The first two authors contributed equally to this project.
Calcium plus vitamin D alters preneoplastic features of colorectal adenomas and rectal mucosa
Article first published online: 13 DEC 2005
Copyright © 2005 American Cancer Society
Volume 106, Issue 2, pages 287–296, 15 January 2006
How to Cite
Holt, P. R., Bresalier, R. S., Ma, C. K., Liu, K.-F., Lipkin, M., Byrd, J. C. and Yang, K. (2006), Calcium plus vitamin D alters preneoplastic features of colorectal adenomas and rectal mucosa. Cancer, 106: 287–296. doi: 10.1002/cncr.21618
- Issue published online: 5 JAN 2006
- Article first published online: 13 DEC 2005
- Manuscript Accepted: 15 AUG 2005
- Manuscript Received: 2 AUG 2005
- National Cancer Institute Awards. Grant Numbers: N01-CN-25439-1, RO1CA69480
- vitamin D;
- adenomatous polyps;
- vitamin D receptor;
- galectin 3
Calcium and vitamin D are chemopreventive agents for colorectal neoplasia. Studies of the effects of calcium and vitamin D on early surrogate markers of reduced risk, such as proliferation, have been limited to evaluation of the flat colorectal mucosa. Biologic changes that may occur in colorectal adenomas after chemopreventive regimens have not been reported.
In the current study, adenomatous polyps were transected, approximately 50% were removed for histologic examination, and the remnants tattooed before the administration of either calcium carbonate (1500 mg 3 times daily) plus vitamin D3 400 IU or a placebo for 6 months. At study end, polyp remnants were resected completely and were used for histologic examination. Immunohistochemical staining was performed in both flat mucosa and in polyp tissue. Proliferation was assessed by MIB-1 staining; apoptosis was assessed by terminal deoxyuridine triphosphate-biotin nick-end labeling, BAK, and Bcl-2 staining; and cytokeratin AE1, vitamin D receptor, MUC5AC mucin, and galectin-3 were assessed by immunohistochemisty.
Nineteen patients, including 11 patients in the treatment group and 8 patients in the control group, completed the study. Proliferative indices fell both in flat mucosa and in polyps in the treatment group, and there were no significant changes in the control group. Apoptosis and Bcl-2 immunostaining were unchanged in both groups, but the frequency of BAK-immunostained cells in the interior of polyps rose significantly. Vitamin D receptor staining increased slightly and significantly in flat rectal tissue in the treatment group. There were no significant changes in galectin-3 staining, but a striking reduction in MUC5AC mucin staining in polyps was observed after treatment with calcium plus vitamin D.
The administration of a calcium plus vitamin D chemopreventive regimen resulted in several changes in adenomatous tissue that may have contributed to reduced polyp formation. Cancer 2006. © 2005 American Cancer Society.
Colorectal carcinoma is the second leading cause of cancer deaths in the United States, and chemoprevention for this tumor has been the focus of considerable research over the past 3 decades. Epidemiologic studies have been concentrated on the potential of calcium and, more recently, vitamin D on surrogate biomarkers of the risk for colorectal neoplasia.
Supplemental calcium administration to volunteers who were at risk for colorectal neoplasia in the form of calcium carbonate (1.2–2.0 g per day) in several studies resulted in a change in proliferation kinetics from a higher risk to a lower risk pattern in the flat mucosa of the colorectum.1, 2 Similar observations were made when low-fat dairy foods were consumed as a source of supplemental calcium.3 These observations led to several prospective studies that demonstrated a reduction in the recurrence of colorectal adenomas after calcium carbonate supplementation in patients who had undergone previous resection for adenomas.4, 5 In a study by Grau et al., calcium was more effective when higher levels of serum 25-hydroxy vitamin D were given,6 an observation that supplemented our own data on the effects of vitamin D on epithelial cell proliferation.7
All previous studies on surrogate biomarkers of risk were conducted on biopsies of flat rectal mucosa from patients who had undergone previous removal of adenomatous polyps. There also is considerable interest in possible changes that may occur in the adenomatous polyps themselves after the administration of a chemopreventive agent. It is clear that neoplastic adenomatous polypoid tissue may differ significantly from that seen in flat mucosa in response to a chemopreventive agent. For example, studies of the effects of a cyclooxygenase 2 inhibitor in patients with familial adenomatous polyposis have shown changes in apoptosis in adenomas after treatment.8
The current study was designed to explore the potential effects of a 6-month chemopreventive regimen of calcium and vitamin D on biomarkers of risk, both in flat rectal mucosa and in polyps that had been left in situ, to obtain new data on individual adenomatous polyps. In this study, polyps were transected prior to administration of the chemopreventive regimen, and remaining tissues were left in situ for 6 months before they were removed. The preliminary data from this study suggest that several changes can be detected with adenomatous polypoid tissue during the administration of a calcium/vitamin D chemopreventive regimen.
