Prostaglandin Levels in Human Colorectal Mucosa

Abstract

Recent evidence suggests that nonsteroidal antiinflammatory drugs (NSAIDs) may prevent colorectal cancer. The mechanism of action of NSAIDs in chemoprevention is unknown but may be linked to their effect on mucosal prostaglandin levels. Levels of five major prostaglandin metabolites were measured by gas chromatography–mass spectrometry in biopsy specimens of flat rectal mucosa from four patients with familial adenomatous polyposis (FAP) before and after sulindac therapy and from five healthy individuals. The prostaglandin present at highest concentration in rectal mucosa from FAP and control subjects was prostaglandin E₂. The concentration of thromboxane B₂ alone was significantly elevated in FAP patients compared to controls (P = 0.016). In FAP patients treated with sulindac, all prostaglandin metabolite levels were significantly reduced compared to pretreatment levels (P < 0.05) except prostaglandin D₂ (P = 0.07). Prostaglandins D₂, E₂, F_2α, and 6-ke to-F_1α levels also were significantly reduced in FAP patients on sulindac compared to healthy controls (P < 0.05). However, interpatient heterogeneity of response to sulindac was evident with changes ranging from +19% to −89%, and the patient with the greatest reductions after sulindac developed colorectal cancer after 35 months of therapy. Sulindac treatment, at drug doses shown to regress colorectal adenomas in FAP patients, has heterogeneous effects on the level of major prostaglandins in their rectal mucosa and may not prevent colorectal cancer due to uncoupling of prostaglandin levels and carcinogenesis.

Several lines of investigation support the concept that nonsteroid al antiinflammatory drugs (NSAIDs), which inhibit the synthesis of prostaglandins, can prevent colorectal cancer (1). Specifically, NSAIDs inhibit cell growth in cell culture (2-4), decrease the multiplicity and incidence of colon tumors in carcinogen-induced murine models (5-9), decrease the relative risk of the incidence and mortality of colorectal cancer in users of NSAIDs in human epidemiologic studies (10, 11), and regress adenomas in patients with familial adenomatous polyposis (12-16).

The mechanism of action of NSAIDs in the chemoprevention of colorectal cancer is unknown but may be linked to the effect of these agents on mucosal prostaglandin levels. Earlier evidence suggests that prostaglandins may modulate tumor cell growth and facilitate tumor cell progression (17-19). Prostaglandins, particularly prostaglandin E₂ (PGE₂), are produced in increased amounts in human colonic tumor cells (20, 21). In animal models of colorectal cancer carcinogenesis, normal-appearing mucosa in cancer-bearing animals has increased levels of PGE₂ (22) compared to controls, and PGE₂ is higher in the tumor tissue compared to the mucosa in carcinogen-treated animals (23). In humans, tissue concentrations of PGE₂ are significantly increased in colorectal adenomas compared to flat mucosa in controls, and even greater levels were found in malignant polyps and cancers (24, 25). An increase of PGE₂ is also found in flat colorectal mucosa of patients with adenocarcinoma (24). On the other hand, prostaglandin levels are found to be independent of the effects of NSAIDs in studies of experiment colorectal carcinogenesis and growth of colorectal cancer cell lines (26, 27).

In patients with familial adenomatous polyposis (FAP), sulindac has been shown to result in regression of colorectal adenomas, but limited data exist on the influence of NSAIDs on colorectal mucosal prostaglandin levels (13, 28, 29). Therefore, to understand the effect of NSAIDs on a profile of colorectal mucosal prostaglandin levels, we utilized gas chromatography and mass spectrometry, a methodology not previously employed, to measure the concurrent levels of five major prostaglandin metabolites in biopsy specimens. Flat mucosa of patients with FAP before and after sulindac treatment and of healthy controls was studied.

