panel, blood samples (0.5 mL) were collected in sterile lithium heparin tubes (#13–680–62, BD Microtainer, Fisher Scientific) and analyzed the same day using standardized protocols at the Texas A&M Veterinary Medical Diagnostic Laboratory (College Station, TX).

Immunoblotting, IHC, and hematoxylin and eosin staining

Published immunoblotting methodologies in the rat (14, 15, 29, 30) used primary antibodies to anti-phospho-Erk1/2 (pErk, Sigma-Aldrich #05–797) and Erk1/2 (Erk, Cell Signaling Technology #9107). Proteins were visualized by Western Lightning Plus-ECL (Perkin Elmer, Inc.) and quantified using a ChemiDoc imaging system (BioRad). For IHC, colon polyps and adjacent normal tissue were sectioned at 4 to 5 microns and processed with primary antibodies to β-catenin (BD Biosciences #61053) and Ki-67 (Dako #M7248), with Dako Universal Negative Control (N1699) as reported (19, 31–33). Max PolyOne Polymer HRP Detection solution (Vector Laboratories) was applied at room temperature, followed by Nova Red as chromogen and Dako hematoxylin as counterstain. Epidermal tissues were sectioned at 4 to 5 microns and stained using hematoxylin and eosin (H&E).

RNA analyses

As reported (31–33), RNA was extracted using an RNeasy kit (Qiagen), reverse-transcribed via SuperScript III (Thermo Fisher), and quantified by qPCR, with target mRNAs normalized to Gapdh. Primer sequences used for qPCR were Tnf: 5′-ATGGGCTCCCTCTCATCAGT-3′ (Forward), 5′-GCTTGGTGGTTTGCTACGAC-3′ (Reverse); early growth response 1 (Egr1): 5′-CGAGGAGCAAATGATGACCG-3′ (Forward), 5′-CAGGGATCATGGGAACCTGG-3′ (Reverse); Mmp7: 5′-ATGGTGAGGACTCAGGAGTGA-3′ (Forward), 5′-TTGCTGGTGTCTGTCGTGTT-3′ (Reverse); Gapdh: 5′-ATGGGAGTTGCTGTTGAAGTC-3′ (Forward), 5′-CCGAGGGCCCACTAAAGG-3′ (Reverse). Three or more independent experiments were performed for each sample.

Ulcerogenicity

The stomach was examined microscopically for signs of toxicity, as reported (34). Ulcers were enumerated under a Nikon SMZ800N stereomicroscope. The ulcer index was scored as follows: 0 = no lesions/normal color; 0.5 = red coloration; 1 = spot lesions (<1 mm); 2 = 1–5 small ulcers (1–2 mm); 3 = >5 small (1–2 mm) or 1 large ulcer (>2 mm); 4 = >1 large ulcer (>2 mm).

Skin toxicity

One to two skin biopsies were taken per animal, mainly from the facial area surrounding the eyes. Based on prior studies, H&E-stained epidermal biopsies were examined for acanthosis, hyperkeratosis, perivascular inflammation, or other reported abnormalities (35).

Statistical analyses

Unless indicated otherwise, tumor data were presented as bar graphs with mean ± SD. Other data were presented as box-and-whisker plots or bar graphs with mean ± SE. ANOVA was used for group comparisons followed by Tukey's multiple comparisons test. In the figures and tables, significant outcomes were shown with P values (*, P < 0.05; **, P < 0.01; and ***, P < 0.001), or with superscript letters designating no significant difference among groups.

After discontinuing ERL dosing, pErk was inhibited for up to 10 days in Pirc colon polyps

Lubet and colleagues reported that pErk protein expression was inhibited in rat mammary tumors 1 week after a single bolus dose of 42 mg ERL/kg BW i.g., but they did not examine the time-course for recovery (29). In the Biomarker Study (Fig. 1A), there was effective inhibition of pErk in colon polyps between 1 and 10 days post-ERL dosing, when daily SUL in the diet (250 ppm) was administered with ERL 6 mg/kg BW i.g. daily (Fig. 1B, red boxes and associated densitometry, P < 0.05 by ANOVA). Dose-response and time-course experiments revealed that pErk inhibition by 6 or 12 mg/kg BW i.g. daily lasted for up to 10 days post-ERL dosing (Fig. 1C and D, P < 0.05 by ANOVA), whereas pErk inhibition was detected for up to 7 days post ERL i.g. dosing at 21 mg/kg twice weekly or 42 mg/kg weekly (Fig. 1E and F, P < 0.05 by ANOVA), based on densitometry measurements of the corresponding immunoblots (Supplementary Fig. S1). Although SUL is used as standard of care in patients with FAP (6–10), SUL+ERL was more effective than SUL alone at inhibiting pErk between 24 hours and 7 to 10 days after discontinuing ERL treatment (Fig. 1C–F).

Short-term ERL+SUL combinations provided good efficacy in the Pirc model of FAP

We next assessed antitumor outcomes of the following dosing regimens: (i) once-per-week i.g. oral gavage administration of ERL at 21 and 42 mg/kg BW (ERL21, ERL42); (ii) 6 mg/kg BW equivalent ERL, but given in the diet (70 ppm) either continuously or every other week; and (iii) 12 mg/kg equivalent ERL, given in the diet (140 ppm) on alternating weeks. Colonoscopy data revealed efficacy in multiple groups, especially for ERL+SUL in combination (Fig. 2B), and assisted in determining the final termination date for the study. Statistical analyses of the colonoscopy data (Supplementary Table S1) indicated that antitumor efficacy started as early as 6 weeks after initiating drug treatments, especially in the ERL+SUL combination arms, and was sustained for the entire study duration. No deleterious effects were noted on BW growth curves or food consumption in any of the groups (Supplementary Fig. S2A and S2B).

