Regional Differences in Normal and Cancer-Associated Glycoconjugates of the Human Colon 1,2,3

Transcription

1 Regional Differences in Normal and Cancer-Associated Glycoconjugates of the Human Colon 1,2,3 Robert S. Bresalier,4,5,6 C. Richard Boland,4,7 and Young S. Kim 4,8 ABSTRACT-Because inherent regional differences in colonic epithelium may determine the biologic behavior of tumors originating from different sites, and inasmuch as colonic epithelial cells secrete mucins that may reflect the state of cell differentiation, colonic goblet cell mucin was analyzed with the use of fluorescein isothiocyanate-conjugated lectins and fluorescence microscopy in normal fetal and adult mucosa and in cancers of the proximal and distal colon. In the adult proximal colon only the goblet cell mucin in the upper portion of the crypts was specifically labeled by the lectin Dolichos biflorus agglutinin (DBA), whereas this gradient was progressively lost distally; mucin in the upper and lower crypts of the sigmoid colon and rectum bound the label uniformly. In fetuses less than 22 weeks of age, DBA bound only to mucin in the crypts of the distal colon. Seven of 12 (58%) cancers originating from the proximal colon bound DBA, whereas only 2 of 23 (9%) from the distal colon bound this lectin (P<'005). Logistic regression analysis suggested that this difference may reflect the occurrence of larger tumors in the proximal colon. Regional differences in the binding of Ulex europaeus agglutinin (UEA-I) to nonneoplastic mucosa was similar to that found by others; predominant binding occurred in the proximal colon in the adult. No difference was noted for UEA-I binding to tumors of the proximal (8 of 12; 66.6%) versus distal (11 of 23; 48%) colon (P=.48). These findings may reflect regional differences in normal and tumor-related carbohydrate structures in mucin of the human colon.-jnci 1985; 75: The human colon is a complex epithelium-lined structure, which is the site of the second most common form of newly diagnosed cancer in the United States (1). The proximal and distal colon in both animals and humans differ in embryologic origin (2), physiologic function (3-5), morphology, and detailed histology (3, 6, 7). There may also be differences in the epidemiology, pathogenesis, and behavior of tumors which develop in the proximal versus distal colon (8-16). Although considerable attention has been paid in the literature to alterations in the colonic mucosa during carcinogenesis, regional differences are often ignored. This is unfortunate, because developmental or oncodevelopmental changes in one region may differ from those elsewhere in the colon. Significant differences have been noted between the proximal and distal colon in glycoconjugate composition and the expression of blood group substances in rodents (17-20) and humans (21-25). In addition, the expression of such glycoconjugates may appear in carcinomas where they cannot be demonstrated in normal adult tissues from given regions of the colon (24, 26-32). Because inherent regional differences between proximal and distal colon epithelium may be critical in the determination of the biologic behavior of tumors originating from those sites, we used the FITC-conjugated lectins DBA and UEA-I to study the mucins produced by normal 249 and neoplastic mucosa from various regions of the colon. Mucins, the major secreted glycoproteins of the colon, are densely glycosylated and are avidly bound by certain lectins. DBA has been previously shown to bind to mucin in the majority of normal colon tissues studied, but in relatively few carcinomas (33). Its binding was also decreased in the transitional mucosa adjacent to colon carcinomas. DBA appeared to bind preferentially to the mucin of mature goblet cells in the upper colonic crypt. This lectin was believed, therefore, to define a carbohydrate structure produced by differentiated cells in the colon. Although intriguing, detailed regional patterns of distribution of DBA binding were not analyzed at that time. In addition, such studies were not performed for the fetal colon. The current study was therefore undertaken. UEA-I has been employed by others for similar studies (24,32) and was used for comparison. MATERIALS AND METHODS Lectins.-FITC-DBA and FITC-UEA-I were used to label tissue sections. Lyophilized preparations of lectins (Vector Laboratories, Inc., Burlingame, CA) were reconstituted in PBS to a concentration of 0.1 mg protein/ml for use. DBA binds to exposed GalNAc residues, and UEA-I to fucosyl residues at the nonreducing terminus of the mucin carbohydrate chains. Tissue samples.-formalin-fixed tissue sections from 35 colon cancers and 35 specimens of histologically normal adult mucosa were studied. Specimens of human ABBREVIATIONS USED: DBA=Dolichos biflorus agglutinin; FITC= fluorescein isothiocyanate; GaINAc=N-acetylgalactosamine; PBS= phosphate-buffered saline; UEA-I= Ulex europaeus agglutinin. I Received October 29, 1984; accepted April 26, Supported by Public Health Service grants CA (to C. R. B.) and CA (to Y. S. K.) from the National Cancer Institute and by the Research Service of the Veterans Administration (VA). 3 This study was presented in part at the annual meeting of the American Gastroenterological Association in New Orleans, LA, in May G.1. Research Laboratory (151M2), VA Medical Center and the University of California, San Francisco, CA. 5 Address reprint requests to: Dr. Bresalier at G.1. Research Laboratory (151M2), VA Medical Center, 4150 Clement St., San Francisco, CA Recipient of a Research and Education Associate Investigatorship of the VA. 7 Current address: G.1. Unit, VA Medical Center, 2215 Fuller Rd., Ann Arbor, MI We thank Ms. Trish Harrington for preparing the manuscript. JNCI, VOL. 75, NO.2, AUGUST 1985

