Brain protection using autologous bone marrow cell, metalloproteinase inhibitors, and metabolic treatment in cerebral ischemia

Abstract

Despite advances in imaging, understanding the underlying pathways, and clinical translation of animal models of disease there remains an urgent need for therapies that reduce brain damage after stroke and promote functional recovery in patients. Blocking oxidant radicals, reducing matrix metalloproteinase-induced neuronal damage, and use of stem cell therapy have been proposed and tested individually in prior studies. Here we provide a comprehensive integrative management approach to reducing damage and promoting recovery by combining biological therapies targeting these areas. In a rat model of transient cerebral ischemia (middle cerebral artery occlusion) gene delivery vectors were used to overexpress tissue inhibitor of matrix metalloproteinase 1 and 2 (TIMP1 and TIMP2) 3 days before ischemia. After occlusion, autologous bone marrow cells alone or in combination with agents to improve NO bioavailability were administered intraarterially. When infarct size, BrdU incorporation, and motor function recovery were determined in the treatment groups the largest beneficial effect was seen in rats receiving the triple combined therapy, surpassing effects of single or double therapies. Our study highlights the utility of combined drug, gene, and cell therapy in the treatment of stroke.

Approximately 400 persons per 100,000 over 45 years of age have a first stroke each year in the United States, Europe, and Australia (1). Approximately 600,000 people per year in the United States suffer a new or recurrent stroke (2). Stroke is also a major cause of long-term disability (1, 3). i.v. thrombolysis with t-PA improves outcome if treatment is initiated within 3 h after symptom onset (4). In specialized stroke units, intraarterial revascularization (within 6 h after symptom onset) may improve flow better than i.v. thrombolysis (4). Innovative strategies include novel fibrinolytics (tenecteplase, desmetolplase, and microplasmin), glycoprotein IIb/IIIa antagonists (abciximab and tirofiban), and externally applied ultrasound to enhance fibrinolysis (5). Novel endovascular reperfusion strategies include intraarterial thrombectomy (clot retrieval and suction thrombectomy) and mechanical disruption (microguidewire passage, laser photoacoustic emulsification, and intracranial angioplasty) (5). Magnetic resonance can rapidly assess infarct core and site of occlusion, penumbra, and tissue hemorrhagic propensity, enabling improved selection of patients for reperfusion. However, stroke units able to do all current therapies are not very disseminated worldwide, and in some cases the therapeutic approach for stroke patients is the same as of some decades ago; e.g., in Italy only 10% of 200,000 stroke patients per year receive venous thrombolysis (E. Stucchi and D. Ovadia, personal communication). Thrombolysis is also associated with a markedly increased risk of symptomatic hemorrhage (4–6). Thus, new and simple therapeutic stroke treatment approaches are needed.

Bone marrow cells (BMCs) contain populations of precursors that are multipotent and can differentiate into bone, cartilage, and mesenchyma (7), neurons and glia (8), and endothelial cells (9, 10). BMCs have been demonstrated to cross the blood–brain barrier (BBB) (11). Functional improvement and reduction in cell death in the ischemic regions have been reported after both intracerebral (12) and i.v (13, 14) BMC delivery. Several mechanisms are involved in the BMC neuroprotective effects after middle cerebral artery occlusion (MCAO) in the rat, including induction of angiogenesis and cytokine secretion (12–15). Conversely, the increase in extracellular matrix protease activity and activation of matrix metalloproteinases (MMPs) play a pivotal role contributing to brain damage after ischemia (16–19). MMP activity in vivo is tightly controlled by the endogenous tissue inhibitors of MMPs (TIMPs), a family of glycosylated proteins, which, in addition to their inhibitory functions on MMPs, may also affect cellular differentiation and proliferation (20, 21). MMPs and TIMPs are also likely to play key roles in the repair phases of ischemia, particularly during angiogenesis and reestablishment of cerebral flow (20, 21). Induction of TIMP overexpression may therefore reduce ischemic damage by either reducing exaggerated MMP activity or activating neuroprotective signals and is likely dependent on the choice of inhibitor. Notably, synthetic MMP inhibitors could not mimic this beneficial effect (22), and studies highlight the importance of a therapeutic window for MMP inhibition with such suppression resulting in exacerbation of lesion size and reduced recovery (23, 24). Gene transfer can result in efficient production of proteins for a transient or long period by a single injection of vector, overcoming disadvantages of infection caused by retention of catheters and brain damage caused by repeated therapeutic injections. To date, virus-mediated overexpression of TIMPs has not been used as a means to confer neuroprotection in vivo. Furthermore, neuroprotection could be improved by metabolic treatment, a concept supported by a great body of evidence indicating beneficial roles of treatment with antioxidants and l-arginine, the natural precursor of NO, during vascular dysfunction (9, 25–28). Antioxidants mitigate oxidative and neuronal damage after ischemic stroke (29–31). Because augmentation of NO production increases cerebral flow, which can lead to neuroprotection during brain ischemia, several modalities that increase NO bioavailability are under investigation (32, 33). Here we investigated a multifaceted approach on brain rebuilding by using an integrative management protocol with autologous BMCs, transgene overexpression of TIMPs, and metabolic treatment in rat MCAO.