When the efficacy study was terminated after 16 weeks of drug treatment, in animals at 22 weeks of age (Fig. 2A), the number of polyps and the total tumor burden (i.e., number plus volume of all tumors combined) were determined in the colon (Fig. 2C and D) and SI (Fig. 2E and F). Key observations were as follows: (i) SUL at 250 ppm in the diet-suppressed colon and SI tumor burden significantly (50%–60% inhibition; P < 0.001) compared with AIN basal diet (yellow vs. white bars in Fig. 2D and F); (ii) ERL 70 and 140 ppm in the diet did not reduce colon polyp number significantly, but in combination with SUL it offered better protection than SUL alone (70%–85% inhibition; P < 0.001, Fig. 2C); (iii) suppression of colon and SI polyp number and total tumor burden was prominent in rats given weekly i.g. doses of ERL21 and ERL42, alone (70%–95% inhibition; P < 0.001) or in combination with SUL (80%–98% inhibition; P < 0.001; pink and purple bars, Fig. 2CF); (iv) a low-dose “window” for antitumor efficacy was defined for ERL70, given either continuously or on alternating weeks in combination with SUL (250 ppm), for which the individual test agents were less effective (Fig. 2CF, E70cS and E70aS vs. SUL, E70a, or E70c).

Based on the densitometry data for pErk/Erk, target modulation occurred in Pirc colon polyps for multiple treatment groups (Fig. 2G). Unlike in the Biomarker study (Fig. 1B), treatment with SUL alone for 16 weeks decreased pErk levels (Fig. 2G, P < 0.05 by ANOVA). Addition of SUL to ERL (ERL+SUL) led to even lower pErk levels, especially for ERL70a and ERL70c, which was in concordance with the antitumor efficacy of the combined agents at these doses (Fig. 2BF).

Long-term ERL+SUL had sustained efficacy with minimal toxicity/resistance

A 1-year study was undertaken, beginning in rats at 6 weeks of age, and with drug treatments lasting 46 weeks (Fig. 3A). Based on the short-term efficacy study, which indicated that oral gavage dosing regimens generally were superior to feeding ERL in the diet (Fig. 2CF), we selected once-per-week i.g. doses of ERL 21 and 42 mg/kg BW, and included one lower i.g. dose, i.e., once-weekly ERL 10 mg/kg BW, to seek the cutoff for efficacy. Monthly colonoscopy data identified good antitumor efficacy for once-per-week ERL10, ERL21, and ERL42 when combined with SUL (Fig. 3B and Supplementary Table S2). At the termination of the study, compared with AIN controls, ERL10 had good efficacy in the SI and colon when combined with SUL. Indeed, ERL10+SUL suppressed significantly in both target tissues with respect to the number of polyps (60%–85% inhibition; P < 0.001; Fig. 3C and E) and the total tumor burden (86%–91% inhibition; P < 0.001; Fig. 3D and F). However, ERL10+SUL was less efficacious than ERL21+SUL and ERL42+SUL with respect to the number of polyps (81%–99% inhibition; P < 0.01; Fig. 3C and E) and the total tumor burden (91%–99% inhibition; P < 0.05; Fig. 3D and F). As in the short-term efficacy study, pErk inhibition was observed in multiple treatment groups, including SUL, ERL10±SUL, ERL21±SUL, and ERL42±SUL (Fig. 3G), coinciding with antitumor activity after 52 weeks.

Cumulative data for lesions from the long-term efficacy study (Supplementary Fig. S3) revealed that the largest polyps in the colon and SI were in the range 0.5 to 1.2 cm3 and 1 to 3 cm3, respectively. The final data included lesions that were resected periodically via colonoscopy-polypectomy (16) to prevent occlusion of the colon and premature termination of the study. The AIN controls required the most resections, with one large polyp removed from each of five separate rats. The Pirc model typically is not associated with invasion (13), and although histopathologic analysis was not performed in the current study, drug treatments lowered β-catenin and Ki-67 overexpression in the colon tumors (Supplementary Fig. S4A), consistent with effective tumor interception.

No deleterious effects were noted on BW growth curves in any of the groups in the long-term study (Fig. 4A). The Kaplan–Meier plot revealed lower overall survival in the AIN group, intermediate survival for SUL, and better survival for all other groups—especially for ERL+SUL in combination (Fig. 4B). Compared with AIN controls, an increase in the ulcer index was observed for ERL21±SUL and ERL42±SUL, but not in SUL, ERL10, or ERL10+SUL groups (Fig. 4C). In contrast to AIN, SUL, ERL10, and ERL10+SUL groups, 30% to 40% of rats given ERL21±SUL or ERL42±SUL had porphyrin staining around the eye (Fig. 4D). Biopsies at the time of necropsy revealed no overt skin pathology, such as acanthosis, hyperkeratosis, perivascular inflammation, or other abnormalities (Supplementary Fig. S4B).

Figure 4.

