The secretome of liver X receptor agonist‐treated early outgrowth cells decreases atherosclerosis in Ldlr−/− mice

Abstract Endothelial progenitor cells (EPCs) promote the maintenance of the endothelium by secreting vasoreparative factors. A population of EPCs known as early outgrowth cells (EOCs) is being investigated as novel cell‐based therapies for the treatment of cardiovascular disease. We previously demonstrated that the absence of liver X receptors (LXRs) is detrimental to the formation and function of EOCs under hypercholesterolemic conditions. Here, we investigate whether LXR activation in EOCs is beneficial for the treatment of atherosclerosis. EOCs were differentiated from the bone marrow of wild‐type (WT) and LXR‐knockout (Lxrαβ−/−) mice in the presence of vehicle or LXR agonist (GW3965). WT EOCs treated with GW3965 throughout differentiation showed reduced mRNA expression of endothelial lineage markers (Cd144, Vegfr2) compared with WT vehicle and Lxrαβ−/− EOCs. GW3965‐treated EOCs produced secreted factors that reduced monocyte adhesion to activated endothelial cells in culture. When injected into atherosclerosis‐prone Ldlr−/− mice, GW3965‐treated EOCs, or their corresponding conditioned media (CM) were both able to reduce aortic sinus plaque burden compared with controls. Furthermore, when human EOCs (obtained from patients with established CAD) were treated with GW3965 and the CM applied to endothelial cells, monocyte adhesion was decreased, indicating that our results in mice could be translated to patients. Ex vivo LXR agonist treatment of EOCs therefore produces a secretome that decreases early atherosclerosis in Ldlr−/− mice, and additionally, CM from human EOCs significantly inhibits monocyte to endothelial adhesion. Thus, active factor(s) within the GW3965‐treated EOC secretome may have the potential to be useful for the treatment of atherosclerosis.


| INTRODUCTION
Cardiovascular disease remains the leading cause of death worldwide. 1 Atherosclerosis is a vascular complication arising from cardiovascular disease and is typically diagnosed in its later stages, after established lipid-laden plaques have deposited in the aorta. Plaque development is initiated by coordinate dysfunction to various vessel wall cell types, including endothelial and vascular smooth muscle cells, as well as hematopoietic cell types. 2 Prior to plaque deposition, the aortic endothelium undergoes pathological activation resulting from systemic inflammation and elevations in plasma lipoproteins. During endothelial activation, selectins and adhesion molecules are upregulated, which facilitate the binding of monocytes to the endothelium. Adherent monocytes then traverse the endothelium to the intima, and differentiate into macrophages which engulf modified lipoproteins to promote the plaque development characteristic of established atherosclerosis. 3 The liver X receptors, LXRα (Nr1h3) and LXRβ (Nr1h2), belong to the nuclear receptor superfamily of transcription factors. 4,5 Activation of the LXRs induces the gene expression of the cholesterol efflux transporters, Abca1 and Abcg1; and represses proinflammatory gene expression, including Mcp-1, Tnfα, and Il-1β. [6][7][8] These roles of LXRs have been particularly well characterized in macrophages, in part through bone marrow transplant experiments that support an anti-atherogenic role for LXRs. [9][10][11][12][13][14] However, recent bone marrow transplantation studies using bone marrow deficient in LXR target genes have demonstrated that LXRs may elicit their anti-atherogenic roles in other bone marrow-derived cell types apart from monocytes/macrophages. [14][15][16][17] In the bone marrow, hematopoietic stem cells (HSCs) can differentiate into myeloid progenitors, which produce the monocytes/macrophages known to contribute directly to plaque progression. 3,18 HSCs, however, can also differentiate into other cell types including endothelial progenitor cells (EPCs). 19,20 EPCs were initially described in a report by Asahara et al in 1997 as peripheral blood cells that differentiate into endothelial-like cells and contribute to vascular repair by direct incorporation. 19 Many studies have since demonstrated that patients with diabetes or cardiovascular disease have dysfunction in EPC numbers and function. [21][22][23] Over the past 20 years, EPC function has largely been characterized using ex vivo cultures, which yield two cell populations: early outgrowth cells (EOCs; 7-10 days in culture) and late outgrowth endothelial cells (14-21 days in culture). EOCs participate in endothelial repair through the secretion of factors, whereas late outgrowth cells differentiate to endothelial-like cells and promote vascular homeostasis through direct incorporation. [24][25][26][27] Consistent with observations made by Asahara and colleagues, EPC studies were initially focused on their endothelial-like phenotype and how they facilitate neovascularization by incorporation into the damaged vessel. 28  CAD was confirmed by coronary angiography and defined as an epicardial stenosis ≥50%. The presence of type 2 diabetes was defined as hemoglobin A1c (HbA1c) levels ≥6.5% or a prior diagnosis or medical therapy for diabetes. Patient characteristics are provided in Table 1.

