Intracerebral transplantation of erythropoietin‐producing fibroblasts facilitates neurogenesis and functional recovery in an ischemic stroke model

Abstract Introduction Erythropoietin (EPO) can enhance neurogenesis and fibroblasts can secrete growth factors; together, they may benefit ischemic stroke. We transplanted EPO‐producing fibroblasts into the rodent infarcted brain to test their effect on neurogenesis and functional recovery. Methods A total of 106 cells of EPO‐producing NIH/3T3 fibroblasts (EPO/EGFP/3T3) or enhanced green fluorescence protein (EGFP)‐expressing fibroblasts (EGFP/3T3) were stereotaxically injected into the infarcted striatum of adult rats that received transient middle cerebral artery occlusion (MCAO) surgery 1 day poststroke. On day 14 after MCAO, the animals were euthanized for the evaluation of neurogenesis via immunohistochemistry and of the expression of growth factors using enzyme‐linked immunosorbent assay. The infarct volume was analyzed using magnetic resonance imaging and the neurological behavior was assessed using the neurological severity scoring performed within 14 days after MCAO. Results The MCAO rats with EPO/EGFP/3T3 treatment showed high EPO expression in the infarcted brain for at least 1 week. The concentration of brain‐derived neurotrophic factor was higher in both hemispheres of MCAO rats with either EGFP/3T3 or EPO/EGFP/3T3 treatment at 14 days poststroke compared with untreated MCAO rats. The number of Ki‐67‐, nestin‐, or doublecortin‐immunoreactive cells in bilateral subventricular zones was higher in EPO/EGFP/3T3‐treated MCAO rats than it was in untreated MCAO control animals, indicating the enhancement of neurogenesis after EPO/EGFP/3T3 treatment. Notably, post‐MCAO EPO/EGFP/3T3 treatment significantly reduced infarct size and improved functional recovery. Conclusion The intracerebral transplantation of EPO‐producing fibroblasts benefited an ischemic stroke model probably via the enhancement of neurogenesis.


| INTRODUC TI ON
Stroke is one of the leading causes of mortality and physical/mental disability worldwide (Benjamin et al., 2017). Regarding ischemic stroke, the current standard treatments, which include thrombolytic therapy and endovascular thrombectomy, only benefit a small group of patients, and most patients who survive stroke suffer from longterm functional deficits (Jung et al., 2010;Sugawara & Chan, 2003).
Although poststroke neuroprotective therapy has been investigated for decades, unfortunately no treatment has shown obvious beneficial effects in clinical trials (Charidimou et al., 2017). The blood-brain barrier (BBB), which protects the brain from systemic toxicity, may prevent the penetration of drugs into brain tissues. Therefore, intracerebral delivery of certain treatments, particularly those with multiple therapeutic mechanisms, might provide an alternative direction for future stroke therapy.
Erythropoietin (EPO), a well-known hematopoietic cytokine, has various pleiotropic effects, such as the promotion of neovascularization, the mobilization of endothelial progenitor cells, and the induction of antiapoptotic and anti-inflammatory processes (Brines et al., 2000;Chong, Kang, & Maiese, 2003). Although preclinical studies have demonstrated that systemic EPO treatment facilitated stroke recovery in experimental stroke models (Gonzalez et al., 2013;Nguyen, Cherry, Scott, Ryou, & Mallet, 2014;Siren et al., 2001;Wang, Zhang, Wang, Zhang, & Chopp, 2004), clinical trials using systemic EPO administration did not consistently show effectiveness and safety in stroke patients (Yao et al., 2017). Systemic delivery of high-dose EPO is required to overcome its poor BBB penetration (Alnaeeli et al., 2012;Zhang et al., 2014) and achieve sufficient brain targeting; however, this approach may increase the risk of systemic thromboembolism (Kirkeby et al., 2008;Meng et al., 2011;Siren et al., 2001).
Fibroblasts are relatively resistant to hypoxic environments (Shinde & Frangogiannis, 2014) and secret several neurotrophic factors, such as the brain-derived neurotrophic factor (BDNF), the vascular endothelial growth factor (VEGF), and the nerve growth factor (NGF) (Dudas et al., 2011;Saito, Hamasaki, & Shibuya, 2003;Young et al., 1975). Therefore, in this study, we attempted to use a fibroblast cell line as a carrier and transplant EPO-producing fibroblasts directly into the infarcted brain of a rodent model of ischemic stroke. The aim of this study was to investigate the therapeutic effect of the intracerebral transplantation of EPO-producing fibroblasts on endogenous neurogenesis and poststroke functional recovery.

