Report

Mice Lacking a Myc Enhancer That Includes Human SNP rs6983267 Are Resistant to Intestinal Tumors

See allHide authors and affiliations

Science  07 Dec 2012:
Vol. 338, Issue 6112, pp. 1360-1363
DOI: 10.1126/science.1228606

Abstract

Multiple cancer-associated single-nucleotide polymorphisms (SNPs) have been mapped to conserved sequences within a 500-kilobase region upstream of the MYC oncogene on human chromosome 8q24. These SNPs may affect cancer development through altered regulation of MYC expression, but this hypothesis has been difficult to confirm. We generated mice deficient in Myc-335, a putative MYC regulatory element that contains rs6983267, a SNP accounting for more human cancer-related morbidity than any other genetic variant or mutation. In Myc-335 null mice, Myc transcripts were expressed in the intestinal crypts in a pattern similar to that in wild-type mice but at modestly reduced levels. The mutant mice displayed no overt phenotype but were markedly resistant to intestinal tumorigenesis induced by the APCmin mutation. These results establish that a cancer-associated SNP identified in human genome-wide association studies has a functional effect in vivo.

Genome-wide association (GWA) studies have identified polymorphisms upstream of the MYC gene that are associated with increased risk for multiple different cancer types [e.g., breast (1), bladder (2), prostate (3), and colon (4)]. Several lines of evidence suggest that the identified polymorphisms affect function or expression of the MYC oncogene. Isolated elements containing the polymorphic sequences have enhancer activity in transgenic mice (5, 6), and chromatin conformation-capture assays suggest that in cultured cells the elements are located close to the MYC promoter sequences (79) despite being more than 300 kb away in linear sequence. Moreover, it has been shown that the cell line DLD-1 that is heterozygous for the polymorphism rs6983267 differentially expresses the MYC alleles (9), suggesting that this polymorphism affects the level of expression of MYC. The single-nucleotide polymorphism (SNP) rs6983267 affects binding of the transcription factor TCF7L2 that is centrally important in colorectal tumorigenesis (5, 10), providing a plausible mechanism of action for this polymorphism. Despite all these lines of evidence, direct analyses comparing MYC expression between individuals with different genotypes have been inconclusive or negative (5, 10).

It is possible that the effect of the polymorphisms on MYC expression has not been robustly detected because single-nucleotide substitutions are expected to alter binding of only one transcription factor out of many that bind to a given regulatory element. Thus, the polymorphisms are expected to have only a minor impact on normal MYC expression. However, even when a polymorphism has minor impact, the activity of the element containing it could be required for tumorigenesis.

Because the region containing rs6983267 is conserved in mice, we used a mouse genetic model to evaluate the functional role of the SNP in Myc expression and tumorigenesis. However, converting the risk allele that mice normally carry to the protective allele would be expected to mimic the effect of the SNP in humans, potentially requiring sample sizes that are similar to those used in human GWA studies (hundreds to thousands of mice). To yield an allele with a stronger effect, we instead generated a gene-targeted mouse in which the entire regulatory element containing rs6983267 (Myc-335 hereafter) can be conditionally deleted. This allele allowed us also to test the possibility that the Myc-335 element would be required for tumorigenesis.

Mice carrying the Myc-335 conditional allele (Myc-335 cKO) were generated by flanking with loxP sites a 1740-bp region that lies 335 kb upstream of the Myc transcription start site. The Myc-335 cKO mice were crossed to a deletor mouse strain to generate the Myc-335 null (Myc-335) allele (Fig. 1A). Deletion of the Myc-335 element was confirmed by polymerase chain reaction (PCR) analysis of the genomic DNA (Fig. 1A), and the mice were further intercrossed to generate Myc-335−/− mice.

Fig. 1

Generation of the Myc-335 null allele. (A) Schematic representation of the targeting strategy. E depicts the EcoRI restriction sites. (Inset) Genotyping PCR analysis of the wild-type and Myc-335 null alleles. (B) Normal intestinal morphology and proliferation in Myc-335−/− mice at p1. Hematoxylin and eosin (HE) stains of sections show normal morphology in Myc-335−/− mice. IHC analysis showed normal proliferation (Ki-67) and restriction of the c-Myc expression to the developing crypts. Scale bars indicate 10 μm. (C) c-Myc expression detected in the crypts of adult Myc-335−/− mice. The arrowhead points to an Myc-positive cell located between two Paneth cells. Scale bars, 20 μm.

Loss of Myc is embryonic lethal and, even in heterozygous form, affects body size of newborn mice by compromising placental function (11, 12). However, the Myc-335−/− mice are viable and fertile and display no overt phenotype, indicating that loss of Myc-335 does not equal loss of Myc. Because rs6983267 increases risk for colorectal and prostate cancers and mice do not develop prostate tumors, we analyzed the intestinal morphology and function of the Myc-335−/− mice in more detail. Histological analysis of the intestines of Myc-335−/− mice, at postnatal day 1 (p1), showed normal morphology. The proliferation of the Myc-335−/− intestinal epithelium, as judged by the immunohistochemistry (IHC) analysis of the Ki-67 marker, was also comparable to that of the wild-type littermates. Furthermore, we detected an apparently normal pattern of Myc expression in the developing intestinal crypts of both the wild-type and the Myc-335−/− animals by IHC analysis (Fig. 1B and fig. S1). The intestinal epithelia of adult mice also showed normal histology and Myc expression in the crypts, including cells located adjacent to Paneth cells (13, 14). Analysis of Paneth cells by morphology (Fig. 1C), Goblet cells by periodic-acid Schiff base (PAS) staining (fig. S2), and Lgr5 expression by quantitative PCR (qPCR) (fig. S3) suggested that the loss of Myc-335 did not have a major impact on intestinal cell differentiation. This result is consistent with the lack of effect of Myc loss on homeostasis of intestinal epithelium (1517).