2 250 Bresalier, Boland, and Kim TABLE I.-FITe-DBA binding to normal goblet cell mucina Site in colon Cecum Ascending colon Transverse colon Descending colon Sigmoid colon Rectum No. of specimens Upper crypt Lower crypt a Numbers represent the range of binding in each region. Specific fluorescence was graded as O=no fluorescence, l+=trace, 2+= weak intermediate, 3+=strong intermediate, and 4+=intense specific fluorescence. Representative photographs comprise fig. 1. fetal colon were also obtained from 12 fetuses ranging in estimated age from 15 to 22 weeks. Twenty-two specimens of normal adult mucosa were obtained from 7 patients known not to have neoplastic disease (14 specimens were Patient No. Tumor location TABLE 2.-Tumors examined Patient's Modified Tumor blood type" Dukes stage" diameter, ern obtained as part of an instant autopsy kidney donor program, 5 as colonoscopic biopsy specimens, and 3 incidentally during surgical procedures). The remainder represented specimens of histologically normal mucosa from patients with colon cancers at least 6 em removed from the site studied. The number and distribution of samples are listed in table 1. "Proximal" colon was considered to include cecum, ascending, and transverse colon, whereas "distal" colon included descending colon and rectum. Colon cancer specimens were obtained after record review at the Pathology Department at the VA Medical Center, San Francisco, CA. The distribution and characteristics of these tumors are listed in tables 2 and 3. Tumor diameters were obtained from pathology reports. Patients' blood types were obtained from review of patient records and blood bank reports. Secretor status was not determined. Histology" Lectin binding" Distal colon #1 Rectum 0 C 2.2 MWD #2 Rectum 0 B 4.0 WD #3 Rectum 0 C 6.0 MWD #4 Rectosigmoid NT D 2.5 MWD + #5 Rectosigmoid 0 D 3.5 MWD + #6 Rectosigmoid A C 8.0 Colloid + #7 Sigmoid 0 D 2.5 WD #8 #9 Sigmoid Sigmoid 0 A D D WD PD #10 Sigmoid A D 3.5 WD #11 Sigmoid B B 3.5 MWD #14 Sigmoid A B 4.0 WD #12 #13 Sigmoid Sigmoid AB NT B D MWD WD.. #15 Sigmoid B B 4.0 WD + #16 Sigmoid 0 C 4.0 PD + #17 Sigmoid 0 C 5.0 PD #18 Sigmoid 0 C 5.0 MWD-PD + #19 Sigmoid NT B 5.5 MWD + + #20 Sigmoid A D 6.0 PD + 0 D 3.5 WD #21 Descending #22 Descending A B 6.0 WD #23 Descending Proximal colon NT C 8.0 PD + 0 D 7.0 MWD + A C 4.0 PD + #26 Ascending A C 5.0 WD + + #27 Ascending 0 C 5.5 PD #24 #25 Transverse Ascending #28 Ascending A B 9.0 PD + + #29 Ascending 0 D 10.0 Colloid + + #30 Cecum NT D 4.0 MWD + #31 Cecum A D 5.0 WD + + #32 Cecum 0 D 7.0 WD #33 Cecum 0 D 9.0 WD + #34 Cecum 0 B 9.0 WD + #35 Cecum 0 C 12.0 Colloid + + DBA UEA-l a NT = not tested. b See (41). cwd=well differentiated, MWD=moderately well differentiated, PD = poorly differentiated. d A specimen was said to be positive (+) when the mucin produced by greater than one-third of tumor glands was specifically labeled by the lectin (well-differentiated tumors) or when the mucin in greater than one-third of the high-power fields was so labeled (less differentiated tumors). -=negative specimen. lnci, VOL. 75, NO.2, AUGUST 1985

