| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Published online before print December 28, 2007
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
From the Translational Oncology Unit Consejo Superior de Investigaciones Cientificas-Universidad Autónoma de Madrid-La Paz,*Centro Nacional de Biotecnología, Madrid; Grup d’Oncogenesi i Antitumorals,Laboratori d’Investigació Gastrointestinal, Institut de Recerca, Hospital de la Santa Creu i Sant Pau, Barcelona; and the Department of Pathology,Hospital Universitario La Paz, Madrid, Spain.
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionReferences |
|---|
The small guanosine triphosphatases (GTPases) Rho proteins compose an important subgroup within the Ras superfamily. To date, 22 Rho family members have been identified, divided into six groups based in homology and/or function.6
The three prototypic Rho GTPases, RhoA, Rac1, and Cdc42, are known for their effects on the actin cytoskeleton7-11
and their regulatory role on adhesion, motility, cell cycle, apoptosis, transcriptional regulation, and membrane trafficking. These functions are mediated through several signaling pathways, like c-Jun NH2-terminal kinase, p38, and phosphatidylinositol 3-kinase, as well as transcription factors, such as serum response factor and nuclear factor-
B.12-14
Their ability to act as general modulators of several cellular processes results in important roles in tumor biology and other human diseases, including cancer.15-17
Rho GTPases influence a variety of processes in cancer, including cell transformation, survival, invasion, metastasis, and angiogenesis.15,16 Accumulating evidence indicates that Rac1-dependent cell signaling is important for malignant transformation. Thus, Rac1 overexpression has been found in breast carcinoma, gastric carcinoma, oral squamous cell carcinoma, non-small cell lung carcinoma, and testicular germ cell tumors.16 Rac1 plays an important role in leptin-mediated invasiveness of colonic epithelial cells18 and in colorectal tumors truncated APC proteins stimulate the activities of Asef, a Rac-specific guanine exchange factor.19 Cross-talk between Rac1 and the dysregulated Wnt signaling pathway in colon adenocarcinoma cells has been found.20-22
We have investigated whether Rac1 plays a role in the tumorigenic properties of human colorectal adenocarcinoma cells using an orthotopic model that closely reproduces the clinical behavior of human colorectal adenocarcinoma.23 Overexpression of Rac1 in SW620 cells, derived from a lymph node metastasis of a primary human colorectal adenocarcinoma, resulted in the acceleration of the tumorigenic process. By contrast, specific inhibition of Rac1 expression by RNA interference completely suppressed tumor formation. These results demonstrate a decisive role of Rac1 in colorectal tumorigenesis.
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Short hairpin RNA (shRNA) against human RAC1 was designed according to Agami and co-workers.24 The 64-nucleotide DNA template oligonucleotides were engineered to produce a 19-nucleotide sequence derived from the target transcript flanked with AA at the 5' end and TT at the 3' end and comprising more than 30% GC content. Candidate sequences were submitted to a BLAST search against the human genome sequence to ensure that only one gene of the human genome was targeted. The oligonucleotide pair was designed to produce restriction site overhangs on annealing (BglII at the 5' end and HindIII at the 3' end) for cloning into vector pSUPER-neo (sense primer, 5'-GATCCCCTATATCCCTACTGTCTTTGTTCAAGAGACAAAGACAGTAGGGATATATTTTTGGAAA-3'; and antisense primer, 5'-AGCTTTTCCAAAAATATATCCCTACTGTCTTTGTCTCTTGAACAAAGACAGTAGGGATATAG-GG-3').24 The underlined characters indicate the mRNA target sequences, and the bold characters refer to the 9-bp hairpin loop and restriction site overhangs.
Cell Culture, Constructs, and Transfection
The human colorectal adenocarcinoma cell line SW620 was cultured in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum, 2 mmol/L glutamine, and 100 U/ml penicillin-streptomycin (Gibco-BRL Life Technologies, Paisley, UK) at 37°C, in 95% humidity and 5% CO2 pressure.
