Adipocyte HIF2α functions as a thermostat via PKA Cα regulation in beige adipocytes

Thermogenic adipocytes generate heat to maintain body temperature against hypothermia in response to cold. Although tight regulation of thermogenesis is required to prevent energy sources depletion, the molecular details that tune thermogenesis are not thoroughly understood. Here, we demonstrate that adipocyte hypoxia-inducible factor α (HIFα) plays a key role in calibrating thermogenic function upon cold and re-warming. In beige adipocytes, HIFα attenuates protein kinase A (PKA) activity, leading to suppression of thermogenic activity. Mechanistically, HIF2α suppresses PKA activity by inducing miR-3085-3p expression to downregulate PKA catalytic subunit α (PKA Cα). Ablation of adipocyte HIF2α stimulates retention of beige adipocytes, accompanied by increased PKA Cα during re-warming after cold stimuli. Moreover, administration of miR-3085-3p promotes beige-to-white transition via downregulation of PKA Cα and mitochondrial abundance in adipocyte HIF2α deficient mice. Collectively, these findings suggest that HIF2α-dependent PKA regulation plays an important role as a thermostat through dynamic remodeling of beige adipocytes.

represses PKACα and therefore a protein kinase A (PKA)-dependent thermogenic/mitochondrial program in adipose tissue. Data are interesting but authors should address the following points.
Major comments: 1.-Authors show that mice treated with the HIF inhibitor YC-1 increased rectal temperature concomitant with an increased expression of thermogenic genes of iWAT upon cold exposure. YC-1 has the potential to inhibit both HIF1 and HIF2 expression (PMID: 19074848). Therefore authors should use the HIF2 specific inhibitors such as PT2385 or PT2399 (PMID: 27595394; PMID: 27595393; PMID: 31999648) to assess (i) whether this inhibitor prevents HIF2 protein accumulation in iWAT and BAT of cold-exposed mice and (ii) whether it increases rectal temperature and surface body temperature, iWAT expression of thermogenic genes as well as miR-3085-3p expression. This suggested experiment might broad the knowledge about potential side effects of this HIF2a inhibitor primarily oriented to treat Vhldeficient renal cell carcinoma. Moreover authors should show whether YC-1 treatment inhibits iWAT and BAT HIF1a and HIF2a protein expression and conversely whether DMOG induces the expression of these two isoforms.
2.-Authors claim that the HIF2α-miR-3085-3p axis is central to reduce the expression of thermogenic genes as well as mitochondrial activity through the repression of PKACα. This conclusion is drawn based on the use of a PKA inhibitor H89. Authors should use another approach with -for example -shRNA o siRNA against PKAC in HIF2α-deficient and control beige adipocytes and -if possible -after in vivo local admistration in iWAT as authors do with miR-3085-3p mimic.
3.-Authors claim that HIF1a AKO mice show a thermogenic phenotype independently of miR-3085-3p. However authors should assess whether HIF1a AKO mice do not have elevated PKACα expression but still can show altered mitochondrial complexes expression in iWAT after cold exposure.
4.-HIF1a and HIF2 deficient mice in brown adipocytes (HIF1BKO and HIF2a BKO) show increased body temperature but without major changes in the RNA expression of the thermogenic genes analyzed especially in HIF1a BKO. These data can lead to consider that metabolic changes described in iWAT might not be central to explain the increased mouse temperature? Or alternatively, do these HIF1BKO and HIF2a BKO mice have increased expression of thermogenic genes and PKACα in iWAT possibly reflecting a possible interplay between BAT and iWAT?
5.-In page 19 (Discussion section) authors mention "Furthermore, it has been reported that HIF1α inhibitor PX-478 and HIF2α inhibitors PT2385 and PT2399 prevent diet-induced obesity and induce thermogenic gene expression in adipose tissues 50, 51, 52". Authors should provide more details about the role of HIF2a in thermogenesis shown in these references since in one of them (PMID: 29035368) its role in related to intestinal HIF2a activity and not locally in the adipose tissue as shown in this manuscript. Moreover authors should cite and discuss the following two references related to the role of HIF signaling in thermogenesis (PMID: 27738746; PMID: 34329568) especially the one in which thermogenesis is impaired upon HIF1α expression in brown adipose tissue (PMID: 27738746).
Minor comment: 1.-In page 7 (Results section) authors mention "Simultaneously, the levels of HIF1α and HIF2α proteins gradually increased during cold exposure in iWAT and BAT (Fig. 1c, d), but not in eWAT (Supplementary fig. 1b)". However a modest HIF2α induction was detected in eWAT upon cold exposure. Therefore authors should change this sentence accordingly. Together, the study provided compelling evidence that cold-induced HIF2a is a crucial negative regulator of PKA signaling and beige fat biogenesis/retention. The research group is a front runner in the field of adipose tissue biology and gained a strong reputation for rigorous research. Overall, the data are compelling, and the manuscript is well-written. The reviewer suggests a few minor points that the authors want to address prior to publication.

