Cancer Letters 356 (2015) 556–560

Contents lists available at ScienceDirect

Cancer Letters j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / c a n l e t

Original Articles

IQGAP1 in rectal adenocarcinomas: Localization and protein expression before and after radiochemotherapy Susanne Holck a, Hans Jørgen Nielsen b, Emilie Hammer b, Ib Jarle Christensen c, Lars-Inge Larsson a,* a

Department of Pathology, Copenhagen University Hospital, Hvidovre, Denmark Department of Surgical Gastroenterology, Copenhagen University Hospital, Hvidovre, Denmark c Finsen Laboratory, Rigshospitalet and Biotech Research Innovation Center (BRIC), University of Copenhagen, Copenhagen, Denmark b

A R T I C L E

I N F O

Article history: Received 7 August 2014 Received in revised form 1 October 2014 Accepted 3 October 2014 Keywords: IQGAP1 ERK Radiochemotherapy Rectal adenocarcinoma Immunohistochemistry

A B S T R A C T

Treatment of rectal adenocarcinoma includes total mesorectal excision, which is preceded by radiochemotherapy (RCT) in cases of advanced disease. The response to RCT varies from total tumor regression to no effect but this heterogeneous response is unexplained. However, both radiation and treatment with 5-fluorouracil may induce treatment resistance through upregulation of the mitogen-activated protein kinase (MAPK) cascade. IQGAP1 is a scaffold protein that appears to be essential to MAPK signaling in cancers. We have therefore studied IQGAP1 protein expression in rectal adenocarcinomas before and after RCT. We demonstrate that cancer cells show increased apical staining for IQGAP1 following RCT. Interestingly, this increase is significantly higher in patients showing poor RCT responses. Our results also suggest that low levels of apical IQGAP1-staining in biopsies may predict the RCT response. Together, these data suggest that both the level and localization of IQGAP1 may influence the treatment response. © 2014 Elsevier Ireland Ltd. All rights reserved.

Introduction Rectal adenocarcinomas are treated with total mesorectal excision (TME). TME is preceded by radiochemotherapy (RCT) in patients having locally advanced disease [1]. Preoperative RCT decreases the incidence of local recurrences and may improve disease-free and overall survival, particularly if a complete pathological response (i.e. no remaining viable tumor cells) is obtained [1]. However, individual patients respond differently to RCT and there is currently no consensus regarding marker(s) that can predict the response, nor is there any explanation as to why some tumors respond poorly to RCT [2,3]. In vitro studies have shown that either radiotherapy [4–6] or chemotherapy with 5-fluorouracil (5-FU) [7] may upregulate signaling through the mitogen-activated protein kinase (MAPK) cascade. This cascade involves the sequential activation of RAF, MEK (MAP/ERK kinase 1 and 2) and ERK (extracellular signal regulated kinase 1 and

Abbreviations: 5-FU, 5-fluorouracil; APC, adenomatous polyposis coli; AUC, area under curve; ERK, extracellular signal-regulated kinase; GAP, GTPase-activating protein; IQGAP1, IQ[isoleucine-glutamine domain]-guanosine triphosphatase activating protein 1; MAPK, mitogen-activated protein kinase; MEK, MAP/ERK kinase; RCT, Radiochemotherapy; ROC, receiver operating characteristic curve; TME, total mesorectal excision; TRG, tumor regression grade. * Corresponding author. Tel.: +45 38 62 66 48; fax: +45 38 62 33 54. E-mail address: [email protected] (L.-I. Larsson). http://dx.doi.org/10.1016/j.canlet.2014.10.005 0304-3835/© 2014 Elsevier Ireland Ltd. All rights reserved.

