Intravascular administration of soluble SARS-CoV-2 Spike protein triggers non-infective heart and lung inflammation and vascular pericytes rarefaction in healthy mice

Background and aim: COVID-19 is a systemic microvascular disease characterised by inflammation and ultimately multiorgan damage in severely ill patients. The SARS-CoV-2 virus initiates the infection, however, still little is known about the potential pathogenic role of the Spike (S) protein presented to human cells as a constituent of the viral capsid or shedding from viral particles. Here we performed a translational preclinical study to verify the harmful effects of a recombinant S extracellular domain (S-ECD).
Methods: The animal study is covered by the British Home Office and follows the principles stated in the Guide for the Care and Use of Laboratory Animals. Healthy 9-week-old CD1 mice (6 male, 6 female) were randomised to receive 10 µg S-ECD or vehicle via a single IV injection in the tail (n=6/group). Experimental groups had equal gender distribution. After 3 days, animals were culled and peripheral blood (PB), hearts and lungs harvested for molecular and histological analyses. Human primary cardiac pericytes (PCs) were used for in vitro studies.
Results: Using an ELISA we confirmed the successful delivery of the S protein (27 ± 7.8 ng/ml plasma), these levels being comparable to those previously measured in COVID patients. Moreover, animals in the S-group showed increased levels of the complement anaphylatoxin C5a (91 vs. 37 pg/ml plasma, S vs. Veh, P = 0.03) and a halved PB neutrophil count (10 vs. 20 % of CD45+ cells, S vs. Veh, P < 0.05). We also documented immunoreactive C5a accumulation in the hearts of the S-group (69 vs. 54 IOD, S vs. Veh, P = 0.04), which was associated with the homing of CD45+ cells to the heart (24 vs. 9 cells/mm2, S vs. Veh, P = 0.0005). Analysis of apoptosis in the hearts revealed enhanced PC death (105 vs. 8 TUNEL+ PC/10000 PC, S vs. Veh, P = 0.0014). Cardiomyocytes and endothelial cells viability was not affected. The increased PC death resulted in a significant PC rarefaction (300 vs. 500 PC/mm2, S vs. Veh, P < 0.0001) and decreased vascular coverage by PCs in the hearts of the S-group. Capillary density remained unchanged. The hearts of S-treated mice were also characterised by augmented ERK1/2 phosphorylation (60 vs. 40% P-ERK+ nuclei, S vs. Veh, P = 0.026), a signalling pathway previously shown to be activated by both S protein and C5a. Histopathological analysis of lungs demonstrated increased cellularity with inflammatory infiltrate and interstitial oedema in the S-group. Moreover, in this latter, dilated and congested blood vessels were present and in one case a fibrin plug was evident within the capillaries. There was no evidence of hyaline membrane formation. Last, after verifying that human cardiac PCs express the C5a receptor in vitro, we collected evidence that PCs challenged with the human recombinant C5a (≥ 10 nM) display increased cell death, as showed by a viability assay (average 3-fold increase of EthD-III+ dead cells in the 10 nM C5a group vs. Veh, n = 5 PC, P < 0.05) and increased caspase 3/7 activity (1.2-fold increase 10 nM C5a vs. Veh, n = 4 PC, P < 0.01).
Conclusions: Our results suggest that soluble S protein released in COVID patients’ blood could induce cardiac and pulmonary harm through non-infective mechanisms, together with triggering a systemic inflammatory response. PCs could have a crucial role by initiating, or contributing to, microvascular damage. C5a could be the key player linking inflammation to vascular damage triggered by PCs. These findings could have therapeutic implications for COVID-19 patients.