MATERIALS AND METHODS

Subjects

The study population consisted of four Caucasian FAP patients with previous colectomy and ileorectal anastomosis who had adenomas in the rectum (mean age 37.4 ± 11 SD; two men, two women). Mucosal prostaglandin levels were measured in these patients immediately before the initiation of sulindac treatment and after therapy with sulindac 150 mg orally twice a day for three months. Polyp regression was noted in all patients after three months of therapy. However, one of the patients (patient 4 in Figure 1 below) developed rectal cancer while on sulindac 35 months after initiation of the drug. The control population was five otherwise healthy Caucasian individuals with normal colonoscopic examinations performed for the evaluation of abdominal pain or hematochezia or for risk of FAP (persons at risk for FAP had negative APC gene tests) (mean age 26.7 ± 15 SD; two men; three women).

Fig 1.

Prostaglandin levels (ng/mg protein) in colorectal mucosa. PGD₂, prostaglandin D₂; PGE₂, prostaglandin E₂; PGF_2α, prostaglandin F_2α; TXB₂, thromboxane B₂; 6-keto-F₁, 6-keto-prostaglandin F_1α. FAP patients 1–4 are noted. * P = 0.016 control vs FAP; ** P < 0.05 FAP vs FAP and sulindac; ^† P < 0.05 control vs FAP and sulindac. Controls are five otherwise healthy Caucasians with normal colonoscopic examinations. FAP patients are four Caucasian patients with previous colectomy and ileorectal anastomosis. Patient 4 developed colorectal cancer after 35 months of sulindac treatment.

All patients were prepared for the endoscopic procedure with a clear liquid diet and oral cathartic solution. Enemas that could influence mucosal biochemistry were not given. In each patient, six rectal mucosal biopsies were taken from flat mucosa 10–12 cm from the anal verge to minimize potential differences that might occur from specimens taken at different colorectal sites. Four rectal mucosal specimens were snap frozen in liquid nitrogen for prostaglandin analysis, and two were placed in formalin for histopathologic examination. Histopathology showed colorectal mucosa without adenomatous epithelium or aberrant crypt foci in all nine individuals.

Specimen Processing

Rectal mucosal biopsy specimens were analyzed for prostaglandin D₂ (PGD₂), prostaglandin E₂ (PGE₂), prostaglandin F_2α (PGF_2α), thromboxane B₂ (TXB₂), and the principal prostaglandin I₂ metabolite, 6-keto-prostaglandin F_1α (6-keto-F_1α). Briefly, specimens were rinsed with cationic-free Hanks' balanced salt solution (HBSS; containing 138 mM NaCl, 5 mM KCl, 4 mM NaHCO₃, 5.6 mM D-glucose, 0.3 mM Na₂HPO₄, and 0.3 mM KH₂PO₄). Samples were then manually homogenized in a glass microhomogenizer in 50 μl HBSS containing 1 mM CaCl₂. The homogenizer was then rinsed with 60 μl HBSS containing 1 mM CaCl₂. Of the combined 110 μl homogenate, 10 μl was used to determine protein concentration by the Bradford protein assay (Bio-Rad). The remaining 100 μl was then incubated at 37°C for 15 min, followed by the addition of 50 μl of deuterated prostaglandin standards. The sample was then equally divided. Following the addition of 250 μl acetone to each half, the specimens were dried under a steady stream of nitrogen gas. At the point of complete drying, 25 μl oximating reagent (2% O-methoxylamine HCl in pyridine) was added immediately to the sample. Each vial was then sealed with a Teflon-lined closure, and the contents were vigorously mixed with the use of a vortex mixer and left at room temperature for 12–16 hr for oximation of the carbonyl moieties of PGD₂, PGE₂, TXB₂, and 6-keto-F_1α. The samples were stored at −20°C until delivery to the mass spectrometry laboratory for quantitation of prostaglandin levels.

Determination of Prostaglan dins by GC-MS

Samples stored at −20°C were first brought to room temperature, followed by evaporation of the pyridine solvent of the oximating reagent under a nitrogen stream. Subsequent to evaporation of pyridine, the residue of each vial was treated with reagents for synthesis of the pentafluorobenzyl ester–trimethylsilyl ether derivatives of the prostanoid analytes for gas chromatography–mass spectrometric (GC-MS) analysis as described (30, 31). Briefly, a Finnagan MAT SSQ710 combined gas chromatograph–mass spectrometer (Finnigan MAT, San Jose, California) was operated under electron capture negative conditions. Characteristic M⁻, −181 fragmentions of each of the prostaglandin derivatives and those of the deuterate d prostaglandin analogs used as internal standards were monitored simultaneously. The levels of each of the prostaglandins in the samples were based upon the use of known quantities of deuterated prostaglandins employed as internal standards. Other instrument parameters for combined GC-MS analysis and commercial sources of reagents for derivatization have been reported previously (30, 31).