Toxicity assessment for ERL±SUL in the Pirc model. A, BWs in the various treatment groups. B, Kaplan–Meier curves for ERL±SUL treatment groups. C, Gastric ulceration (mean ± SE) following drug treatments, with the Ulcer Index determined as described in Materials and Methods. D, Porphyrin staining around the eye, summarized as a percentage of animals in each group (mean ± SE). In C and D, bars sharing the same superscript letter were not significantly different by one-way ANOVA.

Figure 4.

Toxicity assessment for ERL±SUL in the Pirc model. A, BWs in the various treatment groups. B, Kaplan–Meier curves for ERL±SUL treatment groups. C, Gastric ulceration (mean ± SE) following drug treatments, with the Ulcer Index determined as described in Materials and Methods. D, Porphyrin staining around the eye, summarized as a percentage of animals in each group (mean ± SE). In C and D, bars sharing the same superscript letter were not significantly different by one-way ANOVA.

Close modal

Organ weights that were altered in tumor-bearing Pirc controls fed AIN basal diet were normalized by drug treatments (Supplementary Fig. S5A–S5C), except for lung (Supplementary Fig. 5D). Reduced splenomegaly and increased hematocrit with SUL±ERL (Supplementary Fig. S5E and S5F) resembled prior studies in the ApcMin/+ mouse treated with SUL and tea (36, 37). An inverse correlation was noted for spleen/BW (%) versus % hematocrit (Supplementary Fig. S6A), as well as for the number of polyps in the colon and SI versus % hematocrit (Supplementary Fig. S6B and S6C).

A biochemistry panel identified few changes in the circulation (Supplementary Fig. S7). Higher doses of ERL±SUL improved total protein levels to within the normal mean value for wild-type age-matched control F344 male rats (blue dotted line). Albumin was higher for ERL21+SUL, ERL21 produced an increase in ALT, and glucose levels were elevated by ERL42±SUL. No other statistically significant changes were noted.

A summary is provided of key clinical observations (Table 1). Diarrhea and blood in the rectum were observed mainly for AIN, SUL, and ERL10 groups harboring a greater tumor burden and increased mortality/morbidity, whereas higher doses of ERL±SUL produced an eye and/or mild skin phenotype. An apparent low-dose “window”—representing the lowest threshold dose for effective tumor suppression—was defined for ERL10+SUL, exhibiting no skin toxicity, minimal diarrhea/blood in the rectum (Table 1), and sustained antitumor activity (Fig. 3CF).

Table 1.

Clinical observations after 42 weeks of drug treatment.

Clinical observations after 42 weeks of drug treatment.
Clinical observations after 42 weeks of drug treatment.

Gene expression was altered by ERL+SUL in tumors from the Pirc model

Differential gene expression profiling in duodenal polyps from patients with FAP treated with ERL+SUL versus placebo prioritized WNT-, EGFR-, and PGE2-related pathways as mechanistically relevant molecular targets (27), with top genes in each category including MMP7, EGR1, and TNF, respectively. The corresponding murine genes Mmp7, Egr1, and Tnf were examined in Pirc colon polyps from the efficacy study at 16 weeks (Fig. 5A). Higher doses of ERL+SUL inhibited all three genes significantly compared with AIN controls. Notably, ERL70 continuous or alternating dietary treatment plus SUL in the diet (ERL70cS, ERL70aS) also inhibited Mmp7 significantly. Prior studies identified Mmp7 as a key mechanistic target in the Pirc model following treatment with nonsteroidal anti-inflammatory agents (14).

Figure 5.

Tumor-associated gene expression changes in the Pirc model following ERL±SUL treatment. A, Colon tumors from the efficacy study. (B) Colon tumors and (C) small intestine tumors from the toxicity/resistance study. Representative RT-qPCR data (mean ± SE) from two or more independent experiments, with each target normalized to Gapdh. In A, the data were obtained from n = 6–9 polyps per group, whereas in B and C, there were n = 8 biological replicates, except for ERL21, ERL21+SUL, and ERL42 groups (n = 4 tumors per group). The ERL42+SUL group was not determined (“Not done”), due to scarce tissue availability after robust tumor suppression, and prioritization of pErk/Erk assessment (Fig. 3G). Bars sharing the same superscript letter were not significantly different by one-way ANOVA.

Figure 5.

Tumor-associated gene expression changes in the Pirc model following ERL±SUL treatment. A, Colon tumors from the efficacy study. (B) Colon tumors and (C) small intestine tumors from the toxicity/resistance study. Representative RT-qPCR data (mean ± SE) from two or more independent experiments, with each target normalized to Gapdh. In A, the data were obtained from n = 6–9 polyps per group, whereas in B and C, there were n = 8 biological replicates, except for ERL21, ERL21+SUL, and ERL42 groups (n = 4 tumors per group). The ERL42+SUL group was not determined (“Not done”), due to scarce tissue availability after robust tumor suppression, and prioritization of pErk/Erk assessment (Fig. 3G). Bars sharing the same superscript letter were not significantly different by one-way ANOVA.

Close modal

Gene expression analyses were extended to drug treatment groups at 52 weeks in the long-term study, except for ERL42SUL, which had scarce tissue availability after robust tumor suppression (Fig. 3CF). In colon polyps at 52 weeks (Fig. 5B), Mmp7 expression was inhibited significantly by ERL+SUL treatment, including ERL10S, but not by ERL or SUL alone. In the SI tumors at 52 weeks (Fig. 5C), overall trends for Mmp7 paralleled the antitumor outcomes, with higher doses of ERL plus SUL being more effective at lowering Mmp7 levels.