Results

Study Design.

The study design is shown in Fig. 1.

Fig. 1.

The treatment groups subjected to MCAO and reperfusion (45 min later) at day 0. Rats were pretreated with AdTIMP1/TIMP2, l-arginine/vitamin E, or a combination of these. Autologous BMCs were infused 6 h after MCAO and reperfusion. Animals were killed 14 days after MCAO for histologic studies. Group 1 (control) containing rats receiving only MCAO and rats receiving MCAO + control vector virus. Subsequently, for clarity, these groups in Figs. 3–5 and Tables 1 and 2 were kept separately and named Group (MCAO alone) and Group 2 (MCAO + control vector virus).

Infarct Volume [2,3,5-Triphenyltetrazolium Chloride (TTC)], TIMP Transgene Delivery, and Oxidative/NO Pathways.

Normal brain stains red with the TTC method, but the cerebral infarcts (both cortex and basal ganglia) in the brains of the MCAO rats showed reduced staining delineating areas of infarction (data not shown). Histologic analysis of the lesions is shown in Fig. 2. Under experimental conditions we saw a significant reduction in TTC staining when BMCs were injected. Lesions became further progressively reduced when BMCs were administered in association with AdTIMP1/TIMP2 and further more in cotreatment with l-arginine/antioxidants. The group receiving BMCs in association with metabolic treatment alone did not differ significantly from the group receiving BMCs and TIMPs (data not shown). The 3-day pretreatment with AdTIMP1/TIMP2 alone before MCAO resulted in a significant reduction of TTC staining. Preliminary experiments indicated that (i) AdTIMP1 or AdTIMP2 delivery alone, but not in combination, achieved a nonsignificant reduction of brain infarction, and (ii) antioxidant alone achieved a trend toward reduction of TTC staining that did not reach statistical significance (data not shown). At 24 h after TIMP transgene delivery, RT-PCR detection revealed a mean of 1.6 × 10⁷ TIMP1 mRNA and 1.3 × 10⁷ TIMP2 mRNA (n = 3), and we found that TIMP1 transgene expression was earlier than TIMP2 (Fig. 3A). These values increased after 2 days (2.8 × 10⁷ TIMP1 mRNA and 2.6 × 10⁷ TIMP2 mRNA, n = 3) and peaked on day 3 (6.1 × 10⁷ TIMP1 mRNA and 5.4 × 10⁷ TIMP2 mRNA, n = 3). The intraassay variability, determined in three replicates of a brain sample assayed in a single experiment and expressed in terms of a coefficient of a variation of Ct value, was 1.6% for TIMP1 and 1.3% for TIMP2. The interassay variability, obtained by 10 repetitions of the same sample in three different experiments, was 3.0% for TIMP1 and 2.9% for TIMP2. As expected, metabolic treatment (groups 5 and 7) induced a significant amelioration of NO bioactivity and oxidative stress measured as lipid peroxidation (Table 1). Brain infarct size in groups 4 and 6 correlated inversely with nitrite and nitrate (NOx) levels (r = −0.56 and −0.49, respectively; P < 0.01) and positively with lipoperoxides (r = 0.46 and 0.41, respectively; P < 0.05), consistent with a neuroprotective role of reduced oxidative stress and enhanced NO bioactivity.

