Physiological Genomics

Onecut-2 knockout mice fail to thrive during early postnatal period and have altered patterns of gene expression in small intestine

Mary R. Dusing, Elizabeth A. Maier, Bruce J. Aronow, Dan A. Wiginton


Ablation of the mouse genes for Onecut-2 and Onecut-3 was reported previously, but characterization of the resulting knockout mice was focused on in utero development, principally embryonic development of liver and pancreas. Here we examined postnatal development of these Onecut knockout mice, especially the critical period before weaning. Onecut-3 knockout mice develop normally during this period. However, Onecut-2 knockout mice fail to thrive, lagging behind their littermates in size and weight. By postnatal day (d)19, they are consistently 25–30% smaller. Onecut-2 knockout mice also have a much higher level of mortality before weaning, with only ∼70% survival. Interestingly, Onecut-2 knockout mice that are heterozygous for the Onecut-3 knockout allele are diminished even further in their ability to thrive. They are ∼50–60% as large as their normal-sized littermates at d19, and less than half of these mice survive to weaning. As reported previously, the Onecut-2/Onecut-3 double knockout is a perinatal lethal. Microarray technology was used to determine the effect of Onecut-2 ablation on gene expression in duodenum, whose epithelium has among the highest levels of Onecut-2. A subset of intestinally expressed genes showed dramatically altered patterns of expression. Many of these genes encode proteins associated with the epithelial membrane, including many involved in transport and metabolism. Previously, we reported that Onecut-2 was critical to temporal regulation of the adenosine deaminase gene in duodenum. Many of the genes with altered patterns of expression in Onecut-2 knockout mouse duodenum displayed changes in the timing of gene expression.

  • microarray
  • growth defect
  • lipid metabolism
  • transport

the epithelium of the adult mammalian intestine represents a unique, compartmentalized system of cellular proliferation and differentiation (31). This intestinal epithelium has a complex, but highly defined, architecture. The primary components of this architecture in the small intestine are the crypt and the villus. In mouse, morphogenesis of the gut epithelium occurs rather late in development, from embryonic day (e)15 through approximately postnatal day (d)21. A number of dramatic changes in gene expression and function occur during these final stages of development in the intestinal epithelium (12, 30, 32). This is when the final crypt-villus architecture of the small intestine matures and much of its digestive capability is established. The genetic programs that control these developmental processes are poorly understood.

Our previous studies (24) identified a crucial role for Onecut (OC) transcription factors in the temporal regulation of the adenosine deaminase (ADA) gene in small intestine during this critical developmental phase. OC factors bind in a distinct temporal regulatory element and control the timing of enhancer activation by the duodenum-specific enhancer that regulates ADA expression in the epithelium. This observation led us to investigate the possibility that OC factors may play a broader role in orchestrating intestinal gene expression during the final stages of intestinal development, especially in controlling the timing of gene expression.

The Onecut family of factors is comprised of three members in mammals, Onecut 1 (OC-1, originally HNF-6), Onecut 2 (OC-2), and Onecut 3 (OC-3) (4, 14, 16, 21, 33). OC factors are known to have important roles in development and cell differentiation in pancreas and liver, including critical roles in controlling the timing of specific developmental events (2, 6, 7, 13, 15, 25, 28). In contrast, little is known about the function of OC factors in postnatal intestine. Previous studies have shown that only OC-2 and OC-3 are expressed in postnatal small intestine (OC-1/HNF-6 is not) (14, 16, 17, 21, 24, 33, 34). Availability of OC-2 and OC-3 knockout mice (6, 34) has made it feasible to assess the general role of these factors in intestine.

Our initial studies with OC-2 knockout mice showed that they fail to thrive postnatally, are significantly smaller than normal siblings, and have increased mortality during the postnatal developmental period. These problems are exacerbated when the mice are also heterozygous for the OC-3 null allele. Initial microarray studies with duodenum showed dramatic changes in intestinal gene expression in OC-2 knockout mice during this postnatal period, especially in genes associated with epithelium. This included a number of genes related to transport and metabolism. There are many examples where timing of gene expression is altered.


Source, breeding, and genotyping of Onecut knockout mice.