For Tnf, trends also paralleled the antitumor outcomes, with higher doses of ERL+SUL being more effective than SUL alone. Inhibition of Tnf expression was noted in colon tumors from the short-term efficacy study (Fig. 5A), and in colon and SI polyps from the long-term study at 52 weeks (Fig. 5B and C, respectively).

Whereas Egr1 paralleled Mmp7 and Tnf in the short-term efficacy study (Fig. 5A), no significant inhibition of Egr1 was observed in colon polyps at 52 weeks (Fig. 5B), and an apparent statistically significant increase was noted for Egr1 in the SI tumors after ERL21+SUL treatment versus AIN alone (Fig. 5C). We are cautious not to overinterpret the latter findings, given that outcomes could not be corroborated at higher dose combinations of ERL+SUL, due to scarce tumor tissue availability after robust antitumor activity (as noted above).

Collectively, the findings from this report identified ERL 10 mg/kg + SUL 250 ppm (ERL10+SUL) as the lowest, most efficacious dose combination with the least toxicity in the Pirc model. Dose selection was guided by each iterative experiment, as well as by prior work that reported ERL to affect antitumor outcomes and pErk inhibition when continuous daily oral gavage administration was switched to once-a-week cumulative “loading” and “half-loading” i.g. doses of ERL (29). We also determined, for the first time, the temporal recovery of pErk in rat colon polyps, observing pErk inhibition for up to 10 days after discontinuation of ERL dosing in the Pirc model.

Our initial goal was to determine the point at which pErk started an upward trajectory, timing the next ERL dose to maintain pErk inhibition in Pirc colon and SI polyps. In terms of clinical translation, a personalized approach to “precision medicine” in patients with FAP might involve ERL given less frequently, precisely targeting pErk reappearance in adenomatous polyps, while enabling acne-like rash more time for recovery (11). However, ERL10, ERL21, and ERL42 inhibited pErk expression to a similar extent in Pirc colon polyps (Fig. 3G), and pErk levels did not faithfully reflect the degree of antitumor efficacy. Cyclooxygenase-1/cyclooxygenase-2 (Cox1/Cox2) enzymatic activities and PGE2 quantification in target tissues might clarify the role of Cox signaling in the Pirc model.

Notably, SUL alone inhibited pErk expression in adenomatous polyps (Figs. 2G and 3G), consistent with COX-independent actions (38–42) and the cross-talk between WNT and ERK signaling pathways (25, 26, 36, 37, 42). This was less evident in the Biomarker Study (Fig. 1B), probably due to the shorter duration of SUL treatment. These results recommend against the routine use of pErk as a predictive biomarker of response to SUL+ERL combination treatments in future clinical trials. Previous studies in the ApcMin/+ mouse demonstrated tumor promotion by SUL in the colon (43, 44), unlike the colon tumor suppression observed for SUL in the Pirc model (14, 45). This discrepancy between rat and mouse models most likely relates to species, strain, and gender differences. The Apc-mutant mouse develops tumors mostly in the SI, with few in the colon, whereas the Apc-mutant rat has a significant tumor burden both in the SI and in the colon, more closely mimicking FAP in humans (13). However, the degree of inhibition of gross tumors in the Pirc model may be an overestimation, due to the potential presence of microscopic tumors that can only be detected histologically (Supplementary Fig. S4A).

The short-term efficacy study set the stage for the long-term study by establishing antitumor activities for various ERL+SUL combinations in the colon (Fig. 2C and D) and SI (Fig. 2E and F). Notably, ERL10+SUL was the lowest dose combination that had antitumor efficacy at 52 weeks in the colon (Fig. 3C and D) and SI (Fig. 3E and F), with little or no signs of toxicity (Fig. 4 and Table 1). Moreover, antitumor effects in the SI were attributed, in large part, to suppression of duodenum polyps at 16 and 52 weeks (Supplementary Fig. S8). These findings could be pertinent for patients with FAP who have undergone a total colectomy, when duodenal cancer becomes the focus of prevention strategies (6, 11, 12). After allometric scaling, ERL10+SUL equates to one-quarter of the current weekly intake of ERL, combined as a single bolus dose and given once a week with standard-of-care SUL. This work has started to address the need (11, 12) for ERL+SUL dose combinations that retain efficacy in the setting of duodenal cancer prevention, while minimizing toxicity concerns.

Toxicity readouts at higher doses of ERL±SUL included increased gastric ulceration (Fig. 4C) and porphyrin staining near the eye, with no rash or other skin pathology (Supplementary Fig. S4B), which is a common side effect in patients with cancer treated with ERL (46, 47). Whereas patients given ERL can present with diarrhea (46–48), in the current investigation, diarrhea was related to a high-tumor burden rather than ERL dosing (Table 1, dashed box), in AIN and SUL groups that exhibited reduced overall survival after 52 weeks (Fig. 4B).

Other notable observations from the long-term study included a modest impact on blood biochemistry parameters (Supplementary Fig. S7), and the normalization of hematocrit and organ weights, such as liver, kidney, heart, and spleen (Supplementary Fig. S5). One cautionary note concerned the diminished lung weights at 52 weeks (Supplementary Fig. S5D). In some patients given ERL, interstitial lung disease has been reported, although this typically involves a pre-existing pulmonary condition and high-dose drug treatment (49–51). Further studies are needed to corroborate ERL-induced changes in the lung of the Pirc model.

