Although the precise cause of ulcerative colitis and Crohn disease remains unclear, experimental and clinical data suggest that proinflammatory cytokines may mediate mucosal damage (1–3). Recently, elevated serum and mucosal levels of several potent angiogenic factors, including vascular endothelial growth factor (VEGF) (4) and β-fibroblast growth factor (5), have been described in patients with ulcerative colitis and Crohn disease. Angiogenesis-promoting cytokines may be a double-edged sword that may promote inflammation by increasing vascular permeability, but may also mediate tissue repair by activating fibroblasts and promoting wound healing. Another angiogenesis factor not yet thoroughly studied but of potential importance in ulcerative colitis and Crohn disease is hepatocyte growth factor or scatter factor (HGF/SF) (1,6). HGF/SF is produced by fibroblasts, monocytes, and endothelial cells (6,7). Hepatocyte growth factor stimulates proliferation and reconstitution of mucosal cells, causes new vessel formation, and influences ultimate epithelial cell phenotype. Hepatocyte growth factor is required for gastrointestinal development and repair (8,9), for the formation of the pancreatic duct system (10), and for beta-cell proliferation. Supplemental HGF, locally or systemically administered, improves healing of ulcers (11). Mucosal HGF content is also greatly increased in colon and gastric cancer tissue specimens (7,12), suggesting a potential trophic effect on gastrointestinal tumors.
The role of HGF in the pathogenesis of inflammatory bowel disease (IBD) has not been shown. One adult study found no differences in HGF levels in IBD patients compared with controls (13). Our study, performed on a large cohort of pediatric and young adult patients with ulcerative colitis and Crohn disease, determined serum levels of HGF compared with controls using ELISA. We investigated the effect of specific diagnosis, disease activity, age, and medication use on HGF levels. We hypothesized that HGF, produced in the inflamed gastrointestinal tract, would be increased in the serum of IBD patients, and levels would correlate with disease activity.
Serum samples were obtained from 129 patients evaluated and treated in the Children's Hospital pediatric gastroenterology program between November 1992 and June 1999. Sixty patients had Crohn disease, 31 had ulcerative colitis, and 38 were control patients with other gastrointestinal illnesses. One serum sample was obtained per patient.
Patient characteristics are outlined in Table 1. To determine disease severity for patients with Crohn disease, the Pediatric Crohn Disease Activity Index (PCDAI) was used (14). A PCDAI score of 0 to 10 denotes inactive disease, 11 to 30 denotes mild disease, and 31 or greater denotes moderate to severe disease. For the ulcerative colitis patients, disease severity was measured using the clinical score of Kozarek et al. (15). This score is derived from five parameters: number of liquid stools daily, degree of rectal bleeding, number of extraintestinal symptoms, abdominal pain, and well-being. We defined ulcerative colitis patients with a Kozarek score of 0 to 3 as having inactive disease, of 4 to 6 as having mild disease, and a score greater than 6 as having moderate to severe disease. The medications taken by Crohn disease and ulcerative colitis patients at the time of the study are displayed in Table 2.
The 38 control patients had gastrointestinal symptoms but no evidence of inflammatory bowel disease. The largest group consisted of children diagnosed with functional abdominal pain (n = 19) and no evidence of organic disease. The rest of the patients (n = 19) included individuals evaluated by our gastroenterology program with other gastrointestinal illnesses. These included patients with diarrhea (n = 8), esophagitis (n = 4), growth failure (n = 1), irritable bowel syndrome (n = 4), nausea (n = 1), and rectal bleeding (n = 1). Blood samples were obtained during office visits, endoscopies, or hospitalizations. All samples were spun within 1 hour of collection and serum frozen at −70°C until the assays were performed. Informed consent was obtained from each participant under a protocol approved by the Children's Hospital Committee on Clinical Investigation.
Serum assays of HGF were determined by enzyme-linked immunosorbent assay (ELISA) using the Quantikine HS Human HGF immunoassay (R&D Systems, Minneapolis, Minnesota, USA). Fifty microliters of serum were added to wells precoated with a murine monoclonal antibody against HGF. After incubating and washing off unbound proteins, 200 μL of polyclonal antibody conjugated to horseradish peroxidase were added to each well. After another period of incubation, substrate solution and stop solution were added. The colorimetric optical density was then read at 450 nm and compared to a standard curve to determine the HGF concentration in each serum sample. Each serum sample was measured in duplicate, and the value obtained represents the mean of the two samples.