MATERIALS AND METHODS
Potential participants were chosen from patients who had polyps found during a screening sigmoidoscopy for colorectal neoplasia and who then were entered into the study after a regular colonoscopy and partial polypectomy. Eligible patients who agreed to participate were then randomized into one of two regimens.
Patients were included if they were age 18 years or older and if they were scheduled for diagnostic or screening colonoscopy. Exclusion criteria included a known history of familial carcinoma syndromes, a personal history of carcinoma other than nonmelanoma skin carcinoma, intestinal malabsorption or inflammatory bowel disease, gastrointestinal surgery other than appendectomy or surgery of the esophagus, known abnormalities of calcium metabolism, daily ingestion of aspirin > 300 mg, or usual doses of other nonsteroidal antiinflammatory drugs. Furthermore, participants could not be ingesting > 1000 mg of calcium per day.
For study entry, patients had to have small (< 9 mm in greatest dimension), benign-appearing polyps in the distal 55 cm of the colorectum and also had to be cleared of all polyps that measured > 9 mm in greatest dimension regardless of location in the colon. Patients who entered the study had a nutritional evaluation, and blood samples were taken for routine serologic analysis plus 25-hydroxy vitamin D determination. Patients then underwent proctosigmoidoscopy, and 4-quadrant biopsies were taken from the rectum 10–15 cm from the anal verge, and the biopsy specimens were fixed in 10% buffered formalin. Polyps in the distal 55 cm were measured by using open forceps, and transected, and approximately 50% of the polyp was removed. The sites at which the polyps were transected were tattooed with India ink. Specimens were confirmed to show adenomatous features on histologic examination. Randomization of patients was then conducted by using a randomly blocked treatment-assignment scheme.
Description of Study
After randomization, patients started taking either calcium carbonate 1500 mg 3 times daily (to deliver 1800 mg elemental calcium) plus vitamin D3 400 IU daily or placebo tablets similar to the calcium carbonate and vitamin D. The medications were continued for 6 months. Unused pills were returned in preaddressed mailers at 3-month intervals. After 6 months, the patients were interviewed again by the study coordinator to confirm that they had not been taking medications other than the protocol drugs. Blood tests, proctosigmoidoscopy, and rectal biopsies were repeated. The site of the India ink-marked adenoma that had been transected was found, the size was measured with open forceps, and the remaining tissue at that site was removed for analysis. The presence of any other adenomas in the distal 55 cm of the colon was recorded. Biochemical profiles on serum samples were performed in the routine biochemistry laboratory of the institution. Serum levels of 1,25 hydroxy vitamin D were protected from light, rapidly frozen, and sent to Columbia Presbyterian Cancer Center Core Laboratory for analysis. Levels of 25 hydroxy vitamin D were assayed by a radioreceptor assay by using a Nichols Institute assay kit.9 Biopsies were fixed in 10% buffered formalin and were processed into paraffin within 12–18 hours. Histologic sections that measured 6 μm thick were obtained from each sample and were taken for histologic examination after standard staining and several immunohistochemical staining methods, as described below. Evaluation of immunohistochemical staining in rectal mucosal crypts was performed only in well oriented crypts that showed their full length and an open crypt lumen. All well oriented crypts were selected for measurement. Polyps were sectioned perpendicular whenever possible. Immunohistochemical staining was described in surface and deep portions of the crypt and was scored for density and frequency of staining.
Proliferation—mib 1 staining
The proliferation-associated antigen examined was a monoclonal antibody (MAb) to Ki-67 (Mib 1 Immunotech, Marseilles, France). Sections were deparaffinized and then exposed to a microwave pretreatment in 10 mmol/L citrate buffer; pH 6.0, at 850 Watts twice for 5 minutes each to enhance antigenicity, as described previously.10 For negative controls, tissue sections were immunostained in the absence of the primary antibody.
