Contents lists available at SciVerse ScienceDirect
Acta Histochemica
jou rn al homepage: www.elsev ier .de /ac th is
wo initiation sites of early detection of colon cancer revealed by localization ofERK1/2 in the nuclei or in aggregates at the perinuclear region of the tumor cells
Department of Molecular Cell Biology, The Weizmann Institute of Science, Rehovot 76100, IsraelDepartment of Immunology, The Weizmann Institute of Science, Rehovot 76100, IsraelDepartment of Biological Services, The Weizmann Institute of Science, Rehovot 76100, IsraelWolfson Hospital, Holon 58100, IsraelKaplan Medical Center, Rehovot 76100, IsraelThe Hebrew University, Jerusalem, IsraelDepartment of Biological Regulation, The Weizmann Institute of Science, Rehovot 76100, Israel
r t i c l e i n f o
rticle history:eceived 16 October 2012eceived in revised form 9 December 2012ccepted 12 December 2012
eywords:olon cancerAPK
ERK1/2ancer stem cells53RASpiregulin
a b s t r a c t
We have used human specimens and antibodies to pERK1/2 to detect early development of colon cancerusing indirect immunocytochemistry. Two distinct sites were stained; one at the tip of the colon cryptsand the other in the stromal tissue associated with the colonic tissue. These foci represent early stages ofcolon cancer initiation sites as established by enhanced Kirsten Rat Sarcoma Virus (KRAS) and the lack ofp53 staining. The enhanced KRAS coincides with the initiation of tumor growth revealed by pERK1/2, bothin the tip of the colon crypts, as well as in the stromal initiation site of the colon tumors. Foci of pERK1/2staining were also detected in 50% of stromal tissue and tips of colon crypts, which were classified asnormal tissues, adjacent to the malignant tissue according to general morphology. However, in colonspecimens, where no malignancy was observed, no accumulation of pERK1/2 was observed. The stainingof pERK1/2 at the stromal foci of the apparently non-malignant tissue appeared as aggregates in theperinuclear region, while in the colon epithelium it appeared in the cell nuclei. In low-grade colon cancerthat was still free of induced mutated p53, staining of pERK1/2 was prominent in the cell nuclei, both inthe stroma tissue and the tip of the colon crypts. In the intermediate stage, that exhibited significant p53staining, only a fraction of p53-free tumor cells was labeled with pERK1/2 antibody, while in high-gradetumors, all cells of tumors were labeled with antibodies to p53, but not with antibodies to pERK1/2. Wesuggest that the down regulation in pERK1/2 labeling is due to the mitogenic capacity of the tumor cells,which are shifted from being driven by nuclear pERK1/2 to mutated p53 expression. We also found thatthe cytoplasm of low grade tumors was positive for epiregulin, while this labeling decreased in high-grade
tumors. We found that the tumors arising from the stroma demonstrated poor structural differenti-ation, while the tumors initiating from the epithelial cells of the colon demonstrated high structuraldifferentiation. We conclude that pERK1/2 is a sensitive marker of early colon cancer, which disappearsat later stages of cancer development. Moreover, pERK1/2 staining can distinguish between tumor cellsoriginating from the tip of the colon crypts and those developing in the stroma, which is present in the
ithelia
close vicinity to colon ep
ntroduction
Colorectal cancer is the second leading cause of cancer deathsn Western countries. Despite more than two decades of research
nto the molecular genetics of colon cancer, there is still a lackf prognostic and productive molecular biomarkers with proventility in this setting (Baba et al., 2000; Matsumoto et al., 2011;
Manzano and Perez-Segura, 2012). It has been suggested that bal-ance between MAPK pathways could be critical for presenting orpromoting growth in a variety of cancer cell lines, including coloncancer cell lines (Baba et al., 2000). In additional studies, the essen-tial role of mutation of Kirsten Rat Sarcoma Virus (KRAS) and p53 incolorectal cancer was established (Ueda et al., 2003; Klump et al.,2004; Khambata-Ford et al., 2007). It was discovered that loss of
epigenetic control of synucleins serves as a molecular indicator ofmetastasis in a wide range of human cancers, including colon can-cer (Liu et al., 2005, 2010). It was also speculated that UCP2 alteredsuppression, oxidative stress and NF-�B activation may be related
Fig. 1. Percentage of immunolabeled cells with pERK1/2 in the development of coloncancer. (A) Presumptive normal. Both perinuclear and nuclear predominated. (B)
the appropriate IgG fractions at 10 �g/ml. Positive controls for p53,
70 A. Amsterdam et al. / Acta H
o successive events in cancer development, including colon cancerDerdák et al., 2006). Moreover expression of epiregulin and KRAS
utation was found valuable in predicting control of metastaticolorectal cancer, using specific drugs as Cetaxumab (Ueda et al.,003; Li et al., 2010). High expression of ERBB1, ERBB2 and ERBB3,ut not ERBB4, was found in colon cancer cell lines and could help
n the selection of the appropriate chemotherapy (Liu et al., 2005;u et al., 2009). High serum and tissue levels of amphiregulin and
igh tissue levels of epiregulin have been proposed as predictors ofoor prognosis in patients with colorectal carcinoma (Nishimurat al., 2008; Jacobs et al., 2009; Wu et al., 2009; Li et al., 2010).
