Familial colorectal cancer in Ashkenazim due to a hypermutable tract in APC.

Approximately 130,000 cases of colorectal cancer (CRC) are diagnosed in the United States each year, and about 15% of these have a hereditary component. Two well-defined syndromes, familial adenomatous polyposis (FAP) and hereditary non-polyposis colorectal cancer (HNPCC), account for up to 5% of the total new cases of CRC. Truncating APC mutations are responsible for FAP, and defective mismatch repair genes cause HNPCC. However, the genes responsible for most of the familial cases are unknown. Here we report a mutation (T to A at APC nucleotide 3920) found in 6% of Ashkenazi Jews and about 28% of Ashkenazim with a family history of CRC. Rather than altering the function of the encoded protein, this mutation creates a small hypermutable region of the gene, indirectly causing cancer predisposition.

Screening for colorectal cancer with fecal occult blood testing and sigmoidoscopy.

BACKGROUND: The high incidence of and mortality from colorectal cancer (160,000 new cases and 60,000 deaths in the United States each year) are compelling public health concerns. Following the evolution of effective surgery for this disease since the 1960s, the focus has been on improving methods of detection and integrating them into effective screening programs. PURPOSE: This was the first study to evaluate the effectiveness, in a setting of comprehensive medical examinations, of using the fecal occult blood test in conjunction with sigmoidoscopy, rather than sigmoidoscopy alone, to screen for colorectal cancer. Our end points were extent of compliance with fecal occult blood test and sigmoidoscopy, numbers of cancers detected, and mortality rate. METHODS: From 1975 through 1979, a total of 21,756 patients (aged 40 and older) who presented at the Preventive Medicine Institute-Strang Clinic for routine medical examinations were enrolled by calendar period into study and control groups. Study patients were offered annually both rigid sigmoidoscopy examinations and fecal occult blood tests requiring two stool specimens per day for 3 days, while control patients were offered only annual sigmoidoscopy. The majority of fecal occult blood test cards were not rehydrated before assay. Patients with positive tests were referred for double-contrast barium enema and colonoscopy. Two distinct trials were carried out. Trial I was primarily a demonstration of feasibility of using the fecal occult blood test as a supplemental screening method. Of the 9277 participants, 7168 (77%) were assigned to the study group and offered the fecal occult blood test. In trial II, approximately half of the 12,479 patients were assigned to each group. Patients in both trials had follow-up through 1984. RESULTS: Compliance with the fecal occult blood test was initially high in both trials, but diminished such that only 56% of study patients in trial I and 20% of those in trial II returned for second tests. On the initial (prevalence) screen, a substantial number of early-stage cancers were detected by the fecal occult blood test, primarily in trial II. In trial II, survival probability was significantly greater (P < .001) in the study group than in the controls (70% versus 48%), and colorectal cancer mortality was lower (0.36 versus 0.63) with borderline significance (P = .053, one-sided). CONCLUSIONS AND IMPLICATIONS: The screening of average-risk individuals (aged 50 and older) for colorectal cancer through use of the fecal occult blood test in conjunction with sigmoidoscopy can increase the likelihood of early detection of this disease. This practice, coupled with prompt diagnostic work-up following positive tests, will result in treatment of earlier stage cancers and increased survival after treatment.

Proliferation of esophageal epithelial cells among residents of Linxian, People's Republic of China.

Histopathologic and tritiated thymidine labeling subjects were carried out on esophageal biopsy specimens of 44 human subjects with cytologic evidence of dysplasia from Linxian, People's Republic of China, a high-risk area for esophageal cancer. With the use of histopathologic criteria, 10 cases showed evidence of dysplasia, 20 hyperplasia, and 14 a near-normal morphology when compared with 21 normal cases studied previously from Jiaoxian, a low-risk area for esophageal cancer in the People's Republic of China. Significantly increased labeling indices were found in the esophageal mucosa of the dysplasia and hyperplasia subjects. There was a gradient of increased expansion in the basal layer of proliferating cells progressing from normal to hyperplasia to dysplasia, with the expansion twice as high in the epithelial cell lining in dysplasia when compared with the findings in the normal and near-normal groups. The correlation of proliferative abnormalities with the severity of precancerous lesions of the esophagus indicates that labeling studies may provide a sensitive adjunct to evaluate risk status and any modifications that might result from nutritional intervention.