Construction of pCDNA3B-neo directing expression of wild-type (WT) Rac1 protein and conferring neomycin resistance has been previously described.25 For stable transfection, SW620 cells were plated in six-well dishes at 60 to 70% confluence. Transfections were performed with Lipofectamine Plus (Invitrogen, Carlsbad, CA) as directed by the manufacturer. The amount of plasmid DNA was kept constant at 3 µg per dish with the corresponding empty vector. After transfection, cells were cultured for 24 hours in fresh medium before selection in medium containing 2.5 mg/ml neomycin (Sigma Chemical, St. Louis, MO) for 12 days. Two independent clones were selected for each shRNA vector (i1Clone and i2Clone) or GTPase expression vector (wt1Clone and wt2Clone).
Immunoblotting and Antibodies
For detection of Rac1, whole-cell lysates were prepared in lysis buffer (50 mmol/L NaF,10 mmol/L Na4P2O7, 5 mmol/L EDTA, 50 mmol/L Tris-HCl, pH 7.5, 0.5% Triton X-100, 10 mmol/L NaPPi, and 0.5% sodium deoxycholate) containing protease and phosphatase inhibitors: 200 µmol/L orthovanadate, 1 mmol/L dithiothreitol, 2 µg/ml leupeptin, 2 µg/ml aprotinin, 2 µg/ml pepstatin, and 400 µmol/L phenylmethylsulphonyl fluoride. Ten to 20 µg of total protein was analyzed by SDS-electrophoresis on 15% polyacrylamide gels (BioRad Laboratories, Hercules, CA). After transfer of proteins to Immobilon-P polyvinylidene difluoride membranes (Millipore, Bedford, MA), the blots were incubated with the corresponding antibodies. Immunocomplexes were visualized by enhanced chemiluminiscence detection (Amersham Pharmacia Biotech, Buckinghamshire, UK) using an anti-mouse antibody conjugated to peroxidase (Santa Cruz Biotechnology, Santa Cruz, CA). Mouse monoclonal anti-Rac1 was purchased from Upstate Biotechnology (Charlottesville, VA), and mouse monoclonal anti-glyceraldehyde-3-phosphate dehydrogenase was purchased from Chemicon (Temecula, CA). A quantification of Rac1 protein was carried out by densitometry analysis performed by NIH Image software 1.62 (public domain image processing and analysis program for the Apple Macintosh developed at the U.S. National Institutes of Health).
In Vivo Tumorigenesis by Orthotopic Cell Microinjection
Stably transfected tumor cells (1 x 106) were resuspended in 50 µl of culture medium (Dulbecco’s modified Eagle’s medium) and injected directly into the cecal wall of Swiss Nu/Nu male mice, using the orthotopic cell microinjection procedure, as previously described.23 We studied oncogenesis in five animal groups: the SW620 parental cell line (n = 9), the two Rac1 overexpressor clones [wt1Clone (n = 8) and wt2Clone (n = 8)] and the two Rac1-interfered clones [i1Clone (n = 8) and i2Clone (n = 8)]. We recorded tumor take rate and survival time. Mice were monitored every week, and animals died naturally or were sacrificed when signs of advanced disease were detected.