Gene expression analyses in the rat were in general accordance with prior observations from patients with FAP given ERL+SUL, in which WNT, EGFR, and PGE2 signaling pathways were prioritized (27). Thus, Mmp7, Egr1, and Tnf levels were downregulated by the various drug treatments in Pirc colon polyps at the end of the short-term efficacy study (Fig. 5A), and trends were somewhat indicative of antitumor outcomes. Tnf remained inhibited in colon and SI polyps after 46 weeks of drug treatment (Fig. 5B and C), reflecting sustained actions on the PGE2 pathway (27), despite the fact that SUL—a drug known to target COX/PGE2—was kept constant and Tnf changes paralleled the ERL dose applied. In the colon and SI polyps obtained after 1 year, ERL+SUL remained effective at downregulating Mmp7, but ERL as single agent did not attenuate Mmp7 levels in colon polyps (Fig. 5B). One interpretation is that SUL rather than ERL is the primary driver of Wnt-associated gene changes, perhaps via downregulation of β-catenin or β-catenin/Tcf-dependent transcription (8, 36, 37, 42). Although Egr1 was unexpectedly increased by ERL21+SUL treatment in SI polyps at 52 weeks (Fig. 5C), we are cautious not to overinterpret these data. Additional work is needed to verify these observations and any possible drug resistance mechanisms arising via the Egfr pathway. Exome sequencing and transcriptomic analyses of large versus small lesions (Supplementary Fig. S3) might help to prioritize genetic changes and critical signaling pathways related to efficacy versus resistance (8, 14, 27, 36–42, 52, 53).

In summary, we defined an optimized low-dose strategy in the Pirc model, involving i.g. ERL at 10 mg/kg BW once-per-week plus 250 ppm dietary SUL. ERL10+SUL produced minimal toxicity (Table 1), normalized hematocrit and splenomegaly (Supplementary Fig. S5), inhibited tumor-associated pErk (Fig. 3G), Mmp7, and Tnf (Fig. 5B and C), and had significant antitumor activity in the colon (Fig. 3C and D) and SI (Fig. 3E and F). Clinical translation to patients with FAP would entail ERL administration at 125 mg once a week, which is one-quarter of the current recommended therapeutic dose, as a once-per-week regimen with standard-of-care SUL. Expected outcomes would include efficacy against adenomatous polyps in the colon, and in particular the SI, while minimizing the likelihood for skin toxicity, gastric ulceration, or other deleterious outcomes.

P. Rajendran reports grants from NCI/NIH during the conduct of the study. W.M. Dashwood reports grants from NCI/NIH during the conduct of the study. O.F. Yavuz reports grants from NCI/NIH during the conduct of the study. S. Kapoor reports grants from NCI/NIH during the conduct of the study. E. Vilar reports grants from NIH/NCI during the conduct of the study; grants and personal fees from Janssen Research and Development; and personal fees from Recursion Pharma outside the submitted work. R.H. Dashwood reports grants from NCI/NIH during the conduct of the study. No disclosures were reported by the other authors.

The data that support the findings of this study are available from the corresponding author upon reasonable request.

A.M. Ulusan: Investigation, methodology, writing–original draft, writing–review and editing. P. Rajendran: Conceptualization, data curation, formal analysis, supervision, funding acquisition, investigation, methodology, project administration, writing–review and editing. W.M. Dashwood: Conceptualization, formal analysis, investigation, methodology, writing–review and editing. O.F. Yavuz: Investigation, methodology, writing–review and editing. S. Kapoor: Formal analysis, validation, methodology, writing–review and editing. T.A. Gustafson: Investigation, methodology, writing–review and editing. M.I. Savage: Project administration, writing–review and editing. P.H. Brown: Resources, funding acquisition, project administration, writing–review and editing. S. Sei: Conceptualization, formal analysis, supervision, writing–review and editing. A. Mohammed: Conceptualization, investigation, methodology, writing–review and editing. E. Vilar: Conceptualization, formal analysis, supervision, writing–review and editing. R.H. Dashwood: Conceptualization, resources, formal analysis, supervision, funding acquisition, investigation, writing–original draft, writing–review and editing.

This research was supported by Contract No. HHSN261201500018I, Task Order HHSN26100004, from the NCI PREVENT Program. R.H. Dashwood was also supported by grants CA090890 and CA122959 from the NCI, by the John S. Dunn Foundation, by AgriLife Research, and by a Chancellor's Research Initiative from Texas A&M University. Funding was also provided by grant P30 CA016672 (US NIH/NCI) to the University of Texas Anderson Cancer Center Core Support Grant and a gift from the Feinberg Family to E. Vilar.

We thank Gavin S. Johnson and Ying-Shiuan Chen for technical assistance during necropsies, and Yunus E. Demirhan for help with qPCR primer information and sample acquisition.

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.

