Characterization of hERG1 channel role in mouse colorectal carcinogenesis

The human ether-à-go-go-related gene (hERG)1 K+ channel is upregulated in human colorectal cancer cells and primary samples. In this study, we examined the role of hERG1 in colorectal carcinogenesis using two mouse models: adenomatous polyposis coli (Apcmin/+) and azoxymethane (AOM)-treated mice. Colonic polyps of Apcmin/+ mice overexpressed mERG1 and their formation was reverted by the hERG1 blocker E4031. AOM was applied to either hERG1-transgenic (TG) mice, which overexpress hERG1 in the mucosa of the large intestine, or wild-type mice. A significant increase of both mucin-depleted foci and polyps in the colon of hERG1-TG mice was detected. Both the intestine of TG mice and colonic polyps of Apcmin/+ showed an upregulation of phospho-Protein Kinase B (pAkt)/vascular endothelial growth factor (VEGF-A) and an increased angiogenesis, which were reverted by treatment with E4031. On the whole, this article assigns a relevant role to hERG1 in the process of in vivo colorectal carcinogenesis.


Introduction
Colorectal cancer (CRC) is one of the most common cancers among men and women in developed and industrialized countries [1]. Some of these tumors are hereditary [2,3], while the majority are sporadic and linked to environmental factors [4].
In the last few decades, CRC progression was deciphered as an ordered multistep process, which depends on a series of genetic changes, involving the lack of tumor-suppressor and DNA-repair genes accompanied by the activation of oncogenes [5]. Each of these genetic aberrations leads to the accomplishment of specific morphological steps which lead the normal colonic mucosa to a true invasive carcinoma, through specific, ordered lesions [6].
Different mouse models, either genetically modified or chemically induced, have been generated to recapitulate the adenoma-carcinoma sequence that occurs in CRC, with the aim of better defining the molecular basis of the disease and identifying novel drugs for treatment. The genetic mouse model mostly used for this purpose is represented by adenomatous polyposis coli (Apc min/+ ) mice, which carry a dominant heterozygous nonsense mutation at codon 850 of the mouse homologue of the human APC gene. Apc min/+ mice develop spontaneous multiple adenomas throughout the intestinal tract, mainly in the small intestine [7,8].
Among the genes whose expression is altered during the carcinogenetic process, those encoding ion channel and transporters are acquiring a relevant role in the last few years [13]. In particular, K + channels of the ethera-go-go (EAG) family, mainly human ether-a-go-go-related gene (hERG1) [14] and EAG-1 [15], were found to be overexpressed in several types of human cancers [13,16], including CRC [17][18][19][20]. Moreover, the genes encoding either channels were detected in the crypts of murine colon, after carcinogen treatment [18], and an upregulation of oncogenic K + ion channels (BK, Elk1, and EAG) was detected in the colon of Apc min/+ mice [21].
We tested the in vivo relevance of hERG1 channels during colorectal cancerogenesis by studying either Apc min/+ mice or AOM-treated mice. For this purpose, we produced hERG1-transgenic (TG) mice, which overexpress the hERG1 gene in the intestinal mucosa.

Materials and Methods
Mouse strains and production of TG mice Fabp-Cre mice were purchased from National Cancer Institute -Mouse Models of Human Cancers Consortium (NCI-MMHCC) [22] and Apc min/+ mice were obtained from The Jackson Laboratory (stock number: 002020).
The 10-kb vector-free XbaI DNA fragment was microinjected into the male pronucleus of fertilized eggs from FVB mice at LIGEMA, University of Florence, Italy, following standard procedures. TG mice were maintained in a heterozygous state in FVB background.
Animals were housed in plastic cages with a wire mesh providing isolation from the hygienic bed and were kept in temperature-, air-, and light-controlled conditions. They received food and water ad libitum. All experiments involving mice were performed in accordance with the criteria outlined in the Guide for the Care and Use of Laboratory Animals.