| MCAO and stereotaxic intracerebral transplantation
An ischemic stroke model with transient middle cerebral artery occlusion (MCAO) was used as described previously, with modifications (Tsai et al., 2011). Briefly, adult male Sprague Dawley rats (225-260 g) were anesthetized via intraperitoneal injection of Telazol (25 mg/kg) and Xylazine (10 mg/kg) and normal nonlabored breathing was maintained throughout the surgery. After exposure of the right carotid artery in a supine posture, an MCAO suture (MSRC37B280PK50, RWD Life Science, San Diego, CA) was inserted into the right common carotid artery, passed along the internal carotid artery, and brought to the orifice of the right middle cerebral artery (MCA) at a depth of 1.9 cm, to occlude the blood flow of the right MCA. After 1 hr of MCAO, the MCAO suture was withdrawn, to allow reperfusion of the right MCA. The animal was then placed back in its cage (day 0) and received stereotaxic intracerebral cell transplantation on the second day after MCAO (day 1).
For cell transplantation, the MCAO rats were anesthetized again and fixed in the stereotactic apparatus (Kopf Instruments, Tujunga, CA). A hole was drilled through the skull, for cell injection. A total of 10 6 cells in 8 μl of PBS were slowly injected into the infarcted brain at the right striatum (anteroposterior, 0.5 mm; mediolateral, 2.5 mm; dorsoventral, 5.0 mm from the bregma) using a Hamilton syringe (Hamilton Robotics, Reno, NV) for 5 min.
The cells were transplanted into the right striatum near the subventricular zone (SVZ), to trigger the stem cell niche in this zone.
The needle was then withdrawn and the incision was sutured.
In the 2 weeks that followed the procedure, the animals underwent a magnetic resonance imaging (MRI) study and behavioral assessment and were then sacrificed for histological and molecular analyses (Figure 1). The procedures were approved by the Institutional Animal Care and Use Committee (IACUC-20170124) of the National Taiwan University College of Medicine and the College of Public Health, Taiwan.

| Protein extraction and ELISA
On days 3, 7, and 14 after MCAO, the animals underwent transcardial PBS perfusion and the brain was removed. Bilateral striata were dissected using an anatomic microscope (Olympus CH-2; Olympus), homogenized in cold lysis buffer (8 M urea, 1% Triton X-100, 300 mM Na 2 PO 4 , 0.1% 2-mercaptoethanol, and protease inhibitors), incubated on ice for 15 min, sonicated, and centrifuged at 14,000 rpm at 4°C for 15 min. Supernatants were then collected and the protein concentration was measured based on the Bradford protein assay (Bio-Rad, Hercules, CA).
To obtain quantitative data for growth factors secreted in the striatum after MCAO, the following enzyme-linked immunosorbent

| Immunohistochemistry
On days 7 and 14 after MCAO, the animals were sacrificed by transcardial perfusion with 4% paraformaldehyde (PFA) and the brain was removed, cut into 2-mm-thick slices, postfixed in 4% PFA in PBS overnight, and cryoprotected in 30% sucrose solution for 7 days. Brains were then embedded in optimal cutting temperature compound (Leica, Wetzlar, Germany) and rapidly frozen in isobutane with dry ice. Frozen brains were sliced into 14-μm-thick coronal sections on a cryostat (Leica CM3050). Brain sections were fixed with methanol, permeated with 1% Triton-PBS for 30 min, blocked with 3% FBS, and incubated with primary antibodies at The resulting projection image was converted to grayscale, a similar threshold was set for all images, and the area of specific immunoreactivity was measured using LAS X (Leica) and image J.

