Resident Memory T Cells in the Atherosclerotic Lesion Associate With Reduced Macrophage Content and Increased Lesion Stability

BACKGROUND: Tissue resident memory T (TRM) cells are a T-cell subset that resides at the site of prior antigen recognition to protect the body against reoccurring encounters. Besides their protective function, TRM cells have also been implicated in inflammatory disorders. TRM cells are characterized by the expression of CD69 and transcription factors Hobit (homolog of Blimp-1 [B lymphocyte–induced maturation protein 1] in T cells) and Blimp-1. As the majority of T cells in the arterial intima expresses CD69, TRM cells may contribute to the pathogenesis of atherosclerosis as well. Here, we aimed to assess the presence and potential role of TRM cells in atherosclerosis. METHODS: To identify TRM cells in human atherosclerotic lesions, a single-cell RNA-sequencing data set was interrogated, and T-cell phenotypes were compared with that of integrated predefined TRM cells. The presence and phenotype of TRM in atherosclerotic lesions was corroborated using a mouse model that enabled tracking of Hobit-expressing TRM cells. To explore the function of TRM cells during atherogenesis, RAG1−/− (recombination activating gene 1 deficient) LDLr−/− (low-density lipoprotein receptor knockout) mice received a bone marrow transplant from HobitKO/CREBlimp-1flox/flox mice, which exhibit abrogated TRM cell formation, whereafter the mice were fed a Western-type diet for 10 weeks. RESULTS: Human atherosclerotic lesions contained T cells that exhibited a TRM cell–associated gene signature. Moreover, a fraction of these T cells clustered together with predefined TRM cells upon integration. The presence of Hobit-expressing TRM cells in the atherosclerotic lesion was confirmed in mice. These lesion-derived TRM cells were characterized by the expression of CD69 and CD49α. Moreover, we demonstrated that this small T-cell subset significantly affects lesion composition, by reducing the amount of intralesional macrophages and increasing collagen content. CONCLUSIONS: TRM cells, characterized by the expression of CD69 and CD49α, constitute a minor population in atherosclerotic lesions and are associated with increased lesion stability in a Hobit and Blimp-1 knockout mouse model.


A
therosclerosis is a lipid-driven chronic inflammatory disease of the larger arteries resulting in vascular occlusion with clinical complications, including stroke and myocardial infarction.Various immune cell types inside the atherosclerotic lesions have been extensively studied, including T cells.2][3][4] Both CD4 + and CD8 + T cells are highly represented and exhibit a heterogeneous phenotype within the atherosclerotic lesion, encompassing distinct T-cell subsets and memory phenotypes. 1,2,5nterestingly, T cells residing in the plaque demonstrate significantly elevated CD69 expression compared with circulating T cells, with ≈60% of all T cells in the plaque expressing this molecule. 6CD69 is a transmembrane C-type lectin and is upregulated upon TCR (T-cell receptor) stimulation but can also be induced by other external stimuli such as exposure to type I IFN (interferon). 7D69 is involved in a variety of cellular processes, including proliferation and signal transduction, and controls tissue retention by regulating S1PR1 (sphingosine-1phosphate receptor 1) expression. 8Based on its function, CD69 can either be used as a marker for early TCR activation or for tissue residency in T cells.
Tissue resident memory T (T RM ) cells reside at the site of previous antigen encounter without access to the circulation and are primed to provide rapid and localized immune responses upon interacting with familiar antigens.Although T RM cells are efficient in countering secondary infections, they have also been implicated in the involvement of chronic inflammatory diseases.T RM cells are often enriched at sites of chronic inflammation. 9,10oreover, multiple studies have demonstrated the involvement of T RM cells in reoccurring disease flares. 10,11or example, in rheumatoid arthritis, disease reoccurrence was ameliorated by depleting T RM cells. 11et, the involvement of T RM cells in atherosclerosis remains elusive.T cells exhibiting a T RM cell-like phenotype have been identified in human and mouse atherosclerotic lesions, and CD69 + CD103 + CD8 + T cells with a T RM cell-like phenotype were specifically expressed in human atherosclerotic lesions but not in blood. 2,12Moreover, the majority of T cells in the atherosclerotic lesion has an effector memory phenotype, combined with high expression of CD69 (CCR7 low CD45RA high CD69 + ). 1,2,13n the contrary, Depuydt et al 6 demonstrated that these CD69 + effector memory T cells show signs of clonal expansion and recent TCR activation, arguing that T cells in the lesion can also express CD69 as a result of antigen recognition.
In this study, we aimed to assess the phenotype, proportion, and function of T RM cells in the atherosclerotic lesion.To interrogate the phenotype of T cells in the lesion, we used single-cell RNA-sequencing (scRNAseq) data of human atherosclerotic lesions and performed flow cytometry to evaluate the expression of T RM cell-associated molecules on lymphocytes isolated from mouse atherosclerotic lesions.To evaluate the proportion and function of T RM cells in atherosclerotic lesions, a bone marrow transplantation study was performed, in which the formation of T RM cells was hindered.