Reduced capacity for DNA repair synthesis in patients with or genetically predisposed to colorectal cancer.

Peripheral resting mononuclear leukocytes were compared for their capacities to repair DNA lesions induced by a 1-hour exposure to a standardized 10-microM dose of N-acetoxy-N-2-fluorenylacetamide (N-AcO-2-FAA). Leukocytes from the following 3 groups were studied: 39 control subjects, 40 patients after colonic resection because of colorectal cancer (disease-free at the time of this study), and 28 individuals with a hereditary predisposition to colorectal cancer. Although the level of N-AcO-2-FAA that bound to mononuclear leukocyte DNA was the same for the various population groups, the level of N-AcO-2-FAA-induced unscheduled DNA synthesis (UDS) was significantly reduced in the mononuclear leukocytes of individuals who had had colorectal cancer or a genetic predisposition for the disease. These findings indicate that a deficiency in mononuclear leukocyte DNA repair synthesis is associated with the development of colorectal cancer in these populations. Our observation of this nonspecific UDS deficiency (relating to colorectal cancer) was not explained by experimental variations among the sampled groups with regard to individual differences in lymphocyte heterogeneity, age, sex, smoking habits, or blood pressure.

Tyrosine phosphorylation of a Mr 63,000 protein induced by an endogenous mitogen in human colon carcinoma cells, but not in normal colonocytes.

Transformation of normal human colonocytes makes them sensitive to new mitogenic signals. Long-chain diglycerides (LCDGs) found in the human colon are mitogens selective for colon tumor cells, inducing mitogenesis in premalignant cells from each of 13 adenomas and in malignant cells from two of four carcinomas, but having no mitogenic effects on normal colonocytes (E. Friedman, P. Isaksson, J. Rafter, B. Marian, S. Winawer, and H. Newmark, Cancer Res., 49:544-548, 1989). Parallel to this biological activity pattern, LCDGs induce protein phosphorylation only in adenomas and carcinomas. Immunoblotting with an anti-phosphotyrosine monoclonal antibody demonstrated that the LCDG dimyristin, at concentrations found within the body, induced a 6-fold increase of tyrosine phosphorylation of an Mr 63,000 protein found in the particulate fraction of colon carcinoma cells. Tyrosine phosphorylation was maximal 0.5 min after addition of the LCDG, then fell, but remained elevated 40% over constitutive levels for at least 6 h. The Mr 63,000 tyrosine phosphoprotein was found in each of four colon carcinoma cell lines and an adenoma, but not in normal colonocytes, suggesting that the tyrosine kinase is activated only in tumor cells. Constitutive levels of the Mr 63,000 substrate were enhanced 2-fold by incubation of cells for 20 h with sodium orthovanadate, a tyrosine phosphatase inhibitor. This result suggested that carcinoma cells continually phosphorylate and dephosphorylate this tyrosine kinase substrate during growth. Thus, the colon tumor cell mitogen, dimyristin, utilizes a signal transduction pathway, containing the Mr 63,000 tyrosine kinase substrate, which is already in use during cell growth, possibly by other mitogens or growth factors.

A model for human colon carcinoma evolution based on the differential response of cultured preneoplastic, premalignant, and malignant cells to 12-O-tetradecanoylphorbol-13-acetate.

