Final Report Summary - PREPOBEDIA (Novel prep1-dependent transcriptional networks in the control of insulin sensitivity)
Prep1 is a developmentally essential gene but 25 % of the homozygous hypomorphic mutant mice are born and display increased insulin sensitivity and are protected from streptozotocin induced diabetes.
PREPOBEDIA aimed at answering the following questions:
1. elucidate the molecular function of prep1 in insulin sensitivity and the possible connection with adipogenesis
2. dentify prep1 target genes and partners in the mouse and test their role in animal or cellular models
3. study the correlation between prep1 murtations and human type two diabetes mellitus (T2DM) and obesity
4. understand the reasons why the answer to question three was in fact negative.
The results of the work from the participating laboratories could be summarised as follows:
1. Prep1 acted on insulin sensitivity in different ways depending on the tissue. In particular, in muscle and liver the lack of prep1 led to the same result through two different mechanisms. In the muscle, the absence of prep1 destabilised p160MBP1a, an inhibitor of PGHC1a which therefore resulted in the activation of PGC1a, the overexpression of genes required for insulin sensitivity. In the liver, the absence of prep1 reduces also its partner pbx1 and this blocked the expression of the tyrosine phosphatase shp2. Adipogenesis was also connected with prep1, since both lacked and overexpression of prep1 led to a major change in triglyceride synthesis. In this case, the targets of prep1 were the PPARgamma genes.
2. Sequential ribonucleic acid (RNAseq) and ChIPseq analysius of mouse embryos identified the totality of the genes that were bound by prep1 and analysed the effect of the absence or mutation of the gene on their expression. In addition we had the opportunity to compare the results with prep1 with those on meis1, a homolog of prep1. The results showed that prep1 affected in particular basic cellular mechanisms while meis1 affected in particular developmental genes. Nevertheless, meis1 and prep1 co-ordinately regulated a set of important genes, like the Hox genes. The ChIPseq comparison brought out major differences between prep and meis, identified the binding consensus for the different dimers and the different mechanism of action of prep-pbx versus meis-pbx. The first in fact acted at the promoter level while the second at extra-promoter sites.
3. Despite the above results, large scale analysis of human genetic analysis failed to identify prep1 gene as a putative cause of at least some cases of T2DM or obesity. The reason why prep1 was not connected to the diseases was that prep1 had additional, more essential functions which obscured therefore downstream effects. Indeed, cells totally or partially missing prep1 tended to go into apoptosis, more or less prominent depending on the type of cells, since they accumulated deoxyribonucleic acid (DNA) damage and chromosomal abnormalities. This resulted in cells that were more easily transformed as they were capable of bypassing oncogene induced senescence.
As already mentioned, the project studied a few potentially important genes for which preliminary information obtained by the partners of this project strongly indicated their potential participation in the determination of T2DM and obesity.
The gene prep1 was chosen as main player because of its interaction with pbx1, a gene that in turn genetically and molecularly interacted with Isl1 (Pdx1), an essential gene in pancreas development and function. Pbx1 and prep1 were homeodomain transcription factor, members of the same gene family, which dimerized to form transcriptionally active complexes. Since prep1 was the main partner of pbx1 in many molecular activities, it might be considered another potential diabetes associated gene.
Therefore the project was centred on prep1 and was organised in five research groups collaborating in four work packages (WPs). Four groups had an academic background while the fifth was a small and medium sized enterprise (SME) whose interest specifically included diabetes and obesity. The overall objectives were:
1. elucidating the molecular basis of prep1, p160MBP and PGC1a function in insulin sensitivity and adipogenesis
2. providing novel prep1 based animal models to study the role of prep1 in disease
3. identifying prep1 dependent or prep1 related target genes in the mouse
4. determining the presence of prep1, p160MBP and PGC1a RFLP and their association with human T2DM and metabolic syndrome (MS).
The first important result was the elucidation of the molecular basis of the phenotype of the prep1 hypomorphic mice, which displayed increased insulin sensitivity and were protected from streptozotocin induced diabetes. The effect of prep1 on PGC1a expression was indirect and it was due to the direct effect of prep1 in the muscle on transcription factors that influenced PGC1a expression. On the other hand, it became soon apparent that the above mechanism applied to skeletal muscle glucose metabolism and not to the liver. In this organ, in fact p160MBP did not interact efficiently with prep1, which rather interacted with pbx1.