AOM treatment
Two-month old mice, 12 TG, and six controls, maintained in a C57BL6/FVB mixed background, received intraperitoneal (IP) injections of AOM (10 mg/kg body weight) once a week for 6 weeks; in addition, three controls and six TG mice were treated with physiologic solution. Three months after the last injection, all animals were killed to evaluate tumorigenesis. The entire gastrointestinal tract was removed for dissection and flushed with phosphate buffered saline (PBS) to remove intestinal content. The intestine was opened longitudinally and washed extensively with PBS. Colon-rectum was fixed in 4% formaldehyde made in PBS for 24 h, after which the tissues were stained with methylene blue (0.1% for 10 min). The number of polyps was determined under a dissecting microscope (209 power field).
Aberrant crypt foci (ACF) were determined according to Bird [23]. The same methylene blue-stained colons were then restained with high-iron diamine Alcian blue (HID-AB), to identify MDF as described in Ref. [24]. MDF and ACF were identified under a microscope (4009 magnification).

E4031 treatment of Apc min/+ and TG mice
Apc min/+ mice received daily for 3 months IP injections of 20 mg/kg E4031 (TOCRIS, Bristol, U.K.) dissolved in sterile water; control Apc min/+ mice received buffered saline only. After 3 months, animals were killed by cervical dislocation. The entire gastrointestinal tract was removed for dissection and flushed with PBS to remove intestinal content. The colon-rectum was opened longitudinally and washed extensively with PBS, fixed in 4% buffered formaldehyde for 24 h and then stained with methylene blue. The number of polyps was determined under a dissecting microscope (209 power field).
TG mice received daily for 14 days IP injections of 20 mg/kg E4031 (TOCRIS) dissolved in sterile water; control mice received buffered saline only. After 14 days, animals were killed by cervical dislocation. The entire gastrointestinal tract was removed for dissection and flushed with PBS to remove intestinal content. The organ was opened longitudinally and washed extensively with PBS, fixed in 4% buffered formaldehyde for 24 h and embedded in paraffin. Tissue sections (7 lm) were cut from blocks using a microtome (Leica RM2125/ RM2125RT, Nussloch, Germany). Immunohistochemistry (IHC) using antivascular endothelial growth factor (VEGF-A) antibody was performed to evaluate differences between control and treated TG mice.
RNA extraction, reverse transcription, and real-time quantitative PCR RNA was extracted from different tissues of wild-type (WT), TG, and Apc min/+ mice using TRIzol reagent (Invitrogen, Groningen, the Netherlands) according to the manufacturer's protocol. cDNA was obtained from 1 to 2 lg of RNA using 200 U reverse transcriptase Super-Script II (Invitrogen), plus 500 lmol/L each of deoxyribonucleotide triphosphate (dNTP) and 15 ng/lL of random primers, in a 20 lL final reaction volume, for 50 min at 42°C and 15 min at 70°C. cDNA synthesis was monitored by PCR with b-actin primers.
Immunostaining was carried out using a commercially available kit (PicTure Plus kit; Zymed, San Francisco, CA). After extensively washing with PBS, color was developed by incubating the slides with the 3,3′-diamino-benzidine chromogen solution for 2-5 min or until acceptable color intensity had been reached. Slides were then counterstained with Mayer's hematoxylin and mounted using Entellus mounting medium. Images were acquired on a Leica DM 4000B microscope with a Leica DFC 320 camera using Leica Win software (Leica Microsystems; Milan, Italy).

Statistical analysis
Data obtained from AOM-treated mice were reported as mean AE standard error of the mean (SEM) and analyzed by Mann-Whitney U test. A P-value <0.01 (*) was considered statistically significant.
Data obtained from vessel and total vascular area count of WT, TG, and Apc min/+ mice were reported as mean AE SEM and analyzed by Mann-Whitney U test. P-values <0.05 and <0.01 were considered statistically significant.
Data obtained from VEGF-A and pAkt expressions were analyzed by Mann-Whitney U test. P-values <0.05 and <0.01 were considered statistically significant.
Data obtained from RT-qPCR experiments to detect mERG1 expression in TG and WT mice were reported as mean AE SEM and analyzed by Mann-Whitney U test (n: four WT mice and five TG mice). A P-value <0.05 (*) was considered statistically significant.