Results are expressed as mean value ± the standard error of the mean (SEM). Departures from normality in the distribution of HGF for each IBD group and the controls were assessed by the Wilk-Shapiro test. Differences between more than two mean values were assessed using the F test in the analysis of variance followed by the Fisher least significant difference test for multiple comparisons. Pearson correlations were calculated to measure linear association between continuous variables. For patients with Crohn disease, multiple linear regression analysis was performed to identify variables that were independently predictive of HGF level using a backward stepwise method with P < 0.10 for entry and P < 0.05 for variables to be retained in the final model. The coefficient of determination (R 2 ) was used to indicate the proportion of variance in HGF accounted for by the set of predictors in the final model. Statistical significance was set at a two-tailed alpha level of 0.05. Data analysis was conducted with the SPSS/PC statistical package, version 8.0 (SPSS Inc., Chicago, IL, USA).
Serum HGF Levels in IBD Patients and Controls
Figure 1 illustrates scatter plots for the three groups studied (Crohn disease, ulcerative colitis, and controls) with their corresponding serum HGF distribution results. Serum HGF levels (mean ± SEM) were 1,439 ± 84 pg/mL for Crohn disease patients, 1,384 ± 107 pg/mL for ulcerative colitis patients, and 807 ± 50 pg/mL for control patients. Mean HGF levels were significantly higher in the Crohn disease and ulcerative colitis groups than in the control group (P < 0.001 in each case). However, mean HGF levels were did not differ between the ulcerative colitis and Crohn disease group (P = 0.99). In addition, because of the wide spread of the data, HGF levels do not allow a clinician to reliably discriminate inflammatory bowel patients from controls based on serum HGF.
Table 3 lists mean (±SEM) HGF levels and ranges categorized by age and sex for the three study groups. In patients with Crohn disease and ulcerative colitis, age had no influence on serum HGF levels. However, control patients younger than 10 years had significantly lower HGF levels than older patients (621 ± 43 pg/mL for patients <10 years vs. 882 ± 63 pg/mL for patients >10 years, P = 0.02). Sex had no effect on serum HGF in Crohn disease patients, ulcerative colitis patients, or controls.
We evaluated the effect of disease location on HGF levels in Crohn disease and ulcerative colitis. For patients with Crohn disease of the small bowel only (n = 7), large bowel only (n = 12), or small and large bowel (n = 41) mean (±SEM) HGF levels were 1,094 ± 101 pg/mL, 1676 ± 211 pg/mL, and 1428 ± 102 pg/mL, respectively. For patients with ulcerative colitis proctitis, left-sided colitis, and pancolitis, mean HGF levels were 913 ± 103 pg/mL (n = 5), 1,580 ± 454 pg/mL (n = 3), and 1,460 ± 125 pg/mL (n = 23). Analysis of variance revealed no differences between HGF levels and disease location for either Crohn disease or ulcerative colitis. However, the small sample size in each group limits our ability to make meaningful conclusions.
Effect of Disease Activity on Serum HGF
In patients with Crohn disease mean levels of serum HGF were higher with increasing disease activity (Fig. 2). Patients with inactive disease (PCDAI < 10, n = 19) had a mean (±SEM) HGF level of 1,147 ± 87 pg/mL, those with mild disease (PCDAI = 11–30, n = 29) had a level of 1,457 ± 106 pg/mL, and those with moderate to severe disease (PCDAI > 30, n = 12) had a level of 1,858 ± 269 pg/mL. Serum HGF for patients with moderate to severe Crohn disease was significantly higher than the HGF for patients with inactive disease (P < 0.01). No significant differences were identified between patients with inactive vs. mild disease, or mild vs. moderate to severe disease (P > 0.20 in each case). All three groups of patients (inactive, mild, and moderate to severe Crohn disease) had significantly higher HGF levels than did control patients (P < 0.001).
Mean serum HGF levels were also higher with increasing disease activity in ulcerative colitis. Patients with inactive disease (Kozarek score = 0–3, n = 19) had a mean (±SEM) HGF level of 1,121 ± 100 pg/mL, those with mild disease (Kozarek score = 4–6, n = 29) had a mean HGF level of 1,219 ± 104 pg/mL, and those with moderate to severe disease had a mean HGF level of 1,781 ± 235 pg/mL. Patients with moderate to severe ulcerative colitis had significantly higher HGF levels than did patients with inactive ulcerative colitis (P = 0.02). Again the mean HGF level for mild disease was between the levels for inactive and moderate to severe disease, and did not significantly differ from the other two groups. In addition, all three groups of ulcerative colitis patients had HGF levels significantly higher than those of the control patients (P < 0.01).
Relation Between Serum HGF and Other Laboratory Markers
Correlation coefficients between serum HGF and other laboratory parameters commonly measured in patients with Crohn disease are shown in Table 4. Serum HGF levels significantly correlated with erythrocyte sedimentation rate (P = 0.02) and PCDAI (P = 0.01). As expected, there was a positive correlation between erythrocyte sedimentation rate and PCDAI, and negative correlations between albumin versus PCDAI and hematocrit versus PCDAI.
