Several lines of evidence implicate chronic inflammation as a stimulator of the formation of epithelial-derived tumors. This relationship between inflammation and tumorigenesis was shown in various tissues, including the gastrointestinal tract. Patients with chronic inflammatory bowel disease (IBD), especially ulcerative colitis (UC), are at increased risk for colorectal neoplasia and colorectal cancer (CRC) is a serious complication of UC.1 The risk of CRC increases with the extent and duration of UC, which favors the development of epithelial dysplasia. UC-associated CRC progresses from flat dysplastic lesions and especially DALM (dysplasia-associated lesion or mass) and thus differs from sporadic colorectal cancer, which arises predominantly from adenomatous polyps.
The precise mechanism by which chronic mucosal inflammation causes malignancy in UC is not fully delineated, and much remains to be learned regarding the molecular events that facilitate the progression of UC mucosa to dysplasia and carcinoma. Inflammation is characterized among others by the activation of the transcription factor NF-κB, resulting in the expression of various inflammation-regulated genes such as inducible nitric oxide synthase (iNOS) and cyclooxygenase-2 (COX-2) that contribute to the survival of the cells.2, 3 On the other hand, the overproduction of NO via iNOS and prostaglandins via COX-2 contributes to the tissue damage and tumor growth. NO inhibits caspases, the proteases involved in apoptosis, and is able to cause DNA damage and the inhibition of DNA repair mechanisms.4 Similarly, prostaglandins, the products formed by enzymatic activity of COX-2, modulate most cellular processes that govern cell growth, differentiation, and angiogenesis; in addition, the highly reactive intermediates produced by COX-2 may derivatize macromolecules including DNA.2 In colon carcinogenesis, an increased expression of iNOS and COX-2 was found in both human colon adenoma and carcinoma.5–7
However, the sequence normal epithelium-chronically irritated epithelium-dysplasia- carcinoma in the gastrointestinal tract requires the dysregulation of cell proliferation, differentiation, and apoptosis that is associated with expression alterations of not only COX-2 and iNOS but also many other genes. Recent investigations in molecular biology of various cancers including malignancies of the gastrointestinal tract have revealed changes in many genes and their products, such as the family of ras and myc oncogenes, members of the inhibitor of apoptosis protein (IAP) family, elements of the Wnt/β-catenin pathway, metalloproteinases, telomerase, and many others. In particular, the expression of transcription factor c-MYB is increased substantially in colon tumors,8 and, conversely, the downregulation of c-MYB is associated with colon cell differentiation and apoptosis.9 Similarly, survivin, a member of the IAP family, has been shown to be increased in colon cancer compared with adjacent normal mucosa,10, 11 and the expression of integrin-linked kinase (ILK), a key regulator of cellular response to integrin and growth-factor signaling,12 correlates positively with colonic tumor progression.13, 14 Some data also indicate that β-catenin and Tcf-4 genes, which code for factors involved in the Wnt signaling pathway, play a crucial role in many developmental processes of the intestinal epithelium, as well as in colorectal carcinogenesis.15, 16 Similarly, CRC shows increased telomerase activity and mRNA abundance of its catalytic subunit hTERT.17, 18
Although many proproliferative and antiapoptotic genes have been studied as possible molecular biomarkers of neoplasms to reveal their precancerous potential in various types of tumors, including sporadic colorectal neoplasia, their profiles during the multistep progression of inflamed epithelium to the final state of colorectal cancer is poorly understood. The aim of this study was therefore to illuminate the potential relevance of proinflammatory (iNOS, COX-2), and proproliferative and antiapoptotic markers (c-MYB, ILK, survivin, Tcf-4, hTERT) at a precancerous stage of UC and to compare their mRNA abundances in UC-associated and sporadic colon neoplasms in an attempt to define differences and/or similarities between them.
- Top of page
- MATERIALS AND METHODS
Histological analysis revealed that the biopsy samples can be arranged in 4 groups according to the Vienna Classification (Fig. 1): healthy controls, UC patients without dysplasia (category V1), UC patients with indefinite and low-grade dysplasia (category V2+V3), and patients with UC-associated neoplasia (category V5). As shown in Figure 2, all mucosal samples revealed detectable transcripts not only for COX-2 and iNOS but also for survivin, ILK, hTERT, Tcf-4, and c-MYB. When the transcript levels of biopsies were compared, we found in neoplastic mucosa a significant upregulation of all transcripts with the exception of hTERT and ILK. The median expression of COX-2 and iNOS was increased 6.9- and 2-fold, respectively, in UC-associated neoplasia (V5) compared with the V1 group. The neoplastic epithelium also showed marked overexpression of survivin (median: 14.2×), c-MYB (2.8×), and Tcf-4 (52.8×).
