Intestinal mucin antigens in ulcerative colitis and their relationship with malignancy

Intestinal Mucin Antigens in Ulcerative Colitis and Their Relationship With Malignancy M. I, FILIPE,* A, SANDEY,* AND J. MA t The expression of two intestinal mucin-associated antigens large intestine muein antigen (LIMA) and small intestine mucin antigen (SIMA) were investigated by indirect immunoperoxidase staining of rectal mucosa from patients suffering from ulcerative colitis with (n = 6) and without (n = 31) associated carcinoma and in noncolitic controls (n = 40). The aim was to assess the relationship between antigen patterns and malignant change. SIMA, which is localised predominantly in the small intestine, is virtually undetectable in the normal adult colonic mucosa. However, this antigen is present in the foetal colon and colonic carcinoma. LIMA is expressed in normal colonic mucosa, but absent from the small intestine. LIMA staining patterns were not significantly different among the three groups. In contrast, expression of SIMA was significantly higher in the patients who had developed carcinoma (6/6) than in the noncancer group (7/71) (P < 0.001). The presence of SIMA was also significantly related to areas of dysplasia compared to normal (P = .03) or inflammation (P < .05), but it did not differ from mucosa showing "indefinite" atypia. The finding of 31% SIMA-positive biopsies associated with severe inflammation in colitis with active disease, but no evidence of malignancy, is difficult to explain at the present stage. A followup study would be necessary to determine its significance. Perhaps the most important finding is the increased frequency of SIMA-positive loci in histologically normal mncosa in carcinoma patients compared with the noncancer group (P < .00l), suggesting a field change. These observations may be prove useful for the identification of patients who may be at risk of developing carcinoma. HUM PATHOL 19:671--681. 9 1988 by W.B. Saunders Company. From the *Department of Histopathology, The United Medical and Dental Schools of Guy's and St Thomas' Hospitals, Guy's Campus, London Bridge, and tDivision of Laboratory Medicine, Royal Southern Memorial Hospital, Melbourne, Australia. Revision accepted for publication 17 August 1987. Supported by a grant from the Special Trustees of Guy's Campus, London, Address correspondence and reprint requests to Dr Filipe: Department of Histopathology, United Medical and Dental Schools of Guy's and St Thomas' Hospitals, Guy's Campus, London Bridge, London SE 1 9RT, England. Key words: Mucin antigens; ulcerative colitis; carcinoma. 9 1988 by W.B. Saunders Company. 0046-8177/88/1906-000655.00/0

Since the original paper of Morson and Pang 1 in 1967, the presence of dysplasia in large bowel biopsies has been the most commonly used indicator of a precancerous state in ulcerative colitis. However, this approach is known to have limitations. First, the patchiness of the lesion may lead to a false negative result, which can be minimized by multiple biopsies and endoscopy expertise. Second, absence of dysplasia in biopsies preceding the development of carcin o m a raises problems in the followup of these patients. 2-6 Thus, an additional functional criteria needs to be considered. Studies using carcinoembryonic antigen, IgA, and secretory component, 7-1~ and m o r e recently lectins, la'12 to distinguish dysplasia from inflammatory changes in ulcerative colitis, have not been successful in defining increased cancer risk in the individual patient. Altered mucin secretion with excess of sialomucins also has been reported in association with dysplasia and in apparently 'normal' mucosa in carcinoma fields. 13-15 However, its specificity and predictive value as an indicator of malignancy are low when the c o n v e n t i o n a l h i s t o c h e m i c a l m e t h o d s a r e employed. 16--17 This study was undertaken to establish the patterns of expression of two intestinal mucin-associated antigens SIMA and LIMA, ~8,m in the colonic mucosa of patients with ulcerative colitis and its relationship to the development of carcinoma. We were particularly interested in ascertaining whether the nonneoplastic colonic epithelium (normal and diseased) in ulcerative colitis patients who developed carcinoma differed from that of patients who did not, and from normal controls. Small intestine mucin antigen (SIMA) is a mucin glycoprotein originally isolated from mucus secreting colonic carcinoma with histochemical and biochemical evidence indicating that it is a nonsulphated mucin glycoprotein, is It is always present in normal

TABLE 1. Ulcerative Colitis With Cancer: Summary of Clinical Data and Macroscopic Findings in the Resected Specimens

Patient No. 1 2 3 4 5 6

Sex F M F M M F

Duration of disease (years) 40 22 31 39 14 24

Age at operation (years)

Carcinoma Operation

Grade*

67 40 45 68 49 69

Proctectomy Proctectomy Proctectomy Proctectomy Proctectomy Proctectomy

Precancer III II II III II

* I, II, and III: Duke's grades; well, moderately, and poorly differentiated, respectively. t A, B, and C: Duke's stage of invasiveness. Site in the rectal stump. Proximal, near ileorectal anastomosis; distal, near anal/rectal zone.