Fig. 2.

Infarct volume values (TTC) in six groups. ∗, P < 0.05 vs. group 1; ∗∗, P < 0.01 vs. group 1 and P < 0.05 vs. group 2; ○, P < 0.001 vs. group 1 and P < 0.01 vs. groups 2 and 3; ○○, P < 0.05 vs. group 1; §, NS vs. group 1.

Fig. 3.

Brain effects of joint treatments. (A) RT-PCR for TIMP1 and TIMP2 at 1, 2, and 3 days after transgene delivery. (B) Immunostaining for BrdU on coronal sections. BrdU⁺ cells were detected in the ipsilateral cortex near the infarct boundary (B1 and B2) and subventricular area (B3 and B4). (C) Cumulative BrdU staining in the groups. ∗, P < 0.001 vs. groups 1, 2, 6, and 7; ∗∗, P < 0.0005 vs. groups 1, 2, 6, and 7; §, P < 0.05 vs. group 3. Group 1 (control), only MCAO; group 3, autologous BMCs injected 6 h after MCAO; group 4, BMCs in association with AdTIMP1/TIMP2 pretreatment; group 5, BMCs in association with AdTIMP1/TIMP2 and concurrent treatment (120 mg/kg l-arginine i.p. per day and 2.4 mg/kg vitamin E per day); group 6, MCAO in association with AdTIMP1/TIMP2. Group 7 received MCAO with the metabolic treatment. (D) Representative double-immunofluorescent staining (merged) in ischemic brains under laser-scanning confocal microscopy of the BrdU⁺ cells (green) coupled to expression of GFAP (red), VWF (red), MAP-2 (red), and Neu-N (red).

Table 1.

Plasma lipoperoxides and NOx 3 days after MCAO

Group	Lipoperoxides, μmol/liter	NOx, mM
1	0.84 ± 0.22	22.30 ± 5.8
2	0.91 ± 0.21	23.67 ± 6.9
3	0.85 ± 0.30	27.97 ± 7.2
4	0.80 ± 0.32	26.55 ± 6.8
5	0.58 ± 0.35*	40.11 ± 7.4*
6	0.79 ± 0.28	26.66 ± 5.1
7	0.63 ± 0.16*	35.32 ± 6.8*

Data are expressed as mean ± SD. Group 1 (control), only MCAO; group 2, MCAO plus control vector virus; group 3, autologous BMCs injected at 6 h after MCAO; group 4, BMCs in association with TIMP1 and TIMP2 inhibitors; group 5, BMCs in association with TIMP1 and TIMP2 inhibitors and concurrent metabolic with 120 mg/kg l-arginine i.p. per day and 2.4 mg of vitamin E per kilogram of body weight per day; group 6, MCAO in association with TIMP1 and TIMP2 inhibitors. Group 7 received MCAO with the metabolic treatment.

*, P < 0.01 vs. groups 1, 2, 3, and 4.

Autologous BMCs Stimulate Homing to Brain After Cerebral Ischemia.

To determine whether BMCs homed in on the injured brain tissue of treated rats, BrdU labeling was used to follow the engraftment of BMCs injected in the brain circulation. BrdU-immunoreactive (BrdU⁺) cells were detected mainly in the subventricular area of the lateral ventricle in both hemispheres of BMC-treated rats, indicating that they reached these brain areas (Fig. 3B). Cumulative labeling of BrdU also revealed few BrdU⁺ cells in the ipsilateral cortex near the infarcted boundary and subventricular region (Fig. 3B). Quantitative analysis revealed that BrdU⁺ cells in the brain of BMC-treated rats significantly increased 7 and 14 days after MCAO compared with controls. Indeed, BrdU pulse labeling was used to quantify the BrdU⁺ cells in the ischemic hemisphere of rats at 7 and 14 days after MCAO compared with controls (Fig. 3C). Results were consistent with the study hypothesis; BMC-treated groups had significantly higher levels of BrdU⁺ cells that further increased in groups 4 and 5. Finally, the group receiving BMC in association with metabolic treatment did not differ significantly from the group receiving BMC and TIMPs (data not shown).

Combined Treatment of BMCs with TIMPs and Metabolic Supplementation Enhances Neurogenesis and Angiogenesis.

