Abstract Background and aims: Togain insight in intestinal epithelialproliferation, cell death, and geneexpression during experimental coli-tis rats were treated with dextran sulfate sodium (DSS) for 7 days.Materials and methods: Proximaland distal colonic segments were ex-cised on days 2, 5, 7, and 28. Epithe-lial proliferation, cell death, entero-cyte gene expression (carbonic anhy-drase I (CA I) and goblet cell geneexpression (mucin, MUC2; trefoilfactor 3, TFF3) were studied im-munohistochemically and biochemi-cally. Results: Proliferative activitywas decreased in the proximal anddistal colon at the onset of disease(day 2). However, during active dis-ease (days 5–7) epithelial prolifera-tion was increased in the entire proximal colon and in the proximityof ulcerations in the distal colon.During DSS treatment the number of apoptotic cells in the epitheliumof both colonic segments was in-creased. In the entire colon surfaceenterocytes became flattened andCA I negative during active disease(day 5–7). Additionally, CA I levelsin the distal colon significantly de-
creased during this phase. In con-trast, during the regenerative phase(day 28) CA I levels were restored inthe distal colon and up-regulated inthe proximal colon. During all dis-ease phases increased numbers ofgoblet cells were observed in thesurface epithelium of the entire co-lon. In the distal colon TFF3 expres-sion extended to the bottom of thecrypts during active disease. Finally,MUC2 and TFF3 expression was in-creased in the proximal colon duringdisease. Conclusion: DSS affectedthe epithelium by inhibiting prolifer-ation and inducing apoptosis. DSS-induced inhibition of CA I expres-sion indicates down-regulation ofspecific enterocyte functions. Accu-mulation of goblet cells in the sur-face epithelium and up-regulation ofMUC2 and TFF3 expression in theproximal colon underline the impor-tance of goblet cells in epithelialprotection and repair, respectively.
Int J Colorectal Dis (2002) 17:317–326DOI 10.1007/s00384-002-0409-4 O R I G I N A L A RT I C L E
Ingrid B. RenesMelissa VerburgDaniëlle J. P. M. Van NispenJan A. J. M. TaminiauHans A. BüllerJan DekkerAlexandra W. C. Einerhand
Epithelial proliferation, cell death, and gene expression in experimental colitis:alterations in carbonic anhydrase I, mucin MUC2, and trefoil factor 3 expression
Introduction
Since the beginning of the 1990s dextran sulfate sodium(DSS) has been widely used in a variety of animal models to induce colitis. Several clinical symptoms
(diarrhea, bloody stools, and weight loss) and histopath-ological changes (inflammatory infiltrates, erosions, andcrypt loss) that are induced by DSS are similar to thoseobserved in patients with ulcerative colitis (UC) [1, 2, 3,4, 5, 6]. Analogously to UC, DSS-induced colitis is miti-
I.B. Renes · M. Verburg D.J.P.M. Van Nispen · H.A. Büller J. Dekker · A.W.C. Einerhand (✉ )Pediatric Gastroenterology and Nutrition,Department of Pediatrics, Erasmus Medical Center, Dr Molewaterplein 50, and Sophia Children’s Hospital, 3015 GE Rotterdam, The Netherlandse-mail: [email protected].: +31-10-4087444Fax: +31-10-4089486
J.A.J.M. TaminiauDepartment of Pediatrics, Academic Medical Center, Amsterdam, The Netherlands
gated by treatment with sulfasalazine [7], olsalazine [7],and cyclosporin [4], drugs that are widely used to treatinflammatory bowel diseases [2].
The DSS-induced colitis model gives the opportunityto study the dynamic disease process from the onset ofdisease to complete remission. Such a study is not possi-ble in humans, especially since the disease is usually di-agnosed only during advanced stages. However, as inUC, the exact mechanisms underlying the DSS-inducedpathology are largely unknown. It has now been demon-strated that immune cells and bacteria do not play an essential role in the induction of DSS-induced colitis [8, 9]. Furthermore, DSS is known to inhibit epithelialcell proliferation in vitro [9, 10]. In vivo the first signs ofinjury are seen in colonic crypt cells located at the cryptbase [1, 5]. Previously Tessner et al. [11] reported that inmice DSS induces a decrease in the proliferative activityin the cecal epithelium. Collectively these data suggestthat the primary action of DSS on the epithelium is theinhibition of proliferation, thereby deranging epithelialhomeostasis. Nevertheless, it is not known whether in-hibition of proliferation is the only effect of DSS on theepithelium. As DSS induces diarrhea, it is also possiblethat DSS affects the expression of gene products of sur-face enterocytes, such as carbonic anhydrases (CAs),which are involved in electrolyte transport, water ab-sorption, and intracellular pH regulation [12, 13, 14].Along the same line, we cannot exclude the possibilitythat DSS alters the epithelial barrier function and epithe-lial repair capacity by affecting respectively mucin(MUC2) synthesis and trefoil factor 3 (TFF3). To help usunderstand the pathogenic mechanisms of colitis it is essential to unravel the primary effects of DSS on theepithelium and the ensuing response of the epithelium tothe DSS-induced changes.