It is well established that the interaction of EGF ligands ofhe EGF family, epiregulin and ampiregulin, with the appropri-te receptors (ERBB) will lead to phosphorylation of the receptorolecule (Bae and Schlessinger, 2010), which will eventually phos-
horylate MAK and ERK1/2, leading to their translocation into theuclei of the cancer cells (Jaaro et al., 1997; Chuderland et al.,008; Yao and Seger, 2009; Keshet and Seger, 2010; Plotnikov et al.,011).
The intestinal renewal system is tightly controlled and dependsn the spatial organization of signals that emanate from support-ve mesenchymal cells, as well as from differentiated epithelialrogeny. Recent evidence suggests that intestinal cancers maytill contain a hierarchical organization, with cancer stem cellsCSCs) at the apex (Vermeulen et al., 2008; Amsterdam et al.,012a). From the seminal studies of Fearon and Vogelstein (1990)nd Amsterdam et al. (2012b) it is clear that CSC develops as
stepwise accumulation of genetic hits in specific genes andathways. The CSC theory refines this model further and sug-ests that the actual tumorigenic capacity of individual cancerells may be influenced by homeostatic signals derived from theiricroenvironment. These findings are especially exciting in light
f recent developments that have increased our comprehensionf the regulatory mechanisms that control individual stem cellsISCs), and have resulted in new tools to identify and localizeSCs (Amsterdam et al., 2012b,c). Although we clearly do not fullyrasp the complete spectrum of signals and interactions at thisoint, our understanding of normal crypt homeostasis and the
dentification of markers that define ISCs are providing intrigu-ng insights into the organization of intestinal cancers (Fearon andogelstein, 1990; Medema and Vermeulen, 2011; Amsterdam et al.,012c).
In recent studies it was shown how RAS activation causes aber-ant nuclear localization of phosphorylated MAPK and ERK to driveeoplastic transformation in colorectal tumors and human colonancer cells. However, in these reports there was no follow-up ofhe appearance of nuclear pERK1/2 in the different stages of thisisease and there were no studies on the correlation between theppearance of pERK in relationship with the appearance of mutated53 (Duhamel et al., 2012; Zhao et al., 2011).
In the present study we demonstrate that pERK1/2 could serves an early marker for the development of colon cancer and may beresent preferentially in colon cancer stem cells. Moreover, theiristribution in the colonic tissue and associated stromal tissueould explain multiple sites of cancer development that could beecognized by the presence of pERK1/2, which would disappeart later stages of colon cancer development, coinciding with theppearance of mutated p53.
aterials and methods
uman tissue samples
We analyzed normal tissues of colon: 4 �m sections of formalinxed and paraffin embedded tissues that were removed adjacento the tumor (11 samples) and were classified as normal tissue
according to general morphology of sections stained with hema-toxylin and eosin (H&E). In addition, we analyzed 3 colon specimenswith no history of cancer. We also analyzed tissues of low gradetumors (stage I) (12 samples), intermediate (stage II) (11 sam-ples) and high grade tumors (stages III and IV) (13 samples). 4 �mthick sections were stained by the indirect immunocytochemistrymethod (Ginath et al., 2001; Singer et al., 2005; O’Neill et al., 2005;Zeren et al., 2008; Amsterdam et al., 2011, 2012a, 2012b, 2012c)with anti-pERK1/2 or anti-p53 or anti-epiregulin or anti-KRAS anti-bodies followed by staining with hematoxylin for 90 s. It should benoted that the number of samples referred to above indicates thenumber of patients from whom the tissue samples were collected.All biopsies were obtained with permission of the Helsinki Com-mittees of the Wolfson Hospital in Holon and Kaplan Hospital inRehovot, Israel, receiving informed consent.
Reagents
Primary antibodies used
(1) Monoclonal antibodies to p53 (mouse clone 421) that werekindly donated by Prof Moshe Oren from the Weizmann Insti-tute of Science.
(2) Antibodies to pERK1/2 were mouse monoclonal antibodies thatrecognize only the phosphorylated form of ERK 1/2 and notthe non-phosphorylated form, using a dilution 1:200 (M8159;Sigma–Aldrich, St. Louis, MO, USA).
(3) Antibodies to epiregulin were anti-human recombinant epireg-ulin and were affinity purified (1195-EP, R & D Systems, MN,USA).
(4) Rabbit polyclonal antibodies to KRAS, purified (Acris Anti-bodies, San Diego, CA, USA). Following incubation with thefirst antibodies there was incubation with specific secondaryantibodies against the first ones conjugated to peroxidase (N-Histofine, Tokyo, Japan) and the staining of the slides wasperformed as described elsewhere (Ginath et al., 2001; Singeret al., 2005; O’Neill et al., 2005; Zeren et al., 2008; Amsterdamet al., 2011, 2011b, 2012a).
