Impaired vascular remodeling after endothelial progenitor cell transplantation in MMP9-deficient mice suffering cortical cerebral ischemia

Abstract

Endothelial progenitor cells (EPCs) are being investigated for advanced therapies, and matrix metalloproteinase 9 (MMP9) has an important role in stroke recovery. Our aim was to determine whether tissue MMP9 influences the EPC-induced angiogenesis after ischemia. Wild-type (WT) and MMP9-deficient mice (MMP9/KO) were subjected to cerebral ischemia and treated with vehicle or outgrowth EPCs. After 3 weeks, we observed an increase in the peri-infarct vessel density in WT animals but not in MMP9/KO mice; no differences were found in the vehicle-treated groups. Our data suggest that tissue MMP9 has a crucial role in EPC-induced vascular remodeling after stroke.

Introduction

Transplantation of endothelial progenitor cells (EPCs) has become a promising approach to enhance angio-vasculogenic responses after ischemia. EPCs can be mobilized in response to ischemia, can home to sites of neovascularization, and can differentiate into endothelial cells.¹ There are two subsets of EPCs: the so-called circulating angiogenic cells and the outgrowth endothelial cells (OECs).² Both types of EPCs have been demonstrated to increase brain angiogenesis in animal models of cerebral ischemia.^{3, 4, 5}

The degradation and remodeling of the vascular basal membrane is required to allow endothelial cells to migrate, and matrix metalloproteinases (MMPs) have key roles in the initial steps of angio-vasculogenesis. Specifically, MMP9 has been shown to be essential for the invasion of endothelial cells and capillary branching, and in vitro studies have indicated that its absence impairs the angiogenic abilities of the EPCs.⁶ Other authors have described that MMPs have dual roles after ischemia, where their upregulation is detrimental in the acute phase but is essential for effective neurorepair.⁷ Experimental models of cerebral ischemia have proven that inhibiting MMPs can reduce the size of the ischemic lesion.⁸ Published studies on MMP9-deficient mice have demonstrated the pivotal role of this protease in infarct expansion,^{9, 10} although the results remain controversial.¹¹

Our core hypothesis is that tissue MMP9 is a key protease required for an effective and successful cell-based therapy to potentiate vascular remodeling. In this study, we investigated the effects of brain MMP9 deficiency on the cerebral ischemia after EPC-based treatment to potentiate angio-vasculogenesis. To examine this hypothesis, we evaluated the infarct size in the acute phase as well as the long-term peri-infarct angiogenesis in wild-type (WT) and MMP9 knock-out (MMP9/KO) mice after EPC transplantation in a model of cortical stroke.

Materials and Methods

Animals

A total of 50 male WT and MMP9/KO mice (8-to-12-weeks old, FVB background) were obtained from Jackson Laboratories (Sacramento, CA, USA). Both genotypes were bred in-house and the offspring were used for experimentation. All procedures were approved by the Ethics Committee of Animal Experimentation of the Vall d'Hebron Research Institute (58/08/10) and were conducted in accordance with the Spanish legislation and the Directives of the European Union. The ARRIVE guidelines were considered in the design and report of the study.

Permanent Focal Cerebral Ischemia

The left middle cerebral artery was permanently occluded (MCAO) by electrocauterization, affecting the cortex as described.⁶

In Vivo Magnetic Resonance Imaging

Magnetic resonance studies were carried out 24 hours after MCAO using a 7-Tesla horizontal magnet (BioSpec 70/30, Bruker BioSpin, Rheinstetten, Germany). Sixteen coronal slices for T2-weighted fast spin-echo images (T2WI) and eight coronal slices for T2 map and diffusion tensor imaging were acquired. Maps of the trace of the diffusion tensor, also named the apparent diffusion coefficient (ADC), and the fractional anisotropy (FA) were derived using the Paravision software program, v5.0 (Bruker BioSpin). Regions of interest corresponding to the infarct area were traced on T2, ADC, and FA maps in all slices presenting infarcts and in matching contralateral hemispheres. The mean values are expressed as the ipsilateral/contralateral ratios.

