Human mesenchymal stromal/stem cells recruit resident pericytes and induce blood vessels maturation to repair experimental spinal cord injury in rats

Abstract

Angiogenesis is considered to mediate the beneficial effects of mesenchymal cell therapy in spinal cord injury. After a moderate balloon-compression injury in rats, injections of either human adipose tissue-derived stromal/stem cells (hADSCs) or their conditioned culture media (CM-hADSC) elicited angiogenesis around the lesion site. Both therapies increased vascular density, but the presence of hADSCs in the tissue was required for the full maturation of new blood vessels. Only animals that received hADSC significantly improved their open field locomotion, assessed by the BBB score. Animals that received CM-hADSC only, presented haemorrhagic areas and lack pericytes. Proteomic analyses of human angiogenesis-related factors produced by hADSCs showed that both pro- and anti-angiogenic factors were produced by hADSCs in vitro, but only those related to vessel maturation were detectable in vivo. hADSCs produced PDGF-AA only after insertion into the injured spinal cord. hADSCs attracted resident pericytes expressing NG2, α-SMA, PDGF-Rβ and nestin to the lesion, potentially contributing to blood vessel maturation. We conclude that the presence of hADSCs in the injured spinal cord is essential for tissue repair.

Figure 1

hADSCs and CM-hADSC increased the number of blood vessels at the edges of lesion cavities after a compressive spinal cord injury. Low-magnification micrograph photomontages of the spinal cord (horizontal sections, approximately 500 -750 μm down from the dorsal surface, identified as the lesion epicentre by GFAP reactivity) of rats treated with DMEM (A), CM-hADSC (C) and hADSC (E), 1 week after SCI, showing blood vessels and phagocytic cells, identified with Isolectin IB4 (red, A, C, E), and its interaction with astrocytes, identified with anti-GFAP (green, A, C, E), and DAPI-stained nuclei (blue, A, C, E). Confocal images under high magnification of the boxed fields show the nervous tissue at the edges of lesion cavities (B, D, F). Areas marked by in yellow boxes were further amplified and shown in the red channel only to better reveal the morphologies of blood vessels. White arrows indicate blood vessels and white arrowheads indicate phagocytic cells (red round cells). White thick dotted lines delineate lesion cavities. Scalebar: A, C, E 3 mm; and B, D, F 200 μm .

Figure 2

Treatment with CM-hADSC and hADSC promoted increased vascular length in relation to control DMEM animals. (A–D) Quantification of the total length of RECA-1 immunostained blood vessels per tissue area at the grey (GM) (A, C) or white matters (WM) (B, D), in the perilesional region, 1 week (A, B) and 8 weeks (C, D) after injury. (E) Graph representing the number of blood vessels in the acute and chronic phases of SCI. (F) Graph representing vessel junction density, indicating the number of branching points of RECA-1 vessels. All quantitative analyses evaluated 3 animals per experimental group, and were performed in horizontal sections. In all graphs the first, second and third columns refer to DMEM, hADSC and CM-hADSC animals, respectively. Values in graphs represent the means ± standard errors. *P < 0.05; **P < 0.01; ***P < 0.001.

Figure 3

Treatment with hADSC, but not CM-hADSC, promoted functional recovery after spinal cord compression. Graph representing the locomotor performance, which was assessed by the BBB score. Animals were evaluated weekly, during eight weeks after spinal compression. The curves described in the graph in black, blue, green and red line represent animals from the experimental groups SHAM (surgical control, without laminectomy), DMEM, treated with CM-hADSC and hADSC, respectively. Values in graphs represent the means ± standard errors. *P < 0.05; **P < 0.01; ***P < 0.001.

Figure 4

hADSCs favoured maturation of the blood brain barrier. (A–C) Confocal images of the spinal cord (horizontal sections) of rats treated with DMEM (A), CM-hADSC (B) and hADSC (C), labelled with TUJ-1 (green, white-filled arrowheads), Isolectin IB4 (red, white arrows) and DAPI (blue), which recognizes young axons, endothelial and phagocytic cells, respectively. (D, E) High magnification confocal images of TUJ-1 immuno-labelled axonal fibers (green) growing in close contact with blood vessels (red) in hADSC treatment. Note that some axonal fibers were projected into the injury cavity (E, white arrow). (F) Quantification of total length of blood vessels per tissue area (immuno-labelled with specific blood brain barrier antibody anti-SMI-71), 1 WPI. (G, H, I) Confocal images of the spinal cord (horizontal sections) of rats treated with DMEM (G), CM-hADSC (H) and hADSC (I), showing blood vessels identified with the blood brain barrier protein, SMI-71 (red, white arrows), and astrocytes, identified with anti-GFAP (green), and DAPI-stained nucleus (blue). (J, K) High magnification confocal images show that in CM-hADSC animals the BSB protein SMI71 pattern is punctuated (J, white arrowheads), while it is continuous in hADSC–treated animals (K). (L) High resolution confocal image showing the close contact and interaction between GFAP-astrocyte (green) and the blood brain barrier (red) positive cells (white asterisk). White-filled arrowheads indicate axonal fibers (TUJ-1 positive), white arrows indicate blood vessels (anti-SMI-71 or IB4 positive), arrowheads with white outline indicate vascular breakpoints, white asterisk indicates astrocyte-blood vessel interaction, and thick white dotted lines delineate lesion cavities. Black asterisks in graphs indicate statistical differences between DMEM (n = 10), CM-hADSC (n = 10) and hADSC groups (n = 10). Values in graphs represent the means ± standard errors. *P < 0.05; **P < 0.01; ***P < 0.001. Scalebar: C, E, I—50 μm; J, L—20 μm.

