• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of pnasPNASInfo for AuthorsSubscriptionsAboutThis Article
Proc Natl Acad Sci U S A. Aug 26, 2008; 105(34): 12417–12422.
Published online Aug 21, 2008. doi:  10.1073/pnas.0801540105
PMCID: PMC2527926

The H19 locus acts in vivo as a tumor suppressor


The H19 locus belongs to a cluster of imprinted genes that is linked to the human Beckwith-Wiedemann syndrome. The expression of H19 and its closely associated IGF2 gene is frequently deregulated in some human tumors, such as Wilms' tumors. In these cases, biallelic IGF2 expression and lack of expression of H19 are associated with hypermethylation of the imprinting center of this locus. These observations and others have suggested a potential tumor suppressor effect of the H19 locus. Some studies have also suggested that H19 is an oncogene, based on tissue culture systems. We show, using in vivo murine models of tumorigenesis, that the H19 locus controls the size of experimental teratocarcinomas, the number of polyps in the Apc murine model of colorectal cancer and the timing of appearance of SV40-induced hepatocarcinomas. The H19 locus thus clearly displays a tumor suppressor effect in mice.

Keywords: genomic imprinting, Igf2, murine models

The H19-Igf2 locus is subject to genomic imprinting and has often been used as a paradigm for the study of this particular epigenetic regulation. The H19 locus produces a 2.5-kb noncoding, spliced, and polyadenylated RNA of yet-unknown function (1, 2). The Igf2 gene encodes a fetal growth factor, insulin-like growth factor 2. These two genes are located 90 kb apart and are oppositely imprinted: H19 is maternally expressed and Igf2 paternally expressed (1, 3). They belong to a large imprinted domain localized on chromosome 7 in mice and chromosome 11p15.5 in humans. The imprinting of Igf2 and H19 is controlled by a region located 4 kb upstream from the H19 transcription unit, defined as the H19 differentially methylated region (DMR) or imprinting control region (ICR) (4).

The 11p15.5-imprinted domain is linked to the Beckwith-Wiedemann syndrome (BWS), which is characterized by overgrowth phenotypes of affected children as well as a predisposition to develop embryonal tumors such as Wilms' tumor or rhabdomyosarcomas (5). Among the molecular alterations involved in BWS, certain cases (20%) show paternal uniparental disomy (UPD); other cases (5–10%) have hypermethylation of the H19 DMR; and both lead to lack of expression of H19 as well as activation of IGF2. These patients have higher risk of developing tumors than patients with other molecular defects (6). Genetic and epigenetic alterations at 11p15.5 similar to those found in the BWS have also been demonstrated in nonsyndromic Wilms' tumors. A great number of these cases have either loss of the maternal allele (LOH) or LOI (7, 8). It has thus been suggested that the H19 gene could have a possible tumor suppressor function (9). The first direct evidence for this tumor suppressor function was provided by in vitro experiments in which transfection of H19 cDNA into G401-transformed kidney cells resulted in loss of tumorigenicity of these cells (10). Subsequent experiments performed with in vitro culture systems suggested that H19 played a role as an oncogene rather than a tumor suppressor gene (11, 12). This controversy has not yet been resolved, as numerous human tumors have been shown to display either overexpression or lack of H19 expression (1315).

We decided to investigate the potential role of the H19 locus in vivo by producing murine models of tumorigenesis. We used H19Δ3 (16) and H19ΔEnh (17) mice (Fig. 1A) and 3 distinct models of tumorigenesis to investigate the potential tumor suppressor activity of the H19 locus. In the first model, experimental teratocarcinomas induced by grafting embryos under the kidney capsule were compared for size, weight, and histopathology (18). In the second model, the H19Δ3 mice were bred with mutants of the Apc gene, ApcΔ14/+, which represent a murine model for colorectal cancer (19). The double mutants lacking H19 and Apc show an increase in number of polyps compared with their Apc littermates. Finally, using a transgenic SV40 hepatocarcinoma model (20, 21), we show that the delay of appearance of these tumors is greatly reduced in the absence of H19. Interestingly, these models derive from the 3 germ layers (endoderm, mesoderm, and ectoderm) and result in similar phenotypes, showing a tumor suppressor function for the H19 locus.

Fig. 1.
Expression of the H19-Igf2 locus. (A) Maps of the H19-Igf2 locus. Wt, H19Δ3 and H19Δenh mutants are represented with maternal (Mat) and paternal (Pat) alleles. Endodermal enhancers (E Enh) are indicated downstream from the H19 gene. Black ...


