Phenotype of FH pigs. We used 14-month-old FH noncastrated males (n = 5/group) and gilt (never been used for breeding) females (n = 9/group) and administered recombinant human IGF-1 (rhIGF-1), 50 μg/kg/d, twice per day, or saline for 6 months (Figure 1A). Of note, the IGF-1 dose is within the range of the FDA-approved dosage for long-term treatment of growth failure in children with primary IGF-1 deficiency (28), and the amino acid sequence structure of porcine IGF-1 is identical to human IGF-1 (29). Safety, pharmacokinetics, and efficacy of IGF-1 are reported for patients (28, 30, 31). To verify that IGF-1 administration into FH pigs stimulates specific downstream signaling in porcine vasculature and blood cells, we injected recombinant human rhIGF-1 (or saline, control) into pigs and isolated carotid arteries and peripheral blood mononuclear cells (PBMCs) after 4 hours. Akt phosphorylation was increased by almost 4-fold (P < 0.001) in both vascular tissue and PBMC sin IGF-1–injected pigs versus control (Figure 1B), indicating that IGF-1 promotes specific downstream signaling. IGF-1 levels at T0 were higher in males versus females (Figure 1, C and D, P < 0.005). IGF-1–injected pigs had significantly higher IGF-1 levels compared with control pigs at all tested time points. The average increase in IGF-1 group versus saline in males was 88.0% ± 19.4% and 83.3% ± 9.7% in females.

Phenotype of FH pigs. (A) Experimental design. FH pigs were injected daily with 50 μg/kg human recombinant IGF-1 or saline (control) (males, n = 5/group; females, n = 9/group) and fed with high-fat diet (HFD) for 6 months. Coronary atherosclerosis was quantified by intravascular ultrasound (IVUS) before injections (T0), after 3 months (T3), and after 6 months (T6, at sacrificing). (B) IGF-1 stimulated specific downstream signaling in porcine carotid artery and in peripheral blood mononuclear cells (PBMC). IGF-1 (or saline) was injected into pig, carotids and blood were collected 4 hours following injections, and Akt phosphorylation was quantified by immunoblotting. (C–H) Blood was collected at basal level (T0) and each month during injections (total 7 time points). Total plasma IGF-1 level was quantified by ELISA in FH males (C) and females (D). Cholesterol (E and F) and triglyceride (G and H) levels in IGF-1– and saline-injected male (left) and female (right) FH pigs. Males: n = 10 for basal, and n = 5/group for each time point. Females: n = 18 for basal, and n = 9/group for each time point. *P < 0.05, **P < 0.01, vs. saline based on t test, #P < 0.05 vs. basal level based on 3-way ANOVA.

Male and female animals had a similar initial BW on average, and both IGF-1– and saline-injected pigs gained BW steadily throughout the study, with no difference found between saline and IGF-1 groups (Supplemental Figure 1; supplemental material available online with this article; https://doi.org/10.1172/jci.insight.165713DS1). FH males had a larger BW increase compared with FH females (M, 69.2% ± 2.9% increase at T6 versus T0; F, 45.0% ± 5.2% increase, P < 0.01).

Blood pressure and heart rate were measured in sedated pigs at each IVUS procedure. We found no difference in systolic/diastolic blood pressure and heart rate between sexes and between saline and IGF-1 groups (Supplemental Table 1). There was no time-dependent change in blood pressure or in heart rate.

Blood tests were performed monthly. We found no statistically significant difference in complete metabolic profiles and complete blood count (CBC) with differential between sexes and saline versus IGF-1 group (Supplemental Tables 2 and 3). However, we found that females had a striking 2.3-fold higher cholesterol levels at T0 compared with males (P < 0.001, Figure 1, E and F). HFD feeding caused a significant and sustained elevation of total cholesterol levels in both sexes. IGF-1 did not change total cholesterol in both males and females. FH females had significantly higher triglyceride levels compared with males at T0 (P < 0.005, Figure 1, G and H), and HFD feeding did not change triglycerides. IGF-1 reduced triglycerides in FH pigs (3-way ANOVA, P < 0.005, Figure 1, G and H).

