研究目的：

基于单细胞转录组测序数据，从细胞层面剖析巨噬细胞的异质性，探索成纤维细胞表型化的巨噬细胞群。基于免疫组化和流式细胞术等实验技术，对鉴定的巨噬细胞群进行验证。

探究临床治疗手段能否影响或逆转巨噬细胞向成纤维细胞表型转变的过程。

在体外研究探究巨噬细胞的可塑性。

构建小鼠的皮肤纤维化模型，在体内评估巨噬细胞向成纤维细胞的转化能力。通过抑制TGF-β1，验证巨噬细胞-成纤维细胞是依赖TGF-β1的可逆过程。

研究方法：

对瘢痕疙瘩和正常皮肤的单细胞测序数据完成数据质控和标准化，解析瘢痕疙瘩的细胞组成。对巨噬细胞进行聚类分析，鉴定差异表达基因并将其与各个生物学过程之间进行映射，计算显著富集的生物学过程。构建瘢痕疙瘩的细胞间通讯网络，重点聚集巨噬细胞的通讯网络变化。基于免疫荧光和流式细胞术验证出现成纤维细胞表型的巨噬细胞群。

第二部分研究共招募了14名患者，其组成为：3位接受美容手术的患者（正常皮肤对照）、3位接受单纯手术切除瘢痕疙瘩患者（空白治疗对照）、3位接受术前高压氧治疗的瘢痕疙瘩患者、3位接受病灶内曲安奈德注射治疗的瘢痕疙瘩患者和2位接受病灶内曲安奈德联合5-氟尿嘧啶注射治疗的瘢痕疙瘩患者。在Illumina Hiseq X平台完成单细胞测序。对巨噬细胞的转录谱数据进行间充质激活评分。使用“PROGENy”探究治疗前后瘢痕疙瘩中信号传导途径活性变化。

使用TGF-β1诱导THP-1巨噬细胞，探究其表型的可塑性。

构建了小鼠的皮肤纤维化模型，来评估在纤维化环境下巨噬细胞向成纤维细胞的转化能力。通过使用Repsox（TGF-β受体I的抑制剂），验证TGF-β1是调控巨噬细胞-成纤维细胞转变的有效治疗靶点。

研究结果：

一共捕获了58102个细胞的单细胞转录组数据。在巨噬细胞的聚类分析中发现组成上的高度异质性。我们一共识别出簇0到簇6的7个巨噬细胞簇。簇1强烈表达的基因与细胞外基质合成、胶原纤维和骨化有关。通过拟时序分析，在巨噬细胞的谱系发育过程中检测到间充质激活的转录变化，包括ASPN、COL1A2、LUM、DCN的表达。瘢痕疙瘩的免疫荧光中检测出具有胶原合成能力的巨噬细胞。在流式细胞术上，较皮肤相比，瘢痕疙瘩中具有巨噬细胞（CD68+）和成纤维细胞标记（CD90+）的细胞从1.6±0.2%增加到4.6±0.9%。在细胞外基质蛋白的合成功能上，具有巨噬细胞（CD68+）标记和表达COL1A1的细胞从7.2±1.1%增加到10.0±1.3%，表明在功能上巨噬细胞可以直接参与胶原蛋白的分泌和沉积。在硬皮病巨噬细胞中同样发现了成纤维细胞代表性标志物COL1A1、COL1A2、DCN和LUM的表达。

在治疗前后的单细胞转录谱中，也鉴定出强烈表达细胞外基质相关基因的巨噬细胞簇。通过对基线和缓解的瘢痕疙瘩巨噬细胞进行组间差异分析，发现高压氧治疗、曲安奈德病灶内注射、曲安奈德注射联合5-氟尿嘧啶病灶内注射均可以逆转巨噬细胞的间充质激活现象，其间充质激活评分降低至正常皮肤水平。发生间充质激活的巨噬细胞簇存在TGF-β通路的高评分，高压氧治疗和病灶内药物注射均可以抑制TGF-β通路的高活性。

TGF-β1处理后的THP-1巨噬细胞中促纤维化蛋白（包括I型胶原、纤连蛋白、骨桥蛋白）的表达增加。在皮肤纤维化模型中，博来霉素诱导出胶原I+F4/80+的双阳性细胞和FN1+F4/80+的双阳性细胞，这一过程可被Repsox抑制。

研究结论：

本研究构建了瘢痕疙瘩的单细胞转录组图谱，鉴定出了独特的成纤维细胞样的巨噬细胞亚群，该亚群以巨噬细胞和成纤维细胞标志物的双阳为特点。在硬皮病中也观察到了相似的现象，巨噬细胞的成纤维细胞表型化可能是皮肤纤维化的共同机制。

临床治疗手段可以有效抑制巨噬细胞的间充质激活和向成纤维细胞转变的进程。

巨噬细胞向成纤维细胞的表型转变是可逆的，通过消耗TGF-β1的表达可以逆转这一现象。在皮肤的高纤维化环境下，巨噬细胞可能成为成纤维细胞的细胞来源，从而加速细胞外基质的沉积。调控TGF-β1可能是抑制巨噬细胞-成纤维细胞表型转变、减轻瘢痕疙瘩纤维化的有效治疗策略。

Objectives:

To dissect macrophage heterogeneity at the cellular level based on single-cell transcriptome sequencing data and to explore macrophage populations with emerging fibroblast phenotypes. To validate the explored macrophage populations based on experimental techniques including immunohistochemistry and flow cytometry.