Specificity of the staining
Negative controls were performed by staining the sections with
PERK and KRAS were as described by Amsterdam et al. (2011,2011b, 2012b). Positive controls to KRAS antibodies were also pro-vided by Acris Company demonstrating human prostate carcinomastained specifically with anti-KRAS antibodies.
A. Amsterdam et al. / Acta Histochemica 115 (2013) 569– 576 571
Fig. 2. Normal tissue and low grade colon cancer. (A) Normal tissue stained with antibodies to pERK1/2. No staining is visible either in the stroma (St) in the colon crypts(Cr) or the mucosa. (B) Four micrometer thick sections stained with antibodies to pERK1/2. Staining is evident at the edges of the crypts (Cr and arrows) facing the lumen(L). Essentially similar staining was evident both in tissue specified as normal adjacent to the tumor (close to 50% of the cases) and in low grade tumors (St; stroma). (C)Section stained with antibodies to pERK1/2. Several small aggregates of pERK1/2 are evident at initial development of foci of tumors (arrows). Origin of tissue considered tobe normal (close to 50% of the cases). Note the large irregular in shape of nuclei in the small tumors characteristic of cancer cells (St; stroma). (D) Low-grade tumor stainedwith antibodies to pERK1/2. Most of pERK1/2 aggregates are localized in the perinuclear region (arrows). Only in isolated cells, localization was found within the nuclei(double arrows). Cells of the crypt (Cr) are free of labeling (St; stroma). (E) Extended development of tumors in the stroma. Labeling of pERK1/2 is confined to aggregatesm Labelia m; B =
M
awTtdtpltg
ostly in the perinuclear region. Stroma cells (St) are free of labeling (arrows). (F)
lso in tumors (T) initiate from stroma (St; stroma and arrows). Scale bars: A = 20 �
icroscopic examination
For scoring the number of labeled cells and nuclei in normalnd different stages of the colon tumor development, imagesere taken with a Nikon upright microscope (Eclipse E800, Nikon,
okyo, Japan) in bright-field mode using 10× and 40× objec-ives and a Nikon digital camera (model DS-Ri1). Six to twelveifferent areas of stained slides of each different stage of colonumor development (and controls of normal colon tissue) were
hotographed (40× objective) and the number of labeled nuclei or
abeled perinuclear foci and total nuclei were scored, calculatinghe percentage of labeled nuclei or perinuclear foci in each micro-raph, and the mean value ± SD of all fields were also calculated.
ng with antibodies to epiregulin. Label is evident in both in the cell crypts (Cr) and 200 �m; C–F = 50 �m.
At each stage of the colon tumor development (and normal tissue)a total of about 750 cells were scored. It should be noted thatwith the normal colon samples pictures were taken to show thestroma and the epithelial cells, whereas in the colon containingthe tumors images were taken in the main bulk of the tumors.
Statistics
Statistical evaluation of the different percentages of labeled
nuclei among the different stages of ovarian tumors developmentwas performed, as well as the controls of normal colon. Analy-sis of variance (ANOVA) followed by multiple comparison tests orstudent t-test as appropriate, was performed. Calculations were
572 A. Amsterdam et al. / Acta Histochemica 115 (2013) 569– 576
F r staint ady cL ors. S
pCs
R
fplgtbcapcipw(clsrseciwNabtt
tac(dssantd
ig. 3. Staining of apparently normal colon tissue with antibodies to KRAS. (A) Cleaip of the colon crypts (arrows) facing the lumen (L), suggesting that this tissue alrearge nuclei scattered in the stroma are indicative of the development of small tum
erformed using SPSS software (Student’s t-test, Version II, IBM,hicago, IL, USA). Values of P < 0.05 were considered statisticallyignificant (see also Amsterdam et al., 2011).
esults
In order to examine whether pERK1/2 can serve as early markeror colon cancer progression we stained 4 �m thick sections ofaraffin embedded tissue of: normal tissue with normal morpho-
ogical appearance adjacent to the tumor, low grade, intermediaterade, and high grade specimens of colon cancer with antibodieso pERK1/2. Approximately 50% of the normal tissues (determinedy their morphology) showed no staining; neither in the intestinalrypts nor in the stromal tissue cells, as well as in the mucosa layert the bottom of the crypts (Fig. 2A). The remaining 50% of the sam-les showed immunostaining in the nuclei at the tip of the intestinalrypts (Fig. 2B). In the stroma of these apparently normal tissues,ncreasing intensities of pERK1/2 appeared in aggregates in theerinuclear region (Figs. 1 and 2C–E). The appearance of pERK1/2as occasionally detected in the tips of the intestinal crypts
Fig. 2B). Occasionally the pERK1/2 appeared earlier in the stromaells leaving the crypt cells unstained (Fig. 2D) and sometimes theabeling appeared earlier in the colon epithelial cells. Overall, thetaining in the labeled areas amounted to 10.8 ± 4.48%, which rep-esented both the nuclear and the perinuclear areas (Fig. 1). Whenections of the apparently normal tissue were stained with anti-piregulin antibody, there was clear staining in the cytoplasm of theells of crypts as well as in the cytoplasm of tumor cells embeddedn the stroma cells, mainly in tissues that showed positive staining
ith pERK1/2, leaving the non-cancerous cells unstained (Fig. 2F).o positive labeling was evident following staining of sections withntibodies to p53 (not shown). However, when stained with anti-odies to KRAS, only sites of the initial development of tumors athe tips of the colon crypts were stained (Fig. 3A) while the sites ofumor growth in the stroma remained unstained (Fig. 3B).