This study used a rat model of DSS-induced colitisand investigated the successive changes in epithelialmorphology, cell death, proliferation, and cell type spe-cific gene expression (as measure of cell function) in theproximal and distal colon. In addition, clinical symptomswere evaluated during and after DSS administration.Epithelial proliferation and cell death were studied (im-muno)histochemically. Cell type specific gene expres-sion was analyzed both (immuno)histochemically andbiochemically. CA I was used as marker for colonic en-terocyte function. MUC2 the primary constituent of thecolonic protective mucus layer and TFF3, a bioactivepeptide that is involved in epithelial protection and re-pair, were used as marker for goblet cell function [15,16]. By studying all these parameters in conjunction weobtained a more complete picture of the complex patho-genic processes that occur during damage and regenera-tion in the different regions of the colonic mucosa duringexperimental colitis.
Materials and methods
Animals
Eight-week-old, specified pathogen-free male Wistar rats (Broek-man, Utrecht, The Netherlands) were housed at constant tempera-ture and humidity on a 12-h light-dark cycle. One week prior to and during the experiment the rats were housed separately. Therats had free access to a standard pelleted diet (Hope Farms, Woerden, The Netherlands) and sterilized tap water (controls) orsterilized tap water supplemented with DSS. All the experimentswere performed with the approval of the Animal Studies EthicsCommittee of our institution.
Experimental design
Rats were given 7% (w/v) DSS (37–40 kDa, TdB Consultansy,Uppsala, Sweden) in their drinking water for 7 days, followed by a21-day recovery period during which DSS was omitted from thedrinking water. Fresh DSS solutions were prepared daily. To studyproliferation of epithelial cells 50 mg bromodeoxyuridine (BrdU)per kilogram body weight (Sigma, St. Louis, Mo., USA) was injec-ted intraperitoneally 24 h before decapitation. Four animals werekilled per day on days 0 (control), 2, 5, 7, and 28. Duplicate seg-ments 1 cm in length of the proximal colon (adjacent to the ileocecal valve) and of the distal colon (1 cm proximal from the rec-tum) were dissected. One set of segments (i.e., proximal and distalcolon) was washed in phosphate buffered saline (PBS), fixed in 4%(w/v) paraformaldehyde (Merck, Darmstadt, Germany) in PBS, andsubsequently processed for light microscopy. The other set of seg-ments was frozen in liquid nitrogen and stored in –70°C until ho-mogenization for biochemical analysis. Clinical symptoms (weightloss, stool consistency, bloody stool, and presence of gross blood)were assessed during the course of the experiment.
Histology
Sections 5 µm thick were routinely stained with hematoxylin andeosin to study histological changes, i.e., distortion of crypt epithe-lium, erosions, inflammatory infiltrate, cell death, and epithelialrestitution during and after DSS-induced damage. The extent ofcrypt loss and ulcerations was estimated with a micrometer in fivetissue sections per segment of each animal and the area involvedwas expressed as the percentage of the total surface area.
Immunohistochemistry
The 5-µm-thick paraffin sections were deparaffinized through agraded series of xylol-ethanol. To visualize BrdU incorporationsections were incubated with 2 N HCl for 1.5 h, washed in boratebuffer (0.1 M Na2B4O7, pH 8.5), incubated in 0.1% (w/v) pepsinin 0.01 M HCl for 10 min at 37°C and rinsed in PBS. Endogenousperoxidase activity was inactivated by 1.5% (v/v) hydrogen perox-ide in PBS for 30 min, followed by 30 min incubation withTENG-T (10 mM Tris-HCl, 5 mM EDTA, 150 mM NaCl, 0.25%(w/v) gelatin, 0.05% (w/v) Tween-20) to reduce nonspecific bind-ing. This was followed by overnight incubation with a 1:500 dilu-tion of mouse anti-BrdU (Boehringer-Mannheim, Mannheim, Ger-many). Then the sections were incubated for 1 h with biotinylatedhorse anti-mouse IgG (diluted 1:2000, Vector Laboratories, UK)followed by 1 h incubation with ABC/PO complex (VectastainElite Kit, Vector Laboratories) diluted 1:400. Binding was visual-ized after incubation in 0.5 mg/ml 3,3′-diaminobenzidine, 0.02%(v/v) H2O2 in 30 mM imidazole, 1 mM EDTA (pH 7.0). Finally,sections were counterstained with hematoxylin, dehydrated and
318
mounted. MUC2, TFF3, and CA I expression was demonstratedaccording to the above protocol with omission of the HCl incuba-tion, washing with borate buffer, and pepsin treatment. To stainMUC2 and TFF3 sections were boiled in 0.01 M citrate buffer atpH 6.0 for 10 min prior to incubation with a MUC2-specific anti-body (WE9, 1:500) [17] or TFF3-specific antibody (1:6000, Prof.Dr. D.K. Podolsky). To detect CA I rabbit anti-CA I (1:16000,Prof. Dr. W.S. Sly) was used. Biotinylated goat anti-rabbit IgG(1:2000, Vector Laboratories) was used as secondary antibody todetect TFF3 and CA I.