Teratocarcinoma Model.

H19Δ3 phenotype and Igf2 expression.

We originally described that in the H19Δ3 mutants the maternal Igf2 allele was slightly reexpressed in skeletal muscle but not in liver (16). To identify the precise levels of Igf2 expression, we extended our analysis to other organs using a cross between H19Δ3 and SD7 mice. Igf2 is biallelically expressed only in mesoderm-derived tissues (limb muscle, tongue, diaphragm, and heart) and not in endoderm-derived tissues (kidney, lung, and liver) (Fig. 1B). Importantly, maternal Igf2 reexpression reaches at most 20–30% of the paternal allele in 5-day neonates, showing that there is only a slight increase in Igf2 mRNA levels in the H19Δ3 mice.

Production of tumors on wt background.

Experimental teratocarcinomas were produced by grafting E 6.5 embryos under the kidney capsule of syngenic mice. We first compared the weight of tumors obtained after grafting wt or H19Δ3 embryos into wt recipient mice. The results showed a clear difference for the two genotypes (Fig. 2A). Although there was some heterogeneity, the overall weight of H19Δ3 derived tumors was ≈1.6-fold higher than that of wt-derived tumors (P = 0.015).

Fig. 2.
Teratocarcinoma model. (A) Diagram of weights from wt and H19Δ3 derived tumors 35 days after grafting. Wt and H19Δ3 (indicated as H19−/−) are plotted in gray and black respectively, either on wt or Igf2−/− ...

In all cases, the tumor and the kidney were clearly individualized, with no invasion of the tumor into the kidney (Fig. 2B). Histologic analysis revealed the presence of different tissues derived from all 3 embryonic germ layers (ectoderm, endoderm, and mesoderm). There were no striking differences in the type of tissue, suggesting that the absence of H19 does not affect the development of any 1 particular tissue and that all 3 germ layers are involved.

Production of tumors on Igf2−/− background.

To separate paracrine and autocrine effects of IGF2, we performed the same grafting experiment using Igf2−/− recipient mice. There was a strong reduction in the overall size of the tumors for both genotypes (Fig. 2A), suggesting that circulating IGF2 in the recipient mouse plays an important role in the final size of the tumors. Most interestingly, the relative weight difference was maintained, with a 2-fold difference between H19Δ3 tumors compared with wt tumors on this Igf2−/− background (P < 0.05).

Taken together, analysis of the weight of the tumors at 35 days on both wt and Igf2−/− backgrounds showed a significant difference according to their genotype, with the wt tumors being smaller than the H19Δ3-derived tumors. Because the relative weight difference between the two genotypes (wt and H19Δ3) was maintained on both backgrounds (wt and Igf2−/−), this implies that the H19 locus plays a definite role in the production of these experimental tumors.

Tumor Characteristics.

H19 and Igf2 expression.

We investigated the levels of Igf2 mRNA by real-time quantitative RT-PCR in tumors produced both in the wt and the Igf2−/− host background. Igf2 expression was heterogeneous, did not strictly correlate with tumor weight and showed no significant difference between wt and H19−/− derived tumors (Fig. 2C). We concluded from these results (i) that tumor size is controlled by presence or absence of the H19 locus with little correlation to autocrine levels of Igf2 expression, and (ii) that the host background and the level of paracrine IGF2 play a major role in tumor size. Whether this effect is due to the protein itself or to other factors under the control of the Igf2 gene remains to be elucidated.

Methylation status of H19 ICR.

Because hypermethylation of the H19 DMR has been observed in Wilms' tumor samples, we investigated its methylation state in our teratocarcinoma samples. Tumor DNA was digested with SacI and HhaI (a methylation-sensitive enzyme) and the methylation state of the ICR was analyzed using a probe overlapping 1 of the CTCF binding sites (CTCF site 3) (Fig. 2D). These results were quantified and show a slightly higher methylation index (MI; methylated fragment/methylated plus unmethylated fragment) than expected but no significant difference between the two types of tumors. There was therefore no striking shift in the pattern of methylation of the H19 ICR.

The results obtained from this teratocarcinoma study suggest that the lack of H19 expression leads to larger tumors, consistent with its proposed role of tumor suppressor, and independently of Igf2 levels of expression.

