wnt-driven intestinal tumourigenesis is suppressed by chk1 deficiency but enhanced by conditional...

ORIGINAL ARTICLE

Wnt-driven intestinal tumourigenesis is suppressed by Chk1deficiency but enhanced by conditional haploinsufficiencyKR Greenow1, AR Clarke1, GT Williams2 and R Jones1

Chk1 is essential in maintaining genomic stability due to its role in cell cycle regulation. Several recent studies have indicated thatthe abrogation of checkpoints in tumourigenesis through the inhibition of Chk1 may be of therapeutic value. To further investigatethe role of Chk1 in the mouse small intestine and its potential role as a therapy for colorectal cancer, we simultaneously deletedChk1 and Apc in the mouse small intestine. We found that homozygous loss of Chk1 is not compatible with Wnt-driven proliferationand resulted in the suppression of Wnt-driven tumourigenesis in the mouse small intestine. In contrast, heterozygous loss of Chk1in a Wnt-driven background resulted in an increase in DNA damage and apoptosis and accelerated both tumour development andprogression.

Oncogene advance online publication, 16 September 2013; doi:10.1038/onc.2013.371

Keywords: Chk1; Apc; genomic instability; small intestine; apoptosis

INTRODUCTIONGenomic instability appears to be a crucial feature in tumourdevelopment and is thought to be one of the key characteristics oflate-stage colon cancer.1 The loss of genomic stability in tumourprogression through DNA damage, tumour-specific DNA repairdefects and the loss of cell cycle checkpoints has been proposedto create a permissive environment for potential changes in bothoncogenes and tumour-suppressor genes. The accumulation ofthese genetic and epigenetic alterations is thought to lead tothe transformation of normal colon epithelial cells to colonadenocarcinoma cells, resulting in the stepwise development ofcolorectal cancer.2 In addition to driving the disease, the genomicinstability of tumours also provides therapeutic opportunities, andthe exploitation of the reliance of cancer cells on the DNA damageresponse pathways has been shown to be a valid therapeuticstrategy.3–6

The ATR-Chk1 checkpoint pathway is a key DNA damageresponse pathway that has been shown to be an important barrierto the progression of pre-neoplastic lesions and has the potentialto be used as a synthetic lethal target in tumourigenesis.7–9 Chk1is a serine/threonine protein kinase that is active at the intra-S-phase, G2/M and the spindle checkpoints of the cell cycle andis therefore a key factor in maintaining genomic stability.10 Severalrecent studies have indicated that the abrogation of thesecheckpoints through the inhibition of Chk1 may be oftherapeutic value, as the loss of the G2/M checkpoint causes thesensitization of tumour cells that are defective in the G1checkpoint to DNA-damaging agents and results in theredirection of tumour cells to apoptotic pathways.4,6,11

Homozygous loss of murine Chk1 is embryonically lethal12 andtissue-specific deletion of Chk1 results in extensive cell death inproliferating tissues but has been shown to be tolerated inquiescent tissues such as the liver.13–15 This increase in apoptosisin proliferative cells due to Chk1 deletion has been associated withgenomic instability and p53 accumulation, although this apoptoticphenotype has been shown to be p53-independent by several

studies.14–16 Chk1 heterozygosity has also been shown to resultin checkpoint disruption, genomic instability and enhancedtumourigenesis in specific tissues implicating it as a tumoursuppressor.13,17 Recent work has also confirmed the potential roleof Chk1 as a feasible cancer-cell-specific therapy as Tho et al.17

demonstrated that Chk1 is indispensable for the formation ofchemical-induced skin tumours.

We have previously used a conditional model to investigate theconsequences of Chk1 loss in the murine small intestine.11 Thesestudies showed that homozygous loss of Chk1 resulted in theablation of the Chk1 null epithelium and subsequent repopulationby wild-type cells. However, we did not observe an atypicalphenotype in the Chk1 heterozygous epithelium. We have nowextended these studies to determine the effect of Chk1 loss in thecontext of activated Wnt signalling. First, as might be predictedfrom previous studies, we show that the complete loss ofChk1 is able to suppress Wnt-driven proliferation andtumourigenesis, although Chk1 haploinsufficiency promotesWnt-driven tumourigenesis.

RESULTSHomozygous loss of Chk1 suppresses Wnt-driven proliferation inthe mouse small intestineIn order to investigate the effect of Chk1 loss on the Apc phenotype,Chk1 mice carrying the loxP-flanked Chk1 allele were crossed ontothe conditional Apc background.18,19 Cre activity was inducedin experimental (AhCreþApcfl/flChkfl/þ , AhCreþApcfl/flChkfl/fl) andcontrol mice (AhCreþApcþ /þChkþ /þ , AhCreþApcfl/flChkþ /þ , AhCreþ

Apcþ /þChkfl/fl) by intraperitoneal injection of b-naphthoflavone.Previous studies have demonstrated that three injections of 80 mg/kg of b-naphthoflavone within 24 h result in efficient recombinationof both the Chk1 allele14 and the Apc allele20 in the mouse smallintestine. Due to the short-term survival of induced AhCreþ

