Cholecalciferol or 25-Hydroxycholecalciferol Neither Prevents Nor Treats Adenomas in a Rat Model of Familial Colon Cancer^,

Abstract

Background: Epidemiologic studies in humans have shown associations between greater sunlight exposure, higher serum 25-hydroxycholecalciferol [25(OH)D₃] concentrations, and reduced colon cancer risk. However, results from a limited number of vitamin D supplementation trials in humans have not shown a protective effect.

Objective: We sought to determine whether adding to the diet increasing amounts of either 25(OH)D₃, the stable metabolite measured in serum and associated with cancer risk, or cholecalciferol (vitamin D₃), the compound commonly used for supplementation in humans, could reduce emergent adenomas (chemoprevention) or decrease the growth of existing adenomas (treatment) in the colons of vitamin D–sufficient rats carrying a truncation mutation of adenomatous polyposis coli (Apc), a model of early intestinal cancer.

Methods: Apc^Pirc/+ rats were supplemented with either vitamin D₃ over a range of 4 doses [6–1500 μg/(kg body weight · d)] or with 25(OH)D₃ over a range of 6 doses [60–4500 μg/(kg body weight · d)] beginning after weaning. Rats underwent colonoscopy every other week to assess effects on adenoma number and size. At termination (140 d of age), the number of tumors in the small intestine and colon and the size of tumors in the colon were determined, and serum calcium and 25(OH)D₃ measurements were obtained.

Results: At lower doses (those that did not affect body weight), neither of the vitamin D compounds reduced the number of existing or emergent colonic tumors (P-trend > 0.24). By contrast, supplementation at higher doses (those that caused a suppression in body weight gain) with either 25(OH)D₃ or vitamin D₃ caused a dose-dependent increase in colonic tumor number in both males and females (P-trend < 0.003).

Conclusions: No evidence for protection against colon tumor development was seen with lower dose supplementation with either cholecalciferol or 25-hydroxycholecalciferol. Thus, the association between sunlight exposure and the incidence of colon cancer may involve factors other than vitamin D concentrations. Alternative hypotheses warrant investigation. Furthermore, this study provides preliminary evidence for the need for caution regarding vitamin D supplementation of humans at higher doses, especially in individuals with sufficient serum 25(OH)D₃ concentrations.

Introduction

Worldwide, 1.4 million new cases of colon cancer are diagnosed each year, resulting in nearly 700,000 deaths, ∼80/h (1). In the United States, colorectal cancer is the third leading cause of cancer-related death in men and women (2). A disease associated with aging, colon cancer generally affects individuals aged ≥50 y (2). However, the full development of the early adenoma into frank cancer can take a decade or longer (3). Thus, effective chemopreventive agents that are readily available and can be incorporated into a healthy diet in advance of or during early tumor development could reduce long-term disease burden.

Vitamin D is one dietary component that has received a great deal of attention for its potential to reduce colorectal cancer risk. Observations made >3 decades ago by Garland and Garland (4) showed that sunlight exposure is inversely related to colon cancer risk. They subsequently hypothesized that vitamin D might be responsible for reducing that risk. Since then, numerous studies indeed showed that individuals with higher serum 25-hydroxycholecalciferol [25(OH)D₃]⁷ concentrations show a reduced risk of developing colon cancer (5–7). Studies of colon cancer cells in culture give observations consistent with this finding (8–11). Thus, the human association data and mechanistic data in vitro support vitamin D as a strong candidate for a dietary supplement to prevent or even treat colorectal cancer.