Colorectal Cancer Model.

Increased adenoma number in the H19Δ3/+ApcΔ14/+ mutant mice.

To study the effect of absence or presence of the H19 locus in a murine carcinogenesis model, we bred H19Δ3 heterozygous females (outbred C57BL/6/CBA) with the ApcΔ14 heterozygous males. This cross produced 4 genotypes, H19+/+Apc+/+ (wt), H19Δ3/+Apc+/+, H19+/+ApcΔ14/+ and H19Δ3/+ApcΔ14/+. Because of genomic imprinting, H19Δ3/+ progeny completely lack H19 expression. Mice were analyzed at 180 days, as some began to show signs of anemia and rectal bleeding with prolapse. As expected, in the absence of Apc mutation, no polyps were detected in the wt and H19Δ3/+Apc+/+ mice. In ApcΔ14/+mice, the number of adenomas was significantly higher (2.2-fold) in the absence of H19 than in its presence (P = 0.05) (Fig. 3A). Strikingly, the increase in the number of smaller polyps (<2 mm) was even greater (3-fold difference) in the absence of H19. This would suggest that the H19 locus could control the initiation step of tumorigenesis.

Fig. 3.
Colorectal cancer model. (A) Effect of H19 deletion on ApcΔ14/+ intestinal polyps at 180 days. Top shows the number of polyps in H19+/+ApcΔ14/+ (gray) and H19Δ3/+ApcΔ14/+ (black) mice on C57BL/6 background (P = 0.05). ...

We performed the same cross using females carrying the H19Δ3 mutation on the 129/SvPas background. A similar increase in number of adenomas (1.4-fold) in mice lacking the H19 locus was detected. This difference (1.4- vs. 2.2-fold) is consistent with observations suggesting the presence of modifier genes in the 129 strain compared with the C57BL/6 background (22).

Histology of the intestinal lesions.

Close histologic examination of the whole-intestine “Swiss rolls” revealed the presence of adenomas, with some in situ adenocarcinomas (Fig. 3B), which occurred independently from the presence or absence of the H19 locus.

Previous experiments describing the effect of another H19 mutation, H19Δ13 (in which Igf2 mRNA level is increased by 2-fold) on tumor incidence in the Apc min mice had shown that an increase in crypt length was observed due to the increased level of IGF2 (23, 24). To exclude a role for IGF2 in the polyps produced in the H19Δ3 mutants, we evaluated the crypt size on hematoxylin and eosin(H&E) -stained sections of the small intestine, by performing 16 measures per section of H19+/+ApcΔ14/+ (n = 5, 3 on C57BL/6 and 2 on 129/SvPas background) and H19Δ3/+ApcΔ14/+ mice (n = 5, 2 on C57BL/6 and 3 on 129/SvPas background) (Fig. 3B). No difference in crypt length was found, with an average of 2.8 units for the H19+/+ApcΔ14/+ mice and of 2.6 units for the H19Δ3/+ApcΔ14/+ mice (a control wt mouse had a 2.8-unit crypt length). These data suggest that the levels of Igf2 expression did not differ between the wt and the H19Δ3 genotypes and did not affect intestinal growth in these mice.

Igf2 expression levels.

To confirm the absence of effect of the Igf2 gene, we compared the levels of Igf2 expression between wt and H19Δ3 strains by semiquantitative and real-time qRT-PCR. Igf2 expression is very low in colon and intestine in H19Δ3/+ as well as in wt mice, compared with muscle, where it is highly expressed (data not shown). Igf2 expression is also very low and heterogeneous in polyps from both H19Δ3/+ApcΔ14/+ and H19+/+ApcΔ14/+ mice, with no significant difference between the two genotypes (P = 0.06) (Fig. 3C).

ICR methylation analysis.

We also investigated whether the methylation status of the H19 ICR was disrupted in the intestine, colon, and polyp DNA from these mice. The H19 ICR displayed a constant methylation pattern, with the methylation index showing no significant difference between all genotypes and tissues analyzed (Fig. 3D).

Taken together, these results show that the lack of H19 expression causes an increase in the number of polyps in the Apc colorectal cancer model, independent of Igf2 expression. Interestingly, the initiation step of polyp appearance seems to be affected by the absence of H19.

Experimental Liver Carcinogenesis.

Tumor formation.