Pathology Procedures
Complete necropsy examinations of each animal were performed. Any external anomaly and the weight of each animal were recorded. The cervical, thoracic, abdominal, and pelvic organs were extracted en bloc, and the presence of tumor was assessed macroscopically in every organ. All tumor foci were measured, and the presence of any other macroscopic anomalies was recorded. Small samples of tumor were obtained and frozen for biochemical studies (see below). The whole bloc of organs was fixed in 4% formaldehyde in phosphate buffer solution for 24 hours. The brain was also extracted and fixed in the same solution. After fixation, the organs were dissected from the whole bloc, and appropriate photographs of the tumor and the present anomalies were taken with a Nikon Coolpix 5700 digital camera. The tumor and every organ were paraffin embedded in toto. Six-micrometer-thick sections from each paraffin bloc were obtained and stained with H&E. The sections were examined microscopically to describe the tumor and detect its presence in organs in which it was not seen macroscopically. The micrographs shown in Figures 3 to 5
were made using a digital camera (AxioCam HRc; Zeiss, Oberkochen, Germany) attached to a Zeiss Axiophot microscope. The tumor was staged in each animal as follows: local tumor growth at the site of injection was stage I; stage I plus foci of tumor in the visceral and parietal peritoneum was stage II; stage I or II plus mesenteric lymph node and/or pancreatic metastasis was stage III; stage I, II, or III plus hepatic and/or lung metastasis was stage IV.23
Any other microscopic anomalies were recorded. These anomalies, together with the macroscopic ones, served to establish the cause of death in each animal. Sections from selected blocs containing the tumor were also stained with periodic acid-Schiff reaction and immunohistochemistry with the following antibodies: anti-Rac1 (Upstate Biotechnology); anti-AE1/AE3, anti-cytokeratin 7, anti-cytokeratin 20, anti-cytokeratin high molecular weight, anti-epithelial membrane antigen, and anti-vimentin (Progen Biotechnik GMBH, Heidelberg, Germany); anti-cytokeratin low molecular weight (Enzo Life Sciences, New York, NY); and anti-carcinoembryonic antigen (Novocastra, Newcastle, UK). Negative controls were achieved by substituting the primary antibodies with preimmune sera from the species in which they had been raised.
|
The frozen tumor samples from mice were homogenized and lysed with detergent solution (1.5 mmol/L MgCl2, 0.2 mmol/L EDTA, 0.3 mol/L NaCl, 25 mmol/L HEPES, 20 mmol/L β-glycero-phosphate, and 0.1% Triton X-100) containing protease and phosphatase inhibitors mentioned above. The extracts were treated with 1.5 µl (375 U) of Benzonase from Calbiochem (San Diego, CA), for degradation of nucleic acids, which leads to higher resolution in 2DE-gels.
Statistical Analysis
Statistical significance in tumor take rate or survival time between a control group (parental SW620) and each of the genetically modified groups (stably transfected SW620 cells) was evaluated. The Exact Fisher Test was used to statistically analyze the differences in tumor take rates. The likelihood of survival among groups was estimated according to the method of Kaplan and Meier, and survival distributions were determined using the Log Rank Test. P < 0.05 was considered to have statistical significance.
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
To overexpress Rac1, a eukaryotic expression vector containing the cDNA of the wild-type gene was used (pcDNA-RAC1wt), and the small interference RNA technique was used to specifically knockdown gene expression of RAC1. Two neomycin-resistant stable clones from each modulation were selected, wt1Clone and wt2Clone, and i1Clone and i2Clone, respectively, according to the total Rac1 protein levels, determined by Western blot (Figure 1)
.
|
. The modest increase in Rac1wt expression in stable clones is likely due to high intrinsic Rac1 levels in the parental cells. Overexpression or inhibition of Rac1 induced strong phenotypic changes in cells (Figure 1B)
. Furthermore, the apparently moderate levels of overexpression were sufficient to induce genetic changes determined by their transcriptomes.22 Tumorigenic Process Acceleration Induced by Rac1wt
To examine the effect of increased or decreased Rac1 protein levels on tumor growth, 1 x 106 cells of the indicated stable cell lines (control SW620 and RAC1wt- and iRAC1-transfected clones) resuspended in 50 µl of Dulbecco’s modified Eagle’s medium were injected orthotopically into the submucosa of the cecal wall of athymic nude mice. Injection of cells resulted in tumor development, as previously described.23
As shown in Table 1
, local tumor growth was confirmed in 63% of the mice bearing SW620/Rac1wt and 89% of the mice bearing control tumors (SW620). Injection of RAC1wt-transformed cells was highly lethal (Figure 2A)
. Four months after injection, 50% of RAC1wt-injected mice were dead, accelerating the tumorigenic process compared with the parental cells. Although not statistically significant due to the small sample size, as determined by the log rank test (with values of P = 0.44 and 0.42 for wt1Clone and wt2Clone, respectively), there was a clear trend toward a shorter survival in animals bearing tumors derived from Rac1 overexpressor clones. To confirm these results, the orthotopic experiment was repeated, with a similar outcome (Figure 2B)
, and the tendency of Rac1 aggressiveness was confirmed (with values of P = 0.63 and 0.71 for wt1Clone and wt2Clone, respectively). These results indicated that cell lines that overexpress of Rac1wt are more aggressive than the parental cell line SW620.