Statistical Analysis

Differences in prostaglandin levels between controls and FAP patients before sulindac treatment were compared by Mann-Whitney U test. Reductions in FAP patients after sulindac treatment as compared to baseline were compared by paired one-tailed Student's t test. Significance between groups was defined as P < 0.05.

RESULTS

Of the five prostaglandin metabolites measured (Table 1 and Figure 1), PGE₂ was present at the highest concentration in rectal mucosa from FAP patients before sulindac treatment and from controls. The concentration of TXB₂ was statistically elevated in untreated FAP patients compared to controls (P = 0.016), but the concentrations of PGD₂, PGE₂, PGF_2α, and 6KF_1α were not statistically different.

Table 1.

Prostaglandin Levels in Rectal Mucosal Biopsy Specimens^*

	Prostaglandin ng/mg protein			Baseline change (%)†
	Control	FAP	FAP & SUL
PGD2	1.541 ± 0.461	1.136 ± 0.182	0.729 ± 0.450c	−38.8
PGE2	5.355 ± 1.015	4.832 ± 0.454	2.930 ± 1.820bc	−32.3
PGF2a	3.345 ± 0.791	2.917 ± 0.881	1.618 ± 0.888bc	−43.6
TXB2	0.945 ± 0.042	1.673 ± 0.329a	1.005 ± 0.643b	−45.2
6KPGF1a	0.960 ± 0.482	1.184 ± 0.108	0.567 ± 0.307bc	−49.5

^*Results are mean ± SD. FAP is familial adenomatous polyposis; FAP & SUL is FAP patients treated with sulindac.

^aP = 0.016 FAP compared to control

^bP = 0.05 FAP compared with FAP treated with sulindac

^cP = 0.05 FAP treated with sulindac compared to control.

^†Percent change from baseline in patients before and after sulindac treatment.

In FAP patients treated with sulindac for three months, all prostaglandin metabolite levels except PGD₂ were statistically significantly lower than before therapy (P < 0.05); the concentration of PGD₂ level was marginally lower (P = 0.07). Mean PGD₂, PGE₂, PGF_2α, and 6-keto-F_1α concentrations in FAP patients treated with sulindac also were significantly reduced compared to those of healthy controls (P < 0.05) . These effects correlated with complete adenoma regression in all patients. Mean change from base line after sulindac therapy was found to decline for all prostaglandins (Table 1). When interpatient variation was considered, patient 1 (Figure 1) showed changes from base line ranging from +19% for PGD₂ to −34% for PGF₂. The greatest response to sulindac therapy occurred in patient 4 (range : −89% for PGE₂ to −82% for PGF₂), who subsequently developed an ulcerating flat rectal adenocarcinom a but no adenomatous polyps while on sulindac therapy for 35 months (Figure 2A and B).

Fig 2.

(A) Endoscopic view of the rectum in patient 4 at 33 months of sulindac treatment revealing flat erythematous lesion. Mucosal biopsies revealed tubular adenoma and fibroinflammatory exudate. No polypoid adenomas were present. (B) Endoscopic view of the rectum at 35 months of sulindac treatment. Ulce rating rectal adenocarcinoma shown without concomitant re ctal polypoid adenomas.

