Abstract
A retrospective case–controlled analysis was performed to identify drug candidates in the current use that may prevent colorectal cancer, outside of aspirin. A total of 37,510 patients aged ≥20 years were assessed to identify subjects who had been diagnosed with colorectal cancer by colonoscopy without a previous diagnosis of colorectal cancer, inflammatory bowel disease (IBD), or gastrointestinal symptoms; 1,560 patients were identified who were diagnosed with colorectal cancer by colonoscopy. The patients with colorectal cancer were matched with 1,560 age, gender, family history of colorectal cancer and comorbidity-matched control patients who were not diagnosed with colorectal cancer at colonoscopy. The medication histories were compared between the two groups. Next, candidate drugs that were more frequently used by the control patients were selected and their effects on human colorectal cancer cell lines in vitro and an inflammation-induced mouse model of colorectal cancer were tested. Putative colorectal cancer preventative agents were identified, including aspirin, vitamin D, vitamin B, vitamin C, vitamin E, xanthine oxidase inhibitor, alpha-blockers, angiotensin receptor blocker, nateglinide, probiotics, thienopyridine, folic acid, nitrovasodilators, bisphosphonates, calcium channel blockers, steroids, and statins (P < 0.05). Alpha-blockers and xanthine oxidase inhibitors were selected for further study because these agents have not been analyzed previously as factors that may affect colorectal cancer outcomes. In vitro doxazosin (alpha-blocker), but not febuxostat (xanthine oxidase inhibitor), suppressed the proliferation of human colorectal cancer cells. Doxazosin also decreased tumorigenesis in an AOM/DSS mouse colorectal cancer model. Alpha-blockers may prevent colorectal cancer.
Introduction
Drug repositioning involves using existing drugs for new indications. This strategy has attracted attention in recent years (1) as it is much more economical than developing new drugs. Colorectal cancer is the third most prevalent cancer worldwide; 1.36 × 106 people are affected with it and 0.69 × 106 people were expected to die from colorectal cancer in 2012 (2). As such, this disease places a huge social and economic burden on our communities, highlighting the significant benefit of new, low-cost chemopreventative agents for this disease. The colorectal cancer preventative properties of commonly used, existing drugs, aspirin, have been shown (3, 4).
Colorectal cancer is associated with chronic diseases such as obesity and diabetes (5–8). Patients with these conditions are continuously treated with various drugs, some of which are candidates for drug repositioning for colorectal cancer. As occurred with aspirin, a prospective intervention trial for any new agents to assess colorectal cancer outcomes will first require analysis of retrospectively collected data from existing cohorts (9, 10). Patients with colorectal cancer often suffer from multiple chronic diseases, and it is not unusual for them to take more than one drug. Therefore, it can be difficult to conduct retrospective analyses on such a diverse group of patients. Nevertheless, if other existing drugs are identified with colorectal cancer preventive effects similar to aspirin, it could be useful. Using drugs that are already approved for use is far more economical than developing new drugs.
To discover novel drug repositioning candidates for preventing colorectal cancer, we conducted a retrospective multi-center propensity-matched study. On the basis of our findings, we selected an alpha-adrenoceptor blocker and a xanthine oxidase inhibitor as drug repositioning candidates, with aspirin serving as a positive control and beta-blockers as a negative control. We then examined the effect of these agents on human colorectal cancer cells in vitro and an inflammation-associated mouse model of colorectal cancer.
Materials and Methods
Data source
We previously developed a Colorectal Cancer Endoscopy (CCE) Database at Tokyo University. This was a retrospectively recorded database of patients who underwent colonoscopy at one referral hospital and four territorial hospitals in Japan; data were compiled from 57,113 patients between 2001 and 2016 (11). The database includes the following information: patient characteristics, indications for colonoscopy, colonoscopy findings, and colorectal cancer data including the site of cancer, the Union for International Cancer Control cancer stage, cancer therapy, and drug history.
Study design, setting, and participants
We performed a retrospective case–control analysis of Japanese adults using the CCE database. Data were extracted from this database from 37,555 patients aged ≥20 years, including those without a previous diagnosis of colorectal cancer, inflammatory bowel disease (IBD), or other gastrointestinal symptoms. Next, 1,560 patients who were diagnosed with colorectal cancer from a colonoscopy and a control group of 1,560 patients not diagnosed with colorectal cancer were selected (Fig. 1). Colorectal cancer was diagnosed at colonoscopy and confirmed by pathology. The human investigations were performed after approval by Institutional Review Boards of all participating institutions (the University of Tokyo, Tokyo, Japan; the Institute for Adult Diseases, Asahi Life Foundation, Tokyo, Japan; Japanese Red Cross Medical Center, Tokyo, Japan; JR Tokyo General Hospital, Tokyo, Japan; and Yaizu City Hospital, Shizuoka, Japan), according with the policy of the Japanese Ministry of Health, Labor and Welfare. Written informed consent from patients was obtained from each subject in accordance with Declaration of Helsinki.