Necropsy/histopathology findings. IGF-1 levels have been reported to be associated with an increased risk of cancer (26); thus, we performed autopsies and collected all major organs for histopathological analysis. No gross abnormalities were found either in saline-injected or IGF-1–injected FH pigs, and no tissue was considered carcinogenic by a certified pathologist. All female pigs had multiple coalescing yellow to tan plaques on the intimal surface of the aorta and large visible lipid deposits in the right coronary artery (RCA), left anterior descending artery (LAD), and circumflex artery. Two female pigs (saline, 1; IGF-1, 1) had evidence of a myocardial infarction. Abdominal fat deposits, hepatic lipidosis/fatty liver, and fatty lymph nodes were found in several saline- and IGF-1–injected females. Two female pigs in the saline group had pleural adhesions and moderate splenic enlargement, suggesting inflammation. Sections of the lung, liver, kidney, spleen, lymph node, stomach, small and large intestines, urinary bladder, ovary, uterus, pancreas, salivary gland, skeletal muscle, thyroid, and aorta from each animal were examined. No significant microscopic lesion including inflammatory or neoplastic process was noted in all animals (not shown) except atherosclerosis in the aorta and in coronary arteries.

IGF-1 decreases coronary atherosclerosis. IVUS was performed in the RCA and LAD at T0, T3, and T6 time points (Figure 1A). We found no significant difference in vessel volume, lumen volume, and plaque + media volume between RCA and LAD and between sexes at T0 (Table 1). There was a time-dependent increase of the vessel volume, presumably due to normal animal growth and to vascular remodeling concomitant with intimal thickening. In fact, the lumen volume time dependently decreased in both saline- and IGF-1–injected pigs (P < 0.001 for both RCA and LAD), consistent with intimal thickening. IGF-1–injected pigs had a larger time-dependent increase in RCA and LAD artery volume compared with controls (P < 0.05), suggesting vascular hypertrophy. Coronary arteries in IGF-1–injected FH females had larger lumen volume at T3 and T6 (P < 0.005) compared with control.

Pigs had approximately 16% of relative atheroma volume at T0. IGF-1– and saline-injected pigs had a time-dependent increase in absolute plaque + media volume and in relative atheroma volume (P < 0.001 for RCA and LAD). FH females had a significantly larger time-dependent increase in RCA and LAD relative atheroma at T3 (females, 36.6% ± 2.6%, vs. males, 23.3% ± 0.2%) and at T6 (females, 47% ± 4% vs. males, 35.9% ± 1.9%), indicating the presence of a strong sex effect. IGF-1 did not significantly change the absolute plaque + media volume in RCA and LAD. We found no interaction between IGF-1 effect on relative atheroma and sex (3-way ANOVA). As sex does not influence IGF-1’s potential effect on relative atheroma volume, we combined measurements of both sexes and performed a 2-way analysis (treatment vs. time) using Bonferroni’s correction for repeated measurements. IGF-1 time dependently decreased relative atheroma volume in RCA (P < 0.05 vs. saline) and in LAD (P = 0.054, Figure 2). The IGF-1–induced increase in vessel lumen and reduction in relative atheroma indicate that IGF-1 reduces coronary atherosclerosis.

IGF-1 reduces coronary atheroma volume. FH pigs (both sexes) were injected with IGF-1 or saline (control) (males, 5/group; females, 9/group). The coronary atheroma volume was quantified in RCAs and LADs by serial IVUS before injections (T0), after 3 months (T3), and after 6 months (T6). Relative atheroma volume (%) was defined as plaque + media volume divided per the vessel (external elastic membrane, EEM) volume × 100%. Since sex does not influence IGF-1’s effect on relative atheroma volume, atheroma measurements of both sexes were combined and shown. n = 28/group for T0, and n = 14 per RCA or LAD per group for T3 and T6. *P < 0.05 vs. saline, and #P < 0.05 vs. T0 based on 2-way ANOVA.

