Autoimmunity in inflammatory bowel disease

KM DAS. Autoimmunity in inflammatory bowel disease. Can J Gastroenterol 1993;7(2):102-109. Autoimmune injury to the colonic epithelium has been a favoured hypothesis ever since circulating anticolon antibodies were first demonstrated in patients with inflammatory bowel disease (IBO). There is some evidence that anticolon antibodies present in the scrn from patients with ulcerative colitis, but not Crohn's disease, may injure epithelial cells by antibody-dependent cellular cytotoxicity (ADCC) . The ADCC correlated with the disease state in ulcerative colitis and disappeared following total colectomyRecent studies also demonstrated specific immunoglobulin (lg) G deposition along with activated complement components on colonic epithelium in ulcerative colitis and not in Crohn's disease, suggesting a role of anti-epithelial antihody causing cellular damage. The mucosal B cells in ulcerative colitis produce anticolon antibodies. Antineutrophil cytoplasmic antibody in patients with ulcerative colitis is intriguing and identification of the immunoreactive antigen from the neutrophils may explain its association with ulcerative colitis. Detection of intestinal mucosa( T cells reactive against epithelial cell-associated components also supports the existence of cytotoxic cells aimed at intestinal cells. The isolation of colon-bound IgG antibody in ulcerative colitis directed towards a unique colonic antigen, molecular weight fraction (Mr) 40K protein (P-40) suggests a specific form of autoimmunity in ulcerative colitis. P-40 is expressed in colonic epithelium and not in 13 other epithelial organs including other parts of the gastrointestinal tract, However, extraintestinal organs such as skin and biliary epithelium (organs commonly affected in ulcerative colitis) have a unique epitope shared with colonic epithelium. Anti-P-40 antibodies are present in the circulation of patients with ulcerative colitis and peripheral blood lymphocytes show proliferative response to the colon extract enriched in P-40. Recent studies demonstrated amplified lgG 1 antibody in the circulation (as well as in situ) bound to the colonic mucosa! epithelium along with activated complement products in patients with ulcerative colitis and not in Crohn's disease. This lgG 1 autoantibody response appears to be predominantly directed to P-40. (Pour resume, voir page 103)

T HE IDEA TIIAT INFLAMMATORY bowel disease (IBO) might be some form of an autoimmune disea~c has been under consideration ever since the in itial report of the presence of c irculating anticolon antibodies m 1959 (1 ).This idea is particularly at• tractive in relation to ulcerative colitis which is confined to the colonic epithelium .A lthough both Crohn' ~ disease and ulcerative colitis are two well defined and distinct entities, they are commonly referred to as a single entity, IBO.Several recent rcview1 have discussed various immunological aspects and their possible roles in the pathogenesis of .This review will focus o n some of the recent advan• ces in understanding of the general autoimmune reactions in human diseases, wi th particular attention to IBO.
The minimal requirement for activation of the helper T (T11) lymphocytes ( CD4) 1s the inrcractton of their T cell receptors with fragments of processed antigen (foreign or nutoontigen) expressed along with class II MHA molecules on the cell surface of antigen presenting cells (Figure L ).This interaction between T cells and antigen presentmg cells is facilitated by the CD4 molecule, which bmds to MHA II.Class II molecules arc encoded by DP, DQ, and DR genes of the human lymphocyte anugen (HLA) complex in man.The immune system can respond to a wide variety of different protem antigens, anJ any one MHA molecule can associate with a vru;t assortment of ann genic structures.
