`injury in a murine model of experimental colitis
`
`DANIEL J. DRUCKER,1 BERNARDO YUSTA,1 ROBIN P. BOUSHEY,1
`LORRAINE DEFOREST,1 AND PATRICIA L. BRUBAKER1,2
`Department of 1Medicine, Banting and Best Diabetes Centre,
`The Toronto Hospital, and 2Department of Physiology, University of Toronto,
`Toronto, Ontario, Canada M5G 2C4
`
`Drucker, Daniel J., Bernardo Yusta, Robin P. Boushey,
`Lorraine DeForest and Patricia L. Brubaker. Human
`[Gly2]GLP-2 reduces the severity of colonic injury in a murine
`model of experimental colitis. Am. J. Physiol. 276 (Gastro-
`intest. Liver Physiol. 39): G79–G91, 1999.—The pathology of
`Crohn’s disease and ulcerative colitis is characterized by
`chronic inflammation and destruction of the gastrointestinal
`epithelium. Although suppression of inflammatory mediators
`remains the principle component of current disease therapeu-
`tics, strategies for enhancing repair and regeneration of the
`compromised intestinal epithelium have not been widely
`explored. The demonstration that a peptide hormone secreted
`by the intestinal epithelium, glucagon-like peptide-2 (GLP-2),
`is a potent endogenous stimulator of intestinal epithelial
`proliferation in the small bowel prompted studies of the
`therapeutic efficacy of GLP-2 in CD1 and BALB/c mice with
`dextran sulfate (DS)-induced colitis. We report here that a
`human GLP-2 analog (h[Gly2]GLP-2) significantly reverses
`weight loss, reduces interleukin-1 expression, and increases
`colon length, crypt depth, and both mucosal area and integ-
`rity in the colon of mice with acute DS colitis. The effects of
`h[Gly2]GLP-2 in the colon are mediated in part via enhanced
`stimulation of mucosal epithelial cell proliferation. These
`observations suggest that exploitation of the normal mecha-
`nisms used to regulate intestinal proliferation may be a
`useful adjunct for healing mucosal epithelium in the presence
`of active intestinal inflammation.
`intestine; inflammatory bowel disease; epithelium; growth
`factor; inflammation
`
`INFLAMMATION OF THE intestinal epithelium, as exempli-
`fied by Crohn’s disease and ulcerative colitis, results in
`considerable morbidity, and current therapeutic strate-
`gies, generally directed at suppressing components of
`the inflammatory response, remain suboptimal (27).
`The identification of molecules important for maintain-
`ing the growth and integrity of the mucosal epithelium
`has stimulated the development of novel approaches
`toward enhancement of mucosal protection in the gut.
`For example, the observation that trefoil peptides are
`abundantly expressed in the epithelium after intestinal
`injury was followed by studies demonstrating that mice
`deficient in intestinal trefoil factor are more susceptible
`to mucosal injury and recombinant intestinal trefoil
`factor enhances epithelial healing of the murine colon
`in vivo (35). Similarly, the demonstration that the
`
`The costs of publication of this article were defrayed in part by the
`payment of page charges. The article must therefore be hereby
`marked ‘‘advertisement’’ in accordance with 18 U.S.C. Section 1734
`solely to indicate this fact.
`
`keratinocyte growth factor (KGF) stimulates epithelial
`cell proliferation in the gastrointestinal tract (30),
`taken together with increased KGF expression in in-
`flammatory bowel disease (IBD) (51), suggests a pos-
`sible link between KGF and intestinal epithelial func-
`tion in vivo. We have now examined the therapeutic
`potential of a recently described intestinal growth
`factor, glucagon-like peptide-2 (GLP-2), in mice with
`dextran sulfate (DS)-induced colitis.
`Despite ongoing advances in our understanding of
`the cell biology of the gastrointestinal epithelium,
`principal strategies for treatment of IBD remain fo-
`cused on suppression of the cellular and humoral
`inflammatory response. These approaches involve local
`or systemic administration of corticosteroids, aminosa-
`licylates, or immunomodulatory agents such as azathio-
`prine, mercaptopurine, cyclosporin, and methotrexate
`(27). Although these latter agents are generally effec-
`tive they do not specifically target the intestine and
`their side effects may be considerable, precluding long-
`term use in patients with chronic IBD. Newer targeted
`approaches to immunosuppressive and anti-inflamma-
`tory therapy, including use of monoclonal antibodies
`against lymphocyte antigens (24, 48) or tumor necrosis
`factor (TNF) (46, 50), interleukin-4 (IL-4) delivery via
`adenoviral gene transfer (29), and antisense oligonucleo-
`tides for suppression of intercellular adhesion molecule
`activity (ICAM), are currently under evaluation.