Trichrome-stained RCA and LAD cross sections were used for histological analysis. FH females developed larger and more complex plaques in coronaries compared with males, in agreement with IVUS results (Figure 3). Plaques in males contained diffuse homogeneous collagen material and lipid droplets and had neither necrotic cores nor fibrous caps. We classified plaques in FH males as type III pre-atheroma in accordance with histological classification of human atherosclerotic lesions (32). Coronary plaques in FH females were significantly larger (RCA: 2-fold increase in CSA versus males, LAD: 1.5-fold increase, P < 0.05, data for saline group); they contained dense fibrous caps, large acellular/necrotic cores, and multiple cholesterol clefts. We observed the presence of strong calcification (Alizarin Red staining, data not shown) and neovascularization (IHC with CD31, endothelial cell marker, data not shown) in coronary plaques in FH females but not in males. Plaques in FH females were classified as advanced type V fibroatheroma (32).

IGF-1 reduces coronary atherosclerosis and promotes features of stable atherosclerotic plaque. IGF-1 increased vascular media (A and B), reduced atherosclerotic plaque cross-sectional area (CSA) (C), decreased necrotic core (D), and elevated thickness of fibrous cap (E). RCA and LAD were isolated from IGF-1– and saline-injected FH pigs and further cut onto 6 sequential fragments for embedding in paraffin. Trichrome-stained cross sections were obtained from each fragment and used for morphological analysis. n = 30 per RCA or LAD per group for males and n = 54 for females. (A) Representative images of RCA sections obtained from FH males and females. Tunica media (TM), atherosclerotic plaque (AP), fibrous cap (FC), and necrotic cores (NCs) were manually outlined to quantify TM and AP CSA, and results were normalized per EEM area. The thickness of FC was calculated as the mean length of 5 arbitrary lines distributed across the cap area. *P < 0.05, ***P < 0.005 vs. saline based on unpaired 2-tailed t test.

FH males had a thicker tunica media compared with females in both IGF-1– and saline-injected groups (P < 0.05 in each case, Figure 3B). IGF-1 significantly increased medial CSA (males, RCA: 13% increase vs. control; LAD: 12% increase; females, RCA: 34% increase; LAD: 26% increase), consistent with vascular hypertrophy (Figure 3B). IGF-1 did not change total vessel CSA (outlined by external elastic membrane boundary, data not shown). IGF-1 reduced relative atherosclerotic plaque area in males (RCA, 15.2% ± 4.5% decrease; LAD, 16.6% ± 7.1% decrease compared with control) and in females (RCA, 21.5% ± 2.7% decrease; LAD, 17.4% ± 2.1% decrease compared with control) (Figure 3C), consistent with IVUS data. Plaques in IGF-1–injected females were more cellular and contained reduced necrotic cores compared with controls (RCA: 50.1% ± 1.6% decrease; LAD: 47.8% ± 1.4% reduction, Figure 3D). IGF-1 significantly increased the thickness of fibrous caps in coronary plaques in female pigs (Figure 3E). Our results indicate that IGF-1 induces vascular hypertrophy, reduces coronary atherosclerotic burden, and promotes features of plaque stability.

IGF-1 reduces macrophage-like cells and upregulates endothelial-like cells in coronary plaque. We reported previously that IGF-1 increases plaque SMCs (33), downregulates macrophages (MFs), and elevates levels of circulating endothelial cell (EC) progenitors (19) in HFD-fed Apoe–/– mice. SMCs and MFs share cell markers in the atherosclerotic plaques (34), and plaque ECs undergo a change in phenotype toward a mesenchymal cell type (35). Such phenotype switching complicates marker-based cell identification. To validate the IHC protocol, serial RCA sections were stained with a set of cell marker antibodies and immunopositivity pattern was compared. We found that each of 4 SMC marker antibodies stained virtually identical cell populations in the plaque, and a similar conclusion was made for 3 tested MF and 3 EC marker antibodies (Supplemental Figure 2). These data show that IHC with antibody for a single cell marker identified plaque cells expressing multiple markers, increasing confidence in identifying specific plaque cells. We also verified that cells immunopositive for macrophage scavenger receptor A (MSR), an MF marker, were immunonegative for α–smooth muscle actin (α-SMA), an SMC marker, and vice versa (Supplemental Figure 2A), showing that these antibodies have no cross-reactivity.