Terminal deoxyuridine triphosphate-biotin nick-end label staining
Apoptosis was determined with terminal deoxyuridine (dUTP) triphosphate-biotin nick-end label (TUNEL) staining by using a procedure modified from Gavrielli et al.,11 as described previously.12 Briefly, after preincubation with buffer, sections were incubated with buffer containing terminal transferase (0.5 U/mL) and 0.4 μM dUTP. Detection was enhanced with nickel.
Staining of the proapoptotic protein BAK was performed by using a primary antibody against a synthetic peptide that corresponded to amino acids 14–36 of the human BAK protein, as described previously.13 Distribution of BAK was evaluated in rectal crypts divided into three portions. Immunostaining was scored from 0 to 3 + for intensity and for frequency of staining.
Staining of the antiapoptotic protein Bcl-2 was performed by using a murine immunoglobulin G1 (IgG1) MAb (Bcl-2 124) raised against human Bcl-2 protein (Dako Inc., Carpenteria, CA) as the primary antibody. For negative controls, the sections were immunostained in the absence of primary antibodies. Distribution of Bcl-2 was evaluated in rectal crypts divided into three portions. Immunostaining was scored from 0 to 3 + for intensity and frequency of staining.
Cytokeratin AE1 staining
Staining for the cytokeratin AE1 was carried by using an MAb purchased from Signet Pathology Systems, Inc. (Dedham, MA), as described previously.14 A semiquantitative method was used to evaluate expression of AE1 proteins, as described previously,14 in which the crypt column was divided into 3 compartments, and the expression was scored from 0 to 3 +.
Vitamin D receptor staining
Vitamin D receptor (VDR) staining was performed by using a rabbit antiserum to a peptide mapped at the amino terminus of VDR of rat origin (Santa Cruz Biotechnology, Inc., Santa Cruz, CA). After deparaffinization and rehydration, 10% normal horse serum was added for 30 minutes. Next, slides were incubated in 1:50 dilution of the antibody to VDR at 4 °C overnight, incubated in 1% biotinylated horse antirabbit IgG (Vector Laboratories, Burlingame, CA) for 1 hour after washing 3 times with phosphate buffered saline (PBS), then incubated in 1% avidin-biotin-peroxidase complex reagent (Vector Laboratories) for 1 hour. After washing 3 times in PBS, the color was developed in the diaminobenzine solution under microscopic control, counterstained in hematoxylin, dehydrated in graded ethanol, cleared in xylene, and covered with Permount.
VDR brown nuclear staining was scored as the number and position of positive stained epithelial cells in the rectal mucosal crypt. These data were entered into a statistical program that has been used previously to analyze the distribution of proliferated cells and that provides the number and position of all positive cells.
Mucin and galectin-3 staining
Galectin-3 was detected by using the antigalectin-3 MAb TIB 166 (rat MAb; dilution, 1:1000), as described previously.15 To determine the expression of MUC5AC, which is expressed highly in human gastric tissues, antihuman gastric mucin MAb (HGM-45M1; dilution, 1:100; Novacastra Laboratories Ltd.) was used. The avidin-biotin peroxidase system (Vector Laboratories) was used to demonstrate these antigens.15, 16 Quantitative evaluation of mucin and galectin-3 staining was performed under × 10 light microscopy. The number of mucous glands that showed positive or negative expression of the antigen and the distribution of antigen-positive cells in flat mucosal crypts and in polyps were determined and expressed on a scale from 0.0 to 3.0 as follows: A score of 0.0 indicated < 2.5% positive cells, a score of 0.5 indicated 2.5–8.0% positive cells, a score of 1.0 indicated 8–20% positive cells, a score of 1.5 indicated 21–50% positive cells, a score of 2.0 indicated 51–75% positive cells, a score of 2.5 indicated 76–90% positive cells, and a score of 3.0 indicated > 90% positive cells.
All measurements were expressed as mean values ± the standard error of the mean. Paired t tests were used to compare the differences before and after treatment in polyps and rectal biopsy specimens. P values < 0.05 were considered significant. A one-way analysis of variance followed by the Turkey test was used to test for statistically significant differences between normal rectal epithelium mucosa and colorectal polyps in both the control group and the treatment groups P values < 0.05 were considered significant.