671

Stage?

Site~:

C B A B B

Multi focal Distal Distal Proximal Proximal Proximal

HUMAN PATHOLOGY

A nee-Rectal anastomosis

C,,] II

I t

i' ' i

FIGURE t. D i a g r a m of resected rectal stump [6) where

B

I

I I I I I I

0

1 2 025

No. Foci 2

2

2

5

I I I

1

Grades

No. Foci

mm I I I

l

I

I

C

0

404

03

mm

Ano-Rectal region

adult small intestine and has therefore been named "small intestine mucin antigen". SIMA has been shown to be an oncofetal antigen of the large intestine, because it is found in 8-12 week old fetal colon and absent thereafter and in the normal adult colonic mucosa, but is present in carcinomas of the colon (and adjacent transitional mucosa), stomach, gallbladder, a n d ovary. Large intestine mucin a n t i g e n (LIMA) is present in the normal large intestine mucosa and, like SIMA, it is also produced to a varying extent in carcinomas of colon and other parts of the digestive system and ovary. ]8-23 They appear to be distinct from previously reported antigens of normal and malignant gastrointestinal tissues.V24-27 MATERIAL AND

Grades 0

I I

nH

I

Volume 19, No. 6 [June 1988]

~ 3 4

1 1

tumour area [T] and mucosal strips taken are outlined. B and C represent sections of the mucosa, which is divided into zones (10 mm length] using a stage micrometer. Morphologic changes are plotted, grades 0 to 5, for epithetial atypia and counted. Any grade of epithelial atypia within each zone counts as one foci/area regardless of its frequency within the zone. Grades 0 to 5 for epithelial atypia [see text], mm, Muscularis mucosae.

trols) were studied using histologically normal rectal biopsies from 40 patients suffering from various large bowel conditions, ie, irritable bowel syndrome, diarrhoea which included postinfective Schistosomiasis and postcholecystectomy and postvagotomy etiologies, constipation, amyloid, etc. Ulcerative colitis and malignancy were excluded. T h e r e were 29 men and 11 women with a mean age of 40 years (table 2). Methods

All tissues were fixed in 10% formol-saline, routinely embedded in paraffin wax. 5-txm sections were cut and stained with haematoxylin-eosin stain. Eighteen representative tissue blocks (out of 30 total) from six patients in Group A, and all biopsies from control groups B and C, were stained with an indirect immunoperoxidase technique to demonstrate SIMA and LIMA. 2s Sections were trypsinized for 15 minutes before staining. Endogenous peroxidase was blocked with 1% H20 2 in methanol for 30 minutes, and normal rabbit serum, diluted 1:5 with PBS, was used to block crossreacting molecules. The monospecific antisera in this study were prepared by Dr Ma and described elsewhere. 18'19 The appropriate concentrations of the primary antibody (anti-LIMA and anti-SIMA) and the second layer of horseradish-conjugated antibody were determined by testing different dilutions on sections of normal colon and small intestine, known to stain for LIMA and SIMA, respectively. The former acts as a positive control for LIMA and negative for SIMA. The converse reaction is obtained in the small intestine.