|
|
In keeping with the above results, shRNA-mediated depletion of Rac1 strongly inhibited the local tumor growth on injected cells. Although orthotopic injection was successfully performed, only 13 to 14% of the mice injected with the iRAC1 cells died at the end of the experiment, 242 days after injection, and only these mice presented local tumor growth (Table 1)
. Finally, the cumulative survival after orthotopic injections of iRAC1-transformed cells and parental cells demonstrated the suppression of tumorigenesis as measured by the take rate, leading to an increased survival time (Figure 2C)
. Survival time was found statistically significant by the log rank test with values of P = 0.003 and 0.007 for i1Clone and i2Clone, respectively. Data were also statistically significant for the take rate when the exact Fisher test was applied, with values of P = 0.003 for both i1 and i2 clones.
Pathology Findings
The macroscopic features of the tumors generated at the injection site by SW620, wt1Clone, wt2Clone, and the two derived from the i1Clone and i2Clone, were similar in all of the animals. These tumors were solid and whitish with some necrotic foci and extended longitudinally through the cecum. When the cecum was transversally open, the tumor occupied and expanded the cecal wall narrowing the cecal lumen, closing it completely in some animals. The foci of tumor at other locations showed the same macroscopic features. The microscopic features of the tumor at the injection site were also the same in all of the animals. It consisted of a poorly differentiated adenocarcinoma made of sheets of cells with highly atypical nuclei that occasionally formed glands. Some tumor cells were multinucleated, and others had an intracytoplasmic lumen that gave them the appearance of a signet ring. The mitotic rate was higher than 5 mitoses per 10 high-power fields, and many necrotic areas were seen (Figure 3A)
. The tumor cells were periodic acid-Schiff negative and were stained focally in the cytoplasm with antibodies against AE1/AE3, low-molecular-weight cytokeratins, and epithelial membrane antigen, whereas they were not stained with antibodies against cytokeratin 20, 7 and high-molecular-weight cytokeratins. Vimentin was also positive in the cytoplasm of many tumor cells, but carcinoembryonic antigen was negative, due possibly to the poorly differentiated characteristics of the adenocarcinoma. In keeping with these results, Hewitt et al26
had previously described a weak staining for carcinoembryonic antigen in SW620, and the American Type Culture Collection indicates that these cells synthesize only small quantities of carcinoembryonic antigen.
The tumor expanded from its initial location at the submucosa through all of the intestinal wall layers ulcerating the mucosa and reaching the serosal surface in all of the animals (Figure 3B)
. The tumor was also found in all of the animals inside the lumen of lymphatic vessels of the mucosa, the submucosa, the muscular layers, and the serosa (Figure 3C)
. Some animals also showed tumor foci in the visceral and parietal peritoneum that infiltrated the adjacent organs, such as the diaphragm (Figure 3D)
. Others also showed metastasis to mesenteric lymph nodes and pancreas (Figure 3, E and F)
. Tumor foci in the liver were always found in continuity with the peritoneal surface of the organ (Figure 3, G and H)
and were never seen inside vascular lumens, so we considered these tumor foci as stage II. Tumor deposits in the lungs or in the brain were not found. The data of the tumor staging of the animals belonging to each group are summarized in Table 2
. The microscopic and immunohistochemical features of the metastases of the tumor were the same than those found at the injection site.
|
. Rac1 staining was positive in all of the cases and was found mainly in the cytoplasm of the tumor cells. Frozen tumor samples from the cecum of three mice were also analyzed by Western blot to confirm Rac1 protein overexpression (Figure 4D)
.
|
. This finding indicates that tumor growth was the result of instability of shRNA in vivo, and therefore the lack of interference after injection allowed the re-expression of the protein.
