Immunological Studies in Ulcerative Colitis VIII. Antibodies to Colon Antigen in Patients with Ulcerative Colitis, Crohn’s Disease, and Other Diseases
Scand J Gastroenterol Downloaded from informahealthcare.com by V U L Periodicals Rec on 12/29/14 For personal use only.
HANS ERIK CARLSSON, RUTGER LAGERCRANTZ & PETER PERLMANN Dept. of Immunology, Wenner-Gren Institute, University of Stockholm, and Dept. of Pediatrics, Karolinska Hospital, Stockholm, Sweden
Carlsson, H. E., Lagercrantz, R. &. Perlmann, P. Immunological studies in ulcerative colitis. VIII Antibodies to colon antigen in patients with ulcerative colitis. Crohn’s disease, and other diseases. Scand. J. Gastroent. 1977, 12, 707-7 14. Sera from patients with ulcerative colitis or Crohn’s disease had elevated titers to colon antigen from germ-free rats significantly more often than sera from patients with gastroenteritis, irritable colon, non-gastrointestinal diseases, and healthy controls. Elevated anticolon titers in significant frequency were also found in patients with liver cirrhosis, urinary tract infections,and in polyposis coli and their relatives.Females with ulcerative colitis had, on an average, higher titers than men especially in the age group 30 years and over. In Crohn’s disease the antibody titers often increased with time-as opposed to those in ulcerative colitis and non-gastrointestinaldiseases. In conjunction with results published earlier, the present work supports the assumption that the antibodies in ulcerative colitis patients react with antigenic determinants distinct from those recognized by the colon antibodiespresent in other groups, including patients with Crohn’s disease and polyposis. Key-words: Auto-antibodies; colon antigen; Crohn’s disease; polyposis coli; ulcerative
colitis P. Perlmann, Ph.D., Dept. of Immunology, Wenner-Gren Institute, University of Stockholm, S-106 91 Stockholm, Sweden
In sera from patients with ulcerative colitis autoantibodies against colon were demonstrated in the early 1960s (1 9). N o such antibodies were found in sera from healthy controls when antigen from human sterile colon was used (4, 5). However, when human conventional colon or colon from germ-free rats was used, a background titer was found in most sera from control subjects ( 18, 30). The present article reports the results from a retrospective study of 13 10 consecutive patients. A material of patients with polyposis coli and their relatives is also included. A cross-reaction between the Common Antigen (CA) present in most Enterobacteriaceae (6,8, 17) and the colon antigen has been demonstrated (15). It has, therefore, been proposed that autoantibodies against colon in ulcerative colitis may be
formed as a result of breakage of tolerance through cross-reaction between C A and autologous colon tissue (15). In this paper we also examine the reactivity with C A of sera from patients with Crohn’s disease, polyposis coli, and their relatives.
MATERIALS A N D METHODS
Patients. Between 1968 and 1973 the Department of Immunology at the Wenner-Gren Institute received sera for testing of autoantibodies against colon. During this time, sera from 13 10 in- or outpatients were obtained from hospitals all over Sweden. The diagnoses were checked in the case histories obtained from the hospitals. In 273 cases the case histories were incomplete or not obtained.
Scand J Gastroenterol Downloaded from informahealthcare.com by V U L Periodicals Rec on 12/29/14 For personal use only.
108
H . E. Carlsson, R . Lagercrantz & P. Perlmann
Diagnosis, sex, and age of the remaining patients are shown in Table I. Serum. A total of 1900 serum samples were investigated. In 83 cases three or more samples were obtained from the same patient. 280 samples were from healthy blood donors and laboratory personnel and 182 from patients with polyposis coli and their relatives. Sera were stored frozen in aliquots at -20 OC until used. Antigen. Colon antigen was derived from the feces of germ-free rats of the Swedish inbred strain (9, 10). The antigen was extracted with phenolwater and purified as described earlier ( 1 1, 18). E. coli strains 0 : 14 and 0: 75 were grown in synthetic medium ( 17) and the lipopolysaccharide (LPS) extracted with phenol-water (26) and purified as previously described ( 1 1). Passive hemagglutination. Passive hemagglutination was performed as described earlier ( 1 1). Packed and washed sheep erythrocytes (0.025 ml) were sensitized with 8 mg colon antigen, 0.12 mg E . coli 0: 14 LPS, and 1.0 mg E . coli 0:75 LPS, respectively, in 4 ml phosphate-buffered saline (PBS) pH 7.3. All sera were inactivated at 56 OC for 30 minutes and absorbed twice with equal volumes of washed packed sheep red blood cells and, if necessary, with human A, red cells. To ascertain reproducibility 3-5 sera of known titer were included in all experiments. All sera were titrated twice in independent experiments. The titers were reproducible within plus or minus one dilution step.
Hemagglutination inhibition. Hemagglutination inhibition tests were performed as earlier described ( 1 1, 17). For the inhibition experiments sera with titers >1/32 were chosen. To 0.1 ml of sera containing four to eight hemagglutinating units ot‘ antibody, inhibitor was added (0.002- 1 mg/ml). After incubation for 1/2 hr. the sensitized cells were added and hemagglutination recorded as usual ( 1 1).
RESULTS
Clinical data. Diagnosis, sex, and age distribution of the 1037 consecutive patients are shown in Table I. A majority (857) suffered from gastrointestinal Table 11. Incidence of elevated anti-colon hemagglutination titers (>1/16) and geometric means of titers in different patient groups Tot. no.