We used α-SMA, MSR, and CD31 antibodies to quantify SMC-like, MF-like, and EC-like cells, respectively, by IHC. SMC-like cells were abundant in the vascular media, and a mixture of SMC- and MF-like cells was found in the plaque fibrous cap (Figure 4 and Supplemental Figure 2A). In addition, MF-like cells were present in the area surrounding the plaque necrotic core and colocalized with cholesterol clefts in lipid cores. IGF-1–injected pigs had a slight increase in plaque SMC-like cells (P = NS) (Supplemental Figure 3) and a dramatic 2-fold reduction in plaque MF-like cells in females (P < 0.05 for RCA and LAD) (Figure 4). IGF-1 increased EC marker–positive area and this effect reached significance for RCA and LAD in females and LAD in males (Figure 4). Notably, EC marker–positive area was markedly larger in males compared with females (Figure 4A). We noted discontinuous CD31+ staining in endothelium layers of both IGF-1– and saline-injected pigs, suggesting either CD31 downregulation, or focal loss of EC, suggesting reduced endothelial integrity. To further obtain a surrogate index of endothelium layer integrity, we normalized CD31+ area per lumen perimeter. IGF-1 significantly increased CD31+ area/lumen perimeter ratio in the RCA and LAD in the female group (pixels2/pixels, RCA: IGF-1, 5.04 ± 0.84 vs. saline, 2.76 ± 0.64, P < 0.05; LAD: IGF-1, 10.31 ± 1.53 vs. saline, 5.52 ± 0.51, P < 0.01), suggesting that IGF-1 reduced the number of CD31+ endothelium layer breaks.

IGF-1 suppresses MF-like cells and upregulates EC-like cells in coronary plaques. Serial RCA and LAD sections were immunostained with α–smooth muscle actin (α-SMA), macrophage scavenger receptor A (MSR), and CD31 antibody to identify SMC-like, MF-like, and EC-like cells, respectively. The primary antibody signal was amplified by biotin/streptavidin or tyramide systems conjugated to Alexa Fluor 488 (for α-SMA and MSR) or Alexa Fluor 594 (CD31). (A) Representative images of RCA sections obtained from IGF-1– or saline-injected FH females. Yellow square outlines plaque area magnified in B. (B) SMC, MF, and EC marker–immunopositive cells. Yellow arrows in insert indicate breaks in endothelial layer. (C and D) Quantitative data. n = 5 per RCA or LAD per group for males and n = 9 for females. *P < 0.05 vs. saline based on unpaired 2-tailed t test.

Systemic IGF-1 administration increased expression of pro–α 1(I) collagen in aortic lysates (36), and SMC-specific IGF-1 overexpression increased collagen fibrillogenesis in the atherosclerotic plaque in Apoe–/– mice (37). We found that IGF-1–injected pigs had a trend toward increased collagen levels in the vascular media and in coronary plaques (~10% increase, P = NS) (Supplemental Figure 3, B and C). Thus, IGF-1 changed the cellular composition of porcine coronary plaques. Atherosclerotic lesions in IGF-1–injected pigs had decreased levels of MF-like cells and increased endothelial-like cells compared with controls.

IGF-1 decreases cell apoptosis, reduces systemic oxidative stress, and suppresses inflammation. IGF-1 is a mitogen and prosurvival molecule (38) and exerted antiapoptotic, antioxidant, and antiinflammatory effects in Apoe–/– mice (19, 33). Apoptotic cells in porcine coronary plaques were localized on the luminal border, in the fibrous cap (Figure 5A), and around necrotic cores. IGF-1–injected pigs had an almost 3-fold decrease in cell apoptosis in the male group (P < 0.05 for LAD) and an approximately 2-fold reduction in apoptosis rate in females (Figure 5D). Proliferating cell nuclear antigen (PCNA) is a cell proliferation marker. Atherosclerotic plaques in the female group had increased PCNA immunopositivity compared with males. IGF-1 upregulated PCNA levels in coronary plaques in females (P < 0.05 in RCA) and did not change the PCNA signal in the male group (Figure 5, B and E).