The recognition of the anngen(s) 111 coniunctton with self MHA -DR molecules 111 in test111al ep1thcl1um (ur ot her target cells) ,illows ,l populat1on of T cells (TH) acttvared to rcwgrnzc the same• antigen(s) presented by sclf-MHA-DR rnmronenrs at .1 future rime.The activated TH cells can generate cywcox1e T lymphocyte:, and can sttmulatc the B cclb to produce spcufic antibodie~ agamst ep1thclial cell associated protein(s) (Figure 2).lmmunogenetics: Ev1Jence ro suppmt a genetic has1s for an autoimmune d1~-orJer mcluJes fam1 lml clusters of J1scasc and autoantiboJies, a high rate nf concordance 111 identical rwms, and linkage w1rh ,t known genetic locus -111 particular, class I or c lass II gene proJucrs of the maJor h1stncompat1bdity complex (9).Clusters nf cases of inflammatory howcl disease \\'tthtn families have been reporreJ 111 17'\, to 39c}b of index cases in large series and.as re\ 1cwcJ hy Klem ,md colleagues (10), a high concorJancc rate in identical rwms has been founJ.Lmkage J1s-cqutl1hnum 111 chronic IRD has not been found to occur, although lack of an association with class II antigens may simply reflect the microheterogeneity of DR and other genes ( 11).A recent stud y usin g a n e thnic a lly matched population observed more frequent association of ulcerati ve colitis with HLA-DR2 and for Crohn's disease a somewhat weaker linkage with DRl -DQ wS (1 2 ).Absen ce o f a st rong genetic basis for 180 or any human autoimmune disease, and the presence of autoantibodies in apparently healthy family members ( 13 ), may suggest char inheritance of two or more unlinked functio nal classes o f genes may be necessary fo r disease to occur.O ne gene may permit the formation of autoantibo d i cs th ro ug h a d efec t in immunoregulation, whereas the other may allow the d evelo pme nt of lesio ns, thro ugh a disorder in the immune system's effector arm (9 ).Anticolon antibodies: The initial finding of anticolon antibodies ( l) has been confirmed b y ma n y in vestiga to rs ( 14,15).These anticolon antibodies are present in many but no t all patients with 180; they are equally common in ulcerative colitis and Crohn's disease, but cross react with enterobacterial common antigen ( 16) and with human blood group ABH antigens; they do not correlate with disease activity, duration or extent.There was no evidence of binding of these an ticolon antibo<lies with intestinal tissue in vivo and they do not mediate complement lysis of cells or a ntibo dy-d ependent cell -mediated cytolysis (A OCC).They were also found in several illnesses other than IBO that do not have intestinal involvement.The antigen recognized by scrum from patients with IBO is a lipopolysaccharide extractable from human fe tal colon ( l), germ-free rat colon or feces; t he latter materials have frequently been substituted for human material in assays.The colon lipopolysaccharide antigen is re lated to, but is distinct from, the common enterobacterial antigen of Kunin.O ne inherent problem with these serum an ticolon antibo di es h as been the question of whether they have anything to do with the initiation or pathogenesis of the disease in the intestinal mucosa.More recent stu<lics, however, demonstrated the presence of d isease specific antibodies in IBO (17-2 1).
