Lack of p38 activation in T cells increases IL-35 production and protects against obesity by promoting thermogenesis

  1. Ivana Nikolic
  2. Irene Ruiz-Garrido
  3. María Crespo
  4. Rafael Romero-Becerra
  5. Luis Leiva-Vega
  6. Alfonso Mora
  7. Marta León
  8. Elena Rodríguez
  9. Magdalena Leiva
  10. Ana Belén Plata-Gómez
  11. Maria Beatriz Alvarez
  12. Jorge L. Torres
  13. Lourdes Hernández-Cosido
  14. Juan Antonio López
  15. Jesús Vázquez
  16. Alejo Efeyan
  17. Pilar Martin
  18. Miguel Marcos
  19. Guadalupe Sabio

Reviewed by

Read the full article

Listed in

Log in to save this article

Abstract

Obesity is characterized with low grade inflammation, energy imbalance and impaired thermogenesis. The role of regulatory T cells (Treg) in inflammation-mediated maladaptive thermogenesis has not been well established. We discovered that p38 pathway is a key regulator of T cell-mediated adipose tissue (AT) inflammation and browning. Mice with T cells specific deletion of the p38 activators, MKK3/6, were protected against diet-induced obesity and AT inflammation improving their metabolic profile, higher browning and thermogenesis. We identified IL-35 as a driver of adipocyte thermogenic program through ATF2/UCP1/FGF21 pathway. IL-35 limits CD8 + T cell infiltration and inflammation in AT. Interestingly, we found that IL35 was reduced in visceral fat from obese patients. Mechanistically we showed that p38 controls the expression of IL-35 in human and mouse Treg cells through mTOR pathway activation. Our findings highlight p38 signaling as a molecular orchestrator of AT T cell accumulation and function and identify p38 and IL-35 as promising targets for metabolic diseases.

Article activity feed

  1. Review Commons

    Note: This rebuttal was posted by the corresponding author to Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Reply to the reviewers

    Reviewer #1 (Evidence, reproducibility and clarity):

    Major comments:

    1. A control group of mice fed chow diet is needed to distinguish the effects of the genotype from those caused by diet. What is the phenotype of regular chow-fed mice in terms of energy metabolism and thermogenesis?

    We are sincerely grateful to Reviewer 1 for raising an important question regarding the need for a control group of mice fed chow diet.

    To address this concern, we have conducted experiments on mice fed a regular chow diet and measured their phenotype in terms of energy metabolism and thermogenesis. In addition to be sure that the phenotype also is present in when we compared littermates we have included as control both to chow-fed CD4-Cre and littermates (MKK3/6f/f). Our findings reveal that MKK3/6CD4-KO mice fed a chow diet presented an increased brown adipose tissue (BAT) thermogenesis compared with CD4-Cre and littermates. This phenotype is similar to the observed in HFD-fed mice. Also, these results indicate that the same phenotype is observed when we compared with littermates including an extra control in the study.

    To further investigate the effect on energy metabolism, we utilized metabolic cages. The data from these experiments align with the increased thermogenesis observed in MKK3/6CD4-KO mice fed a chow diet, as they also demonstrated increased energy expenditure. We thank the reviewer for this suggestion as we believe that these new data strengthen our conclusion significantly.

    We have thoughtfully incorporated these essential findings into in Supplementary Figure 2C-D of the manuscript.

    1. While an increase in BAT temperature (as demonstrated here by infrared imaging) in line with increased thermogenesis, it will be critical to verify this hypothesis by indirect calorimetry. Energy expenditure, food intake, and activity measures should be added for regular and DIO mice. Please follow the guidelines for ANCOVA analysis and measurements explained in PMID: 22205519 and PMID: 21177944.

    We are grateful to Reviewer 1 for bringing up an essential point concerning the need to verify our hypothesis on increased BAT temperature and thermogenesis through indirect calorimetry. We acknowledge the importance of including energy expenditure, food intake, and activity measures for both regular and DIO mice to strengthen our study.

    To address this valuable suggestion, we have taken immediate action. We utilized metabolic cages in mice under chow diet. The data from these experiments align with the increased thermogenesis observed in MKK3/6CD4-KO mice fed a chow diet, as they also demonstrated increased energy expenditure, without differences in food intake or locomotor activity. We thank the reviewer for this suggestion as we believe that these new data strengthen our conclusion significantly. These new data are now in Supplementary Figure 2A-B.

    In addition, we have initiated a new experimental group of age-matched mice on HFD, which we will carefully feed for 8 weeks. Following this dietary period, we will subject the mice to metabolic cage analysis, allowing us to obtain accurate data on energy expenditure, food intake, and activity levels. These additional measurements will provide a comprehensive understanding of the metabolic changes induced by MKK3/6 deficiency in T cells under different dietary conditions.

    1. That the phenotype is still seen at isothermal housing is interesting but should be backed up by direct assessment of thermogenic capacity (see PMID: 21177944). In the end, it could also be increased heat loss, independently of heat production. If the browning is cause or consequence remains unclear, then.

