Diabetic Wound Keratinocytes Induce Macrophage JMJD3-Mediated Nlrp3 Expression via IL-1R Signaling

Wound healing is a complex process involving coordination through the stages of homeostasis, inflammation, re-epithelization, and resolution for proper healing to occur. Macrophage (Mφ) plasticity is essential during the inflammation phase of tissue repair for wounds to transition into the resolution state (1,2). During the initial inflammatory phase, Mφs demonstrate increased production of inflammatory cytokines and inflammasome activation, while late inflammatory-phase Mφs produce other mediators important in the transition to the proliferative and resolution phases of wound healing; however, this has been shown to not occur in type 2 diabetes (T2D) wounds. In the diabetic state, wound Mφs display a prolonged inflammatory phenotype and promote low-grade inflammation associated with enhanced inflammasome activity that contributes to poor wound healing (1–6). Although keratinocytes can regulate Mφ phenotypes within the wound environment (7), the molecular mechanisms regulating Mφ activity, and, in particular, the increase in inflammasome activity, in diabetic wounds have not been explored.

Activation of the NOD-like receptor protein 3 (NLRP3) inflammasome is closely linked to driving skin inflammation in diabetic wounds (8–12). NLRP3 inflammasome activation consists of two steps, where signal 1 is the priming step and involves the induction of gene transcription of NLRP3 inflammasome components (i.e., NLRP3, caspase 1, pro-IL1β) (13). Signal 2 is the activation step; this entails the cleavage of proform caspase-1, which then cleaves and activates proinflammatory cytokines IL-1β and IL-18 (14). Elevated levels of IL-1β have been found in human and murine diabetic wound Mφs, which correlates with increased NLRP3 inflammasome activity (15–17), yet the mechanism driving the inflammasome activity has remained ill defined. Studies have identified that increased priming of Mφs, resulting in elevated inflammasome gene transcription, can lead to enhanced inflammasome activation (18,19); therefore, the priming step may play a critical role in driving the enhanced Mφ NLRP3 inflammasome activation and IL-1β release seen in pathologic diabetic wounds.

Mediators in diabetic wounds can induce increased NLRP3 inflammasome activity in monocytes (16); however, the reasons for this were not well explored. Keratinocytes are the primary cell type in the epidermis, and are involved via cross-talk with immune cells in the dermis to regulate inflammation during wound healing (20). Following cutaneous injury, keratinocytes release chemokines and inflammatory cytokines that can activate and recruit immune cells to the injury site (7,21–25). In particular, interleukin 1α (IL-1α) is constitutively expressed and induced by keratinocytes following skin injury (26,27). The IL-1α precursor, unlike IL-1β, is biologically active and can bind to the IL-1 receptor (IL-1R) to induce skin inflammation after being released from cells (28,29). IL-1R signaling can regulate Mφ Nlrp3 expression; however, whether IL-1α is dysregulated in keratinocytes in the context of diabetic wounds and the mechanism by which it can regulate Mφ Nlrp3 expression are unknown.

Our group and others have shown that epigenetic regulation plays an essential role in Mφ phenotype during wound healing by suppressing or promoting specific genes important for inflammation (30). Jumanji domain-containing protein-3 (JMJD3), a histone demethylase that regulates the trimethylation of histone H3 on lysine 27 (H3K27me3), has been shown to regulate Nlrp3 expression in Mφs during colitis (31), but this has not been explored in the context of diabetes or wound repair. Further, our group recently identified that JMJD3 is upregulated in diabetic wound Mφs (32,33); however, the influence of keratinocytes on this pathway as well as upstream regulation and the downstream gene expression patterns related to NLRP3 were not examined.