METHODS

Patients

Material was obtained from three groups of patients: (A) Colitis with cancer was previously studied in six patients who underwent total colectomy and ileorectal anastomosis (IRA) for ulcerative colitis and later developed precancer or carcinoma in the retained rectum (table 1). Two or three strips of mucosa were taken along the entire length of the resected rectal stump, and each strip divided into several blocks (fig 1). 5 (B) Colitis without cancer (control) was studied in rectal biopsies obtained from 31 patients suffering from active-x6 or quiescent ~5 disease: There were 16 men and 15 women with a mean age of 49 years for both sexes and disease duration from 2 to 30 years or longer (table 2). (C) Noncolitic patients (con-

TABLE 2o Controls: Summary of Clinical Data a n d M o r p h o l o g y of Rectal Biopsies From Noncolitic a n d Colitic Patients Without C a n c e r

Patients (No.) Group B: Colitis without cancer ( 3 1 ) Group C: Noncolitic(40)

Histology* Active inflammation Non-inflammatory Normal

No. Men 6 10 29

* All biopsies show grade 0 accordingto Riddell et al, 1983. 671

No. Women 10 5 11

Age (years) 49 (28-72) 49 (22-74) 40 (18-79)

Duration of disease (years) 2-29 2-40 --

INTESTINALMUCIN ANTIGENS IN ULCERATIVECOLITIS [Filipe et al]

FIGURE 2. Hyperplastic rectal mucosa characterized by (top left) taller c~/pts lined by goblet cells secreting sulphomucins [darkcolored on HID-AB],(bottom) a normal pattern of LIMAexpression in the goblet cells, and (top right) absence of SlMA. (lop left) Approximate magnification x160', (bottom) indirect immunoperoxidase method; approximate magnification x 160; (top right) approximate magnification x120.

673

HUMAN PATHOLOGY

Volume 19, No. 6 [June 1988]

G r a d i n g of M o r p h o l o g i c Features

Morphologic features were defined according to the classification of cellular atypia in inflammatory bowel disease by Riddell et al. 29 as follows: negative, grade 0, indefinite, grades 1-3 (probably inflammatory, unknown, and probably dysplasia, respectively), and positive, grades 4 and 5 (low and high grades of dysplasia, respectively). In addition, the terms "hyperplasia" and "transitional" were included in grade 0. Hyperplasia is defined by taller crypts, sometimes branched and lined by increased numbers of larger goblet cells secreting predominantly sulphomucins. "Transitional" mucosa shows morphologic features similar to hyperplasia, but goblet cells contain an excess of sialomucins and intermediate cells between them are more conspicuous and produce sulphomucins. Intermediate cells refer to a relatively inconspicuous population of partially differentiated cells that in the normal colonic epithelium occupy the lower crypt. 3~ These show characteristics of both absorptive and goblet cells at the ultrastructural level, ~4'~1 and during the course of their differentiation, one or the other function normally predominates. T h e selection criteria for LIMA and SIMA demonstration of 18 of 30 tissue blocks from rectal stumps were based on the range of morphological atypia. 29 Furthermore, to avoid possible hybrid patterns of ileo-rectal transition, only blocks of distal rectal mucosa away from the anastomosis were included. Hyperplasia and transitional epithelium were characterised by their distinct patterns of mucin secretion, described in a previous study. 5 To map the morphologic features and the varying antigen expression in sections from the rectal stump material, we used a microscope stage micrometer and the areas/foci of change were noted and counted, as illustrated in detail in figure 1. For statistical analysis, X2 and Fisher's exact tests, as well as two tailed tests, were used. RESULTS Clinical Data a n d M o r p h o l o g y

For group A (ulcerative colitis with cancer), a detailed description of the gross appearances of the whole resected specimens has been reported. 5 In patient 1, where no invasive carcinoma was present, the rectal stump displayed several polypoid and plaquelike areas. In patient 5, the distal mucosa had a granular velvety texture, and the tumour, a poorly differentiated Duke's B adenocarcinoma, was located proximally close to the ileo-rectal anastomosis and surrounded by multifocal, plaquelike lesions. Histologically, a range of morphological changes from normal, hyperplasia, and "transitional", through the various grades of indefinite atypia and dysplasia, was seen in both of these patients. A close correlation was noted between the microscopic and gross appearances. In contrast, the rectal mucosa from the other four cases was pale, flat, and smooth. The carcinoma

FIGURE 3. Normal rectal mucosa from ulcerative colitis with carcinoma showing absence of SIMA(top) and a slightly altered pattern of LIMAstaining (bottom) in the luminal border, intermediate cells, and irregular distribution in goblet cells [indirect immunoperoxidase method; approximate magnification x160].