Apcfl/flChkþ /þ mice and the transient phenotype of the induced

1Cardiff School of Biosciences, Cardiff University, Cardiff, UK and 2School of Medicine, Cardiff University, Cardiff, UK. Correspondence: Professor AR Clarke, Cardiff School ofBiosciences, Cardiff University, Museum Avenue, PO Box 911, Cardiff, Wales CF10 3AX, UK.E-mail: [email protected] 18 February 2013; revised 17 May 2013; accepted 7 June 2013

Oncogene (2013), 1–8& 2013 Macmillan Publishers Limited All rights reserved 0950-9232/13

www.nature.com/onc

http://dx.doi.org/10.1038/onc.2013.371

mailto:[email protected]

http://www.nature.com/onc

AhCreþApcþ /þChkfl/fl mice, both experimental and control cohortswere killed at day 2 and day 4 time points post-induction (PI).

In order to assess the phenotypic consequences of theconditional deletion of both Chk1 and Apc, we analysed thecrypt� villus structure. Histological sections from day 2 PI revealedthe previously characterized apoptotic phenotype of Chk1 loss inthe mouse small intestine in the Chk1 homozygous mice (AhCreþ

Apcþ /þChkfl/fl and AhCreþApcfl/flChkfl/fl), whereas the small intes-tine of the Chk1 heterozygous mice (AhCreþApcfl/flChkfl/þ ) showedno atypical phenotype (Supplementary Figure 1a). These observa-tions are consistent with complete loss of the Chk1 protein in theChk1 homozygous mice at day 2 and a reduction in Chk1 proteinlevels in the AhCreþApcfl/flChkfl/þ mice (Figure 1b).14 At day 2 PI,no abnormal phenotype was seen in the AhCreþApcfl/flChkþ /þ

and the AhCreþApcfl/flChkfl/þ mice as the previously characterized‘crypt progenitor-like’ phenotype of Apc loss in the mouse smallintestine does not occur until day 4 PI,21 whereas, at day 4 PI(Figure 1a), both of these cohorts demonstrated an increase incrypt cell proliferation and perturbed migration, consistent withprevious studies.21

The phenotype of the AhCreþApcþ /þChkfl/fl control mice atday 4 PI was also consistent with our previous studies, wherebyChk1 loss resulted in widespread apoptosis, crypt death andsubsequent crypt repopulation (Figure 1a). From the histologicalsections shown (Figure 1a), both crypt death and repopulation ofcrypts can be observed at day 4 PI in both the control AhCreþ

Apcþ /þChkfl/fl and the experimental AhCreþApcfl/flChkfl/fl mice,demonstrating that the loss of Chk1 effectively abrogated thephenotypic consequences of Apc loss in the mouse smallintestine. Repopulation of the mouse small intestine with cellscontaining a functional unrecombined Chk1 allele in AhCreþ

Apcþ /þChkfl/fl mice was confirmed by the presence of nearlywild-type levels of Chk1 protein at day 4 PI (Figure 1b). At day 4PI, the AhCreþApcfl/flChkfl/fl mice also showed the presence ofChk1 protein (Figure 1b), although some heterogeneity withrespect to the protein levels were seen (Supplementary Figure 2),which may be due to a variation in the repopulation levels withinthis cohort.

To quantify the phenotypic changes indicated by the histolo-gical analysis, the number of live crypts and cells per crypt werescored from the haematoxylin and eosin-stained sections asdescribed previously.22 Dying crypts were characterized by a highlevel of apoptosis and loss of the crypt� villus structure, whereaslive crypts consisted of low level apoptosis, a normal crypt� villusstructure and demonstrated cell proliferation. The data for thecontrol mice (AhCreþApcþ /þChkþ /þ , AhCreþApcþ /þChkfl/fl,AhCreþApcfl/flChkþ /þ ; Figure 2a) at day 4 PI were consistent withprevious data,14,18,21 with a significant decrease in crypt survivalfor the AhCreþApcþ /þChkfl/fl mice compared with wild type(n¼ 4, Mann–Whitney U-test P¼ 0.0259) and no significantchange between the wild type and AhCreþApcfl/flChkþ /þ mice(n¼ 3, Mann–Whitney U-test P¼ 0.3313). Similar to theobservations from the histological analysis, the loss of a singleChk1 allele had no effect on the Apc homozygous phenotype withrespect to crypt survival (n¼ 3, Mann–Whitney U-test P¼ 1.00),whereas homozygous loss of Chk1 in combination with Apcdeletion (AhCreþApcfl/flChkfl/fl) resulted in a significant decrease incrypt survival compared with the AhCreþApcfl/flChkþ /þ mice(n¼ 7, Mann–Whitney U-test P¼ 0.0113). In addition,quantification of live crypts allowed us to assess the effect ofthe concurrent loss of Apc and Chk1 on the repopulation of thesmall intestine. Our results above (Figure 1b) indicate that the lossof Apc results in delayed repopulation in Chk1 homozygousmouse; however, this was not reflected in our results as nosignificant difference in live crypts was observed between AhCreþ

Apcþ /þChkfl/fl and AhCreþApcfl/flChkfl/fl at day 4 PI (Figure 2a;Mann–Whitney U-test, P¼ 0.1493), although a heterogeneousphenotype with respect to crypt survival was observed.