To determine whether injected BMCs differentiated into neuronal, glial, or endothelial cells in the brains of treated rats (groups 3–5), double-staining experiments were performed. Several BrdU⁺ cells (green for nucleus identification) colocalized with Abs for neuronal nuclei (Neu-N), microtubule-associated protein 2 (MAP-2), glial fibrillary acidic protein (GFAP), and von Willebrand factor (VWF) (red for cell type–specific markers) (Fig. 3D). Data were quantified (Table 2). Ischemic cortical areas of BMC-treated rats revealed an increase in BrdU⁺ cells coexpressing the neuronal phenotypes of Neu-N⁺ and MAP-2⁺ cells and the glial phenotype of GFAP⁺ cells or vascular phenotype (VWF⁺ cells). Data were more evident in group 5 receiving the triple therapy (BMC, AdTIMP1/TIMP2, and l-arginine/vitamin E) in comparison to groups 3 (BMCs alone) and 4 (BMC and AdTIMP1/TIMP2). Also in this case, a group receiving BMCs in association with metabolic treatment alone did not differ significantly from the group receiving BMCs and TIMPs (data not shown). Enlarged and thin-walled vessels in the cortical ischemic boundary zone (IBZ) are indicative of angiogenesis (34). Treatment with BMCs significantly increased vascular perimeter and density in the cortical penumbra area compared with control rats (Fig. 4), which indicates that BMCs enhance angiogenesis in ischemic brain.

Table 2.

Percentages of BrdU⁺ cells in the ischemic brain and percentage of GFAP⁺, VWF⁺, MAP-2⁺, and Neu-N⁺ cells within the BrdU⁺ cell population in the study groups

Groups	BrdU+	GFAP+	VWF+	MAP-2+	Neu-N+
1	18.40 ± 0.40	12.10 ± 0.35	20.15 ± 0.29	10.08 ± 0.24	24.66 ± 0.34
2	16.92 ± 0.31	13.04 ± 0.25	21.10 ± 0.21	11.66 ± 0.25	23.18 ± 0.25
3	31.85 ± 0.30*	19.22 ± 0.19*	27.55 ± 0.20*	13.22 ± 0.19	33.28 ± 0.29*
4	34.40 ± 0.32*	20.77 ± 0.22*	31.27 ± 0.22*	15.75 ± 0.20*	33.55 ± 0.28*
5	39.33 ± 0.35**	26.58 ± 0.18**	35.68 ± 0.26*	18.68 ± 0.21*	38.43 ± 0.23**
6	25.21 ± 0.24*	10.89 ± 0.20	23.56 ± 0.31	11.57 ± 0.18	23.89 ± 0.25
7	22.28 ± 0.16	15.85 ± 0.26	24.68 ± 0.28	12.66 ± 0.15	27.85 ± 0.27

Data are expressed as mean percentage ± SE. Groups are as in Table 1.

*, P < 0.05 vs. groups 1 and 2;

**, P < 0.01 vs. groups 1 and 2.

Fig. 4.

Vessel perimeter and density. VWF immunostaining shows enlarged thin wall vessels (vessel perimeter and density) along cortical IBZ in the BMC-treated group (group 3) or in combination with TIMPs (group 4) and metabolic supplementation (group 5). ∗, P < 0.05 compared with groups 1, 2, 6, and 7; §, P < 0.05 vs. group 3.

Motor Function Recovery.

MCAO caused marked impairment in the ability to remain on the rod 2 h after ischemia. The BMC treatment alone or in combination with TIMPs and/or metabolic supplementation showed significant improvement in neurological outcome 14 days after MCAO (Fig. 5). The group receiving BMCs in association with metabolic treatment alone did not differ significantly from the group receiving BMCs and TIMPs (data not shown).

Fig. 5.

