Developmental Regulation and Induction of Cytochrome P450 2W1, an Enzyme Expressed in Colon Tumors

Cytochrome P450 2W1 (CYP2W1) is expressed predominantly in colorectal and also in hepatic tumors, whereas the levels are insignificant in the corresponding normal human adult tissues. CYP2W1 has been proposed as an attractive target for colorectal cancer (CRC) therapy by exploiting its ability to activate duocarmycin prodrugs to cytotoxic metabolites. However, its endogenous function, regulation and developmental pattern of expression remain unexplored. Here we report the CYP2W1 developmental expression in the murine and human gastrointestinal tissues. The gene expression in the colon and small intestine commence at early stages of embryonic life and is completely silenced shortly after the birth. Immunohistochemical analysis of human fetal colon revealed that CYP2W1 expression is restricted to the crypt cells. The silencing of CYP2W1 after birth correlates with the increased methylation of CpG-rich regions in both murine and human CYP2W1 genes. Analysis of CYP2W1 expression in the colon adenocarcinoma cell line HCC2998 revealed that the gene expression can be induced by e.g. the antitumor agent imatinib, linoleic acid and its derivatives. The imatinib mediated induction of CYP2W1 suggests an adjuvant therapy to treatment with duocarmycins that thus would involve induction of tumor CYP2W1 levels followed by the CYP2W1 activated duocarmycin prodrugs. Taken together these data strongly support further exploration of CYP2W1 as a specific drug target in CRC.


Introduction
Colorectal cancer (CRC) is an important contributor to cancer mortality and morbidity with limited treatment options for patients with advanced disease. CRC causes half a million deaths annually and the general five years survival rate is around 60% [1]. Currently, surgical resection remains the most curative approach for CRC, but approximately 40% of treated patients will subsequently develop local recurrence or distant metastases [2,3]. Therefore, refinement of the strategies for the treatment of CRC metastases is of crucial importance.
The orphan cytochrome P450 2W1 (CYP2W1) has been found to be expressed preferentially in colorectal tumors while its expression is barely detectable in non-transformed tissues [4]. High CYP2W1 levels were observed in approximately 30% of human CRC specimens [5,6]. Recently significant expression has also been seen in liver cancers with even higher expression in liver metastases than in the parent tumor from the same patient [7]. In both colon and liver cancers the prognosis of survival is worse in patients carrying tumors with higher levels of CYP2W1 expression [5,6,8]. Low expression levels of CYP2W1 are observed also in the other tumors, e.g. adrenal gland, gastric and lung tumors [4,9,10].
Tumor-specific expression of CYP2W1 has been suggested as an attractive target for CRC therapy exploiting either its ability to activate certain prodrugs to cytotoxic metabolites or using CYP2W1 as a new tumor-associated antigen for immunotherapy-based treatment. Thus far, the prodrug monotherapy (PMT) option has been successfully confirmed both in vitro and in a murine xenograft model [11].
In this context, the exploration of the mechanisms that control CYP2W1 expression is of paramount importance as it might lead to the development of a novel combined therapy of CRC that would include a tumor-specific induction of CYP2W1 followed by the treatment with CYP2W1-specific prodrugs, increasing thus the number of patients who would benefit from such approach.
In addition to its expression in colon cancer cells, the CYP2W1 mRNA has been reported in rat fetal colon and in murine embryonic pooled tissues, whereas no significant expression has been reported in adult tissues. However, the developmental expression of human CYP2W1 remains largely unexplored. In CRC tumors, higher expression was associated with increased demethylation of CpG island in the exon 1/intron 1 junction [9]. This implies interesting parallels between the developmental and cancer regulatory mechanisms of CYP2W1 expression that were also found to overlap in many oncofetal genes [12,13].
There have been limited attempts to modulate the intracellular levels of CYP2W1. Previous unpublished study suggested CYP2W1 induction by imatinib in leukaemia cells (GEO dataset: GDS3044) and CYP2W1 was induced in breast cancer cells following the treatment with 5F-203 or GW-610 [14]. However, the extensive analyses of CYP2W1 regulation in colon cancer cells are missing.
The aim of current study was to examine CYP2W1 regulation by (i) studying the murine and human developmental patterns of its constitutive expression and putative epigenetic regulation and (ii) screening of the potential CYP2W1 inducers that could modulate the levels of its endogenous expression.

