Surrogate endpoint biomarker assays in phase II chemoprevention clinical trials.

Surrogate endpoint biomarker (SEB) assays carried out in rodent models have benefitted from large amounts of available colonic tissue, abundant well-aligned colonic crypts, and population groups with fairly uniform biological characteristics. In contrast, SEB assays in human colon studies have often been carried out on small groups of subjects, without the advantages inherent in rodent studies. Some factors that contribute to variations in human colon SEB assays include differences in genetic background, the extent and duration of previous colonic diseases, degree of previous chronic irritation to the colonic mucosa, the initial levels of nutrients ingested prior to the study, administration of large volumes of fluid prior to SEB measurement which induce hypermetabolic and then quiescent changes in the mucosa, failure to use strict morphologic criteria in counting colonic crypts, and availability of only a small number of crypts for analysis. Measurements of adenoma recurrence over short durations are limited by factors that include a large potential miss-rate of small adenomas, a window of observation with short duration which limits the stage of adenoma observed, and the consequent inability to measure mechanisms that a chemopreventive intervention is affecting in a different stage of adenoma development.

Genetic factors controlling inheritance of susceptibility to 1,2-dimethylhydrazine.

Reciprocal crosses were made between AKR/J, a 1,2-dimethylhydrazine (DMH)-resistant mouse strain, and SWR/J, a sensitive strain. The F1 hybrids were tested with DMH and methylazoxymethanol (MAM), two colon carcinogens. Either DMH (20 mg/kg body weight) or MAM (35 mg/kg body weight), a metabolic derivative of DMH, was injected weekly for 10 weeks. In each group of 35 mice, 10 were injected with tritiated thymidine (25 microCi) 1 week after the sixth injection of DMH and MAM for the evaluation of proliferative characteristics and the number of foci of dysplasia occurring in 325 microns of distal colonic mucosa. At 27 weeks after the first injection of the carcinogen, the colons of remaining mice were opened longitudinally and the number of tumors enumerated. Compared with DMH-treated mice, the number of foci of dysplasia per mouse, the percentage of tumor-bearing mice, the number of tumors per animal, and the number of tumors per tumor-bearing animal induced by MAM were severalfold higher. This would suggest the presence of a gene(s) repressing metabolism of DMH to MAM. Moreover, differences in response to the carcinogens were observed between the sexes. In contrast to males, females treated with both DMH and MAM had significantly greater numbers of tumors per animal, tumors per tumor-bearing mice, and a greater proliferative response with extension of S-phase cells to the upper third and luminal surface of crypts. Among males, those with the XAKR/YSWR heritage appeared more resistant than XSWR/YAKR males, particularly in their response to MAM. A twofold difference in the number of foci of dysplasia per mouse, tumors per animal, and the number of tumors per tumor-bearing animals was seen. Analyses of the response to DMH and MAM by F1 reciprocal hybrids of the AKR and SWR strains have shown a complex inheritance pattern governing susceptibility to DMH. Resistance to the carcinogen is provided by at least two specific repressor genes, one governing metabolism of carcinogen from DMH to MAM, and the other controlled by gender. Genetic factors contributed by the AKR female appear to convey additional resistance to male progeny, suggesting more than one gender-related gene.

Susceptibility to 1,2-dimethylhydrazine-induced colonic tumors and epithelial cell proliferation characteristics of F1, F2, and reciprocal backcrosses derived from SWR/J and AKR/J parental mouse strains.