Here, we demonstrate that diabetic wound keratinocytes induce Mφ JMJD3-mediated Nlrp3 expression via IL-1R signaling. We found that stimulation of Mφs with diabetic wound keratinocyte-conditioned media (KCM) induced increased Nlrp3 expression and enhanced inflammasome activation through an IL1-R signaling pathway. Further, we found that IL-1α is increased in human and murine diabetic wound keratinocytes and drives increased Nlrp3 expression in Mφs. Using an epigenetic PCR array on Mφs treated with diabetic wound KCM, we identified that JMJD3 was the most significantly altered enzyme, and pharmacologic inhibition of JMJD3 in Mφs decreased Nlrp3 expression following IL-1α stimulation and treatment with diabetic wound KCM. We then examined this in vivo using our Mφ-specific JMJD3-deficient mice (Jmjd3^f/flyz2^Cre+) and found that JMJD3 regulates Nlrp3 expression and inflammasome activation in wound Mφs following injury. Altogether, our data highlight the keratinocyte-Mφ cross-talk in wounds and suggest a role for IL-1α/IL-1R signaling in driving increased JMJD3-mediated Nlrp3 expression and enhanced inflammasome activation in diabetic wound Mφs.

Male C57BL/6J mice were purchased from The Jackson Laboratory and maintained in breeding pairs on a normal chow diet (13.5% kcal fat; LabDiet) in the Unit for Laboratory Animal Medicine facilities. To initiate a “diabetic” state, male C57BL/6J mice were fed a high-fat diet (60% kcal fat; Research Diets) for 12–16 weeks to generate the DIO model. Jmjd3^f/fLyz2^Cre mice on a C57BL/6J background were created as previously reported by our group to obtain male mice deficient in Jmjd3 in Mφs, monocytes, and granulocytes (32). All mouse protocols were approved by and complied with the guidelines established by the Institutional Animal Care and Use Committee at the University of Michigan, Ann Arbor, MI.

Four-millimeter punch biopsy wounds were created on mouse dorsal skin as previously described (7). In local injection experiments, Anakinra, anti-IL1α, or control were injected subcutaneously at two points along the wound edge, as previously described by our group (7). For the acute wound healing model, photographs were taken following injury, and the wound area was calculated, as described previously (7). The 6-mm wound biopsies (2-mm wound edge) were collected at the indicated time points and prepared for keratinocyte or Mφ isolation.

Primary keratinocytes were isolated from the wounds of mice as previously described (7). Following isolation and culture of primary wound keratinocytes, cells were harvested for RNA, and KCM was collected for Mφ stimulation and protein studies.

For monocyte/Mφ isolation, wounds were digested, and monocyte/Mφs (CD3-CD19-Ly6G-NK1.1-CD11b+) were isolated as described previously by our group (32). Cells were immediately saved or cultured ex vivo for RNA, cDNA, or protein studies.

Bone marrow–derived Mφs (BMDMs) were grown as detailed previously (34). On day 6, the cells were replated and stimulated with control or DIO wound KCM, IL-1α, GSK-J4, IgG, or anti-IL1α. For siRNA experiments, cells underwent transfection using Lipofectamine RNAimax (Invitrogen), as previously described (35). Briefly, 20 pmol of the nontargeting siRNA or JMJD3 pooled siRNA were added to BMDM cultures prior to KCM stimulation. siRNA information is provided in the Supplementary Material. After 5 h of stimulation, the cells were placed in Trizol (Invitrogen) for RNA analysis. After 24 h of stimulation, cells were either immediately saved for western and chromatin immunoprecipitation (ChIP) analysis or continued in culture for IL-1β ELISA experiments.

Wound Mφs and BMDMs were cultured for 4 h in RPMI with lipopolysaccharide (LPS) (100 ng/mL) and 2 h with ATP (5 mmol/L). After stimulation, the cell-free supernatant was collected and analyzed by a specific enzyme immunoassay kit for IL-1β. Primary wound KCM was analyzed by a specific enzyme immunoassay kit for IL-1α or IL-1β. All ELISA kits are from R&D Systems and were performed according to the manufacturer's instructions.

RNA was isolated from BMDMs, wound Mφs, and keratinocytes using chloroform, isopropanol, and ethanol. Superscript III Reverse Transcriptase (Thermo Fisher Scientific) kits were then used to reverse transcribe RNA to cDNA. Real-time PCR was performed with 2 × TaqMan Fast PCR mix via the 7500 Real-Time PCR System. Primers are listed in the Supplementary Material. Data were examined in a relative quantification analysis to 18S (2^–ΔΔCt). All samples were performed in triplicate.

