Using a diet induced obese mouse model we provide an analysis elucidating the effect of paternal obesity on male offspring. We used observations of parental and offspring’s phenotypes to guide transcriptome data interpretation towards understanding the transcriptomic mechanisms of obesity in male mice. Eight F0 and 50 F1 individuals were included for growth curve analysis, and transcriptome array analyses among male offspring revealed two of 18 DEG contrasts to be strongly defined. In BAT, the relevant contrast was that between both parents and a father alone being obese (Fig. 3a). In L the relevant contrast appeared to be between both parents being obese against no parents being obese (Fig. 3b).
Based on the expression pattern of the most prominent DEGs in BAT, male offspring with both parents being obese, when compared to males with only a father being obese, may exhibit a higher propensity to insulin resistance, impaired muscle growth and regeneration, and possibly altered impulse control, while exhibiting expression patterns possibly congruent with enabled physiological mechanisms protective against diet induced obesity. Firstly, Beta-thymosin as produced by TMSB15b2 (Thymosin ß 15b2) are primarily actin monomer sequestering proteins38, and hence in conjunction with ADAM12 activity (encoding disintegrin and metallopeptidase domain 12, a marker of skeletal muscle regeneration39) may suggest an alteration of muscle growth among offspring to obese parents. Secondly, while the function of SERF1 (small EDRK-rich factor 1) so far appears poorly understood, SERF2 knock-out in mice was recently reported to result in unexpected male-specific phenotypes in startle response and pre-pulse inhibition40, so that it is conceivable that SERF1 has a similar effect, even if here only investigated in L. Furthermore, PLIN5 (encoding perilipin 5) influences lipid droplet's biogenesis, structure and degradation, processes which themselves keep balance between lipolysis and lipogenesis. Likewise, PLIN5 deletion related to insulin resistance in muscle 41–43, so that its downregulation, as observed in male offspring to obese mice, may have an according effect. SAT1 (spermidine/spermine N1-acetyl transferase 1) encodes for an vital enzyme which catalyses the acetylation of polyamines44, and may be expressed particularly in offspring to obese parents as a protectant against diet induced obesity 45. In summary, our observations of gene expression male offspring’s BAT may support, and extend on, findings by Dahlhoff et al.10 by corroborating an influential role of obese mothers on negative obesity–related outcomes in male offspring.
In liver tissue (L) the contrast between offspring to obese parents and those to not-obese parents appeared defined by expression changes among a range of well-described genes. For examples, in mice, the fatty acid translocase gene CD36 regulates fatty acid metabolism, has been implicated in heart disease, and appears to influence taste perception and dietary fat processing in the intestine46–50. CD36 upregulation in offspring to obese parents may impair their metabolic health negatively, when compared to offspring of none-obese parents. APQ4 encodes for Aquaporin-4, a strict water channel aquaporin not implicated in the metabolic processes related to obesity51. At the same time, Aquaporin-4 plays an outstanding role in mediating water homeostasis throughout the central nervous system, and other fluid transporting tissues52,53 and has been associated with the pathogenesis of CNS injury, oedema and multiple neuro-degenerative diseases54. As it has been shown that HFD maintenance in rats reduced the amount of water consumed55, we believe AQ4 activity may be upregulated in offspring to obese parents to ensure correct water homeostasis also in times of low fluid intake, possibly due to a lower fluid intake of the parents. LY96 encodes for Myeloid differentiation protein 2 (MD2) which plays an important role in lipopolysaccharide-induced innate immunity response and inflammation, and has been shown, when not knocked out, to be causally related with obesity-associated renal inflammatory injury56. LY96 down-regulation in offspring to obese parents is puzzling, and perhaps constitutes pre-emptive protection mechanism against inflammation in an environment of abundant fatty nutrients. Expression increases in ATP5D, encoding for ATP synthases in the mitochondrial F1 complex has been associated with the recovery from post-natal malnutrition in mice57. Again, it appears to be standing to reason that this gene is downregulated in offspring to obese parents, as post-natal malnutrition would not have occurred during offspring nursing period when fed by obese parents. RPL29 plays an important role in protein synthesis and angiogenesis58. Exercise is a powerful driver of physiological angiogenesis59, and a lack thereof may be observable in downregulation of RPL29 associated with an established impaired angiogenic capacity among obese, which in turn may promote obesity-associated diseases such as diabetes and cardiometabolic disease60. RPL29-donwregulation in offspring to obese parents, independent of their own obesity status, also appears to possibly be passed on from parental to filial generation. Considering the observed expression profiles in our analysis, we believe liver tissue in male offspring to obese mice to have an expression pattern in parts mimicking the regulatory landscape of obesity, possibly consistent with observations of Dahlhoff et al.10.
Corresponding to our inspection of individual genes, both KEEG and GO terms in unison, and in part across both BAT and L, appear to well summarize the interplay of gene expression between the defined contrasts in brown adipose and liver tissue beyond individual genes, and further highlight transcriptomic changes in offspring to obese parents when compared to those of lean parents, as exemplified in the following.
KEGG terms appear to well reflect a regulatory landscape of obesity. Firstly, the ubiquitin proteolytic system, as found associated with transcription in BAT and L, influences basic cellular systems such as regulation of cell cycle, modulation of the immune and inflammatory responses, control of signal transduction pathways, and development and differentiation, and its dysfunction may lead to obesity-induced endoplasmic reticulum stress and insulin resistance in the liver61,62. Also, suppression of BAT thermogenesis, as seen here, is known from hyperphagic obese animals, due to an obligatory raised thermic effect of food intake63. Furthermore, it has been reported that diet-induced obese mice are prone to altered responses to Salmonella infection 64 as here observed in BAT. On the contrary, linkage of L expression patterns to HPV infection may be an artefact, resulting from changes of the hepatic glucose metabolism65. Our view is supported by the observation of changed PI3K/AKT signalling pathway activity in the liver, which may result in reduction of hepatic glucose production and glycogenolysis, while increasing glycogen synthesis and the synthesis of fatty acids for storage and subsequent utilization by other tissues, with a link to diabetes66. Lastly, as observed in BAT, the ErbB signalling is essential for adipogenesis67, and likewise observed in the liver; and the Rap1 signalling pathway regulates glycaemic control, rendering it a potential target for therapeutic intervention in diabetes 68.
A comparison between GO terms associated with obesity in humans and our mouse-derived terms derived from liver and brown adipose tissue revealed no obvious similarity, neither across biological processes nor cellular components, nor and molecular functions69. Specifically, terms in Lu et al. 69 appeared overwhelmingly related to inflammatory biological processes, across all three GO term categories (BP, CC, MF). On the contrary, our GO terms appeared related to basic metabolic functions such as thermal homeostasis, and mitochondrial and ribonucleic acid metabolism. Arguably, we also observe startingly different GO terms not only because we are inspecting the regulatory landscape in mice instead of humans, but also because obesity did not appear prevalent among our F1 mice (see Fig. 2; SI Fig. 1).