HIF1a pathway is known to induce adipose tissue fibrosis, whereas recent studies suggest that beige fat biogenesis is accompanied by repressed fibrosis. It would be insightful if the authors discuss or look into the regulation of adipose tissue fibrosis by the HIF2a-miR-3085 axis.
Thanks for raising the interesting viewpoint. As the reviewer pointed out, adipose tissue fibrosis is negatively correlated with beige adipogenesis. It has been reported that adipose tissue fibrosis could be repressed by the stimuli promoting beige adipocytes including cold exposure and chronic CL-316,243 administration 1, 2, 3 .
Especially, PRDM16 plays an important role in adipose tissue fibrosis via EHMT1-GTF2IRD1 complex-mediated suppression of fibrosis-related gene expression 1 . In contrast, TGFβ signaling, one of the crucial factors promoting fibrosis, attenuates beige adipogenesis and thermogenic programing 4, 5, 6 . In obesity, adipocyte HIF1α stimulates adipose tissue fibrosis, and mural HIF1α also promotes fibrosis-related gene expression and pro-fibrotic phenotypes 3,7,8,9 . In case of HIF2α, it has been shown that HIF2α triggers fibrosis in lung, eye, liver and kidney 10,11,12,13 , but the roles of HIF2α in adipose tissue fibrosis remain elusive. Considering the potential roles of HIF1α and HIF2α in fibrosis, it is feasible to speculate that increasing beige adipocytes in adipocyte HIFα deficient models may lead to suppression of adipose tissue fibrosis. According to the reviewer's suggestion, we analyzed expression profiles of fibrosis-related genes in iWAT of WT and HIF2α AKO mice. Similar to a previous report 1 , cold stimuli inhibited overall fibrosis-related genes, whereas there were no significant differences between the genotypes (Reviewer's only Figure 1). Thus, it is likely that HIF2α-dependent beige adipocyte regulation might not be involved in the suppression of fibrosis upon cold exposure. However, we cannot exclude the possibility that HIF2α-miR-3085-3p axis could regulate adipose tissue fibrosis in other conditions such as CL-316,243 treatment, re-warming, and obesity. It would be interesting to investigate the roles of adipocyte HIF2α in adipose tissue fibrosis on physiological and/or pathophysiological situation for future studies.

The authors wish to mention if the role of miR-3085 is selective to beige adipocytes in the inguinal WAT or also in iBAT.
We appreciate this comment. Our previous manuscript raised the possibility that HIF2α would play thermoregulatory roles in brown adipocytes. According to the reviewer's comment, we analyzed expression levels of miR-3085-3p and Prkaca in BAT of HIF2α BKO mice to test HIF2α-miR-3085-3p-PKA Cα axis. As shown in new Supplementary Fig. 8b, c, cold stimuli promote miR-3085-3p level, and cold-induced miR-3085-3p expression was downregulated in BAT of HIF2α BKO, leading to an increase of Prkaca expression. Similar to beige adipocytes ( Supplementary Fig. 8d, e), HIF2α also promoted miR-3085-3p expression and suppressed Prkaca expression in brown adipocytes (new Supplementary Fig. 8f-i). And then, we tested the thermoregulatory roles of miR-3085-3p in brown adipocytes. Transfection of miR-3085-3p mimic in brown adipocytes attenuated Prkaca, PKA signaling, and PKA-induced thermogenic gene expression (new Supplementary Fig. 8l, m).
Furthermore, we directly injected miR-3085-3p mimic to BAT to evaluate in vivo functions of miR-3085-3p in BAT using liposome-based transfection method (new Supplementary Fig. 8o). After 3 days of cold exposure following miR-3085-3p mimic transfection, BAT was examined. As shown in new Supplementary Fig. 8p, q, miR-3085-3p suppressed the levels of Prkaca and thermogenic genes in BAT. These data suggest that miR-3085-3p regulates thermogenic functions of BAT as well as iWAT. We described these in the revised manuscript (p.12, Data are expressed as the mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001 by two-way ANOVA followed by Holm-Sidak's multiple comparisons test.