RNA was extracted from BMDMs treated with control or DIO KCM using an RNeasy Micro kit (Qiagen), and then column DNAase digestion was used to eliminate genomic DNA contamination. After digestion, the RT2 First Strand kit (Qiagen) was used to prepare cDNA from RNA. Next, the RT2 Profiler PCR array (PAMM-085Z; Qiagen) was used for gene expression analysis of chromatin-modifying enzymes. Data were analyzed using the GeneGlobe web portal (Qiagen).

ChIP was performed on cells using an anti-H3K27me3 antibody or isotype control, as detailed previously by our group (32). H3K27me3 deposition was measured by quantitative PCR (qPCR) using 2× SYBR PCR mix (Invitrogen), and primers are listed in the Supplementary Material.

BMDM protein suspensions in lysis buffer were standardized for protein concentrations and subjected to gel electrophoresis and wet transfer as previously described (32). Nitrocellulose membranes were probed with β-actin primary antibody (8H10D10; Cell Signaling) or NLRP3 primary antibody (D4D8T; Cell Signaling) diluted to 1:1,000 and 1:500, respectively, in 5% BSA in Tris-buffered saline with Tween buffer, followed by incubation with anti-mouse IgG or anti-rabbit IgG horseradish peroxidase–conjugated secondary antibody (Cell Signaling, Inc.) and then visualized with chemiluminescence (Thermo Fisher Scientific). ImageJ (National Institutes of Health) was used to calculate densitometry.

Six-millimeter punch biopsy whole wounds were excised from mice. Wound sections were fixed overnight in 10% formalin before embedding the wound in paraffin. Five-micrometer sections were stained with hematoxylin-eosin as described previously (7). Images were taken using a Zeiss Axioskop 2 microscope at 4× magnification.

Single-cell RNA sequencing (scRNA-seq) was performed on 8-mm punch biopsy samples obtained from normal and diabetic wounds, as previously described (36). The data accession numbers include GSE154557 and GSE179162 (Gene Expression Omnibus). Patient consent for collecting wound tissue was exempt by the institutional review board because the tissue was obtained from discarded surgical material (protocol no. HUM00060733).

All data were analyzed and graphed with GraphPad Prism software. Data comparing differences between two groups were obtained using a two-tailed Student t test for normally distributed groups, and Welch correction was applied for data with unequal variances. Analysis between more than two groups was done using one-way ANOVA. All P values less than 0.05 were considered significant.

Detailed methods are provided in the Supplementary Material.

All data needed to support the conclusions of the paper are present in the paper. All materials are available by request, restricted by institutional material transfer agreements.

Given that increased priming of Nlrp3 gene transcription activation in Mφs is important for enhanced inflammasome activation and keratinocytes in the skin regulate Mφ phenotype through inflammatory cytokine production following injury (7,19), we examined the role of diabetic wound keratinocytes in driving increased Nlrp3 expression and enhanced inflammasome activation in Mφs. First, to establish the kinetics of Nlrp3 expression in Mφs from diabetic wounds, we generated a diet-induced obesity (DIO) murine model of T2D. C57BL/6J DIO mice were wounded with a 4-mm punch biopsy, and Nlrp3 expression was assessed in sorted wound Mφs (CD3-CD19-Ly6G-NK1.1-CD11b+) isolated on days 1, 3, and 5 following injury. We found that Nlrp3 is upregulated by day 5 in DIO wound Mφs, confirming the increased regulation of Nlrp3 gene expression within the diabetic wound environment takes place late following injury (Fig. 1A). Further, Nlrp3 expression is significantly increased in Mφs isolated from DIO wounds compared with control on day 5 post-injury (Fig. 1B). A trending increase in Nlrp3 expression was detected in spleen Mφs (Supplementary Fig. 1). To understand the translational relevance, we examined human Mφs previously analyzed by our group using scRNA-seq from non-T2D and T2D chronic wounds (7). Similar to others, we see that NLRP3 was increased in Mφs from chronic wounds of T2D patients (Fig. 1C). To determine the relevance of keratinocytes in inducing the increased Mφ Nlrp3 expression seen in diabetic wounds, total wound keratinocytes were isolated and cultured on day 5 following injury from control and DIO mice. This period is consistent with the day Nlrp3 expression is elevated in DIO wound Mφs. Following the culture of the wound keratinocytes, the KCM were collected and added to wound Mφs or BMDMs. We identified a significant increase in Nlrp3 expression in wound Mφs and BMDMs stimulated with DIO wound KCM compared with controls (Fig. 1D and E). NLRP3 protein was also increased in BMDMs stimulated with DIO wound KCM compared with control (Fig. 1F). These data suggest that keratinocytes play a role in priming Mφs for increased Nlrp3 gene expression levels within the diabetic wound environment late following injury.