MATERIALS AND METHODS
The data that support the findings of this study are available from the corresponding author upon reasonable request.

Human Studies
Handling of all human samples complied with the Code for Proper Secondary Use of Human Tissue and is in accordance with the Declaration of Helsinki regarding ethical principles for medical research involving human subjects, and all patients signed an informed consent form.To evaluate the proportion of T RM cell marker expressing T cells in the atherosclerotic lesion, atherosclerotic plaque samples were obtained from 20 patients who underwent carotid endarterectomy surgery at the Haaglanden Medical Center, Westeinde, The Hague, the Netherlands (study approval number: 17-046; protocol number: NL57482.098.17).Only atherosclerotic plaques from primary carotid endarterectomy surgeries were included in this study.

Human Atherosclerotic Plaque Cell Isolation
Human endarterectomy samples were digested into a singlecell suspension following a previously described protocol. 6In brief, lesions were digested into a single-cell suspension by cutting the tissue into pieces of ≈1 mm 2 , followed by digestion with 2.5 mg/mL collagenase IV (Thermo Fisher Scientific), 0.25 mg/mL DNAse I (Sigma), 2.5 mg/mL Human Albumin Fraction V (MP Biomedicals) in RPMI 1640 for 30 minutes at 37 °C.After digestion, plaque tissue was mashed over a 70-µm strainer to create a single-cell suspension and washed in RPMI

scRNA-seq on Human Plaques
A previously published scRNA-seq data set was used to interrogate the phenotype of T RM cells in the atherosclerotic lesion. 6n brief, scRNA-seq was performed on PBMCs and atherosclerotic lesions of 3 male patients who underwent carotid endarterectomy.PBMCs and single-cell plaque suspensions were stained with TotalSeq-C antibodies against CD3 and CD14.Plaque suspensions were additionally stained for CD45.scRNA-seq was performed on PBMCs and live CD45 + plaque cells, using 10× Genomics 5ʹ Single Cell Immune Profiling technology.Next, CD3 + T cells were selected from all immune cells and subclustered.

Integration of T RM Cell Data Set
To compare the phenotype of T cells residing in the plaque with T cells previously identified as T RM cells, T RM cells were selected from a published scRNA-seq data set containing gene expression data of T cells from human intestinal transplants, and integrated with the plaque T cells, following the Seurat introduction to scRNA-seq integration vignette. 14In brief, raw data (Gene Expression Omnibus: GSE162687) were processed using Seurat omitting cells with gene expression <200 and >2500 or >10% mitochondrial genes from the data set.
To decrease the dimensionality of the normalized data and execute cell clustering, a principal component analysis (PCA) was conducted.The initial 10 PCA components were utilized for identifying clusters at a resolution of 0.4, following authors' description.
After integration of the predefined T RM cells into the atherosclerosis data set, the curated data set was clustered based on the top 15 PCA components and a resolution of 0.7.Next, classical T RM cells, with high expression of CD69, ITGAE (integrin alpha E), and ZNF683 (zinc finger protein 683), were selected from the integrated T RM cell data set for projection onto and integration into the atherosclerosis scRNA-seq data set.Other T cells from the integrated data set were removed from the curated data set.Clusters were visualized using UMAP (Uniform Manifold Approximation and Projection) and annotated based on differential gene expression between clusters, identified by a Wilcoxon rank-sum test (FindMarkers function Seurat).