Treatments

Growing OECs were obtained and expanded as previously described⁶ from WT FVB mice. The cells were trypsinized, and 5 × 10⁵ cells were prepared in 200 μL of endothelial basal medium (EBM, Lonza, Allendale, NJ, USA). Treatments were randomly administered intravenously (200 μL via retro-orbital sinus) at 30 to 32 hours postischemia to the following groups: WT mice receiving vehicle (EBM, n=15), MMP9/KO mice receiving vehicle (n=8), WT mice receiving OECs (n=15, 5 × 10⁵ OECs in EBM), and MMP9/KO mice receiving OECs (n=8), according to the power calculations described in the 'Statistical Analysis' section. Selected WT and MMP9/KO mice (n=4) received the same cell treatment with green fluorescent OECs labeled in vitro with PKH26 (Sigma, St Louis, MO, USA), a general cell membrane green fluorescent dye.

Measurement of the Infarct Lesion

The hyperintense area corresponding to the injured tissue and the total areas of both ipsilateral and contralateral hemispheres were traced manually on T2W images by investigators who were blinded to the genotype. The edema index and lesion volume were calculated as previously described.⁴

Behavioral Testing

The corner test (used to assess sensorimotor asymmetries) and grip strength meter test (used to measure the maximum strength of the mouse forelimbs) were used to evaluate the functional outcomes. The tests were performed blindly before surgery and were repeated 1, 7, 14, and 21 days after MCAO as described elsewhere.¹²

Evaluation of the Brain Vasculature

To assess the poststroke angio-vasculogenesis, mice were injected intravenously with 80 μg of Dylight 594-labeled tomato (Lycopersicon esculentum) lectin (Vector Laboratories, Burlingame, CA, USA) and were euthanized by cardiac perfusion of 4% paraformaldehyde under deep anesthesia 21 days after the induction of ischemia. Brains were collected, and 12-μm-thick coronal sections were selected from the anterior area (+1 to +0.2 bregma), including the lateral ventricles. Two images ( × 100 magnification) of the peri-infarct cortex were taken with a laser confocal spectral microscope (Olympus FV-1000, Olympus, Tokyo, Japan), and the total area of lectin-positive vessels was calculated (ImageJ software, NIH, Bethesda, MD, USA) by an investigator masked to the genotype/treatment. The results are expressed as the mean vessel density.

Doublecortin (DCX) and NeuN Immunohistochemistry

EPC-induced neurogenesis was assessed in contiguous coronal sections, which were immunostained with antibodies for mature neurons (NeuN; Millipore, Billerica, MA, USA) and neuroblasts (DCX; Abcam, Cambridge, UK). The presence of DCX-positive cells was analyzed in the dorsolateral and subventricular zones of the ipsilateral hemisphere, and mature neurons were measured in the peri-infarct cortical area (n=7 to 8/group).

Statistical Analysis

SPSS 15.0 software (IBM, Armonk, NY, USA) was used for the statistical analyses. All variables were normally distributed, and the data are expressed as the mean±s.d. Statistical significance was assessed by Student's t-test or an analysis of variance followed by the least significant difference post-hoc test. A P-value <0.05 was considered to be statistically significant at a 95% confidence level.

After a pilot study including n=7/8 mice/treatment, power calculations were evaluated using Ene 2.0 (http://sct.uab.cat) considering a P-value <0.05% and 50% statistical power. The number of animals was raised up to 15 for WT vehicle and WT receiving OECs to assess vessel density.

Results

MMP9 Deficiency Reduces the Cortical Ischemic Lesion

The ischemic lesions were evaluated by magnetic resonance imaging 24 hours after permanent MCAO in 32 animals. MMP9 deficiency entailed significant neuroprotection, reducing the infarct volume by 25% in MMP9/KO animals (17±1.5 mm³ in WT versus 12.8±0.9 mm³ in MMP9/KO, P=0.025; Figure 1A). The edema index was also significantly reduced in the MMP9/KO mice (P=0.026), which indicated decreased swelling. In accordance with these findings, the T2 map values showed reduced lesions in the MMP9/KO mice (P=0.007; Figure 1B). Measurements of the ADC and FA in the infarct area confirmed that the severity of the lesion (ADC) and axonal damage (FA) within the infarct core were similar between genotypes (Figure 1B). In addition, there were no significant differences between the treatment subgroups (vehicle vs. EPCs) before the treatment administration.

