C-terminal truncation of Pik3r1 in mice models human lipodystrophic insulin resistance uncoupled from dyslipidemia

Heterodimeric class IA phosphatidylinositol-3-kinases (PI3K) transduce signals from many receptor tyrosine kinases including the insulin receptor. PI3K recruitment to phosphotyrosines is mediated by Pik3r1 gene products including the most intensely studied PI3K regulatory subunit, p85α, which also binds and regulates the PIP3 phosphatase Pten, and the lipogenic transcription factor Xbp1. Mutations in human PIK3R1 cause SHORT syndrome, featuring lipodystrophy and severe insulin resistance which, uniquely, are uncoupled from fatty liver and dyslipidemia. We describe a novel mouse model of SHORT syndrome made by knock in of the Pik3r1 Y657X mutation. Homozygous embryos die at E11.5, while heterozygous mice exhibit pre-and postnatal growth impairment with diminished placental vascularity. Adipose tissue accretion on high fat feeding was reduced, however adipocyte size was unchanged and preadipocyte differentiation ex vivo unimpaired. Despite severe insulin resistance, heterozygous mice were hypolipidemic, and plasma adiponectin, liver weight, cholesterol, glycogen and triglyceride content were unchanged. Mild downregulation of lipogenic Srebp1, Srebp2 and Chrebp transcriptional activity but no suppression of Xbp1 target genes was seen after fasting. These findings give new insights into the developmental role of Pik3r1, and establish a model of lipodystrophic insulin resistance dissociated from dyslipidemia as seen in SHORT syndrome.


11
The critical importance of Class 1A PI3K in human physiology is manifest in

9
PIK3R1 gene products have also been suggested to have signaling roles 10 beyond stabilising PI3K and conferring recruitability to phosphorylated YMXM motifs.

11
Pertinent to metabolic homeostasis, for example, is the ability of p85 in mice to bind

15
In the face of this complexity, the discovery that missense or nonsense 16 mutations in the C-terminal SH2 domain of PIK3R1, which affect all three protein 17 products of the gene, produce SHORT syndrome (21)(22)(23), was of great interest.

18
SHORT syndrome is named according to visible dysmorphic features (short stature,

19
hyperextensibility of joints and/or hernia, ocular depression, Rieger anomaly of the 20 iris, and teething delay), however it is the associated metabolic disorders that are 21 potentially more informative about mechanisms of pandemic metabolic disease.

22
Most patients with SHORT syndrome show insulin resistance that is often very 23 severe, and partial lipodystrophy is also common (24). Highly unusually,  targeting, and these cells were used to generate founder heterozygous knock-in 20 mice (Supplemental Figure S1). Immunoblotting of insulin-responsive liver, 21 skeletal muscle and adipose tissue confirmed the presence of a truncated p85 22 gene product, which in general was more highly expressed than full length, wild type 23 protein in the same tissues, most likely due to loss of a previously described C-  Figure S1).

2
Homozygosity for the Pik3r1 Y657X allele was lethal in utero, with no 3 homozygous embryos identified beyond E11.5. At E11.5 homozygous embryos 4 were smaller, with poorly developed limb buds and reduced eye pigmentation 5 ( Figure 1A), while heterozygous embryos were smaller from E15.5 onwards 6 (Supplemental Figure S2). No difference in mass of wild type and heterozygous 7 placentas examined from wild type dams was seen, however vascularisation of the 8 placental exchange region was severely compromised, with around a 40% reduction 9 in vessel density, volume and length at E15.5 (Supplemental Figure S3), as 10 reported previously for placentas heterozygous for a kinase dead p110 catalytic 11 subunit (28). In contrast to that model, however, the size of the placental exchange 12 region (both volume and surface area) and the thickness of the diffusion barrier were 13 normal (Supplemental Figure S3).
14 Pik3r1 WT/Y657X mice were born at expected Mendelian frequency, however they 15 showed impaired linear growth on a chow diet, with body length of males 6% less 16 than wild-type littermates at 18 weeks old (Figures 1B,C), and bodyweight 17% less 17 ( Figure 1D). Overall body composition assessed by TD-NMR showed no difference 18 between heterozygous males and controls ( Figure 1E), and, in keeping with this, no 19 difference was found in epididymal or inguinal white adipose depot nor interscapular 20 brown adipose depot weights when the reduced bodyweight of the male 21 heterozygous mice was taken into account ( Figure 1F-H 15.9±3.9 g/L on ad libitum feeding (n=13,13; not significant)). Liver weights were 25 also indistinguishable ( Figure 1I), however hearts were significantly heavier in 26 Pik3r1 WT/Y657X mice ( Figure 1J) whether analysed with respect to lean mass or whole 27 body weight by ANCOVA. No significant difference in either food consumption or 1 energy expenditure was seen at 16 weeks old on chow diet (Figure 1K,L). A similar 2 pattern of differences was seen between heterozygous and wild-type female mice for 3 all variables assessed in both sexes (Supplemental Figure S4)    28 1 content was significantly higher than in controls (Supplemental Figure S7).

2
Normalised plasma leptin concentrations, in contrast, were lower in heterozygous 3 mice than controls (Supplemental Figure S7).

4
To evaluate insulin sensitivity in more detail, hyperinsulinemic euglycemic

23
No difference was seen in liver triglyceride content between heterozygous and

17
In the previous study of Pik3ca KD

11
Targeted cells were injected into Bl/6J blastocysts to generate chimeras which were 12 bred onto the C57Bl/6J background to create the mutant strain (Supplemental 13 Figure S1).           Table S3 5 Imaging and quantification of immunoblots were undertaken using the BioRad image 6 system.

7
For RNA expression studies total RNA was extracted by RNeasy kits (Qiagen).
8 500 ng of total RNA was used for cDNA synthesis with the ImProm-II reverse

16
We would also like to thank Mr Gregory Strachan at the Imaging Core, which is            shown relative to total lean mass, and were analysed statistically by ANCOVA. **** 25