The association between elevated serum 25(OH)D₃ concentrations and reduced cancer risk is strong, and plausible mechanisms for the action of vitamin D against colorectal cancer cells have been demonstrated. However, the ultimate test of causation is to determine whether increasing an individual’s serum 25(OH)D₃ concentration decreases future colon cancer risk. The results of the few such supplementation trials are much less conclusive than the association studies. For example, in a randomized, double-blind placebo-controlled trial, >2600 men and women over the age of 65 y given a bolus dose of 100,000 IU cholecalciferol (vitamin D₃) once every 4 mo for 5 y showed no significant difference in the rate of colon cancer development (7 cancers) compared with those given placebo (11 cancers; P = 0.33) (12). Daily supplementation trials also showed no protection. In a randomized, double-blind placebo-controlled trial involving >36,000 postmenopausal women, those given 400 IU vitamin D plus 1000 mg calcium daily for 7 y with follow-up during the same interval developed a similar number of colon cancers (168 cancers) as did those given placebo (154 cancers; P = 0.51) (13). As is well known, similar observations led to the rigorous demonstration by supplementation that vitamin D causes healthy bone development in experimental animals and in humans. But such a stringent demonstration of causation has not yet been shown for a protective role of vitamin D against colon cancer. Despite this, vitamin D supplementation has been enthusiastically embraced to the extent that there are some recommendations to increase serum vitamin D concentrations above those currently deemed sufficient for bone health.

Assessment of the effect of supplements in human populations can encounter many lifestyle and dietary confounders (14). The use of appropriate rodent models of the disease in a controlled setting can overcome many of these factors. Although some studies of carcinogen-induced tumorigenesis in mouse (15, 16) and rat (17–19) models showed protection with vitamin D supplementation, studies in genetic models of the disease did not show a protective effect (20, 21). Recently, we reported that supplementation with a moderate dose of 25(OH)D₃ did not reduce the number of adenomas in the small intestine (SI) or colon or alter the growth patterns of colonic adenomas in 2 animal models that carry truncation mutations of adenomatous polyposis coli (Apc), Apc^Min/+ mice and Apc^Pirc/+ rats (21). In subsequent studies in our laboratory investigating chemopreventive compounds, we chose to use Apc^Pirc/+ rats, which more closely model human colon cancer, because of the predominance of tumors in the colon rather than the SI, unlike the C57BL/6-Apc^Min/+ mouse. The biology of the colon differs from that of the SI—for example, by showing preferential male susceptibility to colonic neoplasia in the Apc^Pirc/+ rat and Apc^Min/+ mouse (22).

To more thoroughly test whether relevant vitamin D compounds can protect specifically against colonic adenoma development, we examined the effect of supplementation with vitamin D₃ over a range of 4 doses [6–1500 μg/(kg body weight · d)] or with 25(OH)D₃ over a range of 6 doses [60–4500 μg/(kg body weight · d)] in vitamin D–sufficient rats. In this study, we sought generality by examining supplementation with vitamin D₃ and also by its stable metabolite 25(OH)D₃. Human dietary supplementation generally involves vitamin D₃. When vitamin D₃ is ingested, it undergoes hydroxylation in the liver to form 25(OH)D₃ (23). Because 25(OH)D₃ is the compound measured in serum that is inversely associated with colon cancer risk, we chose also to test the effect of this compound in our model.

To broaden the assessment of effect on tumors, the rats were followed longitudinally by colonoscopy to assess the impact on both existing adenomas as well as newly emerging adenomas. Then, at termination at 140 d of age, terminal tumor counts for the SI and colon, terminal sizing of colonic tumors, and serum calcium and 25(OH)D₃ measurements were obtained.

Methods

Animal breeding and maintenance.

All protocols were approved by the Animal Care and Use Committee of the University of Wisconsin School of Medicine and Public Health and were consistent with the Guide for the Care and Use of Laboratory Animals (24). Rats were housed in the McArdle Laboratory Vivarium, a facility approved by the American Association of Laboratory Animal Care. Rats were maintained on a 12:12-h light:dark cycle with free access to food and acidified water. F₁ generation rats were created by breeding female August Copenhagen Irish (ACI) Apc^+/+ rats (Harlan) with male Fisher 344N/Tac (F344) Apc^Pirc/+ rats (developed in the laboratory of WFD and available through Taconic) (25). Coisgenic F344-Apc^Pirc/+ rats develop approximately an equal number of tumors in the SI as in the colon; this distribution is shifted toward the colon in congenic ACI-Apc^Pirc/+ rats. However, ACI-Apc^Pirc/+ rats typically show a rapid emergence of a large number of colonic tumors. An F₁ cross between ACI and F344-Apc^Pirc/+ results in rats with adenomas that are predominately in the colon and in more manageable numbers (26).