Figure 1.

Magnetic resonance imaging 14 hours after cerebral ischemia. (A) A decrease in the infarct volume and edema index was observed in matrix metalloproteinase-9 knockout (MMP9/KO) versus wild-type (WT) mice by T2WI at 24 hours after ischemia. Representative images show severe cortical damage in mice of both genotypes. (B) The T2, apparent diffusion coefficient (ADC), and the fractional anisotropy (FA) were measured to assess the extent of brain damage, and relative values (to the contralateral cortex) are shown. The data are expressed as the mean±s.d., and each individual value is represented. N=16 per group; t-test, *P<0.05.

Neurorepair with OECs Enhanced the Cortical Angiogenesis Only in WT Mice

The cortical vessel density was analyzed 21 days after the induction of ischemia. One animal in the MMP9/KO vehicle group died 2 weeks after treatment after suffering severe aggression from the other mice housed in the same cage. Interestingly, no correlation existed between the vessel density and the magnetic resonance imaging parameters (infarct volume and ADC and FA values) in any of the groups. Green fluorescent cells were observed to be associated with the brain vessels in the peri-infarct areas in both WT and MMP9/KO mice receiving cell therapy (see Figure 2A), confirming the arrival of administered cells to the brain and its incorporation to the brain vasculature. When we compared the amount of functionally perfused microvessels in the peri-infarct area, there were no significant differences between the WT and MMP9/KO mice that received the vehicle (P=0.220). Regarding OEC treatment, there was a significant increase in the vessel density in WT mice that received OECs compared with the WT vehicle controls, as expected (P=0.028, Figure 2B). However, there was no significant difference in the vessel density between the MMP9/KO vehicle vs. MMP9/KO OEC groups (P=0.672), although there was an increased vessel density in the WT vs. MMP9/KO mice treated with OECs (P=0.001), which suggested that endogenous MMP9 is required for the pro-angiogenic actions of OECs.

Figure 2.

The brain peri-infarct vasculature in wild-type (WT) and matrix metalloproteinase-9 knockout (MMP9/KO) mice 21 days after cerebral ischemia. Functional blood vessels stained after lectin perfusion were quantified in the ipsilateral peri-infarct cortex 21 days after ischemia. (A) The right panel represents PKH67-positive outgrowth endothelial cells (OECs) in vitro. PKH67-stained OECs were detected in the brain vasculature (lectin-stained vessels) 24 and 72 hours after intravenous administration both in WT and MMP9/KO ischemic brains (bar=50 μm). (B) Representative micrographs of lectin staining in cortical areas (bar=200 μm). (C) A study of the vessel density showed a significant increase in the number of vessels in WT vs. MMP9/KO mice after OEC treatment (analysis of variance followed the least significant difference post-hoc test, n=7 to 15/group). The data are expressed as the mean±s.d., and each individual value is represented; *P<0.05; **P<0.01. DAPI, 4,6-diamidino-2-phenylindole.

Immunohistochemistry studies showed a non-significant decrease in DCX+ and NeuN+ cells in the MMP9/KO mice that received OECs compared with WT mice with the same treatment (0.51±0.26% vs. 0.83±0.42% for DCX+, P=0.087 and 342.4±69.3 vs. 437.9±131.4 for NeuN+, P=0.09). Moreover, when the differences were analyzed among all treatment groups, no statistically significant difference was observed (P=0.243 for the DCX-positive area and P=0.359 for the NeuN+ neuron counts).

The corner test and the grip strength meter test failed to show any impairment after ischemia (pre- vs. 24-hour scores; P>0.05), and no significant differences during the follow-up among the groups were observed (P>0.05).