Small colonic adenocarcinomas can be found in focal areas within benign tumors (adenomas), strongly suggesting an adenoma-to-carcinoma sequence. The induction of plasminogen activator (PA) secretion by the tumor promoter 12-O-tetradecanoylphorbol-13-acetate (TPA) has been used to order histologically distinct classes of human colonic adenomas in primary culture into a sequence from the most benign to the most advanced premalignant state. This ordering is based on the observation that each of five carcinomas examined in an earlier study by E. A. Friedman (Cancer Res., 41: 4588-4599, 1981) and each of seven carcinomas tested in this study released PA in response to TPA, inducing easily scored morphological alterations. Benign tumors either resembled carcinomas in their response to TPA or exhibited no morphological changes. The most benign adenomas by histopathology criteria were the small pure tubular adenomas without dysplasia. Six of seven of these adenomas did not secrete PA in response to TPA. We concluded that malignant cells had acquired the ability to respond to TPA by PA secretion, while tubular adenoma cells were not an advanced enough preneoplastic stage to so respond. TPA treatment of two cultured villous adenomas, one with infiltrating carcinoma and one with focus of moderately dysplastic cells, in the presence of low serum to decrease the plasmin concentration, demonstrated that only a subpopulation of cells secreted PA. Local areas of the monolayer were morphologically altered by the protease, forming clusters of cells loosely attached to the dish. The presence of such subpopulations within cultured adenomas was demonstrated by screening an additional five villous adenomas, 15 villotubular adenomas, and 11 tubular adenomas. The presence of dysplastic cells in 23 of 24 cases correlated with PA secretion. A subpopulation of villous cells, in the absence of dysplastic cells in each of three cases, also secreted PA. We conclude that, during tumor evolution, this villous subpopulation is the first preneoplastic cell type to acquire responsiveness to TPA by PA secretion. This property is maintained as the cells further evolve through premalignant dysplastic stages to carcinoma.

Colonic epithelial cell proliferation in responders and nonresponders to supplemental dietary calcium.

Supplemental dietary calcium decreased and normalized hyperproliferation of colonic epithelial cells in individuals in familial colon cancer kindreds, measured by rates and patterns of [3H]thymidine labeling of epithelial cells in colonic crypts. In whole colonic crypts hyperproliferation was decreased to lower levels in over one-half of the subjects individually studied during the course of the calcium supplementation regimen. The remaining familial colon cancer subjects did not show reductions in cell proliferation measured over the whole crypt. However, when their cell-labeling data were analyzed in regions of the colonic crypt, the size of the proliferative compartment decreased and contracted towards the crypt base after calcium, a pattern typical of individuals at decreased risk for colonic cancer. This contraction of the proliferative region of the crypts occurred through decreased cell labeling in the two crypt compartments closest to the luminal surface and increased cell labeling in the second crypt compartment nearest to the base of the crypt. Following in vitro exposure of colonic epithelial cells to increasing physiological amounts of calcium, cell proliferation in familial colon cancer subjects decreased uniformly and greater heterogeneity in responsiveness was observed in cells from individuals with familial polyposis.

Actin cytoskeletal organization loss in the benign-to-malignant tumor transition in cultured human colonic epithelial cells.

The colonic epithelium in vivo is a highly indented sheet one cell thick. Culture methods have been developed to allow the normal cellular migration of the cells comprising this sheet to flatten it into a patch on the surface of a Petri dish [Friedman, E. A., Higgins, P.J., Lipkin , M., Shinya , H., and Gelb , A.M., In Vitro (Rockville), 17: 632-644, 1981]. Actin cytoskeletal organization was analyzed in such epithelial "patches" derived from several human colonic adenocarcinomas and their precursors, adenomas (benign tumors). The actin cytoskeleton was visualized by fluorescence microscopy after the fixed, permeabilized cells were stained with rhodamine-conjugated phalloidin. This drug has a very high affinity for actin filaments and a much lower affinity for monomeric actin. Actin organization was scored from 0 (no cables) to 5 points (extensive intercellular cable network). The phalloidin-stained actin found in seven adenocarcinomas had a predominantly granular fluorescence pattern with very little cable organization, scoring an average of 0.9 +/- 0.8 (S. D.). Three established cell lines derived from human colon carcinomas contained no cables by this analysis, scoring 0.0 +/- 0.0. In marked contrast, all 12 of the cultured adenomas had extensive actin cable networks, scoring an average of 4.3 +/- 0.4. There was no statistical difference between adenomas of differing histopathology class and malignant potential. However, cytoskeletons of plasminogen-activator-secreting "late-stage" preneoplastic cells from adenomas became disorganized by exposure to 12-O-tetradecanoyl-phorbol-13-acetate or another tumor promoter, teleocidin B. They scored, respectively, average actin organization values of 0.0 +/- 0.0 and 0.4 +/- 0.6. In contrast, nonplasminogen -activator- secreting "early-stage" preneoplastic cells from less advanced benign tumors were unaffected by 12-O-tetradecanoyl-phorbol-13- acetate or teleocidin B and retained extensive actin organization. Most, if not all, adenocarcinomas arise from preexisting preneoplastic adenomatous cells. Thus, loss of actin organization appears to mark the transition of noninvasive benign colonic tumors to invasive malignant tumors in humans. This transition is mimicked in vitro by exposure of certain "late-stage" preneoplastic cells to a tumor promoter which induces secretion of a plasminogen activator.