IGF-1 suppresses plaque cell apoptosis, promotes cell proliferation, and decreases DNA damage. Cell apoptosis was quantified by TUNEL assay, and cell proliferation and DNA damage were quantified by immunostaining with PCNA antibody and pH2A.X antibody, respectively. (A–C) Representative images of RCA sections obtained from IGF-1– or saline-injected FH females. (D–F) Quantitative data. n = 5 per RCA or LAD per group for males and n = 9 for females. *P < 0.05, **P < 0.01 vs. saline based on unpaired 2-tailed t test.

Oxidative stress is a major characteristic of hypercholesterolemia-induced atherosclerosis (39). Oxidative DNA damage promotes cell apoptosis and contributes to formation of unstable plaques. Histone H2A.X phosphorylation is a highly specific molecular marker to quantify DNA damage (40). We found that 15%–35% of plaque cells in coronaries contained detectable levels of phosphorylated S139-histone H2A.X (pH2A.X) (Figure 5, C and F), and higher pH2A.X levels correlated with larger atherosclerotic burden seen in females. IGF-1 significantly decreased the number of pH2A.X+ cells in plaques in the female group. Circulating levels of N-tyrosine and total plasma antioxidant capacity (TAC) were measured as indices of systemic oxidative stress (Figure 6). IGF-1 decreased plasma N-tyrosine levels in females at both T3 and T6 time points (56% and 47% decrease, respectively, vs. saline, P < 0.05) (Figure 6A). FH female pigs had a 2-fold reduction in TAC compared with males at T6. IGF-1 upregulated TAC in males and females (Figure 6B), though the increase in males did not reach statistical significance. Taken together, the N-tyrosine and TAC data indicate that IGF-1 suppressed systemic oxidative stress.

IGF-1 downregulates markers of systemic oxidative stress and decreases C-reactive protein and chemokine CXCL12. Markers of systemic oxidative stress (A and B), C-reactive protein (CRP) (C), and chemokine CXCL12 (D). IGF-1 did not alter circulating monocyte subsets (E). Circulating N-tyrosine, CXCL12, and CRP levels were quantified by ELISA and TAC, by using colorimetric assay. TAC assay results shown in urinary acid (standard) equivalents (UAE). (E) Whole blood was mixed with a cocktail of antibodies against CD163-PE, CD14–Alexa Fluor 488, and porcine CD172a and subsequently with streptavidin-APC/Cy7. CD172a-positive leukocytes were size-gated and further differentiated into subsets based on CD163 and CD14 expression levels using FACS. n = 5 per time point per group for males and n = 9 for females for N-tyrosine, CRP, CXCL12, and monocyte assay. n = 5/males and females for TAC assay. *P < 0.05 and **P < 0.01 vs. saline, &P < 0.05 vs. males based on 3-way ANOVA.

CRP is an acute marker of inflammatory responses, and circulating levels of CRP correlate with progression of CAD (41). Human CRP transgene expression causes accelerated aortic atherosclerosis in Apoe–/– mice (42). FH females had higher circulating CRP levels compared with males. IGF-1 significantly decreased CRP levels of both sexes at T3 and T6 (Figure 6C), suggesting a reduction of inflammatory responses. Macrophage-specific IGF-1 reduces chemokine CXCL12 levels, and this effect is associated with decreased atherosclerotic burden in Apoe–/– mice (43). IGF-1 significantly reduced circulating CXCL12 in FH males at T3 and T6 time points and in the female group at T6 (Figure 6D).

The frequency of monocyte subsets has been linked to severity of atherosclerosis in patients with stable CAD (44). We measured 2 subsets of circulating monocytes, defined by surface expression levels of CD163 and CD14. Monocyte subpopulations were assessed by flow cytometry. CD172a+ myeloid cells were size-gated to monocytes, which were further assessed for CD163 and CD14 expression levels, classifying cells as CD163hiCD14lo monocytes and CD163loCD14hi monocytes. The population size of CD163hiCD14lo monocytes was larger in male animals than in female animals (P < 0.001) (Figure 6). No significant effects were induced by IGF-1 administration.