`The rapid turnover and renewal of differentiated cell
`types that constitute the mucosal epithelium of the
`small and large bowel raise the possibility that stimula-
`tion of epithelial proliferation may be useful for enhanc-
`ing repair of epithelial damage in vivo. Identification of
`growth factors produced locally in the bowel that
`regulate crypt cell proliferation, such as epidermal
`growth factor, transforming growth factor-␣ (TGF-␣),
`and insulin-like growth factor I (18), provides an oppor-
`tunity to manipulate mucosal epithelial regeneration
`in experimental models of intestinal damage or resec-
`tion. The gastrointestinal tract also secretes regulatory
`peptides such as gastrin and gastrin-releasing peptide,
`with intestinal growth-promoting activity (28, 31, 56).
`The observation that injury of the intestinal mucosa is
`frequently associated with increased secretion of the
`proglucagon-derived peptides (PGDPs) (4), taken to-
`gether with increased intestinal growth in patients and
`rodents with glucagon-producing tumors (20, 26, 47),
`resulted in the identification of GLP-2 as the PGDP
`with intestinal growth factor-like activity (20).
`GLP-2 administered to normal mice and rats in-
`creases growth of the mucosal epithelium in small and
`
`0193-1857/99 $5.00 Copyright r 1999 the American Physiological Society
`
`G79
`
`CFAD Exhibit 1023
`
`1
`
`
`
`G80
`
`H[GLY2]GLP-2 IN A MURINE MODEL OF COLITIS
`
`large intestine (21, 53). The increase in small bowel
`mass is attributable in part to activation of crypt cell
`proliferation and inhibition of enterocyte apoptosis
`(53). GLP-2 also promotes intestinal hexose transport
`via upregulation of sodium-dependent glucose trans-
`porter 1 (SGLT-1) activity (11, 12). The importance of
`GLP-2 as a trophic factor for intestinal epithelium is
`illustrated by studies demonstrating that GLP-2 infu-
`sion prevents parenteral nutrition-associated mucosal
`hypoplasia in rats (10). To address the possibility that
`GLP-2 may be therapeutically useful for enhancing the
`endogenous reparative response to mucosal epithelial
`damage, we administered a degradation-resistant hu-
`man GLP-2 analog, h[Gly2]GLP-2 (6, 21), to mice with
`experimental DS-induced colitis.
`
`METHODS
`
`Animals and experimental protocol. Groups of 6- to 8-wk-
`old female CD1 mice, 22–24 g or 8- to 9-wk-old female BALB/c
`mice, 18–21 g (Charles River), were housed in plastic bottom,
`wire-lid cages, maintained on a 12:12-h light-dark cycle, and
`allowed chow and water containing 0 or 5.0% DS ad libitum
`throughout the study. The experiments carried out with CD1
`mice, designated experiments A and B, were carried out with
`treatment groups containing four to five mice housed to-
`gether. The experiments with BALB/c mice were carried out
`with five mice per control group for saline- and h[Gly2]GLP-2-
`treated mice (not receiving DS) and 10 mice per treatment
`group for the DS arm of the study, each BALB/c mouse being
`housed in a separate cage. CD1 mice were injected with either
`0.5 ml saline or 750 ng h[Gly2]GLP-2 in 0.5 ml saline twice
`daily. BALB/c mice were injected with either 0.5 ml saline or
`350 ng h[Gly2]GLP-2 in 0.5 ml saline twice daily.
`For the experiments with BALB/c mice individual water
`intake was recorded every 2 days. DS (mol wt 40,000–50,000;
`United States Biochemicals, Cleveland, Ohio, lot 103811) was
`freshly dissolved in drinking water throughout the study for
`both CD1 and BALB/c mice. Four days before the start of the
`study, mice were weighed using a Mettler PJ300 scale and
`randomly allocated to treatment groups. Subcutaneous injec-
`tions of PBS or h[Gly2]GLP-2 were administered twice daily,
`at 8 AM and 6 PM. Groups of mice received either regular
`autoclaved drinking water or water supplemented with 5.0%
`DS. CD1 mice were killed on the morning of day 11 after
`receiving 10 full days of water alone or water with DS and the
`final timed injection of saline or h[Gly2]GLP-2 was adminis-
`tered 2 h before mice were killed. BALB/c mice receiving DS
`appeared sicker than CD1 mice; after 9 days of DS treatment
`two deaths occurred in the saline-treated group and one
`death in the h[Gly2]GLP-2-treated group before day 10. After
`consultation with veterinary staff, the remaining groups of
`BALB/c mice were killed on day 10, after receiving ⬃91⁄3 days
`of oral DS. Two additional deaths in the BALB/c group (1 in
`saline-treated DS group, 1 in h[Gly2]GLP-2-treated DS group)
`occurred on day 10 on the morning when the mice were killed.