Outcomes and variables
The primary outcome measure was the risk of colorectal cancer incidence associated with drug exposure, as determined by ORs. Colorectal cancer cases were categorized according to the location of the tumor (all colorectal, right-sided colon, and left-sided colon/rectal).
Use of the following 40 classes of drugs was assessed: antihypertensives (angiotensin-converting enzyme inhibitor, alpha-blocker, beta-blocker, angiotensin II receptor blocker, and calcium channel blocker), antidiabetics (biguanide, pioglitazone, insulin, sulfonylurea, alpha-glucosidase inhibitor, dipeptidyl peptidase-4 [DPP4] inhibitor, and nateglinide), diuretics (loop diuretic, benzothiazide diuretic, and thiazide diuretic), xanthine oxidase inhibitor (allopurinol and febuxostat) fibrates, statin, low-dose aspirin, selective COX-2 inhibitors (celecoxib), non-steroidal anti-inflammatory drugs (NSAID; loxoprofen, diclofenac sodium, and others), acetaminophen, thienopyridines (ticlopidine and clopidogrel), cilostazol, non-aspirin antiplatelet drugs (dipyridamole and eicosapentaenoic acid), anticoagulants [warfarin or non-vitamin K antagonist oral anticoagulants (NOAC) including dabigatran, rivaroxaban, and edoxaban], steroids, vitamins (vitamins B, C, D, E, and K), calcium, folic acid, probiotics, thiamazole, and potassium sodium hydrogen citrate. Use of a medication was defined as oral administration starting at least 4 weeks before colonoscopy. Use of NSAIDs included intermittent use within 4 weeks of colonoscopy.
Confounders including age, sex, family history of colorectal cancer, and other comorbidities were evaluated. Age was categorized into quintiles. Comorbidities were evaluated using the Charlson Comorbidity Index (12).
Statistical analysis
Propensity scores were estimated using a logistic regression model for colorectal cancer cases as a function of patient demographics. Eighteen factors were included as potentially clinically significant variables: age, sex, ischemic heart disease, chronic heart failure, peripheral vascular disease, cerebrovascular disease, dementia, chronic obstructive pulmonary disease, collagen disease, peptic ulcer, diabetes mellitus, chronic kidney disease, paresthesia, leukemia, malignant lymphoma, liver cirrhosis, acquired immunodeficiency syndrome (AIDS), and a family history of colorectal cancer. One-to-one matching analyses were also performed between the two groups (colorectal cancer and non-colorectal cancer) using the nearest neighbor method within a caliper width of 0.2 of the SD of the logit of the propensity score. After propensity matching, differences in the prevalence of drug exposure were compared between the two groups’ ORs as an estimate of whether colorectal cancer incidence was associated with drug exposure. We did not perform double adjustment after propensity score matching because covariate balance is already achieved in matched population.
Statistical analyses were performed using SAS software version 9.4 (SAS Institute). P < 0.05 was considered statistically significant.
Mouse tumor assays
Both experimental and control groups consisted of 6 (3 males, 3 females) C57BL6/J (CLEA, Japan) littermate mice. Primary outcome was set to tumor incidence per mouse. Then, the sample power (6 vs. 6) size was calculated by Power and Sample Size Calculation software provided from University of Vanderbilt (Nashville, TN) and approved by animal ethics committee of Asahi Life. Mice were injected intraperitoneally with 12.5 mg/kg AOM (Sigma-Aldrich) on day 1. After 5 days, mice received water supplemented with 2.5% DSS (MP Biomedicals) for 5 days, after which the mice were maintained on regular water for 14 days and then subjected to two identical additional DSS treatment cycles (13). On days 1–64, the experimental mice were injected intraperitoneally with doxazosin (Tokyo-Kasei; 5 mg/kg) 5 days a week. The control mice were injected with PBS. On day 64, the mice were sacrificed and colon tumors were analyzed. Macroscopic colon tumors were counted and the longest diameter of each tumor was measured using a digital caliper in a blinded fashion. The animal room was quarantined by airflow and maintained at a constant temperature and humidity under conventional conditions. All animal experiments were approved by the ethics committee of the Institute for Adult Diseases, Asahi Life Foundation, (Tokyo, Japan) and were performed according to the guidelines for the care and use of laboratory animals of the Institute for Adult Diseases, Asahi Life Foundation (Tokyo, Japan).
Cell culture and growth assays
HCT116 (RIKEN BRC, RCB2979) and RKO (ATCC, CRL2577) human colorectal cancer cells were cultured in McCoy's 5A and DMEM, respectively, containing 10% FBS. Additions to the culture medium included 0.6% DMSO (vehicle), doxazosin (1, 5, 15, and 30 μmol/L), febuxostat (1 and 10 μmol/L), aspirin (0.31 μmol/L, 0.62 μmol/L, 1.25 mmol/L, 2.5 mmol/L, and 5 mmol/L), and bisoprolol (12.5, 25, 50, and 100 μmol/L). Cell growth was measured using Cell Counting Kit-8 (CCK-8) from Dojindo Laboratories. Cells (1.5 × 104/mL) were seeded into 48-well plates and exposed to vehicle, febuxostat, doxazosin, aspirin, or bisoprolol. CCK-8 solution was added to each well at 24, 48, and 72 hours and the absorbance was read at 450 nm using a plate reader (SpectraMax, Molecular Devices). Mycoplasma negative of both cells was confirmed by MycoAlert Kit (Lonza) within a month from experiments.