1.
Torre
LA
,
Siegel
R
,
Ward
EM
,
Jemal
A
. 
Global cancer incidence and mortality rates and trends – an update
.
Cancer Epidemiol Biomarkers Prev
2016
;
25
:
16
27
.
2.
Mork
ME
,
You
YN
,
Ying
J
,
Bannon
SA
,
Lynch
PM
,
Rodriguez-Bigas
MA
, et al
High prevalence of hereditary cancer syndromes in adolescents and young adults with colorectal cancer
.
J Clin Oncol
2015
;
33
:
3544
9
.
3.
Borras
E
,
San Lucas
FA
,
Chang
K
,
Zhou
R
,
Masand
G
,
Fowler
J
, et al
Genomic landscape of colorectal mucosa and adenomas
.
Cancer Prev Res
2016
;
9
:
417
27
.
4.
Fu
Y
,
Zheng
S
,
An
N
,
Athanasopoulos
T
,
Popplewell
L
,
Liang
A
, et al
Beta-catenin as a potential key target for tumor suppression
.
Int J Cancer
2011
;
129
:
1541
51
.
5.
Morin
PJ
,
Sparks
AB
,
Korinek
V
,
Barker
N
,
Clevers
H
,
Vogelstein
B
, et al
Activation of beta-catenin-Tcf signaling in colon cancer by mutations in beta-catenin or APC
.
Science
1997
;
275
:
1787
90
.
6.
Lynch
PM
. 
Chemoprevention of familial adenomatous polyposis
.
Fam Cancer
2016
;
15
:
467
75
.
7.
Weng
W
,
Feng
J
,
Qin
H
,
Ma
Y
. 
Molecular therapy of colorectal cancer: progress and future directions
.
Int J Cancer
2015
;
136
:
493
502
.
8.
Li
N
,
Xi
Y
,
Tinsley
HN
,
Gurpinar
E
,
Gary
BD
,
Zhu
B
, et al
Sulindac selectively inhibits colon tumor cell growth by activating the cGMP/PKG pathway to suppress Wnt/beta-catenin signaling
.
Mol Cancer Ther
2013
;
12
:
1848
59
.
9.
Cruz-Correa
M
,
Hylind
LM
,
Romans
KE
,
Booker
SV
,
Giardiello
FM
. 
Long-term treatment with sulindac in familial adenomatous polyposis: a prospective cohort study
.
Gastroenterology
2002
;
122
:
641
5
.
10.
Brosens
LA
,
Keller
JJ
,
Offerhaus
GJ
,
Goggins
M
,
Giardiello
FM
. 
Prevention and management of duodenal polyps in familial adenomatous polyposis
.
Gut
2005
;
54
:
1034
43
.
11.
Samadder
NJ
,
Neklason
DW
,
Boucher
KM
,
Byrne
KR
,
Kanth
P
,
Samowitz
W
, et al
Effect of sulindac and erlotinib vs placebo on duodenal neoplasia in familial adenomatous polyposis: a randomized clinical trial
.
JAMA
2016
;
315
:
1266
75
.
12.
Samadder
NJ
,
Kuwada
SK
,
Boucher
KM
,
Byrne
K
,
Kanth
P
,
Samowitz
W
, et al
Association of sulindac and erlotinib vs. placebo with colorectal neoplasia in familial adenomatous polyposis: secondary analysis of a randomized clinical trial
.
JAMA Oncol
2018
;
4
:
671
7
.
13.
Amos-Landgraf
JM
,
Kwong
LN
,
Kendziorski
CM
,
Reichelderfer
M
,
Torrealba
J
,
Weichert
J
, et al
A target-selected Apc-mutant rat kindred enhances the modeling of familial human colon cancer
.
Proc Natl Acad Sci U S A
2007
;
104
:
4036
41
.
14.
Ertem
FU
,
Zhang
W
,
Chang
K
,
Dashwood
W
,
Rajendran
P
,
Sun
D
, et al
Oncogenic targets Mmp7, S100a9, Nppb and Aldh1a3 from transcriptome profiling of FAP and Pirc adenomas are downregulated in response to tumor suppression by Clotam
.
Int J Cancer
2017
;
140
:
460
8
.
15.
Rajendran
P
,
Johnson
G
,
Li
L
,
Chen
Y-S
,
Dashwood
M
,
Nguyen
N
, et al
Acetylation of CCAR2 establishes a BET/BRD9 acetyl switch in response to combined deacetylase and bromodomain inhibition
.
Cancer Res
2019
;
79
:
918
27
.
16.
Ertem
F
,
Dashwood
WM
,
Rajendran
P
,
Raju
G
,
Rashid
A
,
Dashwood
R
. 
Development of a murine colonoscopy polypectomy model (with videos)
.
Gastrointest Endosc
2016
;
83
:
1272
6
.
17.
Yun
C
,
Dashwood
WM
,
Li
L
,
Yin
T
,
Ulusan
AM
,
Shatzer
K
, et al
Acute changes in colonic -PGE2 levels as a biomarker of efficacy after treatment of the Pirc rat with celecoxib
.
Inflamm Res
2020
;
69
:
131
7
.
18.
Luceri
C
,
Femia
AP
,
Tortora
K
,
D'Ambrioso
M
,
Fabbri
S
,
Fazi
M
, et al
Supplementation with phytoestrogens and insoluble fibers reduces intestinal carcinogenesis and restores ER-beta expression in Apc-driven colorectal carcinogenesis
.
Eur J Cancer Prev
2020
;
29
:
27
35
.
19.
Okonkwo
A
,
Mitra
J
,
Johnson
GS
,
Li
L
,
Dashwood
WM
,
Hegde
ML
, et al
Heterocyclic analogs of sulforaphane trigger DNA damage and impede DNA repair in colon cancer cells: interplay of HATs and HDACs