| Magnetic resonance imaging
On day 1 (before cell transplantation) and day 14 after MCAO, rats were anesthetized with 2% isofluorane. The body core temperature NA, 4). T2-weighted images were processed and analyzed using the ImageJ system with the Bruker toolbox. The infarct volume is presented as the ratio of the infarct area over the whole brain area.

| Behavioral tests
The animals were subjected to a modified Neurological Severity

| Statistical analysis
All data were presented as mean values ± SEM and plotted using GraphPad Prism ® 7.0 (GraphPad, La Jolla, CA). We used one-way ANOVA followed by Tukey's post hoc comparison test for statistical analysis using the GraphPad software (GraphPad Prism version 7.00). Significance was set at p < 0.05.

| Implantation of EPO/EGFP/3T3 cells in the infarcted brain
To demonstrate that the EPO/EGFP/3T3 cells were successfully transplanted into the infarcted brain, we used an anti-GFP antibody We also used ELISA to analyze the EPO concentration after EPO/EGFP/3T3 transplantation in bilateral striata on days 3, 7, and 14 post-MCAO. In the infarcted striatum, the EPO concentration was higher on post-MCAO day 3 compared with days 7 and 14 (1,284 ± 42 vs. 599 ± 169 and 185 ± 83 pg/mg; p < 0.05) (Figure 2b).
In the contralateral striatum, the EPO concentration was not significantly different on post-MCAO days 3, 7, and 14. These results imply that the implanted EPO/EGFP/3T3 cells expressed EPO for ~7 days after transplantation.

| Increased intracerebral BDNF concentration after cell implantation
Previous studies have shown that 3T3 fibroblasts can secrete some growth factors, such as BDNF, VEGF, and NGF (Dudas et al., 2011;Saito et al., 2003;Young et al., 1975

| EPO/EGFP/3T3 treatment enhanced cell proliferation and differentiation in the SVZ
To investigate the cell proliferation capacity in the SVZ, we used immunohistochemistry to analyze the density of cells that were immunoreactive to Ki-67 (a proliferating marker) and nestin (Figure 4). We also analyzed the cells that were immunoreactive to DCX (a neuroblast marker) and GFAP (an astrocyte marker), to investigate the cell differentiation in bilateral SVZs ( Figure 5). MCAO rats that received EPO/EGFP/3T3 treatment showed a larger DCX-immunoreactive area in the SVZ than did EGFP/3T3-treated and untreated MCAO rats on the infarcted side (53.1% ± 8.9% vs. 40.6% ± 3.1% and 34.6% ± 6.2%, respectively; p < 0.05). On the contralateral side, EPO/EGFP/3T3 and EGFP/3T3 treatment increased the DCX-immunoreactive area compared with untreated MCAO rats (41.6% ± 11.2% and 42.5% ± 6.5%, respectively, vs. 25.0% ± 2.5%; p < 0.05). The GFAP-immunoreactive area in bilateral SVZs was similar among the three groups. These findings indicate that both EPO/ EGFP/3T3 and EGFP/3T3 treatments enhanced cell proliferation in the bilateral SVZs of MCAO rats. EPO/EGFP/3T3 treatment was also able to promote cell differentiation toward neuroblasts in bilateral SVZs.

| EPO/EGFP/3T3 treatment reduced infarct size and improved functional recovery
To investigate the effect of EPO/EGFP/3T3 treatment on infarct volume, we used MRI to analyze the infarct size on days 1 and 14 F I G U R E 3 The concentration of erythropoietin (EPO), brainderived neurotrophic factor (BDNF), vascular endothelial growth factor (VEGF), and bNGF in bilateral striata on post-middle cerebral artery occlusion (MCAO) day 14 was analyzed using enzyme-linked immunosorbent assay (ELISA). The concentration of BDNF was significantly increased in (a) the infarcted and (b) the contralateral sides of EGFP/3T3-and EPO/EGFP/3T3-treated rats compared with untreated MCAO control. The concentration of EPO, VEGF, and bNGF was similar among the groups (n = 3; *p < 0.05; **p < 0.01) post-MCAO in rats with or without EPO/EGFP/3T3 and EGFP/3T3 treatment ( Figure 6). On day 1 before treatment, the cerebral infarct was detected by both T2-weighted imaging (Figure 6a-c) and DWI (Figure 6d-f). The infarct size before treatment was similar among the three groups, as measured using T2-weighted imaging ( Figure 6m). On day 14, however, the residual infarct volume in EPO/ EGFP/3T3-treated rats (Figure 6g-i) was significantly lower than that in untreated rats (6.2% ± 6% vs. 16.5% ± 6%; p = 0.02) (Figure 6n).