The second important result was the effect of prep1 on lipid metabolism. In order to generate these data we had to generate muscle specific prep1 KO mice. These animals showed no differences in body weight and body composition as measured by quantitative nuclear magnetic resonance (NMR). However, transcriptional profiling of skeletal muscle samples from double heterozygous animals revealed significant differences in transcripts of enzymes and proteins controlling oxidative phosphorylation as well as glycogen metabolism. The importance of these transcriptional changes in physiological consequences was validated through the measure of the endurance of mice on a motorized treadmill. This revealed a significant increase in the endurance of dihydrotestosterone (DHT) as compared to wild type mice. This phenotype was further defined measuring the maximal oxidative capacity on a metabolic treadmill coupled with an oxygen sensor. Consistent with the increased endurance we found increased maximal oxygen consumption (VO2max) in the DHT mice. These findings coupled with the observation that all complexes of the respiratory chain showed a decreased expression in our microarray analysis strongly suggested differences in the mitochondrial properties of the double heterozygous mice.
The third important result of this project was the identification of basic, non tissue specific functions of prep1 which might affect, either obscure or enhance, the role of prep1 on the glucose and lipids phenotype. Cells totally missing prep1 went into apoptosis because of the accumulation of DNA damage and chromosomal abnormalities. Indeed, the comparison of wt and Prep1i/i mouse embryonic fibroblasts (MEFs) showed that these cells underwent apoptosis, in particular under conditions of DNA damage caused by irradiation or specific DNA damaging drugs. This effect was due to a complex regulation exerted by prep1 on the apoptosis pathway. The Prep1-p53 connection was also validated in vivo. The phenotype was caused by a specific decrease of the epiblast cells which were shown to undergo massive apoptosis. The apoptosis in the embryo was p53-dependent and in fact a mouse doubly deleted for prep1 and p53 was partially rescued. Apoptosis appeared to be due to the accumulation of DNA damage because the phenotype was exacerbated under conditions in which the level of Atm was decreased or annulled. Hence, the prep1-null phenotype was due to accumulation of DNA damage in the epiblast that caused apoptosis and death of the embryo. DNA damage was an important prerequisite for cancer. Therefore we tested the Prep1i/i mice and observed that a very high percentage of them developed cancer of various kinds. Moreover, we also realised that the haploinsufficiency of prep1 drastically reduced the survival of mice spontaneously developing tumours, like the MNyc-dependent lymphomas in EmMyc mice. These data therefore strongly suggested that prep1 acted as a tumour suppressor. Indeed, we analysed a large number of human cancers for the expression of prep1 and found that it was down-regulated in a very high percent of cases. Therefore we proposed that prep1 was a novel tumour suppressor gene.
The fourth important result was the genome wide identification of prep1 dependent or affected genes. This was particularly difficult in view of the large number of orthologous genes. Since the homeodomain of all these genes was identical within each orthology group we had to devise ways to demonstrate specific and common binding sites. In any case, the results allowed extracting a series of general rules that could be summarised as follows:
1. prep and meis transcription factors bound a mostly different genomic sequence
2. prep and meis acted in different ways, with prep acting on promoters and meis on sites remote from the genomic unit
3. both prep and meis used pbx as a partner; however the presence of pbx did not modify the target sequence.
Because of all this, prep and meis target genes were widely different. Superficially it could be stated that prep affected basic functions of the cells, whereas meis affects in particular embryonic developmental functions. The transcription regulatory function of these genes was evaluated by RNAseq experiments in which the total messenger ribonucleic acid (mRNA) of wild type and Prep1i/i embryos was deeply sequenced. This allowed for the identification of specific genes whose activity was dependent on prep1 or meis1. Despite the mostly non-overlapping target gene landscape of these two transcription factors we observed that there was a certain level of coordination. For example, both prep and meis affected Hox gene expression, but their effect was opposite.
These results represented the basis for a more thorough analysis of the role of prep1 in physiology and pathology. In addition, they were very useful to answer some of the questions arisen from this project.
The final result was the assessment of the genetic association between prep1, p160MBP and PGC1a RFLP to T2DM and obesity. This was tested on over 1 000 carefully selected subjects, using single nucleotide polymorphisms (SNPs). However, the results did not show any statistically significant genetic association with the human diseases. Since this result strikingly opposed all the others that showed a relevant function of in particular prep1 in insulin sensitivity and lipid metabolism, it was very likely that the basic effects of prep1 leading to cancer susceptibility or embryonic death obscured a direct role in T2DM or obesity.
In conclusion, the overall project went extremely well in all its parts. The minor modifications over the original project did not represent a change in the aims of the project or in its basic hypothesis, rather only technical modifications to achieve the same goals. In practice, all objectives were reached.
The results of the project were published in peer reviewed journals. In addition they were presented at several international scientific conferences.
The potential impact of our project was all at the scientific level. It became clear that the genes investigated had also other very important functions in our body and therefore their alteration, i.e. mutation, would affect health in much more drastic terms than diabetes or obesity, hence obscuring their role in these diseases.
