TET2-mediated epigenetic modification promotes stress senescence of pancreatic β cells in type 2 diabetes mellitus

Reviewer #1 (Public review):

Summary:

In this manuscript, the authors set out to determine how a DNA demethylation enzyme TET2 regulates beta cell senescence in the context of Type 2 Diabetes and aging. They analyze public RNA-seq data and found upregulation of TET2 coincident with downregulation of MOF and PTEN, genes involved in chromatin regulation and cell cycle. TET2 is upregulated during aging, high-fat diet feeding, high glucose on rat beta cell line INS1E, and in leptin receptor deficient (db/db) mice islets. This was not found for TET1 and TET3. TET2 global KO mice show improved glucose tolerance during aging, but not TET1 or TET3. The authors show improved beta cell identity genes in TET2 KO islets. They they performed DNA methyalation/hydroxymethylation analyses of TET2 KO transformed rat beta cell line INS1E followed by ChIP-seq of Histone H4K16 acetylation to find this mark relies on TET2 expression. Finally they demonstrate in the cell lines that overexpressing TET2 leads to loss of MOF and increased PTEN and p16, linking TET2 to a regulatory mechanism with these factors that may influence senescence.

Strengths:

The study uses a number of orthogonal approaches and evidence from cell lines and the genetic TET2 KO as well as primary islets. The concept is interesting and potentially useful to the field. Efforts were made to examine TET1 and TET3 paralogues to rule out their compensation.

Weaknesses:

The study has several major weaknesses that mean the data presented did not fully support the main conclusions. These include the following:

(1) From the beginning of the manuscript the authors first sentence does not seem to indicate which datasets were analysed, the rationale behind why public datasets were used and what the main conclusions are being drawn from the plots shown throughout Fig. 1. This section of the manuscript was very hard to follow, and lacked rationale and explanation as to what these data show.

(2) All of the metabolic phenotypic data come from global TET2 KO mice, where TET2 is lost from all cells. The authors need to use a beta cell-specific KO of TET2 to ensure that metabolic changes are not due to cross-talk with other tissues (e.g. liver, adipose, even effects on central control of metabolism). No insulin tolerance tests were done to ascertain phenotypes in other metabolic tissues. This was a major weakness of the study. The authors should also provide clear validation of their global TET2 KO mice demonstrating a total lack of protein in islets and metabolic tissues.

(3) TET2 localization and expression pattern in islets was not clearly demonstrated and the data shown are not convincing from Fig 3 and Fig 4. In Fig 3e the staining for TET2 in green looks ubiquitous in acinar tissue (not nuclear) and not in the islet. In Fig 4d there is an increase in nuclear stain shown during aging, but no INS stain is used to show specificity to beta cells. Thus there is not sufficient data to support the expression pattern and localization of TET2 and specificity of the antibody.

(4) In Fig. 5: The effect sizes for the beta cell identity gene expression differences by qRT-PCR between WT and TET2 KO islets shown in Fig 5 are extremely modest so as to be questionable whether they are biologically meaningful. The same is true of the senescence markers quantified from isolated islets by qRT-PCR in Fig 5f. The immunostains for Pdx1 are hard to see and signal should be quantified. The SA-Bgal staining is quantified but no representative image is shown. The p16 immunostaining is not clear and should be quantified. Given that a lack of truly specific p16 antibodies in mouse immunostainings have been a major issue for the field, the authors would be advised to demonstrate specificity of the antibody if possible on mouse KO tissue, or to at least validate the predicted increase in p16 staining comparing young versus old pancreas as has been shown in other studies.

(5) Throughout the manuscript the figures colors are difficult to see and text difficult to read. Text in the p-values above the bars on most Figures is not legible (particularly Figs 4, 5, and 9). The legends simply do not contain sufficient information to interpret the data panels. This is true from Figures 1 through 9. P-value and specific statistical tests are missing from legends as well. For instance, in Fig 6c, what is being shown in LV-Ctrl vs LV-TET2 and why are these sample labels the same for two sets of images with two different outcomes of the staining? How many cells were quantified here?

(6) There is an over-reliance on cell lines throughout the manuscript. INS1E and BTC6 are not truly representative of mature adult mouse or rat beta cells, and hence the connections between H4K16ac/MOF/PTEN and TET2 must be assessed in primary mouse or rat islets to confirm these phenotypes.