The macroscopic and microscopic anomalies found at the necropsy considered as a whole served to establish the cause of death of each animal. Some animals showed a very narrow lumen of the cecum caused by the tumor and dilation of the small bowel (Figure 5A)
. Microscopically, the intestinal wall had been completely replaced in some areas by necrotic tumor (Figure 5B)
. There were abundant inflammatory cells at the serosal surface of the small and large bowel (Figure 5C)
. The death of these animals was attributed to peritonitis due to intestinal obstruction caused by the tumor. Other animals showed the large bowel filled with blood while the rest of the organs showed an intense pallor and the spleen was enlarged (Figure 5D)
. On the cut surface, part of the cecal wall had been replaced by the local tumor growth, of which the surface was hemorrhagic (Figure 5E)
. Microscopically, the tumor surface was also hemorrhagic. The death of these animals was attributed to intestinal hemorrhage originated in the tumor. This hemorrhage was chronic in many animals because they showed compensatory hyperplasia of the hematopoietic tissue of the spleen (Figure 5F)
. Other less frequent causes of death were sepsis and obstruction of the urinary flow due to tumor compression of pelvic structures. The main causes of death of the animals belonging to each group are summarized in Table 3
.
|
|
|
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Rac1 is involved in membrane ruffling, morphological and growth transformation, and signaling to a variety of downstream transcription factors, including c-Jun, SRF, and nuclear factor-B.13,33-35
Also, Rac1 is involved in cell adhesion and migration, and its activation is associated with increased tumor invasiveness.36
Rac1 is overexpressed in different types of cancers. Rac1-dependent cell signaling may be relevant for malignant transformation16
and may respond to external tissue-specific clues that signal toward promoting proliferation, adhesion, or motility.37
It is now clear that Rac1 is involved in multiple stages of the tumorigenic process, raising the possibilities that Rac1 and its effectors are useful anticancer targets. Thus, interference with Rac1 function might be explored as an attractive and novel anticancer strategy.15
Several evidences support the importance of Rac1 in colorectal adenocarcinoma. Extensive studies into the molecular etiology of cancer have established a central role for the Wnt signal transduction pathway. Mutations in the canonical Wnt signaling pathway leading to its activation are known to cause the majority of intestinal tumors. Recent studies have suggested that Rac1 could modulate β-catenin function in the context of the canonical Wnt signaling pathway and that the activation of Rac1 may represent a common step of colorectal tumorigenesis.20-22,38
It has been suggested that Rac1 and MAPK p38 are potentially involved in key mechanisms leading to malignant transformation in the colonic crypt.39
Moreover, the 
6β4 integrin may mediate motile events required for invasion of colon carcinoma cells through Rac1 and phosphatidylinositol 3-kinase signaling.40
However, there is little evidence so far from animal studies to define how Rac1 contributes to colon cancer cell proliferation, survival, invasion, and metastasis, and we therefore focused on this system. We genetically modulated the expression levels of the human RAC1 gene, by either up- or down-regulation. Characterization of the effects of Rac1 modulation in colorectal carcinoma cell line-derived tumors suggests a critical role of Rac1 in colorectal adenocarcinoma progression.
Animal models for human colorectal cancer can provide insight into the mechanisms that underlie the development and pathogenesis of this type of tumor. The ideal animal model should therefore faithfully replicate all aspects of tumor development in man. These include the sequential acquisition of genetic alterations with consequent changes in cell behavior and tumor biology.41 Moreover, many of the most-used mouse models form adenomas primarily in the small bowel, with relatively few lesions in the descending and sigmoid colon.42
Here, we demonstrate that Rac1 is critically involved in the pathogenesis of human colorectal cancer cells. The colorectal adenocarcinoma-derived cell line SW620 had acquired the sequential mutation of APC, TP53, and SMAD4 and activation of the proto-oncogene KRAS43 . The ectopic expression of Rac1 is sufficient to induce cell transformation.44 To evaluate the in vivo effects of this GTPase, RAC1wt-overexpressing and RAC1-interfering cancer cells were injected orthotopically into the mouse submucosal layer of the cecal wall, and tumor growth, progression, dissemination, and metastasis of the generated colorectal adenocarcinoma were analyzed.