Human and animal studies
Human adult and fetal tissues were obtained from the NICHD Brain and Tissue Bank for Developmental Disorders at the University of Maryland (Baltimore, MD, USA) after approval of our study by the committee of the University. Adult colon (n = 5) and small intestine (SI, n = 5) control samples were from donors between 19 and 61 years of age with heart/pulmonary failure or accidental death diagnosis. The three colon and five SI human fetal tissues were from different female donors, gestational weeks 18 and 19.
Cyp2w1 murine expression from fetal life to adulthood was analyzed by collecting C57BL/6 WT mouse colon and SI tissue samples from embryonic days 13 (E13), E16, E18, followed by postnatal day 0, day 2, 3, 7 and 28. Adult mice were investigated at 12 weeks of age, 7 and 13 months of age (n = 6 per time point). Mice were anesthetized by Ketamine/Xylazine intraperitoneally followed with CO 2 . The tissues were collected immediately after cervical dislocation and frozen in dry ice. Animal experimentation was approved by the local animal ethical committee in Stockholm, Sweden (permits: N147/11 and N505/11).

CYP2W1 mRNA expression
Total RNA and DNA were extracted using AllPrep DNA/RNA Mini Kit (Qiagen, Hilden, Germany) from 10-50 mg of human and mouse colon, SI or confluent HCC2998 cells. Reverse transcription was performed using the SuperScript III first-strand synthesis (Invitrogen, Rockville, USA). Real-time polymerase chain reaction was carried out using ready-to-use CYP2W1 (Hs00908623_m1, Mm01207203_m1 for human and mice, respectively) and CYP1A1 (Hs00153120_m1) TaqMan Gene Expression Assays (Applied Biosystems, Stockholm, Sweden), according to the manufacturer's instructions. An internal control sample on the plate and housekeeping genes, EIF2B2 (Hs00204540_m1), HMBS (Hs00609293_g1) and PPIA (Hs99999904_m1) for human and Tjp1 (Mm00493699_m1), 18S (Mm03928990_g1) and Gadph (4352339E) for mice were used for normalization according to the 2 -ΔΔCt method [16].
The in silico analysis showed enrichment of murine cyp2w1 gene by CpG sites in several areas (S1 Fig). Selected areas (n = 6) within the cyp2w1 gene and upstream region were amplified with two subsequent PCR reactions using either the Taq DNA polymerase and associated reagents (Thermo Scientific, St. Leon-Rot, Germany) or the KAPA HiFi HotStartUracil Ready-Mix (Kapa Biosystems, Inc., Wilmington, USA). Same protocol was used for DNA samples from human fetal and adult colon tissue and also for the HCC 2998 cells treated with CYP2W1 inducing agents (see Cell cultures section, below). All primers were designed using the online bisulfite primer design software (http://bisearch.enzim.hu/) and optimized to amplify a single PCR product. S1 Table describes PCR conditions and primer sequences used. PCR products were then run on a 1.5% TAE gel, the bands were excised and DNA samples were purified using a QIAquick Gel Extraction kit (Qiagen). Direct Sanger sequencing (Eurofins, Ebersberg, Germany) was performed and percentage of methylated cytosines was evaluated by the peak ratio of cytosine or guanine peak height (forward or reverse primer, respectively), to the peak height sum of cytosine and thymine or guanine and adenine (forward or reverse primer, respectively) using Chromas 2.4.1 (Technelysium, South Brisbane, Australia).

Cell cultures
The human colon adenocarcinoma HCC2998 cells obtained from National Cancer Institute (Frederick, MD, USA) were cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum, 100 U/ml penicillin-100 μg/ml streptomycin, including 2 mM L-glutamine. Passage numbers between 11 and 16 were used throughout the study.
HEK 293T cells were cultured in Dulbecco's modified Eagle's DMEM medium with 10% fetal bovine serum, and 100 U/ml penicillin-100 μg/ml streptomycin. The pCMV6-Entry plasmid containing mouse cyp2w1 cDNA (Origene, Rockville, MD, USA), the previously described cDNA construct of human CYP2W1 [17] or an empty plasmid were transfected using 2 μg Lipofectamine 2000 for 48 h. HepG2 cells at passage 13 were used as a second positive control for human CYP2W1 [4] and were maintained in MEM medium supplemented with 10% fetal bovine serum, 100 U/ml penicillin-100 μg/ml streptomycin, 1 mM sodium pyruvate, and nonessential amino acids. All cell culture media and reagents were obtained from Life Technologies (Stockholm, Sweden).