%> 1/16Geometric
mean
5 16 148 71 91 12 13
Ulcerative colitis Crohn’s disease Gastroenteritis Colon irritabile Salmonellosis Liver cirrhosis Polyposis coli Relatives of polyposis patients Urinary tract infections Non-gastrointestinal diseases Healthy controls
23.x 24. I 23.3
122
60 61 41 47 50 62 53
60 16
55 69
23.h
158 280
35 13
33.1
21.2 2.3.2 24.5 23.1
24.5
22.2
Table I. Major diagnosis with sex and age distribution of 1037 consecutive patients with sera submitted for testing of autoantibodies against colon Age range. years Diagnosis Tot. Male Female 0-14 15-29 30-44 45-60 >60 no. c 9 0 a’ Y S d \ U” Ulcerative colitis Crohn’s disease Gastroenteritis Colon irritabile Salmonellosis Liver cirrhosis Urinary tract infection Non-gastrointestinal diseases Total
516 148 71 97 12 13 16 158
261 66 36 41 4 6 3 16
255 82 41 56 8
79 106 35 4 2 1 5 2 1 9 1 2 1 3
13 82
101 82 17 2 1 8 8 17 17 2 2 0 0 3 7 34 25
1037
493
547
182 162
143 203
7
2
2
0
4 13
16
44 39 3 0 1 13 13 7 3 3 5 1 1 9 0 2 1 1 1 1 0 1 0 I I 14 10
19 5 7 8 1 2 1 16
7 0
84
59
81
55
9 1
3
2
1 0 2 0 5
8 0 2 0 14
18
36
Scand J Gastroenterol Downloaded from informahealthcare.com by V U L Periodicals Rec on 12/29/14 For personal use only.
Immunologicat Siudies in Ulcerative Colitis
disorders; of these 5 16 had ulcerative colitis (including 96 cases of proctitis) and 148 Crohn’s disease. Most patients with Crohn’s disease had lesions in colon and ileum (especially the distal part). Serological data. The incidence of hemagglutinating antibody titer 2 I / 16 against colon antigen in the different patient groups is given in Table 11. There is a statistically significant difference in the incidence of elevated titers (> I / 16) between healthy controls and non-gastrointestinal diseases, on the one hand, and ulcerative colitis and Crohn’s disease on the other (p 1/16) in ulcerative colitis with only rectal involvement (proctitis) and with colonic or colonic and rectal involvement (ulcerative colitis). The incidence in male (M) and female ( F ) patients within the groups is also given
%>I116
Sex
N
?h>1/16
420
60
96
59
M F M F
208 212 53 33
54 65 49
Tot.no.
I
Ulcerative colitis Proctitis
I
P p
0.05
p
0.05
12
i
titers in proctitis patients differed statistically from healthy controls (p (0.001) and non0 gastrointestinal disease (p ~ 0 . I). The incidence of elevated titers is significantly higher in females with ulcerative colitis. while no significant difference is seen between the sexes in
disease or the intestinal involvement and colon titer in ulcerative colitis and Crohn’s disease. The number of patients with relevant clinical data was. however. small and too heterogeneous to give conclusive results. There was a decrease of colon titer in female patients with ulcerative colitis or
Table IV. Incidence of elevated colon titers ( > I / 16) in male ( M ) and female (F) patients with ulcerative colitis, Crohn’s disease and non-gastrointestinal diseases
Ulcerative colitis Crohn’s disease Non-gastrointestinal disease
Tot. no.
Sex
N
?h> 1/16
5I6
M F M F M F
26 I 255 66 82 16 82
53 66 58 63 31 31
148 158
P p
0.01
p > 0.05 p
> 0.05
710
H. E. Carlsson. R . Lagercrantr & P. Perlmann
looj
(a) Utcerotive colitis
I
(516)
n
P
v 0,
0,
.-> .= 50ul
.->
n
.= 508 n
s
-$
Scand J Gastroenterol Downloaded from informahealthcare.com by V U L Periodicals Rec on 12/29/14 For personal use only.
0
r
3
, 1/ 16, similar to those observed in ulcerative colitis and Crohn’s disease. The difference from healthy controls and non-gastrointestinal diseases is statistically significant (p t0.001 and p 1/32), and E. coli 0:75 (>1/64), in patients with ulcerative colitis, Crohn’s disease, polyposis coli, relatives to polyposis and healthy controls
Tot. no. Ulcerative colitis Crohn’s disease
50 96
Polyposis coli
122
Polyposis relatives
60
Healthy controls
48
Sex
F M F M M F M F
N
Colon %>1/16
E. coli 0: 14 %>1/32
E. coli 0: 75 %>1/64
24 26 47 49 81 41 31 29 25 23
56 63 68 61 56 50 57 50 9 30
58 58 17 30 20 20 26 21 12 30
85 73 70 68 nd a nd * nd * nd a 76 83
* nd = not determined In 39 sera from patients with Crohn’s disease no inhibition of the reaction between four to eight hemagglutinating units of colon antibodies and colon antigen was observed with 0.002 to 1 mg/rnl of LPS from E. coli 0: 14 or E. coli 0:75.
DISCUSSION The sera were obtained from many different hospitals. All diagnoses were checked, and patients with incomplete case histories were excluded from the investigation. It is inherent in this kind of retro-
In earlier investigations using germ-free rat colon antigen, a hemagglutination titer of >1/16 has been shown to discriminate between ‘normal’ and ‘abnormal’ (14). When this titer limit is used, only about 15% of healthy individuals score as positive, whereas about 60% of patients with ulcerative colitis or Crohn’s disease have elevated titers. This study has confirmed that the incidence of titers >I/ 16 is significantly higher in patients with ulcerative colitis or Crohn’s disease than in healthy controls or non-gastrointestinal diseases ( 14).
