Osteopontin Enhances Endogenous Repair After Neonatal Hypoxic–Ischemic Brain Injury
Background and Purpose—Hypoxic–ischemic (HI) brain injury is a frequent cause of perinatal morbidity and mortality with limited therapeutic options. To identify molecules important for cerebral damage and repair, we investigated the growth factor-related gene expression profile after neonatal cerebral HI. We identified osteopontin (OPN) as the most highly upregulated factor early after HI. We therefore explored the role of endogenous OPN in brain damage and repair.
Methods—Nine-day-old wild-type mice were exposed to cerebral HI; growth factor-related gene expression profiles were analyzed 1 to 7 days later by reverse transcriptase–polymerase chain reaction arrays. To determine the contribution of OPN to brain damage, we used p9 OPN−/− and wild-type mice. HI brain damage, sensorimotor function, and cell proliferation and differentiation were compared.
Results—Gene expression profiling of 150 genes related to growth factors and neurotrophins showed that expression of 52 genes changed during the first 7 days after HI. OPN was the gene with the strongest increase expression at all time points measured. We show here for the first time that in response to neonatal HI, OPN-deficient mice developed increased gray and white matter loss and more pronounced sensorimotor deficits as compared with wild-type littermates. Furthermore, OPN deficiency decreases HI-induced cell proliferation/survival and oligodendrogenesis without affecting neuronal differentiation.
Conclusions—OPN plays an important role in repairing brain injury after neonatal HI by regulating cerebral cell proliferation/survival and oligodendrocyte differentiation after injury. The observed promyelinative effect of OPN may offer novel possibilities for a therapy targeting white matter injury.
Endogenous repair mechanisms exist within the brain.1,2 In response to brain injury, neural stem cells that reside within the subventricular zone and in the subgranular zone of the dentate gyrus migrate toward the damaged area and can differentiate into several cell types. Indeed, formation of new neurons and oligodendrocytes and proliferation of astrocytes and microglia increases in response to neonatal cerebral hypoxia–ischemia (HI).1–3 These processes are regulated by a myriad of growth factors, neurotrophins, cytokines, and neurotransmitters.4,5
Understanding the changes in growth factor environment in the neonatal brain after exposure to HI and its contribution to damage and repair could help identify novel therapeutic targets. Here, we used growth factor reverse transcriptase–polymerase chain reaction array analysis of brain tissue of neonatal mice exposed to cerebral HI and identified osteopontin (OPN) as the most strongly induced gene. OPN, also known as secreted phosphoprotein-1 (Spp1) or early T-lymphocyte activation protein-1 (Eta-1), is synthesized by a variety of tissues and is present in all body fluids. It binds to integrins and isoforms of CD44 to regulate diverse processes, including cell survival, migration, differentiation, and inflammation.6 In several brain pathologies, OPN has been shown modulate injury and repair, including multiple sclerosis,7 Parkinson disease,8 brain tumors,9 and adult ischemic brain injury.10,11 We investigated the contribution of OPN to neonatal HI brain injury. We analyzed the effect of OPN deficiency in vivo on neonatal HI brain damage, sensorimotor function, and brain cell proliferation/survival as well as differentiation of newly formed cells into neurons, oligodendrocytes, astrocytes, and microglia.
Materials and Methods
Experiments were performed according to international guidelines and approved by the local experimental animal committee. We used OPN± mice (B6.129S6[Cg]-Spp1tm1Blh/J; Jackson Laboratory, Bar Harbor, ME) to breed OPN−/− and WT littermates that were used in HI experiments. Genotypes were assessed by genomic polymerase chain reaction analysis.
At p9 OPN−/− and WT littermates of both sexes underwent occlusion of the right common carotid artery under isoflurane anesthesia followed by hypoxia (10% O2) for 45 minutes. Sham controls underwent anesthesia and incision only. Pups from at least 3 different litters were used in each experimental group; both genders were equally distributed among experimental groups and data were obtained in at least 2 independent experiments. All analyses were performed in a blinded setup. Mortality rates (12.5%) did not differ between genotypes or gender and death only occurred immediately after HI.