CYP2W1 activity
The CYP2W1 enzymatic activity was assessed by measuring the metabolism of a CYP2W1-specific non-toxic duocarmycin derivative substrate (ICT2726) as previously described [11]. Five days after reaching the confluent state HCC2998 cells were incubated with ICT2726 (4 μM) for 24 h, after which cells were washed twice with PBS and the reaction was stopped by adding 100μl of ice-cold acetonitrile. After centrifugation at 4°C, 13000 g for 20 min, 30 μl of supernatant was injected into a Luna C18 (2.0x150 mm, 5 μm col HPLC column; Phenomenex, Vaerløse, Denmark), as previously described [17]. The cytotoxic effect of two duocarmycin prodrugs, ICT2705 and ICT2706 [11] was tested on the confluent HCC2998 cells that were incubated for 72 h with the vehicle (0.1% DMSO) or ICT2706 and ICT2705. Cell viability was assessed by the EZ4U assay (Biomedica, Vienna, Austria) according to manufacturer´s instructions.
HCC2998 cells were treated for 48 h with the vehicle (0.1% DMSO or 0.1% ethanol according to the drugs solubility, see S2 Table) or the indicated drugs. The culture medium, including vehicle and investigated drugs was exchanged every 24 h. After cell harvesting, the CYP2W1 mRNA levels were analyzed as described above. All experiments were performed at least twice in at least two replicates per experiment.

Statistics
Cyp2w1 mRNA expression was correlated to the methylation percentage at each CpG site and to the average methylation for each amplicon using Spearman correlation coefficient (R spearman ) for p<0.01. The methylation percentage differences from embryonic (E, E13-E18), early postnatal (PN, D0-D7) to adult ages (D28-13M) were assessed by Kruskal-Wallis in mice and by Mann-Whitney tests between fetal and adult human samples. Two-tailed Student t-test was performed to analyze CYP2W1 fold change in HCC2998 cells. All data were expressed as mean ± SEM and statistical significance was defined as p<0.05 (GraphPad Prism v.5.0; GraphPad Software, La Jolla, CA, USA).

CYP2W1 developmental expression
We have studied the CYP2W1 expression on mRNA and protein levels in pre-and postnatal periods in mice and human colon and SI. Cyp2w1 mRNA appears in the mouse embryonic colon and SI at the gestational day 13 (E13). The expression peaks at E18 and is silenced rapidly after the birth (Fig 1A). The expression curves for both tissues are similar, however, they do not match exactly: mRNA expression starts earlier in SI and sharply decreases already by postnatal day three, whereas corresponding levels in colon tissues were stable until post-natal day 7. At four weeks of age (D28), the tissue-specific Cyp2w1 mRNA expression was silenced in both SI and colon tissues. As expected, no Cyp2w1 transcripts were detected in the liver at any time point (Fig 1A). The Cyp2w1 protein data as shown on Fig 1B confirm the mRNA expression pattern in colon and SI.
The CYP2W1 mRNA was detected in human fetal colon and SI at the gestational weeks 18-19 but not in any of the adult samples analyzed (Fig 2A). Expression in SI was stable over both studied fetal ages while in colon it started to decline from week 19. Consistent with the murine samples the human protein expression followed the CYP2W1 transcript's pattern (Fig 2B).

Immunohistochemistry analysis
Immunohistochemical analysis using antibodies from two different sources was used to identify the cells types expressing CYP2W1 in the human fetal colon samples from gestational week 19. CYP2W1 expression was observed in the crypt cells (Fig 3A and 3B) with both antibodies confirming this pattern (Fig 3A). No signals were observed in the cells constituting lamina propria, muscularis mucosae or submucosa (Fig 3A). A typical endoplasmic reticulum expression of CYP2W1 in the crypt cells was confirmed by its co-localization with the ER marker, an abundant molecular chaperone, GRP78 (Fig 3B).