To determine whether there were any differences in cellularitywithin the live crypt population, the number of cells per crypt wasscored (Figure 2b). Similar to the crypt survival results, the data forthe control mice at day 4 PI (AhCreþApcþ /þChkþ /þ , AhCreþ

Apcþ /þChkfl/fl, AhCreþApcfl/flChkþ /þ ) were consistent withprevious results,14,18,21 with a significant decrease in crypt cellnumber for the AhCreþApcþ /þChkfl/fl mice compared with wildtype (n¼ 4, Mann–Whitney U-test P¼ 0.0404) and a significantincrease in crypt size between the wild-type and AhCreþApcfl/fl

Chkþ /þmice (n¼ 3, Mann–Whitney U-test P¼ 0.0404). Loss of asingle Chk1 allele had no effect on crypt cellularity of the Apchomozygous mice (Figure 2b; AhCreþApcfl/flChkfl/þ versus AhCreþ

Apcfl/flChkþ /þ , Mann–Whitney U test P¼ 0.6625), although homo-zygous loss of Chk1 did result in a significant decrease in crypt cellnumber (n¼ 7, Mann–Whitney U test P¼ 0.0259), confirming thatChk1 loss in combination with Apc loss results in the attenuation ofthe Apc phenotype.

In order to further assess the effect of the simultaneous loss ofChk1 and Apc on the Apc phenotype and also to confirm the lossof Apc, we carried out immunohistochemical analysis of b-catenin.Nuclear relocalization of b-catenin, as a result of dysregulatedWnt signalling, is a widely used surrogate marker of Apc loss.Immunohistochemical analysis confirmed Apc loss in our AhCreþ

Apcfl/flChkþ /þ and AhCreþApcfl/flChkfl/þmice, with the majority ofepithelial cells showing strong nuclear localization of b-catenincompared with wild-type and AhCreþApcþ /þChkfl/fl control mice,where b-catenin is predominantly cytoplasmic or membraneassociated (Figure 2c). Interestingly, the AhCreþApcfl/flChkfl/fl miceshowed heterogeneous localization of b-catenin (Figure 2c).Dying crypts appear to have nuclear b-catenin, whereas the livecrypts appear to have cytoplasmic or membrane-associatedb-catenin, demonstrating that, after crypt death due to theloss of Chk1, crypt repopulation occurs with wild-typeApc-expressing cells.

Chk1 heterozygosity results in a haploinsufficient phenotype whenApc is also lostIn addition to quantifying crypt survival and cell number, apoptosisand mitosis were quantified as described previously.17 The data forthe control mice (AhCreþApcþ /þChkþ /þ , AhCreþApcþ /þChkfl/fl,AhCreþApcfl/flChkþ /þ ) at day 4 PI were consistent with previousdata,14,18,21 with a significant increase in apoptosis between theAhCreþApcþ /þChkfl/fl (Mann–Whitney U-test, P¼ 0.0404) andwild-type mice and an increase in both apoptosis and mitosis inthe AhCreþApcfl/flChkþ /þ (Mann–Whitney U-test, P¼ 0.0404)compared with wild-type mice (Figures 3a and b). Chk1 deletionalso resulted in a significant further increase in apoptosis in theAhCreþApcfl/flChkfl/fl mice compared with the AhCreþApcfl/fl

Chkþ /þ mice (Mann–Whitney U-test, P¼ 0.0259) and resulted inan abrogation of the cell proliferation phenotype of Apc loss asshown by a decrease in mitosis and BrdU-labelled cells (AhCreþ

Apcfl/flChkfl/fl versus AhCreþApcfl/flChkþ /þ , Mann–Whitney U-test,P¼ 0.0404; Figures 3a–c). Notably, although the heterozygous lossof Chk1 had no effect at day 2 PI (Supplementary Figure 1b), atday 4 PI, the loss of the single Chk1 allele in combination withthe homozygous loss of Apc resulted in a significant increasein apoptosis (AhCreþApcfl/flChkfl/þ versus AhCreþApcfl/flChkþ /þ ,Mann–Whitney U test P¼ 0.0404; Figure 3a).

The above observations show that, in addition to thehomozygous loss of Chk1 having a significant effect on the Apcphenotype in the mouse small intestine, conditional heterozygousloss of Chk1 in the context of Apc deficiency elevates cell death.To further investigate whether this increase in apoptosis was dueto Chk1 heterozygosity promoting DNA damage, immunohisto-chemical analysis was carried out for phospho-H2A.X, anestablished DNA-damage marker that becomes phosphorylatedand activated in the presence of DNA damage.23,24 Conditional

Chk1 loss and Wnt-driven intestinal tumorigenesisKR Greenow et al

2

Oncogene (2013) 1 – 8 & 2013 Macmillan Publishers Limited

homozygous loss of Chk1 resulted in an increase in phospho-H2A.X (Figure 3d), indicating an increase in DNA damage. Micehomozygous for Apc loss (AhCreþApcfl/flChkþ /þ ) also showed an

increase in H2A.X staining, which was further increased by the lossof a single Chk1 allele. This increase in phospho-H2A.X stainingwas quantified and confirmed a significant increase in phospho-