Scand J Gastroenterol Downloaded from informahealthcare.com by V U L Periodicals Rec on 12/29/14 For personal use only.
112
H . E . Carlsson, R . Lagercrantz & P . Perlmann
Previous investigations have suggested that a few patients with solely distal engagement in ulcerative colitis may have a low incidence of elevated colon titers (14). The present work on a much larger material has shown that this is not the case. Gastroenteritis and irritable colon showed a relatively high incidence of elevated titers. It is, ‘however, significantly lower than in ulcerative colitis and Crohn’s disease, and significantly higher than in healthy controls. This result is not in complete agreement with earlier studies in which the titers were even lower and not statistically different from healthy controls (14). Our explanation for this discrepancy is that the present gastroenteritis and irritable colon groups are heterogeneous, possibly containing undiagnosed Crohn’s disease or ulcerative colitis. The high frequency of elevated titers seen in salmonellosis is in contrast with previous findings (14), but can be explained by the small number of patients in the present study. The higher incidence of elevated titers in females with ulcerative colitis previously observed (14) has been confirmed in this study. Such a difference was not observed in Crohn’s disease, at least not in patients younger than 3 0 years. A decrease of titers with age was observed in male patients but not in females with ulcerative colitis or Crohn’s disease. Thus, the difference in titers between males and females became more pronounced with age. No relation between clinical activity of disease and anticolon titer was seen in patients with ulcerative colitis or Crohn’s disease. This is compatible with previous investigations (12, 14, 30). The significant increase in titers with duration of Crohn’s disease has not been reported previously. In ulcerative colitis and non-gastrointestinal diseases, the frequencies of increasing and decreasing titers were equal, suggesting that the duration of the disease does not influence the titer values in these groups. Treatment with steroids slightly decreased colon titers in female patients but not in males. However, these patients constituted a relatively small and very heterogeneous group in that dose and duration of therapy varied widely. The suppressing effect on antibody levels is not unexpected, since the drug has an anti-inflammatory effect. The influence of steroids only on female patients’titers is most likely
a consequence of their higher average titer levels as compared to males. The relatively high incidence of elevated colon titers in the polyposis groups is interesting. Similarly to ulcerative colitis and Crohn’s disease, polyposis affects the colonic mucosa and an increased amount of mucus is secreted (1). Moreover, as in ulcerative colitis, the risk for malignancies is high ( I ) . The similarity between polyposis patients and their relatives in respect to antibody titers could be expected since the disease is dominantly inherited with a high degree of penetrance (20). The presence of high levels of colon antibodies in liver cirrhosis has been previously described (24). Raised levels of antibodies to many different antigens are seen in liver cirrhoses (3, 24, 25). These unusually high levels of antibodies may be produced as a response to antigens that should normally be trapped by the liver (23). It has also been suggested that non-specific stimulation results from exposure of immunologically competent tissues to mitogens like endotoxins, which should similarly be eliminated by the liver (23). E. coli 0: 14 contains high amounts of C A present in lower concentrations (or less active form) in most Enrerobacteriaceae (13). Previous investigations have shown significantly increased hemagglutination titers against E . coli 0: 14 LPS in patients with ulcerative colitis compared to healthy controls ( 14, 15, 22). It has therefore been assumed that this antigen cross-reacts with the colon antigen (15). Such cross-reactions have also been demonstrated after immunization of rats with heterologous colon antigen (18) and in hemagglutination inhibition with sera from patients with ulcerative colitis (15). Possibly CA, with which most individuals come into contact, may break the tolerance to autologous colon antigen in certain individuals and thus evoke auto-antibodies (15). The unusually high frequency of colon antibodies seen in patients with urinary tract infections may also be ascribed to this cross-reaction with CA. Such infections are frequently caused by gramnegative bacteria, notably E . coli, and elevated titers against C A in serum from patients with urinary tract infections have been noted by several investigators (2, 7, 21, 27). LPS from E. coli 0:75, which contains low
Scand J Gastroenterol Downloaded from informahealthcare.com by V U L Periodicals Rec on 12/29/14 For personal use only.