2) and is associated with stimulation of cell growth and regulation of apoptosis [8]. Activation of ERK has been shown to interfere with the effects of radiotherapy and 5-FU treatment [5–7]. The IQGAP1 protein appears to be essential to ERK signaling in cancers [9]. The name IQGAP1 was derived from the presence of a calmodulin-binding IQ (isoleucine-glutamine) domain as well as the presence of a domain with homology to GTPase-activating protein (GAP) [10]. However, IQGAP1 does not possess GAP activity, but instead binds to and stabilizes GTP-bound proteins (Rac and cdc42) in their active conformation [11,12]. IQGAP1 interacts with many proteins, including adenomatous polyposis coli (APC) [13], and acts as a scaffold for the RAF-MEK-ERK (MAPK) activation cascade [14–16]. Additionally, it binds actin and participates in the organization of adherens junctions, beta-catenin signaling and cell motility [12]. IQGAP1 is expressed at high levels in placenta, lung and kidney and at lower levels in heart, liver, skeletal muscle and pancreas [10]. Additionally, IQGAP1 is present in endothelial cells, where it has been linked to cell proliferation and angiogenesis [17,18], in the gastrointestinal epithelium [19–21] and in lymphoid cells [22,23]. IQGAP1 expression has been detected in multiple tumors, including gastric and colorectal cancers [19,21,24,25]. Moreover, IQGAP1 expression correlates to increased aggressiveness and invasiveness [19,25–27]. Recent data suggest that the MAPK scaffolding function may be important to growth of neoplastic cells [9]. Thus, interference with IQGAP1’s ability to dock ERK has been shown to

S. Holck et al./Cancer Letters 356 (2015) 556–560

interfere with signaling through the MAPK cascade and with tumor growth [9]. This is of great interest because such interference may target tumors driven by different oncogenic events, including EGF receptor overactivity or activating mutations in RAS or BRAF. Importantly, mice deficient in IQGAP1 are healthy, except from the appearance of gastric hyperplasia at old age [20], suggesting that the functions of IQGAP1 are dispensable for most normal cells. Only two studies on the expression of IQGAP1 in colorectal tumors exist [19,24] and no studies have addressed the expression in rectal adenocarcinomas. We have examined this in biopsies obtained before RCT as well as in matched resections obtained after RCT. Our results show that RCT is associated with a marked increase of staining for IQGAP1 in the apical region of cancer cells and that the patients, which show the least response to RCT show the highest increase in apical IQGAP1. Finally, our results suggest that patients showing a complete response to RCT have a low apical expression of IQGAP1 in pre-treatment biopsies. Materials and methods Patients Fifty-four patients (29 male and 25 female; median age: 63.5 years; range: 19– 86 years) with rectal adenocarcinoma that had been subjected to TME at Hvidovre University Hospital during 2007–2012 were studied. Forty-six patients received preoperative standard treatment (48–60 Gy over 25 days) plus concomitant Capecitabine (Xeloda; metabolized to 5-FU in the body). The remaining eight patients received a preoperative non-standard treatment (radiation as above plus Xeloda, which was interrupted after 14 days (n = 3) or no Xeloda treatment (n = 1), radiation as above plus Xeloda and Oxaliplatin (n = 2) or received chemotherapy, but no radiation (n = 2). In 30 cases, matched pre-treatment biopsies and posttreatment surgical resections were available. Tumor regression grade (TRG) was assessed as described [28,29]. According to this classification TRG1 means complete regression and TRG5 means no regression. TRG was evaluated by an experienced pathologist and included 6 cases of TRG1 (11.1%), 18 cases of TRG2 (33.3%), 19 cases of TRG3 (35.2%) and 11 cases of TRG4 (20.4%). Pathology staging revealed 6 cases with pT0, 5 cases with pT1, 13 cases with pT2, 22 cases with pT3, 7 cases with pT4 and 1 case with pTX. Nodal metastases were detected in 11 patients. Histological grading and recording of tumor infiltrating lymphocytes (TIL) [30], was performed on the biopsy specimens, disclosing 1 highly, 47 moderately, and 6 poorly (including 2 mucinous) differentiated carcinomas. Extramural vascular and perineural invasion was detected in four and seven resections, respectively. Data pertaining to the 30 matched patients are given in Table 1. The study was approved by the Danish Data Protection Agency (2008-412252) and Ethical Committee (H-KF-26288/KF-01–164/03).