| Human EOC culture
Peripheral blood mononuclear cells were isolated by density centrifugation using Ficoll-Paque Plus (GE Healthcare, Marlborough, Massachusetts). The cells were differentiated to EOCs as described above.
For each patient, the mononuclear cells were differentiated to EOCs in both vehicle and 1 μM GW3965.

| CM collection
After 7 days of differentiation, EOCs were washed twice with cold PBS and incubated in factor-and GW3965-free media (EBM-2; Lonza) for 30 minutes at 37 C and 5% CO 2 . The media was discarded and the cells were replenished with fresh EBM-2 media.
The CM was then collected after 24 hours, centrifuged at 700g for 5 minutes to remove any cells, and syringe filtered (0.22 μm). The CM was frozen at −80 C until use. The number of EOCs from each treatment group was unchanged after differentiation (data not shown).     Figure 1A were not increased, suggesting that agonist treatment did not prevent differentiation or induce de-differentiation, but rather produced an altered population of EOCs.

| Splenectomy
To assess the effect of GW3965 at the protein level we performed shotgun proteomics on the EOCs after 1 day or 9 days of differentiation. While CD144 and VEGFR2 proteins were not detectable, we did observe increased EOC expression of another classic endothelial marker, von Willebrand factor (vWF) during differentiation (day 1 vs day 9). Similar to the trends observed for other endothelial markers at the gene expression level, EOC differentiation in the presence of GW3965 resulted in decreased levels of vWF compared to vehicle (Supporting Information Figure S1D). Unsupervised clustering of the data demonstrated that each of the three treatment groups clustered separately, supporting the idea that the GW3965 treatment creates a distinct EOC sub-population (Supporting Information Figure S1E). Time course analyses are shown in the left panels, and day 7 EOC expression levels in the right panel. n = 5-6 per group. Data represent the mean ± SEM. **P < .01, ***P < .001, ****P < .0001 EOC CM. We observed that while the addition of the pro-inflammatory stimulus TNFα did induce the expression of the adhesion molecules VCAM1 and ICAM1, as well as SELE, there was no change in any of these markers of endothelial activation when co-treated with WT GW CM compared the WT Veh CM (Supporting Information Figure S2A). To establish whether the CM was required to "prime" the endothelial cells prior to TNFα administration, we tested whether CM administered only in the pre-TNFα period (20 hours) and then washed out; or only during TNFα treatment (ie, for 4 hours concurrent with TNFα) would be sufficient to elicit the beneficial effect to decrease monocyte-endothelial binding.

| Secretome of GW3965-treated EOCs reduced monocyte endothelial adhesion
Interestingly, CM was effective at decreasing monocyte adhesion under both conditions and the effect was most pronounced when included both before and during TNFα (Supporting Information Figure S2D).  Figure 2B). This therefore demonstrated that GW3965 improved EOC function from CAD patients to a similar extent as that observed in murine cells from WT mice.

| Administration of EOCs and their secreted factors reduce plaque size during atherogenesis
Ex vivo cultured EOCs are currently undergoing clinical trials for the treatment of vascular complications arising from cardiovascular disease, such as pulmonary hypertension and acute myocardial infarction. [38][39][40] In our study, we wanted to assess the therapeutic potential of WT receiving CM from GW3965-treated EOCs in which monocytes were unexpectedly increased (Supporting Information Table S4). Remarkably, treatment with CM from GW3965-treated EOCs reduced atherosclerotic plaques in the aortic sinus by 36% (P < .01) compared to Veh-treated EOC CM and 47% (P < .001) compared to control unconditioned EBM-2 media ( Figure 4C). The CM from the GW3965-treated EOCs significantly reduced VCAM1, the adhesion molecule responsible for leukocyte adhesion to the inflammatory endothelium during atherogenesis ( Figure 4C). The proportions of CD45, a pan-leukocyte marker, were not altered; indicating in part that the CM did not alter the content of the plaque but rather through the effects on the endothelium reduced total lesion area.
No differences in plasma IL-6 levels or spleen weights were found among the treatment groups (Supporting Information Figure S3), F I G U R E 3 LXR agonist-treated EOCs reduce plaque area in the aortic sinus. A, EOCs derived from WT mice were differentiated in the presence of GW3965 and administered to splenectomized Ldlr−/− mice bi-weekly over an 8-week period. B, Plasma cholesterol levels were quantified from the blood collected at sacrifice. C-F, Histological analyses were performed on the aortic sinus sections for total lesion area (C, D) and percent positive areas for VCAM1 (C, E) and CD45 (C, F). n = 5-11 mice per group. Data represent the mean ± SEM. *P < .05 suggesting no overall effect on systemic inflammatory status. While the molecular component(s) contributing to these changes are not clear, these data demonstrate the therapeutic potential of the GW3965-treated EOC secretome to limit the development of atherosclerosis in mice, with in vitro evidence that this response will be conserved in humans.