3 Regional Differences in Lectin Binding by Colon Mucins 251 TABLE 3.-Size of colon cancers studied Cancers Proximal colon All tumors (n=12) DBA+ (n=7) DBA- (n=5) Distal colon All tumors (n=23) DBA+ (n=2) DBA- (n=21) Mean size, cm ± SD 7.2± ± ± ±1.6 ( ) 4.1±1.4 All tissues were obtained according to guidelines approved by the Committee on Human Research of the University of California, San Francisco, CA (Committee approval No , effective April 26, 1979 through May 1982, at which time approval was gained under new regulations). Five-micrometer tissue sections were cut and stained with hematoxylin and eosin and periodic acid-schiff reagent to verify pathologic findings and document the production of mucin in normal tissues and tumors. Adjacent serial sections were cut for fluorescence microscopy studies. Fluorescence microscopy.-the paraffin-embedded tissue sections were serially rehydrated by two 2-minute washes in xylene, and 2-minute washes in 100,95, 80, and 70% ethanol followed by three 2-minute washes in PBS at ph 7.4. A solution of FITC-lectin was prepared in 0.05 M PBS (ph 7.4) at a concentration of 0.1 mg protein solution/m!. FITC-lectin (50 Jll) was applied to the hydrated sections, which were then incubated for 20 minutes, and the excess unbound FITC-lectin was rinsed off with a 5-minute wash in PBS. The tissue sections were mounted in Gelvitol (Monsanto Co., St. Louis, MO), ph 8, and covered with a glass cover slip. The lectin specificity was confirmed by incubating 50 Jll of FITC-lectin together with 50 Jll of the specific inhibitory sugar (0.2 M hapten sugars for each lectin) before application to a duplicate tissue section. Sugar specificity was demonstrated by the complete abolition of specific fluorescence Fetus No. on the tissue section. The prepared slides were examined with a Zeiss epifluorescence microscope (Carl Zeiss, Inc., New York, NY) and the labeling of the mucin was noted. In normal tissues, specific fluorescence in the upper and lower crypt regions was graded as absent (0) to intense (4+) (tables 1,4). For cancers, a specimen was said to be positive when the mucin produced by greater than one-third of the tumor glands was specifically labeled by the lectin (well-differentiated tumors) or when the mucin in greater than one-third of the high power fields was so labeled (less differentiated tumors). At least 20 fields (or the entire specimen) were examined. Statistical analysis.-statistical analysis for regional differences in lectin binding to tumors was performed with the use of the Fisher's exact probability test. Logistic regression models (34) for the occurrence of DBA and UEA-I binding were fitted with the use of SAS procedure Logist (35) to determine a possible contribution of size of tumor versus "sidedness." RESULTS TABLE 4.-FITC-DBA binding to fetal globetcell mucin a Normal Colon Adult colon: Specific fluorescence for binding to colon goblet cell mucin by lectins was graded semiquantitatively from absent (0) to intense (4+). In the proximal colon, binding by DBA to mucin in the upper crypt predominated, whereas little or no binding occurred in the lower crypt (table 1; figs. 1A, 1B). In contrast, there was a progressive loss of this gradient from proximal to distal colon; goblet cell mucin in the sigmoid colon and rectum bound the label more uniformly throughout the crypt (table I, figs. lc-le). The range of DBA binding in various regions of the colon is given in table 1. Although the intensity of specific fluorescence varied slightly between individual specimens, all specimens from a given region demonstrated a similar pattern. We did not note differences between specimens from kidney donors (instant autopsies), endoscopic biopsy specimens, surgi- Estimated Proximal colon Distal colon age, wk Upper crypt Lower crypt Upper crypt Lower crypt NS 2 15 (Cecum) 0 0 (AC) NS NS 5 15 NS (DC) (Rectum) 7 16 NS NS NS a Grading of specific fluorescence as in table 1 except that numbers are for each specimen and represent a range between crypts. NS = no specimen available; AC = ascending colon; DC = descending colon. JNCI, VOL. 75, NO.2, AUGUST 1985