Motor function recovery. A repeated-measures ANOVA revealed that a significant interaction was found among treatments and time (days) for duration on rod. A Mann–Whitney U test revealed that treated rats with BMCs and TIMPs with and without metabolic treatment showed significant improvement in neurological outcome. Group 1 (control), only MCAO; group 2 (control), MCAO and control vector virus; group 3, autologous BMCs injected 6 h after MCAO; group 4, BMCs in association with AdTIMP1/TIMP2 pretreatment; group 5, BMCs in association with AdTIMP1/TIMP2 and concurrent metabolic treatment (120 mg/kg l-arginine i.p. per day and 2.4 mg/kg vitamin E per day; group 6, MCAO in association with AdTIMP1/TIMP2; group 7 received MCAO with the metabolic treatment. ∗, P < 0.05 vs. groups 1 and 2; #, P < 0.01 vs. groups 1 and 2; +, P < 0.05 vs. group 3; §, P < 0.05 vs. groups 3 and 4.

Discussion

We show that intraarterial administration of autologous BMCs alone or in combination with gene transfer to overexpress TIMPs and/or metabolic supplementation into the rat MCAO model resulted in enhanced benefits confirmed by both histologic and behavioral analyses. These results indicate the feasibility of this integrated therapeutic approach in acute cerebral infarction.

BMCs may provide a neuroprotective influence. The trophic influence of transplanted BMCs migrating to the damaged lesions may prevent the ischemic brain from injury. We show that injected BMCs migrated to nondamaged brain tissues, as well as to the moderately damaged tissues, the latter being destined to be irreversibly damaged by the ischemic stress. Their trophic influence continues over an extended period, suggesting strong neuroprotective effects on the damaged CNS tissues. BMCs may integrate into and replace the tissue damaged by the ischemic stress. We show that BMCs accumulated within the extensively damaged tissues in the ischemic lesions and differentiated into both neural and glial phenotypes within the lesions. Woodbury et al. (35) suggested that BMCs can differentiate into neural phenotypes in vitro. BMCs are likely to retain the capacity to respond to local epigenetic signals and to differentiate with a terminal phenotype appropriate for that ectopic site in the ischemic brain that represents a different environment than the intact CNS (11, 36). Together, injected BMCs survived and were likely to differentiate into at least a modest number of neuronal and glial cell phenotypes in lesions, although it remains unclear how much they contributed to the behavioral rescue and improvement. Here the significant therapeutic effect in the acute phase is likely from a more “global” and as yet undefined neuroprotective effect. As we showed in the present study and in another study (14), very early intervention is most effective, and progenitor stimulation and migration would take time. Thus, whether endogenous progenitor stimulation may contribute to later phases of the therapeutic benefits remains to be defined at the molecular level. The intracarotid delivered cells are unlikely to migrate from the vascular system into the CNS tissue under normal conditions because they may not pass through the BBB between the vascular interface and CNS. In the present study, however, the intracarotid transplanted BMCs accumulated in the ischemic brain. The ischemic stress likely resulted in a partial disruption of the BBB that may have allowed the transplanted BMCs access to the parenchyma. Indeed, the permeability of the BBB begins to increase 3 h after ischemia (37).

Chemical homing factors may contribute to the extensive migration of the transplanted BMCs into the ischemic brain lesions, although natural migration of BMCs into the CNS may occur. Hypoxic–ischemic brain injury also stimulates local neural stem or progenitor cells to generate new neurons and oligodendrocytes (38). Ischemic brain tissue extract selectively induces chemotaxis of mesenchymal progenitors, and the expression of chemotactic factors such as monocyte chemoattractant protein 1 (39), intercellular adhesion molecule, vascular adhesion molecule 1, and E-selectin (40) is also high. These molecules may promote the targeting of endogenous (41) or transplanted BMCs to the ischemic lesions. Although the therapeutic effects of bone marrow transplantation after 6 h from MCAO are associated with a variety of events such as cell targeting, disruption of the BBB, neuroprotection, regeneration, and antiinflammation, our data indicate that early treatment facilitates recovery both at the level of infarct volume and at the level of brain function. Increased angiogenesis promotes neurological functional recovery after stroke (34). Our data demonstrate that BMCs injected intraarterially increase vessel diameter and density in the cortical IBZ. This finding is consistent with an important study from Chopp and colleagues (42), which demonstrated BMC-enhanced angiogenesis in the IBZ in the host brain after MCAO. This increase may enhance delivery of oxygen and trophic factors to the local environment and thus promote local axonal sprouting and synaptogenesis. We showed increased effects of BMCs on angiogenesis, synaptogenesis, and gliogenesis in the ischemic brain. Although a variety of immature cells such as embryonic stem cells and neural stem cells are thought to be possible donors for use in stroke therapy, BMCs are relatively accessible and could be harvested in autologous form upon patient admission after insult. Clearly, delivery at 6 h is beneficial, suggesting clinical utility of this approach. Because BMCs are a donor source for autotransplantation therapy, ethical, infectious, and immunological concerns of cell therapy are naturally solved. It will be important to define whether subpopulations of BMCs are responsible for the effect and whether enrichment of such cells provides additional benefit.