.
Mol Nutr Food Res
2018
;
62
:
e1800228
.
20.
Irving
AA
,
Duchow
EG
,
Plum
LA
,
DeLuca
HF
. 
Vitamin D deficiency in the ApcPirc/+ rat does not exacerbate colonic tumorigenesis, while low dietary calcium might be protective
.
Dis Model Mech
2018
;
11
:
dmm032300
.
21.
Jones
JE
,
Busi
SB
,
Mitchem
JB
,
Amos-Landgraf
JM
,
Lewis
MR
. 
Evaluation of a tumor-targeting, near-infrared fluorescent peptide for early detection and endoscopic resection of polyps in a rat model of colorectal cancer
.
Mol Imaging
2018
;
17
:
1536012118790065
.
22.
Buchanan
FG
,
Holla
V
,
Katkuri
S
,
Matta
P
,
DuBois
RN
. 
Targeting cyclooxygenase-2 and epidermal growth factor receptor for the prevention and treatment of intestinal cancer
.
Cancer Res
2007
;
67
:
9380
8
.
23.
Pagán
B
,
Isidro
AA
,
Cruz
ML
,
Ren
Y
,
Coppola
D
,
Wu
J
, et al
Erlotinib inhibits progression to dysplasia in a colitis-associated colon cancer model
.
World J Gastroenterol
2011
;
17
:
4858
66
.
24.
Buck
E
,
Eyzaguirre
A
,
Brown
E
,
Petti
F
,
McCormack
S
,
Haley
JD
, et al
Rapamycin synergizes with epidermal growth factor receptor inhibitor erlotinib in non-small cell lung cancer, pancreatic, colon and breast tumors
.
Mol Cancer Ther
2006
;
5
:
2676
84
.
25.
Zeller
E
,
Hammer
K
,
Kirschnick
M
,
Braeuning
A
. 
Mechanisms of RAS/β-catenin interactions
.
Arch Toxicol
2013
;
87
:
611
32
.
26.
Jeong
WJ
,
Ro
EJ
,
Choi
KY
. 
Interaction between Wnt/β-catenin and RAS-ERK pathways and an anti-cancer strategy via degradations of β-catenin and RAS by targeting the Wnt/β-catenin pathway
.
NPJ Precis Oncol
2018
;
2
:
5
.
27.
Delker
DA
,
Wood
AC
,
Snow
AK
,
Samadder
NJ
,
Samowitz
WS
,
Affolter
KE
, et al
Chemoprevention with cyclooxygenase and epidermal growth factor receptor inhibitors in familial adenomatous polyposis patients: mRNA signatures of duodenal neoplasia
.
Cancer Prev Res
2018
;
11
:
4
15
.
28.
Devasari
N
,
Dora
CP
,
Singh
C
,
Paidi
SR
,
Kumar
V
,
Sobhia
ME
, et al
Inclusion complex of erlotinib with sulfobutyl ether-beta-cyclodextrin: preparation, characterization, in silico, in vitro and in vivo evaluation
.
Carbohydr Polym
2015
;
134
:
547
56
.
29.
Lubet
RA
,
Szabo
E
,
Iwata
KK
,
Gill
SC
,
Tucker
C
,
Bode
A
, et al
Effect of intermittent dosing regimens of erlotinib on methylnitrosourea-induced mammary carcinogenesis
.
Cancer Prev Res
2013
;
6
:
448
54
.
30.
Rajendran
P
,
Dashwood
WM
,
Kang
Y
,
Kim
E
,
Johnson
G
, et al
Nrf2 status affects tumor growth, HDAC3 gene promoter associations, and the response to sulforaphane in the colon
.
Clin Epigenetics
2015
;
7
:
102
.
31.
Wang
R
,
Dashwood
WM
,
Nian
H
,
Löhr
CV
,
Fischer
KA
,
Tsuchiya
N
, et al
NADPH oxidase overexpression in human colon cancers and rat colon tumors induced by 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP)
.
Int J Cancer
2010
;
128
:
2581
90
.
32.
Wang
R
,
Kang
Y
,
Löhr
CV
,
Fischer
KA
,
Bradford
CS
,
Johnson
G
, et al
Reciprocal regulation of BMF and BIRC5 (Survivin) linked to eomes overexpression in colorectal cancer
.
Cancer Lett
2016
;
381
:
341
8
.
33.
Wang
R
,
Löhr
CV
,
Fischer
KA
,
Dashwood
WM
,
Greenwood
JA
,
Ho
E
. 
Epigenetic inactivation of endothelin-2 and endothelin-3 in colon cancer
.
Int J Cancer
2013
;
132
:
1004
12
.
34.
Tammara
VK
,
Narurkar
MM
,
Crider
AM
,
Khan
MA
. 
Morpholinoalkyl ester prodrugs of diclofenac: synthesis, in vitro and in vivo evaluation
.
J Pharm Sci
1994
;
83
:
644
8
.
35.
Brown
AP
,
Dunstan
RW
,
Courtney
CL
,
Criswell
KA
,
Graziano
MJ
. 
Cutaneous lesions in rats following administration of an irreversible inhibitor of erbB receptors, including the epidermal growth factor receptor
.
Toxicol Pathol
2008
;
36
:
410
9
.
36.
Orner
GA
,
Dashwood
WM
,
Blum
CA
,
Diaz
GD
,
Li
Q
,
Dashwood
RH
. 
Suppression of tumorigenesis in the Apc(min) mouse: down-regulation of beta-catenin signaling by a combination of tea plus sulindac
.
Carcinogenesis
2003
;