Immunological Studies in Ulcerative Colitis
amounts of CA, was used as control antigen. As was to be expected (15), there were no differences in the titers to this antigen between patients and controls. As previously shown, patients with Crohn’s disease have titers to E . cofi0: 14 similar to healthy controls (14). This disagrees with the results of Thayer et al. and Eckhardt et al. (8, 22). These authors found that a small number of patients with Crohn’s disease had an increased incidence of elevated antibody titers against E. coli 0: 14,significantly higher than healthy controls. The discrepancy between our results and theirs is hitherto unexplained. The reaction between colon antigen and serum from patients with ulcerative colitis can be inhibited with LPS from E. coli 0: 14 in about 30% of the cases, while there was no inhibition with LPS from E. cofi 0:75 (15). When the same antigen and experimental procedure were used, no inhibition was established with serum from patients with Crohn’s disease using LPS from E . cofi 0: 14 or E . cofi 0:75. These results, together with hemagglutination data, suggest that colon antibodies in Crohn’s disease, polyposis coli, relatives to polyposis, non-gastrointestinal diseases and healthy controls are of a specificity different from that of colon antibodies in ulcerative colitis. Similar differences in antibody specificity have been demonstrated in experimental immunization, as already mentioned ( 17). Statistically significant differences in the frequency of elevated hemagglutination titers to colon between groups of patients with ulcerative colitis and Crohn’s disease and other diseases were demonstrated. The wide distribution of individual titers may, however, limit the use of the test as a diagnostic tool. This limitation is probably due to the complexity of the antigen preparation used. It is a mixture of several similar bloodgroup-like glycoproteins of varying size and charge, containing 6080% carbohydrate (1, 28 and unpublished). The carbohydrate consists of a large number of oligosaccharide side-chains with different terminal sugars and of varying lengths and branching (29, and unpublished). This heterogeneity most likely results in a variety of antigenic determinants. Accordingly, individual patients or groups of them
71 3
may have elevated titers to different antigenic determinants. Further work to elucidate and differentiate the antigenic determinants of diagnostic importance on rat colon antigen is needed. ACKNOWLEDGEMENT The skillful technical assistance of Miss YlvaSjovall is gratefully acknowledged. The authors are also indebted to Mr. Ulf Brodin for valuable discussions and advice concerning the statistical evaluations. Sera from polyposis patients and their relatives were obtained through the courtesy of Dr. T. Alm, St. Erik’s Hospital, Stockholm. The germ-free rat material was kindly provided by Professor B. E. Gustafsson, Department of Germfree Research, Karolinska Institute, Medical School, Stockholm. E . coli strains were obtained from the international Escherichia coli Center in Copenhagen through the courtesy of Dr. F. Orskov and from the University of Gothenburg through the courtesy of Dr. K. Lincoln. This work was supported by Grant B 75-16X159- 11 C from the Swedish Medical Research Council.
REFERENCES I. A h , T. & Licznerski. G. Clin. Gastroent. 1973, 2, 577-602 2. Andersen, H. J. J. Pediat. 1966, 68, 542-550 3. Bjorneboe, M., Prytz, H. & Orskov, F. Lancet 1972, 1, 58-60 4. Broberger, 0. & Perlmann, P. J . Exp. Med. 1962, 115, 13-26 5. Broberger, 0. & Perlmann, P. J. Exp. Med. 1959, 110, 657-674 6. Bull, D. M. & Ignaczak, T. F. Gastroenterology 1973.64. 43-50 7. Diaz, F. & Neter, E. Amer. J. Med. Sci. 1968, 256, 18-24 8. Eckhardt, R., Heinisch, M. & Meyer zum Buschenfelde, K. H. Scand. J. Gastroent. 1976, 11, 49-54 9. Gustafsson, B. E. Acta. Path. Microbiol. Scand. 1948, SUPPI7 3 , 1-130 10. Gustafsson, B. E. Ann. N.Y. Acad. Sci. 1959, 78. 17-28 1 1 . Hammerstrom, S., Lagercrantz, R., Perlmann, P. & Gustafsson, B. E. J . Exp. Med. 1965, 122, 10751086 12. Harrison, W. J. Loneet 1965, I , 1346-1350 13. Kunin, C. M., Beard, M. V. & Halmagyi, N. E. Proc. Soc. Exp. Biol. (N.Y.) 1962, 111, 16&166 14. Lagercrantz, R.. Hammarstrom, S., Perlmann, P. &
714
H . E. Carlsson, R . Lagercrantz & P. Perlmann
Gustafsson, B. E. Clin. Exp. Immol. 1966,I , 263-
276 15.Lagercrantz, R., Hammarstrom, S., Perlmann, P. & Gustafsson, B. E. J. Exp. Med. 1968, 128, 13391352 16.Lagercrantz, R., Perlmann, P. & Hammarstrom, S. Gastroenterology 1971,60. 38 1-389 17.Perlmann, P., Hammarstrom, S., Lagercrantz, R.. & Campbell, D. Proc. Soc. Exp. Biol. (N.Y.) 1967, 125, 975-980
Scand J Gastroenterol Downloaded from informahealthcare.com by V U L Periodicals Rec on 12/29/14 For personal use only.
18. Perlmann, P., Hammarstrom, S., Lagercrantz, R. & Gustafsson, B. E. Ann. N.Y. Acad. Sci. 1965,124,
377-394 19. Perlmann, P., Lagercrantz, R. & Hammarstrom, S. pp. 755-774 in Miecher, T. A. & Muller-Eberhard, H. J. (eds.) Textbook of lmmunopathologv Grune and Stratton, New York, 1976 20.Reed,T. E. & Neel, J. V. Amer.J. Hum. Genet. 1955, 7 . 23G263 Received 12 February 1977 Accepted 23 May 1977
21.Saito, I. Fukushima J. Med. Sci. 1967,14, 45-53 22.Thayer, W. R.,Brown, M.,Sangree,M. H.,Katz, J. & Hersh, Th. Gastroenterology 1969,57, 3 11-3 18 23.Thomas, H. C., McSween, R. N. M. & White, R. G. Lancet 1973,I , 1288-1291 24.Triger, D. R., Alp, M. H. & Wright, R. Lancet 1972, I , 60-63 25.Triger, D.R., Kurtz, J. B. & Wright, R. Gut 1974, 15, 94-98 26. Westphal. O., Luderitz, 0. & Bister, F. Z. Nuturforsch. 1952,7b, 148-155 27.Whang, H. Y.& Neter, E. J. Pediat. 1963,63,4124 19 28. Wold, J. K., Midtvedt, T. & Jeanloz,R. W. Acta Chem. Scand. 1914,B 2 8 , 277,'284 29.Wold, J. K., Smestad, B. & Midtvedt, T. Acta Chem. Scand. 1975,B 29, 703-709 30.Wright, R. & Truelove, S. C. Gut 1966,7 , 32-40
Scand J Gastroenterol Downloaded from informahealthcare.com by V U L Periodicals Rec on 12/29/14 For personal use only.