Animals
C57BL/6J, RAG1 −/− (recombination activating gene 1 deficient), and LDLr −/− (low-density lipoprotein receptor knockout) mice were purchased from The Jackson Laboratory (Sacramento, CA) and bred in-house.Animals were kept under standard laboratory conditions, and food and water were provided ad libitum.All animal work was performed in female mice, to prevent sexdriven variation in atherosclerosis development.All

Bone Marrow Transplant
Forty-five female RAG1 −/− LDLr −/− mice of 8 to 16 weeks of age were exposed to 9 Gy (2×4.5 Gy, 0.19 Gy/min, 200 kV, 4 mA) total body irradiation using an Andrex Smart 225 Röntgen source (YXLON International) with a 6-mm aluminum filter to induce bone marrow aplasia.Mice received 10×10 6 donor bone marrow cells originating from either wild-type mice (C57BL/6 mice), Hobit tdTomato mice (Hobit +/tdTomato mice; bone marrow obtained from Sanquin laboratories), or Hobit KO/CRE Blimp-1 flox/flox mice (Hobit CRE/tdTomato ×Blimp flox/flox ; bone marrow obtained from Sanquin laboratories) in 100-µL saline via tail vein injection.Bone marrow was isolated by flushing the femoralis and tibia with PBS and passing the cells through a 70-µm cell strainer (Greiner Bio-One).During the recovery period of 4 weeks, the drinking water of the mice was supplemented with antibiotics (83 mg/L ciprofloxacin, 67 mg/L polymyxin B sulfate, and 6.5 g/L sucrose).After recovery, the mice were fed a Western-type diet (0.25% cholesterol and 15% cacao butter; Special Diet Services) for 10 weeks to induce lesion formation.Upon euthanization, the mice were anesthetized intraperitoneally with a mix of ketamine (100 mg/mL) and xylazine (10 mg/mL).Blood was collected via orbital bleeding, followed by perfusion with PBS through the left ventricle.The heart, aorta, spleen, and liver were collected for further analysis.

Organ Processing
Single-cell suspensions of the spleen and liver were obtained by using a 70-µm cell strainer (Greiner Bio-One).Red blood cells were removed from the liver, spleen, and blood samples by lysing the red blood cells for 2 minutes at room temperature with lysis buffer.To separate leukocytes from hepatocytes in the liver samples, gradient centrifugation with 35% Percoll (Merck) was applied.Aortas were cut into small fragments of ≈1 mm 2 , and incubated with a digestion mix of collagenase I, 450 U/mL; collagenase XI, 250 U/mL; DNAse, 120 U/mL; and hyaluronidase, 120 U/mL (all Sigma-Aldrich) for 30 minutes at 37 °C while rotating.Digested aorta samples were passed over a 70-µm cell strainer to obtain a single-cell suspension.The upper part of the heart was embedded in Tissue-Tek OCT (Sakura) and stored at −80 °C until further usage.

Cholesterol Assay
Serum was obtained by collecting whole blood in cloth inducing microvette CB300Z tubes (Sarstedt), followed by centrifugation at 2000 RCF (relative centrifugal force) for 10 minutes at 4 °C.Serum was stored at −80 °C until further usage.Total cholesterol levels in serum were measured by performing an enzymatic colorimetric analysis (Roch/Hitachi, Germany), using Precipath standardized serum (1.69 mg/mL; Roche/Hitachi) as an internal standard.

Histological Analysis
The trivalve area of the hearts was sectioned at 7-µm thickness using a Leica CM1950 cryostat.The size of atherosclerotic plaques and the percentage of stenosis were determined using oil red O staining (Sigma-Aldrich).To assess collagen content and necrotic core size in the atherosclerotic lesions, Masson trichrome staining (Sigma-Aldrich) was performed.To determine macrophage content in the lesions, sections were immunohistochemically stained with a primary antibody targeting a macrophage-specific antigen (monoclonal rat IgG2b macrophages/monocytes antibody, clone MOMA-2, diluted 1:1000), followed by secondary antibody staining with biotinylated rabbit anti-rat IgG (BA-4001, Vector, diluted 1:200).The reaction was visualized using the ImPACT NovaRED Peroxidase substrate (Vector).All sections were digitized using a Panoramic 250 Flash III slide scanner (3DHISTECH; Hungary).Data analysis was performed using the ImageJ software.