To investigate whether clinical therapies can influence or reverse the process of macrophage to fibroblast phenotypic transition.

To explore macrophage plasticity in in vitro studies.

To construct a mouse model of dermal fibrosis and assess the ability of macrophages to transform into fibroblasts in vivo. To validate TGF-β1 as an effective therapeutic target to regulate macrophage-fibroblast transition.

Methods:

1. Data quality control and standardization were completed on single-cell sequencing data from keloid and normal skin to resolve the cellular composition. We performed cluster analysis on macrophages to identify differentially expressed genes for mapping to individual biological processes, and calculated the biological processes that were significantly enriched. The intercellular communication network of keloids was constructed, emphasizing on the changes in the communication network of clustered macrophages. The macrophage populations appearing to have fibroblast phenotype were verified based on immunofluorescence and flow cytometry.

2. 14 patients were recruited into this study, which consisted of 3 patients undergoing cosmetic surgery (normal skin control), 3 patients undergoing simple surgical excision of keloid, 3 keloid patients undergoing preoperative hyperbaric oxygen therapy, 3 keloid patients treated with intra-lesional triamcinolone acetonide injections, and 2 keloid patients treated with intra-lesional triamcinolone acetonide in combination with 5-fluorouracil co-injections. Single-cell sequencing was completed on the Illumina Hiseq X platform. Macrophage transcriptional profiling data were scored for mesenchymal activation. "PROGENy" was performed to investigate the changes in signaling pathway activity in keloid scars pre- and post-treatment.

3. THP-1 macrophages were induced by TGF-β1 to investigate their phenotypic and functional plasticity.

4. A mouse model of dermal fibrosis was constructed to assess the ability of macrophages to convert to fibroblasts in a fibrotic environment. Through the application of Repsox, an inhibitor of TGF-β receptor I, TGF-β1 was validated as a potent therapeutic target for regulating macrophage-fibroblast transition.

Results:

1. Single-cell transcriptomic data were captured for a total of 58,102 cells. A high degree of heterogeneity in composition was detected in the cluster analysis of macrophages. In total, we identified seven macrophage clusters from cluster 0 to cluster 6. Cluster 1 strongly expressed genes associated with extracellular matrix organization, collagen fiber organization, and ossification. Transcriptional changes in mesenchymal activation were detected during macrophage lineage development, represented by the expression of ASPN, COL1A2, LUM, and DCN. Macrophages with collagen synthesis capacity were detected in immunofluorescence in keloids, and cells with macrophage markers (CD68+) and fibroblast markers (CD90+) increased from 1.6±0.2% to 4.6±0.9% in keloids compared to skin on flow cytometry. Cells with macrophage (CD68+) markers and COL1A1+ increased from 7.2±1.1% to 10.0±1.3%, confirming that  macrophages can be directly involved in collagen secretion and deposition functionally. Expression of the representative markers of fibroblasts, COL1A1, COL1A2, DCN and LUM, was similarly identified in macrophages in localized scleroderma.

2. In single-cell transcriptional profiles pre- and post- keloid treatment, clusters of macrophages strongly expressing genes associated with the extracellular matrix were also identified. Analysis of intergroup differences in macrophages from baseline and remission keloids revealed that hyperbaric oxygen therapy, intralesional injection of triamcinolone acetonide, and intralesional injection of triamcinolone acetonide in combination with 5-fluorouracil reversed the phenomenon of mesenchymal activation in macrophages, and their mesenchymal activation scores were reduced to a similar level as in normal skin. Macrophage clusters with mesenchymal activation had high scores of the TGF-β pathway, and hyperbaric oxygen therapy and intralesional drug injection could inhibit the high activity of the TGF-β pathway.

3. Expression of pro-fibrotic proteins (including collagen type I, fibronectin, and osteopontin) was increased in TGF-β 1-treated THP-1 macrophages. In a skin fibrosis model, bleomycin induced a high number of collagen I+F4 / 80+ double-positive cells and FN1+F4 / 80+ double-positive cells, and this process could be inhibited by Repsox.

Conclusions:

1. In this study, we constructed a single-cell transcriptome profile of keloid scars and identified a unique fibroblast-like macrophage subpopulation characterized by double-positive macrophage and fibroblast markers. Similar phenomena have been observed in localized scleroderma, and fibroblast phenotyping of macrophages may be a common mechanism of skin fibrosis.

2. Clinical therapies can effectively inhibit the process of mesenchymal activation of macrophages and their transformation to fibroblasts.

3. The phenotypic shift from macrophages to fibroblasts is reversible and can be inverted by depleting TGF-β1 expression. In the hyperfibrotic environment of the skin, macrophages may become a cellular source of fibroblasts, thereby accelerating extracellular matrix deposition. Targeted inhibition of TGF-β1 may be an effective therapeutic strategy to inhibit the macrophage-fibroblast phenotypic transition and reduce keloid fibrosis.