We also show in the setting of MCAO the utility of gene therapy using TIMPs, particularly when given in combination with other agents. Overexpression of TIMP1 and TIMP2 by first-generation adenovirus-mediated gene transfer effectively reduces the extent of neuronal damage induced by MCAO. There is a pivotal role of exaggerated MMP expression in the pathogenesis of several CNS diseases (43). After ischemia the early up-regulation of MMP expression can be detrimental for the brain and the BBB (17–19, 44–46). Nevertheless, a beneficial effect of late MMP expression in recovery from CNS injury cannot be ruled out (17). Rivera et al. (47) reported an induction of TIMP1 expression after global ischemia. Early after the ischemic insult they observed increased TIMP1 expression in regions resistant to the injury, whereas at a late point intense and persistent expression was detected in areas with a great potential for neurogenesis after ischemia (48). These findings suggest that TIMP1 might either prevent apoptosis or repair mechanisms after brain injury, depending on the timing of expression, the regions where expression occurs, and the cellular source. TIMP2 is detected in basal conditions, and induction of TIMP2 has been described some days after ischemia (43). TIMP2 delivery was shown to inhibit MMP2 activity and reduce proteolytic BBB opening in focal ischemia (49). Because MMP activity contributes to the damage that occurs after ischemia, TIMP neuroprotection is likely to occur through classical MMP inhibition. However, TIMPs have a wide variety of other functions, not all of which are attributed to MMP blockade, such as apoptosis/survival-modulating mechanisms (50). These pathways are involved in cerebral ischemia (51, 52). After cerebral injury, astrocytes and microglia are activated and overexpress MMPs (53). The finding of a small reduction in astrocytic reaction in the TIMP-treated rats could be important in understanding the role of TIMPs in the injured CNS. Decreased astrocytosis may be an epiphenomenon or contribute indirectly to neuroprotection after ischemia by ameliorating astrocytic activation. Thus, this section of our study gives additional evidence for the importance of the MMPs and their inhibitors in neuronal death and indicates a therapeutic potential for transgene TIMP1 and TIMP2 overexpression. The effect of gene transfer alone was minimal but was substantially improved by combination treatment. Our experimental regimen used preinjection of gene therapy 3 days before MCAO. This allows for maximal TIMP expression but minimizes the utility of the approach clinically. In future experiments it will be important to assess (i) whether gene delivery can be beneficial at time points closer or after MCAO, and (ii) whether infusion of recombinant TIMP1 and TIMP2 at the time of BMC infusion (6 h after injury) provides the equivalent benefit to preinjection of adenoviruses.

Excessive production of radicals is known to lead to cell injury in a variety of diseases, including cerebral ischemia (29–31). Beneficial effects can be obtained from (i) inhibition of radical production, (ii) scavenging of radicals, and (iii) increase of radical degradation by endogenous antioxidants. Similarly, administration of NO donors and/or l-arginine in stroke may reduce stroke lesions in permanent and transient models (32, 33). Here the combined treatment of BMCs with TIMPs and metabolic supplementation (antioxidants and l-arginine) added further beneficial neuroprotective effects. Thus, autologous BMC treatment alone or in combination with transgene delivery of TIMPs and antioxidants/l-arginine after MCAO significantly improves histologic and functional outcome in brain ischemia. This benefit was related to BMC homing to the brain, enhanced neurogenesis/angiogenesis, NO bioactivity, and decreased systemic oxidative stress and MMP activity.

Methods

Study Design and Cerebral Ischemia Model.