In mice carrying a targeted deletion of the endodermal enhancers located downstream from the H19 gene (H19ΔEnh) (17), it was shown that these enhancers are required in cis for the activation of Igf2 and H19 during liver carcinogenesis (20). To establish whether H19 had a role in tumor development, we analyzed liver carcinogenesis in mice lacking H19 expression (MatΔEnh, Fig. 1A).

The model investigated was that in which mice of the CRP-Tag 60-3 line carry the SV40 T antigen oncogene under the promoter of the liver-specific human C-reactive protein gene (25, 26). The males of this transgenic line have a low but constitutive expression of the SV40 T antigen, which, after formation of hyperplastic foci and neoplastic nodules, eventually leads to the development of multiple hepatocellular carcinomas at 4–5 months of age.

We have previously shown that H19 is activated in the experimental hepatocellular carcinomas (20, 26). We have now analyzed the expression of H19 and Igf2 in the liver tumors arisen in MatΔEnh and wt mice. Northern and in situ hybridization analyses showed that H19 RNA was undetectable in the MatΔEnh tumors but was activated in the majority of the neoplasms found in the wt mice (Fig. 4A and and44B). RNase protection analysis demonstrated that the majority of wt tumors had H19 RNA levels at least 3 orders of magnitude higher than the MatΔEnh tumors (data not shown). As in the wt mice, Igf2 was expressed at variable levels in all tumors with the MatΔEnh genotype (Fig. 4 A and B). RNase protection analysis on 15 wt and 15 MatΔEnh tumors demonstrated very similar Igf2 mRNA levels (4.72 ± 2.6 vs. 4.73 ± 1.6 arbitrary units, data not shown).

Fig. 4.
Experimental liver carcinogenesis model. (A) Expression analysis of Igf2 and H19 transcripts. Northern analysis of 15 wt or MatΔEnh liver tumors. The blot was hybridized sequentially with the H19, Igf2 and ribosomal 28S probes. (B) Igf2 and H19 ...

Latency of tumor development.

The male mice which developed liver tumors larger than 3 mm by 120 days of age were 11/31 (35%) for the MatΔEnh and 5/24 (20%) for the wt genotypes, respectively. By 127 and 134 days, the mice with tumors were 6/8 and 6/9 for the MatΔEnh genotype and 7/19 and 3/7 for the wt genotype, respectively (Fig. 4C). Overall, 23/48 (48%) MatΔEnh and 15/50 (30%) wt male mice (P = 0.037) had liver tumors when analyzed between 120 and 134 days of age. In contrast, PatΔEnh mice, expressing high levels of H19 but low levels of Igf2, developed tumors with a delayed time course (20). These results suggest that the lack of H19 expression causes acceleration in the development of liver tumors, consistent with its proposed role of tumor suppressor.


The presence of a potential tumor suppressor gene in the 11p15.5 chromosomal region has been hypothesized for many years. This stemmed from the observation that in patients with BWS, there was a predisposition toward development of embryonal tumors. Two domains of imprinted genes have been identified in this region, the IC1 domain with the H19-IGF2 locus and the IC2 domain with the KCNQ1 locus associated with several other genes such as CDKN1C (or p57KIP2). The H19 locus, with no known function was one of the candidates for a tumor suppressor (9).

The initial evidence in support of this hypothesis mainly involved in vitro experiments (10). Our aim was to use animal models to investigate the potential tumor suppressor activity of H19 in vivo. The H19Δ3 mice we had produced never developed tumors, whatever background they were bred on (outbred C57BL/6/129 or inbred 129/SvPas). For this reason, we challenged these mutant mice in different tumor models.

The teratocarcinoma model has been an experimental system for producing tumors in the mouse which was described many years ago (27). Teratocarcinomas are composed of highly undifferentiated embryonal carcinoma (EC) cells and of differentiated cells derived from all 3 embryonic layers (mesoderm, endoderm, and ectoderm) (18). Teratocarcinomas can be produced either by grafting embryos or by injecting ES cells under the skin. We chose to graft E 6.5 embryos under the kidney capsule because this approach provided a better control of the number of grafted cells. We were aiming not only to investigate whether tumors were produced in presence or absence of H19, but also whether there was a size difference in these different cases.

The comparison of tumors induced by either wt or H19Δ3 embryos clearly shows a difference in the weight of the tumors that are produced after 35 days. There is, of course, some heterogeneity, but overall H19Δ3 tumors are larger than their wt counterparts. Importantly, the size difference between H19Δ3 and wt induced teratocarcinomas was maintained on the Igf2−/− background (2-fold) compared with the wt background (1.6-fold).