Since IL-1R signaling can activate Nlrp3 transcription in Mφs, we then examined the extent to which diabetic wound keratinocyte induction of Mφ Nlrp3 expression relied on the IL-1R signaling pathway. To determine this, we stimulated BMDMs from IL-1R knockout (KO), MyD88 KO, and wild-type (WT) mice with DIO wound KCM. We observed that KO of the IL-1R and MyD88 in BMDMs resulted in decreased Nlrp3 expression following stimulation with DIO wound KCM compared with WT, suggesting the induction of Nlrp3 expression in Mφs occurs in part via IL-1R/MyD88 signaling (Fig. 1G). We next examined whether this increased priming of Nlrp3 gene transcription in Mφs resulted in enhanced inflammasome activation. To activate the inflammasome, Mφs were stimulated with LPS and ATP following the addition of wound KCM, and the release of IL-1β was used to track inflammasome activation changes. We noted the enhanced release of IL-1β by BMDMs stimulated with DIO wound KCM compared with control and media alone. Additionally, IL-1R KO BMDMs displayed decreased IL-1β release following stimulation with DIO wound KCM, indicating that the enhanced inflammasome activation occurs via IL-1R signaling (Fig. 1H). To further evaluate the role of IL-1R signaling in regulating Mφ Nlrp3 expression and enhanced inflammasome activation in diabetic wounds, DIO mice were subcutaneously injected with PBS or Anakinra (an IL-1R antagonist) following injury. We observed that local injection of Anakinra resulted in decreased Mφ Nlrp3 expression and inflammasome activity, indicated by decreased release of IL-1β (Fig. 1I and J). However, local injection of Anakinra did not result in improved healing. This may be due to Anakinra’s shortened half-life and bioactivity following injection (Supplementary Fig. 2). Together, these data show that Nlrp3 is increased late following injury in DIO wound Mφs, and keratinocytes within the diabetic wound environment can induce Nlrp3 expression and enhance inflammasome activation in part via IL-1R signaling.

Since IL-1α is constitutively expressed by keratinocytes and signals through the IL-1R to induce tissue inflammation (26,27), we investigated the differences in IL-1α levels between keratinocytes from normal and diabetic wounds. We noted that Il1a was increased in DIO keratinocytes isolated from day 5 wounds (Fig. 2A). These cells also demonstrated increased release of IL-1α, but not IL-1β, into the KCM compared with control (Fig. 2B and Supplementary Fig. 3A). Using scRNA-seq from non-T2D and T2D chronic wounds, we found that IL1A was increased in keratinocytes from chronic wounds of T2D patients (Fig. 2C). While IL1B was increased in T2D chronic wound keratinocytes compared with control, a smaller percentage of keratinocytes in T2D chronic wounds express IL1B (Supplementary Fig. 3B). Since IL-1α is increased in DIO wound KCM, we then examined the ability of IL-1α to prime Nlrp3 transcriptional activation in Mφs. We found that stimulation of BMDMs with IL-1α increased Nlrp3 expression (Fig. 2D). To directly examine the effects of IL-1α in DIO KCM on Mφ Nlrp3 expression, we added anti-IL1α or IgG control antibodies to BMDMs following stimulation with DIO KCM. We observed a significant decrease in Nlrp3 expression in BMDMs after the addition of anti-IL1α antibodies following stimulation with DIO KCM compared with control (Fig. 2E), suggesting that keratinocyte IL-1α production may contribute to increased Mφ Nlrp3 expression in chronic diabetic wounds. To further evaluate the role of IL-1α signaling in regulating Mφ Nlrp3 expression in diabetic wounds, DIO mice were subcutaneously injected with anti-IL1α or IgG control following injury. We observed that local injection of anti-IL1α in DIO mice resulted in decreased Mφ Nlrp3 expression and improved healing compared with control (Fig. 2F and G).