24
:
263
7
.
37.
Orner
GA
,
Dashwood
WM
,
Blum
CA
,
Diaz
GD
,
Li
Q
,
Al-Fageeh
M
, et al
Response of Apc(min) and A33(Delta N beta-cat) mutant mice to treatment with tea, sulindac, and 2-amino-1-methylimidazo[4,5-b]pyridine (PhIP)
.
Mutat Res
2002
;
506–507
:
121
7
.
38.
Liggett
JL
,
Min
KW
,
Smolensky
D
,
Baek
SJ
. 
A novel COX-independent mechanism of sulindac sulfide involves cleavage of epithelial cell adhesion molecule protein
.
Exp Cell Res
2014
;
326
:
1
9
.
39.
Gurpinar
E
,
Grizzle
WE
,
Piazza
GA
. 
COX-independent mechanisms of cancer chemoprevention by anti-inflammatory drugs
.
Front Oncol
2013
;
3
:
181
.
40.
Dell'Omo
G
,
Crescenti
D
,
Vantaggiato
C
,
Parravinci
C
,
Borroni
AP
,
Rizzi
N
, et al
Inhibition of SIRT1 deacetylase and p53 activation uncouples the anti-inflammatory and chemopreventive actions of NSAIDs
.
Br J Cancer
2019
;
120
:
537
46
.
41.
Babbar
N
,
Ignatenko
NA
,
Casero
RA
,
Gerner
EW
. 
Cyclooxygenase-independent induction of apoptosis by sulindac sulfone is mediated by polyamines in colon cancer
.
J Biol Chem
2003
;
278
:
47762
75
.
42.
Rice
PL
,
Kelloff
J
,
Sullivan
H
,
Driggers
LJ
,
Beard
KS
,
Kuwada
S
, et al
Sulindac metabolites induce caspase- and proteasome-dependent degradation of beta-catenin protein in human colon cancer cells
.
Mol Cancer Ther
2003
;
2
:
885
92
.
43.
Greenspan
EJ
,
Nichols
FC
,
Rosenberg
DW
. 
Molecular alterations associated with sulindac-resistant colon tumors in ApcMin/+ mice
.
Cancer Prev Res
2010
;
3
:
1187
97
.
44.
Yang
K
,
Fan
K
,
Kurihara
N
,
Shinozaki
H
,
Rigas
B
,
Augenlicht
L
, et al
Regional response leading to tumorigenesis after sulindac in small and large intestine of mice with Apc mutations
.
Carcinogenesis
2003
;
24
:
605
11
.
45.
Femia
AP
,
Soares
PV
,
Luceri
C
,
Lodovici
M
,
Giannini
A
,
Caderni
G
. 
Sulindac, 3,3′-diindolylmethane and curcumin reduce carcinogenesis in the Pirc rat, an Apc-driven model of colon carcinogenesis
.
BMC Cancer
2015
;
15
:
611
.
46.
Townsley
CA
,
Major
P
,
Siu
LL
,
Dancey
J
,
Chen
E
,
Pond
GR
, et al
Phase II study of erlotinib (OSI-774) in patients with metastatic colorectal cancer
.
Br J Cancer
2006
;
94
:
1136
43
.
47.
Wacker
B
,
Nagrani
T
,
Weinberg
J
,
Witt
K
,
Clark
G
,
Cagnoni
PJ
. 
Correlation between development of rash and efficacy in patients treated with the epidermal growth factor receptor tyrosine kinase inhibitor in two large phase III studies
.
Clin Cancer Res
2007
;
13
:
3913
21
.
48.
Yang
Z
,
Hackshaw
A
,
Feng
Q
,
Fu
X
,
Zhang
Y
,
Mao
C
, et al
Comparison of gefitinib, erlotinib and afatinib in non-small cell lung cancer: a meta-analysis
.
Int J Cancer
2017
;
140
:
2805
19
.
49.
Kashiwabara
K
,
Semba
H
,
Fujii
S
,
Tsumura
S
. 
Outcome in advance non-small cell lung cancer patients with successful rechallenge after recovery from epidermal growth factor receptor tyrosine kinase inhibitor-induced interstitial lung disease
.
Cancer Chemother Pharmacol
2017
;
79
:
705
10
.
50.
Qi
WX
,
Sun
YJ
,
Shen
Z
,
Yao
Y
. 
Risk of interstitial lung disease associated with EGFR-TKIs in advanced non-small cell lung cancer: a meta-analysis of 24 phase III clinical trials
.
J Chemother
2015
;
27
:
40
51
.
51.
Fujita
K
,
Hirose
T
,
Kusumoto
S
,
Sugiyama
T
,
Shirai
T
,
Nakashima
M
, et al
High exposure to erlotinib and severe drug-induced interstitial lung disease in patients with non-small cell lung cancer
.
Lung Cancer
2014
;
86
:
113
4
.
52.
Larsen
CA
,
Dashwood
RH
. 
(-)-Epigallocatechin-3-gallate inhibits Met signaling, proliferation, and invasiveness in human colon cancer cells
.
Arch Biochem Biophys
2010
;
510
:
52
7
.
53.
Blum
CA
,
Xu
M
,
Orner
GA
,
Fong
AT
,
Bailey
GS
,
Stoner
GD
, et al
beta-Catenin mutation in rat colon tumors initiated by 1,2-dimethylhydrazine and 2-amino-3-methylimidazo[4,5-f]quinoline, and the effect of post-initiation treatment with chlorophyllin and indole-3-carbinol
.
Carcinogenesis
2001
;
22
:
309
14
.

Supplementary data