To study differences in epithelial proliferation between controland DSS-treated animals the number of BrdU-positive cells in sixwell oriented crypts was counted and expressed per crypt, per in-testinal segment, per time point (±SEM). To determine the numberof BrdU-positive cells per crypt in each intestinal region sectionswere judged twice by two independent and blinded observers.
Protein dot blotting
CA I, TFF3, and MUC2 expression was quantified as describedpreviously [18, 19]. Briefly, segments of proximal and distal colonwere homogenized, protein concentration was measured, and0.30 µg protein of each homogenate was dot-blotted. After blottingthe blots were incubated with anti-CA I (1:4000) to detect CA I ex-pression or with a MUC2-specific antibody (WE9, 1:100) to detectMUC2 expression [17, 20]. Thereafter the blots were incubatedwith 125I-labeled protein A (Amersham, Bucks, UK; specific activi-ty 33.8 mCi/mg). Binding of 125I-labeled protein A to anti-CA I,anti-TFF3 or WE9 was detected by autoradiography using a Pho-sphorImager and quantified using ImageQuant software (MolecularDynamics, B&L Systems, Zoetermeer, The Netherlands).
Statistical analysis
Analysis of variance was performed, followed by an unpaired t test. Differences were considered significant at P<0.05. Data arepresented as the mean ±standard error of the mean.
Results
Clinical symptoms
Rats treated with DSS suffered weight loss compared tocontrol animals (Fig. 1). Immediately after the beginningof DSS treatment rats started to lose weight. Weight lossprogressed during and shortly after DSS treatment. Onday 10, when weight loss of DSS-treated rats was maxi-mal, DSS-treated rats weighed 25% less than controlrats. On day 11, 5 days after the end of the DSS adminis-tration, rats started to gain weight again. Loose stools,diarrhea, and bloody stools were observed within 3 daysafter the start of the DSS treatment. Gross bleeding firstoccurred on day 5, and persisted until day 8. The occur-rence of and recovery from the various clinical symp-toms over time are presented in Fig. 1.
Morphology
The colonic tissue showed dramatic morphologicalchanges depending on time of treatment and localization
319
Fig. 1 Effects of DSS treatment on body weight and the occur-rence of clinical symptoms. Relative changes in body weight ofcontrol (open squares) and DSS-treated (filled triangles) rats. Error bars reflect the standard deviations of the relative changes inbody weight on each day. The body weight of each individual ani-mal was set arbitrarily at 1 on day 0 of the experiment. Clinicalsymptoms, i.e., loose stool, diarrhea, bloody stool, and grossbleeding, are presented as a function of time. Dashed line Days ofDSS/water treatment. Note that changes in body weight coincidedwith changes in clinical symptoms
within the colon, and was divided into three phases: on-set of disease, active disease, and regenerative phase.Onset of disease (day 2) was characterized by a slightflattening of the crypt epithelial cells (Fig. 2B), and aslight increase in the number of apoptotic cells in thecrypts and surface epithelium in both proximal and distalcolon (Fig. 2C). Moreover, in the distal colon focal cryptdistortions were observed that varied from atrophy tocomplete loss (Fig. 2B). During active disease (day 5–7)areas were seen with flattened crypt cells, crypt loss,massive inflammatory infiltrate, flattening of the surfaceepithelium, focal erosions, and necrotic cells in both co-lonic segments. The histological damage was focal in na-ture and most severe in the distal colon on days 5–7(Fig. 2D). Specifically, approx. 50% of the distal colonictissue consisted of areas with crypt loss or erosions,while approx. 20% of the proximal colon was severelyaffected during this phase. During the regenerative phase(day 28) the epithelial morphology of the proximal coloncompletely recovered. In contrast, in the distal colonbranched crypts and pronounced crypt elongations alongside erosions, which involved less than 10% of the totalsurface, were still apparent (Fig. 2E).