Figure 1. Combined conditional loss of Apc and Chk1 in the mouse small intestine ablates the Apc homozygous phenotype. Cre activity wasinduced in experimental mice (AhCreþApcfl/flChkfl/þ , AhCreþApcfl/flChkfl/fl) and control mice (AhCreþApcþ /þChkþ /þ , AhCreþApcþ /þ Chkfl/fl,AhCreþApcfl/flChkþ /þ ) by three consecutive intraperitoneal injections of 80mg/kg b-naphthoflavone within 24 h. Animals were culled at day 2and day 4 time points after induction, and tissue was isolated from both control and experimental animals. (a) H&E-stained histologicalsections of mouse small intestine from experimental mice (AhCreþApcfl/flChkfl/þ , AhCreþApcfl/flChkfl/fl) and control mice (AhCreþApcþ /þChkþ /

þ , AhCreþApcþ /þChkfl/fl, AhCreþApcfl/flChkþ /þ ) on day 4 PI. Recombination of Chk1 and Apc in control mice (AhCreþApcþ /þ Chkþ /þ ,AhCreþApcþ /þChkfl/fl, AhCreþApcfl/flChkþ /þ ) resulted in previously reported phenotypes, whereas the AhCreþApcfl/flChkfl/fl experimental miceshowed a severe phenotype of crypt death, which was not seen in the experimental Chk1 heterozygous mouse (AhCreþApcfl/flChkfl/þ ). Bar,50mm. (b) Loss of Chk1 protein was confirmed by western blot analysis. Protein was isolated from tissue samples enriched for intestinalepithelium cells and subjected to SDS� PAGE. Blotted membranes were incubated with a Chk1 mouse monoclonal antibody overnightand the membrane was re-probed with a b-actin mouse monoclonal antibody to confirm equal protein loading. Densitometry was carriedout to quantify Chk1 protein expression, and optical density values are expressed as a ratio between Chk1 andb-actin. Data shown are representative of a minimum of three separate experiments and are shown relative to the wild-type control ateach time point.


3

& 2013 Macmillan Publishers Limited Oncogene (2013) 1 – 8

H2A.X-positive cells (AhCreþApcfl/flChkfl/þ versus AhCreþ

Apcfl/flChkþ /þ , Mann–Whitney U-test P¼ 0.0404; SupplementaryFigure 3).

Loss of Chk1 modifies Wnt-driven tumourigenesisFrom the above short-term data, it is clear that the loss of Chk1has a significant impact on Wnt-driven proliferation, which couldpotentially perturb Wnt-driven tumourigenesis. In order to testthis hypothesis, we crossed the loxP-flanked Chk1 onto aconditional Apc heterozygous background, a well-established

colorectal cancer model.19 Cre activity was induced in bothexperimental (AhCreþApcfl/þChkfl/þ , AhCreþApcfl/þChkfl/fl) andcontrol mice (AhCreþApcfl/þChkþ /þ ) by three intraperitonealinjections of 80 mg/kg b-naphthoflavone in 24 h. We inducedcohorts of more than 15 experimental and control mice andallowed them to age. The mice were monitored regularly for signs ofintestinal tumours (rectal bleeding, prolapse and anaemia) or otherillness and were killed when they became symptomatic of disease.

As shown in Figure 4a, loss of either a single or both Chk1 alleleshad a significant effect on survival. The Chk1 heterozygous(AhCreþApcfl/þChkfl/þ ) mice had a significantly reduced median

Figure 2. Combined loss of Apc and Chk1 results in crypt death and repopulation of the mouse small intestine with Apc wild-type crypts.(a and b) H&E-stained tissue sections were used to quantify live crypts and number of cells per crypt. The number of live crypts percircumference of the intestine was scored and the data shown are means±s.d. of a minimum of four independent experiments. A significantdecrease in the number of live crypts and cells per crypt between experimental, AhCreþApcfl/flChkfl/fl, and control, AhCreþApcfl/flChkþ /þ , tissuesections can be seen at day 4 PI. *Po0.05. Bar, 50 mm. (c) Immunohistochemical analysis of b-catenin localization within intestinal tissuesections from experimental mice (AhCreþApcfl/flChkfl/þ , AhCreþApcfl/flChkfl/fl) and control mice (AhCreþApcþ /þChkþ /þ , AhCreþApcþ /þChkfl/fl,AhCreþApcfl/flChkþ /þ ) on day 4 PI.