    Thank you for raising this important point. Indeed, it is essential to corroborate the observed phenotype with direct assessments of thermogenic capacity to gain a comprehensive understanding of the underlying mechanisms. The study mentioned in PMID: 21177944 highlights the significance of evaluating thermogenesis directly to support the findings.

    According to your suggestion, we plan to house the animals at 30 ºC for four weeks and subsequently inject norepinephrine to evaluate thermogenesis capacity while measuring brown adipose tissue (BAT) activation. This approach should provide valuable insights into the thermogenic potential of the animals under isothermal conditions.

    However, we will not be able to conduct the experiment in metabolic cages at 30 ºC due to the constraint that our system does not allow 30 ºC temperature. For this reason, we will measure BAT temperature to analyze this experiment.

    1. Regarding the in vitro data, a thermogenic phenotype should be functionally verified by Seahorse analysis.

    We thank Reviewer 1 for raising an important point concerning the need for functional verification of the thermogenic phenotype observed in our in vitro data using Seahorse analysis.

    In response to this valuable suggestion, we performed Seahorse analysis in differentiated adipocytes treated with or without IL-35 for 48 hours. The results demonstrated a slight increase in basal metabolism and a heightened response to isoproterenol (ISO) stimulation of β3 adrenergic receptors in adipocytes after IL-35 treatment. These findings provide functional evidence supporting the thermogenic phenotype induced by IL-35 in adipocytes.

    We have thoughtfully included this essential data in Figure 2 of this revision plan, allowing reviewers and the scientific community to comprehensively evaluate and validate the functional implications of our findings.

    1. Mechanistically, there is epistasis type of experiment that IL-35 influences Ucp1 levels via ATF2 as the data remain associative in nature.

    Thank you for your valuable comment. We agree that to establish a mechanistic link between IL-35 and Ucp1 levels will improve the strength of the manuscript.

    To delve deeper into the mechanism through which IL-35 influences Ucp1 expression, we focused on the role of ATF2, a transcription factor known to be involved in regulating UCP1 levels (PMID: 11369767 and PMID: 15024092). In our investigation, we treated adipocytes with IL-35 both in the presence and absence of an inhibitor targeting the ATF2 pathway. The results were illuminating as we observed a significant reduction in the expression of Ucp1 when the ATF2 pathway was inhibited.

    These findings indicate that ATF2 is indeed a crucial mediator of the effects of IL-35 on Ucp1 levels. By inhibiting the ATF2 pathway, we demonstrate a direct functional link between IL-35 and the expression of Ucp1, providing mechanistic insights into the regulatory role of IL-35 in thermogenesis. We included new results in Figure 7F.

    1. What are other consequences of injecting IL-35? Is it good or bad? What is the therapeutic potential in DIO mice? Also, in these experiments (Fig. 7) indirect calorimetry as described would be supportive of the claims.

    Regarding the consequences of injecting IL-35, we have already performed experiments to analyze its effect. Our findings indicate that IL-35 increases thermogenesis in BAT (Figure 7), suggesting that it may play a role in promoting energy expenditure, which could be beneficial in combating diet-induced obesity (DIO) in mice. Importantly, we did not observe any negative effects of IL-35 in our experiments.

    Based on these promising results, we are expecting the therapeutic potential of IL-35 in DIO mice. By promoting thermogenesis in BAT, IL-35 may offer a novel approach to manage obesity and related metabolic disorders. However, we acknowledge that further comprehensive studies are needed to fully understand its therapeutic benefits and potential side effects.

    In our future works, we plan to evaluate a targeted delivery system for IL-35. We are currently generating IL-35 loaded metal-organic frameworks (MOFs) labeled with adipose tissue-specific peptides. This innovative strategy aims to enhance the delivery of IL-35 to adipose tissue, potentially maximizing its effects in the relevant areas. Our ongoing work with IL-35 loaded MOFs may offer a promising avenue for targeted delivery.

    Minor comments:

    1. The authors claim that their HFD-fed MKK3/6CD4-KO mice are protected against hyperglycemia, but only fasted/fed blood glucose tests are performed. Lower glucose levels could be explained due to a hyperinsulinemic state in response to growing insulin resistance in the presence of HFD. It would be sensible to perform both glucose and insulin tolerance tests to back up your statement.

    Thank you for your insightful comment. We agree that to support our claim of protection against hyperglycemia in HFD-fed MKK3/6CD4-KO mice, further tests are necessary beyond fasted/fed blood glucose measurements.

    In response to your suggestion, we conducted both glucose tolerance tests (GTT) and insulin tolerance tests (ITT) in HFD-fed MKK3/6CD4-KO mice. We did not observed differences in glucose tolerance and but ITT showed significantly enhanced insulin sensitivity compared to control mice. These findings provide evidence that the protection against hyperglycemia in HFD-fed MKK3/6CD4-KO mice is not solely due to a hyperinsulinemic state, but rather indicates genuine improvements in glucose handling and insulin response.