`Synthetic h[Gly2]GLP-2 was obtained from Allelix Biophar-
`maceuticals (Mississauga, Ontario, Canada). The exact pep-
`tide concentrations of different lots of h[Gly2]GLP-2 used in
`experiments A–C were determined using a combination of
`amino acid sequencing and HPLC. All animal experiments
`were carried out following experimental guidelines approved
`by the Animal Care Committee of the Toronto Hospital. The
`DS colitis experiments were carried out on several occasions,
`with similar results, and the data shown here are from three
`representative experiments designated A–C.
`
`Experimental analyses. Intestinal weights, morphology,
`enzymatic activity, and GLP-2 content were assessed as
`described previously (6, 7, 20, 53). For analysis of tissue
`PGDP content 2 cm of distal ileum (2 cm from cecum) and
`distal colon (2 cm from the anus) were homogenized on ice in 5
`ml of extraction buffer (1 N HCl, 5% HCOOH, 1% trifluoroace-
`tic acid, 1% NaCl) and extracted as described previously (13).
`RNA was prepared from homogenates of distal jejunum,
`ileum, and colon (13) and analyzed as previously described
`(7). Blood for GLP-2 RIA was collected in a final volume of
`10% Trasylol, EDTA, Diprotin A (5,000 KIU/ml:32 mM:0.1
`nM), and plasma was stored at ⫺80°C before analysis by RIA
`(6). Semiquantitative RT-PCR was carried out with aliquots
`analyzed from a range (20–30) of cycle numbers to ensure
`linearity for mouse TGF-␣ mRNA as previously described (8,
`9, 34). The PCR conditions were 94°C for 1 min and 68°C for 2
`min for 30 cycles. Primers for TGF-␣ were 58-TGCAGCACCCT-
`GCGCTCGGAAGAT-38 and 58-CCACCTGGCCAAATTCC-
`TCCTCTG-38. Occult blood testing was carried out using
`Hematest reagent tablets (Bayer, Etobicoke, Canada), as per
`the manufacturers’ instructions. Myeloperoxidase (MPO) ac-
`tivity was assayed spectrophotometrically as previously de-
`scribed (5). Statistical differences between treatment groups
`were determined by ANOVA using Tukey’s studentized range
`test for multiple comparisons at P ⫽ 0.05.
`Histological analysis. Intestinal segments for histology
`were taken from proximal
`jejunum (8 cm distal to the
`pylorus), distal jejunum (18 cm distal to the pylorus), proxi-
`mal ileum (10 cm before the cecum), and distal ileum (just
`proximal to cecum) and from the colon (1–3, 3–5, 5–7, and
`7–9 cm distal to the cecum). Tissues were fixed in 10%
`buffered Formalin for 48 h and embedded in paraffin using
`standard techniques. Four- to six-micrometer cross sections
`were cut and stained with hematoxylin and eosin. Intestinal
`micrometry was performed using a Leica Q500MC image
`analysis system. Ten well-oriented villi and 25 well-oriented
`crypts from each small intestinal section were used to deter-
`mine villus height and crypt depth. Disease severity was
`graded on a scale from 0–3 according to a standard scoring
`system (42): 0, normal bowel; 1, focal inflammatory cell
`infiltrate; 2, inflammatory cell infiltrate, gland drop out and
`crypt abscess; and 3, mucosal ulceration. Crypt cell prolifera-
`tion index as assessed by proliferating cell nuclear antigen
`(PCNA) staining and colonic epithelial apoptosis index as
`assessed by percent TUNEL-positive cells was carried out as
`previously described (20, 53).
`
`RESULTS
`
`Control mice not exposed to DS treated with either
`saline or h[Gly2]GLP-2 gained weight over the 9- to
`10-day experimental period (Fig. 1). The slightly in-
`creased body weight gain in the CD1 h[Gly2]GLP-2-
`treated control mice is largely attributable to the
`relatively greater increase in small bowel mass follow-
`ing treatment with the larger dose (750 ng twice daily)
`of h[Gly2]GLP-2 (Figs. 1 and 2).
`Mice receiving 5% DS in the drinking water devel-
`oped loose blood-streaked stools after 4–5 days, became
`progressively more lethargic, and lost ⬃20–25% of
`their body weight at the end of the 9- to 10-day
`experiment (Fig. 1). In contrast, h[Gly2]GLP-2-treated
`mice receiving 5% DS appeared much healthier and
`lost significantly less weight over the 9- to 10-day
`experimental period (P ⬍ 0.05 for both CD1 and
`BALB/c experiments, DS-h[Gly2]GLP-2- vs. DS-saline-
`treated groups, Fig. 1).
`
`2
`
`
`
`H[GLY2]GLP-2 IN A MURINE MODEL OF COLITIS
`
`G81
`
`exposed to DS but treated with h[Gly2]GLP-2, 750 ng
`twice daily for 10 days, had a significant increase in
`small bowel mass (Fig. 2B, 2.2 ⫾ 0.04 vs. 1.3 ⫾ 0.01 g
`h[Gly2]GLP-2 vs. control for experiment A, P ⬍ 0.05).