674

INTESTINALMUClN ANTIGENSIN ULCERATIVECOLITIS [Filipe et all

FIGURE 4. Abnormal patterns of LIMA expression. [Left) Localized predominantly in the supranuclear cytoplasma [? Golgi area] of goblet cells, and particularly intense in luminal border and apical cytoplasma of intermediate cells, The upper crypt retains a normal antigen profile in the goblet cell mucus. (Right) Presence of LIMA in the luminal border only. Indirect immunoperoxidase method; approximate magnification x320.

in the goblet cell mucus all along the crypt epithelium. It may be absent or at the bottom of the crypt and is not conspicuous in the luminal border (fig. 2). Abnormal patterns were characterized by decreased or absent reaction in the goblet cell mucus and the appearance of LIMA in intermediate cells between goblet cells. These intermediate ceils showed staining in the luminal border and apical cytoplasm or in the supranuclear cytoplasm (Golgi area) or, alternatively, the loss of antigenic expression. The changes may involve part or the whole crypt (figs. 3 and 4). SIMA is not expressed in normal large bowel mucosa, though it may occasionally be seen in goblet cells at the bottom of the crypt in patients in the control groups (figs. 2 and 3). The abnormal pattern is defined by the presence of the antigen in goblet cells occupying the lower third of the crypt, scattered along the crypt, or involving the whole crypt epithelium (figs. 5-8). Abnormal expression of both LIMA and SIMA can be focal (1-3 crypts/tissue section), patchy or extensive, and the change from a normal to

was flat and ulcerated in three of the cases and sited distally in two. Apart from the carcinoma, the mucosa was either normal, hyperplastic; or of "transitional" type. No dysplasia was found (table 1). Non-cancer Patients

Table 2 summarizes the clinical data and the histological features of the rectal biopsies from 31 cases of colitis without cancer (group B) and 40 cases of noncolitic patients (group C). In group B, 16 cases showed active inflammation, whereas the other 15 cases showed noninflamed mucosa, consistent with ulcerative colitis in remission. In all of the 31 group B patients and the 40 group C cases, the mucosa revealed no epithelial atypia, and was therefore classified as grade 0. Neither hyperplasia nor transitional features were observed in these 71 patients. Patterns of LIMA and SIMA Expression

In the normal staining pattern observed in the colonic mucosa (normal pattern), LIMA is expressed 675

HUMAN PATHOLOGY

Volume 19, No. 6 [June 1988]

FIGURE 5. Transitional mucosa with taller b r a n c h e d crypts showing a marked though patchy expression of SIMA in the goblet cells. To compare with hyperplasia in Figure 2. Indirect immunoperoxidase method; approximate magnification x120.

an abnormal pattern can be abrupt or gradually increasing in severity.

extensive SIMA staining at some point in the rectal mucosa compared with 2/40 (5%) noncolitic and 5/31 (16%) colitic with no evidence of malignancy (P < .001) (table 3).

LIMA and SIMA in Cancer v, Noncancer Patients

LIMA is expressed in goblet cells of normal adult colonic mucosa. Abnormal intracellular distribution was observed in the mucosa from both carcinoma patients (group A) and controls (groups B and C). Within group A, atypical staining with LIMA was more marked and more frequent in dysplasia (71%) than in those foci labelled as indefinite (43%) or grade 0 (33%, including normal, hyperplastic, and transitional epithelium), but these differences were not significant. LIMA staining in grade 0 epithelium did not differ significantly between carcinoma patients and controls. (fig. 9 and table 3). In contrast, all six patients with ulcerative colitis and carcinoma (fig. 9 and table 4) had intense and

SIMA and Epithelial Alypia in Ulcerative Colitis With Cancer

The expression of SIMA was related to increased severity of epithelial atypia, since it was observed in 12/30 (40%) grade 0 epithelium as opposed to 5/7 indefinite (71%) and 6/7 dysplasic areas (86%), but the numbers are too small for valid statistical significance. However, in this group of cancer patients, grade 0 includes three phenotypes. It is of interest to analyze separately the antigenic expression of normal, hyperplastic and transitional epithelium. A positive SIMA staining was significantly more common in transitional epithelium (100%) than in normal (23%) 676

INTESTINALMUCIN ANTIGENSIN ULCERATIVECOLITIS(Filipe et al)

FIGURE 6. Normal rectal mucosa (left) from ulcerative colitis associated with carcinoma showing (right) SIMApositive goblet cells scattered along the c~/pt. (Left) Hematoxylin-eosin stain, approximate magnification • (Right) Indirect immunoperoxidase method; approximate magnification x120.