Results
Case–Controlled analysis of patients with colorectal cancer
Between 2001 and 2016, 57,113 patients were enrolled in our database at five centers after receiving colonoscopy without any abdominal symptoms. A total of 18,043 patients aged <20 years or with a history of colorectal cancer or IBD were excluded. Of the patients in the database, 1,560 were diagnosed with colorectal cancer and 37,510 were not at the initial colonoscopy (Fig. 1). The baseline data for all patients and data after 1:1 power density matching are presented in Table 1.
Baseline characteristics
. | All patients . | Propensity score–matched patients . | ||||
---|---|---|---|---|---|---|
Variables . | Non–colorectal cancer (n = 37,510) . | Colorectal cancer (n = 1,560) . | P . | Non–colorectal cancer (n = 1,560) . | Colorectal cancer (n = 1,560) . | P . |
Age, years | <0.001 | 0.999 | ||||
<50 | 7,743 (20.64) | 278 (5.45) | 83 (5.32) | 85 (5.45) | ||
50–59 | 8,405 (22.41) | 542 (16.54) | 253 (16.22) | 258 (16.54) | ||
60–69 | 10,427 (27.80) | 942 (29.29) | 462 (29.62) | 457 (29.29) | ||
70–79 | 8,270 (22.05) | 936 (31.86) | 498 (31.92) | 497 (31.86) | ||
≥80 | 2,665 (7.10) | 490 (16.86) | 264 (16.92) | 263 (16.86) | ||
Male | 24,383 (65.00) | 1,003 (64.29) | 0.565 | 1,008 (64.62) | 1,003 (64.29) | 0.852 |
Comorbidities | ||||||
Ischemic heart disease | 4,838 (12.90) | 347 (22.24) | <0.001 | 341 (21.86) | 347 (22.24) | 0.796 |
Congestive heart failure | 3,540 (9.44) | 280 (17.95) | <0.001 | 280 (17.95) | 280 (17.95) | 1.000 |
Peripheral vascular disease | 1,870 (4.79) | 112 (7.18) | <0.001 | 117 (3.75) | 112 (3.59) | 0.728 |
Cerebrovascular disease | 2,288 (6.10) | 189 (12.12) | <0.001 | 192 (12.31) | 189 (12.12) | 0.870 |
Dementia | 402 (1.07) | 48 (3.08) | <0.001 | 40 (2.56) | 48 (3.08) | 0.387 |
COPD | 652 (1.74) | 51 (3.27) | <0.001 | 50 (3.21) | 51 (3.27) | 0.919 |
Collagen disease | 1,449 (3.86) | 64 (4.10) | 0.631 | 55 (3.53) | 64 (4.10) | 0.400 |
Peptic ulcer disease | 11,419 (30.44) | 710 (45.51) | <0.001 | 714 (45.77) | 710 (45.51) | 0.886 |
Diabetes | 9,484 (25.28) | 601 (38.53) | <0.001 | 612 (39.23) | 601 (38.53) | 0.686 |
Chronic kidney disease | 654 (1.74) | 38 (2.44) | 0.042 | 44 (2.82) | 38 (2.44) | 0.502 |
Hemiplegia | 745 (1.99) | 80 (5.13) | <0.001 | 75 (4.81) | 80 (5.13) | 0.680 |
Leukemia | 119 (0.32) | 3 (0.19) | 0.386 | 4 (0.26) | 3 (0.19) | 0.705 |
Malignant lymphoma | 567 (1.51) | 24 (1.54) | 0.932 | 25 (1.60) | 24 (1.54) | 0.886 |
Liver cirrhosis | 762 (2.03) | 35 (2.24) | 0.561 | 28 (1.79) | 35 (2.24) | 0.373 |
AIDS | 155 (0.41) | 27 (1.73) | <0.001 | 27 (1.73) | 27 (1.73) | 0.438 |
Family history of colorectal cancer | 230 (0.61) | 22 (1.41) | <0.001 | 23 (1.47) | 22 (1.41) | 0.881 |
. | All patients . | Propensity score–matched patients . | ||||
---|---|---|---|---|---|---|
Variables . | Non–colorectal cancer (n = 37,510) . | Colorectal cancer (n = 1,560) . | P . | Non–colorectal cancer (n = 1,560) . | Colorectal cancer (n = 1,560) . | P . |