    We have thoughtfully included these crucial data in the revised version of the manuscript, both in the main text and Supplementary Figure 4. We extend our appreciation to the reviewer for this valuable suggestion, which has enhanced the scientific rigor and completeness of our study.

    1. Please provide the loading control for p38 and S6 blots (Figure 6G).

    Thank you for the comment. The loading control we used for P p38 and P S6 blots in Figure 6G is β-actin. Due to the limited amount of sample available, we can only use β-actin as the loading control. The sample amount obtained is very limited, and we can only provide enough lysate to run a couple of blots from the same sample. Running several western blots with the same sample is almost impossible given the constraint of the sample availability. We apologize for this limitation, but it is necessary to avoid using too many mice for ethical reasons, as the samples come from a large number of mice.

    1. Statistical test from Figure 7B should be a t-test, since it is only comparing 2 variables (PBS vs IL-35), and not a 2-way ANOVA as described in the legend.

    We sincerely thank the reviewer for the comment. It was indeed a mistake in the text. While we have performed a t-test, there was an error in the legend that we have now corrected. We apologize for any confusion this may have caused and appreciate the opportunity to rectify the oversight.

    1. Label correctly the panels in the figures -examples: Fig 3, panels C and D are interchanged; reference in the text to Fig S1G even though the figure only as panels A-F; Fig 7 legend referes to the statistical test of panel E when the figure only has A-D.

    We sincerely apologize for any mistakes in our manuscript that may have caused difficulties while reading the article and potentially led to misleading results. We are grateful to Reviewer #1 for bringing these errors to our attention. Thanks to their diligent review, we have been able to identify and rectify the issues in our manuscript. The necessary corrections have been made, ensuring the accuracy and reliability of our research. We greatly appreciate the reviewer's valuable feedback and contribution to improving the quality of our work.

    1. There are several typos along the text, please revise (example: page 4;line 4 -"tremorgenic")

    We apologize for the presence of any typos in the initial version of the article. We have thoroughly revised the manuscript to correct these errors. Thank you for bringing this to our attention and helping us improve the accuracy and clarity of our work.

    Reviewer #1 (Significance):

    The manuscript is well written, and the research conducted properly, even though a thorough analysis of energy metabolism in mice and cells is missing and the mechanistic claims are based on relatively thin data.

    The immune system and inflammation play important roles for obesity and insulin resistance, yet the roles they play in thermogenic adipocytes remains unclear. This work adds novel aspects to this relationship.

    Reviewer #2 (Evidence, reproducibility and clarity):

    This manuscript by Nikolic et al sought to investigate the role of p38 activation in adipose tissue Treg cells and obesity. They found that the expression of p38a, its upstream kinase MKK6, and downstream substrate ATF2 was upregulated specifically in adipose T cells associated with human obesity. They generated T cell-specific knockout MKK3/6 in mice and found these animals were protected from diet-induced obesity as a result of increased BAT thermogenesis. Mechanistically, loss of p38a activation promoted adipose tissue accumulation of Treg cells, leading to elevated IL-35 availability and UCP1 expression.

    Major comments:

    1. They attributed the obesity protection to energy expenditure; however, food intake and intestinal absorption were never tested. Immune cells particularly Treg cells are important modulates of nutrient uptake.

    We are sincerely grateful to Reviewer #2 for this crucial comment, highlighting the importance of assessing not only energy expenditure but also food intake and intestinal absorption in our study.

    In response to this valuable suggestion, we have initiated an HFD experiment to comprehensively examine food intake and intestinal absorption. For food intake analysis, we are employing metabolic cages, which will allow us to monitor and quantify the amount of food consumed by the mice accurately. Additionally, we plan to follow the methodology outlined in the study by Kraus et al. (PMID: 27110587) to measure lipid content in feces, enabling us to evaluate intestinal absorption.

    By conducting these additional experiments, we aim to gain a deeper understanding of the potential role of Treg cells, known immune modulators of nutrient uptake, in our observed obesity protection phenotype.

    1. At thermoneutrality, BAT is inactive even though UCP1 expression is still present (not activated). MKK3/6 deficiency in T cells still confer protection against obesity at thermoneutrality suggests it regulates other energy balance components in addition to BAT thermogenesis.

    Thanks for the comment. We believe that the effects of IL35 on thermogenesis are likely partly mediated by alternative mechanisms, as we did not observe an increase in UCP1 gene expression in BAT in vivo (Figure 3D of the manuscript), and the increase in thermogenesis is still present even at thermoneutrality where UCP1 is inactive (Figure 4E of the manuscript). This suggests that IL35 might regulate other alternative pathways that control BAT thermogenesis.

    While our current findings provide valuable insights, further experiments may be necessary to fully understand the underlying mechanisms. For instance, conducting experiments with transgenic mice expressing IL35 or using IL35 knockout (KO) mice could shed more light on the specific pathways through which IL35 exerts its effects on thermogenesis and energy balance.