In low-grade tumors, the labeling with pERK1/2 was similaro that observed in the labeled part of apparently normal tissues,lthough the labeling in the perinuclear regions increased signifi-antly to 24.75 ± 15.54%, and in the nuclei to 68.84 ± 25.32% p < 001Figs. 1 and 4A–C). As in the normal ones, these tissues had noetectable p53, but the labeling of epiregulin was increased (nothown). In the intermediate stage, the staining with pERK1/2 wascattered. The labeling in the nuclei was reduced to 24.68 ± 13.86%
nd in the perinuclear region to 16.48 ± 15.81%, which is very sig-ificantly lower than in low grade tumors (P < 0.001). The labeling inhe perinuclear region was observed predominantly in the tumorseveloping from colonic stroma. However, there were tumors in
ing with anti-KRAS antibodies is evident in the cytoplasm of epithelial cells at thean be classified as containing low grade tumors. (B) No staining in the stroma (St).cale bars: A = 200 �m; B = 50 �m.
the stroma where the vast majority of pERK1/2 was in the nuclei. Intumors, which apparently originate from the colon epithelial cellsand exhibit the typical organization of carcinoma, cells labeled withantibodies to pERK1/2 were scattered in only some of the cell nuclei.The reduction in pERK1/2 staining seemed to be accompanied byan increase in p53 staining, as anti-p53 antibody revealed scatteredlabeling of nuclei (Fig. 4D and E). Cell cytoplasm was intensivelylabeled when the staining was performed with anti-epiregulin anti-bodies (not shown).
In high-grade carcinomas, almost no labeling to pERK1/2 wasevident in most of the fields examined in highly differentiatedtumors originating from the epithelial cells (Fig. 5A and B). Sta-tistically, the labeling of pERK1/2 in nuclei was 2.54 ± 2.21%, whilein the perinuclear region it was 2.48 ± 2.2% (Fig. 5A and B), which issignificantly lower than in intermediate tumors (P < 0.05). In con-trast, all cell nuclei were stained with anti-p53 antibody (Fig. 5C)and the labeling with antibodies to epiregulin was significantlyweaker, compared to tumor cells of intermediate grade (Fig. 5D).
It should be noted that all the images of tumors of patientsinspected in the present study which originated from the stroma,were structurally poorly differentiated (Figs. 2B–F, 3B, 4A–C, and 5).In contrast, all images of tumors of the patients which originatedfrom the epithelial cells of the colon crypts were structurally highlydifferentiated (Figs. 2B, 3A, 4B–D, 5A, C and D).
Discussion
Colon cancer is one of the most lethal cancers if not diagnosed atearly stages (Baba et al., 2000; Manzano and Perez-Segura, 2012). Inthe present study we propose that immunolocalization of pERK1/2can be considered as a useful marker for early diagnosis, whichmay lead to more accurate treatment of the cancer and to improvedprognosis. It should be noted that the current general dogma is thatcolonic cancer may be initiated exclusively from the colon epithelialcells (Gerlee et al., 2011). However, our results suggest that it mayin fact also develop from the adjacent stromal cells as well as fromthe epithelial cells themselves, since we found accumulation ofnuclear pERK immunolocalization in both sites, even in the prema-lignant stage. It should be noted that cancer stem cells can undergoepithelial–mesenchymal transition and these may exhibit a novelfocus of tumor development in the stroma (Scheel and Weinberg,2012). These observations are consistent with our recent findingsthat the stroma cells contain cancer stem cells expressing mainly
the novel stem cells marker LGR5, and the cancerous epithelial cellsin the crypts express mainly the transcription factor Nanog, whichis also used as a reliable marker for cancer stem cells (Amsterdamet al., 2012c).