A significant difference in Myc RNA expression was not observed by qPCR from the duodenum of the Myc-335−/− mice at p1. However, we detected a decreased expression of Myc transcripts in the colon of the Myc-335−/− mice both by qPCR and exon array analysis at p1 (Fig. 2, A and B). Chromatin immunoprecipitation and sequencing (ChIP-seq) analysis for Tcf7l2 in adult mouse colon confirmed that, also in mice, the highest peak of Tcf7l2 binding within 1 Mb of Myc is located at Myc-335. Binding within this region was completely abolished in the colon of Myc-335−/− mice, and no obvious compensatory changes in binding pattern of Tcf7l2 were observed at other regions within 1 Mb of Myc (Fig. 2C). These results indicate that Myc-335 contains a major Tcf7l2 binding site and that its loss results in a modest decrease of Myc expression in the mouse colon. However, the Myc-335−/− mice are viable and fertile, indicating that the Myc-335 element is dispensable for function of the mouse intestine under standard mouse housing conditions.

Fig. 2

Myc-335 deletion affects Myc transcription. (A) qPCR analysis showing decreased expression of Myc in Myc-335−/− mice (n = 3) compared with wild-type control (n = 3), P < 0.04 (Student’s t test). Error bars denote one standard deviation. (B) Exon array analysis, showing decreased expression of Myc in Myc-335−/− mice. Hexagonal binning density plot of normalized RNA expression signal (log2) for each exon of a wild-type (x axis) and a Myc-335−/− (y axis) male mouse from the same litter are shown. The density scale is at right. Myc exons (red) are expressed at a slightly lower level in the Myc-335−/− mouse as compared with the wild-type colon. (C) Loss of major peak of transcription factor Tcf7l2 binding upstream of Myc in the Myc-335−/− animals (y axes, extended read coverage). (Inset) Tcf7l2 binding site within this peak. Red denotes the SNP rs6983267 associated with the risk allele in humans, which is normally present in mice.

The Myc-335 element was first identified as a conserved element containing a cancer-associated SNP that affects binding of TCF7L2 (5, 8), the transcription factor that is activated in most cases of colorectal cancer (18, 19). Thus, even though Myc-335 is not critical for normal intestinal function, it could still be required for tumorigenesis. To test this, we crossed the Myc-335−/− mice to the Apcmin mouse strain that spontaneously develops tumors in the small intestine and colon. These tumors are dependent on the activity of Myc (2022) and the T cell factor/lymphoid enhancer factor (TCF/LEF) family of transcription factors (2325). We generated Apcmin/+; Myc-335−/− mice and scored the number of polyps in the small intestine and colon of these mice at 4 months of age. We found a lower total number of polyps in the intestines of Apcmin/+; Myc-335−/− (n = 9) mice compared with that in the control Apcmin/+ mice (n = 9; P < 0.00038; Fig. 3). The effect was observed primarily at the level of frequency of occurrence of tumors. From these results, we conclude that, in laboratory mice, Myc-335 functions as a tumor-specific enhancer element that is dispensable for normal intestinal function but required for tumorigenesis.

Fig. 3

Reduced incidence of polyp formation in Apcmin/+; Myc-335−/− mice at 4 months of age. (A) Microscopic image of intestine showing abundant polyps in Apcmin/+mice compared with the Apcmin/+; Myc-335−/− mice. Scale bars, 2 mm. (B) Box plot summarizing the total numbers of polyps from both small intestine and colon observed per animal (n = 9 for each group). The ends of the whiskers denote the minimum and maximum of each group. The line within the box denotes the median. Student’s t test, P < 0.00038.

Most associations identified by using GWA studies map to noncoding regions whose function is largely unknown. This has been used as an argument against the validity of the GWA approach [for example, (26)]. GWA studies have implicated the region containing the SNP rs6983267 as being responsible for more human cancer-associated morbidity than any other known inherited variant or mutation. Our results presented here validate this region by showing that the Myc-335 enhancer has a critical role in intestinal tumorigenesis in mice. Because MYC-335 does not contain other SNPs with similar P values to rs6983267 (5), our use of the Myc-335 deletion allele does not affect the conclusion that rs6983267 is the causative SNP in humans.

Our results also highlight the fact that, although a disease-associated polymorphism typically has a relatively modest effect size, the element that it affects can be critically important for the underlying pathological process. In particular, MYC is a promising target for cancer therapy (27) that has been difficult to target by direct inhibitors (28). Our results suggest that MYC expression could be decreased by inhibitors targeting MYC-335. In a broader context, our finding that the phenotypic effect of Myc-335 loss is tumor-specific suggests that normal growth control and pathological growth induced by cancer can use different mechanisms.

Supplementary Materials

www.sciencemag.org/cgi/content/full/science.1228606/DC1

Materials and Methods

Figs. S1 to S3

Tables S1 and S2

References (2932)

References and Notes

  1. Acknowledgments: We thank S. Miettinen and A. Zetterlund for technical assistance. This work was supported by the Academy of Finland grant 140753, postdoctoral researcher’s project 134073, Finnish Center of Excellence in Cancer Genetics Research 250345, European Union Framework Program 7 project SYSCOL, European Research Council project GROWTHCONTROL, the Sigrid Juselius Foundation, the Finnish Cancer Society, the Swedish Cancer Society, the Center for Biosciences at Karolinska Institutet, and Vetenskapsrådet.
View Abstract

Navigate This Article