(WHS-P3.04) Comprehensive transcriptional characterization of macrophage and monocyte dysregulation in diabetic wound healing
Friday, May 17, 2024
10:30 AM – 11:30 AM East Coast USA Time
Background Diabetes is a huge and growing public health concern. Monocytes (Mo) and macrophages (Mp) play major roles in all phases of wound healing and are particularly affected by diabetes-induced transcriptomic impairments. We characterize the role of Mos and Mps in genetically altered and high-fat diet induced diabetic wound healing models. Methods Full-thickness excisional wounds were created on the dorsum of C57/BL6 (WT) mice, WT mice fed with a high-fat (HF) diet, and leptin-receptor deficient (DB) mice. Tissue was explanted at post-operative days (PODs) 0, 2, 7 and 30 and processed for high-throughput single cell RNA sequencing. Mps and Mos were identified with the canonical cell markers Cd45, Cd68 and Ly6c. Human tissue was collected from non-diabetic, and diabetic wounds and processed for scRNA-seq. Results Both genetic and pathophysiologic diabetes significantly impaired wound healing compared to healthy WT mice. In the early wound healing phase, myeloid cells predominated across all groups (67%), followed by fibroblasts (15%) and epithelial cells (9%). DB demonstrated a distinct anti-inflammatory, proliferating M2 Mp cluster expressing C1qa, C1qb and Timp2 at POD 7, at 43% (vs 13% in WT, 0% in HF). HF mice overexpressed a pro-inflammatory M1 Mp cluster, defined by the genes Dusp2, Cxcl2 and Nfkbia at PODs 2 (71%, vs 2% in both WT and DB) and 7 (67%, vs 11% in WT and 5% in DB). In human diabetic wounds, we saw an upregulation of the same pro-inflammatory DUSP2+, CXCL2+, NFKBIA1+ cells by ~90% in the pre-diabetic and diabetic samples (vs 1% in the non-diabetic human). In non-diabetic WT mice, we identified a proliferating Mo cluster upregulated at POD 2 expressing Ly6c2, Vcan1 and Chil31 at 68% (vs ~9% in DB and HF). In human samples, non-diabetic wounds also demonstrated an 82% upregulation of this cell population, compared to only ~5% in pre-diabetic and diabetic wounds. Conclusion To fully understand the changes in the cellular ecology of diabetic wound healing, we investigated two popular models that imitate the human diabetic physiology. Both DB and HF models demonstrated delayed wound healing similar to diabetic patients, but we discovered that these two models exhibited very different Mo/ Mp responses. An upregulation of Mo in the pro-inflammatory wound healing phase is essential to initiate normal, non-diabetic wound healing in both humans and mice. DB mice exhibit a Mp polarization profile with anti-inflammatory, proliferating M2 Mps driving the transition into chronic wounds. Contrarily, in HF mice, enhanced and prolonged inflammation from M1 Mp proliferation delays wound closure. Confirming the human relevance of our findings, pre-diabetic and diabetic human samples showed a similar dominance of M1 Mps representing over >80% of the Mo/Mp populations. These findings also demonstrate that the HF model causes delays in wound closure through similar cellular mechanisms as seen in human diabetes.