Inhibition of human colonic epithelial cell proliferation in vivo and in vitro by calcium.

Nine patients at high risk of developing colon cancer were placed on daily p.o. supplementation of 1500 mg of calcium for 4-8 weeks. The colonic epithelial cells in six of these patients showed a statistically significant decrease in their [3H]thymidine labeling indices in tissue culture so that they resembled those of patients at low risk of developing colon cancer. The three nonresponders had similar labeling indices before and after calcium supplementation. Biopsies from each of nine high-risk patients exhibited a decrease in proliferation when they were cultured in vitro with a high level of CaCl2 (2.2 mM compared with the 0.1 mM optimum value for proliferation). Two adenomas and two carcinomas showed a different pattern of response than normal cells, exhibiting no inhibition of growth at 2.2 mM CaCl2. These data indicate that the growth inhibition induced by high levels of extracellular calcium levels is lost at a stage in tumor development before cells become malignant.

Heterogeneous responses of human colon carcinomas to hexamethylene bisacetamide.

Primary cultures of resected human colon carcinoma were used to study differentiation agents directly on the biologically relevant cancer cells rather than on highly selected established cell lines. To achieve primary cultures which remained viable and replicating for several days, carcinomas were partly digested to epithelial organoids, which were selectively plated with high efficiency on collagen I-bovine serum albumin films in specially formulated serum-free medium. A monoclonal antibody, 29-15, was identified which binds to a cell surface epitope expressed on 16 of 21 invasive colon carcinomas of the Dukes' B2, C, or D histopathology classes, but not expressed on any of 11 noninvasive benign tumors (adenomas) at identical antibody titer. Noncytotoxic concentrations of the differentiation agent, hexamethylene bisacetamide (HMBA), induced the loss of the 29-15 epitope from HT29 colon carcinoma cells. HMBA also induced HT29 cells to lose the capacity for anchorage-independent growth with a similar dose-response curve and time course to the loss of 29-15 epitope. Twelve primary cultured human colon carcinomas exhibited differential responses when exposed to 1 to 7 mM HMBA for 7 days. Four moderately to well-differentiated carcinomas lost expression of the 29-15 epitope at each HMBA concentration. The tumor growth fraction was decreased in each tumor, with a mean decrease of 76% at 5 mM HMBA. A dose-dependent induction of nonproliferating tumor colonies, lacking [3H]thymidine labeling, occurred in three of the four carcinomas. In six other tumors, including those at less differentiated stages, HMBA induced the opposite effect: a two- to threefold increase in the tumor growth fraction at the optimal value of 5 mM HMBA, an increase in mean colony size, and no loss of the 29-15 malignancy epitope. No effects were observed in the two other carcinomas tested. Thus HMBA was able to induce growth arrest and loss of the malignancy epitope 29-15 in those carcinomas already at an advanced stage of differentiation, and to exert a growth stimulating effect on those carcinomas apparently at more immature stages.

New chemotherapeutic drug sensitivity assay for colon carcinomas in monolayer culture.

Ten previously untreated colon carcinomas were tested for chemotherapeutic drug sensitivity in primary monolayer culture. Colon carcinomas were partly digested to groups of epithelial cells which plated with a mean efficiency of 42 +/- 9% (SE) on a collagen I-bovine serum albumin substrate in serum-free medium, producing patches of tightly adherent epithelial cells. The cultured cells were judged epithelial by the presence of cytokeratins, an epithelial cell surface epitope, junctional complexes, and brush borders. Each carcinoma was plated in 40 to 60 Petri dishes (35 mm), yielding a mean of 28 +/- 8 (SE) colonies per dish (6832 +/- 1952 cells). Drugs tested in duplicate plates were mitomycin C, cisplatin, streptozotocin, and 5-fluorouracil at 0.1, 1, 10, and 100 micrograms/ml, and at 0.1, 1, and 2x the peak tolerated drug concentration in serum. Twenty-four h after plating, any nonadherent cells were removed, and the adherent tumor cells were continuously exposed to the drugs for 3 days. Each drug induced colony lysis in a dose-dependent manner in responsive tumors. Drug-resistant, cycling cells were identified by [3H]thymidine incorporation in colonies which were not lysed by drug treatment. Each of the ten carcinomas exhibited inherent resistance to one or more chemotherapy drugs within the concentration ranges clinically achievable.