Chemicals.

Vitamin D₃ was purchased from Spectrum and Sigma and was determined to be 95% pure by HPLC; 25(OH)D₃ was purchased from Chemvon and was determined to be 97% pure by straight-phase HPLC (27).

Diets.

Control rats were fed a vehicle diet of powdered AIN-76A, which contains 1 IU vitamin D₃/g of diet (Harlan) plus 5% (wt:wt) soybean oil (Wesson) and 0.25% (vol:wt) ethanol (Pharmaco Aaper). The soybean oil and ethanol were necessary to incorporate the vitamin D compounds into the AIN-76A diet and were therefore also added to the vehicle diet. The diet was stored under refrigeration and feed given to animals was consumed within 3–5 d. In this study we chose to supplement with vitamin D₃, the compound commonly consumed as a human dietary supplement, or with 25(OH)D₃, the stable intermediate measured in serum and with which cancer risk has been inversely associated. A consumption rate of 20 g of food per rat per day was used to estimate the daily dose amounts of vitamin D₃ and 25(OH)D₃. Vitamin D₃ was added to the diet to achieve target doses of 6, 60, 500, and 1500 μg/(kg body weight · d); 25(OH)D₃ was added to the diet to achieve target doses of 60, 170, 500, 1500, 3000, and 4500 μg/(kg body weight · d). These concentrations were determined from data from a previous study (21) and from pilot studies that determined doses that caused a small but measureable elevation in serum calcium, demonstrating that the compounds were biologically effective while remaining tolerable to the rats. For this study, “lower” doses are defined as those that do not affect body weight gain, whereas “higher” doses are those that cause a suppression in body weight gain.

Experimental design.

At 33 d of age, (ACI × F344)F₁ Apc^Pirc/+ rats were randomly assigned to either the vehicle diet, 1 of the 6 25(OH)D₃-supplemented diets, or 1 of the 4 vitamin D₃–supplemented diets (Figure 1). Both female and male rats were tested in each of the 25(OH)D₃-supplemented groups, except that only females were given the 3000 μg/(kg body weight · d) dose. Only female rats were used for the study involving vitamin D₃, because sex did not affect the results of supplementation in the 25(OH)D₃ study, and female rats have a lower tumor burden than male rats, which permits facile longitudinal tracking. The group of female rats fed 25(OH)D₃ at 3000 μg/(kg body weight · d) were tested at the same time as those in the vitamin D₃ study; therefore, those results are compared against the contemporaneous vehicle controls from the vitamin D₃ study.

FIGURE 1.

Flow chart of the experimental protocols. Dose units are μg/(kg body weight · d). 25(OH)D₃, 25-hydroxycholecalciferol.

Endoscopy.

Rats underwent endoscopy starting before diet supplementation and continuing on alternate weeks until termination at 140 d of age. This allowed us to determine whether supplementation affected existing adenomas as well as whether it could prevent the emergence of new adenomas. In our experience, endoscopy allows visualization of the distal two-thirds of the rat colon. Body weights were measured at each colonoscopy and at termination.

Scoring of longitudinal tumor fate.

Three observers blinded to the rats’ diet determined the longitudinal fate of each tumor by reviewing the still images captured by endoscopy over the duration of the study. Tumors were examined at each endoscopy, and any tumor that appeared at least twice during the study was assigned to 1 of 3 categories: growing, static, or regressing. A consensus was generated on the basis of agreement between at least 2 of the 3 observers.

Terminal tumor sizing and counts.