YC-1 treatment inhibits iWAT and BAT HIF1a and HIF2a protein expression and conversely whether DMOG induces the expression of these two isoforms.
We appreciate this critique. Following the reviewer's suggestion, we tested the effect of HIF2α-specific inhibitor PT2385 on thermogenic activity. PT2385 cocktail (10 mg/kg, 10% Ethanol, 40% PEG300, 5% Tween-80, and 45% saline) was orally administrated twice a day from 1 day before cold exposure to 3 days after cold exposure. As shown in new Supplementary Fig. 4c, d, PT2385 selectively downregulated the level of HIF2α protein in both iWAT and BAT. When we tested the effects of YC-1 and DMOG on HIFα protein in iWAT and BAT, YC-1 reduced levels of HIF1α and HIF2α protein in both iWAT and BAT (new Supplementary Fig. 4a, b).
In contrast, DMOG elevated the levels of HIF1α and HIF2α protein in both iWAT and BAT (new Supplementary   Fig. 4e, f). In order to address HIF2α-dependent thermogenic regulation, we measured rectal temperature and surface body temperature upon cold stimuli. PT2385-treated mice exhibited cold-resistant phenotypes (new Fig.   2j and new Supplementary Fig. 4g). In addition, thermos genic gene expression and beige adipocyte generation were upregulated by PT2385 treatment in iWAT (new Fig. 2k, l). Also, to examine HIF2α-dependent PKA Cα regulation, the levels of miR-3085-3p and Prkaca were measured. As shown in new Fig. 6g, h, PT2385 treatment suppressed miR-3085-3p expression, leading to an increase of Prkaca expression in iWAT upon cold exposure. As the reviewer mentioned, HIF2α inhibitor has additional effects on metabolism in multiple organs besides VHL-deficient renal cell carcinoma 14,15,16 . Although we cannot exclude that in vivo PT2385 treatment could play systemic roles in the regulation of thermogenesis besides BAT and iWAT, our pharmacological and genetic data suggest that adipocyte HIF2α-dependent PKA regulation is crucial for the regulation of thermogenesis. These results were addressed in the revised manuscript

2.-Authors claim that the HIF2α-miR-3085-3p axis is central to reduce the expression of thermogenic genes as well as mitochondrial activity through the repression of PKACα. This conclusion is drawn based on the use of a PKA inhibitor H89. Authors should use another approach with -for example -shRNA or siRNA against PKACα in
HIF2α-deficient and control beige adipocytes and -if possible -after in vivo local admistration in iWAT as authors do with miR-3085-3p mimic.
Thanks for this suggestion. According to the comment, we performed Prkaca knockdown experiments to examine PKA-dependent thermogenic regulation by HIF2α in beige adipocytes and iWAT. Similar to H89 treatment (new Supplementary Fig. 7f), Prkaca knockdown attenuated thermogenic gene expression in beige adipocytes of HIF2α AKO (new Supplementary Fig. 7g). Moreover, enhanced mitochondrial respiration in HIF2α deficient beige adipocytes was impaired by Prkaca knockdown (new Supplementary Fig. 7h). We also tested in vivo effects of Prkaca knockdown in iWAT via liposome-based transfection method. Local delivery of siPrkaca in iWAT downregulated UCP1 and OXPHOS complexes, along with reduced beige adipocyte formation in HIF2α AKO mice upon cold stimuli (new Fig. 5k, l). Together, these results suggest that HIF2α-PKA Cα axis would mainly regulate thermogenic execution in beige adipocytes. These data were described in the revised manuscript (p.11, line 212-218).