HANS ERIK CARLSSON, RUTGER LAGERCRANTZ & PETER PERLMANN Dept. of Immunology, Wenner-Gren Institute, University of Stockholm, and Dept. of Pediatrics, Karolinska Hospital, Stockholm, Sweden
Carlsson, H. E., Lagercrantz, R. &. Perlmann, P. Immunological studies in ulcerative colitis. VIII Antibodies to colon antigen in patients with ulcerative colitis. Crohn’s disease, and other diseases. Scand. J. Gastroent. 1977, 12, 707-7 14. Sera from patients with ulcerative colitis or Crohn’s disease had elevated titers to colon antigen from germ-free rats significantly more often than sera from patients with gastroenteritis, irritable colon, non-gastrointestinal diseases, and healthy controls. Elevated anticolon titers in significant frequency were also found in patients with liver cirrhosis, urinary tract infections,and in polyposis coli and their relatives.Females with ulcerative colitis had, on an average, higher titers than men especially in the age group 30 years and over. In Crohn’s disease the antibody titers often increased with time-as opposed to those in ulcerative colitis and non-gastrointestinaldiseases. In conjunction with results published earlier, the present work supports the assumption that the antibodies in ulcerative colitis patients react with antigenic determinants distinct from those recognized by the colon antibodiespresent in other groups, including patients with Crohn’s disease and polyposis. Key-words: Auto-antibodies; colon antigen; Crohn’s disease; polyposis coli; ulcerative
colitis P. Perlmann, Ph.D., Dept. of Immunology, Wenner-Gren Institute, University of Stockholm, S-106 91 Stockholm, Sweden
In sera from patients with ulcerative colitis autoantibodies against colon were demonstrated in the early 1960s (1 9). N o such antibodies were found in sera from healthy controls when antigen from human sterile colon was used (4, 5). However, when human conventional colon or colon from germ-free rats was used, a background titer was found in most sera from control subjects ( 18, 30). The present article reports the results from a retrospective study of 13 10 consecutive patients. A material of patients with polyposis coli and their relatives is also included. A cross-reaction between the Common Antigen (CA) present in most Enterobacteriaceae (6,8, 17) and the colon antigen has been demonstrated (15). It has, therefore, been proposed that autoantibodies against colon in ulcerative colitis may be
formed as a result of breakage of tolerance through cross-reaction between C A and autologous colon tissue (15). In this paper we also examine the reactivity with C A of sera from patients with Crohn’s disease, polyposis coli, and their relatives.
MATERIALS A N D METHODS
Patients. Between 1968 and 1973 the Department of Immunology at the Wenner-Gren Institute received sera for testing of autoantibodies against colon. During this time, sera from 13 10 in- or outpatients were obtained from hospitals all over Sweden. The diagnoses were checked in the case histories obtained from the hospitals. In 273 cases the case histories were incomplete or not obtained.
Scand J Gastroenterol Downloaded from informahealthcare.com by V U L Periodicals Rec on 12/29/14 For personal use only.
108
H . E. Carlsson, R . Lagercrantz & P. Perlmann
Diagnosis, sex, and age of the remaining patients are shown in Table I. Serum. A total of 1900 serum samples were investigated. In 83 cases three or more samples were obtained from the same patient. 280 samples were from healthy blood donors and laboratory personnel and 182 from patients with polyposis coli and their relatives. Sera were stored frozen in aliquots at -20 OC until used. Antigen. Colon antigen was derived from the feces of germ-free rats of the Swedish inbred strain (9, 10). The antigen was extracted with phenolwater and purified as described earlier ( 1 1, 18). E. coli strains 0 : 14 and 0: 75 were grown in synthetic medium ( 17) and the lipopolysaccharide (LPS) extracted with phenol-water (26) and purified as previously described ( 1 1). Passive hemagglutination. Passive hemagglutination was performed as described earlier ( 1 1). Packed and washed sheep erythrocytes (0.025 ml) were sensitized with 8 mg colon antigen, 0.12 mg E . coli 0: 14 LPS, and 1.0 mg E . coli 0:75 LPS, respectively, in 4 ml phosphate-buffered saline (PBS) pH 7.3. All sera were inactivated at 56 OC for 30 minutes and absorbed twice with equal volumes of washed packed sheep red blood cells and, if necessary, with human A, red cells. To ascertain reproducibility 3-5 sera of known titer were included in all experiments. All sera were titrated twice in independent experiments. The titers were reproducible within plus or minus one dilution step.
Hemagglutination inhibition. Hemagglutination inhibition tests were performed as earlier described ( 1 1, 17). For the inhibition experiments sera with titers >1/32 were chosen. To 0.1 ml of sera containing four to eight hemagglutinating units ot‘ antibody, inhibitor was added (0.002- 1 mg/ml). After incubation for 1/2 hr. the sensitized cells were added and hemagglutination recorded as usual ( 1 1).
RESULTS
Clinical data. Diagnosis, sex, and age distribution of the 1037 consecutive patients are shown in Table I. A majority (857) suffered from gastrointestinal Table 11. Incidence of elevated anti-colon hemagglutination titers (>1/16) and geometric means of titers in different patient groups Tot. no.