`CD1 mice with DS colitis treated with saline alone had
`a significant reduction in the mass of the small bowel
`(P ⬍ 0.05 for experiments A and B, Fig. 2B). In contrast,
`DS-h[Gly2]GLP-2-treated CD1 mice with colitis (experi-
`ments A and B) exhibited a significant increase in small
`bowel mass (Fig. 2B, 1.85 ⫾ 0.2 vs. 0.86 ⫾ 0.1 g,
`DS-h[Gly2]GLP-2- vs. DS-saline-treated mice for experi-
`ment A, P ⬍ 0.05) and a small but significant increase
`in small bowel length in experiment B (Fig. 2B, P ⬍
`0.05).
`Treatment of healthy control CD1 mice with
`h[Gly2]GLP-2 produced a significant increase in large
`bowel weight (Fig. 2B, 0.27 ⫾ 0.01 vs. 0.35 ⫾ 0.01, P ⬍
`0.05 saline- vs. h[Gly2]GLP-2-treated animals), consis-
`tent with the results of previous studies (19). Similarly,
`mice with DS colitis (experiments A and B) treated with
`h[Gly2]GLP-2 had a significant increment in large
`bowel weight (0.24 ⫾ 0.04 vs. 0.32 ⫾ 0.02 g, DS-saline-
`vs. DS-h[Gly2]GLP-2-treated mice, P ⬍ 0.05, Fig. 2B,
`experiment A). Treatment of normal CD1 mice with
`h[Gly2]GLP-2, 750 ng twice daily, produced a small but
`significant increment in large bowel length (Fig. 2B, b
`and f). Mice with DS colitis also exhibited a significant
`decrease in large bowel length (P ⬍ 0.05, saline-treated
`controls vs. mice with DS colitis, Fig. 2B). Although
`length was greater in h[Gly2]GLP-2-
`large bowel
`treated mice with colitis (Fig. 2B) this difference was
`statistically significant for mice in experiment B (DS-
`saline- vs. DS-h[Gly2]GLP-2-treated mice, P ⬍ 0.05)
`but not in experiment A (Fig. 2B).
`Similar results were observed for BALB/c mice with
`DS colitis treated with a lower dose of h[Gly2]GLP-2
`(350 ng twice daily). The relative magnitude of increase
`in small bowel weight in wild-type BALB/c mice treated
`with 350 ng h[Gly2]GLP-2 was smaller than in CD1
`mice but still significant (P ⬍ 0.05, saline- vs.
`h[Gly2]GLP-2-treated mice, Fig. 2B). The BALB/c mice
`receiving 5% DS appeared more ill than the CD1 mice,
`and a total of five BALB/c DS mice died during this
`experiment (3 in the saline-treated and 2 in the
`h[Gly2]GLP-2-treated group). The small bowel weight
`was significantly reduced in BALB/c DS-saline mice
`and increased significantly in mice treated with
`h[Gly2]GLP-2 (P ⬍ 0.05, saline- vs. h[Gly2]GLP-2-
`treated BALB/c mice, Fig. 2B). The large bowel weights
`of saline-treated BALB/c mice with DS colitis were
`significantly reduced compared with h[Gly2]GLP-2-
`treated mice with colitis, P ⬍ 0.05. Furthermore, the
`large bowel lengths were markedly reduced in both
`saline- and h[Gly2]GLP-2-treated BALB/c mice with
`colitis, but large bowel length was significantly greater
`in the h[Gly2]GLP-2-treated mice (P ⬍ 0.05, h[Gly2]GLP-
`2-treated vs. saline-treated BALB/c mice with DS coli-
`tis, Fig. 2B).
`As small and large bowel wet weights in mice with
`intestinal inflammation potentially reflect cellular infil-
`tration, hyperplasia, increased protein synthesis, and/or
`edema, we compared wet and dry small and large bowel
`
`Fig. 1. Change in body weight in groups of CD1 mice (experiments A
`and B, n ⫽ 5 mice for each treatment group) or BALB/c mice
`[experiment C, n ⫽ 5 mice for each control group and 10 mice for each
`treatment group receiving dextran sulfate (DS) in water] receiving
`either drinking water alone or 5% DS and saline vs. human glucagon-
`like peptide-2 analog (h[Gly2]GLP-2; 750 or 350 ng twice daily for
`CD1 vs. BALB/c mice, respectively) subcutaneous injections for 9–10
`days. *P ⬍ 0.05 for saline-treated control vs. 5% DS-saline groups.
`⫹P ⬍ 0.05 for 5% DS-saline vs. 5% DS-h[Gly2]GLP-2. Statistical
`differences between treatment groups were determined by ANOVA
`using Tukey’s studentized range test for multiple comparisons.