Age, years | <0.001 | 0.999 | ||||
<50 | 7,743 (20.64) | 278 (5.45) | 83 (5.32) | 85 (5.45) | ||
50–59 | 8,405 (22.41) | 542 (16.54) | 253 (16.22) | 258 (16.54) | ||
60–69 | 10,427 (27.80) | 942 (29.29) | 462 (29.62) | 457 (29.29) | ||
70–79 | 8,270 (22.05) | 936 (31.86) | 498 (31.92) | 497 (31.86) | ||
≥80 | 2,665 (7.10) | 490 (16.86) | 264 (16.92) | 263 (16.86) | ||
Male | 24,383 (65.00) | 1,003 (64.29) | 0.565 | 1,008 (64.62) | 1,003 (64.29) | 0.852 |
Comorbidities | ||||||
Ischemic heart disease | 4,838 (12.90) | 347 (22.24) | <0.001 | 341 (21.86) | 347 (22.24) | 0.796 |
Congestive heart failure | 3,540 (9.44) | 280 (17.95) | <0.001 | 280 (17.95) | 280 (17.95) | 1.000 |
Peripheral vascular disease | 1,870 (4.79) | 112 (7.18) | <0.001 | 117 (3.75) | 112 (3.59) | 0.728 |
Cerebrovascular disease | 2,288 (6.10) | 189 (12.12) | <0.001 | 192 (12.31) | 189 (12.12) | 0.870 |
Dementia | 402 (1.07) | 48 (3.08) | <0.001 | 40 (2.56) | 48 (3.08) | 0.387 |
COPD | 652 (1.74) | 51 (3.27) | <0.001 | 50 (3.21) | 51 (3.27) | 0.919 |
Collagen disease | 1,449 (3.86) | 64 (4.10) | 0.631 | 55 (3.53) | 64 (4.10) | 0.400 |
Peptic ulcer disease | 11,419 (30.44) | 710 (45.51) | <0.001 | 714 (45.77) | 710 (45.51) | 0.886 |
Diabetes | 9,484 (25.28) | 601 (38.53) | <0.001 | 612 (39.23) | 601 (38.53) | 0.686 |
Chronic kidney disease | 654 (1.74) | 38 (2.44) | 0.042 | 44 (2.82) | 38 (2.44) | 0.502 |
Hemiplegia | 745 (1.99) | 80 (5.13) | <0.001 | 75 (4.81) | 80 (5.13) | 0.680 |
Leukemia | 119 (0.32) | 3 (0.19) | 0.386 | 4 (0.26) | 3 (0.19) | 0.705 |
Malignant lymphoma | 567 (1.51) | 24 (1.54) | 0.932 | 25 (1.60) | 24 (1.54) | 0.886 |
Liver cirrhosis | 762 (2.03) | 35 (2.24) | 0.561 | 28 (1.79) | 35 (2.24) | 0.373 |
AIDS | 155 (0.41) | 27 (1.73) | <0.001 | 27 (1.73) | 27 (1.73) | 0.438 |
Family history of colorectal cancer | 230 (0.61) | 22 (1.41) | <0.001 | 23 (1.47) | 22 (1.41) | 0.881 |
NOTE: Bold indicates statistical significance (P < 0.05).
Abbreviation: COPD, chronic obstructive pulmonary disease.
The site of colorectal cancer was the right colon for 265 cases (17.0%), left colon for 521 cases (33.4%), and both left and right for one case (0.1%); 773 cases (49.6%) had no data for colorectal cancer site (Table 2). The use history for the 40 drug classes was compared between the colorectal cancer and non–colorectal cancer groups. Use of alpha-blockers (OR, 0.69), xanthine oxidase inhibitors (OR, 0.69), aspirin (OR, 0.63), vitamin D (OR, 0.36), vitamin B, vitamin C, vitamin E, ARB, nateglinide, probiotics, thienopyridine, folic acid, nitrovasodilators, bisphosphonates, Ca-blockers, calcium, steroids, and statins were significantly associated with a decreased risk of colorectal cancer (Table 3). In the CCE, doxazosin, febuxostat, and bisoprolol were the most frequently used alpha-blocker, xanthine oxidase inhibitor, and beta-blocker, respectively.