Since JMJD3 has previously been identified as one of the most relevant epigenetic enzymes in driving the diabetic Mφ inflammatory profile late post-injury and has been shown to regulate Nlrp3 expression in Mφs during colitis (31–33), we sought to determine the extent to which IL-1α induced Mφ Nlrp3 expression via a JMJD3-mediated mechanism. Following stimulation of BMDMs with IL-1α, we identified that Jmjd3 expression was significantly increased (Fig. 2H). Furthermore, addition of GSK-J4, a JMJD3 inhibitor, to BMDMs following IL-1α stimulation resulted in decreased Nlrp3 gene expression and protein levels (Fig. 2I and J), suggesting that increased IL-1α produced by DIO wound keratinocytes late post-injury contributes to priming Nlrp3 gene transcriptional activation in Mφs via a JMJD3-mediated mechanism.

Since diabetic wound keratinocytes produce increased IL-1α, and JMJD3 plays a central role in IL-1α–mediated Nlrp3 expression in Mϕs, we examined whether diabetic wound keratinocytes induced Nlrp3 transcriptional activation via a JMJD3-mediated mechanism. JMJD3 regulates gene expression by demethylating H3K27, so we first examined Mφs following stimulation with DIO and control wound KCM for changes in H3K27me3 levels at the Nlrp3 promoter by ChIP. We found a significant decrease in the repressive H3K27me3 mark at the Nlrp3 promoter in BMDMs stimulated with DIO wound KCM compared with control and media only (Fig. 3A). Since several methyltransferases and demethylases are specific for H3K27, we performed an epigenetic PCR array on BMDMs stimulated with control or DIO wound KCM. We observed Jmjd3 (Kdm6b) was the main epigenetic enzyme that was increased following stimulation with DIO wound KCM (Fig. 3B). Quantitative PCR analysis also revealed a significant increase in Jmjd3 expression in DIO wound KCM-stimulated BMDMs (Fig. 3C). Interestingly, we found that pharmacologic inhibition (GSK-J4) or siRNA knockdown of JMJD3 following the addition of DIO wound KCM to BMDMs resulted in a significant decrease in Nlrp3 expression (Fig. 3D and Supplementary Fig. 4). Concurrently, this decrease was also noted at the protein level (Fig. 3E). Taken together, these data demonstrate that diabetic wound keratinocytes late post-injury can induce Nlrp3 gene transcription in Mφs via a JMJD3/H3K27me3-mediated mechanism.

Our group has demonstrated that Mφ-specific inhibition of JMJD3 improves diabetic wound healing (32). Additionally, as previously identified, we see increased Jmjd3 expression in DIO wound Mφs late following injury (Fig. 4A). To identify whether JMJD3 could regulate Mφ Nlrp3 gene expression in vivo, we isolated Mφs from DIO and control mice from day 5 wounds, and examined changes in H3K27me3 at the Nlrp3 promoter by ChIP. We found a significant decrease in the H3K27me3 mark at the Nlrp3 promoter in DIO wound Mφs (Fig. 4B). Next, to confirm the relevance of JMJD3 in regulating Nlrp3 transcriptional activation in Mφs during wound repair, we utilized the Cre recombinase Lox-P system to generate a myeloid-specific, JMJD3-deficient mouse (Jmjd3^f/fLyz2^Cre+). Wound Mφs isolated on day 5 following injury from Jmjd3^f/fLyz2^Cre+ mice exhibited a decrease in Nlrp3 expression compared with the control mice (Fig. 4C), and ChIP analysis of the Nlrp3 promoter showed an increase in H3K27me3 levels in these Mφs (Fig. 4D). These data indicate that the increased Nlrp3 expression observed in DIO wound Mφs late following injury is mediated by JMJD3/H3K27me3 activity. Finally, we explored the effect of Mφ specific loss of JMJD3 on inflammasome activation during wound repair by examining IL-1β release from wound Mφs on day 5 and found decreased release of IL-1β by Jmjd3^f/fLyz2^Cre+ mice compared with control (Fig. 4E). Taken together, these data suggest a central role for JMJD3 in regulating Mφ NLRP3 inflammasome hyperactivity noted in chronic diabetic wounds.