`
`GLP-1 has recently been shown to inhibit food and
`water intake (49, 55), whereas GLP-2 had no effect on
`food intake in mice over a 10-day experimental period
`(53). Nevertheless, one potential explanation for the
`different degree of illness and weight loss in our
`experiments might be due to theoretical effects of
`h[Gly2]GLP-2 on reduction of water intake and subse-
`quent cumulative intestinal exposure to DS. Saline-
`injected control BALB/c mice not receiving 5% DS had a
`mean daily water intake of 6.5 ⫾ 0.3 vs. 6.7 ⫾ 0.4 ml for
`h[Gly2]GLP-2-treated control mice (P not significant).
`Furthermore, the cumulative intake of 5% DS water
`over the entire 9-day experiment, as well as the 5% DS
`water intake from experimental days 7–9, was signifi-
`cantly greater for h[Gly2]GLP-2-treated compared with
`saline-treated BALB/c mice receiving 5% DS (4.96 ⫾
`0.5 vs. 5.9 ⫾ 0.6 ml/day for saline- vs. h[Gly2]GLP-2-
`treated mice with DS colitis on days 7–9, P ⬍ 0.05).
`These observations demonstrate that the difference in
`disease severity between groups cannot be explained on
`the basis of any putative effects of GLP-2 on water
`intake and hence intestinal exposure to DS.
`To determine the consequences of h[Gly2]GLP-2 ad-
`ministration in mice with DS-induced colitis, we exam-
`ined the gastrointestinal tract from the stomach to the
`colon in control and DS colitis treatment groups. Al-
`though no visible or microscopic pathology was de-
`tected in the stomach of DS-treated CD1 mice, stomach
`weight was reduced in the DS-saline-treated group and
`was restored toward normal in the DS-h[Gly2]GLP-2-
`treated mice (Fig. 2A, P ⬍ 0.05). Control CD1 mice not
`
`3
`
`
`
`G82
`
`H[GLY2]GLP-2 IN A MURINE MODEL OF COLITIS
`
`Fig. 2. A: stomach weight in groups of CD1 mice
`(experiment A) receiving either drinking water
`alone or 5% DS and saline vs. h[Gly2]GLP-2
`subcutaneous injections for 10 days. *P ⬍ 0.05
`for DS-saline-treated vs. either saline-treated
`alone or h[Gly2]GLP-2-treated mice. ⫹P ⬍ 0.05
`for 5% DS-saline vs. 5% DS-h[Gly2]GLP-2. C: wet
`and dry weights in 1-cm segments from jejunum,
`ileum, and colon of control and DS CD1 mice from
`experiment A. *P⬍0.05 for saline-treated nor-
`mal controls vs. all other groups. XP ⬍ 0.05 for
`h[Gly2]GLP-2-treated control vs. h[Gly2]GLP-2-
`treated DS. P ⬍ 0.05 for h[Gly2]GLP-2-treated
`DS vs. saline-treated DS mice.
`
`weights in CD1 mice with and without colitis (Fig. 2C).
`The increase in small bowel wet and dry weights in
`control mice treated with h[Gly2]GLP-2 was most evi-
`dent in the jejunum (P ⬍ 0.05, saline- vs. h[Gly2]GLP-2-
`treated control mice, Fig. 2C). Both saline- and
`h[Gly2]GLP-2-treated mice with DS colitis had in-
`creased wet colon weights (P ⬍ 0.05, control vs. DS
`colitis groups). In contrast, only the h[Gly2]GLP-2-
`treated mice with DS colitis had significantly increased
`dry colon weights (P ⬍ 0.05, saline- vs. h[Gly2]GLP-2-
`treated mice with DS colitis, Fig. 2C).
`
`Control CD1 mice treated with h[Gly2]GLP-2 exhib-
`ited a significant increase in jejunal crypt and villus
`height that was most prominent in the proximal jeju-
`num (P ⬍ 0.05, saline- vs. h[Gly2]GLP-2-treated mice,
`Fig. 3), consistent with previous experiments (19, 54).
`Histological analysis of the intestine from DS mice
`treated with saline injections demonstrated a reduction
`in small bowel villus and crypt height that was most
`marked in the jejunum (Fig. 3A). In contrast, mice with
`DS colitis treated with h[Gly2]GLP-2 exhibited a signifi-
`cant increase in small bowel villus height and crypt
`
`4
`
`
`
`H[GLY2]GLP-2 IN A MURINE MODEL OF COLITIS
`
`G83
`
`Fig. 2. B: intestinal weight and length in mice receiv-
`ing water alone or 5% DS. Small (a and e) and large (b
`and f) bowel weights and lengths (c and g and d and h
`for small and large bowel, respectively); means ⫾ SE
`for CD1 mice in experiments A and B and for BALB/c
`mice in experiment C. *P⬍0.05 for saline-treated vs.