Site of cancer at diagnosis (N = 1,560)
Site . | Number of patients . |
---|---|
Right-sided colon | 265 (17.0) |
Appendix | 2 (0.06) |
Cecum | 56 (1.79) |
Ascending | 133 (8.53) |
Transverse | 74 (4.74) |
Left-sided colon | 521 (33.4) |
Descending | 33 (2.12) |
Sigmoid | 252 (16.15) |
Rectum | 236 (15.13) |
Double | |
Cecum + Rectum | 1 (0.06) |
Data not provided | 773 (49.55) |
Site . | Number of patients . |
---|---|
Right-sided colon | 265 (17.0) |
Appendix | 2 (0.06) |
Cecum | 56 (1.79) |
Ascending | 133 (8.53) |
Transverse | 74 (4.74) |
Left-sided colon | 521 (33.4) |
Descending | 33 (2.12) |
Sigmoid | 252 (16.15) |
Rectum | 236 (15.13) |
Double | |
Cecum + Rectum | 1 (0.06) |
Data not provided | 773 (49.55) |
Association between drug exposure and colorectal cancer
Drugs . | Non–colorectal cancer (n = 1,560) . | Colorectal cancer (n = 1,560) . | OR (95% CI) . | P . |
---|---|---|---|---|
ACE | 55 | 58 | 1.06 (0.73–1.54) | 0.774 |
Acetaminophen | 91 | 72 | 0.78 (0.57–1.07) | 0.126 |
Pioglitazone | 53 | 37 | 0.69 (0.45–1.06) | 0.089 |
aGI | 68 | 74 | 1.09 (0.78–1.53) | 0.607 |
Alpha-blocker | 103 | 73 | 0.69 (0.51–0.95) | 0.021 |
Xanthine oxidase inhibitor | 100 | 64 | 0.63 (0.45–0.86) | 0.004 |
ARB | 266 | 205 | 0.74 (0.60–0.90) | 0.002 |
Aspirin | 153 | 100 | 0.63 (0.48–0.82) | 0.001 |
Beta-blocker | 96 | 77 | 0.79 (0.58–1.08) | 0.137 |
Biguanide | 72 | 65 | 0.89 (0.64–1.26) | 0.541 |
Bisphosphonates | 44 | 21 | 0.47 (0.28–0.80) | 0.005 |
Ca-blocker | 308 | 239 | 0.74 (0.61–0.89) | 0.001 |
Calcium | 2 | 5 | 2.50 (0.49–12.9) | 0.273 |
Cox2-inhibitor | 26 | 15 | 0.57 (0.30–1.09) | 0.088 |
DPP4-inhobitor | 41 | 26 | 0.63 (0.38–1.03) | 0.066 |
EPA | 21 | 10 | 0.47 (0.22–1.01) | 0.052 |
Nateglinide | 31 | 16 | 0.51 (0.28–0.94) | 0.030 |
Fibrates | 24 | 19 | 0.79 (0.43–1.45) | 0.446 |
Insulin | 53 | 59 | 1.12 (0.77–1.63) | 0.564 |
Nitrovasodilator | 70 | 48 | 0.68 (0.47–0.98) | 0.040 |
NOAC | 8 | 4 | 0.50 (0.15–1.66) | 0.257 |
NSAIDs | 212 | 189 | 0.95 (0.80–1.13) | 0.219 |
Dipyridamole | 5 | 5 | 1.00 (0.29–3.46) | 1.000 |
Cilostazol | 23 | 20 | 0.87 (0.48–1.59) | 0.646 |
Diuretics | 124 | 99 | 0.79 (0.60–1.03) | 0.083 |
Selective estrogen receptor modulators | 9 | 4 | 0.43 (0.13–1.42) | 0.164 |
Statin | 216 | 165 | 0.74 (0.59–0.91) | 0.005 |
Steroid | 67 | 45 | 0.66 (0.45–0.97) | 0.006 |
Sulfonylurea | 113 | 125 | 1.12 (0.86–1.45) | 0.419 |
Thienopyridines | 59 | 37 | 0.62 (0.41–0.94) | 0.024 |
Thiamazole | 23 | 15 | 0.65 (0.34–1.25) | 0.192 |
Potassium sodium hydrogen citrate | 12 | 7 | 0.58 (0.23–1.48) | 0.256 |
Vitamin B | 135 | 91 | 0.65 (0.49–0.86) | 0.003 |
Vitamin C | 30 | 12 | 0.40 (0.20–0.78) | 0.007 |
Vitamin D | 54 | 20 | 0.36 (0.22–0.61 | <0.001 |
Vitamin E | 8 | 7 | 0.44 (0.25–0.78) | 0.005 |
Warfarin | 45 | 36 | 0.80 (0.51–1.24) | 0.314 |
Probiotics | 100 | 74 | 0.73 (0.53–0.99) | 0.043 |
Folic acid | 18 | 6 | 0.33 (0.13–0.84) | 0.019 |
Uricosuric | 12 | 21 | 1.76 (0.86–3.59) | 0.120 |
Drugs . | Non–colorectal cancer (n = 1,560) . | Colorectal cancer (n = 1,560) . | OR (95% CI) . | P . |
---|---|---|---|---|
ACE | 55 | 58 | 1.06 (0.73–1.54) | 0.774 |
Acetaminophen | 91 | 72 | 0.78 (0.57–1.07) | 0.126 |