Table 1 Characteristics of patients with matched biopsies and resection specimens in relation to regression grade (TRG).

Age (median; range) Sex difAa (median; range) pT1c (n, %) pT2c (n, %) pT3c (n, %) pT4c (n, %) pN0d (n, %) pN1d (n, %) pN2d (n, %) Differentiation: high (n, %) Differentiation: moderate (n, %) Differentiation: poore (n, %) Vascular invasion (n, %) Perineural invasion (n, %) TIL (n, %) a

TRG 2–3

TRG 4

61; 30–81 years 14 f, 9 m 0.87 (−2.00 to 4.00) 1 (4.3%) 8 (34.8%) 11 (47.8%) 3 (13.0%) 17 (73.9%) 5 (21.7%) 1 (4.3%) 1 (4.3%) 20 (87.0%) 2 (8.7%) 1 (4.3%) 5 (21.7%) 1 (4.3%)

61; 40–86 years 2 f, 5 m 2.75 (1.50–6.00)b 1 (14.3%) 1 (14.3%) 4 (57.1%) 1 (14.3%) 5 (71.4%) 1 (14.3%) 1 (14.3%) 0 (0%) 5 (71.4%) 2 (28.6%) 3 (42.9%) 2 (28.6%) 0 (0%)

Denotes increase in apical IQGAP1 staining between biopsies and resections. The difference between TRG 2–3 and TRG 4 is significant at p = 0.01. If the two patients in the TRG 4 group, who did not receive radiation therapy (but received chemotherapy) are excluded, the difference remains significant at p = 0.03. c pT denotes pathological tumor stage. d pN denotes pathological nodal stage. e Denotes poor/mucinous differentiation. b

557

Immunohistochemistry Biopsies and surgical resections were fixed in formalin and embedded in paraffin by standardized procedures. Deparaffinized sections were stained with HE and with immunocytochemical procedures for detecting IQGAP1 and cytokeratin (AE1/ A3; Dako, Glostrup, Denmark). Initially we tested different demasking procedures (heat retrieval at high and low pH, pepsin treatment and no treatment) and three different mouse monoclonal IQGAP1 antibodies (#AM20036-AF-N; Acris, Herford, Germany, #05–504; Upstate/EMD Millipore, Temecula, CA; #610611; Becton Dickinson, Franklin Lakes, NJ) applied to a training set of six rectal adenocarcinomas. The procedure finally selected included demasking at high pH (PTLink, Dako), staining with a monoclonal anti-IQGAP1 antibody (Acris) at 1 μg/ml and detection with a two-layer Envision (Dako) peroxidase detection system using a Dako autostainer. Sections were counterstained with hematoxylin. Controls included a type-matched control mouse monoclonal antibody derived from an IgG2a-producing myeloma (mouse IgG2a Isotype Control, Clone UPC-10, M5409, Sigma, St. Louis, MO).

Scoring Two independent observers evaluated the specimens blindly. Staining intensities were graded as weak or absent (0), moderate (1) or strong (2) and number of stained tumor cells was graded as 0% (0), below 10% (1), 10–60% (2) and above 60% (3). These cut-offs were established by consensus of both investigators following an initial survey of the entire blind-coded material. For all tumors this grading was applied to three different patterns of IQGAP1 staining in tumor cells: staining of the lateral cell membrane; staining of the apical cell region and staining of the cytoplasm. Final results were computed as the product of staining intensities and cell numbers. Staining of inflammatory cells was present in all specimens and served as an internal positive control. Additionally, tumors stained for cytokeratin were evaluated for the presence of tubular structures graded as 0% (0), below 10% (1), 10– 60% (2) and above 60% (3). In cases where scorings differed by more than one unit, the observers re-evaluated the specimens to reach a consensus. In other cases, means of the scorings were calculated.