Proliferation
Epithelial proliferation was studied by immunohisto-chemical detection and quantification of incorporated
320
Fig. 2A–E Epithelial morphol-ogy and apoptosis before, dur-ing and after DSS treatment.Hematoxylin and eosin stainingof the distal colon. Morphologyof controls (A), after 2 days (B, C), and after 7 days of DSStreatment (D). Distal colon atday 28, 21 days after the end of DSS administration (E). Arrows Apoptotic cells (C). Arrows Flattened surface epi-thelium; arrowheads erosions(D, E)
Fig. 3A–D Effect of DSStreatment on epithelial prolifer-ation as measured by BrdU in-corporation into the distal co-lon. BrdU-positive epithelialcells were confined to the low-er half of the crypts in controltissue (A). The number ofBrdU-positive cells was loweron day 2 in distal colon (B)than in control tissue. Continu-ation of DSS treatment resultedin a strong increase in the num-ber of BrdU-positive cells inthe proximity of ulcerations inthe distal colon on day 7 (C).Note that BrdU-positive cellswere seen from crypt base toalmost the surface epithelium.On day 28, 21 days after thelast DSS administration, eleva-ted levels of BrdU-positivecells were still observed nearulcerations in the distal colon(D)
BrdU (Table 1, Fig. 3). Decreased numbers of BrdU-pos-itive cells were seen in the crypt compartment of eachcolonic segment on day 2, indicating that BrdU incorpo-ration was decreased during the first days of DSS treat-ment (Fig. 3B). As DSS administration continued, aclear increase in BrdU-positive crypt cells was observedin proximal colon (day 7). In the distal colon increasednumbers of BrdU-positive cells were only seen in elon-gated crypts in the proximity of ulcerations and/or areaswith crypt depletion and a flattened surface epithelium(Fig. 3C). However, in the epithelium located more dis-tantly from ulcerations the number of BrdU-positivecells was still decreased at day 5 (Table 1). In addition,due to the DSS-induced crypt loss (approx. 50% of thesurface area distally and approx. 20% proximally) theoverall number of BrdU-positive cells in the proximaland distal colon decreased compared to control tissue.On day 28, 21 days after ending DSS treatment, the levels of BrdU-positive crypt cells had returned to con-trol values in the proximal colon. In the distal colon, in-creased numbers of BrdU-positive cells were still ob-served in elongated crypts adjacent to erosions or areas
covered by a flattened surface epithelium (Fig. 3D). Thenumber of BrdU-positive cells in elongated crypts in thedistal colon located more distantly from ulcerations wascomparable to control values (Table 1).
Enterocyte specific CA I expression
CA I is expressed by surface enterocytes of the proximaland distal colon [13, 21, 22], and was used as marker forcolonic enterocyte function. In the proximal colon theCA I expression appeared unaltered during DSS treat-
321
Fig. 4A–C Effects of DSS onsurface enterocytes in the distalcolon. Surface enterocyteswere identified by CA I expres-sion in the colon of control rats(A), after 7 days of DSS treat-ment (B), and on day 28 duringthe recovery phase (C). ArrowsCA I positive cells during ac-tive disease (day 7) (B)
Fig. 5 CA I expression levels during the different phases of theDSS-induced disease. CA I levels in each colonic segment of con-trol and DSS-treated rats were quantified. Thereafter the values ofthe control rats were averaged per segment and arbitrarily set at 1.The CA I levels in each segment of DSS-treated rats were ex-pressed as relative values compared to control values (day 0). Fi-nally, the mean ±standard error of the mean is presented. Statisti-cal analysis was performed using analysis of variance followed byan unpaired t test. In the proximal colon CA I levels were slightlybut not significantly decreased during DSS treatment. In contraston day 28, CA I levels appeared significantly higher than on days5 and 7 (aP<0.05). In contrast in the distal colon CA I levels weresignificantly decreased on days 2, 5, and 7 (bP<0.05)
Table 1 Number of BrdU-positive cells in the proximal and distalcolon during and after DSS treatment. The number of BrdU-posi-tive cells in six well oriented crypts was counted and expressedper crypt, per intestinal segment, per time point
Days of /after DSS treatment Proximal colon Distal colon
Proximal colon: day 0 vs. day 2, P<0.001; day 0 vs. day 7,P<0.01; day 2 vs. days 5, 7, and 28, P<0.001; day 5 vs. day 7,P<0.05; day 7 vs. day 28, P<0.01. Distal colon: day 0 vs. days 2and 5, P<0.01; day 2 vs. days 7 and 28, P<0.001; day 5 vs. day 7,P<0.01; day 5 vs. day 28, P<0.001
322
Fig. 6A–L Effects of DSS on goblet cells in the proximal and dis-tal colon. Goblet cells in the proximal colon (A–D) and distal co-lon (E–L) were identified by detection of MUC2 (A, B, E–G, K)and TFF3 expression (C, D, H–J, L). Goblet cells in control tissue(A, C, E, H), after 7 days of DSS treatment (B, D, F, G, I, J), andon day 28, 21 days after the end of DSS administration (K, L).Note the accumulation of goblet cells in the surface epithelium of
the proximal colon on day 7 (B, D), the expression of TFF3 at thebottom of the crypts in the distal colon on day 7 (J), the accumula-tion of goblet cells in the surface epithelium on day 7 (F, I) andday 28 (K, L), and the large goblet cells in the elongated crypts onday 28 (K, L). Contr. control; pc proximal colon; dc distal colon;DSS7 day 7 of DSS treatment; DSS28 day 28, i.e., 21 days afterthe end of DSS treatment
ment in areas with intact crypts and normal appearingsurface epithelium (not shown). In the distal colon, how-ever, some surface enterocytes were CA I negative dur-ing DSS administration (days 5 and 7) in areas with in-tact crypts and otherwise normally appearing surfaceepithelium. Moreover, in both proximal colon and distalcolon CA I expression was almost completely absent ondays 5 and 7 in areas with a flattened surface epitheliumand crypt loss (Fig. 4B). During the regenerative phase(day 28) the CA I expression was restored in the entireproximal colon and also in those areas of the distal colonwhere the epithelial morphology appeared normal (notshown). Moreover, flattened surface epithelium adjacentto ulcers was frequently CA I positive (Fig. 4C).