DISCUSSION

We examined a profile of prostaglandin metabolites in human colorectal mucosa. In contrast to other investigators who have utilized radioimmunoassay to determine colorectal mucosal prostaglandin levels (13, 28, 29, 32), we employed gas chromatography coupled with mass spectrometry. The accuracy of radioimmunoassay is dependent upon antibody specificity, and the limitation of this approach is evidenced by the wide variation of absolute values for prostaglandin levels between published studies. In contrast, our methodology allowed the convenient, simultaneous, and accurate measurement of a series of prostaglandins. Difficulties with metabolite stability are minimized by the use of internal standards employed in this technique . Examination of a mucosal prostaglandin profile may also have future importance in understanding which of these compounds might best serve as biomarkers of human colorectal carcinogenesis and as indicators of patient compliance with medication in clinical trials. More over, this approach may also be useful in testing compounds thought to act via pathways independent of prostaglandin synthesis.

In our study, PGE₂ was present in greatest concentration in rectal mucosa not only in those affected by FAP but also in healthy individuals, a finding not previously reported. Of note , a significantly higher level of TXB₂, but not other prostaglandins, was found in the flat mucosa of FAP patients compared to healthy controls. These findings contrast with those in animal models of colorectal cancer, in which carcinogen-treated rats have persistently higher levels of PGE₂ in normal flat mucosa compared to control animals (23). The significance of increased TXB₂ levels in FAP patients awaits further investigation.

We found that sulindac taken orally by FAP patients significantly decreased mean rectal mucosal levels of all the major prostaglandin metabolites except PGD₂ (although PGD₂ had a marginally significant decrease, P = 0.07). This finding is consistent with the work of Rigau et al (13), who measured PGE₂ and 6-keto-F_1α in flat colorectal mucosa of FAP patients before and after sulindac treatment and found a significant decrease in both prostaglandins after sulindac therapy. Winde et al (28) and Nugent et al (29) analyzed flat colorectal mucosal PGE₂ and PGF_2α and also noted a significant reduction in the PGE₂ after treatment. Importantly, in our study, sulindac administration significantly decreased levels of several prostaglandins (D₂, E₂, F₂, 6-keto-F_1α) below concentrations in healthy control subjects. These effects correlated with the regression of colorectal adenomas in the sulindac-treated FAP patients.

Some investigators have indicated that colorectal prostaglandin levels are not affected in patients taking similar NSAID doses that cause tumor regression. Although our data demonstrate a significant decrease in mean prostaglandin levels in rectal mucosa following sulindac treatment, substantial interpatient heter-ogeneity was noted. Patient 1 showed little or no reduction in several prostaglandin levels in the face of regression of adenomas and high medication compliance. Of even greater interest, patient 4, who had the greatest reduction in prostaglandin levels, developed a ulcerating flat rectal adenocarcinoma despite the absence of polypoid adenomas 35 months after starting sulindac. Other patients developing rectal cancer while on sulindac therapy have been reported (33, 34). In addition, investigators have questioned the importance of prostaglandin inhibition in colorectal chemoprevention. Ahnen et al (26) noted that the sulfone metabolite of sulindac (FGN-1) inhibited experimental colon carcinogenesis in the rat by a prostaglandin-independent process (26). Further investigation revealed reduced tumor incidence, multiplicity, and tumor burden in sulindac sulfone -treated rats compared to controls (36). Moreover, there was no decrease in PGE₂ or significant inhibitory effects on cyclooxygenase, lipoxygenase or phospholipase A₂ (36). These observations are supported by Hanif et al, who noted, in colorectal cancer cell lines, that growth inhibition was not associated with prostaglandin inhibition (27). Similar findings have been noted with sulindac sulfone in murine models of mammary carcinoma (37). The uncoupling of prostaglandin levels from carcinogenesis raises the possibility that other biochemical effects of NSAIDs are important. Sulindac has been shown to increase apoptosis without changing the proliferative rate as measured by immunohistoche mistry for proliferating cell nuclear antigen (35). Therefore, NSAIDs may have regulatory effects on the apoptotic pathway independent of prostaglandin synthesis.

Our results suggest that future work evaluating an array of chemoprotective compounds should include an analysis of prostaglandin levels to determine if the semetabolites are affected in unpredicted ways. Long-term studies are also needed to correlate the development of human colorectal neoplasia with mucosal prostaglandin levels on a longitudinal basis in patients randomized to prostaglandin inhibitory drugs and placebo.