To evaluate whether serum HGF correlated with disease activity independently of other variables, we constructed a multivariate linear regression model to identify predictors of disease activity. We used PCDAI as the outcome variable, and sedimentation rate, hematocrit, albumin, HGF level, weight, height, age, and sex as candidate predictors. The final regression model included two predictors: serum albumin and serum HGF level (adjusted R 2 = 0.44, P < 0.01). The multivariate model suggests serum HGF correlates with disease activity in Crohn disease, even after adjusting for other variables.
This study demonstrates that patients with ulcerative colitis and Crohn disease have significantly increased systemic HGF levels compared with healthy controls, and that serum HGF concentration correlates directly with disease activity. These findings suggest that HGF may play a role in the pathogenesis of ulcerative colitis and Crohn disease and their complications. Alternatively, the raised serum HGF identified may be an epiphenomenon of inflammation with HGF release caused by other proinflammatory cytokines. This study's results differ from the findings of Sturm et al. (13), who did not identify a statistically significant difference in serum HGF levels between IBD patients and controls. However, in the Sturm study, mean HGF was 497 pg/mL in controls (n = 28), 865 pg/mL in ulcerative colitis patients (n = 29), and 683 pg/mL in Crohn patients (n= 45). The significant differences detected in our study may reflect the larger sample size. Nevertheless, the extensive overlap between controls and IBD patients makes this cytokine of no use as a diagnostic marker.
Hepatocyte growth factor is a protein consisting of two polypeptide chains of molecular weights 69 kd and 34 kd. Fibroblasts are the principal cellular source of HGF in humans. The HGF receptor, c-MET, is present on neurons, hematopoietic precursors, endothelial cells, and epithelial cells. Several isoforms of c-MET also exist, allowing different effects of HGF specific to target cell type (16). Depending on the cell to which HGF binds, a diverse group of cellular actions may result (16). These actions include increased cellular mitosis, alterations in changes in cell shape, and invasion into the basement membrane (16). Secreted HGF binds with low affinity to matrix proteoglycans such as heparin, providing a potential reserve of growth factor in the interstitia that can be proteolytically freed.
The elevated HGF levels identified in the serum may be secondary to bowel inflammation, with other proinflammatory cytokines inducing HGF release by monocytes and endothelial cells. Increased serum cytokines (TNF-α, IL-1, IL-6) (17,18) correlate with disease activity in ulcerative colitis and Crohn disease. The systemic release of these cytokines may contribute to clinical symptoms exhibited by patients, such as anorexia, cachexia, and fever (19). Both IL-6 and IL-1 increase HGF expression (19). Plasma HGF may also be increased in other conditions causing systemic inflammation, such as arthritis (20). When released, HGF may interact synergistically with other cytokines such as IL-8 or VEGF to increase mucosal blood flow and immune cell recruitment. IL-8 can also function as an angiogenesis factor and synergistically interact with others, such as VEGF or HGF, to increase neovascularization.
In addition to mediating some of the effects of inflammation, HGF may also play a role in the reconstitution of the damaged mucosa in ulcerative colitis and Crohn disease (21). Such a mechanism could help to explain the observed exacerbation of some patients with ulcerative colitis and Crohn disease given nonsteroidal antiinflammatory drugs (22,23) for treatment of arthritis or other conditions. Hepatocyte growth factor mRNA is increased at the site of active disease in the mucosa of ulcerative colitis patients, as it is in gastric ulcers. Nonsteroidal antiinflammatory drugs damage normal gastric mucosa and delay peptic ulcer healing by decreasing prostaglandin E2 and HGF production by fibroblasts (22,23). Studies of gastric ulcer healing in rats show that exogenous HGF supplementation increases epithelial cell proliferation at the ulcer margin and improves healing (24). This suggests that an increase in mucosal HGF might benefit tissues injured in Crohn disease and ulcerative colitis.
Alternatively, mucosal HGF might be a factor in promoting colonic malignancy in Crohn disease and ulcerative colitis. Previous studies suggest an important role for HGF in epithelial malignancies, particularly colorectal cancer (23). Intestinal epithelial cell lines Caco-2 and T84 express high levels of c-MET and increase their proliferative rate with HGF (23), suggesting potential roles for HGF in the development or progression of colorectal cancer in ulcerative colitis and Crohn disease.
In summary, serum HGF is elevated in patients with IBD compared with controls. Elevation of this and other angiogenesis factors in IBD patients suggest that these factors may mediate vascular permeability and wound healing in patients with IBD. To clarify whether HGF has a role in predisposing the IBD patient to malignancy, future studies will need to evaluate HGF levels in intestinal mucosa and to see whether HGF promotes cancer formation in animal models.