Biochemical analysis of CA I protein levels in theproximal colon revealed a slight but not significant de-crease in CA I protein levels in the course of DSS treat-ment (Fig. 5). Remarkably, during the regenerative phase(day 28) an overshoot in CA I expression was seen inthis segment, with CA I levels that were significantlyhigher than on days 5 and 7. In the distal colon morepronounced alterations in CA I protein levels were ob-served. CA I expression levels decreased significantlyuntil day 7, the end of DSS treatment. During the regen-erative phase (day 28) CA I expression returned to levelsthat were not significant different from control values.
Goblet cell specific MUC2 and TFF3 expression
To study goblet cell function MUC2 and TFF3 expressionwas studied. In controls MUC2 was expressed by gobletcells in the proximal and distal colon from crypt base tothe surface epithelium (Fig. 6A, E). In contrast, TFF3was expressed in the upper crypts in the proximal colonand upper two-thirds of the crypts in the distal colon(Fig. 6C, H). Alterations in the localization and numbersof goblet cells were seen during and after DSS treatmentin both colonic segments. During the onset of disease(day 2) MUC2- and TFF3-positive goblet cells accumu-lated in the surface epithelium in both proximal and distalcolon (not shown). As DSS treatment progressed (days 5and 7) the accumulation of MUC2- and TFF3-positivegoblet cells in the surface epithelium became more pro-nounced, particularly in the proximal colon (Fig. 6B, D).In the distal colon the area of TFF3 expression was ex-tended from upper two-thirds of the crypt toward thecrypt bottom in elongated crypts (Fig. 6J). In areas withflattened crypts TFF3 was expressed strongly in the uppercrypts and surface epithelium, and, although very weakly,also in the deeper crypt region (Fig. 6I). MUC2 expres-sion in the proximal and distal colon appeared unaltered,independent of crypt morphology, during active disease.Specifically, Muc2 was expressed in small goblet cellswithin damaged crypts and in small goblet cells in elon-gated crypts (Fig. 6F, G). Due to the massive crypt dam-
age, crypt loss, and loss of surface epithelium in the distalcolon (approx. 50%), the overall number of goblet cellsdecreased in the latter segment. During the regenerativephase (day 28) elevated numbers of MUC2- and TFF3-positive goblet cells were still seen in the surface epitheli-um of distal colon (Fig. 6K, L) but not in the proximalcolon. Especially in the distal colon crypts were elongat-ed and primarily contained large goblet cells.
Biochemical analysis of MUC2 protein levels demon-strated a progressive increase in MUC2 levels in theproximal colon in the course of DSS treatment (Fig. 7A).On day 7 this increase in MUC2 levels in the proximalcolon was fourfold control values and was significantlydifferent from control values and onset of disease. There-
323
Fig. 7 MUC2 (A) and TFF3 expression levels (B) during the dif-ferent phases of the DSS-induced disease. MUC2 and TFF3 levelsin each colonic segment of control and DSS-treated rats werequantified. Thereafter the values of the control rats were averagedper segment and arbitrarily set at 1. The MUC2 and TFF3 levels ineach segment of DSS-treated rats were expressed as relative val-ues compared to control values (day 0). Finally, the mean ±stan-dard error of the mean is presented. Statistical analysis was per-formed using analysis of variance followed by an unpaired t test.In the proximal colon MUC2 expression was significantly higheron day 7 than on day 0 (aP<0.01) and day 2 (bP<0.05). In the dis-tal colon significant differences in MUC2 levels were seen be-tween day 28 and days 5 and 7 (cP<0.05). Significantly lowerTFF3 levels were observed in the distal colon on day 7 than incontrols (aP<0.05), and on days 5 and 7 than on day 28 (bP<0.05)
after, during the regenerative phase (day 28), the MUC2levels in the proximal colon decreased but were still high-er than control levels. In the distal colon, however, atrend to decreased MUC2 expression levels was observedduring active disease (days 5–7), while during the regen-erative phase (day 28) MUC2 levels returned to controllevels. Similarly to MUC2 expression levels, TFF3 levelsseemed to increase in the proximal colon during and afterDSS treatment (Fig. 7B). This increase in TFF3 proteinlevels, however, was not statistically significant. In thedistal colon TFF3 expression levels were significantlylower during active disease (days 5–7) than in controls.During the regenerative phase (day 28) TFF3 levels hadincreased again and were comparable to control levels.