4


survival of 441 days (n¼ 19), compared with the control cohort(AhCreþApcfl/þChkþ /þ ), which had a median survival of 476 days(n¼ 16, Mann–Whitney U-test, P¼ 0.0411), whereas the Chk1homozygous cohort (AhCreþApcfl/þChkfl/fl) had a significantlyincreased median survival of 665 days (n¼ 23, Mann–WhitneyU-test, P¼ 0.0013). All cohorts developed polyp-like lesions withinthe small intestine and the large intestine, and, although nosignificant difference was seen in the number of adenomas in thesmall intestine between either experimental cohort (AhCreþ

Apcfl/þChkfl/þ or AhCreþApcfl/þChkfl/fl) and the control cohort(AhCreþApcfl/þChkþ /þ ; Figure 4b), a significant decrease wasobserved between the number of adenomas in the large intestineof the AhCreþApcfl/þChkfl/fl compared with AhCreþ Apcfl/þChkþ /þ

(Mann–Whitney U-test, P¼ 0.015; Figure 4b). In addition, tumour sizewas measured, although no significant difference was seen betweenour experimental and control cohorts in either the small intestine orthe large intestine. Intestinal tumour burden was also assessed(Figure 4d) and a significant reduction was seen in the large intestine

Figure 3. Chk1 heterozygosity in the context of Apc loss results in an increase in apoptosis and DNA damage. (a and b) H&E-stained intestinaltissue sections from experimental mice (AhCreþApcfl/flChkfl/þ , AhCreþApcfl/flChkfl/fl) and control mice (AhCreþApcþ /þChkþ /þ , AhCreþApcþ /þ

Chkfl/fl, AhCreþApcfl/flChkþ /þ ) on day 4 PI were used to quantify apoptosis and mitosis. Data shown are means±s.d. of a minimum of fourindependent experiments. *Po0.05. (c) To examine S-phase labelling in vivo, control (AhCreþApcþ /þChkþ /þ , AhCreþApcþ /þChkfl/fl, AhCreþ

Apcfl/flChkþ /þ ) and experimental (AhCreþApcfl/flChkfl/þ , AhCreþApcfl/flChkfl/fl) animals were injected with 100 mg/kg BrdU and culled 2 h afterlabelling. BrdU incorporation was determined by immunohistochemistry and quantified by scoring positively stained cells.(d) Immunohistochemical analysis of intestinal tissue sections from experimental mice (AhCreþApcfl/flChkfl/þ , AhCreþApcfl/flChkfl/fl) andcontrol mice (AhCreþApcþ /þChkþ /þ , AhCreþApcþ /þChkfl/fl, AhCreþApcfl/flChkþ /þ ) at day 4 PI. Tissue sections stained with antibodiesdetecting phosphorylated-H2A.X. Bar, 50 mm.


5


of the AhCreþApcfl/þChkfl/fl versus AhCreþApcfl/þ ; Chkþ /þ (Mann–Whitney U-test, P¼ 0.0067), indicating that homozygous deletion ofChk1 strongly suppresses Wnt-driven tumour formation.

Chk1 status influences tumour severity in the mouse smallintestineIn order to understand the decrease in survival seen in the Chk1heterozygous cohort, we histologically examined and character-ized the small intestine and large intestine lesions of both theAhCreþApcfl/þChkfl/þ and the AhCreþApcfl/þChkþ /þ cohorts(Figure 5a). b-catenin staining confirmed Wnt-driven tumourigenesis(Figure 5a), and Chk1 protein expression was confirmed in polypsisolated from both experimental (AhCreþApcfl/þChkfl/þ , AhCreþ

Apcfl/þChkfl/fl) and control mice (AhCreþApcfl/flChkþ /þ ) indicatingthat Chk1 expression is vital for Wnt-driven tumourigenesis(Figure 5c).

We classified the lesions into microadenomas (T1) and adenomasthat were confined to the mucosa (T2), early invasive adenocarci-nomas showing invasion into the submucosa but not themuscularis propria (T3) and advanced invasive adenocarcinomathat penetrated into or through the muscularis propria (T4). Fromour observations, the control (AhCreþApcfl/þChkþ /þ ) miceanalysed had either benign microadenomas (44.4%), adenomas(100%) confined to the submucosa or adenocarcinomas (66.7%);however, only 11.1% had tumours invading into the muscularispropria (Figure 5b). In contrast, 76.9% of the AhCreþApcfl/þChkfl/þ

mice had muscle-invading adenocarcinomas in which the tumourshad penetrated the muscularis propria to reach the serosa orsubserosa, sometimes with the loss of differentiation, features thatwere never seen in the control mice. These data thereforedemonstrate that the loss of a single Chk1 allele enhances tumourprogression.

DISCUSSIONHuman tumours frequently have defects in the maintenance ofgenomic integrity, which involves a loss of DNA damage responsepathways and checkpoints, which causes cancer cells to be more

reliant on intact cell cycle checkpoints and DNA repair pathways.This reliance can provide an opportunity for therapeutic interven-tion, and it may be possible to target tumour cells specifically, withlimited toxicity, in normal tissue. The success of this strategy isillustrated by the progress of Chk1 inhibitors in early-phase clinicaltrials.6 Although Chk1 inhibition does appear to be a promisingstrategy in cancer therapy and several studies have shown thatChk1 inhibition enhances tumour cell death,17,25 a potentialhaploinsufficient role for Chk1 has also been shown12,13,15–17

indicating that further studies are required to fully understand theuse of Chk1 inhibition as a cancer therapy.