Epigenetic regulation of CYP2W1
It has been suggested that the expression of CYP2W1 in colon cancer cells requires the demethylation of CpG island in the exon1/intron 1 junction [9]. We hypothesized that a similar mechanism may govern CYP2W1 gene expression during the development.
Scanning of the regulatory and gene coding regions of murine Cyp2w1 gene could not locate CpG islands that are characteristic for the human gene. However, recent analyses attached an important role also for CpG sites outside the canonical CpG islands [21]. Six of such areas (amplicons) with the relatively high CpG density from -13 kb to the Cyp2w1 gene were selected for further analyses (Fig 4A, upper panel). The percentage of methylated cytosines in these amplicons was evaluated in the genomic DNA isolated from mouse colon tissue. Fig 4A (  and, therefore the correlations were considered as non-significant. Such correlation was not found in the other amplicons and the distal promoter was consistently demethylated in all samples (data not shown).
In addition, the statistically significant differences in murine Cyp2w1 methylation levels between the different age groups (from adult to fetal age) were observed at a number of CpG sites. The major differences were seen in the close promoter, exon 1/intron 1, intron 1 and exon 5. Hypomethylation of CpGs in fetal samples was observed at one site in the promoter (-254 bp) and in exon 5 (3311, 3324, 3356, 3380 bp, amplicon 6) (Fig 4B).
The methylation state of the human CYP2W1 gene in pre-and post-natal human colon samples was analyzed using the same methodology. We chose the exon 1/intron 1 region, demethylation of which was previously shown to correlate with the expression of CYP2W1 in colon tumors. The methylation of six different CpG sites in this region was found to be significantly higher in adult samples as compared to the fetal colon (Fig 5). This is consistent with the methylation state of the gene as induced by CRC [9].

Drug-induced regulation of CYP2W1 expression
Increased expression of CYP2W1 can be beneficial for the specific prodrug based treatment of cancer (see Introduction). However, an appropriate colon cancer cell model with a constitutive CYP2W1 expression suitable for the in vitro regulation studies has hitherto not been identified. Scanning of a number of colon adenocarcinoma cell lines (results not shown) led to identification of the HCC2998 cell line (National Cancer Institute, Frederick, USA) with a significant level of CYP2W1 expression and activity. The HCC2998 cells express similar or even higher amounts of CYP2W1 than HepG2 cells (Fig 6A), a hepatoma cell line with constitutive expression of CYP2W1 [4]. Interestingly, the CYP2W1 mRNA and protein levels in these cells were  substantially increasing upon prolonged cultivation for 4-25 days after the cells reached the confluent state (Fig 6A and 6B). Therefore, all subsequent analyses were done on confluent cells, which were then treated with drugs for additional five days.
CYP2W1 expressed in HCC2998 cells appears to be catalytically active as it was found to metabolize the CYP2W1 specific substrate, non-toxic duocarmycin derivative ICT2726 [11]. Analyses of these cells revealed formation of ICT2726 metabolite with the same retention time as in the positive control, HEK 293T cells overexpressing CYP2W1 (Fig 6C).
Moreover, incubation of HCC2998 cells with the duocarmycin prodrugs ICT2706 and ICT2705 that were shown to be converted by CYP2W1 into potent cytotoxins [11] led to a significant decrease of cell viability (Fig 6D).
Using the HCC2998 cell model, we examined a series of biologically active substances for the induction of CYP2W1 expression. Treatment of the cells with DAC, imatinib, LA and the two conjugated (9Z11E-LA, 11E12Z-LA) significantly increased the CYP2W1 mRNA levels (p<0.05) as compared to vehicle treated cells (Fig 7). In order to test whether this effect might be determined by the drug mediated demethylation of CYP2W1 we tested the methylation state of the same CpG sites in the exon 1/intron 1 region of the human CYP2W1 gene (CpG positions 70,76,101,111,119,123,129,133,136,155,162,173,203,217) in the cells treated with indicated drugs or vehicle. The percentage of methylated cytosines in all of these CpG sites was found unaltered between the drug treated and control cells (data not shown).
Induction of CYP1A1 was also observed by the previously reported CYP1A1 inducers, citco, TCDD and DAC [22][23][24] (S2 Fig) validating thus our cell induction model. No induction of CYP2W1 or CYP1A1 was observed by other drugs used in this study (see Methods) including previously reported CYP1A1 inducer GW-610 [14].