Gene Expression Profiling
Pups were decapitated at 1, 2, 3, 4, 5, and 7 days post-HI or sham treatment. Total RNA was isolated using Rneasy kit (Qiagen) and cDNA was synthesized with Superscript Reverse Transcriptase (Invitrogen). Expression of 154 genes was analyzed in pooled cDNA of 5 to 10 brains per experimental condition using the growth factors (PAMM-041) and neurotrophin and receptor (PAMM-031; SABiosciences) RT2 profiler polymerase chain reaction arrays. Data were normalized using multiple housekeeping genes and analyzed by comparing 2−ΔCt of the normalized data. Data for OPN were confirmed by quantitative reverse transcriptase–polymerase chain reaction analysis on individual samples (primers: forward, 5′-AGCAAGAAACTCTTCCAAGCAA-3′, reverse, 5′-GTGAGATTCGTCAGATTCATCCG-3′) normalized for expression of glyceraldehyde-3-phosphate dehydrogenase and β-actin.12
Histology and Immunohistochemistry
Animals were euthanized with pentobarbital and perfused with 4% paraformaldehyde in phosphate-buffered saline. Coronal paraffin sections (6 μm) were incubated with mouse anti-microtubule associated protein 2 (MAP2), mouse-anti-NF200 (both Sigma-Aldrich), or mouse-anti-myelin basic protein (MBP; Sternberger Monoclonals) followed by biotin–horse–antimouse antibody and visualized with a Vectastain ABC kit (Vector Laboratories). Staining was outlined on full section images and the ratio of ipsi- and contralateral areas was calculated.3
For identification of OPN+ and BrdU+ cells, we used: rabbit-anti-OPN (IBL), biotinylated sheep-anti-BrdU (Abcam), rabbit-anti-Olig2, rabbit-anti-CNPase, mouse-anti neuronal nuclei (NeuN), mouse-anti-S100β (all Millipore), goat-anti doublecourtin (Dcx; Santa Cruz), or rabbit-anti-Iba1 (Wako) followed by Alexa-Fluor 488 or 594 conjugated secondary antibodies (Molecular Probes). Quantification of BrdU+ cells was done as described before.12 Sections were examined with a Zeiss Apotome microscope.
The cylinder rearing test was used to assess forelimb use asymmetry as described.3 The weightbearing forepaw(s) to contact the wall during a full rear were recorded during 3 minutes. Paw preference was calculated as ([nonimpaired−impaired]/[nonimpaired+impaired+ both forepaws])×100%.
Data are expressed as means±SEM and were analyzed using analysis of variance with Bonferroni post-tests.
Changes in Gene Expression of Growth Factors in Response to Neonatal Cerebral HI
P9 mice were exposed to unilateral cerebral HI and expression of 154 genes related to growth factors and neurotrophins was analyzed in the ipsilateral hemisphere from 24 to 168 hours after HI or sham treatment using polymerase chain reaction array analysis. With an arbitrary cutoff of >2-fold change, expression of 27 genes was upregulated, whereas expression of 18 genes was downregulated in response to HI (Supplemental Table I; http://stroke.ahajournals.org). An additional 7 genes were up- or downregulated dependent on the time of analysis. When the 27 upregulated genes were classified with respect to their biological function, 30% of the genes were involved in regulation of cell proliferation (GO: 0008283), 29% in immune responses (GO: 0006955) and 25% in cell death (GO: 0008219). Gene expression of OPN showed the highest increase at all time points measured. Therefore, we designed a study to further establish the contribution of OPN to neonatal HI brain damage. First, we confirmed the polymerase chain reaction array data by real-time reverse transcriptase–polymerase chain reaction analysis on individual samples (Figure 1A). OPN mRNA expression was induced only in the ipsilateral hemisphere already at 12 hours after HI. At 1 day after HI, expression was increased approximately 20-fold compared with sham-operated mice and the peak in OPN gene expression (112-fold increase versus sham) was reached at 5 days after HI. At 7 days after HI, OPN mRNA level had sharply dropped to the level present in sham animals.
Immunostaining for OPN at 3 days after HI showed a scattered staining pattern that was restricted to the ipsilateral hemisphere (Figure 1B). OPN-expressing cells were identified as microglia and astrocytes. In the more severely affected mice with cortical damage, also some cortical neurons were OPN-positive. No OPN expression was detected in OPN−/− mice at 3 days after HI.
Effect of OPN Deficiency on Neonatal HI Brain Damage and Functional Outcome
To determine the contribution of OPN to neonatal HI brain damage and endogenous repair processes, we used homozygous OPN-deficient mice (OPN−/− mice) and littermate WT controls. HI brain damage was induced in 9-day-old WT and OPN−/− mice and the extent of damage was analyzed. At 10 and 21 days after HI, myelin basic protein loss analyzed as a measure of white matter damage was significantly larger in OPN−/− than in WT mice (Figure 2A). The increased myelin basic protein loss in OPN-deficient mice was associated with increased loss of MAP2, which is a marker for neuronal dendrites. Furthermore, we also observed that the absence of OPN was associated with increased axonal loss as determined by decreased NF200 staining (Figure 2B–C). In line with previous reports, we did not observe gross cerebral abnormalities in sham-operated OPN−/− mice.13
HI-induced lateralizing motor deficits in OPN−/− and WT mice were analyzed using the cylinder rearing test. Sham-operated mice showed symmetrical use of both forepaws when rearing. At 10 and 21 days after HI, WT mice showed a preference for use of the nonimpaired (ipsilateral) paw. Importantly, in OPN−/− mice, this impairment of forepaw use was significantly larger than in WT mice (Figure 2D).