Statistics Matched differences were tested using the Wilcoxon signed rank test. Differences in the IQGAP1 staining between tumor regression, dichotomized with 3 as the threshold, was done using a general linear model (GLM). The probability of no tumor versus residual tumor was estimated using logistic regression analysis and illustrated by the receiver operating characteristic curve (ROC). The significance level was set at 5%. All calculations were done using SAS (version 9.3, SAS Institute, Cary, NC).

Results Initially we used three different monoclonal mouse antibodies to IQGAP1 and different pre-treatment techniques. With all three antibodies, optimal results were obtained using heat retrieval at high pH. All antibodies stained the same structures including endothelial cells, inflammatory cells (including lymphocytes) and mucosal epithelial cells, which all are known to contain IQGAP1. In addition, tumor cells stained for IQGAP1, but the frequency, intensity and staining pattern varied between patients. One IQGAP1 monoclonal antibody was selected for staining of all rectal cancers using automated, standardized methods. Control, type-matched mouse monoclonal antibodies produced no staining (Fig. 1). In the single case of TIL in this series, the tumor infiltrating lymphocytes also stained strongly for IQGAP1 (Fig. 2b). Three different staining patterns were observed in rectal cancer cells and included staining of the lateral cell membrane, the apical cell region and of the cytoplasm. Individual tumors showed either only one of these patterns or showed different combinations (Fig. 2). There was no staining of nuclei. Typically, there was heterogeneity in staining of the tumors and this was more marked in resections than in biopsies (Fig. 2). In individual tumors, strongly IQGAP1positive cells sometimes occurred intermingled with unstained tumor cells. The degree of apical IQGAP1-staining was significantly stronger in resections (3.0-fold with respect to medians) than in matched biopsies (signed rank test: p = 0.0001; n = 30), whereas membrane

558

S. Holck et al./Cancer Letters 356 (2015) 556–560

Fig. 1. Rectal adenocarcinoma stained for IQGAP1 (a, c) or with a type-matched control monoclonal antibody (b, d). In (a) are groups of cancer cells (exemplified by arrows), which show strong IQGAP1-staining of the lateral cell membrane and the cytoplasm but no staining with the control antibody (b). Additionally, inflammatory cells, part of which represent lymphocytes, show strong cytoplasmic IQGAP1-staining (exemplified by arrowheads) and no staining in the control (b). In (c) is shown stromal tissue in which endothelial cells (exemplified by arrows) in an arteriole and a venule (arrowhead) show staining for IQGAP1. Additionally, inflammatory cells/lymphocytes show strong IQGAP1-staining. No staining occurs in the control (d).

staining was 0.67-fold lower (p = 0.02) and cytoplasmic staining was unchanged (p = 0.9). Patients showing less than 50% regression (TRG4) showed a significantly larger increase in apical IQGAP1-staining following RCT (GLM: p = 0.01, mean 2.08, 95% CI: 0.53–3.62; n = 7) than patients showing more than 50% regression (TRG2-3; n = 23) (Figs. 2 and 3). Following exclusion of the two patients who did not receive radiotherapy this difference remained significant (Table 1). There was no significant correlation between the cytoplasmic, membrane and apical staining patterns in either biopsies or resections and there was no significant correlation between the staining patterns in matched biopsies and resections. Additionally, there was no significant correlation to age, sex or between prominence of tubules and apical staining. Interestingly, patients showing no residual tumor (TRG1) after RCT had lower levels of apical IQGAP1-staining in pretreatment biopsies than patients with TRG2-4. This difference reached border-line significance (logistic regression, Wald p = 0.06). Although the patient number was small, a ROC plot revealed a potential utility of apical IQGAP1-staining in biopsies as a predictor of response to RCT with an AUC of 0.78 (Fig. 4).