(7) In the in vitro studies of senescence markers, it is not convincingly shown that the cells are actually senescent. Even though there changes found in expression of p16 and SA-Bgal in the cultures, the authors did not evaluate key senescence phenotypes such as the actual cell cycle arrest, SASP proteins or apoptosis resistance. Are the cells actually senescent or are these markers simply increasing? Hence much of the changes driven by TET2 overexpression in the in vitro cell lines could likely changes in p16 protein but not actually a senescence phenotype. BTC6, INS1E, and MIN6 are cell lines that are transformed, and while they can undergo some senescence-like changes in response to specific stressors like lipotoxicity, DNA damage, or oxidative stress, the authors did not evaluate these, only senescence genes/proteins in otherwise unstressed cells. Thus the claim that TET2 modifies senescence of beta cells remains unsubstantiated from the in vitro studies. It was not clear how any of these studies related to beta cell senescence in T2DM where there is metabolic and/or gluco-lipotoxic stress. Although it is claimed from Fig 9 that TET2 regulates PTEN/MOF axis to regulate beta cell function, no functional data (e.g. GSIS) are shown.

(8) There were issues and difficulties with the writing in the introduction and discussion in that they did not clearly or adequately describe, discuss or interpret the main conclusions and their significance. The work is not positioned within the current state of the field and it is very difficult to follow the rationales for the study and the advances in knowledge provided.

Reviewer #2 (Public review):

Summary:
Epigenetic regulation is critical for maintaining cellular function, and its dysregulation contributes to senescence and disease. This manuscript investigates the role of TET2 in β cell aging, proposing that TET2-mediated PTEN DNA methylation promotes H4K16 acetylation (H4K16ac) through MOF, driving β cell senescence. Using TET2 inhibitors, RNA interference, lentiviral overexpression, and knockout mouse models, the authors aim to establish TET2 as a key player in β cell aging and a potential therapeutic target in type 2 diabetes mellitus (T2DM).
However, significant limitations reduce the manuscript's impact. Figures are poorly presented, with illegible fonts and unquantified staining panels, while key analyses, such as β cell specificity and senescence inducers, are missing. The rationale for focusing on H4K16ac and MOF is unclear, and the authors fail to address whether β cell identity gene changes reflect altered gene expression or mass. Additionally, critical controls, such as low-fat diet cohorts, are absent, and the writing lacks clarity and coherence. Together, these weaknesses undermine the validity of the findings.

Main Comments
Figures 1 and 2:
The fonts in Figures 1 and 2 are barely visible and should be improved for readability. Additionally, do TET2 protein levels change in mouse and human β cells with aging? Is there evidence from regression analyses using single-cell RNA sequencing on human islets that TET2 expression correlates with age-associated gene signatures in β cells? Are these correlations specific to β cells, or do they extend to other islet cell types? It would also be informative to assess whether TET2 levels increase with senescence inducers such as DNA damage agents (e.g., bleomycin, doxorubicin) or reactive oxygen species (e.g., H₂O₂).
Figure 3:
Why do TET2 protein levels appear stronger in acinar cells? Additionally, the predominant cellular localization of TET2 seems to be cytoplasmic. Can the authors clarify or expand on this observation?
Figure 4:
The data on the impact of TET2 insufficiency in vivo is compelling. There are several quality control experiments to validate their model and main hypothesis (That T2t2 expression increases with aging in beta-cells). Here, authors have the right system to validate their initial Tet2 protein dynamics in the mouse, since they have a KO mouse model. Here, it would be useful to co-stain Tet2 with insulin and glucagon, to infer the dynamics of Tet2 in the two most abundant islet cell types.
Figure 5:
The upregulation of β-cell identity genes in the KO mouse model raises an important question: Is this effect due to an actual increase in gene expression or simply a higher proportion of β cells? Quantifying β-cell mass and performing gene expression analyses on FACS-sorted β cells would help address this. Additionally, the staining panels lack quantification. For instance, GLUT2 staining appears cytoplasmic when it should be membranous. The authors focus on cellular senescence, but does apoptosis increase in wild-type mice under a high-fat diet (HFD)? Including animals on a low-fat diet (LFD) for comparison would add valuable context.
Figure 6:
The data suggest an increase in cell numbers in TET2-overexpressing cells. Does this indicate an effect on β-cell proliferation? Quantification would provide clarity.
Figure 8:
The rationale for focusing on H4K16ac is insufficiently discussed. What is the mechanism linking TET2-induced changes to decreased H4K16ac levels? Including a more thorough explanation in the introduction and discussion would enhance the manuscript.
Figure 9:
The introduction lacks any discussion of H4K16ac or MOF. The discussion paragraph (lines 530-540) that elaborates on these points should instead be moved to the introduction to improve the manuscript's flow. Furthermore, the authors should cite their 2022 paper on H4K16ac as part of the rationale for focusing on this histone modification.