Role of transforming growth factor beta 1 in induction of colon carcinoma differentiation by hexamethylene bisacetamide.

The differentiation agent hexamethylene bisacetamide (HMBA) increased expression of transforming growth factor beta 1 (TGF beta 1) mRNA in HT29 colon carcinoma cells. The increase was evident after 24 h and was maintained at levels 4-5-fold the control levels for at least 5-13 days. No increase in expression of TGF beta 2 or TGF alpha mRNA was observed. Both TGF beta 1 and HMBA induced loss of expression of a cell surface malignancy marker on HT29 cells, and both decreased cell growth in serum-free medium. Exogenously applied TGF beta 1 mimicked the growth-arresting effect of HMBA on three surgically resected moderately differentiated colon carcinomas in serum-free primary culture. Both TGF beta 1 and HMBA increased the tumor growth fraction in a second group of three more aggressive colon carcinomas, while neither agent had any measurable growth-modulating activity on two other colon carcinomas. The induction of TGF beta 1 mRNA by HMBA along with the parallel biological effects of HMBA and exogenously applied TGF beta 1 on resected carcinomas and on HT29 cells suggest that the effects of HMBA on colon carcinoma cells may be mediated in part by induction of TGF beta 1.

Fecal diglycerides as selective endogenous mitogens for premalignant and malignant human colonic epithelial cells.

Diglycerides (DGs) have been found in fecal extracts at concentrations which induce mitogenesis of adenoma and some carcinoma cells but not normal cells in primary culture. DGs containing stearic, oleic, palmitic, and myristic acid side chains were found in fecal extracts from each of eight subjects. Synthetic 1,2-DGs, containing the fatty acids found in endogenous fecal DGs, induced mitogenesis in cultures of premalignant cells from each of 13 adenomas, covering all histological classes, and in cultures from two of four carcinomas. The potent adenoma mitogen, dimyristin, had no mitogenic activity on cultures of normal colonic epithelial cells from seven different subjects. These results suggest DGs may act as endogenous mitogens in the development of human colon cancer. The extent of adenoma mitogenesis was correlated with the chain length of the saturated R-groups: 16 greater than 14 greater than 12 greater than 10 greater than 8 much greater than 18. DGs with oleic acid residues, C18:1, were among the most active, while substitution of even one fatty acid residue with a stearic acid residue, C18:0, reduced or eliminated mitogenic activity. Dimyristin also induced enhanced levels of urokinase secretion from carcinoma cells, in parallel to the phorbol ester tumor promoter, 12-O-tetradecanoylphorbol-13-acetate. These results imply that DGs found in the colon induce a selective growth of benign colonic tumors and some carcinomas, and may enhance the invasive capacity of carcinomas, while leaving normal cells unaffected.

Urokinase secretion from human colon carcinomas induced by endogenous diglycerides.

Colon tumor cells are more responsive to certain growth modulators in their local environment in vivo than are normal colonocytes. Examples of this class of compounds are the fecal diglycerides (DGs)(E. Friedman et al., Cancer Res., 49: 544-548, 1989), which may act as endogenous tumor promoters. At the concentration found in vivo, fecal DGs composed of oleic, myristic, and palmitic fatty acids induced mitogenesis of all classes of benign tumor cells and of half of the resected carcinomas tested in primary culture, but induced no detectable mitogenesis of normal colonocytes. Colon tumor cells also exhibit selective responses to these endogenous modulators as measured by another biological parameter, secretion of urokinase from carcinomas than from normal colonocytes. Fecal DGs also induced a 13-fold increase in urokinase mRNA synthesis in colon carcinoma cells and induced secretion of active urokinase from each of five resected carcinomas. Colon carcinomas, at both the primary site and metastatic to the liver, secreted the Mr 55,000 form of urokinase constitutively and secreted the same form upon treatment with fecal DGs. An increase in the steady-state level of urokinase secretion by saturated-chain DGs exhibited a strong dependency on the chain length of the fatty acid residues, those of 14 and 16 carbons having the greatest activity. Thus, fecal DGs composed of oleic, myristic, and palmitic acid residues induce two biological activities selectively in colon tumor cells, each of which would enhance tumor development. Selective mitogenesis would increase adenoma and carcinoma cell number relative to normal colonocyte number, and induction of the proteolytic enzyme urokinase would aid local invasion of the carcinoma within the bowel wall.