Discussion
CYP2W1 is one of the enigmatic P450 isoforms characterized by the barely detectable expression in adult organisms and unknown endogenous function in development. At the same time cancer-specific expression and discovery of prodrugs metabolized by CYP2W1 to cytotoxins makes it an interesting target for cancer therapy and therefore a subject for intensive studies on its regulation and function in ontogeny and neoplasia. Here we provide with information on the developmental pattern of CYP2W1 expression in mice and humans and the epigenetic regulation, which characterizes CYP2W1 as a novel oncofetal gene. In addition, screening of a broad spectrum of biologically active molecules yielded a list of potential CYP2W1 inducers that might be instrumental for its up-regulation in cancer tissues, increasing thus the efficacy of prodrug therapy. Embryonic expression of Cyp2w1 in rodents has been suggested previously, albeit only at mRNA level. The data were somewhat inconclusive due to a large standard deviations reported in the rat study [4] and the use of pooled samples of different murine tissues [25]. Here, we show in detail the expression of Cyp2w1 mRNA and protein in fetal murine colon and in SI. It should be noted that the developmental expression curves in these tissues do not overlap: Cyp2w1 mRNA is detected earlier in SI than in colon. This might indicate synchronization with the development of the whole GI tract. However the colon expression remains at a plateau for a longer period of time. In both tissues, gene silencing occurred within the first 28 days of post-natal life.
Information on the human embryonic expression of CYP2W1 is scarce. Using a human fetal mRNA MTC panel lacking GI tissues, Choudhary et al. found qualitatively a transient CYP2W1 mRNA expression in lungs, liver, skeletal muscle and kidney at gestational week 30 [25]. Here, we could describe the CYP2W1 mRNA and protein levels of CYP2W1 in human fetal colon and in SI, and using the immunohistochemical analysis, show the regio-selective distribution of CYP2W1 in the crypt cells of embryonic human colon, reminiscent of the CYP2W1 expression in colon cancer cells [7].
Tumorigenesis and embryogenesis are suggested to share common pathways [12,13]. A number of CRC biomarkers and cancer therapy target genes, such as carcinoembryonic antigen (CEA) [26], Coding Region Determinant-Binding Protein [27] and Cripto-1 [28], Trophoblast glycoprotein (5T4) [29] are shown to be regulated in such manner, i.e. expressed in tumors, whereas their expression in normal tissues is mainly limited to pre-natal period. Moreover, an overlap between murine gut developmentally regulated genes and those reactivated in human CRC tumors was reported previously, including for instance the insulin-like growth factor II (IGF-II), which is known to enhance tumor growth and suppress apoptosis [30].
Among the P450s, similar developmentally regulated expression pattern was reported for CYP1B1, which is involved in eye development [31] and is also expressed in various primary tumors and metastases, while its expression is generally undetectable in corresponding normal cells [32,33]. Likewise, CYP1A1 gene is highly expressed in several tumors but much less in corresponding normal cells [34].
One important mechanism controlling the expression of oncofetal genes is epigenetic regulation, the CpG methylation in particular. Here we present detailed data on the methylation state of murine and human CYP2W1 in GI tissues from fetus to adult age. The developmental shift from hypomethylation to hypermethylation in the exon 1/intron 1 CpG island of human CYP2W1 correlates with the gene expression levels of CYP2W1 and is perfectly consistent with similar changes found in normal versus colon tumor tissues [9]. Interestingly, such methylation shifts in the promoters of many intestine-specific genes, including Cyp2w1 were recently related to intestine maturation in preterm pig [35].
The methylation state of genomic regions outside of CpG islands and the canonical promoter sites as shown in the mouse Cyp2w1 gene (Fig 4A) is receiving increased attention [21]. Recent findings revealed positive correlation between intragenic DNA methylation level and gene expression [21,[36][37][38].