Rats were killed by asphyxiation with carbon dioxide. The SI and colon were cut open longitudinally, laid flat, washed with sterile PBS, and fixed with formalin. Adenoma counts for the SI and colon were obtained on a dissecting microscope (10× magnification) from formalin-fixed tissues. Tumor sizes were determined by using 1 of 2 methods: either with a reticule on a dissecting microscope or by alginate molding and dental stone casting (28). For the 25(OH)D₃ study, after fixation, adenoma dimensions were measured to the nearest tenth millimeter by using an eyepiece reticule (10× magnification). Two measurements were taken in the plane of the epithelium for each adenoma: the longest tumor dimension was measured first (A), followed by a second measurement perpendicular to the first (B). To determine an area for each adenoma, the following equation for the area of an ellipse was used: Area = π • (A/2) • (B/2). In the vitamin D₃ study, tumor volume was determined by using the alginate gel method to create a negative mold of each adenoma followed by creating a positive cast of the original tumor with dental stone, which was then weighed. Tumor volume was calculated from the weight of the cast, as previously described (28).

Biochemical assays.

Blood was obtained after 3 wk with the diet from the lateral ocular orbit (29) and at termination by cardiac puncture. At least 3 biological replicates per diet were analyzed for blood calcium and 25(OH)D₃ concentrations. Serum calcium was measured by atomic absorption spectrometry on the Perkin Elmer 3110. To determine a rat’s true vitamin D status, serum 25(OH)D₃ concentrations were considered the gold standard, rather than calculation from diet consumption. Serum 25(OH)D₃ concentrations were measured by using HPLC. The HPLC method has been confirmed to separate 25(OH)D₃, 24,25-hydroxycholecalciferol, and 1,25-hydroxycholecalciferol using standards. Extraction recoveries are checked by using tritiated 25-hydroxycholecalciferol each time samples are processed and averages ∼90%. For every 2 experimental samples analyzed, two 25(OH)D₃ standards are analyzed, flanked by a background check after the 2 experimental samples are processed. For the recovery controls, ∼740 Bq of tritiated 25-hydroxycholecalciferol are added to 150 μL and then extracted with 1 mL ethyl acetate 3 times. The pooled supernatants were dried under nitrogen, resuspended in HPLC mobile phase (60% acetonitrile and 40% water), and then filtered through a 0.2-mm hydrophilic polypropylene membrane filter before placing into an auto-sampler vial. The extracted and filtered samples (total volume of 250 μL) were loaded on a Symmetry C18 column (3.9 × 150 mm; Waters) at 30°C, and UV absorbance was measured at 265 nm at a flow rate of 1 mL/min. Blood concentrations of 25(OH)D₃ were calculated on the basis of the extraction/filtration/resuspension recoveries and a standard curve spanning 2.5 to 50 ng.

Statistical methods.

In Apc^Pirc/+ rats, the 100% penetrance of colonic tumors and increased proportion of colonic tumors compared with those in the SI allowed us to use fewer rats while still achieving statistical significance by individually following colonic tumors by endoscopy every other week. Significant differences were observed between females and males for adenoma multiplicities in both the SI and colon, as described previously (25). Therefore, data for each SI and colon from female and male rats were analyzed separately. By contrast, no difference between females and males was observed for either serum calcium or serum 25(OH)D₃ measurements. Therefore, data for both of these measures were combined for the sexes. A 2-sided Jonckheere-Terpstra test was used to test for a trend in the number of adenomas or serum values across a dose series. Tumor multiplicity data are not normally distributed and fit overdistributed Poisson models. Therefore, a 2-sided Wilcoxon rank-sum test was used to determine whether tumor size was affected by supplementation. A Bonferroni-corrected P value has been used in instances with multiple comparisons to the same control. A 2-sided Kendall’s rank correlation test was used to test for an association between the number of adenomas and each of the serum measurements. For each of the tests, P ≤ 0.05 was considered significant. To ensure that the highest doses did not mask any protective effect of the lowest doses, analysis was performed both as a complete dose series and separately for the lowest doses (no effect on body weight gain) and the highest doses (suppression in body weight gain).

Results

Serum calcium and 25(OH)D₃ measurements.