NLRP3 inflammasome activity is elevated in diabetic wound Mφs and contributes to impaired wound healing (5,6); however, the molecular mechanism regulating this is unclear. In this study, we showed that Mφ Nlrp3 expression is increased late post-injury in diabetic wounds, and stimulation of wound Mφs with conditioned media from keratinocytes isolated from diabetic wounds late following injury results in both increased Nlrp3 expression and enhanced inflammasome activation that is regulated in part by IL-1R signaling. Next, we identified that IL-1α is increased by human and murine diabetic wound keratinocytes, and neutralization of IL-1α inhibits induction of Mφ Nlrp3 expression by diabetic wound keratinocytes. Furthermore, we demonstrated that diabetic wound keratinocytes induce Nlrp3 expression through a JMJD3/H3K27me3-mediated mechanism. Significantly, Mφ-specific KO of JMJD3 (Jmjd3^f/flyz2^Cre+) reduced Nlrp3 expression and inflammasome activity in wound Mφs late following injury. Together, our data suggest a role for keratinocyte IL-1α/IL-1R signaling in driving the JMJD3-mediated increased Nlrp3 expression and enhanced inflammasome activity seen in chronic diabetic wound Mφs (Fig. 5).

Current literature has demonstrated an important role for epigenetic modifications, including histone modification, DNA methylation, and RNA interference in regulating proper wound healing through controlling structural and immune cell phenotype (2,30,37). In particular, our group and others have examined the role of histone demethylase JMJD3 on Mφ phenotype and keratinocyte re-epithelization during wound healing (32,33,38–40). In normal wound healing, JMJD3 drives Mφs toward an inflammatory phenotype early after injury, whereas, in diabetic wounds, this occurs late following injury contributing to impaired wound healing. In this work, we found that JMJD3 is important for regulating the increased Nlrp3 expression and the resulting enhanced inflammasome activation seen in chronic diabetic wounds. In diabetic wound Mφs, Jmjd3 expression can be regulated by janus kinase/signal transducer and activator of transcription and MyD88 signaling (32,33). This pathway differs depending on the ligand and the downstream targeted genes of interest. This study suggests a role for IL-1α/IL-1R signaling via MyD88 in driving Mφ JMJD3-mediated NLRP3 inflammasome activity in diabetic wounds.

Similar to others, we show Mφs in diabetic wounds have increased Nlrp3 expression and inflammasome activity, suggesting these cells are primed for enhanced inflammasome activation. Our data identify that keratinocytes within diabetic wounds late after injury can drive increased Nlrp3 transcriptional activation in BMDMs and wound Mφs. While BMDMs display a stronger response, this difference may occur because BMDMs are stimulated with KCM from baseline. This increased priming of Mφs by diabetic wound keratinocytes results in enhanced inflammasome activation in the presence of pathogen-associated molecular patterns. To this end, diabetic patients exhibit increased polymicrobial colonization, which is associated with impaired healing (41). Neutrophils are also associated with contributing to Mφ NLRP3 inflammasome activation at the posttranslational modification level through the induction of reactive oxygen species in Mφs by the secretion of neutrophil extracellular traps (42), which are increased in diabetic injured skin (43).