Figure 1. Histology of the large intestine from UC patients: (A) negative for dysplasia (category 1), ×40; (B) close-up of (A) ×100; (C) indefinite for dysplasia (category 2), ×100; (D) tubulovillous adenoma with low-grade dysplasia (category 3), ×60; (E) presence of focus of intramucosal carcinoma (category 5), ×40; (F) close-up of (E) ×100; (G) invasive carcinoma (category 5), ×40; and (H) close-up of (G) with demonstration of invasive type of growth, ×100; all stained H&E.
Download figure to PowerPoint
Figure 2. Quantitative RT-PCR analysis of inducible NO-synthase (iNOS), cyclooxygenase-2 (COX-2), transcription factors c-MYB, and Tcf-4, integrin-linked kinase (ILK), telomerase catalytic subunit (hTERT), and survivin in human colonic biopsies of healthy controls (CTRL) and patients with UC negative for dysplasia (V1), with indefinite and low-grade dysplasia (V2+V3), and with neoplasia (V5). In all, 70 biopsies (UC: 42; CTRL: 28) obtained from 13 patients with UC and 13 healthy controls were analyzed, and the data were normalized for TATA box-binding protein gene expression. Median values, 25th and 75th percentiles, and min./max. ranges are denoted by small squares, boxes, and bars, respectively. Kruskal–Wallis ANOVA identified statistical significance in expression of iNOS (P < 0.001), COX-2 (P < 0.002), c-MYB (P < 0.006), Tcf-4 (P < 0.001), and survivin (P < 0.004).
Download figure to PowerPoint
To verify gene expression changes at the protein level, we performed immunohistochemical study of the UC samples. The results of overall immunoreactivity are summarized in Table 3 and Figure 3. Different patterns of immunostaining were observed in our samples using individual antibodies. Nuclear staining was observed in case of c-MYB and Tcf-4. c-MYB detection exhibited zonal pattern of positivity in UC samples without neoplasia, with maximum expression at the bottom of the crypts, decreasing along the longitudinal axis to the luminal region. However, in neoplastic crypts we observed irregular distribution of c-MYB expression. Contrary to c-MYB, positive immunostaining for Tcf-4 was marked in the superficial epithelium of the nonneoplastic crypts, while heterogeneous expression of Tcf-4 was present in the neoplastic crypts. Cytoplasmic staining was evident in COX-2 and iNOS immunodetection. In nonneoplastic crypts iNOS and COX-2 protein was expressed in the majority of less differentiated cells, contrary to weak reaction in goblet cells. In neoplastic epithelium, strong irregular positivity for iNOS and COX-2 protein was demonstrated. Survivin showed predominantly cytoplasmic localization in UC mucosa without neoplasia; however, strong positivity was found in both cytoplasmic and nuclear compartment in neoplastic cells.
Figure 3. Expression of survivin, c-MYB, Tcf-4, COX-2, and iNOS in UC with progressive stages of neoplastic transformation (V1-V5) and in positive controls (sporadic colon cancer, CRC). For further details, see Fig. 2. Original magnification was ×400 (iNOS-V1; survivin-CRC), ×100 (COX2-CRC), or ×200 (other images).
Download figure to PowerPoint
Table 3. Mean Values of Overall Immunoreactivity Scoring Based on the Assessment of the Percentage of Positive Cells in Patients with Ulcerative Colitis
|V1||2.5 (1.0-4.0)||0.0 (0.0-1.0)||1.0 (0.0-2.0)||1.5 (0.0-2.0)||1.0 (1.0-1.0)|
|V2+V3||4.0 (1.0-4.0)||2.5 (1.0-4.0)||3.0 (1.0-4.0)||3.0 (1.0-4.0)||1.0 (0.0-3.0)|
|V5||4.0 (4.0-4.0)||2.0 (1.0-3.0)||3.0 (2.0-4.0)||3.5 (3.0-4.0)||3.0 (3.0-3.0)|
To investigate if any correlation exists between the expression of proinflammatory, proproliferative, and antiapoptotic transcripts, the data were analyzed using the Spearman's rank correlation test. The test revealed a statistically significant positive correlation between the expression levels of the target genes. Levels of mRNA for both COX-2 and iNOS correlated with each other (RS = 0.641; P < 0.001) and with that of mRNA for survivin (COX-2: RS = 0.653; iNOS: RS = 0.529; P < 0.001), c-MYB (COX-2: RS = 0.441; iNOS: RS = 0.425; P < 0.002), ILK (COX-2: RS = 0.424; iNOS: RS = 0.457; P < 0.002), and Tcf-4 (COX-2: RS = 0.592; iNOS: RS = 0.485; P < 0.002), whereas correlations of COX-2 and iNOS with hTERT were not statistically significant. The expression of survivin also showed a relatively strong positive correlation (P < 0.001) with c-MYB (RS = 0.555), ILK (RS = 0.531), and Tcf-4 (RS = 0.623); also strongly correlated were the expressions of c-MYB with ILK (RS = 0.790) and Tcf-4 (RS = 0.601) and the expression of ILK with Tcf-4 (RS = 0.509). However, in the investigated biopsies there was no obvious trend regarding the expression of hTERT and survivin, c-MYB, ILK, or Tcf-4.