In total, > 2400 patients were screened prior to colonoscopy: Polyps were detected in approximately 400 patients, but only 21 patients could be enrolled into the study, including 12 patients for the treatment group and 9 patients for the control group. Originally, we planned to enroll 40 patients. The main reason for low enrollment was refusal on the part of patients and their physicians to participate.
One patient in each group dropped out of the study, so that 19 patients (10 females and 9 males) completed the study. The mean age was 58.5 years, and 17 patients from the total group were African American, reflecting the racial distribution of patients at the institution. Compliance, as determined by pill counts, was > 96% for the treatment group. No clinical side effects related to the medications were reported during the study, and there were no significant differences in the chemical safety monitoring of blood samples between baseline and study end. No polyps other than tissue marked with India ink were detected at study end in the approximately 60 cm of the distal colorectum examined by sigmoidoscopy.
At baseline, there were no significant differences between the two study groups in their nutritional intake. Energy intake was approximately 1900 kcal, fat intake was approximately 72 g, fiber intake was approximately 16 g, calcium intake was approximately 515 mg, and vitamin D intake was approximately 2.2 μg per day. Despite all the efforts by the nutritionists, at study end, treated patients consumed a significantly lower total energy intake, approximately 400 mg kcal less than at baseline (P < 0.02), and had significant reductions in carbohydrate, fiber, and vitamin D intake (vitamin D, 3.5–1.55 μg/day).
Polyp Histologic Data
At the start of the study, the mean size of the polyps examined was 7.0 mm ± 0.42 mm, and no significant differences were observed between the control group and the treatment group. The polyp tissue from the tattooed areas that were resected at study end were 3.42 mm ± 0.4 mm in mean size. At study start, it was observed that 4 of 19 polyps were tubulorvillous, and the remaining 15 polyps were tubular adenomas. At study end, the tissue removed showed 12 tubular adenomas, 1 tubulorvillous adenoma, and 1 hyperplastic plus tubular adenoma. In addition, histologic examination showed three hyperplastic polyps and two areas of normal mucosa, including one with a lymphoid follicle below.
Proliferation and Apoptosis Data
In the normal-appearing, flat mucosal biopsy specimens, proliferating cells were detected by Ki67 immunostaining and were evaluated by using standard criteria. Data were evaluated both for all patients who had tissue available at study end and for patients who had only adenomatous tissue available at study end. Table 1 shows that, at the study start, there were no significant differences in total crypt epithelial cell numbers or in the overall labeling index. However, patients in the treatment group had greater numbers of labeled cells per crypt column (P = 0.06) and a higher labeling index. Patients in the control group showed small but significantly greater fractions of proliferating cells in the upper 40% of the colonic crypt adjacent to the luminal surface (phi h) and numbers of proliferating cells in the upper 40% of the colonic crypt adjacent to the luminal surface (L h) levels at study start.
|Variable||Control group||Treatment group|
|Study start||Study end||P value||Study start||Study end||P value|
|All patients with tissue at study end||N = 8 patients||N = 8 patients||N = 10 patients||N = 10 patients|
|Labeling cells/column||3.60 ± 0.15||4.18 ± 0.33||0.09||4.14 ± 0.17||3.50 ± 0.14||< 0.001|
|Labeled index (%)||6.30 ± 0.21||6.72 ± 0.54||NS||7.61 ± 0.45b||6.16 ± 0.36||< 0.001|
|Phi h||0.12 ± 0.09||0.33 ± 0.25||NS||0.02 ± 0.01b||0.02 ± 0.02||NS|
|L h||0.13 ± 0.03||0.04 ± 0.01||0.04||0.06 ± 0.03b||0.03 ± 0.01||NS|
|Patients with adenoma at study end||N = 6 patients||N = 6 patients||N = 8 patients||N = 8 patients|
|Labeled cells/column||3.58 ± 0.20||3.96 ± 0.39||NS||4.26 ± 0.19||3.57 ± 0.17||< 0.001|
|Labeling index (%)||6.28 ± 0.29||6.59 ± 0.66||NS||7.86 ± 0.52b||6.36 ± 0.42||< 0.001|
|Phi h||0.16 ± 0.12||0.43 ± 0.33||NS||0.02 ± 0.01||0.03 ± 0.02||NS|
|L h||0.14 ± 0.03||0.04 ± 0.02||0.08||0.07 ± 0.03||0.04 ± 0.02||NS|
At study end, the control group showed a modest but insignificant increase in the labeled cells per crypt column, the labeling index, and in the phi h level when considering either all tissues examined at the study end or only patients who had solely adenomatous tissue at study end. In striking contrast, all patients in the treatment group showed a highly significant reduction in labeled cells per column and the labeling index (P < 0.001) (Table 1, Fig. 1A) in addition to the patients who had only adenomatous tissue at study end. The low phi h and L h levels at study start were unchanged at study end. Differences in the 2 groups between study start and study end are shown in Table 2. It was found that labeled cells and the labeling index were reduced highly significantly among patients in the treatment group.