Effect of OPN Deficiency on Brain Cell Proliferation/Survival
To determine the effect of OPN deficiency on HI-induced cell proliferation and survival of newly formed cells in the brain, OPN−/− and WT mice were injected with BrdU at Days 3, 4, and 5 after HI. These time points were selected because the largest increase in OPN mRNA and protein expression in the ischemic hemisphere was observed during this timeframe (Figure 1).
In the hippocampus of WT animals, the number of BrdU+ cells was increased at 10 and 21 days after HI when compared with sham-operated animals. However, HI did not induce an increase in cell proliferation/survival in OPN−/− mice because the number of BrdU+ cells in the hippocampus was equal to that of sham-operated WT and OPN−/− mice at both time points analyzed (Figures 3A and 3C). In the cortex, the number of BrdU+ cells was significantly lower in OPN−/− mice when compared with WT or sham-operated mice after HI at both time points measured (Figures 3B and 3D). OPN deficiency had no effect on cell proliferation/survival in sham-operated animals, because the number of BrdU+ cells in the cortex and hippocampus did not differ between sham-operated OPN−/− and WT mice.
Effect of OPN Deficiency on Expression of Phenotypic Markers by BrdU+ Cells
We first determined the effect of OPN deficiency on differentiation into oligodendrocytes. Therefore, BrdU+ cells were analyzed for expression of the oligodendrocytes precursor markers, Olig2, NG2, and PDGFRα. The percentage of BrdU+/Olig2+ cells was significantly lower in the hippocampus of OPN−/− mice at 10 days after HI when compared with WT mice (Figure 4). At 21 days after HI, the percentage BrdU+/Olig2+ cells in the hippocampus did not differ between OPN−/− and WT mice. In the cortex, OPN deficiency decreased the percentage of BrdU+/Olig2+ cells at both time points measured. We observed similar effects of OPN deficiency on the percentage of BrdU+/NG+ and BrdU+/PDGFRα+ cells after HI.
The percentage of BrdU+/CNPase+ cells, that is, more mature oligodendendrocytes, in the hippocampus did not differ between OPN−/− and WT mice at 10 days after HI, whereas at 21 days after HI, the percentage of BrdU+/CNPase+ cells was significantly lower in OPN−/− mice compared with WT mice. In the cortex, the percentage of BrdU+/CNPase+ cells was lower in OPN−/− mice than in WT mice at both 10 and 21 days after HI.
OPN deficiency had no effect on differentiation of BrdU+ cells toward neurons in the hippocampus and cortex as no differences were detected in percentages of BrdU+/Dcx+ cells and BrdU+/NeuN+ cells between OPN−/− and WT mice at 10 and 21 days after HI (Figure 5; data not shown).
The percentage of BrdU+/S100β+ astrocytes in the hippocampus was not affected by OPN deficiency. However, in the cortex, the percentage of BrdU+/S100β+ cells was increased in OPN−/− mice at 21 days after HI.
Deletion of OPN also affected the percentage of BrdU+/Iba1+ cells in the hippocampus. In OPN−/− mice, the percentage of BrdU+ microglia was significantly higher than in WT mice at both 10 and 21 days after HI. In the cortex, there was no difference in the percentage of BrdU+/Iba1+ cells between genotypes.
In this study, we investigated changes in global cerebral gene expression in a murine model of neonatal HI brain damage. From 24 to 168 hours after HI, 52 genes were differentially expressed in the ischemic hemisphere. Comparing these results with the results of Hedtjarn et al reveals a similar effect of neonatal HI on the expression pattern of several genes, including OPN, IGF1, CXCL1, and S100A6.11 In our study, OPN was the highest upregulated gene at all time points tested. Furthermore, we show that in response to HI, OPN-deficient mice develop increased white and gray matter brain damage and more pronounced lateralizing motor defects than WT mice. Increased damage in OPN-deficient mice was associated with decreased new cell formation/survival and reduced oligodendrogenesis.