Both the oc2−/− and oc3−/− mouse strains have been described previously (6, 34) and were obtained from Patrick Jaquemin and Frederique Lemaigre (Hormone and Metabolic Research Unit, Université Catholique de Louvain, Brussels, Belgium). Founder heterozygous oc2+/− and oc3+/− male and female mice were bred to 129 mice (Charles River) in our colony to establish the lines. All subsequent breeding and crossbreeding of these lines described in our studies were carried out with these mice or their descendants. All of the studies described in this study were carried out under an animal use protocol approved and monitored by the Cincinnati Children's Hospital Medical Center (CCHMC) internal Institutional Animal Care and Use Committee (IACUC). Genotyping of mice for the OC-2 knockout allele was carried out with following PCR primers: wild-type OC-2 allele: FL OC2-F, 5′-GCCACGCCGCTGGGCAAC-3′ and FL OC2-R, 5′-CAGCTGCCCGGACGTGGC-3′; puromycin in the OC-2 knockout allele: FL PUR-F, 5′-GACCGAGTACAAGCCCACG-3′ and FL PUR-R 5′-GTCCGCGACCCACACCTTG-3′. Genotyping of mice for the OC-3 knockout allele was carried out with PCR primers described previously (34), with the exception of the primers used to detect the wild-type OC-3 gene. Those primers were OC3-F5, 5′-ATGGGTATGGCGTGCGAGG-3′ and OC3-R7, 5′-TGAGGGTGAGCGTGTGGATG-3′. PCR amplification was carried out with the Failsafe PCR system (Epicentre Biotechnologies, Madison, WI) and the following buffer premixes: premix E (OC-2), premix J (PUR), premix G (OC-3), and premix A [green fluorescent protein (GFP)].

Immunohistochemical staining to localize OC protein expression in small intestinal sections.

Immunohistochemical detection was performed on mouse intestinal sections essentially as described previously (10). Antibodies used were directed against OC-2 (Abcam no. ab28466; 6–8 μg/ml) or OC-1 (Santa Cruz Biotechnology no. SC-13050; 4 μg/ml). Staining was performed with the ABC kit from Vector Laboratories.

Real time RT-PCR analysis of mRNA levels.

Total RNA was prepared as described previously (1) from the appropriate small intestinal segments (duodenum, jejunum, ileum) at d2, d15, d18, or d30. Reverse transcription of the total RNA for first-strand synthesis was carried out as described previously (24), with the Invitrogen-Superscript III protocol. Real-time PCR was carried out with a Qiagen QuantiTect SYBR Green PCR system in a Bio-Rad PTC-200 DNA Engine thermal cycler (normally 36 cycles). Samples were normalized against GAPDH, as determined from a GAPDH standard curve, and expressed as relative values after normalization. PCR primer pairs for analysis of OC-1, OC-2, OC-3, GAPDH, sucrase isomaltase (SI), and lactase-phlorizin hydrolase (LPH) were obtained from Qiagen QuantiTect.

Evaluation of pup growth and relative pup weight in litters with variable genotypes.

In general, average pup weight in litters with different numbers of pups is inversely related to the size of the litter. Therefore, relative pup weight was used to compare among litters and to generate statistical comparisons. Relative pup weight was determined as actual pup weight divided by average pup weight in the litter. All statistical analyses of relative weight differences, including unpaired t-tests and estimation of significance, were carried out with GraphPad software (GraphPad Software).

Microarray-based comparison of gene expression in duodenum of oc2−/− and wild-type mice.

Total RNA for microarray analysis was isolated as described above from duodenum samples harvested at d15 and d30 from oc2−/− and wild-type 129 mice. RNA isolations were carried out from three separate sets of animals for each experimental group, and each RNA preparation from a given set included duodenal tissue from five to eight individual mice of the appropriate age and genotype. Before probe preparation, RNA samples were subjected to quality control testing on an Agilent Bioanalyzer. The microarray analysis was then carried out by standard Affymetrix protocols. RNA samples underwent first- and second-strand cDNA synthesis, amplification, purification, quantification, fragmentation, and biotin labeling. Labeled cRNA was then hybridized to the Affymetrix microarray, which was subsequently washed, stained, and scanned. For each sample, a number of quality controls were checked after the data analysis file was created (including examination for RNA degradation, inefficient transcription or labeling, background, and noise).