|Variable||Control group||Treatment group||P value|
|Flat mucosa from all patients||N = 8 patients||N = 10 patients|
|Labeled cells/column||0.57 ± 0.29||− 0.64 ± 0.10||< 0.01|
|Labeling index||0.42 ± 0.47||− 1.45 ± 0.22||< 0.005|
|Phi h||0.21 ± 0.28||0.08 ± 0.02||NS|
|L h||0.09 ± 0.03||− 0.02 ± 0.03||NS|
|Flat mucosa from patients with adenomas||N = 6 patients||N = 8 patients|
|Labeled cells/column||0.38 ± 0.35||− 0.69 ± 0.12||< 0.005|
|Labeling index||0.31 ± 0.57||− 1.50 ± 0.27||< 0.005|
|Phi h||0.28 ± 0.38||0.01 ± 0.03||NS|
|L h||0.10 ± 0.05||− 0.03 ± 0.04||NS|
Data on proliferation in resected polypoid tissue are shown in Table 3 using a semiquantitative scale. At study start, there were no significant differences either in the staining intensity or in the frequency of proliferating cells in polypoid tissue between the control group and the treatment group. Furthermore, in the control group, there was no change in either parameter during the study. In contrast, patients in the treatment group showed a significant reduction in both the intensity (Fig. 2A) and the frequency of Ki67-positive cells in polypoid tissue during the study (Table 3) (intensity, P = 0.025; frequency, P < 0.02). Proliferating cells were distributed in clusters throughout the polypoid tissue, so that changes in geographic distributions within the tissue could not be determined.
|Variable||Control group (n = 8)||Treatment group (n=11)|
|Study start||Study end||P value||Study start||Study end||P value|
|Intensity of Ki-67 staining||1.94 ± 0.24||1.94 ± 0.27||0.96||2.23 ± 0.18||1.68 ± 0.12||0.025|
|Frequency of Ki-67 staining||1.79 ± 0.31||1.40 ± 0.22||0.33||2.23 ± 0.23||1.64 ± 0.18||0.02|
Apoptosis, as determined by TUNEL staining, showed a total of 2.87 ± 0.4 and 2.25 ± 1.1 positive cells per crypt in the control group and the treatment group, respectively, at study start. At study end, the numbers of crypt-positive cells in the 2 groups were 3.75 ± 0.8 and 3.32 ± 1.1, respectively. There were no differences between the groups or within the study in either group (Fig. 1B).
TUNEL-positive apoptotic cells also were seen scattered throughout polypoid tissue, often in small clumps. Apoptosis did not differ between study groups or between study start and study end (Table 4) (Fig. 2B). The number of TUNEL-positive cells per high-power field in different polyps varied from approximately 8.0 cells per high-power field to 10.5 cells per high-power field.
|Group||Study start||Study end||Study end vs. study start|
|Control group (n = 8)|
|Treatment group (n = 11)||10.53||9.95||0.58|
|P value (difference)||0.19||0.47||0.81|
BAK protein also was measured, again using a semiquantitative score for intensity and the percentage of positive cells in or near the surface of the polypoid tissue or separately, deep inside the polypoid tissue (Table 5). The intensity of BAK immunostaining did not differ at study start between the groups or in the control group between study start and study end. However, BAK immunostaining intensity fell significantly in surface cells but increased in internal cells during treatment (Table 5) (Fig. 2C,D). Furthermore, the frequency of positive cells did not differ between the 2 groups at study start or in the control group between study start and study end; however, in the treatment group, the frequency of BAK protein-immunostained cells increased from 3.2% to 20% in the interior of polypoid tissue. No differences in Bcl-2-immunostained cells were detected in flat mucosa or in resected polypoid tissue in either group.