A. Amsterdam et al. / Acta Histochemica 115 (2013) 569– 576 573
Fig. 4. Low grade colon cancer. (A) Low magnification of stromal area (St) are loaded with many initiation points of tumor growth, identified by perinuclear aggregates ofpERK1/2 (arrows). In a significant portion of the tumors labeling is evident in the nuclei (double arrows). (B) Part of large undifferentiated tumor within the stroma stainedwith antibodies to pERK1/2. Immunostaining is evident in the cancer cell nuclei (arrows). Stroma cells are free of labeling (St). (C) Another part of tumor in the stroma(St) stained with antibodies to pERK1/2. Nuclei are clearly labeled (arrows). (D) Colon carcinoma stained with antibodies to pERK1/2. Labeled nuclei are scattered withinthe tumors (arrows). A significant portion of the nuclei remained unlabeled. (E) Staining with antibodies to p53. Clear labeling in part of the nuclei is evident (arrows).T ree ofa rs: A =
fcb(cadtS
cc
he remaining nuclei remained unstained (double arrows). Stromal tissue (St) is fnti-epiregulin (arrows) most of the stromal cells (St) remained unstained. Scale ba
Internalization of pERK1/2 into the nucleus occurs since growthactors of the family of EGF and their receptors are produced in theolon cancer cells (Li et al., 2010). These growth factors are cleavedy metalloproteinase, which binds and activates the receptor to EGFNishimura et al., 2008). The activated EGFR then activates the ERKascade, including phosphorylation and activation of ERK 1/2. Uponctivation, the latter is translocated to the nucleus by a recentlyiscovered specific mechanism, which involves binding of ERK-NTSo importin7 (Jaaro et al., 1997; Chuderland et al., 2008; Yao and
eger, 2009; Keshet and Seger, 2010; Plotnikov et al., 2011).
We found that prior to entry into the nucleus, pERK1/2 was con-entrated in distinct cores adjacent to the nuclei. Such a perinuclearoncentration was characteristic only in cancer cells originating
labeling. (F) Heavy labeling in the cytoplasm of the carcinoma cells stained with 200 �m; B–F = 50 �m.
from the stromal tissue associated with the colon, which wouldsuggest that the origin of cancer from the stromal tissue is dif-ferent from cancer cells which originate from the colon epithelialcells. It is still not clear whether such aggregates of pERK1/2 aresurrounded by a membrane. This could probably be resolved inthe future using specific antibodies to pERK1/2 attached to col-loidal gold particles and visualized at the electron microscope level(Mayhew and Lucocq, 2008).
It should be noted that a simplified model of tumor growth pro-
gression from adenoma to carcinoma has been proposed, whichincludes the stepwise accumulation of genetic events to severalkey genes and genetic loci; disruption of WNT signaling, activationof the KRAS oncogene, allelic imbalance (A1) on chromosome 18q,
574 A. Amsterdam et al. / Acta Histochemica 115 (2013) 569– 576
Fig. 5. High grade colon cancer. (A) Carcinoma tissue treated with antibodies to pERK1/2 by indirect immunocytochemistry. No staining is evident in the carcinoma cells(arrows) nor stromal cells (St). (B) Tumor tissue (double arrows) embedded in the stromal tissue (St) treated with antibodies to pERK1/2 by the indirect immunocytochemistry.The tumor cells can be identified as clusters of cells with large nuclei. No staining is evident. (C) Carcinoma tissue stained with antibodies to p53. All the nuclei of the carcinomac cells ao = 50 �
rpacssgB
lecowccnfttnnNoaesim
ells are labeled (arrows). No label is evident in the stroma cells (St). (D) Carcinomaf the cells (arrows). No labeling is evident in the stromal cells (St). Scale bars: A–D
educed expression of SMAD4, and mutation of the p53 tumor sup-ression gene. The appearance of pERK1/2 probably coincides withctivation of KRAS, probably only in the tip of the colon epithelialells of the crypts and not in the initiation of tumor growth at thetroma, as was demonstrated in the present study, which woulduggest different genetic alterations in the latter site of tumor pro-ression (Baselga, 2001; Figueras et al., 2001; Jonker et al., 2007;okemeyer et al., 2008; Hurwitz, 2009; Kosakowska et al., 2010).
The disappearance of nuclear pERK1/2 from the nucleus duringater stages of tumor development of the highly differentiated cellsmphasizes the role of pERK1/2 as an earlier marker of colon can-er development. Moreover, the fact that in the intermediate stagenly some of the nuclei are labeled with anti pERK1/2 suggests thate can also identify the shift between low-grade and high-grade
arcinoma of colon tumors. We suggest for the first time that theancer cells “know” their root of development in a somewhat eco-omic manner, since when p53 is already mutated there is no need
or an additional factor such as pERK1/2 to stimulate cell prolifera-ion. However, the mechanisms by which the mutation of p53 affecthe disappearance of pERK1/2 has to be explored. It should also beoted that the positive staining to anti-epiregulin was reduced sig-ificantly in colon tumor cells already expressing the mutated p53.evertheless, it should be noted that although the disappearancef pERK1/2 occurred in the differentiated tumor cells, a small but
significant fraction of undifferentiated tumor cells continued to
xpress both pERK1/2 and mutated p53. It may well be that themall fraction of these cells may exhibit a better potential of spread-ng and metastasizing than tumor cells that express only ERK or
utated p53.
re stained with antibodies to epiregulin. Weak staining is evident in the cytoplasmm.