Figure 5

Vascular immaturity such as bleeding, plasma leakage and fenestrations in CM-hADSC-treated SCI animals. (A, B) Macroscopic images of fresh horizontal sections of rats treated with CM-hADSC (A) or hADSC (B), 1 WPI. (C, D) Confocal bright-field reconstructions of fresh horizontal sections at the lesion epicentre (approximately 250–500 μm down from the dorsal surface) of rats treated with CM-hADSC (C) or hADSC (D), 1WPI, showing hemorrhagic areas only in CM-hADSC (A, C, red arrows). (E, F) High magnification images of the boxed areas in C and D, respectively, where red arrows point to a large amount of red blood cells in CM-hADSC animals (E), and substantially less in hADSC—treated animals (F). (G, I) Confocal superimposed bright-field and fluorescence images of histological sections from CM-hADSC (G) and hADSC-treated animals (I), 1 WPI, which were stained with fluorescent anti-rat IgG (red) to detect IgG extravasation. (H, J) High magnification images of the boxed areas in G and I, demonstrated plasma extravasation (red, white asterisk) close to the hemorrhagic region in CM-hADSC animals (H) but not in hADSC treated animals (J). Scalebar: A, C: 500 μm, E, H: 50 μm; G: 200 μm.

Figure 6

Pericytes were distributed along the blood vessels and in close contact with the vascular wall only in hADSCs treated animals. Confocal images of the spinal cord (horizontal sections) of rats treated with DMEM (A, D, G, J), CM-hADSC (B, E, H, K) and hADSC (C, F, I, L, M, N, O), 1WPI. (A, B, C) Histological sections were labelled with anti-alpha smooth muscle actin antibody (red, α-SMA), Isolectin IB4 (green, IB4) and DAPI (blue), which recognize pericytes, endothelial and phagocyte cells and nuclei, respectively. (D, E, F) Histological sections were immuno-labelled with pan-laminin antibody (in green), alpha smooth muscle actin antibody (red, α-SMA,) and DAPI (blue), to identify the basement membrane of the blood vessels, pericytes and nuclei, respectively. (G, H, I) Alternative pericyte marker, platelet-derived growth factor receptor-beta antibody (in red, PDGFR-β) and RECA-1 (green) blood vessel immunolabeling was performed to identify pericytes and endothelial cells, respectively. (J, K, L) Histological sections were labelled with neural/glial antigen 2 antibody (red, NG2), Isolectin IB4 (green, IB4) and DAPI (blue), to identify pericytes, endothelial and phagocytic cells and nuclei, respectively. (M, N, O) To certify that these vessel-associated cells were pericytes, we performed a triple staining: (1) anti-neural/glial antigen 2 (anti-NG2, red); (2) anti-α-SMA (green) and (3) Isolectin IB4 (white). The first two are considered reliable pericyte markers (anti-NG2 and/or anti-α-SMA), and the latter (IB4) identifies endothelial cells. The orthogonal view of the optical slice of confocal microscopy demonstrated the positive colocalization between anti-NG2 and anti-α-SMA (N, O). Arrowhead with white fill indicates blood vessels (IB4 positive), white arrows indicate pericytes (α-SMA, NG2, or PDGFR-β positive), arrowheads with white outline indicate close association between pericyte markers (anti-NG2 and/or anti α-SMA) and white dotted lines delineate lesion cavities. Scalebar: C, L: 100 μm, F, I, M: 20 μm.

Figure 7

hADSCs altered the secretion of soluble pro-angiogenic factors when transplanted into the injured nervous tissue. (A) Table describing the 55 angiogenesis-related human proteins contained in the human-specific angiogenesis array. (B) Table reporting the semi-quantification of angiogenesis-related factors (first column) produced by hADSCs either in culture (“hADSCs in vitro”, second column) or in the rat spinal which had received the injection of the hADSCs (“hADSCs in vivo”, third column) or conditioned medium of cultured hADSCs (“CM-hADSC in vivo”, fourth column), after normalization relative to the positive control of the assay. C) Graph representing the relative concentrations of angiogenesis-related factors secreted in vitro (blue column) or in vivo, in hADSCs-treated animals (red column) or CM-hADSC-treated animals (green column). Note that the profile of angiogenic factors detected in vitro and in vivo are largely different.