Using double colour immunofluorescencc, deposition of specific lgG, together wi th activated complement products along the apical aspects of the colonic epithelial cells and a predominant corresponding subclass bearing B cells in the lamina propria in patients with ulcerative colitis, were <lemonstrated (22 ).In Cro hn's disease the antibody response was mainly lgG2 and a different path way of activation of complement was noted (23 ).T h e mechanism of these specific lgG responses, in particular the antigen(s) involved to initiate these responses, and the precise role of these antibodies are unknown.Complement activation as shown in both ulcerative colitis (22 ) and C rohn's Jisease ( 23 ), with release of C3b anJ CSa molecules may cause migration of neutrophils, macrophages and o ther effecror cells by chemotaxis (Figure 3).Such influx and activation of the effecror cells can cause release of leukotrienes, prostaglandins, platelet activating fac tor, oxygen radicals and proteases which lead to inflammatory response and epithelial cell dcstruc• tion.S pecific antibodies binding to the antigenic molecules on the target cells may also initiate AOCC, causing cytolysis as shown in ulcerative colitis (17.20).Specific epithelial antigens: Isolation of specific cellular components and tissue-bo und a ntibodies again st the specific tissue components have concr i bu ted to underst anding of the pa thogenes is of o th er diseases of chronic nature and unknown ctiologi considered to be 'autoimmune' (24)(25)(26)(27).A h ypo thesis that presently enjoys much popularity among clinical im, muno logists is tha t of 'mo lecular mimicry' associated with a 'hit-and-run event' (28).In this concept, a primary agen t, such as a virus, a bacterium or foreign protein, may attack a specific o rgan , trigge ring a local immune response directed aga inst t he aggressor.This nox io us o ffe nde r may be eliminated or destroyed, but in do ing so the immune system is fo rced to produce antibodi es or generate cells scnsitizc<l to some of the offender's antigens (pep, tides ) which share the same molecular configuration with components of the host's tissues ( mo lecular mimicry) (Figure 2).As a consequence, the immune system is now able to recognize and attack no rma l ce lls, triggering an autoaggressivc reactio n even though the initial culprit has completely disappea red (hit-and-run even t).This t heory has been suggested to explain some <liseases such as gluten-sensitive enteropathy, ankylosmg spond ylit1s, and Reiter's syndrome and autoimmune myocard iti s, based on ev idence of molecular mimicry of the wheat protein A-gliadin with the ElB protein of the human adenovirus Ad-12 for celiac disease (29), the HLA-B27 allele with However, in IBO, neither the tnggermg agent(s) inciting rhe autn1mmune reaction nor the precise cellular ep1copc:;, rhc target molecules, are known.Several candidate an 1 ,gens more specific to IBO have been proposed.
The 1solat1on of purified and well characteriie<l mt est inal antigens has added more sc1ent1fic crcdihility to the investigation of auwunmune phenomena in mo dunng the last several vears (1 U8,l9).S11111larly, the detecnon of intestmal mucosa!T celb reactive against gur-spec1ftc epithelial cell assoctatl•d proteins also k•nd support to the existence of cycotox11.. cel ls aimed at intestinal cells (32,33 ).
In earl ier studies, cells cytotoxic for colonic ep1thelial cel ls were first found m the penphL•ral blond of patients with ulcerative coitus 04.35) and sub-~quently m Crohn's col1t1s (36).The existence of such cells raised the possibility that a cellular ,1uto1mmune reacnon to the colonic epllhelium may be 11wolved mt he pathogenesis of IBD.Thts is an attract1Vl' hypothesis to explain wh'y ulcerative Lol1t1s remains confined m the colnn1c mucosa.Lack llf long rerm cultures nf human colonic epithelial cells and adu lt colonic epithelial cells frl•shly isolated by trypsmizmg mmccd colonic mucosa wnh a l11nited life span have grearly lurnred the numbers l1f studies in this ilfea.The ce ll mediating this cytotoxicity was thought w be killer cell m rhc natural ki ller cell subset (37,38).Thus, the qtotoxic1ry reprcscms e1thl•r ADCC or natural ki ller cell cyrotnx1c1t y.This ts umsistent with the observatilm that a high