|Study start||Study end||Study start||Study end|
|Control group (n = 8)|
|Treatment group (n = 11)|
Flat rectal mucosa showed that approximately 40–50% of epithelial cells showed VDR immunostaining (Table 6). Overall, immunostaining was greater in surface epithelial cells (approximately 65% of positive cells) with reduced staining toward the base of the crypt (approximately 40% of positive cells). Patients in the treatment group showed a small but significant increase in staining index (Fig. 1D) of approximately 10%, but no changes were seen in the control group (Table 6). This increase was expressed mainly in lower crypt epithelial cells. VDR immunostaining was greater in polyps (approximately 60–65% labeled cells) than in flat mucosa.
|Variable||No. of patients: Mean ± SEa||P valueb|
|Study start||Study end|
|Control group (n = 8)|
|No. of crypt cells||58.0 ± 2.7||57.6 ± 1.6||NS|
|Crypt VDR positive cells (no.)||28.1 ± 2.8||28.9 ± 2.4||NS|
|Crypt VDR positive index (%)||49.0 ± 4.0||50.0 ± 4.0||NS|
|Treatment group (n = 11)|
|No. of crypt cells||58.4 ± 2.5||57.4 ± 0.9||NS|
|Crypt VDR positive cells (no.)||24.0 ± 1.3||28.9 ± 2.1||< 0.05|
|Crypt VDR positive index (%)||44.0 ± 3.0||50.0 ± 4.0||< 0.1|
Rectal epithelium strongly expressed galectin-3 immunostaining, with > 80% positive cells. There was a significant trend toward an increase in galectin-3 staining of rectal mucosa in the control group over the course of the study (staining score, 2.44 ± 0.24 at baseline vs. 2.88 ± 0.08 at exit; P = 0.06), but this was not observed in the treatment group (Fig. 1C). There was less staining of galectin-3 in polyps than in flat rectal mucosa (Table 7). Galectin-3 staining of polyps increased in the control group (1.88 ± 0.18 at baseline vs. 2.44 ± 0.20 at exit; P = 0.08), but there was no increase (2.18 ± 0.18 at baseline vs. 1.95 ± 0.23 at exit; P value not significant) in the treatment group (Fig. 2E).
|Variable||Staining score (mean ± SE)|
|Control group||Treatment group|
|Study start||Study end||Study start||Study end|
|Galectin-3, rectal mucosa||2.44 ± 0.24||2.88 ± 0.08||2.64 ± 0.15||2.64 ± 0.14|
|Galectin-3, polyps||1.88 ± 0.18||2.44 ± 0.20||2.18 ± 0.18||1.95 ± 0.23|
|MUC5AC, rectal mucosa||0.0 ± 0.0||0.31 ± 0.13||0.09 ± 0.06||0.14 ± 0.14|
|MUC5AC, polyps||1.31 ± 0.23||1.50 ± 0.16||1.15 ± 0.24||0.50 ± 0.18|
Few rectal epithelial cells showed any staining with the HGM 45 M1 antibody (Fig. 3, Table 7), whereas strong expression of HGM 45M1 protein (staining score, > 1.0) was seen in 15 of 19 adenomas at the study start. Treatment with calcium and vitamin D reduced immunoreactivity of this protein in 6 of 11 patients; and, in 5 of those patients, no MUC5AC mucin could be detected after treatment. In the control group, 3 of 8 patients had an increase and 2 of 8 patients had a decrease in MUC5AC expression over the course of the study, but no patients in the control group were MUC5AC-negative at exit. At the study end, the treatment group had a lower MUC5AC staining score than the control group (staining scores, 0.50 ± 0.18 in the treatment group vs. 1.50 ± 0.16 in the control group; P = 0.001) (Fig. 2F), with no difference in the MUC5AC staining score at the study start (P = 0.64). In pairwise comparisons of baseline versus exit specimens, there was a significant decrease in the MUC5AC staining score in the treatment group (P = 0.02) over the course of the study, and there was no significant change in MUC5AC staining in the control group (P = 0.4).