1. Beck PL, Podolsky DK. Growth factors in inflammatory bowel disease
. Inflamm Bowel Dis 1999; 5: 44–60.
2. Fiocchi C. Inflammatory bowel disease
: etiology and pathogenesis. Gastroenterology 1998; 115: 182–205.
3. MacDermott RP. Chemokines in the inflammatory bowel disease
. J Clin Immunol 1999; 19: 266–72.
4. Bousvaros A, Leichtner AM, Zurakowski D, et al. Elevated serum vascular endothelial growth factor in children and young adults with Crohn's disease. Dig Dis Sci
5. Bousvaros A, Zurakowski D, Fishman SJ, et al. Serum basic fibroblast growth factor in pediatric Crohn's disease; implications for wound healing. Dig Dis Sci 1997; 42: 378–86.
6. Jiang WG, Hiscox S. Hepatocyte growth factor
, a cytokine playing multiple and converse roles. Histol Histopathol 1997; 12: 537–55.
7. To CT, Tsao MS. The roles of hepatocyte growth factor
/scatter factor and met receptor in human cancers. Oncol Rev 1998; 5: 1013–24.
8. Kato Y, Yu D, Schwartz MZ. Enhancement of intestinal adaptation by hepatocyte growth factor
. J Pediatr Surg 1998; 33: 235–9.
9. Kermorgant S, Walker F, Hormi K, et al. Developmental expression and functionality of hepatocyte growth factor
and c-Met in human fetal digestive tissues. Gastroenterology 1997; 112: 1635–47
10. Vila MR, Nakamura T, Real FX. Hepatocyte growth factor
is a potent mitogen for normal human pancreas in vitro. Lab Invest 1995; 73: 409–18.
11. Takahashi M, Ota S, Hata Y, et al. Hepatocyte growth factor
as a key to modulate anti-ulcer action of prostaglandins in stomach. J Clin Invest 1996; 98: 2604–11.
12. Nabeshima K, Shimao Y, Inoue T, et al. Hepatocyte growth factor
/scatter factor induces not only scattering but also cohort migration of human colorectal-adenocarcinoma cells. Int J Cancer 1998; 78: 750–59.
13. Sturm A, Schulte C, Schatton R, et al. Transforming growth factor beta and hepatocyte growth factor
levels in patients with inflammatory bowel disease
. Eur J Gastroenterol Hepatol 2000; 12: 445–50.
14. Hyams JS, Ferry GD, Mandel FS, et al. Development and validation of a pediatric Crohn's disease activity index. J Pediatr Gastroenterol Nutr 1991; 12: 439–47.
15. Kozarek RA, Patterson DJ, Gelfand MD, et al. Methotrexate induces clinical and histologic remission in patients with refractory inflammatory bowel disease
. Ann Intern Med 1989; 110: 353–6.
16. Comoglio PM. Structure, biosynthesis and biochemical properties of the HGF receptor in normal and malignant cells. EXS 1993; 65: 131–65.
17. Noguchi M, Hiwatashi N, Liu Z, et al. Secretion imbalance between tumor necrosis factor alpha and its inhibitor in inflammatory bowel disease
. Gut 1998; 43: 203–9.
18. Reimund JM, Witterscheim C, Dumont S, et al. Increased production of tumor necrosis factor-α, interleukin 1-β and interleukin-6 by morphologically normal intestinal biopsies from patients with Crohn's disease. Gut 1996; 39: 684–89.
19. Sakon, M, Kita Y, Yoshida T, et al. Plasma hepatocyte growth factor
levels are increased in systemic inflammatory response syndrome. Surg Today 1996; 26: 236–41.
20. Yuioka K, Inaba M, Furumitsu Y, et al. Levels of hepatocyte growth factor
in synovial fluid and serum of patients with rheumatoid arthritis and release of hepatocyte growth factor
by rheumatoid synovial cells. J Rheumatol 1994; 21: 2184–9.
21. Taha AS, Curry GW, Morton R, et al. Gastric mucosal hepatocyte growth factor
in Helicobacter pylori gastitis and peptic ulcer disease. Am J Gastroenterol 1996; 91: 1407–9.
22. Bamba H, Ota S, Kato A, et al. Nonsteroidal anti-inflammatory drugs may delay the repair of gastric mucosa by suppressing prostaglandin-mediated increase of hepatocyte growth factor
production. Biochem Biophys Res Comm 1998; 245: 567–71.
23. Ota S, Tanaka Y, Bamba H, et al. Nonsteroidal anti-inflammatory drugs may prevent colon cancer through suppression of hepatocyte growth factor
expression. Eur J Pharmacol 1999; 367: 131–8.