Discussion

The current results suggest that MMP9 has a key role in the tissue vessel remodeling mediated by cell therapy with EPCs after cerebral ischemia. The upregulation of MMPs, including MMP9, after cerebral ischemia and the contribution of the MMPs to brain injury have been widely documented.^{7, 8} Although pharmacological or immunological inhibition of MMP9 has proven to be neuroprotective, as evidenced by a decreased infarct size, discrepant results have been reported in MMP9-deficient mice.^{8, 9, 11} These discrepancies could be attributed to the use of different experimental models of cerebral ischemia (proximal and distal MCA occlusion in the presence or absence of reperfusion) and/or to the use of different techniques to assess the lesion extension (tissue staining for 2,3,5-triphenyl-tetrazolium-chloride). The present study includes the largest sample size of WT and MMP9/KO mice used to evaluate the infarct volume (n=16 per group), and our results support the previous findings showing that MMP9 inhibition reduces the inflammatory response, blood–brain barrier disruption, and the development of edema after cerebral ischemia.^{9, 10}

In addition to its detrimental effects in the acute phase, MMPs also participate in tissue remodeling after brain injury, and MMP9 has been suggested to participate in the unique tissue biobridge between neurogenic and non-neurogenic sites.¹³ After a stroke, the pharmacological inhibition of MMPs with broad-spectrum drugs demonstrated that they have essential roles to ensure an effective neurorepair after cerebral ischemia and poststroke angiogenesis.¹⁴ However, there have been no previous studies that have analyzed the specificity of MMP9 using genetically modified ischemic mice with poststroke EPC treatment to enhance angiogenesis.

We have recently described that there was an increase in the peri-infarct vessel density in WT ischemic mice compared with sham mice, which was not observed in ischemic MMP9/KO individuals.⁶ In the present study, WT and MMP9/KO ischemic mice subjected to long-term brain ischemia showed a similar vessel density 21 days after cortical ischemia. The discrepancy between this finding and those of other studies analyzing the effects of MMP inhibitors in cortical remodeling after stroke¹⁴ may be attributed to the small infarct volumes typical of the FVB strain in our model of cerebral ischemia,¹² the role of other MMPs, the longer time point for the tissue analysis in our study, or to differences in the neurorepair mechanisms in mice vs. rats. In this regard, a possible dual role of MMP9 in angiogenesis that depends on the endogenous tissue vascularization has been observed by other authors, as this protease was required for ischemia-induced angiogenesis after hindlimb ischemia,¹⁵ whereas in cardiac tissue MMP9-deficiency facilitated angiogenesis after myocardial infarction.¹⁶

The most important finding of our study, which we believe will be of great importance to the development of future cell-based neurorepair strategies after stroke, is the MMP9 tissue dependency after EPC therapy. To date, several different subsets of EPC populations have been demonstrated to have therapeutic potential in experimental models of cerebral ischemia.^{3, 4, 5} In the present study, we used OECs (a homogenous and highly proliferative cell population of endothelial-like cells) as a pro-angiogenic treatment for stroke, which was shown to be MMP9 dependent for tubulogenesis in an in vitro Matrigel substrate.⁶ Here we showed that the intravenous administration of OECs enhanced the vessel density in the peri-infarct area in WT animals, but this augmentation did not occur in MMP9-deficient mice. These results suggest that tissue MMP9 has an important role in the vascular remodeling as part of pro-angiogenic and vascular remodeling treatments. Because the administration route was systemic (intravenous), we hypothesize that the MMP9 in the tissue could inhibit the downstream pro-angiogenic effects of transplanted OECs at the ischemic site and/or the expected basal lamina remodeling during angiogenesis. Additionally, coupled EPC-induced neurogenesis was not altered in the MMP9 lacking brains, and this was not modified by EPC therapy, although further investigation will be necessary to confirm this data. A major limitation of our study is the lack of behavioral data supporting the findings for angiogenesis. In this regard, long-term behavioral assessments are difficult in distal MCAO models and might differ among strains,¹² thus suggesting the need for confirmation using other strains or ischemia models; multicentric studies would be an excellent platform for this purpose.

In summary, MMP9 deficiency provides protection against the adverse effects of cortical cerebral ischemia in mice, although the protease is unequivocally required for the success of neurorepair therapies based on vascular remodeling, such as EPC transplantation.

The authors declare no conflict of interest.