Hybrid crosses were performed between SWR/J, a strain highly sensitive to 1,2-dimethylhydrazine (DMH), and AKR/J, a strain highly resistant to the carcinogen. F1 and F2 and reciprocal backcrosses were tested to determine if proliferative characteristics such as high activity, wide compartment (PC), and a large S-phase population in the middle third of crypts were linked to susceptibility and inherited as a dominant autosomal trait as was reported for DMH tumor response. A blend of resistant and sensitive tumor and proliferative characteristics was observed in the F1 and F2 crosses. A tumor incidence of 43.7% in the F1 and 52% in the F2 was obtained rather than the respective 100% and 75% expected frequencies. One week after the sixth injection of DMH, the incidence of focal areas of atypism (FAA) in the backcross to resistance (BCR) and the backcross to sensitivity (BCS) was the same (4.1 per FAA-bearing animal). This suggested that the response to the carcinogen was similar in both groups up to this point. Yet 20 weeks later, the BCR had a 7.3% tumor incidence, far lower than the 50% incidence expected. The BCS had an incidence of 98.6%, not unlike SWR frequencies and close to the expected 100% tumor incidence. Proliferative characteristics in the backcrosses, however, did not revert to parental levels. Instead, the labeling index (LI) or percentage of S-phase cells to total cells scored was significantly higher in the BCR than in the BCS (10.2% +/- 3.2% versus 8.1% +/- 2.2%, P less than 0.02). This study has shown that in crosses between these two strains (SWR/J and AKR/J), susceptibility to DMH-induced tumor is not inherited as a dominant trait. Neither are the proliferative characteristics of the colonic mucosa inherited in a simple Mendelian manner nor are the kinetic properties of the epithelial cells linked to DMH tumor susceptibility. It is suggested that the parental AKR/J strain may contribute a protective or resistant factor, that is, a repressor gene, which impedes the progression of carcinogen-induced foci of dysplasia to colonic neoplasia.

Differential susceptibility of inbred mouse strains forecast by acute colonic proliferative response to methylazoxymethanol.

The proliferative response of colonic epithelial cells to methylazoxymethanol (MAM) was followed in 1,2-dimethylhydrazine (DMH)-sensitive SWR/J, moderately resistant C57BL/6J, and resistant AKR/J strains. Untreated AKR mice had a significantly lower labeling index (L1) a shorter proliferative compartment (PC), and a smaller percentage of DNA synthesizing cells in the middle third of the crypts than did the SWR strain. SWR had the highest Ll, the widest PC, and the largest percentage of DNA synthesizing cells in the middle third of the crypts. C57BL/6 mice had characteristics that lay between the sensitive and resistant strains. Pooled data from 1 week after the fifth and sixth injections and 12 weeks after the first MAM injection revealed that extension of the PC had occurred in all strains, but it was only the DMH-sensitive SWR strain that showed extension of the PC to the upper third of the crypt, the greatest shift of proliferating cells to the middle and upper third of crypts, and the greatest increase in Ll. The AKR strain demonstrated these proliferative alterations least, and the C57BL/6 strain fell between them. This differential response to MAM among the strains mimicked that previously reported by us when DMH was investigated. The similarity in response of the colonic epithelial cells of each strain to either a direct-acting (MAM) or indirect-acting carcinogen (DMH) would support the concept that susceptability to this family of carcinogens is directly related to the genetically controlled indigenous proliferative characteristics of the colon.

Differential susceptibility of AKR, C57BL/6J, and CF1 mice to 1,2-dimethylhydrazine-induced colonic tumor formation predicted by proliferative characteristics of colonic epithelial cells.

The proliferative characteristics of the colonic mucosae of various mouse strains were examined to determine their value in forecasting the differential susceptibility of each to 1,2-dimethylhydrazine (DMH)-induced tumor formation. The control labeling index (LI) was highest (9.9 +/- 2.6) and the proliferative compartment (PC) widest in the DMH-sensitive CF1 strain, whereas the LI was lowest (7.3 +/- 1.3) (P greater than 0.01) and the PC shortest in the AKR resistant strain. The size of the PC and the LI for moderately resistant C57BL/6J mice lay between these values (8.5 +/- 1.1). Pooled data from 1 week after the fifth and sixth injections and 12 weeks after the first injection of six showed elevated LI in the distal colons of all DMH-treated mice. Distribution analyses of [3H]thymidine-labeled cells indicated extension of the PC (Stage II abnormality) in all strains with a shift of DNA-synthesizing cells from the major zone or lower third to the middle and upper thirds of the crypt (Stage II abnormality) occurring primarily in the CF1 strain. Cytotoxicity measurements at 6 hours and compensatory DNA synthesis at 3 and 4 days after DMH injection (20 mg/kg body wt) revealed similar relative response levels in CF1 and AKR mice, which suggested induction initially of like numbers of mutations. Indigenous conditions shown to influence the expression of colonic neoplasia are the level of DNA synthesis and the dimensions of the PC within the mucosa.