In order to further elucidate the potential involvement of survivin, ILK, hTERT, Tcf-4, and c-MYB in neoplasia and colon cancer progression, we also examined their expression in CRC samples. Figure 4 shows that the expression of survivin, c-MYB, hTERT, and Tcf-4 was significantly increased in tumor tissue as compared to the adjacent autologous normal colonic tissue. The median of hTERT expression was enhanced 6.2-fold, c-MYB 5.2-fold, survivin 4.8-fold, and Tcf-4 1.9-fold. Conversely, the ILK expression was not changed between the neoplastic and adjacent tissue. A comparison of iNOS mRNA and especially COX-2 mRNA in the same samples demonstrated that both enzymes are highly upregulated in tumor samples relative to their respective normal tissue samples (median COX-2: 15.3×; iNOS: 3.8×).
Figure 4. Quantitative RT-PCR analysis of iNOS, COX-2, c-MYB, Tcf-4, ILK, hTERT and survivin in samples of sporadic CRC and normal adjacent tissue. For further details, see Fig. 2. Mann–Whitney U-test identified statistical significance in expression of iNOS (P < 0.003), COX-2 (P < 0.001), c-MYB (P < 0.001), Tcf-4 (P < 0.03), hTERT (P < 0.001), and survivin (P < 0.001).
Download figure to PowerPoint
The specificity of gene expression studies in tissue samples is limited by heterogeneous cell populations; therefore, to overcome the problem of stromal cell contamination in homogenized tissue samples and to determine which genes are activated or damped down in the malignant epithelial compartment, we also studied the expression of the investigated gene panel in a xenograft model of primary tumors induced by subcutaneous injection of colon cancer SW620 cells into nu/nu mice. This procedure allowed for the analysis of the investigated transcripts in neoplastic human cells, eliminating the contribution from murine stroma. The mRNA transcripts of all investigated genes were found to be present in the xenografts of human tumors (Table 4). Whereas the abundance of c-MYB mRNA was not elevated in murine tumor in comparison with nonmalignant human colonic samples, the levels of survivin, ILK, hTERT, Tcf-4, COX-2, and iNOS were significantly upregulated.
Table 4. Relative Expression of Survivin, c-MYB, ILK, hTERT, Tcf-4, COX-2, and iNOS Measured by Quantitative RT-PCR in Primary Tumors of SW620 Cells in Athymic Mice and in Normal Human Colonic Samples
|Tumor||6||5.11 (1.99-15.45)||0.25 (0.22-0.71)||5.39 (2.20-10.48)||0.33 (0.12-1.69)||0.28 (0.07-0.90)||0.51 (0.33-2.22)||0.75 (0.19-1.52)|
|Colon||6||0.27 (0.05-0.95)||0.44 (0.05-0.64)||1.43 (0.32-2.06)||0.02 (0.00-0.03)||0.02 (0.01-0.06)||0.27 (0.03-1.14)||0.16 (0.03-0.23)|
- Top of page
- MATERIALS AND METHODS
CRC is a serious complication in patients with UC, and chronic inflammation accompanied with changes of colonic epithelium appears to be a key predisposing factor. As summarized by Itzkowitz and Yio,1 the risk for carcinoma development increases with duration and extent of colitis, whereas antiinflammatory drugs such as 5-aminosalicylates are able to reduce the risk for colorectal dysplasia. Identification of disease activity at the early state of progression from inflamed epithelium to the final stage of colorectal carcinoma is difficult, as is the histologic diagnosis of the initial dysplastic/neoplastic changes. Therefore, various genes and proteins have attracted attention as promising markers for prediction of early stages of malignant transformation. However, they have been previously studied predominantly in sporadic tumorigenesis and much less in UC-associated tumorigenesis.