Statistical Analysis
Data analysis was performed using Prism 9.0 (GraphPad Software, Inc, San Diego, CA).Data are expressed as mean±SD for all analyses.Outliers were identified by an ROUT (Robust Regression and Outlier Removal Test) or Grubbs outlier test.Shapiro-Wilk normality test was used to test data for normal distribution.Normally distributed data were analyzed using a 2-tailed unpaired Student t test when comparing 2 groups or a 2-way ANOVA with Bonferroni multiple comparisons test when comparing 2 groups with multiple conditions.Non-normally distributed data were analyzed using a Mann-Whitney U test.Statistical analysis was performed using GraphPad Prism.Probability values of P≤0.05 were considered to be significant.
To evaluate differential gene expression between T RM -like T cells in the atherosclerotic lesion and other lesion-derived T cells, the FindMarkers function was used.Differential gene expression was computed using a nonparametric Wilcoxon rank-sum test, and Bonferroni-corrected P values were determined considering the overall gene count in the data set.Visualization of the differential gene expression was performed using a Volcano plot, facilitated by the EnhancedVolcano package.

Quantification of the Proportion of T RM Cells in Human Atherosclerotic Lesions
T RM cells have been identified at the site of former inflammation in a large variety of tissues.Here, we interrogated T cells from human atherosclerotic lesions on a single-cell level, to thoroughly evaluate the presence of T RM cells in atherosclerosis.T RM cells are notoriously heterogenic between tissues, making it challenging to identify T RM cells in unexplored tissues by flow cytometry.Therefore, we used an in-house scRNA-seq data set encompassing information on the transcriptome of 3 human atherosclerotic plaque samples, obtained via endarterectomy of the carotid artery, along with patient-matched blood samples. 6Subclustering of T cells residing in the atherosclerosis lesion resulted in 11 clusters, including 2 mixed T-cell clusters (clusters 0 and 7), 1 regulatory cluster (cluster 5), 2 memory T-cell clusters (clusters 1 and 6), 2 exhausted T-cell clusters (clusters 9 and 10), 1 γδ T-cell cluster (cluster 8), 2 cytotoxic lymphocyte clusters (clusters 3 and 4), and 1 GZMK + T RM cell-like cluster (cluster 2; Figure 1A).
To identify T RM cells within the atherosclerotic lesions, we analyzed the expression of the canonical T RM cell marker CD69, as well as markers strongly associated with a T RM cell phenotype, such as CD49Α and ITGAE (encoding CD103), and transcription factors essential for the formation of T RM cells, including ZNF683 (encoding HOBIT), PRDM1 (encoding BLIMP-1), and RUNX3.T RM cell-associated markers were mainly expressed by exhausted T-cell clusters 9 and 10 and by cytotoxic GZMK + cells from cluster 2 (Figure 1B through 1D).Projection of a T RM gene signature, including core T RM genes identified by Kumar et al and T RM cell-associated transcription factors, onto the plaque-derived T cells identified clusters 0, 1, 2, 8, and 9 as clusters that potentially contain T RM cells (Figure 1E).This T RM gene signature was mainly expressed by T cells from the atherosclerotic lesion, and to a minor extent by T cells from circulation, confirming its specificity for T cell residing in tissue (Figure S1).Moreover, coexpression of multiple T RM cell markers, including CD69, CD49α, and CD103, was also confirmed on a protein level in human endarterectomy samples by flow cytometry (Figure 1F).
To more comprehensively evaluate how the phenotype of plaque-derived T cells relates to that of T RM cells, T cells with a canonical T RM cell phenotype (expressing CD69, ITGAE, and ZNF683) were selected from a previously published T RM cell data set and integrated (Figure S2). 14 This T RM cell data set consisted of scRNA-seq data of donor-derived T RM cells extracted from intestinal transplant tissue 1 year after transplant using HLA (human leukocyte antigen) allele congenic cell tracking, ensuring the selection of long-lived T RM cells.Reclustering of the curated data set resulted in 10 distinct T-cell clusters, with the predefined T RM cells predominately coclustering with cells from cluster 3, demonstrating that a small proportion of T cells residing in the plaque exhibits a T RM -like phenotype (≈4.6%; Figure 2A through 2C).The plaquederived T cells from cluster 3 mainly originated from cluster 2 in the original clustering (Figure 2E), a cluster that was already associated with T RM marker expression.To identify the expression profile of plaque-derived T cells from cluster 3, we excluded integrated T RM cells from further analysis.T cells from cluster 3 were characterized by relatively high expression of T RM -associated genes CD69, ITGAE, ITGA1, ZNF683, and RUNX3 (Figure 2D).Moreover, genes associated with T-cell activation and cytotoxicity, including KLRK1 (encoding KNG2D [natural killer group 2 member D]), CCL5, MATK, KLRC4, and CTSW (encoding cathepsin C), were among the top differentially expressed genes (Figure 2E), underlining the high cytolytic potential of the T RM -like T cells in the lesion.