- -

.

:/4

~

:

_'

" \.

;

~

FIGURE 7. Rectal mucosa showing (left) indefinitegrade epithelialalypia with slightlyenlarged and crowded nuclei, fewer goblet cells, and conspicuous intermediate cells.(Right) Expression of SIMA is seen in the cytoplasm and luminal border in two of the crypts. (Left) Haematoxylin-eosin stain,approximate magnification x160; (Rlght) Indirectimmunoperoxidase method; approximate magnification x160.

677

HUMAN PATHOLOGY

Volume 19, No. 6 [June 1988]

FIGURE 8. Rectal mucosa showing marked SIMA expression in (left) low grade and (right) high grade dysplasia, in ulcerative colitis associated with carcinoma, Indirect immunoperoxidase method; approximate magnification x160.

or hyperplastic foci (20%) (P = .004). In contrast, there was no significant difference in SIMA expression in transitional, indefinite, and dysplastic epithelium, probably reflecting a higher percentage of immature cells in these three categories. Thus, it seems valid to classify transitional epithelium as a separate grade. If so, positive SIMA phenotypes are significantly related with dysplasia (P < .01) and indefinite atypia (P = .05), compared with normal or hyperplastic epithelium.

ent from the incidence of SIMA in the noninflamed mucosa (colitis in remission 0/15, noncolitis 2/40, P = .03) and significantly lower to the incidence in dysplastic mucosa (6/7, P < .05) (table 3). Differences in SIMA staining between severely inflamed mucosa in colitis (5/16) and indefinite atypia in the cancer group (5/7) failed to reach statistical significance due to the relatively small number of subjects studied.

SIMA in Grade 0 Epithelium in Carcinoma v Non-carcinoma Patients [Table 3)

In assessing cancer risk in the individual patient with ulcerative colitis, importance has been attached to the presence of dysplasia. However, the knowledge that severe dysplasia is sometimes absent in endoscopic biopsies from colitic patients who later developed carcinoma should prompt us to pay attention to both morphologic and functional mucosal abnormalities that may antedate the development of malignancy. We demonstrated a gradation of aberrant distribution or loss of LIMA and the resurgence of an oncofetal antigen (SIMA) in all patients with ulcerative colitis and carcinoma. Changes in LIMA staining pattern tended to be more common in patients with carcinoma than in controls and more marked in dysplasia than in both

DISCUSSION

SIMA staining was more frequent (40%) in grade 0 foci (normal, hyperplastic, and transitional) in the six patients with ulcerative colitis associated with cancer than in patients without cancer, both colitics (5/31, 16%) and noncolitics (2/40, 5%) (P < 0.001). The incidence of SIMA staining was similar in both control groups as a whole (table 3). However, the presence of inflammation seems to influence SIMA expression, as reported below. Inflammation

I n colitis with active inflammation, SIMA was observed in 5/16 biopsies (31%). This result was differ678

INTESTINAL MUCIN ANTIGENS IN ULCERATIVECOLITIS [Filipe et al]

I ~

II

LIMA

SIMA

100 10O

O

90

~

8o

z m

60

~

so

FIGURE 9. Ulcerative colitis with c a n c e r : expression of intestinal mucin antigens in relation to morphology, Statistics refer to SIMA expression only. Dysplasia v [normal + hyperplastic], P < .01; indefinite v [normal + hyDerplasia], P = .05; and transitional v [normal + hyperplasia], P = .004.

4O

~

3o

0

NORMAL N=13 I

HYPERPLASIA TRANSITIONAL INDEFINITE N=10 N=7 N:7 t

't

'

p:

O.OOq

'

p: 0.05 p~[ 0.01

DYSPLASIA

N=7

Z

I

positive colitis would be in a high risk group for the development of carcinoma. Further studies with a larger number of patients and longterm followup are required to correlate SIMA expression both with activity and duration of diseases and with development of malignancy. The absence of SIMA in 14% o f dysplastic loci is consistent with the heterogeneity of SIMA expression observed in gastric and colonic carcinomas and may reflect cell lineage differentiation. We believe that the availability of new markers for colonic cell differentiation represents a novel approach to the definition of cancer risk, looking at functional atypia related to cell proliferation and impaired or aberrant cell differentiation rather than reliance solely on the morphologic features of dysplasia. In fact, the present study clearly demonstrates that morphologically normal epithelium (grade 0) from carcinoma patients reveals a significantly higher proportion of altered cell populations expressing SIMA than in the noncarcinoma group (table 3) (P < .001). This is not surprising, as neoplastic transformation occurs in a background of disorderly cell proliferation and differentiation. The presence of SIMA in cases 2, 3, 4, and 6, shown in table 4, is of interest in these four patients; atypia or dysplasia were absent in the rectal stumps but SIMA is expressed in the nondysplastic epithelium. Whether or not the resurgence of SIMA means that the tissue is invariably committed to malignant change is open to question. At the moment, we can only speculate that the probability of hitting SIMApositive foci in the absence of dysplasia in colitis associated carcinoma seems higher than in controls.

indefinite and grade 0 epithelium, but these differences were not significant. In contrast, expression of SIMA was significantly more common in dysplasia (6/7, 86%) than in morphologically normal biopsies from noncolitics and colitics in remission (2/55, 3.6%) (P < .001), or colitis with severe inflammation, (5/16, 31%) (P = .05), but not different from that observed in indefinite foci (5/7, 71%). T h e presence of SIMA in 31% of cases of colitis without evidence of malignancy may reflect an early stage in the process that may lead to carcinoma or, alternatively, the expression of SIMA in the colon may be related to cell proliferation and nonspecific. T h e s e patients in the colitis group showing SIMA-positive biopsies had disease duration of >9 years. It is known that the risk of developing carcinoma in ulcerative colitis increases after the first decade of disease. T h e r e f o r e , these five cases of SIMATABLE3. A b n o r m a l Expression of LIMA a n d SIMA in Normal Mucosa From C a r c i n o m a Patients a n d Controls

Histology LIMA SIMA

Ulcerative colitis with carcinoma, n = 30*

Ulcerative colitis, n = 167

Ulcerative colitis, n = 15?

Noncolitis, n = 40t

Noninflamm:~ 10 12

Inflamm 5 5

Noninflamm 1 0

Noninflamm 8 2

NOT~: P < .001. * No. o f foci in 6 patients. t No. o f biopsies. :~ G r a d e 0 a c c o r d i n g to Riddell et al 1983. It includes normal, hyperplastic, a n d transitional epithelium.

679

HUMAN PATHOLOGY TABLE 4. Histology patient no. 1

2 3 4 5 6 Total 6

Volume 19, No. 6 [June 1988]

SIMA Distribution in the Six Patients With Ulcerative Colitis and Carcinoma

Normal, n = 13

Hyperplasia, n = 10

Transitional, n = 7

Indefinite, n = 7

Dysplasia, n = 7

0/2 0/1 1/3 1/3 1/2 0/2

1/7 0 0/1 0 0/1 1/1

1/1 1/1 3/3 1/1 1/1 0

1/1 0 2/2 0 2/4 0

5/6 0 0 0 1/1 0

3/13

2/10

7/7

5/7

6/7

NOTE: Patients 2, 3, 4, and 6 had no dysplasia in the resected rectal stump and patients 2, 4, and 6 had no indefinite atypia. ABBREVIATION:n, number of foci.