4 252 Bresalier, Boland, and Kim cal specimens, and those taken from normal appearing mucosa distant from colon cancers. These regional differences were independent of blood group A activity as determined in our laboratory by immunohistochemical analysis with the use of monoclonal anti-a (36). UEA-I bound only to goblet cell mucin in the proximal colon; binding was predominant in the upper crypt. Fetal colon: Because of the regional differences in DBA binding to goblet cell mucin in normal adult colon mucosa, we tested the ability of DBA to bind to fetal colonic goblet cell mucin. Specimens were obtained from 12 fetuses ranging in estimated age from IS to 22 weeks. The distribution of DBA binding in these specimens is shown in table 4. DBA bound predominantly to goblet cell mucin of the distal colon; binding was slightly greater in the upper crypt (fig. 2A). In contrast to the adult colon, no binding was observed in the proximal colon in most fetal specimens (fig. 2B). We noted substantial binding in the proximal colon only in the single specimen from a fetus with an estimated age of 22 weeks. Colon Cancers Regional differences in specific lectin binding to the mucin of colon cancers are depicted in table 2. Seven of 12 or 58% of cancers originating from the proximal colon bound DBA, whereas 2 of 23 or 9% from the distal colon bound this lectin. This result was highly statistically significant (P=.003, Fisher's two-tailed exact probability test) and appeared independent of tumor histology and the patient's blood group. We did not note a significant difference for UEA-I binding between tumors of the proximal (8 of 12 or 66.6%) and distal (II of 23 or 48%) colon (P =.48). Because tumors of the proximal colon were significantly larger than those of the distal colon (table 3), we performed a logistic regression analysis to determine the contribution of size versus sidedness to the difference in DBA binding to neoplastic glycoconjugates in these regions. Logistic regression models (34) for occurrence of DBA and UEA-I were fitted with the use of the SAS procedure Logist (35). For UEA-I, occurrence was not significantly related to size or to size and side of colon jointly (P>.03). For DBA, however, occurrence was significantly related to size (P=.003). Furthermore, size and side of colon jointly being considered, occurrence was significantly related to size (P =.014) but not to side of colon (P=.20). A graphic check (34) of the model for DBA indicated that it was adequate for the data. DISCUSSION The lectin derived from DBA binds specifically to carbohydrate structures containing exposed, nonreducing GalNAc residues (33). Previous work from this laboratory (33) suggested that this lectin bound selectively to goblet cell mucin of well-differentiated cells in the upper colon crypts of normal adult mucosa, but only rarely to mucin in neoplastic tissues. The present study extends this work and demonstrates regional differences in the binding of this lectin to goblet cell mucin both in normal fetal and adult mucosa and in cancers of the proximal and distal colon. In the normal proximal adult colon DBA bound predominantly to goblet cell mucin in the upper crypts and showed little or no binding in the lower crypt regions. A progressive loss of such a gradient was evident from proximal to distal colon; more uniform binding was evident in the upper and lower crypts of the sigmoid colon and rectum. These differences could be due in part to differences in turnover rates of cells in these regions (37,38) with the addition of terminal GalNAc residues to the carbohydrate chains from more differentiated goblet cells. Previous studies have suggested that cellular and regional differences in lectin binding to mammalian small intestinal (38) and colon (l9, 33) epithelial cells might reflect cell differentiation as well as regional differences in cellular carbohydrate components. The binding of DBA to fetal goblet cell mucin differed from that in the normal adult colon. In addition, specific binding of this lectin occurred in a regional distribution dissimilar to the expression of another group of epithelial cell glycoconjugates, the blood group antigens. These glycoproteins and glycolipids are present throughout the colon in the embryo and fetus, but disappear from the distal colon in the adult (36, 39). By contrast, DBA bound to fetal goblet cell mucin in the distal colon in a pattern similar to that of the adult but did not bind in general to the proximal fetal colon. Only in the specimen from the oldest fetus studied, age 22 weeks, did substantial binding occur in the proximal colon. Thus unlike the blood group substances, the glycoconjugates, to which DBA binds, may only appear in the proximal colon with maturation. Although blood group A activity also depends on the presence of carbohydrate chains with terminal GalNAc residues, it appears that the glycoconjugate structure, to which DBA binds, differs from that which imparts blood group A activity. Further evidence for this was provided by studies that employed monoclonal anti-a to examine the distribution of blood group antigens throughout the colon (36). Regional differences were also demonstrated for the binding of DBA to cancers of the proximal versus distal colon. Seven of 12 (58%) cancers originating from the proximal colon bound DBA, whereas only 2 of 23 (9%) from the distal colon bound this lectin (P<.005). This difference appeared independent of tumor histology or of the patient's blood group; however, most of the cancers studied were advanced (modified Dukes C and D). Since tumors of the proximal colon tend to be larger than those of the distal colon, we performed a logistic regression analysis to define the contribution of tumor size to DBA binding to the mucin from colon cancers. From our data, it would appear that the difference in DBA binding to tumors of the proximal versus distal colon may reflect, at least in part, the occurrence of larger tumors in the JNCI, VOL. 75, NO.2. AUGUST 1985