Expansion of the epithelial cell proliferative compartment and frequency of adenomatous polyps in the colon correlate with the strength of family history of colorectal cancer.

Expansion of the proliferative compartment of epithelial cells in colonic crypts and colonic adenomas have been described as phenotypic precursors to colon cancer in individuals affected with hereditary or sporadic colon cancer. This study measured the size of the proliferative compartment in colonic crypts and the frequency of adenomas in asymptomatic members of families having sporadic colorectal cancer. The subjects were divided into 2 groups according to the frequency of colorectal cancer in their families. A shift of the compartment of proliferating epithelial cells toward the lumenal surface of colonic crypts was seen in the group of subjects with a stronger family history of colorectal cancer, with significant differences in the numbers of proliferative cells in the upper and the lower crypt compartments (P < 0.05) and in the fraction of proliferative cells at the highest compartment at the lumenal surface of the crypts (P < 0.05). Cell proliferation patterns in normal-appearing mucosa of the 2 groups revealed no difference in whole crypt [3H]thymidine labeling index. Colonoscopic examination of the 56 subjects revealed an overall prevalence of adenomas of 21%; when stratified by frequency of colorectal cancer in their families, 3 of 22 subjects (14%) with a weaker family history had adenomas, while 9 of 34 (26%) with a stronger family history had adenomas. Thus, parallel abnormalities of colonic epithelial cell proliferation and neoplasia were seen in individuals with a family history of colorectal cancer, both of which were more pronounced with increasing strength of family history. This observation provides further evidence of relationships among these factors in the etiology of "sporadic" colorectal cancer.

Unscheduled DNA synthesis in mononuclear leukocytes from patients with colorectal polyps.

The mononuclear leukocytes from peripheral blood samples of individuals with (n = 30) and without (n = 48) colonic polyps were examined for their abilities to carry out unscheduled DNA synthesis (UDS) induced by N-acetoxy-N-2-fluorenylacetamide (N-AcO-2-FAA). Individuals with polyps had significantly reduced UDS values compared to the nonpolyp group (P less than 0.01). Furthermore, in a more comprehensive study, patients with hyperplastic polyps had N-AcO-2-FAA-induced UDS values not significantly different from control individuals who were asymptomatic and free from colonic disease as judged by complete colonoscopy. However, patients who had had adenomatous polyps in their large bowel had significantly reduced levels of N-AcO-2-FAA-induced UDS in their mononuclear leukocytes (P less than 0.005). When N-AcO-2-FAA binding to DNA determinations were made in parallel and DNA repair proficiency indices were calculated (i.e., N-AcO-2-FAA-induced UDS/N-AcO-2-FAA binding to DNA), the patients with adenomatous polyps were still shown to be deficient in carrying out DNA repair synthesis. Since adenomatous polyps of the large bowel are considered the premalignant lesion for colorectal cancer, we postulate that reduced UDS may be a genetically sensitive marker that is useful in studying the mechanisms of genetic predisposition to colorectal cancer.

Classification and risk assessment of individuals with familial polyposis, Gardner's syndrome, and familial non-polyposis colon cancer from [3H]thymidine labeling patterns in colonic epithelial cells.