A major advantage of the orthotopic cell microinjection model is the fidelity of emulating human disease. It allows the possibility to modify the experimental conditions to get new methods for cancer diagnosis and prognosis as well as novel strategies to develop new selective drugs. Another advantage of the orthotopic cell microinjection model is the high incidence of the tumor take rate and that the microenvironment is optimal to influence the tumor behavior. Furthermore, this model of colorectal adenocarcinoma in mouse resembles the behavior of human colorectal adenocarcinoma and the cause of death of patients who have not undergone surgical or chemotherapeutic interventions.23 We observed a reduction of survival time in mice with RAC1wt-induced colorectal tumors when compared with control mice, suggesting the high lethality produced by this protein.
Metastasis is the most common complication in human tumors leading to patient death. The variety of cellular functions attributed to Rho proteins includes cell morphology, formation of stress fibers and focal adhesions, membrane ruffling, cell motility, and invasion.45 Activation or overexpression of Rho proteins has been associated with enhanced invasion and motility.36,40,46,47 We have previously reported that cells overexpressing Rac1 acquire metastatic properties in vivo.29 In keeping with this, we have found tumoral cell spreading in the visceral and parietal peritoneum, metastasis to the pancreas and to mesenteric lymph nodes, and tumor foci in the liver. As previously described,26 SW620 cells metastasize to the liver, but metastases were only present on the peritoneal surface of the liver and not within the parenchyma. Because SW620 cells were derived from a lymph node metastasis, the cells may have acquired the ability for lymphatic metastasis but not the ability for hematogenous metastasis to the liver parenchyma. It may be relevant that SW620 cells are epidermal growth factor receptor-negative,48 because there is evidence that epidermal growth factor receptor expression is necessary for colon cancer metastasis to the liver parenchyma.49
Even though no sustained differences in colonized organs or number of metastasis appeared between the control and modified groups (probably tumors killed the animals before metastasis grew further due to the aggressiveness of Rac1 overexpression), there is increasing evidence of the role of Rac1 or downstream effectors in invasion and metastasis.29,50 In addition, most studies have been performed with activated mutants, which presented a more profound effect in adhesion, spreading, and membrane ruffling of colon cancer cells,51 in agreement with our previous findings that although both wild-type and activated mutants of Rho proteins are tumorigenic, the mutated version is more potent.28
Short hairpin RNAs can be used in vitro to produce sequence-specific gene silencing of mammalian cells,52,53 resulting in the effective in vivo suppression of gene expression in adult mice.54-56 Several studies have shown the efficacy of interfering Rac1 expression in different cancer cell lines.57,58 SW620 expresses relatively high levels of Rac1, but when its expression was inhibited, tumor growth in mice was completely abolished. In keeping with this observation, Malliri et al59 reported that Tiam1-null mice have a reduction in Ras-induced skin tumors, suggesting that Rac1 contributes primarily to tumor development by promoting cell proliferation.
Our data suggest that Rac1 plays a crucial role in local growth of colorectal adenocarcinoma and has a causal role in promoting tumor progression. Furthermore, interference of its expression or activity may provide a positive clinical benefit for patients with colorectal carcinoma. Recently, inhibition of Rac1 activity has been evaluated as a possible method for cancer treatment.60-64 Specific Rac1 inhibition in colorectal adenocarcinoma may represent a powerful tool for the in vivo inhibition of tumor growth and may have potential therapeutic value.
|
Acknowledgements |
|---|
|
Footnotes |
|---|
Supported by grants to J.C.L. from Ministerio de Educación y Ciencia (SAF2004-0577 and SAF2005-06195-C02-01), Comunidad de Madrid (S-BIO/0280/2006), Ministerio de Sanidad (RTICC RD06/0020/0016), and TCD Pharma SL and by Ministerio de EducaciónyCíencia (SAF03-07437) and Ag
ncia de Géstió d’Ajuts de Recerca (SGR05-1050) to R.M. C.E. is a fellow from Ministerio de Educación y Ciencia.
Accepted for publication September 14, 2007.
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