These tumors displayed cells derived from endoderm, mesoderm, and ectoderm. This is interesting with regard to observations made on teratocarcinomas produced from androgenetic ES cells, which display lack of H19 (maternally expressed) as well as disruption of many other imprinted genes (28). These tumors consisted predominantly of striated muscle. Since we find no difference in the type of tissues of H19Δ3 or wt derived teratocarcinomas, the H19 locus is probably not involved in the overproduction of muscle cells found in the androgenetic tumors. This could be of interest with respect to the occurrence of rhabdomyosarcomas in BWS patients and suggests that perhaps other genes of the 11p15.5 region are responsible for this type of tumor.

The choice of our second colorectal cancer model was prompted by data showing increase in size and number of polyps in mice carrying the H19Δ13 mutation compared with wt mice (24). These mutant mice lack H19 but also overexpress maternally derived Igf2 because of the ICR deletion. To discriminate between the effects of each one of these genes, we performed the cross between H19Δ3 females (in which the 3-kb transcription unit only is deleted) and Apc mutant males.

The number of polyps is >2-fold higher in absence of H19, with an increase in the number of small polyps, suggesting that lack of H19 may play a role in the initiation step of tumorigenesis. Only low levels of Igf2 mRNA were found in normal tissue (intestine and colon) and in the polyps, with no significant difference between mutant and wt polyps. No difference in the crypt size was detected, whereas previously published data suggested that relative levels of Igf2 are responsible for crypt depth (24). Taken together, the results obtained from both our H19Δ3 mutants and the H19Δ13 mutants allow to postulate that H19 is playing a definite role in the production and size of polyps, whereas Igf2 may be contributing to the growth of these polyps by affecting the intestinal crypt size.

The SV40 induced hepatocarcinomas interestingly revealed an acceleration in the latency of appearance of the tumors in the absence of H19 expression. Expression of the Igf2 and H19 genes is completely shut off in the liver of adult CRP-Tag mice, but is reactivated in a coordinate manner during liver carcinogenesis with conservation of their imprinted expression (26). In addition, loss of the maternal and duplication of the paternal copy of the chromosomal region bearing the Igf2 and H19 genes occur at high frequency in the hepatocellular carcinomas. These genetic events resemble the LOH occurring at chromosome 11p15.5 loci in human cancers and result in activation of IGF2 and lack of H19 expression.

It has been recently reported that the H19 gene may act as an oncogene in studies using human cells maintained in culture and injected into mice to produce tumors (15). The discrepancy with our results could be explained by the difference in the systems. The main interest of our study resides in the use of mouse genetics. Our models reproduce a situation in which the potential oncogenic teratocarcinomas, Apc−/+ polyps or SV40 induced hepatocarcinomas are challenged with mice in which the H19 locus is present (or absent) throughout embryogenesis and the whole life of the mouse. Its effect is therefore constant and this may represent a more biologic situation than a cell culture system. It must also be acknowledged that the H19 locus may play a more complex role in humans than in mice.

Because H19 KO mice never spontaneously develop tumors in vivo, as other murine models of tumor suppressor genes, H19 may play the role of a “modifier gene” suppressing tumorigenesis. It could act either through its long noncoding RNA or through the microRNA (miR-675) that has been recently described in exon 1 (29, 30). Targets of H19 remain to be identified and linked to a biologic function possibly related to pathways involved in tumorigenesis.

Materials and Methods

Mouse Strains and Genotyping.

The H19Δ3 strain carries a 3kb deletion of the H19 transcription unit and was initially established on an outbred C57BL/6/CBA background (16). Because an isogenic 129 background was required for the teratocarcinoma experiment, we also produced a 129 H19Δ3 strain by reinjection of the original H19Δ3/+ ES cell line into blastocysts. The 129/SvPas wt strain and the 129 Igf2−/− strain (31) were used as recipients for the production of teratocarcinomas. The H19ΔEnh mice (17) were maintained on a C57BL/6 background. These mice can be bred as maternal heterozygotes (MatH19ΔEnh) which lack H19 expression since the deletion is carried on the maternal allele (Fig. 1A). The SD7 strain is a C57BL/6/CBA strain carrying the distal part of Mus spretus chromosome 7. The ApcΔ14/+ mice were bred on a C57BL/6 background (19) and were crossed with either the outbred H19Δ3 or the inbred 129 H19Δ3. The CRP-Tag 60-3 strain (25) was maintained on a BALB/c background. The protocol of animal handling and treatment was performed in accordance with the guidelines of the animal ethics committee of the Ministère de l'Agriculture of France.