In the whole-genome DNA methylation and gene expression study of leukemia Kulis et al. showed significant both positive and negative correlation of methylation status and gene expression in the gene body (intragenic) of approximately 900 genes with absence of methylation changes in the promoters [21]. Similar studies on cell cultures demonstrated that the lowest level of intragenic methylation was correlated with both the lowest and highest gene expression [39]. Such intragenic methylation is suggested to regulate gene expression by promoting alternative promoters, enhancer activation, transcript elongation or regulating intragenic non-coding RNA [21].
Transcriptional regulation of gene expression is commonly studied in the cultured cell lines with the constitutive expression of the gene of interest. The colon adenocarcinoma cell line HCC2998 used in this study seems to be an adequate model in that respect as it constitutively produces relatively high amounts of functional CYP2W1, the CYP2W1 levels are also increasing with the enhanced cell density. Similar density-dependent upregulation of CYP3A4 was observed in post-confluent human hepatoma cells [40,41] and in adenocarcinoma cell line Caco2-TC7 [42], apparently due to cell adhesion triggered re-differentiation of these cells. The HCC2998 cell model was successfully used to test whether the CYP2W1 expression can be exogenously induced, which might be useful for augmenting the effects of CYP2W1-dependent cancer therapy [11]. Among the tested substances, imatinib, DAC and LA were found to be the most potent inducers of CYP2W1 expression. The highest induction was found in cells treated with imatinib, which was previously known as a P450 inhibitor [43]. However, an unpublished transcriptomic analysis of leukemia cells after imatinib treatment (Geo Dataset: GDS3044) also revealed induction of CYP2W1 in line with our data. DAC-mediated CYP2W1 induction has previously been reported in Caco2-TC7 cells [4]. DAC is a commonly used demethylating agent and therefore similar mechanism of CYP2W1 induction cannot be ruled out.
CYP2W1-mediated metabolism of certain lipids including arachidonic acid and lysolecithin and their stereoisomers was reported earlier [4,20]. In our study CYP2W1 was significantly induced by LA and the 2 conjugated-LA, however, no stereoselectivity was found.
Is has previously been reported that the in vitro anti-cancer effects of 9Z11E-LA and 10E12Z-LA isomers [44,45] involve numerous pathways, including PPARγ [46], CAR [47] or ERα pathway [48]. In our model, the CAR and PPARγ agonists citco and ciglitazone, respectively did not show any induction. Therefore, in HCC2998 cells, the CYP2W1 induction may involve other, yet unknown mechanisms. These, most probably do not include epigenetic regulation, the mechanism that governs the developmental expression of CYP2W1, as neither of the CYP2W1 inducing drugs had any effect on the methylation state of CpG sites in the critical exon 1/intron 1 region of the CYP2W1 gene. It is likely that the CYP2W1 expression is mediated by the nuclear proteins that are activated in a phosphorylation cascade alteration induced by the treatment of the tyrosine kinase inhibitor imatinib. The details of this signal transduction mechanisms remains to be elucidated. It can be also speculated that the endogenous function of CYP2W1 might be connected with the lipid metabolism.
It has been previously suggested that AhR may mediate the CYP2W1 induction [4]. However, despite the strong induction of CYP1A1 by TCDD, a potent AhR ligand and CYP1A1 inducer [49], CYP2W1 expression remained unaffected (Figs 7 and S2). Interestingly, in addition to TCDD, the CYP1A1 expression was upregulated also by imatinib, which is to our knowledge the first report of such induction (S2 Fig). Taken together with the CYP1A1 involvement in the metabolism of imatinib [50] and in activation of other antitumor agents (Phortress), this finding might be of a clinical relevance and should be taken in account upon combined treatment with imatinib and other CYP1A1 metabolized drugs.
In conclusion we show the developmental expression of CYP2W1 in the GI tract, to a great extent mediated by epigenetic modifications, its specific location, and regulation by anticancer drugs which can be considered as an adjuvant therapy of colon cancer metastases and hepatic cancers in the future. Additional information is required regarding the mechanisms of specific regulation of the enzyme in primary colon cancer tumors.