In nondysplastic mucosa, SIMA is more frequently found in transitional than in normal or hyperplastic epithelium (P = .004). Although hyperplastic and transitional epithelium are morphologically similar and labelled grade 0 in Riddell et al classification, 2~ the latter has distinct histochemical features in common with those of immature colonic cells of the adult and foetal gut and neoplasia ]5'32-34 and we believe that the presence of transitional epithelium should not be included in the grade 0 classification. The resurgence of SIMA in adult colonic epithelium is highly atypical since SIMA is a glycoprotein normally produced by fetal gut. In adult life, SIMA persists in the goblet cells of the small intestine, but is absent in gastric and colonic epithelium, is'l~ It is reexpressed in intestinal metaplasia of gastric mucosa and in both gastric and colonic carcinoma. ~~ Recent studies with lectins and monoclonal antibodies to carbohydrate structures of glycoproteins during embryogenesis and oncogenesis support the concept that aberrant oligosaccharides in the glycoproteins of malignant cells are related to incomplete or inappropriate synthesis. This may be a result of incomplete glycosylation, availability of the substrates, or differences in glycosidase activities, 35-37 or it may be related to the reexpression of a native glycoprotein of the fetal gut. An increased population of immature and intermediate cells has been noted in foci of morphologically normal mucosa at various distances from colorectal carcinoma, 31 and it is possible that the altered p h e n o t y p e s described above in ulcerative colitish a r b o r i n g c a r c i n o m a r e f l e c t this shift in cell differentiation.S8 We conclude that markers for colonic cell differentiation, such as SIMA, may enable detection of altered cell populations in an unstable epithelium prone to malignant transformation. The findings of persistent abnormal expression of SIMA in multiple followup biopsies in ulcerative colitis may be useful for the identification of patients who may be at risk for developing cancer. This is being evaluated in our prospective studies. Acknowledgment. T h e a u t h o r s t h a n k Mrs R o s e m a r y B r o w n for statistical analysis. 680

REFERENCES 1. Morson BC, Pang LSC: Rectal biopsy as an aid to cancer control in ulcerative colitis. Gut 8:423, 1967 2. Cook MB, Goligher JC: Carcinoma and epithelial dysplasia complicating ulcerative colitis. Gastroenterology 68:1127, 1975 3. Lennard-Jones JE, Morson BC, Ritchie JK, et al: Cancer in colitis: Assessment of the individual risk by clinical and histological criteria. Gastroenterology 73:1280, 1977 4. Lennard-Jones JE, Morson BC, Richie JK, et al: Cancer surveillance in ulcerative colitis: Experience over 15 years. Lancet i:149, 1983 5. Filipe MI, Edwards M, Ehsanullah M: A prospective study of dysplasia and carcinoma in the rectal biopsies and rectal stumps of 8 patients following ileorectal anastomosis for ulcerative colitis. Histopathology 9:1139, 1985 6. Allen DC, BiggartJD, Pyrer PC: Large bowel mucosal dysplasia and carcinoma in ulcerative colitis. J Clin Pathol 38:30, 1985 7. Isaacson P: Tissue demonstration of carcinoembryonic antigen (CEA) in ulcerative colitis. Gut 17:561, 1976 8. Isaacson P: Immunoperoxidase study of the secretory immunoglobulin system in colonic neoplasia. J Clin Pathol 34:14, 1982 9. Rognum TO, Elgjo K, Fansa O, et al: Immuno-histochemical evaluation of carcinoembryonic antigen, secretory component, and epithelial IgA in ulcerative colitis with dysplasia. Gut 23:123, 1982 10. Allen DC, Biggart JD, Orcin JC, et al: An immunoperoxidase study of epithelial marker antigens in ulcerative colitis with dysplasia and carcinoma. J Clin Pathol 38:18, 1985 11. Boland CR, Lance P, Levin B, et al: Abnormal goblet cell glycoconjugates in rectal biopsies associated with an increased risk of neoplasia in patients with ulcerative colitis; early results of a prospective study. Gut 25:1364, 1984 12. Pihl E, Peura A, Johnson WR, et al: T-Antigen expression by Peanut agglutinin staining relates to mucosal dysplasia in ulcerative colitis. Dis Colon Rectum 28:11, 1985 13. Filipe MI, Branfoot AC: Abnormal patterns of mucous secretion in apparently normal mucosa of large intestine with carcinoma. Cancer 34:282, 1974 14. Filipe MI, Mughal S, Bussey HJ: Patterns of mucin secretion in the colonic epithelium in familial polyposis. Invest Cell Pathol 3:329, 1980 15. Filipe M I: Mucins of normal, premalignant and malignant coIonic mucosa. In Wolman SR, Mastromarino AJ (eds): Markers of Colonic Cell Differentiation, Vol 29, Progress in Cancer Research and therapy. New York, Raven, 1984 16. Ehsanullah M, Filipe MI, Gazzard B: Mucin secretion in inflammatory bowel disease. Correlation with disease activity and dysplasia. Gut 23:485, 1982 17. Ehsanullah M, Nauton-Morgan M, Filipe MI, et al: Sialomucins in the assessment of dysplasia and cancer-risk patients

INTESTINALMUCIN ANTIGENSIN ULCERATIVECOLITIS[Filipe et al]

18. 19. 20. 21. 22. 23. 24. 25.