5 proximal colon. How this relates to the biologic behavior of the tumors is, as of yet, unclear. In a previous study (40), several metastatic foci were shown to produce mucin that bound DBA. It is possible that DBA binds to a tumor-related carbohydrate structure that contains an exposed GalNAc residue, but is different from that present in normal mucosa. Our results with UEA-I are similar to those of Yonezawa et al. (24), with this lectin binding to mucin in crypts of the normal proximal, but not distal colon. These workers demonstrated more extensive binding throughout the crypt, which may be explained by differences in the concentration of lectin employed (l mg lectin/ml vs. 100 J.(g/ml). No difference was noted for UEA-I binding to tumors of the proximal versus distal colon. Tumors of the proximal and distal colon may differ in their epidemiology, pathogenesis, and behavior. Cancers of the proximal colon have increased in relative proportion over the past 2 decades (9-14). They occur more commonly in the setting of familial colon cancer (15) and have different patterns of metastasis from more distal cancers (16). There is also a relatively greater incidence of proximal colon cancer in population groups that are at low risk for colon cancer (11). Inherent regional differences between proximal and distal colon epithelium may be important in the determination of the pathogenesis and biologic behavior of tumors that originate from these sites. We provide further evidence for differences in the quantitative and qualitative expression of glycoconjugates in the human proximal versus distal colon both in normal and neoplastic mucosa. REFERENCES (1) SILVERBERG E. Cancer statistics, CA 1984; 34:7-23. (2) LEWIS FT. The development of the large intestine. In: Manual of human embryology. Vol 2. Philadelphia: Lippincott, 1912: (3) CHRISTIANSEN J. Mot.ility of the colon. In: Johnson LR, ed. Physiology of the gastrointestinal tract. New York: Raven Press, 1981: (4) FROMM M, HEGEL V. Segmental heterogeneity of epithelial transport in rat large intestine. Pflugers Arch 1978; 378: (5) SELLIN J, DESOIGNIE R. Rabbit proximal colon: A distinct transport epithelium. Gastroenterology 1983; 84:1304. (6) SHAMSUDDIN AK, TRUMP BF. Colon epithelium. I. Light microscopic, histochemical, and ultrastructural features of normal colon epithelium of male Fischer 344 rats. JNCI 1981; 66: (7) SHAMSUDDIN AK, PHELPS PC, TRUMP BF. Human large intestinal epithelium: Light microscopy, histochemistry, and ultrastructure. Hum Pathol 1982; 13: (8) BOYD D, LANGMAN M, DOLL R. The epidemiology of gastrointestinal cancer with special reference to causation. Gut 1964; 5: (9) WYNDER EL. The epidemiology of large bowel cancer. Cancer Res 1975; 35: (10) CORREA P, HAENSZEL W. Colon and rectum. Comparative international incidence and mortality. In: Schottenfeld D, ed. Cancer epidemiology and prevention-current concepts. Springfield, IL: Thomas, 1975: (11) SIDNEY S, FORGUHER JW. Cholesterol, cancer and public health policy. Am J Med 1983; 75: (12) VERNICK LJ, KULLER LH, LOHSOONTHORN P, et al. Relationship Regional Differences in Lectin Binding by Colon Mucins 253 between cholecystectomy and ascending colon cancer. Cancer 1980; 45: (13) NETSCHER DT, LARSON GM. Colon cancer: The left to right shift and its implications. Surg Gastroenterol 1983; 2: (14) MANAZZA J, GORDON PH. The changing distribution of large intestinal cancer. Dis Colon Rectum 1982; 25: (15) LYNCH PM, LYNCH HT. Hereditary proximal colonic