Metabolic Profiling of Multiorgan Samples: Evaluation of MODY5/RCAD Mutant Mice

Abstract

In the present study, we performed a metabolomics analysis to evaluate a MODY5/RCAD mouse mutant line as a potential model for HNF1B-associated diseases. Gas chromatography time-of-flight mass spectrometry (GC-TOF-MS) of gut, kidney, liver, muscle, pancreas, and plasma samples uncovered the tissue specific metabolite distribution. Orthogonal projections to latent structures discriminant analysis (OPLS-DA) was used to identify the differences between MODY5/RCAD and wild-type mice in each of the tissues. The differences included, for example, increased levels of amino acids in the kidneys and reduced levels of fatty acids in the muscles of the MODY5/RCAD mice. Interestingly, campesterol was found in higher concentrations in the MODY5/RCAD mice, with a four-fold and three-fold increase in kidneys and pancreas, respectively. As expected, the MODY5/RCAD mice displayed signs of impaired renal function in addition to disturbed liver lipid metabolism, with increased lipid and fatty acid accumulation in the liver. From a metabolomics perspective, the MODY5/RCAD model was proven to display a metabolic pattern similar to what would be suspected in HNF1B-associated diseases. These findings were in line with the presumed outcome of the mutation based on the different anatomy and function of the tissues as well as the effect of the mutation on development.

Hnf1b haploinsufficiency differentially affects developmental target genes in a new renal cysts and diabetes mouse model

“…Metabolic profiling of different organs, including kidney, pancreas and liver, as well as plasma of adult Hnf1 Sp2/+ mice, was previously reported ( Torell et al, 2018 ). We found evidence of impaired amino acid renal metabolism and reduced plasma levels of total free amino acids and increased myo-inositol, which are metabolic parameters reflecting impaired renal function, together with disturbed hepatic metabolism.…”

Application of Multiblock Analysis on Small Metabolomic Multi-Tissue Dataset

“…We considered JIVE (joint and individual variation explained), DISCO-SCA (distinctive and common components with simultaneous-component analysis), and JUMBA (joint and unique multiblock analysis; a method based on OnPLS) [ 31 ]. From our previous research, we have found the dataset to contain the locally joint variation since the gut was not as affected by the mutation [ 25 ], therefore we were interested in methods that can handle both globally joint and locally joint, as well as unique variation. DISCO-SCA and JUMBA are both able to extract all three kinds of variations while JIVE is only able to extract the globally joint and unique variation.…”

“…The MODY5/RCAD mice were found to be characterized by higher levels of amino acids and fatty acids in the organs (kidney, liver, muscle, and pancreas) and lower levels of amino acids in the plasma. It has previously been shown that the gut was less affected by the mutation than the other studied tissue types [ 25 ]. This is in part due to the fact that the mutation affects a protein, hepatocyte nuclear factor 1b (hnf1b).…”

“…This can apply in case you are working with expensive samples or a pilot study. Our example dataset contains a small metabolomic multiblock dataset (six different tissue types; gut, kidney, liver, muscle, pancreas, and plasma) that has been previously studied using different approaches [ 24 , 25 ]. Samples were obtained from wild type mice and mice with a human point mutation in the locus of HNF1B , referred to as maturity onset diabetes of the young 5/renal cyst and diabetes syndrome (MODY5/RCAD).…”

Application of Multiblock Analysis on Small Metabolomic Multi-Tissue Dataset

“…We considered JIVE (joint and individual variation explained), DISCO-SCA (distinctive and common components with simultaneous-component analysis) and JUMBA (joint and unique multiblock analysis; a method based on OnPLS) [31]. From our previous research we have found the dataset to contain locally joint variation since gut was not as affected by the mutation [25], therefore we were interested in methods that can handle both globally joint, locally joint and unique variation. DISCO-SCA and JUMBA are both able to extract all three kinds of variation while JIVE is only able to extract globally joint and unique variation.…”