IGF-1 changes the global transcriptomic profile of coronary plaques. To obtain further insights into the effects of IGF-1 on coronary atherosclerosis, we performed an exploratory analysis of advanced plaques from FH females, using the newly developed technology spatial transcriptomics. ST provides an unbiased picture of entire transcriptome changes within a spatial context (27). Prior to running ST analyses, we confirmed the quality of RNA preparations. Total RNA was extracted from plaque-containing RCA cryosections (n = 4), and RNA integrity number (RIN) (45) was quantified using an Agilent Bioanalyzer. RIN was greater than 7, which is considered suitable for ST analysis according to the manufacturer’s (Visium, 10x Genomics) recommendations. RCA cryosections from IGF-1– and saline-injected FH females (n = 2/group) were processed, and ST quality controls (Supplemental Table 4) were consistent with a successful experiment. Correlation of ST gene expression with protein expression (measured by IHC) was verified for selected gene/protein combinations, including IGF-1 binding protein 7 (IGFBP7) and α-SMA (data not shown). Furthermore, changes in gene expression of MMP9, IGFBP7, and FOS measured by ST were validated by real-time PCR using aliquots of mRNA isolated from tissue sections (data not shown).

We performed unsupervised clustering of all ST spots in IGF-1 and saline specimens by using a manufacturer-suggested algorithm with R toolkit Seurat (46). ST spots were grouped into 9 clusters (numbered 0–8) based on their transcriptome (Figure 7B). We identified the top 10 genes overexpressed in each cluster (versus all other clusters) to obtain the heatmap (Figure 7C). In parallel, plaque FC, necrotic/lipid core, tunica media, and tunica adventitia were outlined using H&E images, and transcriptome clusters and histological annotations were compared side by side. FC contained mainly (94.4%) transcriptome cluster 1 and 2, and tunica adventitia cluster 0 and 3 (Figure 7B and Supplemental Figure 4A).

IGF-1 alters global transcriptomic profile of coronary plaque. The RCA cryosections were obtained from IGF-1–injected (n = 2) and saline-injected (n = 2) FH females, and spatially resolved global transcriptome was assessed by ST. ST spots in IGF-1 and saline specimens were grouped into 9 clusters (numbered 0–8) based on their transcriptome, and the heatmap with top 10 upregulated genes/cluster was generated. (A) H&E-stained image with FC outline (white curve). (B) Representative transcriptome clustering. Yellow squares in A and B outline FC fragment shown magnified between panels. Cluster 1 and 2 were identified within a histologically homogeneous FC area. (C) Heatmap. (D) Cell type ratio was calculated for each ST spot to identify spots enriched by SMCs, MFs, or fibromyocytes (FMs). (E) Cell type ratio for IGF-1 versus saline specimens for cluster 1 and 2. (F) Violin plots show expression levels of CXCL14, MMP9, cathepsin D (CTSD), and vimentin (VIM) comparing IGF-1 versus saline specimens in cluster 1 and 2. IGF-1 downregulated CXCL14, MMP9, VIM, and CTSD within cluster 1 and decreased expression of CXCL14, MMP9, and CTSD within cluster 2 compared with saline.

The mixed cell deconvolution algorithms use single-cell RNA-sequencing (scRNA-Seq) data as a reference to characterize cellular heterogeneity in a spatial context (47). However, to our knowledge, there are no scRNA-Seq data available for pig atherosclerotic specimens. Therefore, we used scRNA-Seq data obtained for human atherosclerotic RCA (Gene Expression Omnibus GSE131778) (48) as a reference to assess the cellular composition in our porcine ST data set. Using a deconvolution algorithm (48), we calculated the cell type ratio for each ST spot to identify spots enriched in SMCs (SMC-high), MFs (MF-high), or fibroblast-like cells, termed fibromyocytes (FM-high) (Figure 7D). Of note, we observed good agreement between localization of ST-defined SMC- and MF-high spots and IHC-detected immunopositivity pattern obtained for cell markers. The cell type ratio shows that transcriptomic cluster 4 contained around 80% MF-high spots, cluster 5 had more than 70% SMC-high spots, and virtually all FM-high spots (>95%) were assigned to cluster 2 (Supplemental Figure 4B).

