De novo adipocyte differentiation from Pdgfrβ+ preadipocytes protects against pathologic visceral adipose expansion in obesity

Pathologic expansion of white adipose tissue (WAT) in obesity is characterized by adipocyte hypertrophy, inflammation, and fibrosis; however, factors triggering this maladaptive remodeling are largely unknown. Here, we test the hypothesis that the potential to recruit new adipocytes from Pdgfrβ+ preadipocytes determines visceral WAT health in obesity. We manipulate levels of Pparg, the master regulator of adipogenesis, in Pdgfrβ+ precursors of adult mice. Increasing the adipogenic capacity of Pdgfrβ+ precursors through Pparg overexpression results in healthy visceral WAT expansion in obesity and adiponectin-dependent improvements in glucose homeostasis. Loss of mural cell Pparg triggers pathologic visceral WAT expansion upon high-fat diet feeding. Moreover, the ability of the TZD class of anti-diabetic drugs to promote healthy visceral WAT remodeling is dependent on mural cell Pparg. These data highlight the protective effects of de novo visceral adipocyte differentiation in these settings, and suggest Pdgfrβ+ adipocyte precursors as targets for therapeutic intervention in diabetes.

the context of the lengths of the studies where authors used HFD for several weeks. How are the total Pparγ levels at the e nd of the experiments, and how is the Pparg expression in the GFP+ cells (from the lineage tracing experiment, eg. shown in Fig 2s,t? Authors responce: We thank the reviewer for this question. This is an important point worth emphasis. Pdgfrb is not expressed in mature adipocytes-this is apparent from the new qPC R data in Supplementary Figure 2D. Activation of the TRE-Pparg2 transgene is dependent on the presence of rtTA, which in turn, is dependent on active Pdgfrb expression. As such, as cells differentiate into adipocytes, the Pdgfrb is shut off and rtTA is no longer expressed. On the other hand, mGFP expression in this system is coming from the Rosa26R locus; once C re mediated activation of the Rosa locus occurs, mGFP expression is constitutive and no longer dependent on Pdgfrb expression. Of course, new adipocytes (mGFP+) emerging after HFD feeding may differ in gene expression from pre-existing adipocytes or adipocytes from control animals; however, this would not be attributed to active transgene expression.
Additional remark: This explanation describes how their transgenic model works, and this was clear in the original manuscript, but it does not answer my concern. The question is not whether new and pre -existing adipocytes might have different gene expression, but whether the new adipocytes that originated due to Pparg2 overexpression in their precursors differ compared to the naturally occurring new adipocytes in the control mice due to HFD.

Specifically, the results of new experiments that address metabolism in the absence of adiponectin and tissue specific insulin sensitivity by clamp and Akt phosphorylation provide compelling support for the premise.
In its current form, this manuscript represents a convincing and conceptual advance.
We thank the reviewer for taking the time to review the manuscript and provide very constructive feedback.

This work by Gupta and colleagues was previously reviewed by Nature. In the revised manuscript, the authors have very carefully addressed all major points raised by the reviewers, including revisions of the conclusions that can be drawn from the experiments. The manuscript is very interesting and well written. Below are a few points that should be addressed.
We thank the reviewer for carefully reviewing the manuscript and for providing very constructive feedback.

The authors have added an additional animal model, mural-PparγTG bred to adiponectin-deficient mice and show that the improvement of glucose tolerance and insulin sensitivity are dependent on adiponectin. They use this as evidence that the improved metabolic function in mural-PpargTG is dependent on adipose tissue function. However, this is a rather indirect piece of evidence, e.g. it is unclear whether adiponectin is actively involved in the "signal" elicited by the increased level of mural PPARγ or whether it is just required for the insulin sensitizing function of increased mural PPARγ. I don't think the authors need to outline the mechanism, but they should modify their claims.
This is an excellent point. We have now added/modified the following text of the discussion section (Pages 22-23) to explicitly state this caveat: "The observation that the improved insulin sensitivity observed in the Mural-Pparg TG mice correlates with increased serum adiponectin levels, and depends on the presence of this adipokine, does strongly implicate improved adipocyte function as a major driver of the metabolic phenotype in this model; however, it is still unclear whether increased adiponectin secretion per se directly from the healthy visceral WAT depots of these animals is the primary driver of improved systemic glucose homeostasis. Additional studies are needed to understand the precise mechanisms by which healthy visceral WAT can mediate improvements in glucose homeostasis. In fact, the Mural-Pparg TG mice described here may be a useful tool to identify adipokines and/or secreted metabolites linked to healthy vs. unhealthy adipocyte function in obesity. Furthermore, one important question that remains is whether adipocytes emerging in response to HFD feeding in adults are molecularly and functionally distinct from preexisting visceral adipocytes."

It is also not clear which adipose depot secretes increased adiponectin levels in response to ectopic mural-PPARg, and it cannot be concluded that it is from the visceral depots. Did the authors investigate whether Pdgfrβ+ cells in other tissues start to express adipocyte marker genes? Does mural PPARγ affect the bone marrow?
These are also great questions. As it relates to bone marrow specifically, we are currently investigating whether Pdgfrb expression identifies adipocyte precursors giving rise to adipocytes residing within bone marrow, and whether TZDs trigger bone marrow adipocyte differentiation through these cells. This is on-going study that will be published in the future.