`either h[Gly2]GLP-2-treated, DS-saline-treated, or DS-
`h[Gly2]GLP-2-treated mice. #P ⬍ 0.05 for h[Gly2]GLP-
`2- vs. 5% DS-saline-treated mice. ⫹P ⬍ 0.05 for 5%
`DS-saline- vs. 5% DS-h[Gly2]GLP-2-treated groups.
`
`5
`
`
`
`G84
`
`H[GLY2]GLP-2 IN A MURINE MODEL OF COLITIS
`
`Fig. 3. A: villus height and crypt depth
`from CD1 mouse small intestine, experi-
`ment A. PJ and DJ, proximal and distal
`jejunum; DI, distal ileum. *P ⬍ 0.05 for
`saline-treated vs. either h[Gly2]GLP-2-
`treated, 5% DS-saline-treated, or 5% DS-
`h[Gly2]GLP-2-treated mice. #P ⬍ 0.05 for
`h[Gly2]GLP-2-treated vs. 5% DS-saline-
`treated or 5% DS-h[Gly2]GLP-2-treated
`mice. ⫹P ⬍ 0.05 for 5% DS-saline- vs. 5%
`DS-h[Gly2]GLP-2-treated mice. B: crypt
`depth in colon of CD1 mice, experiment A.
`*P ⬍ 0.05 for saline- vs. h[Gly2]GLP-2-
`treated mice. #P ⬍ 0.05 for h[Gly2]GLP-2-
`vs. 5% DS-saline-treated mice. ⫹P ⬍ 0.05
`for 5% DS-saline- vs. 5% DS-h[Gly2]GLP-2-
`treated mice. C: crypt cell proliferation
`rate in colon as assessed by percentage
`of proliferating cell nuclear antigen
`(PCNA)-positive labeled cells (n ⫽ 15–25
`histological sections for CD1 mice in each
`arm of experiment A). ⫹P ⬍ 0.001 for 5%
`DS-saline- vs. 5% DS-h[Gly2]GLP-2-treated
`mice. # P ⬍ 0.01 for h[Gly2]GLP-2-
`treated control vs. 5% DS-h[Gly2]GLP-2-
`treated mice. D: apoptosis in mucosal epi-
`thelium of CD1 mice from group A as
`assessed by TUNEL immunopositivity
`(n ⫽ 20–25 histological sections from each
`experimental group). *P ⬍ 0.05 for saline-
`treated vs. h[Gly2]GLP-2-treated control
`mice.
`
`depth in both proximal and distal jejunum and ileum
`(Fig. 3A, P ⬍ 0.05 for DS-h[Gly2]GLP-2- vs. DS-saline-
`treated mice). In the colon, crypt depth was reduced in
`DS colitis mice treated with saline alone; however,
`crypt depth was significantly increased in mice with
`colitis treated with h[Gly2]GLP-2 (P ⬍ 0.05, DS-
`h[Gly2]GLP-2- vs. DS-saline-treated mice, Fig. 3B).
`Histological analysis of murine small bowel demon-
`strated an increase in both crypt cell proliferation and a
`decrease in epithelial cell apoptosis following treat-
`ment with GLP-2 (20, 53). Although increased numbers
`of PCNA-positive cells were observed in the colon of
`control CD1 mice treated with h[Gly2]GLP-2, the in-
`crease was not statistically significant (P ⫽ 0.1, Fig.
`3C). In contrast, a highly significant increase in the
`number of PCNA-positive cells was detected in histologi-
`cal sections from h[Gly2]GLP-2-treated mice with DS
`colitis (Fig. 3C, DS-h[Gly2]GLP-2- vs. DS-saline-
`treated mice, P ⬍ 0.001). Although h[Gly2]GLP-2-
`treated control mice exhibited a significant decrease in
`apoptotic cells in the colon (P ⬍ 0.05), no significant
`change in the percentage of apoptotic cells was ob-
`
`served in DS-h[Gly2]GLP-2- vs. DS-saline-treated mice
`(Fig. 3D).
`The colons from both CD1 and BALB/c mice with DS
`colitis contained blood and exhibited varying degrees of
`mucosal infiltration with leukocytes, loss of normal
`glandular architecture, and areas of both crypt erosion
`and destruction. All BALB/c mice receiving DS were
`occult blood positive when they were killed. Gross blood
`in the colon was visible at necropsy in four saline-
`treated and two h[Gly2]GLP-2-treated BALB/c mice
`with colitis. The damage to the epithelial mucosa was
`quantified by assessment of the total mucosal area in
`multiple histological segments from proximal, middle,
`and distal colon from CD1 mice. An increase in mucosal
`surface area in all three regions of the colon was
`observed in healthy control mice treated with
`h[Gly2]GLP-2 (Fig. 4A, P ⬍ 0.05 for h[Gly2]GLP-2- vs.