Discussion
As DSS-induced colitis is commonly used as a model forhuman inflammatory bowel disease, knowledge is essen-tial to be able to understand the mechanisms underlyingthe pathology of DSS-induced colitis. Therefore we ana-lyzed the occurrence of clinical symptoms, epithelialgene expression, proliferation, and cell death duringDSS-induced colitis. In our study DSS-induced clinicalsymptoms, such as body weight loss, loose and bloodystools, diarrhea, and gross bleeding, started within sever-al days after the beginning of the DSS treatment, weremost severe at the end of DSS treatment, and disap-peared only gradually during the recovery period. Previ-ous studies in mice and rats confirm this pattern of oc-currence and recovery of clinical symptoms [1, 5, 8, 23].The pathological features such as crypt loss, ulcerations,and inflammatory infiltrate that appeared in our rat DSScolitis model are similar to those in UC [3, 24] and inDSS-treated mice [3, 5, 24]. Furthermore, DSS-induceddamage in rat started and was most severe within the dis-tal colon, and was focal in nature, and was comparable toDSS-induced colitis in mice and UC in human [3, 5, 24].
Our study is the first to describe the response of thecolonic epithelium with respect to proliferation and ap-optosis during the different phases of DSS-induced dis-ease. DSS administration induced a decrease in prolifera-tive activity and an increase in the number of apoptoticcrypt cells, in the proximal colon as well as in the distalcolon, during the onset of disease. These data demon-strate that DSS exerts its toxic effects on the epitheliumvia the relatively undifferentiated crypt cells. This issimilar to humans with active UC, in which the numbersof apoptotic cells and proliferating cells in the cryptcompartment are also increased [25, 26]. The increasednumber of apoptotic cells in the surface epithelium suggests that DSS also affects mature surface cells.However, we cannot exclude the possibility that thesesurface cells were already damaged by DSS when theywere relatively immature crypt cells. When DSS feedingwas continued, the initial decrease in proliferative activi-
ty was followed by epithelial hyperproliferation in theentire proximal colon and in the proximity of ulcerationsin the distal colon. These data clearly demonstrate thatthe inhibiting effect of DSS on epithelial proliferationcan be overruled, leading to hyperproliferation. Howev-er, the mechanism(s) and growth factor(s) responsiblefor this phenomenon remain to be identified.
Apart from the DSS-induced alterations in epithelialproliferation and cell death, epithelial cell functionsmight also be affected. Therefore we examined the effectsof DSS on enterocyte specific CA I expression and gobletcell specific MUC2 and TFF3 expression. The effects ofDSS feeding on CA I expression levels were most pro-nounced in the distal colon, where it induced a significantdecrease in CA I levels during onset of disease and activedisease. This indicates that specific enterocyte functionswere down-regulated and thus that colonic functions areimpaired during DSS treatment. Immunohistochemicalanalysis of the distal colon revealed CA I negative sur-face cells during active disease in areas with apparentlynormal crypts and surface epithelium. In addition, themajority of flattened surfaces epithelial cells overlyingmucosa with crypt distortions were CA I negative, in boththe proximal colon and the distal colon. These data sug-gest that during transition of areas with intact crypts andsurface epithelium into areas with crypt loss and flattenedsurface epithelium, specific enterocyte functions aredown-regulated before restitution is initiated. Later in thecourse of the disease, during the recovery phase, the sur-face epithelial cells of the distal colon near ulcers werestill flattened but were frequently CA I positive.
Moreover, during this disease phase CA I protein levels restored to control levels in the distal colon andwere even up-regulated in the proximal colon. These dataimply that, in addition to epithelial proliferation, epithelialdifferentiation restored during the recovery phase to accel-erate the recovery of colonic functions. Similar resultshave been reported by Fonti et al. [27] who observed a re-duction in CA I expression in active UC, whereas in hu-mans with UC in remission CA I expression was restoredto control levels. Moreover, in rat the down-regulation ofCA I expression during DSS treatment was correlatedwith the occurrence of loose stools and diarrhea, suggest-ing that CA I plays a role in DSS-induced diarrhea.