Discussion IQGAP1 serves at the intersection between actin cytoskeletal organization, calcium/calmodulin, MAPK signaling and the betacatenin/APC pathway [10–12,14–16]. Accordingly, its relation to tumor progression has been much studied. IQGAP1 is highly expressed in virtually all types of tumors studied and has been linked to tumor invasiveness and aggressiveness [19,25–27]. Our study is the first to provide a detailed analysis of IQGAP1 expression in rectal carcinomas and to relate such expression to RCT. Three different monoclonal antibodies produced identical results and revealed IQGAP1 immunoreactivity in both cancer cells and in adjacent cells. The latter included epithelial cells, endothelial cells and inflammatory cells – structures, which already are known to contain IQGAP1 [17–23]. No staining was obtained with a typematched control monoclonal antibody. IQGAP1-staining varied between tumors and three distinct staining patterns associated with, respectively, the lateral cell membrane, the apical region and the cytoplasm were observed. Individual tumors could present with

S. Holck et al./Cancer Letters 356 (2015) 556–560

559

Fig. 2. Different expression patterns of IQGAP1 in different rectal adenocarcinomas. In (a) is a biopsy taken before RCT and in (b) is the corresponding resection specimen after RCT. Note weak or no apical staining (exemplified by arrows) for IQGAP1 in the biopsy and strong apical staining (exemplified by arrows) in the resection (TRG4). In the biopsy, most staining occurs in the lateral cell membrane and the cytoplasm, while the resection specimen shows cytoplasmic staining in addition to the apical staining. Also note staining of lymphocytes, which surround and sometimes infiltrate the cancer tissue in both specimens. In (c) is shown resection tissue following RCT. Note cancer cells present in a mucinous pool (left), which show weak staining for IQGAP1, along with another group of cancer cells (right) showing strong IQGAP1-staining of the lateral cell membrane and cytoplasm. In (d) is resection specimen with cancer cells growing in cords with little or no evidence of lumina. Note the marked heterogeneity of staining with groups of cells showing strong cytoplasmic staining and others showing little or no staining for IQGAP1.

either one or a combination of these patterns. Interestingly, apical IQGAP1-staining was significantly increased in resections as compared with biopsies taken before RCT. Since lateral membrane staining was significantly reduced and cytoplasmic staining was unchanged following RCT, it is unlikely that the changes reflect differences in fixation and processing of biopsies and resections. This is also underlined by the staining of inflammatory cells, which served as an internal control in all specimens. Another explanation could be that RCT affects the number of tubular or acinar structures in the tumors. However, we found no correlation between tubules/ acini (as scored on cytokeratin-stained sections) and apical IQGAP1staining. We posit that RCT induces changes in the subcellular localization and abundance of IQGAP1, possibly by redistributing it from the lateral cell membrane to the apical cell membrane. Importantly, we found that patients, who showed least tumor regression following RCT, showed the significantly highest increase in apical IQGAP1-staining. Moreover, we also observed that TRG1 patients had lower apical IQGAP1-staining in biopsies taken

before treatment. However only six TRG1 patients were available and these results only reached borderline significance. Together, these data suggest that apical expression of IQGAP1 may be important to the response to RCT. The biological background to this is unclear. However, both radiation and 5-FU have been reported to increase ERK activation and such activation has been found to be radio- and chemopreventive [4–7]. Since IQGAP1 appears to be an essential scaffold for ERK activation in cancer cells [9] it may participate in RCT resistance. Acknowledgements We thank Anja Alex Petersen for expert technical assistance. Conflict of interest None.

560

S. Holck et al./Cancer Letters 356 (2015) 556–560

Fig. 3. Box-and-whiskers plot illustrating that patients achieving more than 50% tumor regression (2–3) following RCT show significantly less increase in apical IQGAP1staining than patients achieving less than 50% regression (4). Boxes indicate interquartile ranges, whiskers indicate ranges of maximal and minimal values and diamonds indicate means.

Fig. 4. ROC curve illustrating the potential utility of apical staining for IQGAP1 in biopsies as a predictor for response to RCT.