    In conclusion, we hypothesize that IL35's effects on thermogenesis are mediated partly by alternative mechanisms beyond UCP1 activation, and its ability to enhance thermogenesis even at thermoneutrality highlights its potential as a regulator of energy balance. We plan to further investigate the specific mechanisms through which IL35 impacts thermogenesis and energy balance. To achieve this, we will consider conducting experiments with transgenic mice expressing IL35 or using IL35 knockout (KO) mice in follow up studies. This is now discussed in our manuscript.

    1. Loss of adipose Treg cells (such as Pparg KO, Foxp3-DTR) did not lead to obvious obesity phenotypes. Gain-of-function Treg cells (such as adoptive transfer, IL-2/IL-2 Ab) did not results in profound obesity protection as observed in MKK3/6 CD4-KO mice. It suggests that MKK3/6 KO in T cells causes other immune defects (besides Tregs).

    We agree with the referee's assessment that the lack of obvious obesity phenotypes in above mentioned animal models. The results we observed in our MKK3/6CD4-KO mice suggest that p38 signaling pathway in T cells may modulate their function, leading to an upregulation of IL35 expression, which could be a contributing factor to the significant obesity protection observed in MKK3/6CD4-KO mice. We believe that IL35's effects on energy balance and thermogenesis are critical components of the observed protection against obesity in this model.

    Regarding the studies with PPAR KO in Treg cells, it is important to note that they did not specifically focus on the effect of thermogenesis. While they observed a general tendency of increased fat deposition when treated with a PPAR agonist in the Treg deficient PPAR KO mice, these findings were not extensively studied in that particular paper. Thus, additional research is necessary to specifically evaluate thermogenesis in these mice and further understand the role of PPAR in Treg-mediated thermogenic processes.

    We also acknowledge the presence of contradictory results from loss-of-function experiments of Treg cells in mice. The observed metabolic changes may be context-dependent, and the impact of Treg cells on metabolism might vary under different physiological conditions. For instance, in lean conditions where adipose tissue inflammation is low, a decrease in VAT Treg cells might not lead to significant metabolic changes. However, under certain circumstances, such as obesity, VAT Treg cells may play a critical role in regulating metabolism. In this context increasing that population that is reduced during obesity could results in improve metabolic performance.

    In conclusion, our findings suggest that the lack of p38 activation in Treg cells may prevent the dramatic down-regulation and loss of function observed in Treg cells during obesity. This preservation of Treg function could be a significant factor driving the observed protection against obesity in MKK3/6CD4-KO mice.

    While further studies are required to elucidate the precise timing and spatial aspects of the specific functions of adipose-resident Treg cells, it is evident that these cells play a crucial role in maintaining immune and metabolic homeostasis. They achieve this, in part, by regulating adipose inflammation, insulin sensitivity, lipolysis, and thermogenesis. This is now discussed in our manuscript.

    1. The increase in IL-35 seemed to be very moderate, compared to the metabolic phenotypes. It raises the question if IL-35 is responsible for BAT activation and reduced weight gain. It is unclear what systemic and local levels of IL-35 were reached after recombinant IL-35 treatment (Fig. 7B). IL-35 antibody blockade experiment in KO mice is recommended.

    Physiological changes in cytokines can indeed have a significant impact on the metabolic profile due to their continuous and intricate interactions. Even minor alterations in the overall cytokine milieu can result in substantial changes in metabolism (doi.org/10.1073/pnas.1215840110). In fact, it is well-established that in humans, small changes in cytokine profiles between genders, in obesity, and during aging can play a critical role in the development of pathology. These cytokines often operate in a chronic manner, exerting long-term effects on various physiological processes (doi.org/10.1038/s41467-020-14396-9).

    In summary, the dynamic interplay of cytokines in metabolism can lead to significant metabolic changes even with subtle alterations in their levels. While the increase in IL-35 may appear moderate, our findings using recombinant IL35 indicate that IL-35 increases thermogenesis in BAT, suggesting that it may play a role in promoting energy expenditure, which could be beneficial in combating diet-induced obesity (DIO) in mice. Importantly, we did not observe any negative effects of IL-35 in our experiments.

    1. IL-35 induced p-ATF2 is acute and transient (Fig. 7D) and it was able to increase BAT temperature in just 4 h (Fig. 7B). However, Ucp1 transcription and translation generally take much longer time (e.g. 2d in Fig. 7C). IL-35 may increase energy expenditure through UCP1-independent mechanisms.

    Thanks for the comment. As previously mentioned, we believe that the effects of IL35 on thermogenesis are might be mediated by alternative mechanisms, as we did not observe an increase in UCP1 gene expression in BAT, and the increase in thermogenesis is still present even at thermoneutrality where UCP1 is inactive. This suggests that IL35 might regulate other alternative pathways that control BAT thermogenesis.

    While our current findings provide valuable insights, further experiments may be necessary to fully understand the underlying mechanisms. For instance, conducting experiments with transgenic mice expressing IL35 or using IL35 knockout (KO) mice could shed more light on the specific pathways through which IL35 exerts its effects on thermogenesis and energy balance. We plan to further investigate the specific mechanisms through which IL35 impacts thermogenesis and energy balance. To achieve this, we will consider conducting experiments with transgenic mice expressing IL35 or using IL35 knockout (KO) mice in follow up studies. This is now discussed in our manuscript.