Further information on the project could be obtained at http://www.prepobedia.org/(öffnet in neuem Fenster).
PREPOBEDIA aimed at answering the following questions:
1. elucidate the molecular function of prep1 in insulin sensitivity and the possible connection with adipogenesis
2. dentify prep1 target genes and partners in the mouse and test their role in animal or cellular models
3. study the correlation between prep1 murtations and human type two diabetes mellitus (T2DM) and obesity
4. understand the reasons why the answer to question three was in fact negative.
The results of the work from the participating laboratories could be summarised as follows:
1. Prep1 acted on insulin sensitivity in different ways depending on the tissue. In particular, in muscle and liver the lack of prep1 led to the same result through two different mechanisms. In the muscle, the absence of prep1 destabilised p160MBP1a, an inhibitor of PGHC1a which therefore resulted in the activation of PGC1a, the overexpression of genes required for insulin sensitivity. In the liver, the absence of prep1 reduces also its partner pbx1 and this blocked the expression of the tyrosine phosphatase shp2. Adipogenesis was also connected with prep1, since both lacked and overexpression of prep1 led to a major change in triglyceride synthesis. In this case, the targets of prep1 were the PPARgamma genes.
2. Sequential ribonucleic acid (RNAseq) and ChIPseq analysius of mouse embryos identified the totality of the genes that were bound by prep1 and analysed the effect of the absence or mutation of the gene on their expression. In addition we had the opportunity to compare the results with prep1 with those on meis1, a homolog of prep1. The results showed that prep1 affected in particular basic cellular mechanisms while meis1 affected in particular developmental genes. Nevertheless, meis1 and prep1 co-ordinately regulated a set of important genes, like the Hox genes. The ChIPseq comparison brought out major differences between prep and meis, identified the binding consensus for the different dimers and the different mechanism of action of prep-pbx versus meis-pbx. The first in fact acted at the promoter level while the second at extra-promoter sites.
3. Despite the above results, large scale analysis of human genetic analysis failed to identify prep1 gene as a putative cause of at least some cases of T2DM or obesity. The reason why prep1 was not connected to the diseases was that prep1 had additional, more essential functions which obscured therefore downstream effects. Indeed, cells totally or partially missing prep1 tended to go into apoptosis, more or less prominent depending on the type of cells, since they accumulated deoxyribonucleic acid (DNA) damage and chromosomal abnormalities. This resulted in cells that were more easily transformed as they were capable of bypassing oncogene induced senescence.
As already mentioned, the project studied a few potentially important genes for which preliminary information obtained by the partners of this project strongly indicated their potential participation in the determination of T2DM and obesity.
The gene prep1 was chosen as main player because of its interaction with pbx1, a gene that in turn genetically and molecularly interacted with Isl1 (Pdx1), an essential gene in pancreas development and function. Pbx1 and prep1 were homeodomain transcription factor, members of the same gene family, which dimerized to form transcriptionally active complexes. Since prep1 was the main partner of pbx1 in many molecular activities, it might be considered another potential diabetes associated gene.
Therefore the project was centred on prep1 and was organised in five research groups collaborating in four work packages (WPs). Four groups had an academic background while the fifth was a small and medium sized enterprise (SME) whose interest specifically included diabetes and obesity. The overall objectives were:
1. elucidating the molecular basis of prep1, p160MBP and PGC1a function in insulin sensitivity and adipogenesis
2. providing novel prep1 based animal models to study the role of prep1 in disease
3. identifying prep1 dependent or prep1 related target genes in the mouse
4. determining the presence of prep1, p160MBP and PGC1a RFLP and their association with human T2DM and metabolic syndrome (MS).
The first important result was the elucidation of the molecular basis of the phenotype of the prep1 hypomorphic mice, which displayed increased insulin sensitivity and were protected from streptozotocin induced diabetes. The effect of prep1 on PGC1a expression was indirect and it was due to the direct effect of prep1 in the muscle on transcription factors that influenced PGC1a expression. On the other hand, it became soon apparent that the above mechanism applied to skeletal muscle glucose metabolism and not to the liver. In this organ, in fact p160MBP did not interact efficiently with prep1, which rather interacted with pbx1.