Genotyping was done by PCR on tail DNA. Primers used for detecting the H19Δ3 allele were neo primers: 5′ -GTCCTGATAGCGGTCCGCCA-3′ and 5′-GTGTTCCGGCTGTCAGCGCA-3′ (500 bp). The ApcΔ14 allele was detected using primers that distinguish the wt allele (180 bp) from the exon14-excised allele (160 bp): Primer 1: 5′CTGTTCTGCAGTATGTTATCA-3′; Primer 2: 5′-CTATGAGTCAACACAGGATTA-3′; Primer 3: 5′-TATAAGGGCTAACAGTCAATA-3′.

Teratocarcinoma Production.

Wt or H19Δ3 embryos were dissected at 6.5dpc and the ectoplacental cone was taken off. These embryos were introduced under the kidney capsule of isogenic males (8–10 weeks old), as previously described (18). The recipient mice were either 129/SvPas wt or 129 Igf2−/− strain. 35 days after grafting, tumors were surgically dissected from the mice and weighed. A fraction of the tumor was used to prepare DNA and RNA. The remaining part was fixed in Bouin's fixative, embedded in paraffin, and 5 μm sections were stained with hematoxylin and eosin (H & E).

Polyp Analysis and Tumor Scoring.

The progeny from the cross between H19Δ3/+ and ApcΔ14/+ mutant mice were killed at 180 days and genotyped. The entire intestinal tract was removed, flushed with PBS, and stained with Indigo carmine (0.08%). The small intestine and colon were opened longitudinally, flattened on filter paper, and fixed in 4% PFA. The number and size of the polyps were determined by double counting on mice blinded for genotype. The whole intestine was then rolled and embedded in paraffin for histologic analysis (“Swiss rolls”). Crypt length and tumor grading were performed on 5-μm H & E sections.

Analysis of Liver Tumorigenesis.

Homozygous CRP-Tag 60-3 males were mated with MatH19ΔEnh females, and liver carcinogenesis was analyzed in their progeny. The mice were genotyped for the presence of the H19ΔEnh, as described in ref. 21. The males were killed between 120 and 134 days of age. Livers were dissected from the mice and carefully examined for the presence of tumors. Only nodules larger than 3 mm were considered.

Statistical Analyses.

Data are shown as averages and s.e.m. We used ANOVA analysis and Student t test with Excel X and Statview.

RNA Preparation and Analysis.

Total RNA was extracted from 5-day neonate organs or tumors with TRIzol reagent (Invitrogen). For RT-PCR analysis, DNase I treated RNA (0.5 μg) was reverse-transcribed with SuperScript II and random hexamer primers (Invitrogen). For semiquantitative RT-PCR, 50 ng of cDNA were amplified using gene-specific primers and TaqDNA polymerase (Invitrogen) during 20 cycles. Detection of Igf2 transcripts derived from the SD7 cross was performed using primers: forward 5′-GACGTGTCTACCTCTCAGGCCGTACTT-3′ and reverse 5′ GGGTGTCAATTGGGTTGTTTAGAGCCA-3′. The 517-bp product was digested with BsaA1, yielding a 241-bp paternal fragment detected by the internal primer 5′- TCAAATTTGGTTTTTTAGAA-3′. RT-PCR products were separated on 1% agarose gels and transferred onto Hybond N+ membranes in 0.4 M NaOH. Blots were probed with γP32ATP kinase-labeled primers at 42°C in Church buffer. Membranes were washed in 0.4×SSC, 0.5% SDS at 42°C and results were quantified using PhosphorImager analysis and ImageQuant software.

Quantitative q-PCR was performed on a Light Cycler system using Sybr Green PCR kits (Roche). 1 to 5 ng of cDNA were amplified in duplicate using primers for H19, Igf2, GAPDH and TBP. H19 F 5′-GGAGACTAGGCCAGGTCTC-3′; H19 R 5′-GCCCATGGTGTTCAAGAAGGC-3′; Igf2 F 5′-GGCCCCGGAGAGACTCTGTGC-3′; Igf2 R 5′-TGGGGGTGGGTAAGGAGAAAC-3′; GAPDH F 5′-ACAGTCCATGCCATCACTGCC-3′; GAPDH R 5′-GCCTGCTTCACCACCTTCTTG-3′; TBP F 5′-GCAATCAACATCTCAGCAACC-3′ and TBP R 5′-CGAAGTGCAATGGTCTTTAGG-3′. Genorm calculations were used for normalization.