We have previously deleted Chk1 in the mouse small intestineand have demonstrated that, although the homozygous loss of thischeckpoint protein has a highly traumatic, albeit transient effect, theheterozygous loss of Chk1 is well tolerated.14,25 In order to furtherinvestigate the potential use of Chk1 inhibition in cancer therapy,particularly in colorectal cancer, we used an early-stage Wnt-activated model to characterize the effect of Chk1 deletion on theinitial stages of aberrant Wnt signalling in the mouse small intestine.

The study presented in this paper clearly shows that thehomozygous deletion of Chk1 in the mouse small intestine incombination with Apc prevents Wnt-driven proliferation, one ofthe key events in colorectal cancer. Our short-term studiesdemonstrated that the loss of both Chk1 alleles in the Apchomozygous mouse resulted in widespread apoptosis and cryptdeath and subsequent repopulation of the intestine with wild-type cells positive for Apc and Chk1. These results are inaccordance with previous studies whereby Chk1 has been shownto be essential for the viability of proliferating cells due to its keyrole in cell cycle regulation.13–15 In addition, our short-term studiesdemonstrate that Chk1 has a haploinsufficient phenotype whenlost in a Wnt-driven environment. Heterozygous loss of Chk1 incombination with Apc resulted in an increase in apoptosis andDNA damage. Our previous studies have demonstrated that theloss of a single Chk1 allele in the mouse small intestine has nogross effect, and therefore this difference may be partly due to thecrypt progenitor-like phenotype of Apc deletion in the mousesmall intestine. Loss of Apc results in a decrease in the number ofterminally differentiated cells, an increase in proliferation and

Figure 4. Homozygous Chk1 deletion suppresses Wnt-driven tumourigenesis and increases survival. Cre activity was induced in control(AhCreþApcfl/þChk1þ /þ ) and experimental (AhCreþApcfl/þChk1fl/þ and AhCreþApcfl/þChk1fl/fl) mice. Animals were culled when they showedsigns of illness, and tissue was isolated from both control and experimental animals. (a) Kaplan–Meier survival analysis of the control andexperimental cohorts. Loss of Chk1 in the experimental cohorts (AhCreþApcfl/þChkfl/þ and AhCreþApcfl/þChkfl/fl) showed a significantdifference in lifespan when compared with the control cohort (AhCreþApcfl/þChkþ /þ ). (b–d) Formalin-fixed tissue was used to quantifytumour number, size and burden. The quartered circle symbol denotes the mean value.


6


endogenous DNA damage, and the intestine may be less resilientto the Chk1 heterozygosity, which is consistent with previousstudies demonstrating that the loss of a single allele of keycomponents of the DNA damage response pathway becomescritical under conditions of increased endogenous DNA damage inpre-malignant lesions, a theory described by Bartek et al.9 as‘conditional haploinsufficiency’.

In addition to our short-term studies, we used the Apcheterozygous tumour model to assess the effect of Chk1 loss ontumourigenesis in the mouse intestine. A recent study by Thoet al.17 has demonstrated that Chk1 deletion suppresses mouseskin tumourigenesis, although Chk1 heterozygosity resulted intumour progression. We show that both heterozygous andhomozygous deletion of Chk1 resulted in a significant effect onWnt-driven tumourigenesis. Homozygous loss of Chk1 resulted inan increase in survival and a decrease in tumour number in thelarge intestine. This decrease in Wnt-driven tumourigenesis maybe due to the ability of Chk1 loss to suppress Wnt-drivenproliferation through apoptosis, crypt death and the repopulationof the intestine with wild-type Apc, eliminating Apc-deficient cellsfrom the tissue. Although the homozygous loss of Chk1 did notcompletely suppress tumorigenesis and no difference was seen intumour burden in the small intestine at time of death, it isimportant to note that tumorigenesis was slowed as Chk1homozygous mice survived 50% longer than control Chk1wild-type mice. This inability to completely suppress Wnt-driventumourigenesis may be explained by floxing inefficiency of Chk1resulting in the survival of some Apc-homozygous cells andalso may be due to spontaneous secondary mutations in thistumour-permissive environment.

In contrast, Chk1 heterozygosity resulted in a significant decreasein survival compared with Chk1 wild-type mice. No increase in

tumour number or size was seen, although, consistent with previousstudies,17 loss of a single Chk1 allele resulted in enhancedtumourigenesis with over a 60% increase in the number of micedeveloping advanced invasive adenocarcinomas compared withcontrols. This enhanced tumourigenesis may be due to an increasein genomic instability due to the loss of Chk1, as was suggested bythe increase in DNA damage observed in our short-term studies.This increased susceptibility to tumour progression by heterozygousdefects in components of DNA damage response pathways hasbeen shown by previous studies and is thought to be due to thecombination of the DNA-damage threshold being exceeded and theloss of cell cycle checkpoints, both of which facilitate tumourprogression.9,11 In addition, the impairment in Chk1 function couldalso increase the probability of additional genetic changes; forexample, it is known that the progression of the intestinaltumourigenesis is associated with mutation of additional genessuch as Ras,26 although analysis of our tumours showed this not tobe the case in this study (data not shown).