A probabilistic analysis has been developed to assist the binary classification and risk assessment of members of familial colon cancer kindreds. The analysis is based on the microautoradiographic observation of [3H]thymidine-labeled epithelial cells in colonic mucosa of the kindred members. From biopsies of colonic mucosa which are labeled with [3H]thymidine in vitro, the degree of similarity of each subject's cell-labeling pattern measured over entire crypts was automatically compared to the labeling patterns of high-risk and low-risk reference populations. Each individual was then presumptively classified and assigned to one of the reference populations, and a degree of risk for the classification was provided. In carrying out the analysis, a linear score was calculated for each individual relative to each of the reference populations, and the classification was based on the polarity of the score difference; the degree of risk was then quantitated from the magnitude of the score difference. When the method was applied to kindreds having either familial polyposis or familial non-polyposis colon cancer, it effectively segregated individuals affected with disease from others at low risk, with sensitivity and specificity ranging from 71 to 92%. Further application of the method to asymptomatic family members believed to be at 50% risk on the basis of pedigree evaluation revealed a biomodal distribution to nearly zero or full risk. The accuracy and simplicity of this approach and its capability of revealing early stages of abnormal colonic epithelial cell development indicate potential for preclinical screening of subjects at risk in cancer-prone kindreds and for assisting the analysis of modes of inheritance.

Transforming growth factor beta 1 (TGF-beta 1) inhibits retinoblastoma gene expression but not pRB phosphorylation in TGF-beta 1-growth stimulated colon carcinoma cells.

The response of the retinoblastoma (RB) gene and its product (pRB) to transforming growth factor beta 1 (TGF-beta 1) was studied in three types of colon carcinoma cells derived from the same parental line. TGF-beta 1 was a growth inhibitor for two enterocytic-differentiated lines, a growth stimulator for two undifferentiated lines, and had no effect on two goblet cell-differentiated lines. TGF-beta 1 treatment for 3 days decreased RB gene expression and pRB level two- to threefold in each responsive line. When treated with TGF-beta 1 beginning in early G1, enterocytic cells were arrested in G1 and pRB remained under-phosphorylated and in low abundance. Neither goblet cell line exhibited these responses to TGF-beta 1 because they were shown to lack TGF-beta 1 type I and II receptors. Thus during colonocyte differentiation goblet cells lose responsiveness to TGF-beta 1 by down-regulating TGF-beta 1 receptors, while enterocytic cells retain and exhibit responsiveness to TGF-beta 1 through modulations of pRB. Both of the undifferentiated lines exhibited mixed responses to TGF-beta 1: a decrease in total amount of RB mRNA and pRB protein yet an increase in pRB phosphorylation consistent with increased cell cycling. Therefore, TGF-beta 1 controls RB function by two separable mechanisms, the regulation of pRB phosphorylation and the control of RB mRNA and protein level.

The capacity for growth stimulation by TGF beta 1 seen only in advanced colon cancers cannot be ascribed to mutations in APC, DCC, p53 or ras.

Human colon cancer development is associated with the accumulation of mutations and deletions in the suppressor genes DCC, APC and p53 and mutations in the dominant oncogene K-ras, with loss of wild type alleles. In earlier studies we had observed that about half of the resected human colon cancers placed into primary culture were growth stimulated by TGF beta 1. This group included the more advanced cancers which were either poorly differentiated primary-site cancers or metastases. In contract, the more differentiated colon cancers were inhibited or unaffected by TGF beta 1, indicating that a switch in response to TGF beta 1 occurs during colon cancer progression. Different sublines of the HT29 colon carcinoma cell line model the resected cancers, responding to TGF beta 1 by proliferation, inhibition or no growth modulation. The current study shows that while the poorly differentiated, TGF beta 1-stimulated sublines are most tumorigenic, all the sublines have the same spectrum of mutations: truncating mutations in both APC (adenomatous polyposis coli) alleles, no activated ras genes, mutated and thus overexpressed p53, and very low expression of DCC compared to normal colon cells. Genes other than the four already implicated in colon carcinoma evolution are responsible for the mitogenic response to TGF beta 1 found in the more advanced cancers.

Beneficial effect of human growth hormone on stress ulcers.

Human growth hormone was effective in healing erosions and controlling hemorrhage in six of eight patients with stress ulcers. This approach was based on known beneficial effects of growth hormone on nucleic acid and protein synthesis, demonstration of deleterious effects of stress on nucleic acid and protein synthesis, and demonstrated protective effects of growth hormone on animals subjected to stress. Bleeding cessation within 24 hours of its administration in two patients suggests the possible role of additional mechanisms involved in hemostasis. In a comparable group of eight patients with stress ulcer hemorrhage not treated with this agent, six died with continued bleeding. This high mortality represented the usual outcome in our patients with stress ulcer hemorrhage. These observations need to be extended to additional patients with stress ulcers in a larger randomized study.