Phenotyping T RM Cells in the Atherosclerotic Environment
As the scRNA-seq analysis suggests the presence of a minor subset of T RM in atherosclerotic lesions of patients, we next aimed to address their presence and role in the pathogenesis of atherosclerosis in mice.To corroborate the presence of T RM in atherosclerotic lesions, a bone marrow transplant study was performed in RAG1 −/− LDLr −/− mice, receiving bone marrow of Hobit reporter mice (Hobit tdTomato ).In Hobit tdTomato mice, the Znf683 promotor of one allele was inserted with the tdTomato gene, encoding a red fluorescent protein. 19fter reconstitution of the bone marrow, mice were fed a Western-type diet for 10 weeks to induce atherosclerosis.
To address whether the Hobit reporter allowed detection of T RM after bone marrow transplantation, we first examined T RM in the liver, which naturally contains these cells.We observed that the percentage of tdTomato + T cells in the liver varied from 2.5% to 8% for CD4 + T cells and between 7% and 26% for CD8 + T cells, which is consistent with the previously reported numbers of T RM in livers of naive mice (Figure 3A; Figure S3A and  S3B). 20TdTomato + CD4 + and CD8 + T cells from the liver almost uniformly expressed CD69, whereas tdTomato − T cells harbored a limited number of CD69-positive cells.Moreover, tdTomato + CD4 + and CD8 + T cells expressed significantly elevated levels of CD49α, compared with their tdTomato − counterparts.The increased expression of CD69 and CD49α suggests that tdTomato + T cells indeed positively identify as T RM cells.
Next, we analyzed the number of T RM cells in the aortic arch, a primary site of atherosclerotic lesion development.The proportion of tdTomato expressing T cells in the atherosclerotic lesion ranged from 0% to 6% for CD4 + T cells and between 1% to 7% for CD8 + T cells (Figure 3B; Figure S3C and S3D).The proportion of CD69 + cells was significantly increased in the tdTomato + T cells compared with the tdTomato − T cells.However, CD69 was not exclusively expressed by tdTomato + T cells in the aorta, in line with previous reports suggesting that CD69 is upregulated in lesion-derived T cells as a response to antigen-specific interactions.Hence, CD69 cannot be considered as a standalone marker for identifying T RM cells in the lesion.
Additional markers that were evaluated as potential T RM markers in the lesion were CD49α and CD103.TdTomato + T cells in the lesion trended toward an upregulated expression of CD49α.Moreover, almost all CD49α + cells also coexpressed CD69 (Figure S3E and S3F).Integrin CD103, on the other hand, is a canonical T RM marker for T RM cells in epithelial and mucosal tissues and is differentially expressed by T RM cells in the lesion.Hobit + CD4 + T cells expressed elevated CD103 levels compared with their Hobit − counterparts.Although CD103 was highly expressed by CD8 + T cells in the atherosclerotic lesion, it appeared to be mainly expressed by tdTomato − CD8 + T cells as opposed to tdTomato + CD8 + T cells, demonstrating CD103 alone is not a suitable marker for identifying CD8 + T RM cells in atherosclerosis.Coexpression of CD103 together with CD69, however, was almost exclusive to tdTomato + T cells, compared with tdTomato − cells (Figure S3E and S3F).Taken it all together, tracking of T RM cells revealed CD69 and CD49α as markers characteristic for T RM in the atherosclerotic lesion.