%> 1/16Geometric
mean
5 16 148 71 91 12 13
Ulcerative colitis Crohn’s disease Gastroenteritis Colon irritabile Salmonellosis Liver cirrhosis Polyposis coli Relatives of polyposis patients Urinary tract infections Non-gastrointestinal diseases Healthy controls
23.x 24. I 23.3
122
60 61 41 47 50 62 53
60 16
55 69
23.h
158 280
35 13
33.1
21.2 2.3.2 24.5 23.1
24.5
22.2
Table I. Major diagnosis with sex and age distribution of 1037 consecutive patients with sera submitted for testing of autoantibodies against colon Age range. years Diagnosis Tot. Male Female 0-14 15-29 30-44 45-60 >60 no. c 9 0 a’ Y S d \ U” Ulcerative colitis Crohn’s disease Gastroenteritis Colon irritabile Salmonellosis Liver cirrhosis Urinary tract infection Non-gastrointestinal diseases Total
516 148 71 97 12 13 16 158
261 66 36 41 4 6 3 16
255 82 41 56 8
79 106 35 4 2 1 5 2 1 9 1 2 1 3
13 82
101 82 17 2 1 8 8 17 17 2 2 0 0 3 7 34 25
1037
493
547
182 162
143 203
7
2
2
0
4 13
16
44 39 3 0 1 13 13 7 3 3 5 1 1 9 0 2 1 1 1 1 0 1 0 I I 14 10
19 5 7 8 1 2 1 16
7 0
84
59
81
55
9 1
3
2
1 0 2 0 5
8 0 2 0 14
18
36
Scand J Gastroenterol Downloaded from informahealthcare.com by V U L Periodicals Rec on 12/29/14 For personal use only.
Immunologicat Siudies in Ulcerative Colitis
disorders; of these 5 16 had ulcerative colitis (including 96 cases of proctitis) and 148 Crohn’s disease. Most patients with Crohn’s disease had lesions in colon and ileum (especially the distal part). Serological data. The incidence of hemagglutinating antibody titer 2 I / 16 against colon antigen in the different patient groups is given in Table 11. There is a statistically significant difference in the incidence of elevated titers (> I / 16) between healthy controls and non-gastrointestinal diseases, on the one hand, and ulcerative colitis and Crohn’s disease on the other (p 1/16) in ulcerative colitis with only rectal involvement (proctitis) and with colonic or colonic and rectal involvement (ulcerative colitis). The incidence in male (M) and female ( F ) patients within the groups is also given
%>I116
Sex
N
?h>1/16
420
60
96
59
M F M F
208 212 53 33
54 65 49
Tot.no.
I
Ulcerative colitis Proctitis
I
P p
0.05
p
0.05
12
i
titers in proctitis patients differed statistically from healthy controls (p (0.001) and non0 gastrointestinal disease (p ~ 0 . I). The incidence of elevated titers is significantly higher in females with ulcerative colitis. while no significant difference is seen between the sexes in
disease or the intestinal involvement and colon titer in ulcerative colitis and Crohn’s disease. The number of patients with relevant clinical data was. however. small and too heterogeneous to give conclusive results. There was a decrease of colon titer in female patients with ulcerative colitis or
Table IV. Incidence of elevated colon titers ( > I / 16) in male ( M ) and female (F) patients with ulcerative colitis, Crohn’s disease and non-gastrointestinal diseases
Ulcerative colitis Crohn’s disease Non-gastrointestinal disease
Tot. no.
Sex
N
?h> 1/16
5I6
M F M F M F
26 I 255 66 82 16 82
53 66 58 63 31 31
148 158
P p
0.01
p > 0.05 p
> 0.05
710
H. E. Carlsson. R . Lagercrantr & P. Perlmann
looj
(a) Utcerotive colitis
I
(516)
n
P
v 0,
0,
.-> .= 50ul
.->
n
.= 508 n
s
-$
Scand J Gastroenterol Downloaded from informahealthcare.com by V U L Periodicals Rec on 12/29/14 For personal use only.
0
r
3
, 1/ 16, similar to those observed in ulcerative colitis and Crohn’s disease. The difference from healthy controls and non-gastrointestinal diseases is statistically significant (p t0.001 and p 1/32), and E. coli 0:75 (>1/64), in patients with ulcerative colitis, Crohn’s disease, polyposis coli, relatives to polyposis and healthy controls
Tot. no. Ulcerative colitis Crohn’s disease
50 96
Polyposis coli
122
Polyposis relatives
60
Healthy controls
48
Sex
F M F M M F M F
N
Colon %>1/16
E. coli 0: 14 %>1/32
E. coli 0: 75 %>1/64
24 26 47 49 81 41 31 29 25 23
56 63 68 61 56 50 57 50 9 30
58 58 17 30 20 20 26 21 12 30
85 73 70 68 nd a nd * nd * nd a 76 83
* nd = not determined In 39 sera from patients with Crohn’s disease no inhibition of the reaction between four to eight hemagglutinating units of colon antibodies and colon antigen was observed with 0.002 to 1 mg/rnl of LPS from E. coli 0: 14 or E. coli 0:75.
DISCUSSION The sera were obtained from many different hospitals. All diagnoses were checked, and patients with incomplete case histories were excluded from the investigation. It is inherent in this kind of retro-
In earlier investigations using germ-free rat colon antigen, a hemagglutination titer of >1/16 has been shown to discriminate between ‘normal’ and ‘abnormal’ (14). When this titer limit is used, only about 15% of healthy individuals score as positive, whereas about 60% of patients with ulcerative colitis or Crohn’s disease have elevated titers. This study has confirmed that the incidence of titers >I/ 16 is significantly higher in patients with ulcerative colitis or Crohn’s disease than in healthy controls or non-gastrointestinal diseases ( 14).