This study was conducted according to the Guidelines for Animal Experiments of the American Heart Association and rules of the National Institutes of Health (publication No. 85-23, revised 1985) and approved by the institutional care of experimental animals committees. All efforts were made to minimize the number of animals used and their suffering. Quality standards of laboratories at the University of Naples are in accordance with rules established by the Italian Ministry of Health and the European College of Laboratory Animal Medicine, and Laboratories of the Mayo Clinic, University of California (Los Angeles), and the Universities of Glasgow are in accordance with standards of the Association for Assessment and Accreditation of Laboratory Animal Care. We induced transient MCAO for 45 min by using intraluminal vascular occlusion, as described (32). Briefly, adult male Wistar rats weighing 250–300 g were anesthetized and maintained under anesthesia with mechanical ventilation. A length of 20.0-mm 3-0 surgical suture with the tip rounded was advanced from the external carotid artery into the lumen of the internal carotid artery until it blocked the origin of the middle cerebral artery. After 45 min of MCAO, reperfusion was performed by withdrawal of the suture. Physiological parameters (rectal temperature, blood pH, pO₂, pCO₂, and blood pressure) were maintained within normal ranges during surgery and transplantation procedures for all animals. Experiments consisted of six groups (n = 66) (Fig. 1). Doses of l-arginine and vitamin E were based on preliminary pilot experiments and literature (29, 33).

Neuroprotective Treatment with Autologous BMCs and Transgene Delivery of TIMPs.

BMCs were obtained autologously from the femurs of rats 72 h before transplantation, as described (9). The same rats from which BMCs were obtained (groups 2–4) were subjected later to MCAO. Approximately 2 × 10⁷ BMCs in 250-μl total fluid volume (L-15 medium) were injected into the carotid artery 6 h after MCAO. This time was chosen on the basis of preliminary experiments and those performed in another study (14). Briefly, anesthesia was reinstituted, and a modified polyethylene catheter (PE-50; Becton Dickinson, Sparks, MD) was advanced through the ECA into the lumen of the ICA for a distance of <15.0 mm, lodging <2 mm proximal to the origin of the middle cerebral artery. Construction of replicant-deficient human adenovirus serotype 5 encoding human TIMP1 (Ad TIMP1) and TIMP2 (Ad TIMP2) under control of the cytomegalovirus major immediate early promoter (CMVIEP) was described (54). In preliminary experiments, we evaluated gene infection at different doses of virus (10⁹ to 10¹¹ viral particles). These experiments showed a dose-dependent increase of viral vector expression that was maximal 3 days after viral vector administration (see Results). In the present study, groups 3–5 received via the carotid artery (same route as BMC delivery) an equal mix of AdTIMP1 and AdTIMP2 totaling 3 × 10¹¹ viral particles 3 days before MCAO (to have the maximal transgene overexpression of TIMPs during ischemia). Efficiency of TIMP overexpression was detected by using a multiplex quantitative real-time RT-PCR method based on TaqMan technology. Probe and primers were selected by the proprietary software Primer Express (PE Applied Biosystems, Milan, Italy). Enzymes and reagents used for reverse transcription and PCR were from PE Applied Biosystems. Four hundred nanograms total RNA was reverse-transcribed by MultiScribe transcriptase. PCR was performed in the ABI Prism 7700 Sequence Detector (PE Applied Biosystems) starting from 25 ng of cDNA, in a reaction mix containing 300 nM forward and 900 nM reverse primers (for both genes). Results were expressed as femtograms of HT1080 RNA per microgram of total RNA. Primers used were as follows: TIMP1 sense, 5′-TCTGGCATCTGGCATCCTC-3′; TIMP1 antisense, 5′-AGCAAAGTGACGGCTGCTCTGG-3′; TIMP2 sense, 5′-GCGGTCAGTGAGAAGGAAGTGG-3′; and TIMP2 antisense, 5′-CTTGCACTCGCA-GCCCATCTG-3′ (54, 55).

BrdU Labeling.

A cumulative pulse labeling of proliferating cells in the rat brain was obtained after daily administration of BrdU (50 mg/kg i.p.; Sigma, St. Louis, MO) from the second day after MCAO until after 14 days of ischemia (56).

TTC Staining and Quantitative Analysis of Infarct Volume and IBZ.