Our data support the notion that colon cancer is initiated at leastat two different loci. The most common pathway of colorectal can-cer development is thought to be the adeno-carcinoma sequence, inwhich carcinoma develops from an adenomatous polyp (Vogelsteinet al., 1988). The current practice of removing adenoma polyps ofthe colon and rectum is based on the belief that this will prevent orreduce the risk for colorectal cancer (Bertelson et al., 2012). How-ever, reports have described flat and depressed colonic neoplasms(Rembacken et al., 2000), leading to a proposal of an alterna-tive pathway of de novo colon carcinogenesis, which involves anaggressive growth phenotype and rapid infiltration into neighbor-ing tissue and lymph nodes (Shimoda et al., 1989; Nakajima et al.,2007; Matsuda et al., 2009). The most common site of metasta-sis of these cancers is the liver, followed by the lung (Matsumotoet al., 2011). Our observation of islets of cancer cells labeled withpERK1/2 embedded in the stroma, sometimes even distal to theepithelial cells of the crypts, in contrast to pERK1/2 in epithelialcells, add support to this viewpoint. It is proposed that the distribu-tion of intracellular pERK1/2 can distinguish between two origins.In the first case in the low grade stage the pERK1/2 appears in thefoci in the perinuclear region embedded in the stromal cells dis-tal to the epithelial crypt cells, while in the second case pERK1/2appears directly in the nuclei of the epithelial cells. However, onecannot exclude the possibility that in the appropriate homeostaticconditions, in clusters of cells that already contain pERK1/2 in
the perinuclear region, or already in the nuclei, the developmentof aggressive tumors could be arrested. This issue needs to beexplored in the future. Moreover, it could be that because of thehigh turnover of intestinal cells, some stem cells may also contain
istoch
nst2
mmtwomcdtmtd
atc
A
oKuCd
R
A
A
A
A
B
B
B
B
B
A. Amsterdam et al. / Acta H
uclear pERK1/2. Using other typical biomarkers of colon cancertem cells such as LGR5 and Nanog could assist in such a situa-ion (Vogelstein et al., 1988; Shimoda et al., 1989; Vermeulen et al.,008, 2010; Amsterdam et al., 2012b).
It is important to note that in the present study we found thatost of the colon tumors that develop in the stroma showed poororphological organization (i.e., poorly differentiated), in con-
rast to tumor cells initiating from epithelial cells of the colon,hich showed in general a much higher degree of morphological
rganization (highly differentiated). It has been reported that theortality of patients with poorly differentiated colon tumors was
learly and significantly higher than patients demonstrating highlyifferentiated tumors (Shaikh et al., 2009). We therefore suggestedhat in most of the cases of tumors initiating from the stroma, they
ay demonstrate much more aggressive development and metas-asis than in cases of tumors initiated from the colonic epithelium,emonstrating a high degree of differentiated morphology.
Overall, the appearance of pERK1/2 in the colon tissue couldssist in early detection of this severe disease as well as can dis-inguish between the proposed two types of development of colonancer and may assist in selecting of the appropriate treatment.
cknowledgements
We thank Ms. Danielle Sabah-Israel at the Weizmann Institutef Science for typing this manuscript. We also thank Dr. Fortuneohen and Ms. Rina Tzoref, from the Department of Biological Reg-lation and Ms. Daphna Freeman from the Department of Molecularell Biology at the Weizmann Institute of Science, for constructiveiscussions and for editing this manuscript.
eferences
msterdam A, Shezen E, Raanan C, Slilat Y, Ben-Arie A, Prus D, et al.Epiregulin as a marker for the initial steps of ovarian cancerdevelopment. Int J Oncol 2011;39:1165–72.
msterdam A, Shezen E, Raanan C, Schreiber L, Slilat Y, Fabrikant Y,et al. Differential staining of gamma-synuclein in poorly differ-entiated compared to highly differentiated colon cancer cells.Oncol Rep 2012a;27:1451–4.
msterdam A, Raanan C, Schreiber L, Freyhan O, Schecht-man L, Givol D. Localization of the stem cell markersLGR5 and Nanog in the normal and the cancerous humanovary and their inter-relationship. Acta Histochem 2012b.,http://dx.doi.org/10.1016/j.acthis.2012.09.004.
msterdam A, Raanan C, Schreiber L, Freyhana O, Fabrikantd Y,Melzerd E, et al. Differential localization of LGR5 and Nanogin clusters of colon cancer stem cells. Acta Histochem 2012c.,http://dx.doi.org/10.1016/j.acthis.2012.09.003.
aba I, Shirasawa S, Iwamoto R, Okumura K, Tsunoda T, NishiokaM, et al. Involvement of deregulated epiregulin expression intumorigenesis in vivo through activated Ki-Ras signaling path-way in human colon cancer cells. Cancer Res 2000;60:6886–9.
ae JH, Schlessinger J. Asymmetric tyrosine kinase arrangements inactivation or autophosphorylation of receptor tyrosine kinases.Mol Cells 2010;29:443–8.