Microarray analyses were performed at the Cincinnati Children's Hospital Microarray Core Facility with the Affymetrix Mouse Gene 1.0 ST Array platform and standard labeling kit. Image files were processed with Affymetrix Expression Console version 1.1.1 and the robust multichip average (RMA) normalization method. The mouse ST array allows whole genome, gene-level expression studies for >20,000 well-characterized genes to be carried out. Each of these genes is represented on the array by ∼27 probes spread across the full length of the gene.

Microarray data analysis.

RMA expression values for each probe set for each sample were transformed to linear relative intensity values (I = 2**RMA) and then to a ratio relative to the median value for each probe set relative to that of the six wild-type samples at d15 and d30. In this way, expression changes relative to this median can be detected for both the effects of the knockout, postnatal maturation, and individual variation between the replicate samples within a group. With the use of GeneSpring software (Agilent Technologies) to analyze the microarray data, probe sets were first identified that had a raw RMA expression value >6.0 in at least three samples (26,794 probe sets) and then that had differential expression between any of the four groups by a Welch ANOVA with FDR < 0.05 using Benjamini-Hochberg false discovery rate (FDR) correction (5,277 probe sets). Statistically significant differences in the expression of individual genes in the oc2−/− mice at different developmental time points and changes were then compared with wild type at each time point. Thus both genes that exhibited differential expression in the duodenum between d15 and d30 of postnatal mouse development and those that exhibited altered expression in the oc2−/− mice could be identified. From these, genes that exhibited the greatest degree of differential regulation were identified by group-group average fold differences. Gene enrichment analysis for shared functional associations or gene-associated features of the genes most differentially altered in the oc2−/− mice was performed with the ToppGene Suite for gene list enrichment analysis and candidate gene prioritization (5). The data discussed in this publication have been deposited in NCBI's Gene Expression Omnibus (11) and are accessible through GEO Series accession number GSE18074 (

The validation of microarray results for selected genes was accomplished by real-time RT-PCR conducted as described above with gene-specific primers. This validation by real-time PCR was carried out for a number of genes, including OC-3, OC-1, SI, and LPH. Both SI and LPH showed the expected developmental changes in expression between d15 and d30, but these were not altered by the ablation of the OC-2 gene (data not shown).


Evaluation of postnatal intestine in Onecut knockout mice.

Mouse gene ablation studies have been reported previously for both of Onecut transcription factors OC-2 (6) and OC-3 (34), which are known to be expressed in small intestine. The published studies with knockout mice have focused principally on the role of these Onecut factors in embryonic development of the pancreas and liver (6, 25, 34). Since both OC-2 and OC-3 are highly expressed in mouse postnatal small intestine, we were interested in their potential role there.

Expression of Onecut factors in postnatal mouse intestine.

It has been previously shown that OC-2 and OC-3 delineate specific nonoverlapping domains of the gastrointestinal tract during embryonic development (34). At e12.5, OC-3 is expressed in an anterior segment from the common bile duct to the pylorus, while OC-2 is expressed from the common bile duct into the posterior small intestine. However, by the time of birth both OC-2 and OC-3 are expressed along the length of the small intestine in different patterns. OC-2 is widely expressed in most cells of the crypt and villous epithelium, while OC-3 expression appears to be limited to enteroendocrine cells (34). We examined OC-2 expression in normal postnatal and adult mouse small intestine and found that the newborn pattern persisted. The relative levels of OC-2 and OC-3 at different postnatal time points in the various segments of the small intestine are shown in Fig. 1. Both OC-2 and OC-3 mRNA are higher in the proximal small intestine than in the more distal segments at all postnatal developmental time points. OC-2 was widely expressed in most villous epithelial cells, as well as many cells in the crypt (especially the transit amplifying zone) (Fig. 2B). OC-3 was much more difficult to visualize by immunohistochemical means. Real-time RT-PCR estimations indicated that absolute levels of OC-2 mRNA were 10–30 times the levels of OC-3 mRNA in the small intestine.

Fig. 1.

mRNA levels of the 3 Onecut (OC) factors in small intestinal segments assessed by quantitative real-time PCR. Total RNA samples were prepared from duodenal, jejunal, and ileal samples from mice at postnatal day (d)2, d18, and d30. OC-1, -2, and -3 mRNA levels were quantified by real-time RT-PCR with factor-specific primer pairs. In each case, OC values are expressed relative to the d30 duodenal samples, after normalization to GAPDH mRNA levels.