Figure 8

hADSCs attracted endogenous pericytes to the lesion site but did not differentiate directly into pericytes or endothelial cells. (A) Confocal images of transverse sections of the spinal cord, 1WPI, with GFP-transduced hADSCs (green). The nervous tissue was immuno-labelled with alpha smooth muscle actin antibody (anti-α-SMA) to identify pericytes (red) and counterstained with DAPI for cell nuclei (blue). Note that GFP + hADSCs accumulated predominantly in the central canal and in perilesional area. (B) High magnification confocal image of the yellow boxed area in A, where GFP + positive cell (green) clusters were seen surrounded by α-SMA positive cells (red), suggesting that hADSCs can attract pericytes (red, α-SMA). (C) Confocal image of the spinal cord of the animal treated with GFP + hADSCs (green) immuno-labelled with anti-pan-laminin (white), to identify the vascular basement membrane, anti-platelet derived growth factor receptor-β (PDGFR-β, red), a pericyte marker, and counterstained with DAPI for cell nuclei (blue). White arrowhead indicates PDGFR-β positive cells inside the perivascular space, between endothelial and astrocytic basement membranes. (D, E, F, G). Confocal images of a transverse (D) and a horizontal section (E, F, G) of the spinal cord from GFP + hADSC (green) animals immuno-labelled with -α-SMA (red, D, F), RECA-1 antibody (red, E) and counterstained with DAPI, to identify pericytes, endothelial cells and cell nuclei (blue), respectively. (G) Confocal image of a vascular structure, where a GFP + hADSC (green) is seen at the tip of the sprout. In images in F and G the histological sections were visualized in confocal microscopy superimposed with bright field. White arrows indicate GFP + hADSCs (D, E, F). CC indicates the central canal and MD, the spinal midline. White dashed lines indicate the midline of the spinal cord and thick white dotted lines delineate lesion cavities (A) or lumens of blood vessels (D, F, G). Scalebar: A: 100 μm; B–G: 50 μm.

Figure 9

hADSCs attracted nestin positive cells, possibly expressed by resident pericytes that migrated along the perivascular space. (A–C) Fluorescence microscopy images of the spinal cord (horizontal sections) of rats treated with DMEM (A), CM-hADSC (B), hADSC (C), counterstained with DAPI for cell nuclei (blue), 1WPI. (D–I) Confocal images of horizontal sections of rats treated with DMEM (D, G), CM-hADSC (E, H), hADSC (F, I), immuno-labelled with anti-nestin (green), anti-pan-laminin (red) and counterstained with DAPI (blue) to identify neural precursor cells, the basal lamina of blood vessels and cell nuclei, respectively. The white arrowheads indicate the separation between the two basal laminas of the blood vessels and the white arrows indicate the presence of nestin positive cells. (J–L) Confocal images of consecutive spinal cord sections of hADSC-treated animals, 1WPI, immuno-labelled with anti-nestin, anti-NG2, anti-Ki67 and counterstained with DAPI, respectively, to identify neural precursor cells (green), pericytes (red), proliferative activity (red), and cell nuclei (blue), respectively. Note the presence of nestin positive cells that are proliferating in the perivascular region (L). (M) To certify that these pericytes express nestin, we performed a double immunostaining: (1) anti-nestin (green); (2) anti-α-SMA (red), and DAPI-stained nucleus (blue). Note that there is a positive colocalization for anti-nestin and anti-α-SMA immunostaining, suggesting that pericytes after SCI express nestin. The asterisks mark lumens of blood vessels, arrowhead with white outline indicates nestin/Ki67 double positive cells. White dotted lines delimit cavity areas, and the continuous trace in yellow indicates the lumens of blood vessels. Scalebar: C: 500 μm; F, J, K: 100 μm; G, H, I: 10 μm; L, M: 20 μm.

Figure 10

Mesenchymal cell therapy increases angiogenesis in the spinal cord after a moderate balloon-compression injury. The scheme represents the angiogenic process after compression of the spinal cord in control animals (A), and in animals treated with culture media conditioned by human adipose tissue-derived stromal/stem cells (CM-hADSC) (B) or human adipose tissue-derived stromal/stem cells (hADSC) (C). (A) In the control animal (DMEM medium) few blood vessels grow spontaneously around the lesion site. (B) The injection of the CM-hADSC, rich in pro-angiogenic growth factors, promotes an increase in the number of blood vessels, but they do not have an organized vascular structure, the vascular tubules are very branched, thin, and are not covered with mature pericytes. The lack of vascular maturity results in hemorrhages scattered in the spinal cord and is highlighted by a patchy expression of the BSB marker SMI71. (C) Treatment with hADSCs promotes an increase in the vascularization of the spinal cord, with a mature and well-defined vascular architecture. Vessels are characterized by a continuous expression pattern of the BSB marker SMI71. The blood vessels are wider, less branched, and covered by pericytes. The modification of the proteomic profile of the injected cells, mediated by the lesion environment itself, directs the vascular maturation process, attracting pericytes, which stabilize the new blood vessels. Both pericytes and neural precursors that express nestin, which are attracted by hADSCs, migrate through the perivascular space between the two layers of the basal vascular lamina. This promotes functional recovery after SCI.