Pioglitazone | 53 | 37 | 0.69 (0.45–1.06) | 0.089 |
aGI | 68 | 74 | 1.09 (0.78–1.53) | 0.607 |
Alpha-blocker | 103 | 73 | 0.69 (0.51–0.95) | 0.021 |
Xanthine oxidase inhibitor | 100 | 64 | 0.63 (0.45–0.86) | 0.004 |
ARB | 266 | 205 | 0.74 (0.60–0.90) | 0.002 |
Aspirin | 153 | 100 | 0.63 (0.48–0.82) | 0.001 |
Beta-blocker | 96 | 77 | 0.79 (0.58–1.08) | 0.137 |
Biguanide | 72 | 65 | 0.89 (0.64–1.26) | 0.541 |
Bisphosphonates | 44 | 21 | 0.47 (0.28–0.80) | 0.005 |
Ca-blocker | 308 | 239 | 0.74 (0.61–0.89) | 0.001 |
Calcium | 2 | 5 | 2.50 (0.49–12.9) | 0.273 |
Cox2-inhibitor | 26 | 15 | 0.57 (0.30–1.09) | 0.088 |
DPP4-inhobitor | 41 | 26 | 0.63 (0.38–1.03) | 0.066 |
EPA | 21 | 10 | 0.47 (0.22–1.01) | 0.052 |
Nateglinide | 31 | 16 | 0.51 (0.28–0.94) | 0.030 |
Fibrates | 24 | 19 | 0.79 (0.43–1.45) | 0.446 |
Insulin | 53 | 59 | 1.12 (0.77–1.63) | 0.564 |
Nitrovasodilator | 70 | 48 | 0.68 (0.47–0.98) | 0.040 |
NOAC | 8 | 4 | 0.50 (0.15–1.66) | 0.257 |
NSAIDs | 212 | 189 | 0.95 (0.80–1.13) | 0.219 |
Dipyridamole | 5 | 5 | 1.00 (0.29–3.46) | 1.000 |
Cilostazol | 23 | 20 | 0.87 (0.48–1.59) | 0.646 |
Diuretics | 124 | 99 | 0.79 (0.60–1.03) | 0.083 |
Selective estrogen receptor modulators | 9 | 4 | 0.43 (0.13–1.42) | 0.164 |
Statin | 216 | 165 | 0.74 (0.59–0.91) | 0.005 |
Steroid | 67 | 45 | 0.66 (0.45–0.97) | 0.006 |
Sulfonylurea | 113 | 125 | 1.12 (0.86–1.45) | 0.419 |
Thienopyridines | 59 | 37 | 0.62 (0.41–0.94) | 0.024 |
Thiamazole | 23 | 15 | 0.65 (0.34–1.25) | 0.192 |
Potassium sodium hydrogen citrate | 12 | 7 | 0.58 (0.23–1.48) | 0.256 |
Vitamin B | 135 | 91 | 0.65 (0.49–0.86) | 0.003 |
Vitamin C | 30 | 12 | 0.40 (0.20–0.78) | 0.007 |
Vitamin D | 54 | 20 | 0.36 (0.22–0.61 | <0.001 |
Vitamin E | 8 | 7 | 0.44 (0.25–0.78) | 0.005 |
Warfarin | 45 | 36 | 0.80 (0.51–1.24) | 0.314 |
Probiotics | 100 | 74 | 0.73 (0.53–0.99) | 0.043 |
Folic acid | 18 | 6 | 0.33 (0.13–0.84) | 0.019 |
Uricosuric | 12 | 21 | 1.76 (0.86–3.59) | 0.120 |
NOTE: Forty Drugs. Bold indicates statistical significance (P < 0.05).
Abbreviations: ACE, angiotensin-converting enzyme inhibitor; ARB, angiotensin II receptor blocker; aGI, alpha-glucosidase inhibitor; Ca-blocker, calcium channel blocker; DPP4-inhibitor, dipeptidyl peptidase-4 inhibitor; EPA, eicosapentaenoic acid.
Cell proliferation assay
Doxazosin (alpha-blocker) suppressed the proliferation of RKO in a concentration-dependent manner (Fig. 2). There were significant differences at the following points (24 hours: vehicle >15 μmol/L >30 μmol/L, 48 and 72 hours: vehicle >5 μmol/L >15 μmol/L >30 μmol/L; P < 0.05). At high concentrations, the proliferation of HCT116 cells was also significantly suppressed at 24, 48, and 72 hours; P < 0.05. In contrast, febuxostat (xanthine oxidase inhibitor) and bisoprolol did not inhibit the proliferation of either HCT116 or RKO cells. Aspirin suppressed the proliferation of RKO and HCT116 in a concentration-dependent manner. There were significant differences at the following points (RKO 24 hours: vehicle >5 mmol/L, 42 hours: vehicle >2.5 mmol/L >5 mmol/L, 72 hours: vehicle >1.25 mmol/L >2.5 mmol/L >5 mmol/L, HCT116 24 hours: vehicle >5 mmol/L, and 42 and 72 hours: vehicle>2.5 mmol/L >5 mmol/L; P < 0.05).