Role of hERG1 in colonic polyp development of Apc min/+ mice
We studied the role of hERG1 channels in CRC tumorigenesis, using Apc min/+ mice as a model. On the basis of previous observations [26] indicating that the transcript encoding the murine homologue of the hERG1 gene, mERG1, was expressed in the colon of Apc min/+ mice, we first determined mERG1 expression levels in various tracts of the intestine of such mice, by RT-qPCR.
We found that the colon and rectum of Apc min/+ mice showed an increase in mERG1 expression, compared to that in WT mice. Such increase was more evident in colonic and rectal polyps, which spontaneously develop in these mice. Interestingly, mERG1 expression increased along with the size of polyps (Fig. 1A). The mERG1 protein was also detected by IHC in colonic polyps of Apc min/+ mice (Fig. 1B). Polyps displayed a high mERG1 expression in adenomatous epithelial cells (see arrows in When Apc min/+ mice were treated with the specific hERG1 blocker E4031, daily for 3 months, such long-term mERG1 current inhibition [28][29][30][31][32], that, in agreement Apc min/+ mice and in colonic and rectal polyps developed in Apc min/+ mice after a normalization for mouse myh11, characteristic of myofibroblasts and smooth muscle cells, to detect the only mERG1 epithelial expression [27]. Distal SI, distal small intestine; C, colon; R, rectum; C polyp, colonic polyp; R polyp, rectal polyp. Data relative to colon and rectum derived from two different experiments, each carried out in triplicate, are reported as the mean AE SEM and were calibrated to the expression levels determined in the rectum of WT mice. Data relative to distal small intestine and polyps derived from a single experiment, carried out in triplicate, are reported as the mean, and were calibrated to the expression levels determined in the rectum of WT mice. (B) mERG1 expression in control (WT) and Apc min/+ polyps was evaluated by IHC. An anti-hERG1 monoclonal antibody was used as detailed in Materials and Methods. Upper panels: 509 magnification, bar: 200 lm; lower panels: 4009 magnification, bar: 20 lm. (C) The number of colonic polyps obtained after E4031 treatment of Apc min/+ mice. Four 1-month-old Apc min/+ mice received daily IP injections of E-4031 for 3 months, while two Apc min/+ mice received buffered saline only. After death, the number of polyps that developed in colon of Apc min/+ mice was determined under a dissecting microscope (209 power field). Data were expressed as mean AE SEM.
with a previous paper [33], did not lead to drug-induced torsades de pointes in E4031-treated mice, produced an impairment in colonic lesion development (Fig. 1C). No effects on the number of polyps in the small intestine were observed. Consistently, no overexpression of the mERG1 transcript was detected in the small intestine of Apc min/+ compared to WT mice (Fig. 1A, left most bar).

Effects of AOM treatment in hERG1 TG mice
We then investigated the possible role of hERG1 in a chemically induced mouse model of CRC, utilizing AOM as carcinogen, and treating either WT or genetically modified mice, hERG1-TG mice, which overexpress the hERG1 gene in the intestinal mucosa. The TG mouse model was developed by us, and the procedure we adopted is detailed in the supplementary information section. Briefly, the hERG1 cDNA, tagged with the myc epitope and a poly-histidine (His) flag at the protein C-terminal, was put under the control of the human b-actin minimal promoter, with an intercalated floxed reporter EGFP gene, which should block transgene transcription. However, hERG1-EGFP Floxed mice expressed the hERG1 transcript both in the colon ( Fig. 2A, white bars) and in the liver (Fig. S1D), even in the absence of Cremediated recombination (compare white and black bars in Fig. 2A). Hence, a transcriptional readthrough phenomenon occurred, and no further significant increase in hERG1 expression was triggered by Cre, for example after mating hERG1-EGFP Floxed mice with Fabp-Cre mice. The latter mice express the Cre recombinase under the control of the Fabp promoter, hence in the whole digestive tract [22]. Therefore, hERG1-EGFP Floxed mice, due to the transcriptional control exerted by the b-actin promoter, can be considered to overexpress the hERG1 transcript ubiquitously. Both the hERG1 transcript ( Fig. 2A) and the hERG1 protein (Fig. 2B) were overexpressed in TG mice (either hERG1-EGFP Floxed or hERG1-EGFP Floxed -Cre) belonging to different TG lines (801, 883, and 886), compared to WT mice. In TG mice, hERG1 expression was strongly detectable in colonic epithelial cells and not limited to the stroma and myofibroblasts, as in WT mice. A slight, significant difference in mERG1 expression in TG compared to WT mice (1.88 AE 0.2 mERG1 expression in TG vs. 1.16 AE 0.1 mERG1 expression in WT; P = 0.03, Mann-Whitney U test) was detected, even if such mERG1 expression increment in TG mice resulted to be very minimal when compared to the artificial hERG1 expression. The generated TG mice did not show any apparent phenotype, even at old ages, and presented a normal life span.
Both TG and WT mice were treated with AOM (or physiologic saline), according to the schedule in Figure 3A, and the occurrence of colonic lesions was analyzed 3 months after the last injection. The macroscopic inspection of the large intestine of treated mice revealed, in the colon of AOM-treated TG mice, the presence of polyps that was, on the contrary, only barely detectable in WT mice (Fig. 3B). No lesions were observed in the large intestine of mice injected with physiologic saline. After staining the large intestine with methylene blue, the number of carcinogen-induced ACF, and after restaining with HID-AB, that of MDF was determined. A statistically significant increase in the number of MDF lesions in TG mice compared to WT mice was detected (Fig. 3B). The increase in MDF lesions paralleled the increased number of polyps, as evidenced by macroscopic inspection. No significant difference was detected in the number of ACF between TG and WT mice.