The current study was associated with major problems in recruitment. Greater than 2400 patients were approached about participating prior to their colonoscopy, including approximately 2000 patients who had no adenomatous polyps detected or who had evidence of colorectal carcinoma or adenomas with high-grade dysplasia. Among the potential participants, approximately 90% declined entry either on their own volition or because of the advice of their physicians. This largely was because of the study design, which left residual adenoma in situ for 6 months. We planned to enroll 20 patients in each group, but only 21 patients were enrolled, and 19 completed the study. However, the demographics of the enrolled individuals did not differ from other patients treated at the institution.
At study end, only 14 of 19 tattooed specimens contained adenomatous tissues; in 1 specimen, no polypoid tissue was present; and 3 patients only had hyperplastic tissue detected. Whether this represented replacement of adenomatous tissue after injury with hyperplastic tissue remains unclear. It is possible that pluripotent cells within the mucosa may have led to the conversion of a neoplastic adenoma to a hyperplastic lesion.
Although patients in the control group showed no significant increase in proliferative kinetics in flat rectal tissue at the end of the study, the treatment group showed a highly significant reduction in labeled cells per column and in the labeling index in all patients and in the patients who showed only adenomatous tissue at study end. In the polyps, treated patients also showed a reduction in the immunologic intensity and frequency of proliferative cells that was not observed in the control group. Although reduced proliferation frequently has been detected previously in flat mucosa among patients at risk for colorectal neoplasia,1, 17 to our knowledge, altered proliferation in adenomatous polyps after the administration of a chemopreventive agent, such as calcium and vitamin D, has not been reported. Apoptosis determined by TUNEL staining showed no differences within the group or between study start and study end either in flat mucosa or in adenomatous tissue. Furthermore, BAK protein measured by immunohistochemistry did not change at study end in the control group, although it increased in internal polypoid tissue during treatment with calcium and vitamin D. Such enhanced expression of BAK protein can accompany increased apoptosis.18 Failure to demonstrate changes in apoptosis may be due to the small number of patients enrolled in the current study. No differences in Bcl-immunostained cells were detected, nor were differences observed in cytokeratin AE1 expression in rectal flat mucosa.
To our knowledge, VDR expression has not been evaluated previously in a clinical chemopreventive program. In the current study, approximately 50% of epithelial cells throughout rectal mucosal crypts showed VDR expression without a change in the control group at study end. However, the treatment group demonstrated a slightly lower VDR index and lower numbers of VDR-labeled cells per crypt column at study start that increased significantly after the calcium/vitamin D treatment. Whether this up-regulation of VDR reflects the administration of calcium or vitamin D is unclear.
The antibody to MUC5AC (HGM 45M1) detects a high-molecular-weight glycoprotein with O-linked carbohydrate. MUC5AC protein is encoded by a gene on chromosome 11p15.5 and is expressed in the surface epithelium of the normal stomach and in lung. This glycoprotein is not expressed in normal colorectal epithelial cells, but expression of MUC5AC has been reported in colorectal adenomas.19, 20 In the current study, there was no staining of MUC5AC in normal rectal epithelium, but strong expression of MUC5AC protein was seen in adenomas at the study start. Treatment with calcium and vitamin D reduced immunoreactivity of this protein in the majority of patients in the treatment group; and, in 5 of 11 patients, no MUC5AC mucin could be detected after treatment. In the control group, only 2 of 8 patients had a decrease in MUC5AC expression over the course of the study, and none were MUC5AC-negative at study exit. At the end of the study, there was a lower MUC5AC staining score in the treatment group than in the control group, so that the phenotype of the residual adenomas was more like that of normal rectal epithelium.
Galactin-3, which is an endogenous lectin that is specific for β-galactosyl carbohydrates, was expressed strongly in the majority of epithelial cells throughout rectal crypts. Significantly fewer epithelial cells in adenomatous tissue expressed the galactin-3 protein. In the control group, there was a trend toward increased galectin-3 expression in the residual polyps at the end of the study compared with baseline galectin-3 expression. There also was an increase in galectin-3 staining of normal-appearing, flat rectal mucosa. In the treatment group, however, there was no increase in galectin-3 staining over the course of the study. The apparent prevention by treatment with vitamin D plus calcium of further increases in galectin-3 expression is noteworthy, because galectin-3 expression in the colon was associated previously with the progression of colon carcinoma.15 However, because this observation was quite unexpected, the current study did not provide an opportunity to determine the mechanisms for this phenomenon.