26. 27. 28.

with ulcerative colitis treated with colectomy and ileo-rectal anastomosis. Histopathology 9:223, 1985 Ma J, de Boer WGRM, Ward HA, et al: Another oncofetal antigen in colonic carcinoma. Brit J Cancer 41:325, 1980 Ma J, de Boer WGRM, Nayman J: The presence of oncofetal antigens in large bowel carcinoma. Anst NZ J Surg 52:30, 1982 de Boer WGRM, Ma J, Nayman J: Intestine-associated antigens in ovarian tumours--An immunohistological study. Pathology 13:547, 1981 de Boer WGRM, Ma J, Rees JN, et al: Inappropriate mucin production in gallbladder metaplasia and neoplasia--an immunohistological study. Histopathology 5:295, 1981 Ma J, de Boer WGRM, Nayman J: Intestinal mucinous substances in gastric intestinal metaplasia and carcinoma studied by immuno-fluorescence. Cancer 49:1664, 1982 Filipe MI, Barbatis C, Sandey A, et al: Expression of intestinal mucin antigens in the gastric epithelium and its relationship with malignancy. HUM PATHOL 19:19, 1988 Goldenberg DM, Pant KD, Dahlman HL. Antigens associated with normal and malignant gastrointestinal tissues. Cancer Res 36:3455, 1976 Bara J, Paul-Gardais A, LoisilIier F, et ah Isolation of sulfated glycopeptidic antigen from human gastric tumours. Its localization in normal and cancerous gastrointestinal tissues. I n t J Cancer 21:133, 1978 Pant KD, Dahlman HL, Goldenberg DM: Further characterization of CSAp, an antigen associated with gastrointestinal and ovarian tumours. Cancer 42:1626, 1978 Nardelli J, Bara J, Rosa B, et ah Intestinal metaplasia and carcinomas of the human stomach: an immunological study. J Histochem Cytochem 31:366, 1983 Filipe MI, Lake BD: Immunohistochemical methods. Appendix 6. In Filipe MI, Lake BD (eds): Histochemistry in Pathology. Edinburgh, Churchill Livingstone, 1983

29. Riddell RH, Goldman H, Ransohoff DF, et ah Dysplasia in inflammatory bowel disease: Standardized classification with provisional clinical applications. HuM PATHOL 14:931, 1983 30. Kaye GI, Fenoglio CM, Pascal RR, et al: Comparative electron microscopic features of normal, hyperplastic and adenomatous human colonic epithelium. Gastroenterology 64:926, 1973 31. Dawson PA, Filipe MI: Ultrastructural and histochemical study of the mucous membrane adjacent and remote from carcinoma of the colon. Cancer 37:2388, 1976 32. Cooper HS: Peanut lectin binding sites in large bowel carcinoma. Lab Invest 47:383, 1982 33. Yonezawa S, Nakamura T, Tanaka S, et al: Glycoconjugate with Ulex europaeus agglutinin-l-binding sites in normal mucosa, adenoma and carcinoma of the human large bowel. J Natl Cancer Inst 69:777, 1982 34. Bara J, Burtin P: Mucus-associated gastrointestinal antigens in transitional mucosa adjacent to human colonic adenocarcinomas: Their "fetal-type" association. Europ J Cancer 16:1303, 1980 35. Kim YS, Boland CR, McIntyre LJ: Cell-surface glycoconjugates in differentiation and neoplasia of the colon. In Wolman SR, Mastromarino AJ (eds): Markers of Colonic Cell Differentiation, vol 29, Progress in Cancer Research and Therapy. New York, Raven, 1984 36. Feizi T, Gooi HC, Childs RA, et al: T u m o u r associated and differentiation antigens on the carbohydrate moieties of mucin-type glycoproteins. Biochem Soc Transations 12:592, 1984 37. Feizi T: Demonstration by monoclonal antibodies that carbohydrate structures of glycoproteins and glycolipids are oncodevelopmental antigens. Nature 314:53, 1985 38. La Mont JT, Ventola A: Galactosyltransferase in fetal, neonatal and adult colon: Relationship to differentiation. Amer J Physiology 235:E213, 1978

681