Scand J Gastroenterol Downloaded from informahealthcare.com by V U L Periodicals Rec on 12/29/14 For personal use only.
112
H . E . Carlsson, R . Lagercrantz & P . Perlmann
Previous investigations have suggested that a few patients with solely distal engagement in ulcerative colitis may have a low incidence of elevated colon titers (14). The present work on a much larger material has shown that this is not the case. Gastroenteritis and irritable colon showed a relatively high incidence of elevated titers. It is, ‘however, significantly lower than in ulcerative colitis and Crohn’s disease, and significantly higher than in healthy controls. This result is not in complete agreement with earlier studies in which the titers were even lower and not statistically different from healthy controls (14). Our explanation for this discrepancy is that the present gastroenteritis and irritable colon groups are heterogeneous, possibly containing undiagnosed Crohn’s disease or ulcerative colitis. The high frequency of elevated titers seen in salmonellosis is in contrast with previous findings (14), but can be explained by the small number of patients in the present study. The higher incidence of elevated titers in females with ulcerative colitis previously observed (14) has been confirmed in this study. Such a difference was not observed in Crohn’s disease, at least not in patients younger than 3 0 years. A decrease of titers with age was observed in male patients but not in females with ulcerative colitis or Crohn’s disease. Thus, the difference in titers between males and females became more pronounced with age. No relation between clinical activity of disease and anticolon titer was seen in patients with ulcerative colitis or Crohn’s disease. This is compatible with previous investigations (12, 14, 30). The significant increase in titers with duration of Crohn’s disease has not been reported previously. In ulcerative colitis and non-gastrointestinal diseases, the frequencies of increasing and decreasing titers were equal, suggesting that the duration of the disease does not influence the titer values in these groups. Treatment with steroids slightly decreased colon titers in female patients but not in males. However, these patients constituted a relatively small and very heterogeneous group in that dose and duration of therapy varied widely. The suppressing effect on antibody levels is not unexpected, since the drug has an anti-inflammatory effect. The influence of steroids only on female patients’titers is most likely
a consequence of their higher average titer levels as compared to males. The relatively high incidence of elevated colon titers in the polyposis groups is interesting. Similarly to ulcerative colitis and Crohn’s disease, polyposis affects the colonic mucosa and an increased amount of mucus is secreted (1). Moreover, as in ulcerative colitis, the risk for malignancies is high ( I ) . The similarity between polyposis patients and their relatives in respect to antibody titers could be expected since the disease is dominantly inherited with a high degree of penetrance (20). The presence of high levels of colon antibodies in liver cirrhosis has been previously described (24). Raised levels of antibodies to many different antigens are seen in liver cirrhoses (3, 24, 25). These unusually high levels of antibodies may be produced as a response to antigens that should normally be trapped by the liver (23). It has also been suggested that non-specific stimulation results from exposure of immunologically competent tissues to mitogens like endotoxins, which should similarly be eliminated by the liver (23). E. coli 0: 14 contains high amounts of C A present in lower concentrations (or less active form) in most Enrerobacteriaceae (13). Previous investigations have shown significantly increased hemagglutination titers against E . coli 0: 14 LPS in patients with ulcerative colitis compared to healthy controls ( 14, 15, 22). It has therefore been assumed that this antigen cross-reacts with the colon antigen (15). Such cross-reactions have also been demonstrated after immunization of rats with heterologous colon antigen (18) and in hemagglutination inhibition with sera from patients with ulcerative colitis (15). Possibly CA, with which most individuals come into contact, may break the tolerance to autologous colon antigen in certain individuals and thus evoke auto-antibodies (15). The unusually high frequency of colon antibodies seen in patients with urinary tract infections may also be ascribed to this cross-reaction with CA. Such infections are frequently caused by gramnegative bacteria, notably E . coli, and elevated titers against C A in serum from patients with urinary tract infections have been noted by several investigators (2, 7, 21, 27). LPS from E. coli 0:75, which contains low
Scand J Gastroenterol Downloaded from informahealthcare.com by V U L Periodicals Rec on 12/29/14 For personal use only.