Northern analysis and RNase protection assays were carried out as previously described (20).

Methylation Assay.

DNA from tumors, polyps or control tissue was incubated at 55°C in lysis buffer (Tris 100 mM pH 8, EDTA 5 mM, SDS 0.2%, NaCl 20 mM, and 0.4 mg/ml Proteinase K (Sigma), followed by phenol-chloroform extraction and ethanol precipitation. DNA was digested with SacI and HhaI, separated on 1% agarose gels and transferred onto Hybond N + membranes. Southerns were probed with a 200-bp PCR product corresponding to the region covering the CTCF site No. 3: 5′ CTGTTATGTGCAACAAGGGAA and 3′ GGTCTTACCAGCCACTGA. Blots were washed at 65°C and quantified as described above.

In situ Hybridization.

In situ hybridization on liver sections was performed as previously described (26).


We thank Arg Efstratiadis, Wolf Reik, and Shirley Tilghman for their kind gifts of Igf2−/− and SD7 mice to L.D and H19Δenh mice to A.R. We thank Béatrice Romagnolo and Christine Perret for advice on the ApcΔ14 colorectal cancer model. We are very grateful to Davor Solter, Jean Gaillard, and Michel Huerre for helpful comments and a critical eye on the teratocarcinoma sections. We thank Debra Wolgemuth and Edith Heard for constructive discussions on the manuscript. We thank Charles Babinet, Wolf Reik, and Azim Surani for constant encouragement. This work was supported by funding from the Ministère de la Recherche (ACI), Association de la Recherche contre le Cancer (ARC), Ligue contre le Cancer, Association Française contre les Myopathies (AFM), MIUR PRIN 2005 (AR), Istituto Superiore di Sanità (AR), Associazione Italiana Ricerca sul Cancro (AR), Telethon-Italia Grant GGP07086 (AR), and fellowships from ARC and AFM to A.M. and to T.Y.


The authors declare no conflict of interest.

This article is a PNAS Direct Submission. B.T. is a guest editor invited by the Editorial Board.