    Minor comments:

    1. The gating of Treg cells should exclude CD25- cells. Single positive (CD25+ or Foxp3+) cells are progenitors of Tregs. In addition to number, phenotypic activation of Treg cells should also be determined.

    Thank you for the comment. We have reanalyzed our data by excluding CD25- cells and included now in the figure 5A of the manuscript and new supplementary figure 7 of revised manuscript. We also checked CD69+ and KLRG1+ Treg cells and observed no differences between genotypes. We also included figures in this revision plan (Figure 5 and 6).

    1. ATF is also important for adipogenesis, is the adipogenic differentiation of BAT SVF cells affected by MKK3/6 KO or IL-35 treatment?

    We appreciate the reviewer's observation regarding the importance of ATF in adipogenesis. To investigate this aspect further, we performed in vitro differentiation of adipocytes and treated them with IL-35 in the presence or absence of an inhibitor targeting the upstream activator of ATF.

    The results were compelling, as IL-35 treatment led to an increase in the expression of adipogenic markers, including Pparg, Adipoq, Leptin, and Perilipin. In contrast, inhibiting ATF activation resulted in a reduction of these adipogenic markers. These findings provide strong evidence that ATF plays a significant role in mediating the effects of IL-35 on adipogenesis.

    We have thoughtfully included these essential data in Figure 7G of the manuscript. We extend our gratitude to the reviewer for their keen observation, which has enhanced the scientific depth and completeness of our study.

    1. Metabolic cage experiments are desired to determine whole-body energy balance, including food intake, physical activity, and heat production.

    To address this valuable suggestion, we have taken immediate action. We utilized metabolic cages in mice under chow diet. The data from these experiments align with the increased thermogenesis observed in MKK3/6CD4-KO mice fed a chow diet, as they also demonstrated increased energy expenditure, without differences in food intake or locomotor activity. We thank the reviewer for this suggestion as we believe that these new data strengthen our conclusion significantly. The new data are included in Supplementary figure 2 A-B.

    In addition, we have initiated a new experimental group of age-matched mice on HFD, which we will carefully feed for 8 weeks. Following this dietary period, we will subject the mice to metabolic cage analysis, allowing us to obtain accurate data on energy expenditure, food intake, and activity levels. These additional measurements will provide a comprehensive understanding of the metabolic changes induced by MKK3/6 deficiency in T cells under different dietary conditions.

    1. Total UCP1 expression (both RNA and protein) in the whole BAT from an animal should determined (since BAT is smaller in KO mice).

    Thank you for this comment. Yes, we have measured UCP1 expression in the whole BAT from the animals. It is in the figure 3C and 3D and here. Although in vitro studies indicated that IL35 increase UCP1 in adipocytes we were not able to find an increase of this protein in BAT

    We believe that the effects of IL35 on thermogenesis are likely partly mediated by alternative mechanisms, as we did not observe an increase in UCP1 gene expression in BAT in vivo, and the increase in thermogenesis is still present even at thermoneutrality where UCP1 is inactive (Figure 4E of the manuscript). This suggests that IL35 might regulate other alternative pathways that control BAT thermogenesis.

    1. Fig. 6C, IL-35-expressing Treg cells should be quantified from adipose tissue.

    We appreciate the referee's suggestion to quantify IL-35-expressing Treg cells from adipose tissue in Fig. 6C. While we agree that this would be valuable information, we encountered technical challenges that made it impractical to measure IL-35 directly in Treg cells from the visceral adipose tissue (VAT).

    One of the main technical challenges we encountered is the low number of Treg cells present in the adipose tissue, making it difficult to obtain sufficient cell material for accurate quantification of IL-35. Treg cells are relatively rare compared to other immune cell populations in the adipose tissue, and their extraction and analysis can be technically demanding.

    Reviewer #2 (Significance):

    The manuscript is innovative in define the novel role of p38 activation in the T cell compartment and its metabolic regulation. The involvement of Treg cells in adipose tissue homeostasis has been well documented and Treg cell-derived IL-35 has been demonstrated in immune regulation. The authors provided a relatively thorough description of the altered metabolism in these Mkk3/6 CD4-KO mice; however, the reviewer has doubts if Treg cells and IL-35 are primary mechanisms of the observed protection from obesity. The manuscript would be much stronger if the model were Treg cell-specific KO and/or IL-35 deficiency in Treg cells reverses obesity resistance conferred by MKK3/6 deficiency. It also suspected that BAT thermogenesis is not the major reason, as BAT deficiency or UCP1 KO results in much milder phenotypes in mice, even at thermoneutrality.