Our observations here are consistent with previous studieswhereby the deletion of Chk1 suppresses tumourigenesis.17 Thisdependence is currently being exploited in the treatment oftumours through the use of Chk1 inhibitors. Although thecomplete loss of Chk1 appears to reduce tumourigenesis, bothour results and previous studies demonstrate that Chk1heterozygosity results in spontaneous DNA damage and anincrease in tumourigenesis within specific tissues.13,15–17

Therefore, Chk1 clearly has a gene-dosage effect, a conceptwhich has been shown for other components of the DNA damageresponse pathway,9,11 indicating that, although complete deletionof Chk1 may be a successful cancer therapy, the level of Chk1inhibition is critical and further studies are required to fullyunderstand the effect Chk1 haploinsufficiency.

Figure 5. Chk1 heterozygosity results in tumour progression. (a) Histological examination of small intestine tumours from AhCreþApcfl/þChkfl/þ

mice revealed adenomas and invasive adenocarcinoma. Immunohistochemical analysis of b-catenin localization within intestinal tissuesections from experimental mice (AhCreþApcfl/þChkfl/þ ) and control mice (AhCreþApcfl/þChkþ /þ ). Bar, 200 mm. (b) Classification of tumoursfrom AhCreþApcfl/þChkfl/þ and AhCreþApcfl/þChkþ /þ mice. (c) Chk1 expression in tumour tissue was confirmed by western blot analysis.Protein was isolated from tumour tissue and subjected to SDS–PAGE. Blotted membranes were incubated with a Chk1 mouse monoclonalantibody overnight, and the membrane was re-probed with a b-actin mouse monoclonal antibody to confirm equal protein loading.


7


MATERIALS AND METHODSExperimental miceAll procedures were conducted according to the UK Home Officeregulations. Mice carrying the floxed Chk1 allele were kindly supplied byDr Stephen J. Elledge13 and all mice were maintained on an outbredbackground. Experimental mice were genotyped as previously describedfor the targeted Chk1 allele,13 the targeted Apc allele,19 the Rosa26R allele27

and the AhCre transgene.20 Cre activity was induced in control andexperimental mice by three consecutive intraperitoneal injections of80 mg/kg b-naphthoflavone (Sigma, Dorset, UK) within 24 h. In addition,selected animals were injected with 100mg/kg Bromo-deoxyuridine(Sigma) and culled at indicated time points after labelling.

Histology and immunohistochemistryIntestinal tissue was fixed in ice-cold 10% neutral-buffered formalin for nolonger than 24 h before being processed into paraffin blocks according tostandard procedures. Tissue sections (5 mm) were either stained usinghaematoxylin and eosin for histological analysis or were used forimmunohistochemical analysis. The following antibodies were used forimmunhistochemistry: anti-Caspase 3 (1:750; R&D Systems, Abingdon, UK),mouse anti-H2A.X (Ser139) (1:200, Millipore, Billerica, MA, USA), anti-b-catenin (1/50; Becton Dickinson, Oxford, UK) and mouse anti-BrdU (1:100;Becton Dickinson).

Western blot analysisProtein was isolated from either intestinal epithelial-enriched pellets28 orpolyps isolated from experimental mice, and subsequent protein analysis,SDS� PAGE and western blotting were carried out following standardprotocols. The Chk1 mouse monoclonal antibody (Abgent, Oxfordshire, UK)was used at 1:1000 and the mouse monoclonal b actin (Sigma) was used at1:12000, and the appropriate horseradish peroxidase-conjugatedsecondary antibodies (GE Healthcare, Buckinghamshire, UK) were used at1:3000. Densitometry was carried out using Image J software (v1.46 d).

CONFLICT OF INTERESTThe authors declare no conflict of interest.

ACKNOWLEDGEMENTSThis work was supported by Cancer Research UK. Particular thanks go to Mark Bishop,Lucie Pietzka and Derek Scarborough for technical assistance. The authors declare noconflicts of interest.

REFERENCES1 Grady WM. Genomic instability and colon cancer. Cancer Metastasis Rev 2004; 23:

11–27.2 Soreide K, Janssen E, Soiland H, Korner H, Baak J. Microsatellite instability in

colorectal cancer. Br J Surg 2006; 93: 395–406.3 Venkatesha VA, Parsels LA, Parsels JD, Zhao L, Zabludoff SD, Simeone DM et al.

Sensitization of pancreatic cancer stem cells to gemcitabine by Chk1 inhibition.Neoplasia 2012; 14: 519–525.

4 Montano R, Chung I, Garner KM, Parry D, Eastman A. Preclinical development ofthe novel Chk1 inhibitor SCH900776 in combination with DNA-damaging agentsand antimetabolites. Mol Cancer Ther 2012; 11: 427–438.

5 Plummer R. Poly(ADP-ribose) polymerase inhibition: a new direction for BRCA andtriple-negative breast cancer? Breast Cancer Res 2011; 13: 218.

6 Ma CX, Janetka JW, Piwnica-Worms H. Death by releasing the breaks: CHK1inhibitors as cancer therapeutics. Trends Mol Med 2011; 17: 88–96.