Minor Comments:
The manuscript would benefit from language refinement. Examples include:
Line 183: Replace "the blood included" with a more precise description.
Line 315: "treated with RNA seq" should be rephrased to clarify methodology (e.g., "analyzed via RNA sequencing").
Line 456: Replace "expression of H4K16ac" with "levels of H4K16ac."
Line 496: The phrase "can solve scientific problems from multiple dimensions" sounds vague and overly broad; consider rephrasing to be more specific.

Reviewer #3 (Public review):

Summary:
This study advances the field of β cell dysfunction by unveiling an epigenetic mechanism of β cell senescence. By connecting TET2-mediated DNA methylation to histone acetylation and cellular aging, it opens promising new avenues for therapeutic intervention. In particular, the authors aimed at identifying the mechanisms of pancreatic β cell senescence by epigenetic regulation. They conclude that increased TET2 expression in β cells is associated with ageing and metabolic dysfunction in type 2 diabetes by inducing β cell senescence. The authors further propose that TET2-mediated PTEN promoter methylation promotes β cell senescence by regulating H4K16ac. Last, the authors suggest that this could represent new molecular mechanism and therapeutic target against β cell senescence during type 2 diabetes.

Strengths:
The major strengths of the study are the use of both biased and unbiased experimental tools to approach the topic. The authors also provide in vivo and in vitro mechanistic approaches to answer their questions. All of these approaches are valuable and provides robustness to their study. The authors provide solid evidence that TET2 is associated with ageing and that its absence improves glucose metabolism in ageing and β cell senescence. In addition, the mechanistic studies showing that TET2 regulates the PTEN/MOF/H4K16ac signaling pathway in β cell lines is convincing.

Weaknesses:
Although the use of such a variety of tools is a strength, the outcome of each individual tool is somehow superficial. For instance, the authors focus on very specific targets emanating from their omics studies without a clear or logical justification. In addition, the metabolic studies are inaccurate and the authors do not follow an understandable and rational examination of the ageing versus their obesity cohorts. Last, the mechanistic studies using model cell lines are not validated in the available mouse models.

In my opinion, the evidence that TET2 regulates β cell senescence during obesity is not very strong. This is because the effect of deletion of TET2 in senescence markers is the same under 24weeks of age or 52 weeks of age (16 weeks HFD). Both ageing and HFD promoted the same extent of reduction of senescent markers and increase in β cell markers in the absence of TET2. There is no comparison between young glucose tolerant mice and old glucose intolerant mice. There is also no direct comparison of aged matched lean or obese mice. It may seem as if the mechanism by which TET2 regulates senescence in β cells is independent of the diabetic status but it is more related to ageing. Given that there is evidence that TET2 expression in β cells coordinates inflammatory responses in autoimmune diabetes, it would have been interested to check whether this is also the case for T2DM. Also, considering that expression of TET2 in Figure 3 does not seem to be in β cells in db/db mice but rather in the exocrine pancreas. In addition, senescent marker p16 in Figure 5 in the presence of TET2, seems to be localized in alpha cells or immune cells but not in β cells.
Regarding the mechanistic studies, the authors convincingly show that TET2 regulates the PTEN/MOF/H4K16ac signaling pathway in β cell lines and that this is important for β cell senescence. However, there is no validation of whether this holds true in aged, or prediabetic, mice. Given the availability of mice and model samples, this should be possible and meaningful. Last, in the genome-wide bisulfite sequencing (Figure 7), it seems that the authors are cherry picking for PTEN and in the RNAseq, the same applies for MOF. Thus, although the mechanism seems valid, the lack of in vivo validation, and a proper rational for the selected targets in the omics studies, renders the mechanistic studies rather correlative.

In sum, I believe that the study in its current version, unfortunately, does not bear the conceptual advance or the robustness that is required to offer a strong impact on the field. The methods, on the other hand, mainly the omics analyses provided here, could be of potential benefit for the field of epigenetics in β cell biology. However, in the benefit of the current study, the relevance of this data could be more rigorously assessed experimentally. I believe that the study has the potential to provide the required impact, should the authors work on it further to provide more solid functional and mechanistic validation.