Introduction

Increased leukocytosis, associates with worse cardiovascular outcomes in large clinical cohorts1. Specifically, bone marrow (BM) derived monocytes and neutrophils accumulate in the failing myocardium. Once recruited, leukocytes release inflammatory cytokines and proteases that destabilize inducing heart failure2. Because myeloid cells, like monocytes and neutrophils, circulate only for a few days after release from the BM3, examining the pathways that instigate their overproduction in cardiovascular disease (CVD) may provide a key to curbing cardiovascular morbidity. Central to the pathogenesis of ischemia, are structural and functional vascular changes which occur system-wide. Yet, how this condition affects the BM vasculature, which regulates hematopoietic stem and progenitor (HSPC) activity and leukocyte trafficking4, is unknown.

Methods

We studied the BM vasculature in mice with acute MI. We examined the femoral vasculature with immunofluorescence and flow cytometry. Next, we performed intravital microscopy at the skull before and after MI to examine structural and functional changes, specifically (i) angiogenesis, (ii) integrin activation, (iii) vascular leakage and (iv) blood flow. (i) We imaged AplnCreER;Rosa26ZsGreen/+ reporter mice5, in which sprouting endothelial cells and their progeny express ZsGreen. We visualized (ii) fluorescent RGD probe that binds to integrin αVβ3. We studied (iii) endothelial barrier function with intravital time lapse imaging after labeling the blood pool with fluorescent albumin6. We assayed (iv) blood velocity in the skull BM. Next, we translated our microscopy studies towards non-invasive whole mouse imaging, enabling system-wide in vivo assessment of all hematopoietic sites. Therefore, an RGD-containing gallium-68 PET reporter (68Ga-RGD) and a gadolinium-labeled albumin derivative were co-injected into mice with MI, followed by a joined PET and dynamic contrast enhanced MRI protocol6. PET data were validated with autoradiography and scintillation counting. Finally, we used two mouse models: (i) anti-Vegfr2 antibody treatment in mice after MI; (ii) Cdh5CreERT2;Kdrfl/fl animals, in which endothelial cells lack expression of Vegrf2. We performed immunofluorescence, flow cytometry and intravital microscopy for permeability and integrin activation in these two models. Data are displayed as mean±SEM. Statistical analysis was done using two-tailed Mann-Whitney unless otherwise stated.

Results

We hypothesized that post-MI angiogenesis in the bone marrow facilitates emergency hematopoiesis, perhaps by expanding the vascular niche in which myelopoiesis occurs. We detected in mice after MI increased BM vessel density (Fig. 1a,b). Serial intravital microscopy documented increasing marrow vascularity post-MI (Fig. 1c-e). Flow cytometry revealed an increased number of BM endothelial cells (BMEC) and their proliferation rates with a peak at day 6 after MI(Fig. 2a, b)6. Using confocal imaging on day 6 post-MI, we identified newly developed ZsGreen+ BM vessels (Fig. 2d) together with an increased amount of vascular sprouts (Fig. 2e). We observed in mice at day 2 after MI integrin activation and accelerated vascular leakage (Fig. 3a-d). To probe endothelial dysfunction, we next quantified blood velocity in skull marrow microvessels after intravenous acetylcholine injection. For healthy vasculature7, acetylcholine increased velocity in BM arterioles of control mice and decreased velocity decreased in mice post-MI (Fig. 3 e, f).Next, we complemented our intravital microscopy studies with non-invasive whole mouse imaging, enabling system-wide in vivo assessment of all hematopoietic sites. Using 68Ga-RGD PET and a gadolinium-labeled albumin, which relied on imaging agents corresponding to the fluorescent companions used for intravital microscopy, indicated increased integrin binding and higher vascular permeability in the femur after MI (Fig. 4a-d). Autoradiography (Fig. 4e) and scintillation counting (Fig. 4f) confirmed higher 68Ga-RGD accumulation in the bones of mice with acute MI. Analysis of 68Ga-RGD in various hematopoietic sites identified the highest activity in the sternum (Fig. 4g), possibly due to its proximity to the infarcted heart.We next hypothesized that post-MI angiogenesis in the BM facilitates emergency hematopoiesis, perhaps by expanding the vascular niche in which myelopoiesis occurs. Inhibiting the cognate receptor Vegfr2 with a neutralizing antibody abolished the post-MI increase in BM arterioles, and sinusoids (Fig. 5a-c). In line with these ex vivo data, intravital microscopy detected reduced integrin presence and less vascular leakage, which increase during angiogenesis45, in post-MI mice treated with Vegfr2 blocking antibody (Fig. 5d, e). More importantly, Vegfr2 inhibition after MI reduced the amount of BMECs and promoted HSPC quiescence and lowered the amount of circulating Ly6Chigh monocytes (Fig. 5f). To verify if reduced myelopoiesis could be secondary by inhibiting BM angiogenesis or alternatively caused by blocking the Vegfr2 receptor on hematopoietic cells52, we bred Cdh5CreERT2 with Kdrfl/fl mice for tamoxifen-inducible, endothelial cell-specific deletion of Vegfr2. In these mice, in which baseline hematopoiesis was unaffected, post-MI hematopoiesis and circulating Ly6Chigh monocytes (Fig. 5g) declined, indicating that Vegf signaling affects leukocytosis via BM endothelial cells.

Discussion

The changes in BMEC structure and function may facilitate trafficking of cells and signals across the BM endothelium, potentially supporting HSPC activation, processes that augment leukocytosis. Our data on BMEC Vegfr2 is supported by prior reports that Vegfr2 is essential for vascular recovery after irradiation8, as well as clinical data in patients after MI have high blood VEGF levels9.

Conclusion

In vivo imaging revealed that cardiovascular disease affects the BM vasculature's structure and, most importantly, expands the vascular niche after MI and this effect is mediated by endothelial Vegfr2.

Acknowledgements

The authors thank Robert S. Fujinami (University of Utah School of Medicine, Salt Lake City, UT, USA) for sharing mouse strains, the MGH Mouse Imaging Program for assistance with imaging, Hye-Yeong Kim for radiolabeling imaging tracers, the Center for Skeletal Research Core (NIH P30 AR066261) for histological processing, the HSCI-CRM Flow Cytometry Core for assistance with cell sorting and Kaley Joyes for editing the manuscript. This work was funded in part by U.S. federal funds from the National Institutes of Health (HL142494, HL139598, HL128264, HL131478, HL125428 and HL131495), the Global Research Lab (GRL) program (NRF-2015K1A1A2028228) within the National Research Foundation by the Korean government, the MGH Research Scholar Program; the Deutsche Forschungsgemeinschaft (RO5071/1-1 and HO 5279/1-2) and the Ministerio de Economía y Competitividad y Fondo Europeo de Desarrollo Regional (SAF2014-56546-R and RTI2018-101105-B-I00).