    Reviewer #3 (Evidence, reproducibility and clarity):

    Specific comments:

    1. It's important to use proper controls for mouse metabolic studies. The authors stated that CD4-Cre and MKK3/6 CD4-KO mice are all in the C57B/6L background. However, it would appear that these two lines were bred separately. The difference in the genetic background, despite minor, can lead to the observed phenotype, notably weight gain. Since the metabolic phenotypes seem to be driven by the weight difference, it is even more critical to include additional controls to validate the findings. For instance, crossing MKK3/6 f/f with one copy of CD4-Cre with MKK3/6 f/f to generate age-matched MKK3/6 CD4-KO and MKK3/6 f/f controls should be used to repeat major in vivo studies similar to those in Fig. 2-4.

    We thank the reviewer for the comment. Although, every control is important using conditional mice, there are several papers indicating that all the cre expression lines have for their own effects that could be important in metabolism and there are several articles that strongly recommended to use cre+ lines as a control. For that reason, we have used the cre expressing line as a control because we really think is the best one (Jonkers and Berns, 2002). In fact, Jackson laboratory recommend to use cre expressing line as a control to avoid side effects that cre overexpression could have in the tissue of interest (https://biokamikazi.files.wordpress.com/2014/07/cre-lox-imp-notes.pdf).

    However, as this reviewer suggested, we checked that similar results were obtained using littermates as controls and we have now included these data in the manuscript (Supplementary Figure 2D).

    1. The assessment of adipose tissue immune cell population in Fig. 5 was conducted after HFD-induced obesity. As mentioned above, the change in Treg and M2 cell percentage could be due to the body weight difference. The experiment should be repeated (with proper controls) in normal chow and after a few weeks of HFD when Treg numbers start to decline.

    Thank you for the comment. We currently performing short HFD experiment to check Treg and M2 cell population in adipose tissue using the littermates as controls.

    In addition, we checked those cell populations in adipose tissue infiltrates in mice fed chow diet and observed no differences in M2 macrophage population between mice, while the percentage of Treg cells was actually lower in MKK3/6CD4-KO mice ND-fed mice (Fig 12 of revision plan). This result suggests that higher accumulation of Treg cells in mice lacking p38 activation in T cells are specific of obese state and strengthen our hypothesis that DIO protection in MKK3/6CD4-KO mice is due to Treg cell population.

    1. Data related to the mechanistic link in Fig. 6/7 are not robust and require a large amount of additional work to substantiate the claim. First of all, the role of IL-35 in BAT thermogenesis remains unclear. It's somewhat surprising to see a single dose of IL-35 i.v. injection is sufficient to increase BAT temperature in Fig. 7B. Minimally, the authors need to demonstrate that IL-35 treatment (perhaps after a few daily doses) is able to increase browning/beiging of fat cells and improve cold tolerance when placing the mice at 4 degree of several hours (and up to 3 days). Serum FGF21 level should also be measured after/during IL-13 treatment. Secondly, ATF2 knockout or knockdown in brown preadipocytes should be employed to demonstrate that IL-35 induced UCP1 and FGF21 expression is ATF2 dependent. Another key experiment is to use IL-35 deficient Treg model to definitively demonstrate the requirement of Treg IL-35 to maintain thermogenesis. However, this can be done in a follow up study.

    We are grateful for all the insightful comment provided by Reviewer #3. We understand the concern, but we have the limitations in performing several sequential i.v. injections in our animal facility due to ethical permissions. In light of this constraint, we have devised an alternative approach to evaluate the role of IL-35 in adaptive thermogenesis.

    To address this, we conducted a cold tolerance test in both control mice and MKK3/6CD4-KO mice, which express higher levels of IL-35. Our findings revealed that MKK3/6CD4-KO mice exposed to cold conditions were able to preserve their body and brown adipose tissue (BAT) temperature, while the temperature of control CD4-Cre mice gradually dropped during the cold challenge.

    The data from this cold tolerance test support our hypothesis and demonstrate the role of IL-35 in promoting adaptive thermogenesis, leading to enhanced temperature maintenance in MKK3/6CD4-KO mice. These observations have been included in Figure 7B of the manuscript, and detailed results are available in Figure 11 of this revision plan.

    We appreciate the reviewer's valuable input, which has encouraged us to explore alternative experimental approaches to address the research question effectively.

    We agree with the reviewer #3 that using IL-35 deficient Treg model would be great approach to confirm our results, but we think that now with the additional experiments we have performed, we strength our findings that IL-35 has a novel role in controlling adipose tissue thermogenesis.

    Reviewer #3 (Significance):

    Dissipating energy as heat through brown or beige adipocyte-mediated thermogenesis is believed to be an effective way to combat obesity. The current study aims to characterize the p38 signaling pathway in T cells as a potential target to modulate browning or beiging of adipose tissues. This would be of interest to the basic biomedical research community, particularly in the area of immunometabolism. A major limitation is the concern of improper controls for the mouse models, which makes data interpretation difficult. In addition, the mechanistic studies lack in depth analyses to support the conclusion.