`saline-treated control mice). Mice treated with DS and
`saline injections exhibited a significant decrease in
`mucosal surface area, most prominent in the proximal
`colon (Fig. 4A). In contrast, h[Gly2]GLP-2 administra-
`tion to mice with DS colitis was associated with an
`
`6
`
`
`
`H[GLY2]GLP-2 IN A MURINE MODEL OF COLITIS
`
`G85
`
`increase in mucosal area that was most significant in
`the proximal and middle colon (Fig. 4A, P ⬍ 0.05 for
`proximal and middle colon, DS-h[Gly2]GLP-2- vs. DS-
`saline-treated mice).
`To further quantify the extent of epithelial disruption
`in mice with colitis, the presence or absence of intact
`epithelial mucosa was scored in multiple sections. Mice
`treated with h[Gly2]GLP-2 consistently exhibited a
`
`Fig. 5. Myeloperoxidase (MPO) activity normalized to protein con-
`tent in colon of BALB/c control and DS mice treated with saline or
`h[Gly2]GLP-2. *P ⬍ 0.05 for control vs. mice with DS colitis.
`
`greater proportion of intact mucosal epithelium, and
`this difference was statistically significant for both
`proximal and distal colon (Fig. 4B, P ⬍ 0.05 for proximal
`and distal colon, DS-h[Gly2]GLP-2- vs. DS-saline-
`treated mice). Furthermore, the percentage of histologi-
`cal sections in the proximal colon exhibiting a pathologi-
`cal index (Fig. 4C) of two or three was greater in
`saline-treated mice with DS colitis compared with mice
`treated with h[Gly2]GLP-2 (Fig. 4C, P ⬍ 0.05 for grade
`II lesions, DS-h[Gly2]GLP-2- vs. DS-saline-treated
`mice). Analysis of MPO activity as an indirect indicator
`of neutrophil infiltration demonstrated a marked induc-
`tion of MPO activity in the colon of BALB/c mice with
`DS colitis. No significant differences in the relative
`levels of MPO activity were detected in the colons from
`h[Gly2]GLP-2-treated control or DS mice (Fig. 5).
`To determine the effects of DS colitis and h[Gly2]GLP-
`2 treatment on gene expression in the small and large
`bowel, RNA prepared from CD1 mice in experiment A
`was analyzed by Northern blotting and RT-PCR. Al-
`though the levels of SGLT-1 mRNA were similar in
`jejunum and ileum from CD1 control mice and mice
`
`Fig. 4. A: mucosal surface area in CD1 mouse colon, experiment A.
`*P ⬍ 0.05 for saline-treated vs. either h[Gly2]GLP-2-treated, 5%
`DS-saline-treated, or 5% DS-h[Gly2]GLP-2-treated mice. #P ⬍ 0.05
`for h[Gly2]GLP-2- vs. 5% DS-saline-treated mice. ⫹P ⬍ 0.05 for 5%
`DS-saline- vs. 5% DS-h[Gly2]GLP-2-treated mice. B: mucosal integ-
`rity (solid bar) expressed as percentage of total circumferential
`mucosal area in proximal, middle, or distal colon of CD1 mice,
`experiment A, receiving 5% oral DS and saline or h[Gly2]GLP-2. *P ⬍
`0.05 for saline- vs. h[Gly2]GLP-2-treated mice. C: severity of ulcer-
`ative colitis in histological sections (n ⫽ 27–36 bowel sections) from
`proximal colon scored for degree of inflammation as described in
`METHODS. *P⬍0.05 for DS-saline- vs. DS-h[Gly
`2]GLP-2-treated
`CD1 mice, experiment A. Percentage score refers to percentage of total
`sections examined with normal scores and with grades I–III colitis.
`
`7
`
`
`
`G86
`
`H[GLY2]GLP-2 IN A MURINE MODEL OF COLITIS
`
`Fig. 6. A: Northern blot analysis of proglucagon, sodium-dependent glucose trans-
`porter 1 (SGLT-1), interleukin-1␣, 18S RNA, and keratinocyte growth factor (KGF) in
`small intestine and colon of CD1 mice. S, saline. Relative densitometric values for
`Northern blots from 3 separate mice are shown in B. Relative densitometric signals
`were not corrected for variations in 18S RNA as latter signals were highly comparable
`(⬍10% variation) in each lane.
`
`with DS colitis, a slight decrease in SGLT-1 mRNA was
`detected in the colon of mice with DS colitis, consistent
`with destruction of epithelial mucosa (Fig. 6, A and B).