Fig. 2.

Expression of OC-1/HNF-6 and OC-2 protein in the Brunner's glands and duodenal epithelium. Immunohistochemical analysis of sections of adult mouse duodenum with anti-OC-1/HNF-6 antibody (A) and anti-OC-2 antibody (B) are shown. A: developing crypts and villi in left half of section are negative for OC-1/HNF-6. Brunner's gland region (BG) on right shows high levels of nuclear OC-1/HNF-6 expression throughout. B: high levels of expression of OC-2 are observed throughout the duodenal epithelium in most cells, including many in the crypt and transit amplifying zone. Very little, if any, OC-2 expression is observed in the cells of the Brunner's gland region. The section in A has been counterstained with hematoxylin, whereas the section in B has not.

OC-1/HNF-6 mRNA is essentially undetectable in all segments of the small intestine at all time points, except duodenum at d30. We reported this previously (24), but it was somewhat surprising because OC-1/HNF-6 had been reported by others to be absent from the small intestine (33). To investigate further, we performed immunohistochemistry to localize OC-1/HNF-6 expression within intestinal segments. As predicted from the mRNA studies, OC-1/HNF-6 was absent from jejunal and ileal sections (data not shown). In adult duodenal sections, OC-1/HNF-6 was found to be present exclusively within the Brunner's glands and absent from the normal absorptive epithelium (Fig. 2A). Fully developed Brunner's glands do not appear in duodenum until the final stages of mouse intestinal development (27), corresponding to the timing of appearance of OC-1/HNF-6 at d30 and later. Interestingly, immunohistochemistry showed that OC-2 was not expressed at detectable levels in the Brunner's glands (Fig. 2B). The role or significance of OC-1 expression in the Brunner's glands is uncertain at this time.

Onecut-2 knockout mice fail to thrive during postnatal development, displaying significantly reduced size and increased mortality.

It has been reported previously that the gastrointestinal tracts of oc2−/− and oc3−/− mice show no abnormalities during the various stages of embryonic development (34). The intestinal epithelia of newborn oc2−/− and oc3−/− mice were reported to have no histological defects or alterations in crypt-villus architecture, as well (34). There are, however, a number of dramatic changes that normally occur in the intestinal epithelium during the postnatal developmental period, before weaning. Therefore, we examined the epithelia of both oc2−/− and oc3−/− mice compared with wild type at several developmental time points during this postnatal period. The epithelia of both oc2−/− and oc3−/− mice continued to be structurally and histologically normal during the postnatal period (Supplemental Fig. S1).1

As reported previously (6), oc3−/− mice are viable and fertile and are essentially indistinguishable from wild-type animals in appearance, average weight, and behavior. In contrast, oc2−/− mice demonstrate a distinct failure to thrive during the postnatal period and soon fall behind their littermates in size and weight (Fig. 3). While they are of normal size and weight at birth, by d15 and d30 they are significantly smaller (Fig. 3A). oc2−/− mice are consistently 25–30% smaller (P < 0.0001) than their oc2+/+ and oc2+/− littermates at weaning on d19 (Fig. 3B). This difference in size persists into adulthood. In addition, oc2−/− mice show diminished survival as well. As many as 25–30% of oc2−/− mice do not survive to weaning (depending on the genotype of the parents) (Fig. 4). The oc2−/− mice that do survive to adulthood are fertile but seem to have reduced fecundity and smaller litter sizes. Adult oc2−/− mice appear to have a normal life span.

Fig. 3.

OC-2 knockout mice are significantly smaller than their littermates. A: visual size comparison of OC-2 knockout (KO) mice to wild-type (wt) littermates at d15 and d30. B: relative weights of 78 pups from oc2+/− × oc2+/− breeding pairs at the time of weaning at d19 compared with average pup weight within a litter: oc2+/+, 1.08 ± 0.08; oc2+/−, 1.06 ± 0.14; oc2−/−, 0.78 ± 0.06. The average relative weight of oc2−/− mice differed from that of both oc2+/+ and oc2+/− mice, with a relative significance of P < 0.0001 in both cases.