Immunological Studies in Ulcerative Colitis
amounts of CA, was used as control antigen. As was to be expected (15), there were no differences in the titers to this antigen between patients and controls. As previously shown, patients with Crohn’s disease have titers to E . cofi0: 14 similar to healthy controls (14). This disagrees with the results of Thayer et al. and Eckhardt et al. (8, 22). These authors found that a small number of patients with Crohn’s disease had an increased incidence of elevated antibody titers against E. coli 0: 14,significantly higher than healthy controls. The discrepancy between our results and theirs is hitherto unexplained. The reaction between colon antigen and serum from patients with ulcerative colitis can be inhibited with LPS from E. coli 0: 14 in about 30% of the cases, while there was no inhibition with LPS from E. cofi 0:75 (15). When the same antigen and experimental procedure were used, no inhibition was established with serum from patients with Crohn’s disease using LPS from E . cofi 0: 14 or E . cofi 0:75. These results, together with hemagglutination data, suggest that colon antibodies in Crohn’s disease, polyposis coli, relatives to polyposis, non-gastrointestinal diseases and healthy controls are of a specificity different from that of colon antibodies in ulcerative colitis. Similar differences in antibody specificity have been demonstrated in experimental immunization, as already mentioned ( 17). Statistically significant differences in the frequency of elevated hemagglutination titers to colon between groups of patients with ulcerative colitis and Crohn’s disease and other diseases were demonstrated. The wide distribution of individual titers may, however, limit the use of the test as a diagnostic tool. This limitation is probably due to the complexity of the antigen preparation used. It is a mixture of several similar bloodgroup-like glycoproteins of varying size and charge, containing 6080% carbohydrate (1, 28 and unpublished). The carbohydrate consists of a large number of oligosaccharide side-chains with different terminal sugars and of varying lengths and branching (29, and unpublished). This heterogeneity most likely results in a variety of antigenic determinants. Accordingly, individual patients or groups of them
71 3
may have elevated titers to different antigenic determinants. Further work to elucidate and differentiate the antigenic determinants of diagnostic importance on rat colon antigen is needed. ACKNOWLEDGEMENT The skillful technical assistance of Miss YlvaSjovall is gratefully acknowledged. The authors are also indebted to Mr. Ulf Brodin for valuable discussions and advice concerning the statistical evaluations. Sera from polyposis patients and their relatives were obtained through the courtesy of Dr. T. Alm, St. Erik’s Hospital, Stockholm. The germ-free rat material was kindly provided by Professor B. E. Gustafsson, Department of Germfree Research, Karolinska Institute, Medical School, Stockholm. E . coli strains were obtained from the international Escherichia coli Center in Copenhagen through the courtesy of Dr. F. Orskov and from the University of Gothenburg through the courtesy of Dr. K. Lincoln. This work was supported by Grant B 75-16X159- 11 C from the Swedish Medical Research Council.
REFERENCES I. A h , T. & Licznerski. G. Clin. Gastroent. 1973, 2, 577-602 2. Andersen, H. J. J. Pediat. 1966, 68, 542-550 3. Bjorneboe, M., Prytz, H. & Orskov, F. Lancet 1972, 1, 58-60 4. Broberger, 0. & Perlmann, P. J . Exp. Med. 1962, 115, 13-26 5. Broberger, 0. & Perlmann, P. J. Exp. Med. 1959, 110, 657-674 6. Bull, D. M. & Ignaczak, T. F. Gastroenterology 1973.64. 43-50 7. Diaz, F. & Neter, E. Amer. J. Med. Sci. 1968, 256, 18-24 8. Eckhardt, R., Heinisch, M. & Meyer zum Buschenfelde, K. H. Scand. J. Gastroent. 1976, 11, 49-54 9. Gustafsson, B. E. Acta. Path. Microbiol. Scand. 1948, SUPPI7 3 , 1-130 10. Gustafsson, B. E. Ann. N.Y. Acad. Sci. 1959, 78. 17-28 1 1 . Hammerstrom, S., Lagercrantz, R., Perlmann, P. & Gustafsson, B. E. J . Exp. Med. 1965, 122, 10751086 12. Harrison, W. J. Loneet 1965, I , 1346-1350 13. Kunin, C. M., Beard, M. V. & Halmagyi, N. E. Proc. Soc. Exp. Biol. (N.Y.) 1962, 111, 16&166 14. Lagercrantz, R.. Hammarstrom, S., Perlmann, P. &
714
H . E. Carlsson, R . Lagercrantz & P. Perlmann
Gustafsson, B. E. Clin. Exp. Immol. 1966,I , 263-
276 15.Lagercrantz, R., Hammarstrom, S., Perlmann, P. & Gustafsson, B. E. J. Exp. Med. 1968, 128, 13391352 16.Lagercrantz, R., Perlmann, P. & Hammarstrom, S. Gastroenterology 1971,60. 38 1-389 17.Perlmann, P., Hammarstrom, S., Lagercrantz, R.. & Campbell, D. Proc. Soc. Exp. Biol. (N.Y.) 1967, 125, 975-980
Scand J Gastroenterol Downloaded from informahealthcare.com by V U L Periodicals Rec on 12/29/14 For personal use only.
18. Perlmann, P., Hammarstrom, S., Lagercrantz, R. & Gustafsson, B. E. Ann. N.Y. Acad. Sci. 1965,124,
377-394 19. Perlmann, P., Lagercrantz, R. & Hammarstrom, S. pp. 755-774 in Miecher, T. A. & Muller-Eberhard, H. J. (eds.) Textbook of lmmunopathologv Grune and Stratton, New York, 1976 20.Reed,T. E. & Neel, J. V. Amer.J. Hum. Genet. 1955, 7 . 23G263 Received 12 February 1977 Accepted 23 May 1977
21.Saito, I. Fukushima J. Med. Sci. 1967,14, 45-53 22.Thayer, W. R.,Brown, M.,Sangree,M. H.,Katz, J. & Hersh, Th. Gastroenterology 1969,57, 3 11-3 18 23.Thomas, H. C., McSween, R. N. M. & White, R. G. Lancet 1973,I , 1288-1291 24.Triger, D. R., Alp, M. H. & Wright, R. Lancet 1972, I , 60-63 25.Triger, D.R., Kurtz, J. B. & Wright, R. Gut 1974, 15, 94-98 26. Westphal. O., Luderitz, 0. & Bister, F. Z. Nuturforsch. 1952,7b, 148-155 27.Whang, H. Y.& Neter, E. J. Pediat. 1963,63,4124 19 28. Wold, J. K., Midtvedt, T. & Jeanloz,R. W. Acta Chem. Scand. 1914,B 2 8 , 277,'284 29.Wold, J. K., Smestad, B. & Midtvedt, T. Acta Chem. Scand. 1975,B 29, 703-709 30.Wright, R. & Truelove, S. C. Gut 1966,7 , 32-40