aselga J. The EGFR as a target for anticancer therapy – focus oncetuximab. Eur J Cancer 2001;37(4):S16–22.
ertelson NL, Kalkbrener KA, Merchea A, Dozois EJ,Landmann RG, De Petris G, et al. Colectomy forendoscopic ally unrespectable polyps: how often
is it cancer? Dis Colon Rectum 2012;55:1111–6,http://dx.doi.org/10.1079/DCR.0b013e3182695115.
okemeyer C, Bondarenko I, Makhson A, Hartmann JT, Aparicio J,de Braud F, et al. KRAS tus and efficacy of first-line treatment of
emica 115 (2013) 569– 576 575
patients with metastatic colorectal cancer (mCRC) with FOLFOXwith or without cetuximab: the OPUS experience. J Clin Oncol2008(26), 178s (suppl, abstr 4000).
Chuderland D, Konson A, Seger R. Identification and character-ization of a general nuclear translocation signal in signalingproteins. Mol Cell 2008;31:850–61.
Derdák Z, Fülöp P, Sabo E, Tavares R, Berthiaume EP, Resnick MB,et al. Enhanced colon tumor induction in uncoupling protein-2 deficient mice is associated with NF-kappaB activation andoxidative stress. Carcinogenesis 2006;27:956–61.
Duhamel S, Hebert J, Gaboury L, Bouchard A, Simon R, Sauter G, et al.Sef downregulation by Ras causes MEK1/2 to become aberrantlynuclear localized leading to polyploidy and neoplastic transfor-mation. Cancer Res 2012;72:626–35.
Fearon ER, Vogelstein B. A genetic model for colorectal tumorigen-esis. Cell 1990;61:759–67.
Figueras J, Valls C, Rafecas A, Fabregat J, Ramos E, Jaurrieta E, et al.Resection rate and effect of postoperative chemotherapy onsurvival after surgery for colorectal liver metastases. Br J Surg2001;88:980–5.
Ginath S, Menczer J, Friedmann Y, Aingorn H, Aviv A, Tajima K, et al.Expression of heparanase Mdm2, and erbB2 in ovarian cancer.Int J Oncol 2001;18§:1133–44.
Hurwitz HI. The clinical benefit of bevacizumab in metastatic colo-rectal cancer is independent of KRAS mutation status: analysis ofphase III study of bevacizumab with chemotherapy in previouslyuntreated metastatic colorectal cancer. Oncologist 2009;1:22–8.
Jaaro H, Rubinfeld H, Hanoch T, Seger R. Nuclear translocation ofmitogen-activated protein kinase kinase (MEK1) in responseto mitogenic stimulation. Proc Natl Acad Sci USA 1997;94:3742–7.
Jacobs B, De Roock W, Piessevaux H, Van Oirbeek R, Biesmans B, DeSchutter J, et al. Amphiregulin and epiregulin mRNA expressionin primary tumors predicts outcome in metastatic colorectalcancer treated with cetuximab. J Clin Oncol 2009;20:5068–74.
Jonker DJ, O’Callaghan CJ, Karapetis CS, Zalcberg JR, Tu D, Au HJ,et al. Cetuximab for the treatment of colorectal cancer. New EnglJ Med 2007;357:2040–8.
Keshet Y, Seger R. The MAP kinase signaling cascades: a system ofhundreds of components regulates a diverse array of physiolog-ical functions. Methods Mol Biol 2010;661:3–38.
Khambata-Ford S, Garrett CR, Neropol NJ, Basik M, Harbison CT,Wu S, et al. Expression of epiregulin and amphiregulin andK-ras mutation status predict disease control in metastatic colo-rectal cancer patients treated with cetuximab. J Clin Oncol2007;25:3230–7.
Klump B, Nehls O, Okech T, Hsieh CJ, Gaco V, Gittinger FS, et al.Molecular lesions in colorectal cancer: impact on prognosis?Original data and review of the literature. Int J Colorectal Dis2004;19:23–42.
Kosakowska EA, Stec R, Charkiewic R, Skoczek M, ChyczewskiL. Molecular differences in the KRAS gene mutation betweena primary tumor and related metastatic sites – case reportand a literature review. Folia Histochem Cytobiol 2010;48:597–602.
Li XD, Miao SY, Wang GL, Yang L, Shu YQ, Yin YM. Amphireg-ulin and epiregulin expression in colorectal carcinoma and thecorrelation with clinicopathological characteristics. Onkologie2010;33:353–8.