Doxazosin (alpha-blocker) but not febuxostat inhibits human colorectal cancer cell proliferation. A, Growth curves of RKO and HCT116 cells treated with 0.6% DMSO, doxazosin (1, 5, 15, or 30 μmol/L), febuxostat (1 and 10 μmol/L), aspirin (0.31 μmol/L, 0.62 μmol/L, 1.25 mmol/L, 2.5 mmol/L, and 5 mmol/L), bisoprolol (12.5, 25, 50, and 100 μmol/L) as determined using cell proliferation assays. Untreated cells were used as a control. Results depict a single biological replicate (n = 8). Data shown are means and SDs.
Doxazosin (alpha-blocker) but not febuxostat inhibits human colorectal cancer cell proliferation. A, Growth curves of RKO and HCT116 cells treated with 0.6% DMSO, doxazosin (1, 5, 15, or 30 μmol/L), febuxostat (1 and 10 μmol/L), aspirin (0.31 μmol/L, 0.62 μmol/L, 1.25 mmol/L, 2.5 mmol/L, and 5 mmol/L), bisoprolol (12.5, 25, 50, and 100 μmol/L) as determined using cell proliferation assays. Untreated cells were used as a control. Results depict a single biological replicate (n = 8). Data shown are means and SDs.
Mouse tumor assay
Next, the inflammation-associated colorectal cancer mouse model, AOM/DSS, was used to test whether doxazosin treatment affected colorectal cancer in vivo. As shown in the schematic in Fig. 3A, wild-type C57BL6 mice received AOM/DSS to establish distal colon tumors. Doxazosin treatment began on day 1 and continued until the animals were sacrificed on day 64. Macroscopic observation of dissected colons from the mice revealed that a mean of three colon tumors were present in the middle to distal colon of the doxazosin-treated group; the average tumor size was 3 mm. In the control group, approximately six colon tumors sized approximately 5 mm were present (Fig. 3B). The tumor number and size were significantly smaller in the doxazosin group than the control group (P < 0.05; Fig. 3C). Both groups had tumors with similar pathology on hematoxylin and eosin (H&E) staining: tubular adenomas or well-differentiated adenocarcinomas (Fig. 3D).
A, Schematic representation of the mouse assay of doxazosin administration on AOM/DSS-induced colon tumors. B, Typical macroscopic AOM/DSS-induced colorectal tumors in mice treated with doxazosin or PBS. Scale bars = 1 cm. C, Tumor number and tumor size were determined in PBS- (n = 6) and doxazosin-treated mice (n = 6). Data shown are means and SEM. *, P < 0.05 by Mann–Whitney U test. D, Representative histopathologic images of staining for H&E in doxazosin- or PBS-treated mice. Low magnification (100 ×; left) and high magnification (400 ×; right) images are shown. Scale bar = 200 μm (left), 50 μm (right). E, Body weight curves for each data point shown in the right. All data analyses were not significant (ns) according to Mann–Whitney U tests. F, Representative IHC images for Ki67 staining in AOM/DSS-derived colon tumors in mice treated with PBS or doxazosin. Scale bars = 50 μm. G, The proportion of Ki67-positive cells in AOM/DSS-derived tumors from PBS- and doxazosin-treated mice. Data shown are means ± SEM (n = 3; ns, no significant difference; Mann–Whitney U test).
A, Schematic representation of the mouse assay of doxazosin administration on AOM/DSS-induced colon tumors. B, Typical macroscopic AOM/DSS-induced colorectal tumors in mice treated with doxazosin or PBS. Scale bars = 1 cm. C, Tumor number and tumor size were determined in PBS- (n = 6) and doxazosin-treated mice (n = 6). Data shown are means and SEM. *, P < 0.05 by Mann–Whitney U test. D, Representative histopathologic images of staining for H&E in doxazosin- or PBS-treated mice. Low magnification (100 ×; left) and high magnification (400 ×; right) images are shown. Scale bar = 200 μm (left), 50 μm (right). E, Body weight curves for each data point shown in the right. All data analyses were not significant (ns) according to Mann–Whitney U tests. F, Representative IHC images for Ki67 staining in AOM/DSS-derived colon tumors in mice treated with PBS or doxazosin. Scale bars = 50 μm. G, The proportion of Ki67-positive cells in AOM/DSS-derived tumors from PBS- and doxazosin-treated mice. Data shown are means ± SEM (n = 3; ns, no significant difference; Mann–Whitney U test).
In the AOM/DSS model, weight loss correlates with the severity of colitis. There was no significant difference in body weight between treatment groups (Fig. 3E). Because doxazosin inhibited colorectal cancer cell proliferation in vitro, Ki67 IHC was performed to assess cell proliferation in the mouse tumors; there was no difference between the treatment groups (Fig. 3F and G).