T RM Cells Constitute Only a Minor Population in Atherosclerotic Lesions
Next, we set out to discover the role of T RM cells in the pathogenesis of atherosclerosis.Blimp-1 and Hobit are central regulators of the conversion of T cells to a tissue resident phenotype in both classical and innate lymphocytes. 20To investigate the function of T RM in atherosclerosis, we performed a bone marrow transplant experiment with bone marrow originating from Hobit KO/ CRE Blimp-1 flox/flox mice (Hobit CRE/tdTomato ×Blimp-1 flox/flox mice) or wild-type mice (C57BL/6 mice) transferred to RAG1 −/− LDLr −/− mice (Figure S4A).By knocking out Hobit and conditionally disrupting Blimp-1 in cells with an active Znf683 promotor, the formation of T RM cells can be severely reduced. 18,20Hobit CRE/tdTomato ×Blimp-1 flox/flox mice exhibit a significantly reduced number of T RM . 18Moreover, the tdTomato + T RM progenitor cells that remained expressed reduced levels of retention molecule CD69, as well as increased levels of tissue egress markers S1PR1 and CCR7 (C-C chemokine receptor type 7), demonstrating that these cells are not capable of obtaining a mature T RM phenotype.After recovery of the bone marrow transplant, mice were fed a Western-type diet for 10 weeks to induce atherosclerosis formation.Throughout the experiment, mouse weights and cholesterol levels were similar between groups (Figure S4A through S4C).At the end of the experiment, the hampered formation of T RM cells was confirmed in the liver by evaluating CD69 + T-cell counts (Figure S5).
The blocked T RM formation resulted in some variations in circulating immune cell levels, while leukocyte populations in secondary lymphoid organs remained unaffected (Figure S4D and S4E).Given that Hobit and Blimp-1 are important for the differentiation of innate lymphoid cells as well, the distribution of natural killer (NK) cells and natural killer T (NKT) cells in the liver was evaluated. 20,21The fraction of NK and NKT cells, which are low due to the bone marrow transplantation, were significantly reduced in the livers of Hobit KO/CRE Blimp-1 flox/flox mice (Figure S4F).Also, the cytolytic capacity of NK cells was affected by the Hobit/Blimp-1 knockout, apparent from a significantly reduced GZMB + NK-cell population.
In the lesion, however, general leukocyte levels and T-cell numbers were not affected by the knockdown of T RM cells (Figure 4A through 4E and 4G).Interestingly, we also did not observe changes in the CD69 + or CD49α + T-cell populations and only a modest decrease in the percentage of CD69 + CD49α + double-positive CD8 + T cells (Figure 4F and 4H; Figure S6).The marginal changes in T RM marker expressing T-cell populations underlines that T RM cells only encompass a minor population in the atherosclerotic lesion.

T RM Cells Associate With Reduced Macrophage Content and Increased Collagen Deposition in Atherosclerotic Lesions
To verify the effects of the Hobit/Blimp-1 conditional double knockout on the progression of atherosclerosis, plaque size and composition were assessed in the aortic root lesions of the heart by histological analysis.Neutral lipid staining of the lesions by oil red O demonstrated similar absolute plaque sizes, with a plaque size of ≈400 000 µm 2 and comparable vascular occlusion of ≈40% between groups (Figure 5A).Although plaque size was not altered by the lack of T RM cells, plaque composition was affected in the Hobit KO/CRE Blimp-1 flox/flox group.The macrophage-positive stained area as a relative to plaque size was significantly increased in Hobit KO/CRE Blimp-1 flox/flox mice, with a 1.33-fold increase compared with the wild-type control (Figure 5B).Interestingly, the increased macrophage content did not translate to increased necrotic core sizes, evaluated by Masson trichrome staining (Figure 5C).Collagen content, on the other hand, trended toward a reduction in Hobit KO/CRE Blimp-1 flox/flox mice.