`In contrast, IL-1 mRNA transcripts, indirect markers
`of intestinal inflammation, were markedly induced in
`the colon of mice with DS colitis (Fig. 6, A and B),
`and the levels of IL-1 mRNA transcripts were clearly
`lower in the colons of mice with colitis treated with
`h[Gly2]GLP-2. The reduction in IL-1 mRNA in
`h[Gly2]GLP-2-treated mice is unlikely due to a dilu-
`
`tional effect secondary to increased bowel mass, as we
`did not observe corresponding reductions in the colonic
`levels of proglucagon, SGLT-1, KGF, and 18S RNAs in
`RNA analyses from the same mice.
`Because both KGF and TGF-␣ have been shown to have
`therapeutic activity in experimental models of intestinal
`inflammation (22, 58) we examined whether the effects of
`h[Gly2]GLP-2 treatment might be mediated in part via
`local induction of these growth factors. Northern blot
`analysis detected induction of colonic KGF gene expres-
`
`8
`
`
`
`H[GLY2]GLP-2 IN A MURINE MODEL OF COLITIS
`
`G87
`
`Fig. 6. C: RT-PCR analysis of transforming growth factor-␣ (TGF-␣) gene
`expression in colon of CD1 mice. Data from 3 different groups of mice are
`shown, following PCR for 26 cycles as described in METHODS. The 28S and
`18S ribosomal bands from each RNA sample used for PCR are shown. kb,
`Relative migration position from electrophoresed size markers of RT-
`PCR products. D: tissue levels of immunoreactive GLP-2 in CD1 mouse
`ileum and colon. *P ⬍ 0.05 for saline-treated vs. 5% DS-saline-treated or
`5% DS-h[Gly2]GLP-2-treated mice. #P ⬍ 0.05 for h[Gly2]GLP-2-treated
`vs. 5% DS-saline-treated or 5% DS-h[Gly2]GLP-2-treated mice.
`
`sion in mice with colitis (Fig. 6A). Nevertheless, there
`were no significant differences in the levels of intestinal
`KGF mRNA in saline- vs. h[Gly2]GLP-2-treated mice
`with DS colitis (Fig. 6A). The levels of TGF-␣ mRNA
`transcripts were easily detectable and comparable in
`control saline- and h[Gly2]GLP-2-treated healthy mice
`(Fig. 6C). In contrast, a marked reduction of TGF-␣
`mRNA was observed in colon RNA from both saline-
`and h[Gly2]GLP-2-treated mice with DS colitis, consis-
`tent with the presence of significant destruction of the
`intestinal epithelium and/or toxicity from DS on TGF-␣-
`producing cell types. Thus no evidence for h[Gly2]GLP-
`2-mediated induction of intestinal KGF or TGF-␣ gene
`expression was observed in these experiments.
`Because previous studies have suggested that some
`forms of small bowel injury are associated with increased
`levels of circulating enteroglucagons (4), we measured
`tissue levels of immunoreactive GLP-2 in the ileum and
`colon and levels of proglucagon mRNA transcripts in
`jejunum, ileum, and colon of control and DS-treated mice.
`No significant change in the levels of proglucagon mRNA
`transcripts was detected in the jejunum, ileum, or colon of
`mice with DS colitis (Fig. 6, A and B). Immunoreactive
`GLP-2 levels were not increased in the ileum; however,
`a significant decrease in the levels of GLP-2 was
`observed in the colon of DS-treated mice, possibly
`
`consistent with the colitis-associated destruction of the
`mucosal epithelium and/or increased secretion and
`decreased storage of tissue GLP-2 (Fig. 6D, P ⬍ 0.05,
`control vs. DS-treated mice). In contrast to the reduc-
`tion in tissue levels of GLP-2, analysis of the levels of
`circulating GLP-2 demonstrated increased levels in
`mice with DS colitis (Fig. 7), with a significant increase
`in plasma GLP-2 observed in the DS mice treated with
`h[Gly2]GLP-2 (P ⬍ 0.05, h[Gly2]GLP-2-treated mice
`without colitis vs. h[Gly2]GLP-2-treated mice with DS
`colitis).
`
`DISCUSSION
`
`The mechanisms responsible for the development of
`ulcerative colitis in human patients remain incom-
`pletely understood and likely include an inappropriate
`immune response to dietary or microbial antigens,
`ultimately leading to cytokine activation and epithelial
`damage. The reproducible induction of experimental
`colitis in rodents following exposure to DS provides an
`opportunity to study the therapeutic efficacy of specific
`interventions on disease progression as assessed by
`histopathology of inflamed tissues in vivo. DS-induced
`rodent colitis pathologically resembles human ulcer-
`ative colitis, with development of mucosal edema, crypt
`erosions, and abscesses, leading to polyp formation and
`ultimately, progression to dysplasia and adenocarci-
`
`9
`
`
`
`G88
`
`H[GLY2]GLP-2 IN A MURINE MODEL OF COLITIS
`
`illustrated by increased ICAM-1 expression in inflamed
`colon and antisense oligonucleotides against ICAM-1
`RNA inhibited ICAM-1 expressi