Discussion
This large case–controlled study suggests that, like aspirin and vitamin D, alpha-blockers, xanthine oxidase inhibitors, vitamin B, vitamin C, vitamin E, ARB, nateglinide, probiotics, thienopyridine, folic acid, nitrovasodilators, bisphosphonates, Ca-channel blockers, steroids, and statins may have colorectal cancer preventive effects. Aspirin is a well-validated colorectal cancer chemopreventative (3, 4, 9, 10). Also, vitamin D and calcium are supported by some large observational studies (14–19). However, recent clinical trial results did not show the association between vitamin D supplementation and invasive colorectal cancer (20). In this study, only 7 patients received calcium by prescription and it was not significantly associated with colorectal cancer prevention. Many of the remaining hits in this study have been previously assessed for colorectal cancer preventative effects, including vitamin B, vitamin C, vitamin E, folic acid, bisphosphonates, Ca-channel blockers, ARB, thienopyridine, nitrovasodilators, probiotics, statins, and steroids (21–32). Vitamin B, vitamin C, vitamin E, folic acid, bisphosphonates, and probiotics may have some colorectal cancer preventive effects, although the evidence is not yet strong.
Of the list of drugs that were potentially associated with colorectal cancer prevention in this study, only alpha-blockers and xanthine oxidase inhibitors have not been studied in detail in a colorectal cancer setting to date. Thus, we assessed their potential as colorectal cancer chemopreventative agents in vitro and in vivo. The alpha-blocker doxazosin both inhibited the growth of human colorectal cancer cell lines in culture and decreased tumor size and number in a colorectal cancer mouse model. Doxazosin targets alpha-adrenoceptor signaling and its use has not been studied in any great detail in colorectal cancer. Nevertheless, doxazosin suppressed bladder and prostate carcinogenesis in cohort studies (33, 34) and induced cancer cell apoptosis in vitro and in vivo (35, 36). According to a recent report, doxazosin reduced VEGF levels and angiogenesis in clinical prostate tumors (37). Importantly, tumor angiogenesis is maintained through nerve activation via adrenergic signaling (38). In this cell proliferation assays, doxazosin suppressed cancer cell proliferation; however, the proportion of Ki67-positive tumor cells in mice was not changed. Therefore, the cancer-suppressing effect of doxazosin might be due to apoptosis, necrosis, or changes to the microenvironment rather than cancer proliferation. This case–controlled study suggested that not only aspirin but also other drugs may be effective chemopreventative agents for colorectal cancer. Some drugs, such as Ca-channel blockers and steroids, had no such effect in previous studies (32, 39, 40). As these previous studies were retrospective analyses, they could have been affected by bias. To reduce the potential bias, we performed propensity matching using a comorbidity index.
Limitation
However, other factors also associated with colorectal cancer incidence were not accounted for, such as lifestyle factors including tobacco and alcohol use and body mass index (BMI) and socioeconomic status. Information regarding lifestyle factors and BMI were not collected for the CCE database, and the cohort does contain patients from diverse backgrounds. Patient economic status may be a strong bias as it affects the number of hospital visits and the ability of the patient to receive medications. As such, we used experimental approaches to assess the potential of doxazosin and febuxostat as chemopreventative agents for colorectal cancer. However, the concentration of these drugs in the cell proliferation assay was much higher than human plasma levels after administration. Doxazosin (30 μmol/L) corresponds to 13 ug/mL, which is higher than the interview form Cmax 40 ng/mL. Although we validated doxazosin as an agent of interest for further testing, we have not verified its mechanism of action. As such, we plan to confirm this retrospective case–controlled study in a subsequent prospective trial and undertake further basic research in this area.
In conclusion, alpha-blockers are potential chemopreventatives for colorectal cancer.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Disclaimer
The funders had no role in the design of the study, data collection and analysis, decision to publish, or preparation of the manuscript.
Authors' Contributions
Conception and design: N. Suzuki, R. Niikura, S. Ihara, Y. Hayakawa, Y. Hirata, R. Nakata
Development of methodology: N. Suzuki, Y. Hirata, R. Nakata
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): N. Suzuki, S. Ihara, Y. Hikiba, H. Kinoshita, N. Higashishima, Y. Hirata, R. Nakata, M. Okamoto, M. Sano
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): N. Suzuki, R. Niikura, Y. Hayakawa, R. Nakata, D. Worthley
Writing, review, and/or revision of the manuscript: N. Suzuki, R. Niikura, S. Ihara, N. Higashishima, A. Yamada, Y. Hirata, M. Okamoto, M. Sano, M. Ichinose, S.L. Woods, D. Worthley, K. Koike
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): N. Suzuki, Y. Hikiba, A. Kushiyama
Study supervision: N. Suzuki, Y. Iwamoto, K. Koike
Acknowledgments
The authors thank NY and M. Okamoto of the Japanese Red Cross Medical Center, and SH, KY, and YK of Yaizu City Hospital for their help with data collection. The English in this document has been checked by at least two professional editors, both native speakers of English. For a certificate, please see: http://www.textcheck.com/certificate/wl8SGj. This study was supported by the KAKENHI Grant-in-Aid for Scientific Research, 15K19315, the fellowship grant of Astellas Foundation for Research on Metabolic Disorders, the fellowship grant of Uehara Memorial Foundation, the research grant of Japan Foundation for applied enzymology, the research grant of Smoking Research Foundation, the research grant of Takeda Science Foundation Medical.
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.