DISCUSSION
T RM cells are a relatively newly discovered T-cell subset that forms the first line of defense against secondary infections.Despite the growing body of knowledge, the precise role of this specialized T-cell subset in chronic inflammatory diseases remains largely elusive. 22,23This study contributes valuable insight on the presence, phenotype, and functionality of T RM in both human and mouse atherosclerotic lesions.We demonstrated that T RM cells constitute a minor fraction of the lymphocyte population in the atherosclerotic lesion and that mice in which the formation of T RM was hindered exhibit increased lesion stability.
In human atherosclerotic lesions, we identified a minor population of T RM -like cells that expressed elevated levels of T RM -associated genes and coclustered with predefined T RM cells for intestinal epithelium.Although the phenotype of these predefined T RM cells might be affected by other microenvironmental circumstances (different tissue, disease, and medical background of the patient), they specifically clustered together with the plaque-derived T cells that expressed a T RM -like phenotype in the unintegrated analysis, suggesting a common core T RM signature for T RM cells derived from different tissues. 15The TRM-like cells in the atherosclerotic lesion were characterized by the expression of CD69, ITGA1, and ITGAE, on both transcriptional and translational levels, and a transcription factor profile associated with the T RM phenotype.
The upregulated expression of CD69 and CD49α was corroborated in mice.Although human T RM cells might be characterized by the expression of CD103, as shown here and by others 2 ; this marker was inconsistently expressed by Hobit-expressing CD4 + and CD8 + T cells in mouse atherosclerotic lesions.Various studies demonstrated species-specific differences in the T RM phenotype within the same tissue. 24Moreover, T RM cells located in nonepithelial/mucosal tissues often refrain from expressing CD103, as its ligand, E-cadherin, is not expressed there. 25 Controversially, a large proportion of Hobit − CD8 + T cells did express CD103 in the atherosclerotic lesion.CD103 expression is enhanced by exposure to TGF-β (transforming growth factor beta), a cytokine expressed in atherosclerotic lesions and associated with protection against plaque rupture. 26,27ike the expression of CD103, the expression of CD69 was also not confined to T RM cells.A substantial population of ≈20% of the CD4 + and CD8 + T cells expressed CD69 in mouse atherosclerotic lesions.However, on average, only 2.4% of the T cells expressed Hobit, suggesting that the majority of CD69-positive cells did not identify as T RM .Lesional T cells have been reported to express an activated phenotype and display signs of clonal expansion and tissue enrichment, implying recent activation through TCR-specific interactions. 6,28Hence, it seems that merely a fraction of CD69-expressing cells in the lesion possess a T RM phenotype, while for the remaining cells potentially upregulated CD69 as a response to antigen recognition.
One limitation of our approach is that by identifying T RM cells based on Hobit expression, we may have underestimated the T RM population.Although Hobit is exclusively expressed by T RM , and not by other antigenexperienced T-cell populations, this marker might be downregulated over time by some T RM cells, rendering identification in this study setup impossible. 16,18Nonetheless, the knockdown of T RM cells did not impact the proportion of T cells in the lesion nor the percentage of CD69 expressing T cells, indicating that T RM population is indeed a small fraction of the T-cell population.
In spite of their modest representation, we did observe a significantly reduced plaque stability and increased macrophage content when T RM cells were significantly reduced.However, these effects might not be fully attributable to the lack of T RM cells, as we also observed changes in circulating leukocyte levels and a significant reduction in innate lymphocyte populations.Although NK and NKT cells have been detected inside atherosclerotic lesions, their function in atherosclerosis remains controversial.A recent study on the function of NK cells in atherosclerosis concluded that depletion of NK cells does not affect the progression of atherosclerosis. 29In addition, most reports focusing on the function of NKT cells in atherosclerosis describe an aggravating effect of NKT cells in atherosclerosis, which does not explain the effects observed in this study. 30oreover, similar protective effects were observed when all CD8 + T cells, including all cytotoxic T cells, were depleted from advanced atherosclerotic lesions 31 or when CD4 + T-cell numbers were strongly reduced and their interaction with MHCII (major histocompatibility complex class II) was negated. 32In line with previous reports focusing on the phenotype of T RM cells, scRNA-seq of human atherosclerotic lesions revealed that T RM -like cells in the lesion expressed high levels of granzymes and other cytotoxicity-associated molecules. 33These cytolytic T cells might control intralesional macrophage content by specifically targeting macrophages.Moreover, antigen-specific CD8+ T cells have also been shown to exert atheroprotective effects. 34,35he observed changes in collagen content might be a result of the relatively increased macrophage population, as macrophages express a variety of matrix degrading enzymes, like metalloproteinases, that contribute to the breakdown of the collagen in the fibrous cap. 36,37nterestingly, in most immune disorders, like rheumatoid arthritis, inflammatory bowel disease, and psoriasis, T RM cells are associated with disease acceleration.Yet, unlike atherosclerosis, these diseases are characterized by a relapsing-remitting course, and T RM cells have been implicated to contribute to these reoccurring disease flares. 10,11,38Atherosclerosis, on the other hand, is an ongoing chronic inflammatory disease characterized by constant immune cell infiltration, persistent chemokine and cytokine production, and continuous antigen presentation. 6The chronic nature of the disease might explain the limited formation of T RM cells in the lesion.Although some reports argue persistent low-gradient antigen presentation induces memory inflation and increases T RM populations, 39 others showed compromised T RM formation after chronic viral infection, compared with acute infection. 40n conclusion, this study provides valuable insights into the presence, phenotype, and function of tissue-T RM cells in human and mouse atherosclerotic lesions.We demonstrate that intralesional T RM cells are characterized by the expression of CD69 and CD49α and that although these T RM cells only constitute a minor T-cell population within the lesion, their presence was associated with reduced atherosclerotic burden by decreasing the amount of intralesional macrophages and enhancing lesion stability.