Table 2 shows the top up- and downregulated genes in IGF-1 versus saline specimens identified in all ST spots, and Table 3 contains a list of DEGs in SMC-, MF-, and FM-high spots. IGF-1 dramatically (>10-fold reduction vs. saline) downregulated FOS and FOSB proto-oncogenes in all ST spots and in SMC-high spots. IGF-1 downregulated expression of the cytoskeletal molecule desmin in all spots and upregulated ribosomal protein S27 (Table 2), a component of the translational machinery (49). Activation of translation is in line with known growth-stimulating IGF-1 action (38). MMP9 gene expression level was reduced by IGF-1 in all ST spots and in MF-high spots. This is consistent with our recent report that MF-specific IGF-1 overexpression downregulates MMP9 in peritoneal MFs and decreases aortic atherosclerosis in Apoe–/– mice (43). Cytokine CXCL14 was the top IGF-1–downregulated gene in MF-high spots (Table 3). CXCL14 was upregulated in MF-derived foam cells, and CXCL14 peptide-induced immunotherapy suppresses atherosclerosis in Apoe–/– mice (50), suggesting a proatherogenic role of CXCL14.

IGF-1 versus saline top differentially expressed genes in all ST spots

We observed a clear boundary between transcriptome cluster 1 and 2 within FC histological area (Figure 7, A and B), showing the capability of ST analysis to identify plaque areas with different gene expression patterns that are not histologically distinguishable. To take advantage of this likely unique ST feature and considering the importance of the FC for overall stability of atheroma (51), we focused our subsequent analysis on cluster 1 and 2. First, we found discrete differences in the cellular composition of cluster 1 and 2. More T cells, B cells, and MFs were in cluster 1, whereas fibroblast levels were lower (Supplemental Figure 4B). The top upregulated genes in cluster 1 (versus other clusters) included components of the complement system (C3, C1qA, C1qC, and C1qB) and cathepsin D (CTSD; Figure 7C). The top genes in cluster 2 included CCN2 and CCN3 (CCN molecules are involved in wound healing and fibrosis, ref. 52) and fibronectin 1, which is known to play a vital role in tissue repair (53). A comparison of DEGs between cluster 1 and 2 showed that upregulated CTSD was the top-ranked DEG in cluster 1 versus 2 (~2.1-fold higher expression in cluster 1 vs. 2; adjusted P value = 1.71 × 10–220). CTSD is a proapoptotic molecule and collagen-degrading protease that is highly expressed in MFs (54). Higher cathepsin activity is associated with unstable plaque (54). Ingenuity Pathway Analysis (IPA; QIAGEN) predicted that necrosis and apoptosis pathways are upregulated in cluster 1 versus 2 (data not shown), consistent with increased expression of CTSD. Importantly, cluster 1 is the thinnest part of the FC. Taken together, these data suggest that cluster 1 is a tissue-degrading and less fibrotic compartment of the FC compared with cluster 2. We speculate that cluster 1 represents a plaque site with potentially increased vulnerability and propensity to erode or rupture.

We also compared cell type ratio and DEGs in cluster 1 and 2 in IGF-1 versus saline specimens. IGF-1 slightly decreased MFs in both cluster 1 and 2, and IGF-1 upregulated FMs in cluster 2 (Figure 7E). SMC transition into FMs prevents FC thinning (48), suggesting that higher FM levels in the FC are beneficial during the disease process. IGF-1 reduced CTSD expression in both clusters, whereas vimentin expression (VIM) was decreased only in cluster 1. The top-ranked DEGs in comparing IGF-1 versus saline groups in cluster 1 and 2 included downregulated chemokine CXCL14 and MMP9 (Figure 7F). Thus, use of ST combined with deconvolution algorithms provided, to our knowledge for the first time, a profile of the spatial transcriptome of advanced atherosclerotic plaque and changes induced by IGF-1.