Fig. 4.

oc2−/−;oc3+/− mice are further reduced in size and have increased mortality compared with OC-2 knockout mice. A: surviving pups from (oc2−/−) × (oc2+/−;oc3+/−) breeding pairs were weighed at weaning, and their relative weight compared with the average weight of pups in the litter was determined. B: survival rate for 136 pups born to the (oc2−/−) × (OC2+/−;OC3+/−) breeding pairs was determined at weaning.

It has been reported previously that oc2−/−;oc3−/− mice die at birth for unknown reasons (their intestines develop normally) (34). We confirmed that this genotype is a perinatal lethal but investigated the effect of breeding a single oc3− allele onto the oc2−/− background. This was accomplished by breeding oc2−/− mice to oc2+/−;oc3+/− mice. The resulting oc2−/−;oc3+/− mice are even smaller than oc2−/− mice and have an even higher mortality rate (Fig. 4). They are approximately half the size of normal mice at d19 (weaning), and less than half of them survive to weaning and maturity. Therefore loss of one functional oc3 allele exacerbates the developmental problems that oc2−/− mice have during the postnatal period. The oc2−/−;oc3+/− mice that survive to maturity are fertile but not very productive as parents. Interestingly, the double heterozygous mice, oc2+/−;oc3+/−, show a slight, but significant (P < 0.03), decrease in size as well, compared with the oc2+/− mice from the same litters.

No compensatory expression of OC factors in Onecut-2 knockout mice.

The three mammalian Onecut factors are all relatively homologous with very similar DNA binding domains and binding site recognition (33). Therefore one must always consider the possibility of redundant function in areas of coexpression and compensatory or ectopic expression in those cases in which the expression of one family member is ablated. To begin to assess this possibility in the postnatal OC-2 knockout intestine, we used real-time RT-PCR to examine both OC-3 and OC-1/HNF-6 expression in the postnatal duodenum of wild-type and oc2−/− mice. OC-3 does not show compensatory expression in the duodenum of oc2−/− mice (Fig. 5). Instead, OC-3 mRNA shows a moderate decrease in expression levels in d15 duodenum (P ≤ 0.007) of oc2−/− mice compared with wild type. However, by d30 no difference is observed. OC-1/HNF-6 does not display ectopic or compensatory expression in the duodenum, either (data not shown). These results agree with those described previously for embryonic and newborn oc2−/− mice (34), which were shown to lack compensatory effects from other Onecut factors.

Fig. 5.

OC-3 expression levels in duodenal samples from d15 and d30 OC-2 knockout and wild-type mice. Relative levels were determined by real-time RT PCR on 3 independent total RNA samples from both OC-2 knockout and wild-type mice. Each total RNA sample was isolated from the combined tissue of 5–8 duodena. *d15 wild-type and OC-2 knockout samples differ significantly with a P value of ≤0.007.

Onecut-2 knockout mice show dramatic alterations in gene expression patterns in proximal small intestine.

Next we sought to determine the overall effect of OC-2 gene ablation on the levels and patterns of gene expression in the small intestine. Since the highest levels of OC-2 expression are observed in the proximal small intestine, we compared mRNA levels in duodenum of wild-type and oc2−/− mice by microarray analysis using the Affymetrix platform (see materials and methods for details). The samples were analyzed as biological triplicates, with tissue samples from five to eight mice in each replicate group, at both d15 and d30. Analysis of two postnatal developmental time points allowed us to produce an atlas of temporal/developmental changes in gene expression in the normal duodenum as well identifying temporal specific changes caused by the OC-2 gene ablation.

This approach produced very reproducible and consistent results, which in turn allowed reliable detailed analysis of the results. In the wild-type mice, 618 genes showed consistent changes of twofold or greater (P ≤ 0.05) between d15 and d30 (Fig. 6). This represented <3% of the 21,814 genes on the Affymetrix microarray chip. Among this group of genes, 63 genes showed alterations of twofold or greater in the OC-2 knockout mice at d15, d30, or both (Fig. 6A). In addition, 45 genes that normally did not change significantly between d15 and d30 were altered by twofold or greater in the OC-2 knockout mice, at one or both developmental time points. Figures 7 and 8 show heatmaps and cluster analyses that visually summarize the major changes that occur in gene expression in the duodenum of oc2−/− mice. Of perhaps greatest interest, a high percentage of the 108 genes with altered expression of twofold or greater in OC-2 knockout mice have products integral to or associated with the epithelial membrane, including many involved in transport and metabolism (Fig. 6B).