Figure 1 .
Figure 1.Resident memory T (T RM ) cell-associated marker expression in human atherosclerotic lesions.A, UMAP representation of single-cell RNA-sequencing data of T cells from human atherosclerotic lesions (n=4548).B, Dot plot visualizing average T RM -associated marker gene expression per cluster.C, Feature plot representation of T RM marker expression.D, Heatmap displaying average expression of T-cell function-associated genes per cluster.E, Projection of T RM gene signature onto UMAP representation of plaque-derived T cells.F, Flow cytometry analysis of CD69, CD49α, and CD103 expression by plaque-derived CD4 + and CD8 + T cells, demonstrating the percentage of CD69 + CD49α + CD103 + coexpressing cells in a bar graph and tSNE representation.ITGA1 indicates integrin alpha 1; ITGAE, integrin alpha E; PRDM1, PR domain zinc finger protein 1; Prolif., proliferation; RUNX3, runt-related transcription factor 3; TF, transcription factor; Treg, T regulatory; tSNE, t-distributed stochastic neighbor embedding; UMAP, Uniform Manifold Approximation and Projection; and ZNF683, zinc finger protein 683.

Figure 2 .
Figure 2. Predefined resident memory T (T RM ) cells cluster together with plaque-derived T cells.A, UMAP representation after integration of (top) plaque-derived T cells from the atherosclerosis single-cell RNA-sequencing (scRNA-seq) data set, and (bottom) predefined T RM cells.B, Projection of predefined T RM cells on unsupervised T-cell clustering of the curated data set.C, Percentage of integrated T RM cells per cluster, with the number of integrated cells per cluster on top of each bar and the number of plaquederived T cells in the middle of each bar.D, Heatmap displaying average expression of T-cell function-associated genes per cluster.E, UMAP representation of scRNA-seq data of T cells from human atherosclerotic lesions (clustering before integration), highlighting plaque-derived T cells from cluster 3 (clustering after integration).Highlighted cells are indicated with a read halo.F, Volcano plot displaying significantly differentially expressed genes of plaque-derived T cells from cluster 3, compared with T cells from all other clusters.Significantly differentially expressed genes were identified by the Wilcoxon rank-sum test, reaching statistical significance from P<10 −6 .FC indicates fold change; Prolif., proliferation; TF, transcription factor; Treg, T regulatory; and UMAP, Uniform Manifold Approximation and Projection.

Figure 3 .
Figure 3. Lesion-derived resident memory T (T RM ) cells are characterized by elevated expression of CD69 and CD49α.The percentage of Hobit (homolog of Blimp-1 [B lymphocyte-induced maturation protein 1] in T cells) expressing tdTomato + CD4 + and CD8 + T cells and the expression of CD69, CD49α, and CD103 by tdTomato + (pink) and tdTomato − (blue) in the liver (A) and in the aorta (B).

Figure 5 .
Figure 5. Hindered resident memory T (T RM ) cell formation results in a relative increase in intralesional macrophages and reduction in collagen content.A, Quantification of lesion size and vascular occlusion in the trivalve area determined by oil red O staining and representative pictures of the wild-type and Hobit (homolog of Blimp-1 [B lymphocyte-induced maturation protein 1] in T cells) KO/CRE Blimp-1 flox/flox groups.B, Quantification of absolute and relative monocyte/macrophage content in the plaque by MOMA (monocyte/macrophage) staining and representative pictures of the lesions.C, Quantification of relative necrotic core area and relative collagen content in the lesions after Masson trichrome staining and representative pictures of the lesions.Plots contain individual data points with mean±SD.Significance was determined by unpaired t test.Outliers were removed after an ROUT (Robust Regression and Outlier Removal Test).
Nonstandard Abbreviations and Acronyms