As expected, serum calcium increased in a dose-dependent manner in rats supplemented with vitamin D₃ (P-trend < 0.0001) or 25(OH)D₃ (P-trend < 0.0001; Table 1). These concentrations were reached at or before 3 wk of being fed the diet (data not shown). For rats fed supplemental 25(OH)D₃, the weights of females administered the 60-, 170-, and 500-μg/(kg · d) doses and of males administered the 60-, 170-, 500-, and 1500-μg/(kg · d) doses did not statistically differ from their respective controls. For female rats fed supplemental cholecalciferol, the weights of those administered the 6- and 60-μg doses did not statistically differ from controls. Rats supplemented with vitamin D₃ or 25(OH)D₃ showed a dose-dependent increase in serum 25(OH)D₃ compared with controls (P < 0.0001; Table 1). Both serum calcium concentrations (P < 0.0001) and serum 25(OH)D₃ concentrations (P = 0.008) showed a correlation with the number of colonic adenomas (Figure 2).

TABLE 1.

Body weight and serum concentrations of 25(OH)D₃ and calcium in ^Pirc/+ rats fed diets supplemented with vitamin D₃ or 25(OH)D₃¹

	Females		Males		Serum 25(OH)D3			Serum calcium
Dietary compound and dose2	Rats, n	Weight,3 g	Rats, n	Weight,3 g	Samples, n	Concentration,4 μg/L	P-trend	Samples, n	Concentration,4 mg/dL	P-trend
25(OH)D3 [μg/(kg body weight · d)]							<0.00001			<0.00001
0	17	221 ± 13	6	375 ± 24	10	33 ± 5		17	11.6 ± 0.1
60	6	218 ± 14	4	369 ± 4	4	115 ± 11		10	12.2 ± 0.1
170	5	218 ± 8	5	323 ± 26	4	106 ± 12		10	12.7 ± 0.1
500	4	210 ± 10	4	333 ± 9	4	119 ± 15		8	13.0 ± 0.3
1500	9	195 ± 13*	4	316 ± 34	9	147 ± 22		13	13.3 ± 0.1
3000	6	183 ± 14*	0	NA	5	195 ± 13		5	13.1 ± 0.1
4500	8	169 ± 27*	4	239 ± 57*	9	233 ± 27		12	14.0 ± 0.2
Vitamin D3 [μg/(kg body weight · d)]							<0.00001			<0.0001
0	8	219 ± 28	0	NA	8	37 ± 6		8	11.8 ± 0.2
6	5	219 ± 7	0	NA	3	57 ± 22		3	11.7 ± 0.2
60	10	227 ± 25	0	NA	8	119 ± 11		8	12.0 ± 0.3
500	5	172 ± 9*	0	NA	5	535 ± 33		5	13.6 ± 0.2
1500	4	181 ± 40*	0	NA	4	780 ± 19		4	15.3 ± 0.4

¹*Significant difference compared with control values, P ≤ 0.05. NA, not applicable; 25(OH)D₃, 25-hydroxycholecalciferol.

²Dose = the supplementation amount of base AIN-76A diet, which contains 1 IU vitamin D₃/g diet.

³Values are means ± SDs.

⁴Values are means ± SEMs.

FIGURE 2.

Terminal serum 25(OH)D₃ (A) and calcium (B) concentrations vs. colonic tumor number for Apc^Pirc/+ rats fed diets supplemented with 25(OH)D₃ or vitamin D₃. Each point represents data for a single rat; lines show the regression line for each data set. Each serum measure showed a significant positive association with terminal colonic tumor number. Shaded regions indicate normal serum value ranges: 19.1–53.9 for 25(OH)D₃ and 11.1–12.5 for calcium, calculated from control rats by using the mean and SD. Kendall’s rank correlation coefficients: serum 25(OH)D₃ vs. tumor count τ = 0.2264; serum calcium vs. tumor count τ = 0.2761. 25(OH)D₃, 25-hydroxycholecalciferol.

Adenoma counts in 25(OH)D₃-supplemented Apc^Pirc/+ rats.