TG and Apc min/+ mice overexpress pAkt and VEGF-A in the epithelial lining of the large intestine
Finally, we tried to decipher whether a common molecular mechanism could underline the effect of hERG1 overexpression in the process of colorectal carcinogenesis, as evidenced in either the genetic (Apc min/+ ) or chemical (AOM treated) mouse model. In different types of cancers, including CRC [20,34,35], hERG1 is functionally linked to the pathway that promotes the secretion of the VEGF-A, hence contributing to tumor angiogenesis [36].
On the basis of this assumption, we analyzed the expression of both pAkt (i.e., the kinase which regulates VEGF expression [37]) and VEGF-A in the large intestine of TG and in the mERG1-expressing polyps of Apc min/+ mice. A higher expression of both pAkt (Fig. 4A) and VEGF-A (Fig. 4B) was detected in the proximal colon and rectum of TG compared to WT mice. Similarly and consistent with previous reports [38,39], a clear pAkt (Fig. 4E) and VEGF-A (Fig. 4F) immunostaining was detected in the polyps of Apc min/+ mice with significantly higher levels compared with control mice (Table 1). In both TG mice and Apc min/+ polyps, VEGF-A displayed a peculiar expression pattern, different from that observed in control mice. In fact, VEGF-A expression was not limited to the stroma, but was significantly present in the cells of the epithelial lining. The increased expression of VEGF-A in TG mice and Apc min/+ polyps was accompanied by a significant increase in angiogenesis, evaluated as microvessel density and total vascular area measured after staining with an anti-CD34 antibody (Fig. 4C and Table 2). Control mice showed smaller vessels mainly localized in the muscolaris propria, while TG mice and Apc min/+ polyps were characterized by larger vessels with a less ordinate distribution.
Finally, we verified whether the upregulation of VEGF-A in TG mice was directly linked to a higher hERG1 activity and, hence, could be reverted by treatment with the hERG1 blocker E4031. Indeed, treatment of TG mice with E4031 for 2 weeks, led to a significant decrease in VEGF-A staining (Fig. 4D). This indicates that hERG1 channels are not only overexpressed in TG mice, and drive VEGF-A secretion and an increased angiogenesis but they are active and their activity is more or less directly responsible for the VEGF-A-enhanced production observed in these mice. A B Figure 3. Effect of AOM treatment in hERG1 TG mice. (A) Outline of AOM treatment: six control mice and 12 TG mice, maintained in a C57Bl6/ FVB mixed background, received, at 2 months after birth, IP injections of AOM (10 mg/kg body weight) once a week for 6 weeks and were killed 3 months after the last injection. (B) The number of ACF, MDF, and polyps that developed in control (white bars) and TG (black bars)-treated mice was determined. Data were expressed as mean AE SEM. Statistical analysis was conducted using the Mann-Whitney U test (*significantly different with a P-value of <0.01).