  2. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #3

    Evidence, reproducibility and clarity

    Nikolic et al. examine the metabolic outcome of T cell specific deletion of MKK3/6 (MKK3/6 CD4-KO), which are the main activators of p38. Previous studies have demonstrated that MKK3/6 CD4-KO leads to Treg expansion and that Tregs in adipose tissues are associated with improved metabolic homeostasis. In line with these observations, the authors show that MKK3/6 CD4-KO mice gain less weight and have more active brown fat thermogenesis on a HFD both at the room temperature and 30C housing conditions. They also find more Tregs and M2 macrophages in eWAT of MKK3/6 CD4-KO. All of the metabolic parameters are compared to CD4-cre mice as the wild type controls. Mechanistically, the authors suggest that reduced p38 activation by MKK3/6 CD4-KO leads to increased IL-35 production by Tregs, which induces beiging/browning of adipose tissues to promote metabolic health.

    The authors have spent most of the efforts conducting metabolic phenotyping of MKK3/6 CD4-KO mice. One potential issue is whether the non-littermate CD4-cre mice are the proper controls for the comparison. In addition, the mechanistic link of the IL-35-ATF2-UCP1/FGF21 axis has only been superficially addressed.

    Specific comments:

    1. It's important to use proper controls for mouse metabolic studies. The authors stated that CD4-Cre and MKK3/6 CD4-KO mice are all in the C57B/6L background. However, it would appear that these two lines were bred separately. The difference in the genetic background, despite minor, can lead to the observed phenotype, notably weight gain. Since the metabolic phenotypes seem to be driven by the weight difference, it is even more critical to include additional controls to validate the findings. For instance, crossing MKK3/6 f/f with one copy of CD4-Cre with MKK3/6 f/f to generate age-matched MKK3/6 CD4-KO and MKK3/6 f/f controls should be used to repeat major in vivo studies similar to those in Fig. 2-4.
    2. The assessment of adipose tissue immune cell population in Fig. 5 was conducted after HFD-induced obesity. As mentioned above, the change in Treg and M2 cell percentage could be due to the body weight difference. The experiment should be repeated (with proper controls) in normal chow and after a few weeks of HFD when Treg numbers start to decline.
    3. Data related to the mechanistic link in Fig. 6/7 are not robust and require a large amount of additional work to substantiate the claim. First of all, the role of IL-35 in BAT thermogenesis remains unclear. It's somewhat surprising to see a single dose of IL-35 i.v. injection is sufficient to increase BAT temperature in Fig. 7B. Minimally, the authors need to demonstrate that IL-35 treatment (perhaps after a few daily doses) is able to increase browning/beiging of fat cells and improve cold tolerance when placing the mice at 4 degree of several hours (and up to 3 days). Serum FGF21 level should also be measured after/during IL-13 treatment. Secondly, ATF2 knockout or knockdown in brown preadipocytes should be employed to demonstrate that IL-35 induced UCP1 and FGF21 expression is ATF2 dependent. Another key experiment is to use IL-35 deficient Treg model to definitively demonstrate the requirement of Treg IL-35 to maintain thermogenesis. However, this can be done in a follow up study.

    Significance

    Dissipating energy as heat through brown or beige adipocyte-mediated thermogenesis is believed to be an effective way to combat obesity. The current study aims to characterize the p38 signaling pathway in T cells as a potential target to modulate browning or beiging of adipose tissues. This would be of interest to the basic biomedical research community, particularly in the area of immunometabolism. A major limitation is the concern of improper controls for the mouse models, which makes data interpretation difficult. In addition, the mechanistic studies lack in depth analyses to support the conclusion.

  3. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #2

    Evidence, reproducibility and clarity

    This manuscript by Nikolic et al sought to investigate the role of p38 activation in adipose tissue Treg cells and obesity. They found that the expression of p38a, its upstream kinase MKK6, and downstream substrate ATF2 was upregulated specifically in adipose T cells associated with human obesity. They generated T cell-specific knockout MKK3/6 in mice and found these animals were protected from diet-induced obesity as a result of increased BAT thermogenesis. Mechanistically, loss of p38a activation promoted adipose tissue accumulation of Treg cells, leading to elevated IL-35 availability and UCP1 expression.

    Major comments:

    1. They attributed the obesity protection to energy expenditure; however, food intake and intestinal absorption were never tested. Immune cells particularly Treg cells are important modulates of nutrient uptake.
    2. At thermoneutrality, BAT is inactive even though UCP1 expression is still present (not activated). MKK3/6 deficiency in T cells still confer protection against obesity at thermoneutrality suggests it regulates other energy balance components in addition to BAT thermogenesis.
    3. Loss of adipose Treg cells (such as Pparg KO, Foxp3-DTR) did not lead to obvious obesity phenotypes. Gain-of-function Treg cells (such as adoptive transfer, IL-2/IL-2 Ab) did not results in profound obesity protection as observed in MKK3/6 CD4-KO mice. It suggests that MKK3/6 KO in T cells causes other immune defects (besides Tregs).
    4. The increase in IL-35 seemed to be very moderate, compared to the metabolic phenotypes. It raises the question if IL-35 is responsible for BAT activation and reduced weight gain. It is unclear what systemic and local levels of IL-35 were reached after recombinant IL-35 treatment (Fig. 7B). IL-35 antibody blockade experiment in KO mice is recommended.
    5. IL-35 induced p-ATF2 is acute and transient (Fig. 7D) and it was able to increase BAT temperature in just 4 h (Fig. 7B). However, Ucp1 transcription and translation generally take much longer time (e.g. 2d in Fig. 7C). IL-35 may increase energy expenditure through UCP1-independent mechanisms.