A reduction in the number of adenomas in the SI was seen for both female (P-trend < 0.01) and male (P-trend < 0.04) rats with increasing doses of 25(OH)D₃ (Figure 3). On average, rats fed the vehicle diet exhibited 3- to 4-fold more adenomas in the SI compared with those fed the highest dose of 25(OH)D₃. The majority of tumors in the SI form in the proximal 2 of 4 equally sized sections. Only the second most proximal section in both females and males showed a significant reduction in tumor number with increasing doses of 25(OH)D₃ (P-trend = 0.03; data not shown).

FIGURE 3.

Number of tumors in the SI and colon of Apc^Pirc/+ rats fed diets supplemented with 25(OH)D₃ (A) or vitamin D₃ (B). Each point represents the total number of tumors for the designated section of intestine in a single rat. (A) Supplementation with 25(OH)D₃ in either female or male rats did not reduce the terminal number of colon tumors but instead increased tumor number; 25(OH)D₃ supplementation marginally repressed tumor development in the SI. (B) Similarly, supplementation with vitamin D₃ in female rats did not decrease but instead increased colonic tumor number. Tumor number in the SI was marginally reduced. SI, small intestine; 25(OH)D₃, 25-hydroxycholecalciferol.

No protection against the development of colonic adenomas was evident with low-dose 25(OH)D₃ supplementation. However, male rats given the lowest dose [60 μg/(kg · d)] of 25(OH)D₃ developed fewer colonic tumors than did vehicle-treated rats, although this was not significant (P = 0.66). When the 3 lowest dose groups were considered as a series, the lack of protection remained evident (females, P = 0.11; males, P = 0.18). Instead, we observed an increase in the number of colonic adenomas with increasing doses of 25(OH)D₃ for both female (P-trend < 0.0001) and male (P-trend < 0.003) rats (Figure 3). On average, rats administered the highest dose of 25(OH)D₃ had up to 3-fold more colonic adenomas compared with those fed the vehicle diet.

Adenoma counts in vitamin D₃–supplemented Apc^Pirc/+ rats.

No difference in the number of adenomas of the SI was seen between rats fed any of the vitamin D₃–supplemented diets and those fed the vehicle diet (P-trend = 0.33; Figure 3). Again, the majority of the small number of tumors in the SI form in the most proximal 2 of the 4 sections, but in each of these sections no difference was detected between rats fed any supplemented diet and those fed the vehicle diet (data not shown).

Similar to the results with the 25(OH)D₃-supplemented diet, no protection against the development of colonic tumors was evident with supplementation with vitamin D₃. When considering only the 2 lowest dose groups, there was no evidence for any protection (P = 0.76). Instead, an increase in the number of colonic tumors with increasing doses of vitamin D₃ was observed (P-trend < 0.003; Figure 3).

Growth profile analysis of colonic adenomas.

Tumors from female rats were evaluated for longitudinal growth patterns by endoscopy; male rats develop a large number of tumors that become difficult to annotate accurately between endoscopic visits. No consistent difference was observed in the number of growing vs. nongrowing tumors between rats fed any supplemented diet and those fed the vehicle diet (Table 2).

TABLE 2.

Tumor fates in female ^Pirc/+ rats fed diets supplemented with 25(OH)D₃ or vitamin D₃¹

		Tumor fate, %
Dietary compound and dose2	Total colonic tumors followed, n	Grew	Remained static	Regressed	P-value
25(OH)D3 [μg/(kg body weight · d)]
0	40	75	20	5
60	28	64	36	0	0.10
170	25	96	0	4	<0.0001
500	24	83	13	4	0.15
1500	37	78	19	3	0.57
3000	28	71	29	0	<0.0013
4500	67	76	21	3	0.86
Vitamin D3 [μg/(kg body weight · d)]
0	22	91	9	0
6	14	93	7	0	0.60
60	30	73	27	0	0.001
500	31	94	6	0	0.49
1500	11	100	0	0	0.002

¹25(OH)D₃, 25-hydroxycholecalciferol.