In this study we analyzed gene expression of proproliferative and antiapoptotic genes, which have been associated with early stages of tumorigenesis, and compared their expression levels between sporadic CRC and chronic colitis of patients with and without detected neoplastic lesions. Although the detailed mechanisms of the UC-related tumorigenesis are largely unclear, 1 important mechanism seems to be the generation of NO through iNOS followed by the formation of peroxynitrite, nitrates, and nitrites that may exert oncogenic effects through several mechanisms.24 Our results demonstrate that the expression of iNOS is significantly increased not only in bioptic samples of neoplastic colonic mucosa of UC patients but also in CRC and in tumor xenografts; moreover, our results are in accordance with previous findings of increased expression and activity of iNOS in patients with IBD and colonic tumors.5, 25 The highest expression level of iNOS in human tumor xenograft indicates that iNOS expression is localized not only to surrounding inflammatory cells but also to malignant epithelial cells, as was shown in a murine dextran sodium sulfate model of UC.26
The upregulation of iNOS was similar to COX-2, the key enzyme in the conversion of arachidonic acid to prostaglandins, which promote cell proliferation and angiogenesis, inhibit the induction of apoptosis, and may enhance neoplastic progression.2, 27 Our study demonstrates that COX-2 overexpression in UC-associated neoplasia is much higher than in nonneoplastic mucosal samples of patients with chronic UC in the nonactive stage; additionally, this is in agreement with the concept of the tumor growth-promoting role of prostaglandin derivatives.2 This upregulation of COX-2 is localized not only to stromal but also to epithelial cells.28 One pathway that can activate gene targets involved in tumorigenesis, such as COX-2 and the oncogene c-MYC, requires the transcription factor c-MYB,9, 29 whose inappropriate expression was found in tumors of rodents and humans.30 Our results suggest that c-MYB and COX-2 are overexpressed not only in developed CRC but also in UC-associated neoplasia when c-MYB shows relatively strong positive correlation with other transcripts participating in the neoplastic transformation, such as survivin, ILK, and Tcf-4. In agreement with previous findings,31 there was no apparent overexpression of c-MYB in SW620 cells.
The hyperactivation of the Wnt/β-catenin signaling pathway is a typical feature of colon cancer when the stabilization of β-catenin and its complexes with various nuclear Tcfs results in the expression of a number of genes that have been implicated in oncogenesis.32 In agreement with this model for colorectal carcinogenesis, we found activation of the Wnt-Tcf pathway at the level of T-cell factor 4. One of the important observations made by this study is the identification of Tcf-4 upregulation not only in developed CRC15 but also in preneoplastic lesions of patients with UC.
In agreement with the fact that survivin is a target gene of the Wnt/β-catenin signaling pathway and Tcf-4 is 1 of the major transcription factors involved in the regulation of survivin promoter activity,33 we observed strong upregulation of survivin mRNA and protein expression not only in sporadic CRC, as was also shown by others,10, 11 but also in UC-associated neoplasia. Thus, the upregulation of survivin, an inhibitor of apoptosis and mitotic regulator, may promote the progression of cellular dysplasia from chronically irritated mucosa to carcinoma.
Another protein whose activation is thought to play an important role in permanent cell growth and immortality is telomerase. Although the activity of this enzyme was recently shown to be upregulated by augmenting the gene transcription of hTERT by survivin,34 contradictory results have been reported for the gastrointestinal tract. The majority of studies found upregulation of telomerase activity and mRNA expression in colorectal tumors,17, 35–37 while others found increased activity only in late events during carcinogenesis but not in preneoplastic lesions.38, 39 In this context, contradictory results have been found in patients with UC.40–42 In our study, the level of hTERT mRNA was significantly increased in CRC samples and in human tumor xenografts but not in the samples of UC patients with and without neoplasia. Thus, our quantitative RT-PCR data, together with previous activity measurements,40, 41 indicate that upregulation of hTERT mRNA does not occur in the early stage of inflammation-dysplasia-carcinoma progression but seems to be associated with the late-stage development of CRC, as suggested by Kanamaru et al.39
Similarly, we can assume that prominent upregulation of the ILK expression occurs during the later stages of colon cancer progression, especially in malignant cell compartments endowed with high invasivity and metastatic potential.43 This is supported by 3.8-fold enhanced expression of ILK in xenograft tumors of SW620 malignant cells originally derived from metastatic colon carcinoma.
In summary, we used a quantitative RT-PCR approach and immunohistochemistry to clarify whether the expression of a panel of important proproliferative and antiapoptotic genes connected with tumorigenesis is changed during the early stages of transition of UC to epithelial neoplasia and whether these molecular events are analogous to those described in CRC. We have demonstrated for the first time the quantitative relation of these genes between neoplastic and nonneoplastic mucosa of patients with chronic inflammation, and we showed the neoplasia-specific alterations in the expression of survivin, c-MYB, Tcf-4, COX-2, and iNOS in patients with chronic UC. These quantitative differences of transcripts in mucosal biopsies offer new possibilities for diagnostic purposes, and further studies to evaluate the significance of these genes as a potential multimarker panel of tumorigenesis in patients with chronic UC are currently under way in our laboratory.