    Minor comments:

    1. The gating of Treg cells should exclude CD25- cells. Single positive (CD25+ or Foxp3+) cells are progenitors of Tregs. In addition to number, phenotypic activation of Treg cells should also be determined.
    2. ATF is also important for adipogenesis, is the adipogenic differentiation of BAT SVF cells affected by MKK3/6 KO or IL-35 treatment?
    3. Metabolic cage experiments are desired to determine whole-body energy balance, including food intake, physical activity, and heat production.
    4. Total UCP1 expression (both RNA and protein) in the whole BAT from an animal should determined (since BAT is smaller in KO mice).
    5. Fig. 6C, IL-35-expressing Treg cells should be quantified from adipose tissue.

    Referees cross-commenting

    I agree with Reviewer #1. In addition to energy metabolism and mechanistic action of IL-35, more rigor characterization of adipose Treg cells is needed.

    Significance

    The manuscript is innovative in define the novel role of p38 activation in the T cell compartment and its metabolic regulation. The involvement of Treg cells in adipose tissue homeostasis has been well documented and Treg cell-derived IL-35 has been demonstrated in immune regulation. The authors provided a relatively thorough description of the altered metabolism in these Mkk3/6 CD4-KO mice; however, the reviewer has doubts if Treg cells and IL-35 are primary mechanisms of the observed protection from obesity. The manuscript would be much stronger if the model were Treg cell-specific KO and/or IL-35 deficiency in Treg cells reverses obesity resistance conferred by MKK3/6 deficiency. It also suspected that BAT thermogenesis is not the major reason, as BAT deficiency or UCP1 KO results in much milder phenotypes in mice, even at thermoneutrality.

  4. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #1

    Evidence, reproducibility and clarity

    In this study, Nikolic et al. show a novel role for p38 signaling in Treg cells, which impacts adipocytes through IL-35. This mechanism seems to be important for adipose tissue browning and metabolic health and could be potentially therapeutically exploited.

    Major comments:

    1. A control group of mice fed chow diet is needed to distinguish the effects of the genotype from those caused by diet. What is the phenotype of regular chow-fed mice in terms of energy metabolism and thermogenesis?
    2. While an increase in BAT temperature (as demonstrated here by infrared imaging) in line with increased thermogenesis, it will be critical to verify this hypothesis by indirect calorimetry. Energy expenditure, food intake, and activity measures should be added for regular and DIO mice. Please follow the guidelines for ANCOVA analysis and measurements explained in PMID: 22205519 and PMID: 21177944.
    3. That the phenotype is still seen at isothermal housing is interesting but should be backed up by direct assessment of thermogenic capacity (see PMID: 21177944). In the end, it could also be increased heat loss, independently of heat production. If the browning is cause or consequence remains unclear, then.
    4. Regarding the in vitro data, a thermogenic phenotype should be functionally verified by Seahorse analysis.
    5. Mechanistically, there is epistasis type of experiment that IL-35 influences Ucp1 levels via ATF2 as the data remain associative in nature.
    6. What are other consequences of injecting IL-35? Is it good or bad? What is the therapeutic potential in DIO mice? Also, in these experiments (Fig. 7) indirect calorimetry as described would be supportive of the claims.

    Minor comments:

    1. The authors claim that their HFD-fed MKK3/6CD4-KO mice are protected against hyperglycemia, but only fasted/fed blood glucose tests are performed. Lower glucose levels could be explained due to a hyperinsulinemic state in response to growing insulin resistance in the presence of HFD. It would be sensible to perform both glucose and insulin tolerance tests to back up your statement.
    2. Please provide the loading control for p38 and S6 blots (Figure 6G).
    3. Statistical test from Figure 7B should be a t-test, since it is only comparing 2 variables (PBS vs IL-35), and not a 2-way ANOVA as described in the legend.
    4. Label correctly the panels in the figures -examples: Fig 3, panels C and D are interchanged; reference in the text to Fig S1G even though the figure only as panels A-F; Fig 7 legend referes to the statistical test of panel E when the figure only has A-D.
    5. There are several typos along the text, please revise (example: page 4;line 4 -"tremorgenic")

    Referees cross-commenting

    I think we three reviewers are pretty much on the same page - mouse energy metabolism explored too little and the mechanistic insight a bit thin considering the relatively strong claims.

    Significance

    The manuscript is well written, and the research conducted properly, even though a thorough analysis of energy metabolism in mice and cells is missing and the mechanistic claims are based on relatively thin data.

    The immune system and inflammation play important roles for obesity and insulin resistance, yet the roles they play in thermogenic adipocytes remains unclear. This work adds novel aspects to this relationship.

  5. Version published to 10.1101/2023.08.04.551982v1 on bioRxiv