²Dose = the supplementation of base diet, which contains 1 IU vitamin D₃/g diet.

³Dose-treated compared with contemporaneous vehicle-treated rats from the vitamin D₃ study.

Sizing of colonic adenomas.

Tumor sizes from rats in the 25(OH)D₃ supplementation study, calculated from measurements acquired by using a dissecting microscope, did not differ significantly between the vehicle group and any supplemented group (Table 3). When the doses were analyzed as a series, we found a significant association between increased 25(OH)D₃ dose and decreased tumor area in male (P-trend < 0.03) but not female (P-trend = 0.14) rats (Table 3). Tumor sizes in rats in the vitamin D₃ study, determined by alginate molding, did not differ significantly between the vehicle group and any supplemented group. However, low-dose vitamin D₃ supplementation was associated with a trend for smaller tumors (P-trend < 0.04; Table 3).

TABLE 3.

Tumor sizes in ^Pirc/+ rats fed diets supplemented with 25(OH)D₃ or vitamin D₃¹

Dietary compound and dose2	Total colonic tumors followed, n	Size	P-value
Tumor areas obtained from measurements by using a microscope reticule
25(OH)D3 [μg/(kg body weight · d)]
Males
0	97	13 ± 15 mm2	—
60	90	15 ± 14 mm2	0.06
170	146	15 ± 14 mm2	0.07
500	108	12 ± 14 mm2	0.77
1500	99	9 ± 7 mm2	0.61
4500	117	9 ± 8 mm2	0.29
Females
0	70	11 ± 11 mm2	—
60	39	10 ± 14 mm2	0.72
170	38	11 ± 9 mm2	0.31
500	31	11 ± 8 mm2	0.24
1500	99	9 ± 7 mm2	0.87
3000	65	9 ± 9 mm2	0.64
4500	148	8 ± 6 mm2	0.95
Tumor volumes obtained by using the alginate molding method
Vitamin D3 [μg/(kg body weight · d)]
Females
0	10	20 ± 36 mm3	—
6	5	6 ± 5 mm3	0.40
60	42	7 ± 7 mm3	0.50
500	56	12 ± 17 mm3	0.89
1500	77	13 ± 16 mm3	0.69

¹Values are means ± SDs unless otherwise indicated. Although a trend for decreased tumor area with increasing 25(OH)D₃ dose was apparent, no individual dose differed significantly from vehicle controls. Low-dose vitamin D₃ supplementation appeared to decrease average tumor volume compared with controls; however, this was not significant. 25(OH)D₃, 25-hydroxycholecalciferol.

²Dose = the supplementation of base diet, which contains 1 IU vitamin D₃/g diet.

Discussion

Our study examined the effect of supplementation with vitamin D₃ or 25(OH)D₃ on 3 different measures of colonic adenoma development: number, size, and growth patterns. Supplementation with lower doses of either vitamin D₃ or 25(OH)D₃ each did not suppress any of these measures of colonic adenoma development. The data presented here, consistent with evidence from some studies in humans, indicate that supplementation of already sufficient individuals does not provide any added benefit to protect against the development of adenomas.

The power of our study was increased by studying tumors over time in living rats. We did not detect any significant trends between colonic tumor growth rate and supplementation. However, there was suggestive evidence for an effect of vitamin D supplementation on tumor size. Huerta et al. (20) showed that supplementation can decrease tumor size without affecting tumor number. Further research using quantitative measures of tumor size applied in a longitudinal setting is needed to better understand the role of vitamin D during each stage of the tumor development process.

An added strength of our study was the ability in the Apc^Pirc/+ rat model to separately analyze effects on the SI and colon. Our results highlight the important difference between tumors of the SI and those of the colon (30). In contrast to our findings in the colon, we observed a trend for decreased tumor number in the SI with increasing doses of either vitamin D₃ or 25(OH)D₃. Focusing on the colon is especially pertinent because the majority of intestinal tumors in the human form preferentially in the colon and not in the SI.