Liu H, Liu W, Wu Y, Zhou Y, Xue R, Luo C, et al. Loss of epige-netic control of synuclein-gamma gene as a molecular indicatorof metastasis in a wide range of human cancers. Cancer Res2005;65:7635–43.
iu C, Dong B, Lu A, Qu L, Xing X, Meng L, et al. Synuclein gamma pre-dicts poor clinical outcome in colon cancer with normal levelsof carcinoembryonic antigen. BMC Cancer 2010;7(10):359–68.
anzano A, Perez-Segura P. Colorectal cancer chemoprevention:is this the future of colorectal cancer prevention? Sci World J2012., http://dx.doi.org/10.1100/2012/327341.
atsuda T, Saito Y, Fujii T, Uraoka T, Nakajima T, Kobayashi N, et al.Size does not determine the grade of malignancy of early inva-sive colorectal cancer. World J Gastroenterol 2009;15:2708–13.
atsumoto M, Nakajima T, Kato K, Kouno T, Sakamoto T, MatsudaT, et al. Small invasive colon cancer with systemic metastasis: acase report. BMC Gastroenterol 2011;20(11):59–66.
ayhew TM, Lucocq JM. Developments in cell biology for quan-titative immunoelectron microscopy based on thin sections: areview. Histochem Cell Biol 2008;130:299–313.
edema JP, Vermeulen L. Microenvironmental regulation ofstem cells in intestinal homeostasis and cancer. Nature2011;474(7351):318–26.
ishimura T, Andoh A, Inatomi O, Shioya M, Yagi Y, Tsujikawa T,et al. Amphiregulin and epiregulin expression in neoplastic andinflammatory lesions in the colon. Oncol Rep 2008;19:105–10.
’Neill CJ, Deavers MT, Malpica A, Foster H, McCluggage WG.An immunohistochemical comparison between low-grade andhigh-grade ovarian serous carcinomas: significantly higherexpression of p53, MIB1, BCL2 HER-2/neu, and C KIT in high-grade neoplasms. Am J Surg Pathol 2005;29:1034–41.
lotnikov A, Zehorai E, Procaccia S, Seger R. The MAPK cascades:signaling components, nuclear roles and mechanisms of nucleartranslocation. Biochim Biophys Acta 2011;1813:1619–33.
embacken BJ, Fujii T, Cairns A, Dixon MF, Yoshida S, Chalmers DM,et al. Flat and depressed colonic neoplasms: a prospective study
of 1000 colonoscopies in the UK. Lancet 2000:1211–4.
cheel C, Weinberg R. Cancer stem cells and epithelial–mesenchymal transition: concepts and molecular links. SeminCancer Biol 2012;22:396–403.
emica 115 (2013) 569– 576
Shaikh AJ, Raza S, Shaikh AA, Idress R, Kumar S, Rasheed YA, et al.Demographics, pathologic patterns and long-term survival inoperable colon cancers: local experience in Pakistan. Asian PacJ Cancer Prev 2009;10:361–4.
Shimoda T, Ikegami M, Fujisaki J, Matsui T, Aizawa S, Ishikawa E.Early colorectal carcinoma with special reference to its devel-opment de novo. Cancer 1989;64:1138–46.
Singer G, Stöhr R, Cope L, Dehari R, Hartmann A, Cao DF, et al.Patterns of p53 mutations separate ovarian serous border-line tumors and low- and high-grade carcinomas and providesupport for a new model of ovarian carcinogenesis: a muta-tional analysis with immunohistochemical correlation. Am JSurg Pathol 2005;29:218–24.
Ueda K, Arakawa H, Nakamura Y. Dual-specificity phosphatase 5(DUSP5) as a direct transcriptional target of tumor suppressorp53. Oncogene 2003;28:5586–91.
Vermeulen L, Sprick MR, Kemper K, Stassi G, Medema JP. Can-cer stem cells - old concepts, new insights. Cell Death Differ2008;15:947–58.
Vermeulen L, De Sousa E, Melo F, van der Heijden M, CameronK, de Jong JH, et al. Wnt activity defines colon cancer stemcells and is regulated by the microenvironment. Nat Cell Biol2010;12:468–76.
Vogelstein B, Fearon ER, Hamilton SR, Kern SE, Preisinger AC,Leppert M, et al. Genetic alterations during colorectal-tumordevelopment. N Engl J Med 1988;319:525–32.
Wu WK, Tse TT, Sung JJ, Li ZJ, Yu L, Cho CH. Expression of ErbBreceptors and their cognate ligands in gastric and colon cancercell lines. Anticancer Res 2009;29:229–34.
Yao Z, Seger R. The ERK signaling cascade – views from differentsubcellular compartments. Biofactors 2009;35:407–16.
Zhao SL, Hong J, Xie ZQ, Tang JT, Su WY, Du W, et al. TRAPPC4–ERK2interaction activates ERK1/2, modulates its nuclear localizationand regulates proliferation and apoptosis of colorectal cancercells. PLoS ONE 2011;6(8):e23262.
Zeren T, Inan S, Seda Vatansever H, Ekerbicer N, Sayhan S. Signif-icance of tyrosine kinase activity on malign transformation ofovarian tumors: a comparison between EGF-R and TGF-alpha.Acta Histochem 2008;110:256–63.
![Guidance - Aggregates[1]](https://static.documents.pub/doc/80x56/543d5902b1af9f410a8b474f/guidance-aggregates1.jpg)