| DISCUSSION
Atherosclerosis is a progressive disease that manifests over several decades. The gold standard treatment for atherosclerosis relies on lipid-lowering, such as the administration of statin drugs and more recently, PCSK9 inhibitors. 50 However, results from the Canakinumab The secretome of GW3965-treated EOCs decreases atherosclerotic plaque burden. A, The conditioned media from GW3965 treated EOCs were harvested and administered to Ldlr−/− mice bi-weekly over 8 weeks. B, Plasma cholesterol levels were assessed in samples collected at sacrifice. C-F, Histological analyses were performed on the aortic sinus sections for total lesion area (C, D) and percent positive areas for VCAM1 (C, E) and CD45 (C, F). n = 5-12 mice per group. Data represent the mean ± SEM. **P < .01, ***P < .001 Anti-inflammatory Thrombosis Outcome Study (CANTOS) trial found that targeting the inflammatory axis can reduce cardiovascular-related mortality, independent of lipid lowering. 50,51 These data, among others, emphasize the opportunity for the development of novel therapeutic avenues for the treatment of atherosclerosis that go beyond traditional lipid lowering modalities.
During the early stages of atherosclerosis, damage that occurs within the atherosclerotic environment overwhelms endogenous endothelial repair mechanisms, resulting in an overall proinflammatory response that allows for an increase in endothelial expression of adhesion molecules that facilitate leukocyte binding. 3 While the endothelium has the capability to repair itself, endothelial repair can also be supported by EOCs through the secretion of vasoreparative factors. We previously described a role for LXRs in mitigating the negative effects of high cholesterol-induced defects on EOC differentiation and the secretome. 41   CAD produces a secretome that also reduces monocyte-endothelial adhesion ( Figure 2B), indicating that LXR agonist treatment to EOCs could potentially be a translatable therapeutic for patients with cardiovascular disease.
The administration of EOCs to atherosclerotic mouse models has been previously shown to promote plaque stabilization and decrease plaque formation, primarily through endothelial engraftment. [66][67][68][69] However, those studies were all performed in mouse models of plaque regression that targeted the later stages of atherosclerosis. EOC intervention at the early stages of atherosclerosis, as described herein, has not, to our knowledge, been previously explored. As a complement to our in vitro monocyteendothelial adhesion assay, we evaluated the therapeutic potential reports in which the EOC secretome has been packaged in nanoparticles or hydrogels to treat acute ischemic diseases. 71,72 Previous studies determined that in the absence of the cholesterol efflux transporters (Abca1 and Abcg1) in myeloid cells of the bone marrow was insufficient to explain the anti-atherogenic effects of LXRs described in whole bone marrow transplant experiments. [9][10][11][12][13][14] Our data support the role of another effector cell (EPC) in contributing to the anti-atherogenic effects of LXR. Thus, these studies enhance our understanding of LXR cell targets in the bone marrow and provide a unique mechanism to inhibit the development of atherosclerosis. A major drawback to the in vivo use of LXR agonists is the development of hypertriglyceridemia and hepatosteatosis via upregulation of the LXR target gene and lipogenic factor Srebp1c, in the liver. [73][74][75] Our data demonstrate that the cell-free secretome of LXR-treated EOCs is sufficient to protect against endothelial damage and early atherosclerosis.
These data therefore support a role for ex vivo treatment of patient-derived EOCs as an avenue of intervention that would elicit the anti-atherogenic effects of LXRs in the EOCs without the hepatosteatosis associated with systemic LXR agonism.

| CONCLUSION
Our data demonstrate the potential of ex vivo treatment of patientderived EOCs with LXR agonists as a novel therapy to decrease the endothelial defects associated with atherosclerosis and reduce plaque burden during its progression. Further experiments will need to be performed to examine the long-term efficacy and any potential toxicity that may be associated with administering CM prior to human studies.