Fig. 6.

Microarray comparison of duodenal RNA from wild-type and OC-2 knockout mice at d15 and d30. RNA replicate samples for microarray analysis were independent biological triplicates, with 5–8 mice/replicate. A: samples included in this Venn diagram analysis had reproducible gene expression differences of ≥2-fold (P ≤ 0.05). B: there is functional overlap between the genes included in the lists for membrane, transport, and metabolism.

Fig. 7.

Gene expression patterns in postnatal maturing duodenum samples that are altered by the loss of OC-2. Gene expression patterning is depicted across the series of duodenum samples derived from d15 and d30 wt and oc2−/− mice. After robust multichip average (RMA) normalization, relative expression of any probe set in any sample was transformed to a ratio relative to the mean of its expression in the 6 different d15 and d30 wild-type samples. Probe sets were filtered first for level of detection on any 3 of the arrays from any of the 4 experimental groups exceeding RMA value 6.0 and then those that exhibited statistical and fold different expression in wild type vs. oc2−/− at d15 and/or d30. Probe sets were pooled and subjected to hierarchical clustering using the Standard Correlation similarity measure and the Average Linkage clustering algorithm. A: expression pattern clusters of 763 probe sets with P < 0.05 and fold difference >1.3 between wild-type and oc2−/− mice at either d15 or d30. B: expression pattern clusters of 129 probe sets with P < 0.05 and fold difference >2 between wild-type and oc2−/− mice at either d15 or d30. This probe set number differs from the number of genes indicated in Fig. 6B (108 genes) because several of these genes are represented by >1 probe set in the array.

Fig. 8.

Cluster-averaged K-means clusters of OC-2-regulated genes across the 4 sample types. The 763 probe sets from Fig. 7A were subjected to K-means clustering and divided into 5 groups with Pearson correlation. The average relative expression of each cluster is shown to depict the overall pattern that each gene cluster follows. The numbers of genes in each cluster are: cluster 1, 212; cluster 2, 144; cluster 3, 161; cluster 4, 112; cluster 5, 134.

The results provide support for OC-2 having an integral role in small intestinal gene regulation. A significant number of genes with altered levels of expression in oc2−/− mice demonstrated either a consistent generalized activation or generalized repression at both d15 and d30 (for examples, see Fig. 9). Some of these alterations in expression represented relatively dramatic changes in level (i.e., the decreases observed in serine protease 7 and α-fetoprotein levels). It is, of course, uncertain at this time whether the changes observed imply a direct or indirect role for OC-2 in regulating their expression.

Fig. 9.

Altered patterns of gene expression observed in the microarray comparison of duodenal RNA from wild-type and OC-2 knockout mice at d15 and d30. These examples are genes that demonstrate either generalized repression (top) or generalized activation (bottom) at both d15 and d30, caused by ablation of the OC-2 gene.

As mentioned above, we have previously proposed a temporal role for OC-2 in intestinal gene regulation during the postnatal developmental period (24). A number of the genes with perturbed expression in oc2−/− mice display temporal specific alterations in their expression patterns (see Fig. 10). The majority of these represent premature repression at d15 of a gene that is normally significantly repressed at d30 compared with d15 (Fig. 10, top). In addition, some genes display delayed repression at d15 or temporal specific activation at d30 (Fig. 10, bottom). These results support an important role for OC-2 in orchestrating the levels and timing of gene expression during the critical postnatal developmental period.

Fig. 10.

Temporal alterations in patterns of gene expression in the microarray comparison of duodenal RNA from wild type and OC-2 knockout mice at d15 and d30. These examples display premature repression (top), delayed repression (bottom left), or temporal specific activation (bottom right) at d30, resulting from the ablation of the OC-2 gene.

As shown in Fig. 6, the ablation of the OC-2 gene displays its effects on a discrete subgroup of genes expressed in the small intestine. As described above, many of these genes encode products that are integral to or associated with the epithelial membrane. We used the GeneSpring suite of analytical software (Agilent Technologies) to analyze the group of affected genes, looking for biological networks. This analysis of our initial microarray study indicates that the process and pathways of lipid metabolism and transport are dramatically altered in the OC-2 knockout d15 duodenum (Table 1). The reduction of all of these genes in concert is indicated as highly significant (P ≤ 0.00001) by the analytical programs used, with potentially very significant impact on lipid utilization by the proximal small intestine at d15. This aberration is temporal specific and is no longer evident at d30.