At 14 days after MCAO, rats were killed, and brains were removed and dissected into coronal 1-mm sections by using a vibratome. The fresh brain slices were immersed for 30 min in a 2% solution of TTC (57). The cross-sectional area of infarction in each brain slice was examined with a dissection microscope and measured by using imaging analysis software (9, 25, 26). For measurements in the cortex, eight fields of view from the cortical ischemic penumbra (cortical IBZ) were digitized. The total infarct volume for each brain was calculated by summing the infarcted area of all brain slices.

Immunohistochemistry of Brain Tissue and Laser Confocal Microscopy for Double-Immunofluorescence Analysis.

For BrdU immunostaining, DNA was denatured (58). The immunostaining procedure was performed with the use of the labeled streptavidin-biotin method (DAKO LASB-2 Kit; Milan, Italy). After deparaffinization, tissue slides were incubated with the appropriate diluted Abs to BrdU (for nuclear identification; 1:400; Roche, Basel, Switzerland) at room temperature for 1 h, then washed with TBS containing 0.1% Tween 20 and incubated for 10–30 min with biotinylated anti-rabbit and anti-mouse immunoglobulins (1:200; R & D Systems, Abingdon, U.K.) and peroxidase-labeled streptavidin. Sections were counterstained with hematoxylin. Quantification of BrdU-immunoreactive cells was performed on paraffin-embedded tissue sections and counted digitally (9, 25, 26, 59). To identify expression of cell type-specific markers in BrdU⁺ cells, double immunofluorescence was performed (9, 59). Expression of GFAP, VWF, MAP-2, and Neu-N was tested. Each coronal section was first treated with primary BrdU Ab conjugated with FITC (1:500; Jackson ImmunoResearch, Suffolk, U.K.) staining, followed by treatment with cell-specific Abs: GFAP for astrocytes (1:400; Sigma), VWF for endothelial cells (1:400; Sigma), Neu-N for neuronal nuclei (1:200; Chemicon, Hampshire, U.K.), and MAP-2 for neuronal dendrites (1:200; Roche) with Cy3 (1:500; Jackson ImmunoResearch) staining. Tissue sections were analyzed with a Leica TCSS82 laser-scanning confocal microscope, and green (FITC) and red (Cy3) fluorochromes were excited by laser beam at 488 and 543 nm, respectively. For measurement of vascular density and perimeters in the cortical IBZ, each VWF-positive coronal section was digitized and counted.

Systemic NOx and Lipoperoxides.

NO was measured as nitrites and nitrates, which are stable metabolites of NO, and the concentration of NO in the blood (50 μl from the tail vein) was assessed. NOx levels were measured with Griess reagent according to the manufacturer's instructions (Calbiochem, San Diego, CA) (9). Similarly, plasma lipid peroxidation was evaluated by the Lipid Peroxide Kit (Kamiya, Thousand Oaks, CA), as described (60).

Motor Function Test.

By using the rotarod (IITC Life Science) test (61), rats were given training sessions on an accelerating rod from 5 to 15 rpm for 4 days before MCAO, and only the rats that were able to stay on the rotating rod at 15 rpm for 200 s were subjected to MCAO. Test sessions consisted of five trials at 15 rpm and were performed just before MCAO and at 2 h and at 2, 8, and 14 days after MCAO by an investigator who was blinded to the experimental groups (61, 62). If the rats managed to maintain their balance for 200 sec the trial was ended. The final score was expressed as the mean time that a rat was able to remain on the rod for the five trials.

Statistical Analysis.

All values are expressed as mean ± SD. Statistical analysis was performed by using SPSS system (9.0J). Cell numbers were analyzed by the Student t test. Rotarod data were analyzed by repeated-measures ANOVA and the Mann–Whitney U test. P < 0.05 was considered significant.

Abbreviations

BBB

blood–brain barrier

BMC

bone marrow cell

MCAO

middle cerebral artery occlusion

MMP

matrix metalloproteinase

TIMP

tissue inhibitor of matrix metalloproteinase

IBZ

ischemic boundary zone

TTC

2,3,5-triphenyltetrazolium chloride

GFAP

glial fibrillary acidic protein

VWF

von Willebrand factor

MAP-2

microtubule-associated protein 2

Neu-N

neuronal nuclei

NOx

nitrite plus nitrate.