m1ilecular weight focrnr is 111 the scrum 11f patients with I BD that can mduce cytotox1c1ty m lymphocytes nt normal individuals (39).Although this matenal has never heen conclustvely identified, it is thought to be some type of cytophilic anubody or small immune complexes that 'arms' the ki ller cells.A sunilar type of cytotox1city was 1den-t1fied m the lamina proprrnl cells of patients with IBO (38).Intestina l epithelial cell-associated components: lntcstmal ep1thd1al cell asscK1.1tedlOmpnnents (ECAC) anrigens desert bed hy Roche and col league, ( 13) arc associated with hoth ulcerative colitis and Crohn's disease.Usmg ,ln ADCC ass;1y with ch1ckm erythrocytes coated with ECAC li( whm1l nrigm (ECAC-C) and human sew, it was found that prevall'ncc of anuhodies to ECAC-C was 69.7 1 \1 ,1mong patienrs Wll h chronic IF\[), both u lcCral iVl' coliris and Crohn\ d1,l'ase, and 55.7 1 \, among reln n ves; hoth prevalences were sigmficantly h igher than that nf thl control group (8.0%, P<0.001 ).U:,mg small-bowcl-Jemed ECAC. the prevalence nt ant 1boJ1l's among patients with lB[) and rdatiw, was also significantly higher than rhar of controls.React1,•1ty of s<:rn was directed tn a 160and a 137-KDa macnimolccule.l 60,000, 120,000 ,mJ 110,000wen: detected (41 ).Using immunotranshlot expL flllll'nts, two glycoprorems ,tssociated with Crohn\ d 1,e,l',e tissue of approx1m,1te molcLuL1r weight 200,000 ,mJ 160,000 were sub sequently 1dentif1ed (42).These proteins rcaued with Crohn 's dlSl:asl' sera hur nor w1th ulcerative colitis or other control sera.The sourCl' of these unmunoreacnvc prolctm and I heir role in Crohn's disease ts unknown.A recent study demonsrrated the presence of 1..1r1..ularmg ,111t1hody 111 two thirds of the patients wtth Crnhn's d1se,1sc against a mycohactenal 45/48 kDa doublet anugen ( 4 3 ).There was a pos1 rive 1..nrrdat1on nf the annbody tmc with aCtl\ tt) of the disease.Ir ts unclear whether mycobnctern1l anngcm arc cuologtc.11lyuwolved 111 Cmhn\ disease or if rh1s 1s ,1 moleudar m1m1Cry response.l g G ---------- Anticolon antibody production by mucosa! and peripheral blood lymphocytes in ulcerative colitis: Using ulcerative eolith sernm antiboJy and isolated rat colon ep11hclial cel ls, H1hi et a l (5) found the frequency ol scrum ant icolon ant 1b1xly to he 71 % in 41 patients wirh ulcerative colius.When Epstein-Barr virus was used as a R cell polyclonal act ivator, they further demonstrated rhe frequencies of lgG anticolon antibody secreting B cells ro he 1 .5 to 12.5/10 6 cells 111 the colonil mucosa and 0.1 to 0.5/10 6 cells 1n peripheral bl(1od, from patients with ulcerative colitis.A lthough the nature of the anugcm recognized hy these an-t1bod1es arc unknown, they a rc localized in the colon epithelial cells.Pcrinuc !ear a ntineu troph il-cytoplasmic antibody: Recently perinuclcar ant 111cu rrophil-cytnplasm 1c ant ihody (PANCA) has heen detected tn the scrn nf the majcmry of patients with ulcerative colit1 ~ ,md in a much smaller percentage of the scra of patients with Cn1hn's disease (44).Pat1cnts with pnmary sclcrosing cholangitis with or without !BL) also h:-1d These studies indicate that the Mr 40K protein is a normal colonic anti• gen, and CCA-lgG eluted from colonic tissue seems to he specific rn ulccrauvc coitus patients and represents an 'autoantibnJy' against the colon specific tissue protein.

lgG-Mediated Cellular Injury
The Mr 40K protein was highly enric hed, and subsequently murme monoclonal antibodies against the protein were Jcvclopcd (46).Using one of the monoclonal antibodies designated 7El2Hl2, the Mr 40K protein was localized co colomc cp1 , thclial cells and not in 13 other epithelial o rgans, including the small intestinal cnterocytes and stomach (46).These results indicated epithelial origin of the protein and confirmed an earlier observation nf the organ specitk1ty of rhc Mr 40K protein, as shown by immunotramblot analys1.1 ( 18).The immuno rcaccivity was localized on the plasma membrane mostl1 along the basolateral surface and apical domain of the colo01c epithelial cells.A majority of the patients with ulcerative colitis h ave c irculating antibodies to the cohm extract enriched m the Mr 40K protein further sugge~ting its na, cure as an autoanugen (47).
T he Mr 40K protein is distrihutcd