cancer. Dis Colon Rectum 1977; 20: (16) WELCH JP, DONALDSON GA. The clinical correlation of an autopsy study of recurrent colorectal cancer. Ann Surg 1978; 189: (17) FREEMAN HJ, KIM Y, KIM YS. Glycoprotein metabolism in normal proximal and distal rat colon and changes associated with 1,2 dimethylhydrazine-induced colonic neoplasia. Cancer Res 1978; 38: (18) ESSNER E, SCHREIBER J, GRIEWSKI PA. Localization of carbohydrate components in rat colon with fluoresceinated lectins. J Histochem Cytochem 1978; 26: (19) FREEMAN HJ, LOTAN R, KIM YS. Application of lectins for detection of goblet cell glycoconjugate differences in proximal and distal colon of the rat. Lab Invest 1980; 42: (20) GOLDFARB JP, BRASITUS TS, Hsu KC, et al. Characterization of rat cecum, ascending and descending colon cell surface glycoconjugates by fluoresceinated lectins. Gastroenterology 1981; 80:1159. (21) FILIPE MI, BRANFOOT AC. Mucin histochemistry of the colon. Current Top Patho11976; 63: (22) SZULMAN AE. The histological distribution of blood group substances A and B in man. J Exp Med 1960; 111: (23) ---. The histological distribution of blood group substances in man as disclosed by immunofluorescence. II. The H antigen and its relation to A and B substances. J Exp Med 1962; 115: (24) YONEZAWA S, NAKAMURA T, TANAKA S, et al. Glycoconjugate with Ulex europaeus agglutinin-i-binding sites in normal mucosa, adenoma, and carcinoma of the human large bowel. JNCI 1982; 69: (25) WILEY EL, MURPHY P, MENDELSOHN G, et al. Distribution of blood group substances in normal human colon: Use of the unlabeled antibody (PAP) immunoperoxidase technique to identify A and B blood group substances. Am J Clin Pathol 1981; 76: (26) COWAN WK. Blood group antigens on human gastrointestinal carcinoma cells. Br J Cancer 1962; 16: (27) DENK H, TAPPEINER G, HOLZNER JH. Blood group substances (BG) as carcinofetal antigens in carcinomas of the distal colon. Eur J Cancer 1974; 10: (28) DENK H, TAPPEINER G, DAVIDOVITS A, et al. Carcinoembryonic antigen and blood group substances in carcinomas of the stomach and colon. J Natl Cancer Inst 1974; 53: (29) COOPER HS, HAESLER WE. Blood group substances as tumor antigens in the distal colon. Am J Clin Patho11978; 69: (30) ABDELFATTAH-GAD M, DENK H. Epithelial blood group antigens in human carcinomas of the distal colon: Further studies on their pathologic significance. JNCI 1980; 64: (31) WILEY EL, MENDELSOHN G, EGGLESTON Jc. Distribution of carcinoembryonic antigens and blood group substances in adenocarcinoma of the colon. Lab Invest 1981; 44: (32) SCHOENTAG R, WILLIAMS V, KUHNS W. The distribution of blood group substance Hand CEA in colorectal carcinoma. Cancer 1984; 53: (33) BOLAND CR, MONTGOMERY CK, KIM YS. Alterations in colonic mucin occurring with cellular differentiation and malignant transformation. Proc Nat! Acad Sci USA 1982; 79: (34) Cox DR. Analysis of binary data. London: Methoven, 1970: (35) SAS Institute. SUGI supplemental library users guide. Cary, NC: SAS Institute, (36) YUAN M, ITZKOWITZ SH, PAI.EKAR A, et al. Distribution of blood group antigens A, B, H, Le" and Le b in normal, fetal and malignant colonic tissues. Cancer Res In press. (37) LIPKIN M, BELL B, SHERLOCK P. Cell proliferation kinetics in the gastrointestinal tract of man. I. Cell renewal in colon and rectum. J Clin Invest 1963; 42: (38) LIPKIN M, SHERLOCK P, BELLB. Cell proliferation kinetics in the gastrointestinal tract of man. II. Cell renewal in stomach, ileum, JNCI. VOL. 75, NO.2, AUGUST 1985