If supplementation with vitamin D compounds does not exhibit a protective effect, what is the cause underlying the epidemiologic association between sunlight exposure, high serum vitamin D concentrations, and reduced incidence of colon cancer in human populations? Clearly, the original association between sunlight and cancer risk should be revisited to consider alternate hypotheses; vitamin D status may simply serve as a surrogate for sunlight exposure. Sunlight exposure provides benefits beyond those offered by vitamin D. UV radiation was shown to strongly affect the immune system (31, 32); these systemic effects could suppress tumorigenesis independent of vitamin D. A way forward has been suggested by recent investigations of the protective role of UV light in an experimental mouse model of multiple sclerosis: exposure to narrow-band UVB radiation had no effect on the concentrations of vitamin D in serum but nonetheless dramatically reduced the clinical symptoms of the disease (27, 33).

In contrast to the lack of protection with lower dose supplementation, this study raises the possibility for harm with higher dose supplementation of already sufficient subjects. At higher doses of either vitamin D₃ or 25(OH)D₃, adenoma number was significantly increased in the colon. The highest doses tested in this study increased serum 25(OH)D₃ concentrations >7-fold over control values. Although humans may never reach concentrations this high, the detrimental effects of high-dose supplementation demonstrated here may be possible in already sufficient individuals given additional vitamin D. The observed enhancement in the number of colonic adenomas may be directly or indirectly related to the increase in serum calcium. Regardless of the mechanism, this observation provides reason for caution against supplementing already sufficient subjects with high doses of vitamin D or 25(OH)D₃.

This caution for high-dose supplementation specifically to prevent cancer is supported by some association data from human cohorts, in whom an increased risk of colon cancer at high concentrations of vitamin D is suggested. The Health Professionals Follow-Up Study showed that individuals in quintile 2 [median plasma 25(OH)D concentration of 25.0 μg/L; OR: 0.97; 95% CI: 0.55, 1.70], quintile 3 (29.1 μg/L; OR: 0.66; 95% CI: 0.35, 1.24), and quintile 4 (33.3 μg/L; OR: 0.51; 95% CI: 0.27, 0.97) each had a lower risk of developing frank colon cancer compared with individuals in quintile 1 (18.4 μg/L; OR: 1.00; referent) (7). At the other extreme, individuals in quintile 5 [median plasma 25(OH)D concentration of 39.4 μg/L] appear to have an increased risk compared with quintiles 3 and 4, although still lower than for individuals with the lowest serum concentrations (OR: 0.83; 95% CI: 0.45, 1.52). Likewise, an increase was observed in the number of tumors in the SI of Apc^Min/+ mice fed a diet containing an excess amount of multivitamins (34). This suggests that an intermediate amount of vitamins is associated with the lowest risk of adenoma development, and deviation above or below that can increase risk. Clinical trials investigating high-dose supplementation for cancer prevention or treatment should consider the possibility of detrimental effects and closely monitor study patients for any indications of such an adverse effect on cancer outcome.

Our finding that high-dose supplementation may increase cancer risk is timely. Proponents of vitamin D supplementation for protection against disease, particularly cancer, have stated that “universal intake of up to 40,000 IU vitamin D per day is unlikely to result in vitamin D toxicity” (35). However, the authors only accounted for frank vitamin D toxicity and did not consider the ramifications of such recommendations on disease risk. In support of this caution, the Institute of Medicine recently recommended against changes to the current daily recommended intake of vitamin D, stating that serum concentrations above those currently recommended do not consistently show an association with increased benefit and may even show cause for concern (36). Particular care must be taken regarding health claims for supplements, because they are readily available without prescription.

This study of vitamin D compounds in Apc^Pirc/+ rats represents a starting point for investigating these compounds as chemopreventive agents against colon cancer. Several critical questions remain that have not been addressed by this study. Does vitamin D deficiency affect colon cancer risk in a controlled animal model? And, crucially, can UV light exposure reduce risk, especially at exposure levels that do not affect vitamin D concentrations?