The second important result was the effect of prep1 on lipid metabolism. In order to generate these data we had to generate muscle specific prep1 KO mice. These animals showed no differences in body weight and body composition as measured by quantitative nuclear magnetic resonance (NMR). However, transcriptional profiling of skeletal muscle samples from double heterozygous animals revealed significant differences in transcripts of enzymes and proteins controlling oxidative phosphorylation as well as glycogen metabolism. The importance of these transcriptional changes in physiological consequences was validated through the measure of the endurance of mice on a motorized treadmill. This revealed a significant increase in the endurance of dihydrotestosterone (DHT) as compared to wild type mice. This phenotype was further defined measuring the maximal oxidative capacity on a metabolic treadmill coupled with an oxygen sensor. Consistent with the increased endurance we found increased maximal oxygen consumption (VO2max) in the DHT mice. These findings coupled with the observation that all complexes of the respiratory chain showed a decreased expression in our microarray analysis strongly suggested differences in the mitochondrial properties of the double heterozygous mice.
The third important result of this project was the identification of basic, non tissue specific functions of prep1 which might affect, either obscure or enhance, the role of prep1 on the glucose and lipids phenotype. Cells totally missing prep1 went into apoptosis because of the accumulation of DNA damage and chromosomal abnormalities. Indeed, the comparison of wt and Prep1i/i mouse embryonic fibroblasts (MEFs) showed that these cells underwent apoptosis, in particular under conditions of DNA damage caused by irradiation or specific DNA damaging drugs. This effect was due to a complex regulation exerted by prep1 on the apoptosis pathway. The Prep1-p53 connection was also validated in vivo. The phenotype was caused by a specific decrease of the epiblast cells which were shown to undergo massive apoptosis. The apoptosis in the embryo was p53-dependent and in fact a mouse doubly deleted for prep1 and p53 was partially rescued. Apoptosis appeared to be due to the accumulation of DNA damage because the phenotype was exacerbated under conditions in which the level of Atm was decreased or annulled. Hence, the prep1-null phenotype was due to accumulation of DNA damage in the epiblast that caused apoptosis and death of the embryo. DNA damage was an important prerequisite for cancer. Therefore we tested the Prep1i/i mice and observed that a very high percentage of them developed cancer of various kinds. Moreover, we also realised that the haploinsufficiency of prep1 drastically reduced the survival of mice spontaneously developing tumours, like the MNyc-dependent lymphomas in EmMyc mice. These data therefore strongly suggested that prep1 acted as a tumour suppressor. Indeed, we analysed a large number of human cancers for the expression of prep1 and found that it was down-regulated in a very high percent of cases. Therefore we proposed that prep1 was a novel tumour suppressor gene.
The fourth important result was the genome wide identification of prep1 dependent or affected genes. This was particularly difficult in view of the large number of orthologous genes. Since the homeodomain of all these genes was identical within each orthology group we had to devise ways to demonstrate specific and common binding sites. In any case, the results allowed extracting a series of general rules that could be summarised as follows:
1. prep and meis transcription factors bound a mostly different genomic sequence
2. prep and meis acted in different ways, with prep acting on promoters and meis on sites remote from the genomic unit
3. both prep and meis used pbx as a partner; however the presence of pbx did not modify the target sequence.
Because of all this, prep and meis target genes were widely different. Superficially it could be stated that prep affected basic functions of the cells, whereas meis affects in particular embryonic developmental functions. The transcription regulatory function of these genes was evaluated by RNAseq experiments in which the total messenger ribonucleic acid (mRNA) of wild type and Prep1i/i embryos was deeply sequenced. This allowed for the identification of specific genes whose activity was dependent on prep1 or meis1. Despite the mostly non-overlapping target gene landscape of these two transcription factors we observed that there was a certain level of coordination. For example, both prep and meis affected Hox gene expression, but their effect was opposite.
These results represented the basis for a more thorough analysis of the role of prep1 in physiology and pathology. In addition, they were very useful to answer some of the questions arisen from this project.
The final result was the assessment of the genetic association between prep1, p160MBP and PGC1a RFLP to T2DM and obesity. This was tested on over 1 000 carefully selected subjects, using single nucleotide polymorphisms (SNPs). However, the results did not show any statistically significant genetic association with the human diseases. Since this result strikingly opposed all the others that showed a relevant function of in particular prep1 in insulin sensitivity and lipid metabolism, it was very likely that the basic effects of prep1 leading to cancer susceptibility or embryonic death obscured a direct role in T2DM or obesity.
In conclusion, the overall project went extremely well in all its parts. The minor modifications over the original project did not represent a change in the aims of the project or in its basic hypothesis, rather only technical modifications to achieve the same goals. In practice, all objectives were reached.
The results of the project were published in peer reviewed journals. In addition they were presented at several international scientific conferences.
The potential impact of our project was all at the scientific level. It became clear that the genes investigated had also other very important functions in our body and therefore their alteration, i.e. mutation, would affect health in much more drastic terms than diabetes or obesity, hence obscuring their role in these diseases.
Further information on the project could be obtained at http://www.prepobedia.org/(öffnet in neuem Fenster).