Quantitative analysis of MUC2 and TFF3 demonstrat-ed a trend toward decreased MUC2 levels and significant-ly decreased TFF3 levels in the distal colon. As DSS-in-duced crypt loss and ulcerations were most severe in thedistal colon (approx. 50% distally vs. approx. 20% proxi-mally), the decreased MUC2 and TFF3 levels in this seg-ment during active disease are largely due to the dramaticdecrease in the overall number of goblet cells. Reducednumbers of goblet cells have also been reported in the co-lon of humans with active UC [28]. Additionally, MUC2levels in the sigmoid colon have also been found to be sig-nificantly lower in active UC than in controls and UC inremission [18, 19]. Yet, despite the DSS-induced damage
324
a progressive increase in MUC2 levels and maintenanceof TFF3 levels were observed in the proximal colon dur-ing DSS treatment. Furthermore, immunohistochemicalanalysis revealed an accumulation of MUC2- and TFF3-positive goblet cells in the surface epithelium of the proxi-mal and distal colon, suggesting selective sparing of gob-let cells during DSS-induced colitis. In addition to the ac-cumulation of goblet cells in the surface epithelium, alter-ations in goblet cell morphology and gene expressionwere observed in the crypts during active disease and theregenerative phase. Specifically, goblet cells in flattenedor elongated crypts of the proximal and distal colon weresmall but remained Muc2-positive. During this diseasephase TFF3 expression was extended toward the cryptbottom especially in elongated crypts of the distal colon,implying an up-regulation of the epithelial repair capacityin these crypts. Finally, during the regenerative phasecrypts were elongated and primarily contained large gob-let cells. In conjunction, these data imply that goblet cellsplay a pivotal role in epithelial defense against luminalsubstances and organisms, and in epithelial repair viaMUC2 and TFF3 synthesis, respectively.
Conclusions
DSS-induced clinical symptoms and morphological alterations in rat are comparable with clinical symptomsand morphological changes seen in DSS-induced colitis
in mice and UC in humans. The primary in vivo actionsof DSS on the epithelium are inhibition of proliferationand the induction of apoptosis. The epithelium respond-ed to these changes by a rapid hyperproliferation to re-establish epithelial homeostasis and functions. Epithe-lial hyperproliferation was accompanied by flattening ofthe surface epithelium, i.e., restitution. Collectively,these data underline that maintenance and restoration ofepithelial barrier function are of primary importance.During the recovery phase when erosions were stillpresent, the increase in epithelial proliferation coincidedwith a reestablishment of enterocyte gene expression(i.e., CA I expression), indicating that, next to epithelialbarrier function, epithelial cell function plays an impor-tant role in the restoration of full colonic integrity. Gob-let cells are of significant importance in the protectionand repair of the intestinal surface epithelium as dur-ing DSS colitis: (a) MUC2 and TFF3 expression in the proximal colon were up-regulated or at least main-tained, (b) MUC2 and TFF3-positive goblet cells accu-mulated in the surface epithelium, (c) small goblet cellsin the flattened colonic crypts continued to expressMUC2, and (d) TFF3 expression extended to the cryptbottom.
Acknowledgements This work was financed in part by NumicoBV, Zoetermeer, The Netherlands. The authors thank Prof. Dr.W.S. Sly and Prof. Dr. D.K. Podolsky for kindly providing anti-ratCA I antibodies and WE9 (anti-human MUC2 antibodies), respec-tively.