References [1] M. Maas, P.J. Nelemans, V. Valentini, P. Das, C. Rödel, L.J. Kuo, et al., Long-term outcome in patients with a pathological complete response after chemoradiation for rectal cancer: a pooled analysis of individual patient data, Lancet Oncol. 11 (2010) 835–844. [2] P. Kapur, Tailoring treatment of rectal adenocarcinoma: immunohistochemistry for predictive biomarkers, Anticancer Drugs 22 (2011) 362–370. [3] J.G. Kuremsky, J.E. Tepper, H.L. McLeod, Biomarkers for response to neoadjuvant chemoradiation for rectal cancer, Int. J. Radiat. Oncol. Biol. Phys. 74 (2009) 673–688.

[4] J.K. Leach, G. Van Tuyle, P.S. Lin, R. Schmidt-Ullrich, R.B. Mikkelsen, Ionizing radiation-induced, mitochondria-dependent generation of reactive oxygen/ nitrogen, Cancer Res. 61 (2001) 3894–3901. [5] A. Yacoub, R. McKinstry, D. Hinman, T. Chung, P. Dent, M.P. Hagan, Epidermal growth factor and ionizing radiation up-regulate the DNA repair genes XRCC1 and ERCC1 in DU145 and LNCaP prostate carcinoma through MAPK signaling, Radiat. Res. 159 (2003) 439–452. [6] A. Yacoub, J.S. Park, L. Qiao, P. Dent, M.P. Hagan, MAPK dependence of DNA damage repair: ionizing radiation and the induction of expression of the DNA repair genes XRCC1 and ERCC1 in DU145 human prostate carcinoma cells in a MEK1/2 dependent fashion, Int. J. Radiat. Biol. 77 (2001) 1067–1078. [7] Y. Zhao, S. Shen, J. Guo, H. Chen, D.Y. Greenblatt, J. Kleeff, et al., Mitogen-activated protein kinases and chemoresistance in pancreatic cancer cells, J. Surg. Res. 136 (2006) 325–335. [8] T. Wada, J.M. Penninger, Mitogen-activated protein kinases in apoptosis regulation, Oncogene 23 (2004) 2838–2849. [9] K.L. Jameson, P.K. Mazur, A.M. Zehnder, J. Zhang, B. Zarnegar, J. Sage, et al., IQGAP1 scaffold-kinase interaction blockade selectively targets RAS-MAP kinase-driven tumors, Nat. Med. 19 (2013) 626–630. [10] L. Weissbach, J. Settleman, M.F. Kalady, A.J. Snijders, A.E. Murthy, Y.X. Yan, et al., Identification of a human rasGAP-related protein containing calmodulin-binding motifs, J. Biol. Chem. 269 (1994) 20517–20521. [11] M.D. Brown, D.B. Sacks, IQGAP1 in cellular signaling: bridging the GAP, Trends Cell Biol. 16 (2006) 242–249. [12] M. Fukata, K. Kaibuchi, Rho-family GTPases in cadherin-mediated cell-cell adhesion, Nat. Rev. Mol. Cell Biol. 2 (2001) 887–897. [13] T. Watanabe, S. Wang, J. Noritake, K. Sato, M. Fukata, M. Takefuji, et al., Interaction with IQGAP1 links APC to Rac1, Cdc42, and actin filaments during cell polarization and migration, Dev. Cell 7 (2004) 871–883. [14] J.G. Ren, Z. Li, D.B. Sacks, IQGAP1 modulates activation of B-Raf, Proc. Natl. Acad. Sci. USA 104 (2007) 10465–10469. [15] M. Roy, Z. Li, D.B. Sacks, IQGAP1 binds ERK2 and modulates its activity, J. Biol. Chem. 279 (2004) 17329–17337. [16] M. Roy, Z. Li, D.B. Sacks, IQGAP1 is a scaffold for mitogen-activated protein kinase signaling, Mol. Cell. Biol. 25 (2005) 7940–7952. [17] R.D. Meyer, D.B. Sacks, N. Rahimi, IQGAP1-dependent signaling pathway regulates endothelial cell proliferation and angiogenesis, PLoS ONE 3 (2008) e3848. [18] M. Yamaoka-Tojo, M. Ushio-Fukai, L. Hilenski, S.I. Dikalov, Y.E. Chen, T. Tojo, et al., IQGAP1, a novel vascular endothelial growth factor receptor binding protein, is involved in reactive oxygen species – dependent endothelial migration and proliferation, Circ. Res. 95 (2004) 276–283. [19] H. Hayashi, K. Nabeshima, M. Aoki, M. Hamasaki, S. Enatsu, Y. Yamauchi, et al., Overexpression of IQGAP1 in advanced colorectal cancer correlates with poor prognosis-critical role in tumor invasion, Int. J. Cancer 126 (2010) 2563–2574. [20] S. Li, Q. Wang, A. Chakladar, R.T. Bronson, A. Bernards, Gastric hyperplasia in mice lacking the putative Cdc42 effector IQGAP1, Mol. Cell. Biol. 20 (2000) 697–701. [21] A. Walch, S. Seidl, C. Hermannstädter, S. Rauser, J. Deplazes, R. Langer, et al., Combined analysis of Rac1, IQGAP1, Tiam1 and E-cadherin expression in gastric cancer, Mod. Pathol. 21 (2008) 544–552. [22] J.A. Gorman, A. Babich, C.J. Dick, R.A. Schoon, A. Koenig, T.S. Gomez, et al., The cytoskeletal adaptor protein IQGAP1 regulates TCR-mediated signaling and filamentous actin dynamics, J. Immunol. 188 (2012) 6135–6144. [23] X. Meng, O. Krokhin, K. Cheng, W. Ens, J.A. Wilkins, Characterization of IQGAP1-containing complexes in NK-like cells: evidence for Rac 2 and RACK1 association during homotypic adhesion, J. Proteome Res. 6 (2007) 744–750. [24] K. Nabeshima, Y. Shimao, T. Inoue, M. Koono, Immunohistochemical analysis of IQGAP1 expression in human colorectal carcinomas: its overexpression in carcinomas and association with invasion fronts, Cancer Lett. 176 (2002) 101–109. [25] H. Takemoto, Y. Doki, H. Shiozaki, H. Imamura, T. Utsunomiya, H. Miyata, et al., Localization of IQGAP1 is inversely correlated with intercellular adhesion mediated by e-cadherin in gastric cancers, Int. J. Cancer 91 (2001) 783–788. [26] P. Dong, K. Nabeshima, N. Nishimura, T. Kawakami, T. Hachisuga, T. Kawarabayashi, et al., Overexpression and diffuse expression pattern of IQGAP1 at invasion fronts are independent prognostic parameters in ovarian carcinomas, Cancer Lett. 243 (2006) 120–127. [27] Z. Liu, D. Liu, E. Bojdani, A.K. El-Naggar, V. Vasko, M. Xing, IQGAP1 plays an important role in the invasiveness of thyroid cancer, Clin. Cancer Res. 16 (2010) 6009–6018. [28] H. Bouzourene, F.T. Bosman, W. Seelentag, M. Matter, P. Coucke, Importance of tumor regression assessment in predicting the outcome in patients with locally advanced rectal carcinoma who are treated with preoperative radiotherapy, Cancer 94 (2002) 1121–1130. [29] A.M. Mandard, F. Dalibard, J.C. Mandard, J. Marnay, M. Henry-Amar, J.F. Petiot, et al., Pathologic assessment of tumor regression after preoperative chemoradiotherapy of esophageal carcinoma. Clinicopathologic correlations, Cancer 73 (1994) 2680–2686. [30] J.M. Michael-Robinson, A.-E. Biemer-Huttmann, D.M. Purdie, M.D. Walsh, L.A. Simms, K.G. Biden, et al., Tumour infiltrating lymphocytes and apoptosis are independent features in colorectal cancer stratified according to microsatellite instability status, Gut 48 (2001) 360–366.