Discussion
Substantial evidence indicates that cancer can be partially attributed to ion channel malfunction. Numerous studies included hERG1 in the list of ion channels mis/overexpressed in cancer cells, where it plays the role of regulator of tumor cell proliferation and progression [40,41]. In this article, we analyzed the role of hERG1 in colorectal carcinogenesis in vivo, using either genetic (Apc min/+ mice) or chemical (AOM treated) models of CRC. In both models, we found a relevant role of hERG1 channels, which could be traced back to a hERG1-dependent control of angiogenesis.
Colonic and rectal polyps, which spontaneously develop in Apc min/+ mice, showed an evident overexpression of the murine homologue of hERG1, mERG1 and the long-term treatment of Apc min/+ mice with the specific hERG1 blocker, E-4031, suppressed polyp formation in the large intestine. It is known that the main function of Apc is to degrade cytosolic levels of b-catenin, whose dysregulation is considered a major cause of tumor development. As previous studies from Carlos Munoz's laboratory showed that b-catenin increased the hERG1 protein levels within the oocyte cell membrane [42], the increased hERG1 channel activity detected in the polyps of Apc min/+ mice could be attributed to the overexpression  of b-catenin, widely described in this animal model [43]. It is worth noting that, at difference from what happens in the small intestine, the loss of function of Apc is not sufficient per se to trigger the development of tumors in the colon, where adjunctive genetic events are required for the transition from microadenomas to macroscopic tumors to be accomplished [44]. Our data could suggest considering hERG1 as one factor which cooperates with Apc loss to trigger colorectal tumor progression. This conclusion is further supported by data obtained in carcinogen-treated mice. In this case, in order to analyze the role of hERG1, we generated a TG mouse model overexpressing hERG1. On the bases of the construct we used, we expected to obtain a conditional hERG1expressing mouse. On the contrary, due to a readthrough phenomenon, the mice we generated showed a ubiquitous expression of hERG1. Even when hERG1 was expressed at high levels, the hERG1-TG mice did not show any overt phenotype and presented a normal life span. Hence, hERG1 overexpression per se is not life-threatening and does not induce tumor development. Although hERG1-TG mice did not develop spontaneous tumor, they displayed an accelerated process of tumorigenesis, when treated with AOM, as witnessed by an increased number of preneoplastic lesions (mainly MDF) and polyps in the colon. It is worth noting that MDF, that is dysplastic lesions, characterized by a defective mucin production, are considered precursors of CRC both in humans [8,45,46] and in experimental models [24]. Our finding that the concomitant hERG1 and, probably transgene induced, mERG1 overexpression in the large intestine increases the number of AOM-induced MDF and polyps, strongly indicates that an upregulation of hERG1 accelerates the process of colorectal carcinogenesis, further stressing the role of hERG1 as a progression gene in CRC.
On the whole, the hERG1 gene can be considered a "tumor progression" gene, as it strongly cooperates with genetic (loss of the tumor-suppressor gene Apc) or environmental (chemical carcinogen) factors in triggering CRC progression.
the signalling pathways, which regulate VEGF-A secretion and neoangiogenesis. In CRC cells, hERG1 channels regulate pAkt, through the formation, on the plasma membrane, of a macromolecular complex between hERG1, the b1 integrin, and the p85 subunit of phosphatidyl inositol-3-kinase, which leads to the activation of Akt (O. Crociani et al., pers. comm.). Consistently, we found an upregulation of pAkt and VEGF-A expression in both Apc min/+ polyps and hERG1-TG mice. In the latter, the increased VEGF-A expression causes an increased angiogenesis, which was reverted by blocking hERG1 with a specific blocker. Hence, VEGF-A expression in vivo depends on hERG1 activity. Therefore, when hERG1 is aberrantly overexpressed, VEGF-A secretion and angiogenesis are concomitantly upregulated; through the increased angiogenesis, hERG1 activity may influence tumor development and progression. Taken together, data reported in this article indicate a significant role of hERG1 in colorectal carcinogenesis in vivo, confirming indications, obtained in the human setting, that an early overexpression of the hERG1 gene marks those precancerous lesions of the upper gastrointestinal tract which undergo malignant progression [47]. Moreover, data here provided further stress the inclusion of hERG1 blockers in the treatment of CRC.