Following injury, keratinocytes secrete NFkB-regulated cytokines and type I IFNs (T1IFN) that can regulate immune cell phenotype (7,21–25). Our group has previously shown that IFNκ, a keratinocyte-produced T1IFN, is increased early during normal wound healing and is decreased in diabetic wound keratinocytes (7). While this decreased keratinocyte IFNκ contributes to the impaired Mφ inflammatory phenotype seen initially in diabetic wounds, the factors driving the delayed chronic increase in Mφ inflammatory phenotype have remained undefined. This study demonstrates that IL-1α is increased in diabetic wound keratinocytes late following injury, which is significant since IL-1α is associated with driving skin inflammation (26,27,44). Our study suggests IL-1α drives inflammation through priming the Mφ NLRP3 inflammasome in diabetic wounds.

Although this study provides insight into the mechanisms behind the dysregulation of Mφ NLRP3 inflammasome activity in diabetic wounds, a few limitations must be addressed. First, we acknowledge that IL-1α is produced by other cell types, such as activated Mφs; however, nonhematopoietic cells are considered the primary source of pathogenic IL-1α, and we show it is elevated in diabetic wound keratinocytes. Second, while this paper focuses on changes in Mφ Nlrp3 expression in diabetic wounds, we see a trending increase in Nlrp3 expression in spleen Mφs; however, it is unclear whether keratinocytes are the source of this systemic effect. Third, we acknowledge that no cre-specific transgene line is perfect for Mφs, and that using the Lyz2Cre system generates a myeloid-specific KO of JMJD3. While this paper focused on Mφs, other cell interactions could also be affected within the Lyz2Cre system. Fourth, even though JMJD3 appears to regulate Mφ Nlrp3 expression via an H3K27-mediated mechanism, we recognize that other epigenetic enzymes may also regulate Nlrp3 gene expression. Finally, although IL-1α influences Mφ Nlrp3 expression, further studies assessing the role of IL-1α in Mφs using ChIP sequencing and RNA sequencing would be useful to determine other genes influenced by IL-1α.

In summary, our study suggests an important role for keratinocytes within the diabetic wound environment in inducing Mφ NLRP3 inflammasome activity in chronic diabetic wounds. To our knowledge, this study is the first to examine keratinocyte-produced IL-1α in both normal and diabetic wounds. It provides important mechanistic information that IL-1R signaling regulates JMJD3-mediated Nlrp3 transcriptional activation. Given that JMJD3 plays a central role in regulating NLRP3 inflammasome activity during wound healing, Mφ-specific targeting of JMJD3 as well as IL-1α/IL-1R signaling may be a viable therapeutic for decreasing NLRP3 inflammasome-mediated inflammation in chronic diabetic wounds.

Acknowledgments. The authors thank Robin G. Kunkel for the artwork. The authors thank the Taubman Institute Innovative Program at the University of Michigan for our scRNA-seq analysis.

Funding. S.J.W. has received funding from the National Institutes of Health grants K99-DK133828-01 and K12-GM111725, and the Michigan Nutrition Obesity Research Center grant P30-DK089503. L.C.T. and J.E.G. have received funding from the National Institute of Health grant P30-AR075043. K.A.G. has received funding from the National Institutes of Health grants R01-HL137919, R01-DK124290, R01-AR0798630, R01-HL156274-01A1, and R01-DK12753101A1.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. S.J.W. designed the experiments, collected research data, and wrote, reviewed, and edited the manuscript. C.O.A., J.Y.M., A.D.J., W.J.M., E.C.B., K.M., G.S.d.J., S.R., S.B., Z.A., and R.W. contributed to research data, and reviewed and edited the manuscript. J.M.K. contributed to designing experiments, and reviewed and edited the manuscript. L.C.T. and J.E.G. contributed to research data, and reviewed and edited the manuscript. K.A.G. contributed to the experiment design, and wrote, reviewed, and edited the manuscript. S.J.W. and K.A.G. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Parts of this study were presented in abstract form at the 83rd Scientific Sessions of the American Diabetes Association, San Diego, CA, 23–26 June 2023.