1. Bartolomei MS, Zemel S, Tilghman SM. Parental imprinting of the mouse H19 gene. Nature. 1991;351:153–155. [PubMed]
2. Gabory A, Ripoche MA, Yoshimizu T, Dandolo L. The H19 gene: Regulation and function of a non-coding RNA. Cytogenet Genome Res. 2006;113:188–193. [PubMed]
3. DeChiara TM, Robertson EJ, Efstratiadis A. Parental imprinting of the mouse insulin-like growth factor II gene. Cell. 1991;64:849–859. [PubMed]
4. Thorvaldsen JL, Duran KL, Bartolomei MS. Deletion of the H19 differentially methylated domain results in loss of imprinted expression of H19 and Igf2. Genes Dev. 1998;12:3693–3702. [PMC free article] [PubMed]
5. Weksberg R, Smith AC, Squire J, Sadowski P. Beckwith-Wiedemann syndrome demonstrates a role for epigenetic control of normal development. Hum Mol Genet. 2003;12:R61–R68. [PubMed]
6. Cooper WN, et al. Molecular subtypes and phenotypic expression of Beckwith-Wiedemann syndrome. Eur J Hum Genet. 2005;13:1025–1032. [PubMed]
7. Moulton T, et al. Epigenetic lesions at the H19 locus in Wilms' tumor patients. Nat Genet. 1994;7:440–447. [PubMed]
8. Feinberg AP. Imprinting of a genomic domain of 11p15 and loss of imprinting in cancer: An introduction. Cancer Res. 1999;59:1743s–1746s. [PubMed]
9. Rump P, Zeegers MP, van Essen AJ. Tumor risk in Beckwith-Wiedemann syndrome: A review and meta-analysis. Am J Med Genet A. 2005;136:95–104. [PubMed]
10. Hao Y, Crenshaw T, Moulton T, Newcomb E, Tycko B. Tumor-suppressor activity of H19 RNA. Nature. 1993;365:764–767. [PubMed]
11. Lustig-Yariv O, et al. The expression of the imprinted genes H19 and IGF-2 in choriocarcinoma cell lines. Is H19 a tumor suppressor gene? Oncogene. 1997;15:169–177. [PubMed]
12. Verkerk A, et al. Unique expression patterns of H19 in human testicular cancers of different etiology. Oncogene. 1997;14:95–107. [PubMed]
13. Dao D, et al. Multipoint analysis of human chromosome 11p15/mouse distal chromosome 7: Inclusion of H19/IGF2 in the minimal WT2 region, gene specificity of H19 silencing in Wilms' tumorigenesis and methylation hyper-dependence of H19 imprinting. Hum Mol Genet. 1999;8:1337–1352. [PubMed]
14. Frevel MA, Sowerby SJ, Petersen GB, Reeve AE. Methylation sequencing analysis refines the region of H19 epimutation in Wilms tumor. J Biol Chem. 1999;274:29331–29340. [PubMed]
15. Matouk IJ, et al. The H19 non-coding RNA is essential for human tumor growth. PLoS ONE. 2007;2:e845. [PMC free article] [PubMed]
16. Ripoche MA, Kress C, Poirier F, Dandolo L. Deletion of the H19 transcription unit reveals the existence of a putative imprinting control element. Genes Dev. 1997;11:1596–1604. [PubMed]
17. Leighton PA, Saam JR, Ingram RS, Stewart CL, Tilghman SM. An enhancer deletion affects both H19 and Igf2 expression. Genes Dev. 1995;9:2079–2089. [PubMed]
18. Damjanov I, Solter D, Skreb N. Teratocarcinogenesis as related to the age of embryos grafted under the kidney capsule. Wilhelm Roux' Archiv. 1971;107:288–290.
19. Colnot S, et al. Colorectal cancers in a new murine model of familial adenomatous polyposis: Influence of genetic and environmental modifiers. Lab Invest. 2004;84:1619–1630. [PubMed]
20. Vernucci M, et al. The H19 endodermal enhancer is required for Igf2 activation and tumor formation in experimental liver carcinogenesis. Oncogene. 2000;19:6376–6385. [PubMed]
21. Vernucci M, et al. Developmentally regulated functions of the H19 differentially methylated domain. Hum Mol Genet. 2004;13:353–361. [PubMed]
22. Hassan AB, Howell JA. Insulin-like growth factor II supply modifies growth of intestinal adenoma in Apc(Min/+) mice. Cancer Res. 2000;60:1070–1076. [PubMed]
23. Leighton PA, Ingram RS, Eggenschwiler J, Efstratiadis A, Tilghman SM. Disruption of imprinting caused by deletion of the H19 gene region in mice. Nature. 1995;375:34–39. [PubMed]
24. Sakatani T, et al. Loss of imprinting of Igf2 alters intestinal maturation and tumorigenesis in mice. Science. 2005;307:1976–1978. [PubMed]
25. Ruther U, Woodroofe C, Fattori E, Ciliberto G. Inducible formation of liver tumors in transgenic mice. Oncogene. 1993;8:87–93. [PubMed]
26. Casola S, et al. Loss of heterozygosity of imprinted genes in SV40 t/T antigen-induced hepatocellular carcinomas. Oncogene. 1995;11:711–721. [PubMed]
27. Stevens LC. The development of transplantable teratocarcinomas from intratesticular grafts of pre- and postimplantation mouse embryos. Dev Biol. 1970;21:364–382. [PubMed]
28. Mann JR, Gadi I, Harbison ML, Abbondanzo SJ, Stewart CL. Androgenetic mouse embryonic stem cells are pluripotent and cause skeletal defects in chimeras: Implications for genetic imprinting. Cell. 1990;62:251–260. [PubMed]
29. Mineno J, et al. The expression profile of microRNAs in mouse embryos. Nucleic Acids Res. 2006;34:1765–1771. [PMC free article] [PubMed]
30. Cai X, Cullen BR. The imprinted H19 noncoding RNA is a primary microRNA precursor. RNA. 2007;13:313–316. [PMC free article] [PubMed]
31. DeChiara TM, Efstratiadis A, Robertson EJ. A growth deficiency phenotype in heterozygous mice carrying an insulin-like growth factor II gene disrupted by gene targeting. Nature. 1990;345:78–80. [PubMed]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...