3.-Authors claim that HIF1a AKO mice show a thermogenic phenotype independently of miR-3085-3p. However authors should assess whether HIF1a AKO mice do not have elevated PKACα expression but still can show altered mitochondrial complexes expression in iWAT after cold exposure.
The points are well taken. According to this suggestion, we investigated expression levels of mitochondrial genes and OXPHOS complexes in iWAT of HIF1α AKO. As shown in Supplementary Fig. 7c, d, the expression of mitochondrial genes and OXPHOS complexes was upregulated in iWAT of HIF1α AKO upon cold exposure.
However, PKA Cα expression was not altered by HIF1α in beige adipocytes and iWAT (Fig. 4e, Supplementary   Fig. 6h, k, and new Supplementary Fig. 6i). With these data, we affirmed that HIF1α could regulate thermogenic functions, probably, via PKA Cα-independent manner. We included these results in the revised manuscript (p.10, line 205-206).
4.-HIF1a and HIF2 deficient mice in brown adipocytes (HIF1BKO and HIF2a BKO) show increased body temperature but without major changes in the RNA expression of the thermogenic genes analyzed especially in HIF1a BKO. These data can lead to consider that metabolic changes described in iWAT might not be central to explain the increased mouse temperature? Or alternatively, do these HIF1BKO and HIF2a BKO mice have increased expression of thermogenic genes and PKACα in iWAT possibly reflecting a possible interplay between BAT and iWAT?
We are grateful for intriguing suggestions. As the reviewer pointed out, both HIF1α BKO and HIF2α BKO mice were cold tolerant without significant changes in thermogenic gene expression. Since expression levels of thermogenic genes were comparable in iWAT of WT, HIF1α BKO, and HIF2α BKO mice (new Supplementary   Fig. 3f), we assumed that there might be BAT own thermogenic regulatory mechanisms rather than interplay with iWAT. We found that miR-3085-3p expression was increased in BAT as well as iWAT upon cold exposure (new Supplementary Fig. 8b). Moreover, miR-3085-3p expression was downregulated in BAT of HIF2α BKO, accompanied with an increase of Prkaca expression (new Supplementary Fig. 8b, c and Review's only Figure 2a).
As brown adipocytes also showed HIF2α-miR-3085-3p-PKA Cα axis, (new Supplementary Fig. 8f-i, m), it seems that PKA Cα regulation would be an important process in BAT of HIF2α BKO model. In HIF1α BKO mice, Prkaca expression was not altered in BAT upon cold exposure (Review's only Figure 2a). Although histological differences were not largely observed at 3 days of cold exposure in BAT of WT, HIF1α BKO, and HIF2α BKO ( Supplementary Fig. 3c, d), lipid droplets were relatively smaller in BAT of brown adipocyte HIFα deficient models than that of WT at 6 hours of cold exposure (Review's only Figure 2b, c). These data raise the possibility that HIF1α deficiency in brown adipocytes might activate thermogenic and/or lipid catabolic pathways without regulation of thermogenic gene expression in BAT. It has been reported that thermogenic activity could be reinforced via modifications and/or modulation of UCP1 regardless of the transcriptional changes 17,18,19 .
Additionally, accumulating evidence has suggested that there are several processes of UCP1-independent thermogenesis 20,21,22 . In future study, it will be intriguing to investigate how HIF1α regulates thermogenic function in BAT. We described above data in the revised manuscript (p.7, 12, line 137-139, 235-240). Thanks for this critique. As the reviewer suggested, we included detailed description of the references 14,15,16,23 and addressed in vivo roles of PT2385 and intestinal HIF2α in the revised manuscript. Moreover, we provided additional references 16,24 and discussed roles of HIF signaling in thermogenesis. The text was modified in the discussion section (p.18, line 370-380).
Minor comment: 1.-In page 7 (Results section) authors mention "Simultaneously, the levels of HIF1α and HIF2α proteins gradually increased during cold exposure in iWAT and BAT (Fig. 1c, d), but not in eWAT (Supplementary fig. 1b)". However, a modest HIF2α induction was detected in eWAT upon cold exposure. Therefore, authors should change this sentence accordingly.