6 254 Bresalier, Boland, and Kim colon and rectum. Gastroenterology 1963; 45: (39) SZULMAN AE. The histological distribution of the blood group substances as disclosed by immunofluorescence. III. The A, B and H antigens in embryos and fetuses from 18 mm in length. J Exp Med 1964; 119: (40) BRESALIER RS, BOLAND CR, KIM YS. Characteristics of colorectal carcinoma cells with high metastatic potential. Gastroenterology 1984,87: (41) BEAHRS OH. Colorectal cancer staging as a prognostic feature. Cancer 1982,50: FIGURE l.-fitc-dba binding to goblet cell mucin in crypts of the normal adult colon. In the cecum (IA) and ascending colon (lb) specific labeling occurs in the upper (arrowheads), but not lower (arrows) crypt regions. In the distal colon [descending colon (lc), sigmoid colon (ID), rectum (IE) labeling is progressively more uniform throughout the crypt. CEC=cecum; AC=ascending colon; DC=descending colon; S=sigmoid colon; R=rectum. X625 JNCI, VOL. 75. NO.2, AUGUST 1985

7 Regional Differences in Lectin Binding by Colon Mucins 255 jnci, VOL. 75, NO.2, AUGUST 1985

8 256 Bresalier, Boland, and Kim JNCI, VOL. 75, NO.2, AUGUST 1985

9 Regional Differences in Lectin Binding by Colon Mucins 257 lnci, VOL. 75, NO.2, AUGUST 1985

10 258 Bresalier, Boland, and Kim JNCI, VOL. 75, NO.2, AUGUST 1985

11 Regional Differences in Lectin Binding by Colon Mucins 259 FIGURE 2.-FITC-DBA binding to goblet cell mucin in crypts of the fetal colon. Specific labeling occurs in the distal (2A), but not in the proximal (2B) colon. X625 JNCI, VOL. 75, NO.2, AUGUST 1985

12 260 Bresalier, Boland, and Kim JNCI, VOL. 75, NO.2, AUGUST 1985