7 Bartek J, Mistrik M, Bartkova J. Thresholds of replication stress signaling in cancerdevelopment and treatment. Nat Struct Mol Biol 2012; 19: 5–7.

8 Bartkova J, Horejsı Z, Koed K, Kramer A, Tort F, Zieger K et al. DNA damageresponse as a candidate anti-cancer barrier in early human tumorigenesis. Nature2005; 434: 864–870.

9 Bartek J, Lukas J, Bartkova J. DNA damage response as an anti-cancer barrier:damage threshold and the concept of ’conditional haploinsufficiency’. Cell Cycle2007; 6: 2344–2347.

10 Bartek J, Lukas J. Chk1 and Chk2 kinases in checkpoint control and cancer. CancerCell 2003; 3: 421–429.

11 Murga M, Campaner S, Lopez-Contreras AJ, Toledo LI, Soria R, Montana MF et al.Exploiting oncogene-induced replicative stress for the selective killing of Myc-driven tumors. Nat Struct Mol Biol 2011; 18: 1331–1335.

12 Liu Q, Guntuku S, Cui XS, Matsuoka S, Cortez D, Tamai K et al. Chk1 is an essentialkinase that is regulated by Atr and required for the G(2)/M DNA damagecheckpoint. Genes Dev 2000; 14: 1448–1459.

13 Lam MH, Liu Q, Elledge SJ, Rosen JM. Chk1 is haploinsufficient for multiplefunctions critical to tumor suppression. Cancer Cell 2004; 6: 45–59.

14 Greenow KR, Clarke AR, Jones RH. Chk1 deficiency in the mouse small intestineresults in p53-independent crypt death and subsequent intestinal compensation.Oncogene 2009; 28: 1443–1453.

15 Zaugg K, Su YW, Reilly PT, Moolani Y, Cheung CC, Hakem R et al. Cross-talkbetween Chk1 and Chk2 in double-mutant thymocytes. Proc Natl Acad Sci USA2007; 104: 3805–3810.

16 Fishler T, Li YY, Wang RH, Kim HS, Sengupta K, Vassilopoulos A et al. Geneticinstability and mammary tumor formation in mice carrying mammary-specificdisruption of Chk1 and p53. Oncogene 2010; 29: 4007–4017.

17 Tho LM, Libertini S, Rampling R, Sansom O, Gillespie DA. Chk1 is essential forchemical carcinogen-induced mouse skin tumorigenesis. Oncogene 2012; 31:1366–1375.

18 Sansom O, Reed K, Hayes A, Ireland H, Brinkmann H, Newton I et al. Loss of Apcin vivo immediately perturbs Wnt signaling, differentiation, and migration. GenesDev 2004; 18: 1385–1390.

19 Shibata H, Toyama K, Shioya H, Ito M, Hirota M, Hasegawa S et al. Rapid colorectaladenoma formation initiated by conditional targeting of the Apc gene. Science1997; 278: 120–123.

20 Ireland H, Kemp R, Houghton C, Howard L, Clarke AR, Sansom OJ et al.Inducible Cre-mediated control of gene expression in the murine gastroin-testinal tract: effect of loss of beta-catenin. Gastroenterology 2004; 126:1236–1246.

21 Sansom OJ, Reed KR, Hayes AJ, Ireland H, Brinkmann H, Newton IP et al. Loss ofApc in vivo immediately perturbs Wnt signaling, differentiation, and migration.Genes Dev 2004; 18: 1385–1390.

22 Merritt AJ, Allen TD, Potten CS, Hickman JA. Apoptosis in small intestinalepithelial from p53-null mice: evidence for a delayed, p53-independentG2/M-associated cell death after gamma-irradiation. Oncogene 1997; 14:2759–2766.

23 Bakkenist CJ, Kastan MB. DNA damage activates ATM through intermolecularautophosphorylation and dimer dissociation. Nature 2003; 421: 499–506.

24 Rogakou EP, Nieves-Neira W, Boon C, Pommier Y, Bonner WM. Initiation of DNAfragmentation during apoptosis induces phosphorylation of H2AX histone atserine 139. J Biol Chem 2000; 275: 9390–9395.

25 Bolderson E, Richard DJ, Zhou BB, Khanna KK. Recent advances in cancer therapytargeting proteins involved in DNA double-strand break repair. Clin Cancer Res2009; 15: 6314–6320.

26 Salhab N, Jones DJ, Bos JL, Kinsella A, Schofield PF. Detection of ras genealterations and ras proteins in colorectal cancer. Dis Colon Rectum 1989; 32:659–664.

27 Soriano P. Generalized lacZ expression with the ROSA26 Cre reporter strain. NatGenet 1999; 21: 70–71.

28 Bjerknes M, Cheng H. Methods for the isolation of intact epithelium from themouse intestine. Anat Rec 1981; 199: 565–574.

Supplementary Information accompanies this paper on the Oncogene website (http://www.nature.com/onc)


8


http://www.nature.com/onc

wnt-driven intestinal tumourigenesis is suppressed by chk1 deficiency but enhanced by conditional...

Documents