View this table:
Table 1.

Lipid-related genes whose expression is decreased in d15 OC-2 knockout mice


The Onecut family of transcription factors has two conserved domains involved in DNA binding, a cut domain that is located amino-terminal to a novel type of homeodomain. The initial identification of the cut domain was made in the Drosophila melanogaster Cut protein, which has three copies of the cut domain amino terminal of the homeodomain (3). The gene that encodes this Cut protein was the founding member of the cut superclass of homeobox genes (4). The mammalian Onecut (OC) family is comprised of a relatively small number of genes (OC-1, -2, and -3). However, little is known about their role and function in a number of tissues in which family members are expressed, including intestine. The founding member, OC-1, was designated as HNF-6 when it was originally identified (18, 21). In initial adult rodent expression studies, OC-1/HNF-6 was found in liver, pancreas, brain, spleen, and testis but not in a number of other tissues, including small intestine (21, 33). Our results confirm that OC-1/HNF-6 is essentially absent from mouse postnatal small intestine but identified moderately high expression of OC-1/HNF-6 in Brunner's glands in the most proximal part of the adult mouse small intestine.

Most of the developmental and functional studies for OC-1/HNF-6 have focused on liver and pancreas, and they have shown that this factor is a key regulator in both tissues (6, 7, 15, 22, 25, 28, 29). OC-1/HNF-6 functions as a transcriptional activator in the regulation of several genes expressed in the liver and pancreas (13, 17, 1921, 26). In a study that is especially pertinent to our Onecut temporal studies, OC-1/HNF-6 was shown to be required for time-specific expression of the glucose-6-phosphatase gene during liver development (2). Activation of this late marker of hepatocyte maturation was delayed by several days in OC-1/HNF-6 knockout mice. OC-1/HNF-6−/− mice survive to birth, but 75% of them die between d1 and d10 (probably of liver failure) (13). Those who survive to adulthood are diabetic.

The critical roles of OC-1/HNF-6 in the endodermal derivatives liver and pancreas provided the impetus to investigate the potential roles of OC-2 and OC-3 in endodermally derived tissues, once these factors were identified. OC-2 and OC-3 were both discovered by homology to OC-1/HNF-6 (14, 16, 33). They have Onecut and homeodomain regions highly similar to OC-1/HNF-6, with similar if not identical DNA recognition and binding preferences. OC-2 is tissue restricted in adult rodents, with significant expression in liver, brain, and small intestine (but not pancreas) (14, 24, 33, 34). Significant adult expression of OC-3 is observed only in small intestine, stomach, and brain (but not pancreas or liver) (14, 24, 34). The presence of both OC-2 and OC-3 in small intestine, where OC-1/HNF-6 is absent, led us to investigate their role there. We observed the highest intestinal levels of both OC-2 and OC-3 mRNA in the proximal small intestine throughout the postnatal period examined (d2–d30) (Fig. 1). The predominance of OC-2 expression over OC-3 expression led us to investigate first the role of OC-2 in intestinal gene regulation. We chose to use microarray technology in those initial studies.

Because we had previously identified (24) a role for OC-2 in the temporal regulation of the human ADA gene, we examined gene expression by microarray analysis at two different time points during postnatal development, d15 and d30. The higher levels of OC-2 expression in the proximal small intestine led us to focus initially on duodenum. One important auxiliary benefit of the study conducted in this manner was generation of a catalog of postnatal changes in gene expression levels in normal duodenum. Of the total list of 21,814 genes surveyed, <3% (618 genes) showed significant, reproducible changes of twofold or greater. This information will be valuable as we begin to try to understand the genetic basis for the important functional and metabolic changes that occur during the final stages of development of the small intestine.