325
References
1. Cooper HS, Murthy SN, Shah RS, Sedergran DJ (1993) Clinicopathologicstudy of dextran sulfate sodium experi-mental murine colitis. Lab Invest69:238–249
2. Greenfield SM, Punchard NA, TeareJP, Thompson RPH (1993) The modeof action of the aminosalicylates in in-flammatory bowel disease. AlimentPharmacol Ther 7:369–383
3. Hamilton SR (1983) Diagnosis andcomparison of ulcerative colitis andCrohn’s disease involving the colon.In: Norris HT (eds) Pathology of thecolon, small intestine, and anus. Chur-chill Livingstone, New York
4. Murthy SN, Cooper HS, Shim H,Shah RS, Ibrahim SA, Sedergran DJ(1993) Treatment of dextran sulfate so-dium-induced murine colitis by intra-colonic cyclosporin. Dig Dis Sci38:1722–1734
5. Okayasu I, Hatakeyama S, Yamada M,Ohkusa T, Inagaki Y, Nakaya R (1990)A novel method in the induction of re-liable experimental acute and chroniculcerative colitis in mice. Gastroenter-ology 98:694–702
6. Yamada M, Ohkusa T, Okayasu I(1992) Occurrence of dysplasia and ad-enocarcinoma after experimentalchronic ulcerative colitis in hamstersinduced by dextran sulphate sodium.Gut 33:1521–1527
7. Axelsson LG, Landstrom E, Bylund-Fellenius AC (1998) Experimental co-litis induced by dextran sulphate sodi-um in mice: beneficial effects ofsulphasalazine and olsalazine. AlimentPharmacol Ther 12:925–934
8. Bylund-Fellenius A-C, Landström E,Axelsson L-G, Midtvedt T (1994) Ex-perimental colitis induced by dextransulphate in normal and germfree mice.Microb Ecol Health Dis 7:207–215
9. Dieleman LA, Ridwan BU, TennysonGS, Beagley KW, Bucy RP, Elson CO(1994) Dextran sulfate sodium-inducedcolitis occurs in severe combined im-munodeficient mice. Gastroenterology107:1643–1652
10. Ni J, Chen SF, Hollander D (1996) Ef-fects of dextran sulphate sodium on in-testinal epithelial cells and intestinallymphocytes. Gut 39:234–241
12. Charney AN, Dagher PC (1996) Acid-base effects on colonic electrolytetransport revisited. Gastroenterology111:1358–1368
13. Fleming RE, Parkkila S, Parkkila AK,Rajaniemi H, Waheed A, Sly WS(1995) Carbonic anhydrase IV expres-sion in rat and human gastrointestinaltract regional, cellular, and subcellularlocalization. J Clin Invest96:2907–2913
14. Sowden J, Leigh S, Talbot I, DelhantyJ, Edwards Y (1993) Expression fromthe proximal promoter of the carbonicanhydrase 1 gene as a marker for differentiation in colon epithelia. Differentiation 53:67–74
15. Tytgat KMAJ, Bovelander FJ, OpdamFJ, Einerhand AWC, Büller HA, DekkerJ (1995) Biosynthesis of rat MUC2 incolon and its analogy with humanMUC2. Biochem J 309:221–229
16. Mashimo H, Wu DC, Podolsky DK,Fishman MC (1996) Impaired defenseof intestinal mucosa in mice lacking intestinal trefoil factor. Science274:262–265
17. Tytgat KMAJ, Klomp LW, BovelanderFJ, Opdam FJ, Van der Wurff A, Einer-hand AWC, Büller HA, Strous GJ, Dekker J (1995) Preparation of anti-mucin polypeptide antisera to studymucin biosynthesis. Anal Biochem226:331–341
18. Tytgat KMAJ, van der Wal JW, Einer-hand AWC, Büller HA, Dekker J(1996) Quantitative analysis of MUC2synthesis in ulcerative colitis. BiochemBiophys Res Commun 224:397–405
19. Van Klinken BJW, Van der Wal JW,Einerhand AWC, Büller HA, Dekker J(1999) Sulphation and secretion of thepredominant secretory human colonicmucin MUC2 in ulcerative colitis. Gut 44:387–393
20. Podolsky DK, Fournier DA, Lynch KE(1986) Human colonic goblet cells.Demonstration of distinct subpopula-tions defined by mucin-specific mono-clonal antibodies. J Clin Invest77:1263–1271
21. Bekku S, Mochizuki H, Takayama E,Shinomiya N, Fukamachi H, IchinoseM, Tadakuma T, Yamamoto T (1998)Carbonic anhydrase I and II as a differ-entiation marker of human and rat co-lonic enterocytes. Res Exp Med (Berl)198:175–185
22. Lonnerholm G, Midtvedt T, SchenholmM, Wistrand PJ (1988) Carbonic anhy-drase isoenzymes in the caecum andcolon of normal and germ-free rats.Acta Physiol Scand 132:159–166
23. Takizawa H, Shintani N, Natsui M,Sasakawa T, Nakakubo H, Nakajima T,Asakura H (1995) Activated immuno-competent cells in rat colitis mucosainduced by dextran sulfate sodium andnot complete but partial suppression ofcolitis by FK506. Digestion 56:259–264
25. Arai N, Mitomi H, Ohtani Y, IgarashiM, Kakita A, Okayasu I (1999) En-hanced epithelial cell turnover associ-ated with p53 accumulation and highp21WAF1/CIP1 expression in ulcer-ative colitis. Mod Pathol 12:604–611
26. Iwamoto M, Koji T, Makiyama K, Kobayashi N, Nakane PK (1996) Apoptosis of crypt epithelial cells inulcerative colitis. J Pathol 180:152–159
27. Fonti R, Latella G, Caprilli R, Frieri G,Marcheggiano A, Sambuy Y (1998)Carbonic anhydrase I reduction in co-lonic mucosa of patients with active ul-cerative colitis. Dig Dis Sci 43:2086–2092
28. Jacobs LR, Huber PW (1985) Regionaldistribution and alterations of lectinbinding to colorectal mucin in mucosalbiopsies from controls and subjectswith inflammatory bowel diseases.J Clin Invest 75:112–118
