Assistant Professor University of Virginia, United States
Introduction: Tissue damage due to injury, disease, or congenital defect continues to pose critical clinical challenges to human health. Biomaterial-based therapies show promise to heal this damage, but their success depends on two crucial determinants: immune response and fibrosis. In tissue repair, inflammation and fibroblast activation are tightly controlled to guide remodeling. However, if the body rejects an implanted biomaterial, a fibrotic capsule forms around the implant. Activated myofibroblasts deposit excessive amounts of extracellular matrix (ECM) components, causing scarring and hindered implant function. Understanding the mechanisms behind this fibrotic response could help improve biomaterial design and the success of implanted biomaterials.
Due to the excessive inflammation, myofibroblast accumulation, and ECM deposition in fibrosis, it is crucial to investigate both the fibroblast populations that emerge at the implant site and the immune cell populations. Investigating immune-stromal crosstalk can lead to the refinement or development of biomaterial therapies to avoid fibrotic outcomes. Thus, we examined cell populations present at the biomaterial implant site for a fibrotic or regenerative biomaterial. We used microporous annealed particle (MAP) hydrogels that have been shown to evade fibrosis and enhance tissue regeneration, comparing this system to a traditional fibrotic bulk nanoporous (NP) hydrogel of the same chemistry. Using single-cell RNA sequencing (scRNA-seq), we revealed distinct cell subpopulations up-regulated in MAP or NP implants and differentially expressed gene pathways. These findings can guide the design of future biomaterials to encourage or avoid the emergence of certain cell phenotypes to achieve regenerative outcomes.
Materials and
Methods: MAP and NP gels were fabricated from four-arm PEG-Maleimide with an RGD peptide and MMP-2 degradable crosslinker. After performing subcutaneous MAP and NP implantation in mice, samples were harvested at days 7 and 21 and cells were collected from the harvested implants via a Liberase digestion. Cell viability was confirmed to be >85% for each sample. The cells from each sample underwent automated single cell barcoding and library preparation using the 10X Genomics Chromium Platform with an 8-channel microfluidics chip. Then the libraries were used for high throughput sequencing on the Illumina NextSeq 500 Sequencing System allowing for 400M reads. The sequencing raw data was aligned, filtered, and underwent barcode and UMI counting in Cell Ranger.
With the Gene-barcode matrices, we used PCA and UMAP dimensionality reduction along with clustering approaches in our already established Seurat (v5) pipeline to identify all cell subpopulations present at the implant site, focusing on immune and stromal cells.
Using a Wilcoxon rank sum test with multiple hypothesis correction, we found differentially expressed genes for each cluster of cells. Using those differentially expressed genes, we determined the cell types of each cluster. We performed sub-clustering analysis on the fibroblast and macrophage subpopulations, using PCA and UMAP again on the respective cell subsets to reveal distinct fibroblast and macrophage clusters between MAP and NP implants. Then, we performed gene set enrichment analysis (GSEA) to investigate up-regulated pathways within these cell subpopulations.
Results, Conclusions, and Discussions: Our dimensionality reduction revealed several distinct subpopulations of cells, including macrophages, fibroblasts, mast cells, T cells, and an intermediate cell population. Gene expression analysis depicted many differentially expressed genes between samples from MAP and NP hydrogels. Initial UMAP visualization highlighted the intermediate cell population upregulated in NP implants, which displayed markers characteristic of both macrophages and fibroblasts. This population could implicate macrophages transitioning to myofibroblasts and merits future investigation.
Sub-clustering unveiled distinct macrophage subpopulations exhibiting varied cellular compositions across each implant type, alongside the presence of two fibroblast subpopulations upregulated in MAP gels. The differentially expressed genes were corroborated by GSEA to determine subpopulation phenotypes. Macrophage populations upregulated in NP implants displayed increased gene expression of immune defense; for example, one cluster expressed many genes triggered by interferon exposure, and another expressed high levels of natural killer cell receptors. In contrast, the one subpopulation of macrophages highly present in MAP displayed a migratory phenotype, potentially indicating how the higher porosity of MAP improves cell infiltration and thus regeneration. The two fibroblast populations up-regulated in MAP were associated with macrophage and T cell interaction respectively, which could signal coordination with both the innate and adaptive immune systems by fibroblasts to promote long-term healing. These analyses point to possible pathways to improve biomaterial design.
Future work will involve delving deeper into the intermediate cell population to further characterize its genotype and phenotype, which will clarify why it is almost exclusively present in NP implants. We also aim to decipher cell-cell crosstalk using Connectome and map cell trajectory with RNA velocity analysis. Together, these analyses will give insight into the dynamic cell-cell relationships that influence either regeneration or fibrosis in biomaterial implants.
In summary, our work revealed different macrophage and fibroblast subpopulations that point to important immune-stromal crosstalk for fostering regenerative outcomes of biomaterial implants.
Acknowledgements (Optional): We would like to thank the Genome Analysis and Technology Core at UVA for their help with single-cell RNA sequencing. We would also like to thank the Griffin Lab at UVA for their help with biomaterial implantation.