Examination of the effect of OC-2 gene ablation on gene expression at d15 and d30 provides strong support for an important role for OC-2 in orchestrating the changes in gene expression that occur during this critical developmental period (see Figs. 7 and 8). Of the 618 genes that normally show significant changes in expression between d15 and d30, >10% (63 genes) show significant changes at one or both time points in oc2−/− mice (Fig. 6). In a number of cases, the result is premature repression at d15 of genes that are normally downregulated in duodenum at d30 (Fig. 10). Some of the changes in gene expression, such as delayed repression (Fig. 10), could be explained by the overall delay in development of the oc2−/− mice, but many, i.e., those showing premature repression and temporal specific activation in Fig. 10, cannot.

The microarray data also demonstrate a regulatory role for OC-2 in genes that normally do not show developmental changes in expression. There were 45 genes of this type that showed changes of twofold or greater at d15, d30, or both time points in the oc2−/− mice (Fig. 6). Five of the eleven genes that showed alterations of twofold or more at both time points are shown in Fig. 9. The results summarized by the examples shown in Fig. 9 indicate that OC-2 can have a positive or negative influence on gene expression, since removal of functional OC-2 causes a generalized decrease or increase in expression for particular genes. At least one of these genes has been reported previously to be directly regulated by Onecut factors. OC-1/HNF-6 has been shown to be a direct activator of α-fetoprotein gene expression in liver cells (26). Our results indicate that OC-2 may have a similar role in small intestine.

One of the more striking observations to come out of the microarray studies is that a high percentage of the genes with significantly altered expression patterns are associated with or integral to the membrane. Many of these are involved with transport and/or metabolism of various types of ions and metabolites. Perhaps the most striking changes are observed in genes involved in lipid metabolism and transport at d15 (Table 1). Network, process, and pathway analysis indicate that these changes in gene expression are extremely likely to have dramatic effects on the ability of the duodenum to process and utilize lipids. Lipid metabolism, lipid absorption, and lipid homeostasis are all predicted to be altered significantly.

Thus the microarray studies provide support for OC-2 having important roles in regulating gene expression in proximal small intestine during the postnatal developmental period. One role may be a temporal role in orchestrating the timing of changes in gene expression that are integral to establishing the mature function of the intestine. This may be especially true of transport and metabolic capability, including lipid metabolism and homeostasis.

The other major observation to come out of our studies is the failure to thrive of the oc2−/− mice during the postnatal period. We observed a significant reduction in size and weight, as well as increased mortality. The reduction in size is extremely consistent from individual to individual among the oc2−/− mice. This reduction in size persists into adulthood, where the surviving oc2−/− mice remain smaller than average throughout life. They appear to have a normal life span once they survive the critical developmental period through weaning. Adult oc2−/− mice are fertile, although they are somewhat diminished in their fecundity, being somewhat delayed in producing litters and producing smaller litters in general (data not shown).

Any relationship between alterations in small intestinal gene expression and the diminished size and survival in the oc2−/− mice is strictly speculative and unproven at this point in time. Significant additional studies will be required to determine whether the types of alterations in intestinal gene expression observed might have any impact on growth and survival for the oc2−/− mice. It will be critical to evaluate the metabolic consequences of the alterations in gene expression already observed. However, it is clear that loss of one OC-3 allele greatly exacerbates the growth and survival problems encountered by oc2−/− mice. This is further emphasized by the fact that oc2−/−;oc3−/− mice demonstrate perinatal lethality. The basis for this combinatorial effect when both genes are ablated remains to be determined. It could represent a redundancy of function between OC-2 and OC-3 or the cumulative effect of loss of different functions for the two factors. The answers to these questions, and related ones, will await additional studies. These would include analysis of changes in intestinal gene expression patterns in oc3−/− mice and in oc2−/−;oc3+/− mice. Interpretation of these studies may require a better delineation of the cell-specific expression patterns for OC-2 and OC-3 in intestine.


This work was supported, in part, by National Institute of Diabetes and Digestive and Kidney Diseases Grants RO1-DK-52343 (D. A. Wiginton) and P30-DK-078392.


No conflicts of interest, financial or otherwise, are declared by the author(s).


We sincerely appreciate the gift of the oc2−/− and oc3−/− mouse strains from Dr. Patrick Jaquemin and Dr. Frederique Lemaigre (Hormone and Metabolic Research Unit, Université Catholique de Louvain, Brussels, Belgium) without which these experiments would not have been possible. We appreciate the technical assistance of Bhuvana Sakthivel in the analysis of microarray results.


  • 1 The online version of this article contains supplemental material.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
View Abstract