| D ISCUSS I ON
Although EPO is a potential medication for acute ischemic stroke, its BBB penetration complicates the clinical usage of EPO via systemic delivery for stroke therapy. In this study, we directly trans- EPO is a well-known hematopoietic cytokine that has been widely used to treat patients with anemia (Subiros, Del Barco, & Coro-Antich, 2012). Recently, EPO has been shown to enhance endogenous neurogenesis and improve functional recovery in a rodent model of ischemic stroke (Cho et al., 2010;Gonzalez et al., 2013;Nguyen et al., 2014;Siren et al., 2001;Wang et al., 2004). (a-f) Coronal sections of the forebrain of MCAO rats without treatment (a, d) or with EGFP/3T3 (b, e) or EPO/EGFP/3T3 (c, f) treatment were immunostained with an anti-doublecortin (DCX) (green, neuroblasts) and an anti-glial fibrillary acidic protein (GFAP) (red, astrocytes) antibody, and counterstained with Hoechst 33342 (blue, nuclei). The density of DCX-or GFAP-positive cells was determined based on the immunoreactive area of DCX-positive or GFAP-positive cells/the area of the SVZ on the infarcted side (a-c) and on the contralateral side (d-f). The DCX-immunoreactive area was larger in EPO/EGFP/3T3-treated rats than it was in EGFP-3T3-treated or untreated rats on the infarcted side (g) and on the contralateral side (h). (i, j) The GFAP-immunoreactive area was similar among the groups. Scale bar, 10 μm (n = 4 for each group; *p < 0.05) study, after we singly transplanted EPO-producing fibroblasts into the infarcted striatum, a high EPO concentration was detected in the infarcted brain for at least 1 week, implying the secretion of EPO from fibroblasts at the early stage of stroke. In addition, treatment with EPO-producing fibroblasts was associated with a higher density of proliferating cells, NSPCs, or neuroblasts in the SVZ after cerebral F I G U R E 6 Assessment of infarct volume in middle cerebral artery occlusion (MCAO) rats using magnetic resonance imaging (MRI). Hyperintensity signals (cerebral infarct) were detected by T2-weighted imaging and diffusion-weighted imaging (DWI) on days 1 and 14 post-MCAO in rats without treatment (a-c) or with EGFP/3T3 (e-g) or EPO/EGFP/3T3 (i-k) treatment. (m) The infarct volume is presented as the infarct size/whole brain size in each group. There was no significant difference in infarct volume between the groups on post-MCAO day 1. (n) The residual infarct volume on post-MCAO day 14 was smaller in EPO/EGFP/3T3-treated rats than it was in untreated rats. (o) The ratio of the residual infarct volume over the initial infarct volume was lower in MCAO rats that received either EPO/EGFP/3T3 or EGFP/3T3 treatment compared with untreated MCAO rats (n = 3; *p < 0.05; ***p < 0.001) (a) (e) (i) infarct. These results were in accordance with previous reports that showed that poststroke EPO treatment enhanced not only the proliferation, but also the neuronal differentiation of NSPCs (Gonzalez et al., 2013). A previous study also showed that conditional EPO receptor knockdown reduced NSPC proliferation, migration, and neurogenesis in MCAO rats (Tsai et al., 2006). We also found numerous nestin-immunoreactive cells appearing around the transplanted cells at the injection route. This result may imply that EPO also attracts NSPCs that migrate toward the infarcted region at the injection site.
Genetically modified cells, such as immortalized cells, can offer a local sustainable intracerebral delivery of trophic substances, to rescue injured neurons (Jin, Fischer, Tessler, & Houle, 2002;Rossner et al., 1996). Previous studies have demonstrated that NIH-3T3 fibroblasts were less sensitive to contact inhibition and are less likely to induce tumorigenesis (Jainchill, Aaronson, & Todaro, 1969). These transplanted 3T3 fibroblasts were also able to survive for 2-8 weeks in the brain or spinal cord of rodent models of intracerebral hemorrhage, Parkinson's disease, and spinal cord injury, with various beneficial effects (Grandoso et al., 