OPN was mainly expressed by microglia and astrocytes in the ischemic hemisphere, which is in line with previous reports.10,11 The time course of the upregulation of OPN expression after HI is remarkably similar to cell proliferation kinetics in the dentate gyrus after HI,3 suggesting that OPN may influence cell proliferation in the ischemic hemisphere. Indeed, we show here that OPN−/− mice do not develop the increased new cell formation in the dentate gyrus and cortex that occurs in response to neonatal HI in WT mice. These findings are in agreement with the reported decreased cell proliferation after hemorrhagic stroke in adult OPN−/− mice.14 Moreover, the transient induction of OPN after HI is in agreement with the changes in OPN gene expression that have been reported previously in ischemic brain injury.10
Interestingly, under baseline conditions, cell proliferation in the brain did not differ between OPN−/− and WT mice. This finding is in line with the observation that OPN−/− mice have a normal phenotype.15 Moreover, infusion of OPN into a noninjured adult rat brain does not increase cell proliferation in the dentate gyrus or subventricular zone.16 Together these data indicate that OPN contributes to brain cell proliferation after an injurious stimulus but is not required for normal brain development.
The most important finding in this study is that OPN deficiency led to a decrease in the percentage of recently divided oligodendrocyte precursors (Olig2+, NG2+, and PDGFRα+ cells) and newly formed mature oligodendrocytes (CNPase+ cells). OPN exerts its function by an interaction with several integrins and the predominant receptor for OPN is αv β3 integrin.6,15 It has been shown that binding of OPN to αv integrins on oligodendrocyte precursors cells induces proliferation as well as differentiation of these cells.17,18 In vitro, OPN has been shown to stimulate proliferation of oligodendrocyte cell lines and recombinant OPN has been shown to stimulate the synthesis of MBP and of myelin sheath formation in mixed cortical cultures.17 In vivo, OPN is upregulated during both demyelination and remyelination in the cuprizone model for remyelination.17 Therefore, it is conceivable that OPN might also support myelin formation after injury.
Interestingly, no differences were detected in the percentage BrdU+/Dcx+ cells and BrdU+/NeuN+ cells between both genotypes. This suggests that OPN deficiency does not affect differentiation of newly formed cells toward neurons.
In OPN-deficient mice, increased cerebral damage in an adult model of ischemic brain injury was associated with increased activation of microglia and macrophages.13 We show here that OPN deficiency increased the number of proliferating microglia in the hippocampus (12.4±1.3 BrdU+/Iba1+ cells in OPN−/− versus 8.6±0.4 in WT mice, P<0.05), which is the most severely damaged area in the murine neonatal HI model we used. Because microglial activation contributes to neuronal cell death after ischemic brain injury, OPN may also protect the brain against neuronal loss by downregulating neuroinflammation through decreasing the number of microglia in the brain.19,20
Ischemic brain injury in the absence of OPN resulted in increased gray and white matter loss, which was associated with increased sensorimotor impairment. The increased white matter injury in OPN-deficient mice might be related to diminished proliferation of oligodendrocyte precursor cells and decreased differentiation of these cells toward mature oligodendrocytes. Loss of myelin reduces synaptic transmission of long tracts,21 which may well be responsible for the observed increase in forepaw impairment in OPN-deficient HI mice.
The aggravated neuronal MAP2 and NF200 loss in OPN-deficient mice might be related to increased white matter loss in absence of OPN and/or increased microglia-mediated neuroinflammation and subsequent neuronal cell death. The latter hypothesis is in agreement with the literature because OPN has been shown to have neuroprotective effects by reducing neuronal ischemic cell death after adult stroke.22,23
In summary, our data indicate that OPN may contribute to repair processes after neonatal HI brain injury by influencing proliferation/survival of newly formed cells. Moreover, we show here that OPN may contribute to regulation of the differentiation of newly formed cells into oligodendrocytes. The possibility to use OPN to treat ischemic injury to the newborn brain where white matter is particularly sensitive to damage warrants further investigation.
Sources of Funding
This study was funded by grants from the Wilhelmina Children's Hospital Research fund and the European Union (LSHM-CT-2006-036534, NEOBRAIN).
The online-only Data Supplement is available at http://stroke.ahajournals.org/cgi/content/full/STROKEAHA.110.608315/DC1.
- Received November 10, 2010.
- Revision received February 28, 2011.
- Accepted March 1, 2011.
- © 2011 American Heart Association, Inc.
- Chabas D,
- Baranzini SE,
- Mitchell D,
- Bernard CC,
- Rittling SR,
- Denhardt DT,
- et al
- Wang X,
- Louden C,
- Yue TL,
- Ellison JA,
- Barone FC,
- Solleveld HA,
- et al
- van Velthoven CT,
- Kavelaars A,
- van Bel F,
- Heijnen CJ
- Schroeter M,
- Zickler P,
- Denhardt DT,
- Hartung HP,
- Jander S
- Ekdahl CT,
- Claasen JH,
- Bonde S,
- Kokaia Z,
- Lindvall O
- Yrjanheikki J,
- Keinanen R,
- Pellikka M,
- Hokfelt T,
- Koistinaho J