2007;Jeon, An, Kim, Park, & Lee, 2008;Jin et al., 2002;Liu, Himes, Murray, Tessler, & Fischer, 2002;Rossner et al., 1996). Thus, the 3T3 fibroblasts were carrier cells that were appropriate for intracerebral transplantation. However, we found that most of our implanted fibroblasts survived less than 14 days, and that the concentration of EPO was reduced to the baseline level on post-MCAO day 14, which may result from the harsh environment of the infarcted brain, in which the transplanted cells were surrounded by inflammatory cells and hypoxic conditions (Savitz, Dinsmore, Wechsler, Rosenbaum, & Caplan, 2004).
Fibroblasts secrete several neurotrophic factors, such as BDNF, VEGF, and NGF (Dudas et al., 2011;Saito et al., 2003;Young et al., 1975). In both EPO/EGFP/3T3-and EGFP/3T3-treated MCAO rats, we found that the concentration of BDNF, but not of VEGF or NGF, was elevated in the infarcted brain. BDNF is a neurotrophic factor that plays an important role in endogenous neurogenesis in physiological conditions and in the presence of neurological diseases (Wang et al., 2004). Although the post-MCAO functional improvement was better and the neuronal differentiation capac- F I G U R E 7 Behavior and body weight of middle cerebral artery occlusion (MCAO) rats. (a) The functional performance of MCAO rats was assessed using modified Neurological Severity Score (mNSS) evaluation from post-MCAO days 0-14. MCAO rats that received EPO/EGFP/3T3 or EGFP/3T3 treatment had better functional recovery than did the untreated rats (n = 8 in each group; *p < 0.05 between the EPO/EGFP/3T3 and untreated control groups; +p < 0.05 between the EGFP/3T3 and control groups; #p < 0.05 between the EPO/EGFP/3T3 and EGFP/3T3 groups). (b) The body weight was similar among the groups After single unilateral fibroblast implantation at the striatum on the infarcted side, intracerebral BDNF concentration was increased and NSPC proliferation was facilitated in the contralateral hemisphere. As there is no brain-CSF barrier and proteins are able to diffuse throughout the brain via the CSF (Chen et al., 2012), we postulated that the BDNF produced in the infarcted brain might diffuse toward the contralateral hemisphere and stimulate the proliferation of NSPCs at the contralateral SVZ.
Although transplantation of modified autologous fibroblasts may reduce the possibility of immune reaction following intracerebral injection, the time requiring cell harvest, cell expansion, and genetic modification is too long, impeding cell transplantation at the acute stage of stroke. Therefore, allogenic transplantation may be more appropriate for future application. While the transplanted allogenic fibroblasts were eliminated within 14 days after MCAO in our experiment, the transplanted cells have already released sufficient trophic factors for enhancement of neurogenesis and functional recovery in MCAO rats. As the result, transplantation of genetic modified allogenic fibroblasts may be tested at first in the future.

| CON CLUS ION
We demonstrated that intracerebral treatment with EPO-producing fibroblasts (as carriers) can reduce infarct size and improve poststroke functional recovery. These benefits may stem from the enhancement of NSPC proliferation and neuronal differentiation via the effect of EPO and BDNF secreted from fibroblasts. This study contributes to future research on EPO-related therapy in stroke and fibroblast-assisted cell therapy in neurological diseases.

ACK N OWLED G M ENTS
C.-L. C. gratefully acknowledges support from the Ministry of Science and Technology (MOST 105-2320-B-002-008-MY3) and the National Taiwan University Hospital, Taipei, Taiwan (grant number 107-S3740, 108-A147 and 108-S4110). We thank the staff of the imaging core at the First Core Labs, National Taiwan University College of Medicine, for technical assistance. We thank the 7T Animal MRI Core Lab of the Neurobiology and Cognitive Science Center, National Taiwan University for technical and faculty support.