Intravenous Administration of Hair Follicle Mesenchymal Stem Cells and Insulin Growth Factor-1 Improves the Neurological Performances of Rats Following Focal Cerebral Ischemia

Background: Focal cerebral ischemia is a common cerebrovascular disease with limited treatment options, and new treatments are therefore urgently needed. Hair follicle mesenchymal stem cells (HF-MSCs) are considered ideal cells for the treatment of neurological disorders. Insulin growth factor-1 (IGF-1) is an effective neuroprotective compound. Methods: In the present study, we used middle cerebral artery occlusion (MCAO) model to evaluate the therapeutic effects of HF-MSCs and IGF-1 in focal cerebral ischemia. After middle cerebral artery occlusion (MCAO), rats were randomly divided into six groups. HF-MSCs and IGF-1 were transplanted into rat models by tail vein injection. The fate of transplanted HF-MSCs in the rat brain was assessed using immunouorescence, immunohistochemistry, Western blot analysis, and reverse transcription polymerase chain reaction (RT-PCR). Beam balance tests and neurological severity scores were used to assess neurological recovery. Results: HF-MSCs labeled with the green uorescent dye PKH67 were found to colocalize with 4',6-diamidino-2-phenylindole (DAPI) and neural-specic markers. Rats in the HF-MSCs, IGF-1 and HF-MSCs + IGF-1 groups exhibited neural differentiation marker expression, with those in the HF-MSCs + IGF-1 group exhibiting the highest levels. Conclusions: These results suggest that the combined treatment of HF-MSCs and IGF-1 can enhance neurological recovery, representing a new therapeutic strategy for cerebral ischemia.


Introduction
Stroke is one of the leading causes of high disability and is associated with high mortality rates worldwide [1]. Ischemic stroke is secondary to occlusion of the arteries that supply blood to the brain and accounts for more than 80% of all strokes [2]. The majority of stroke patients have some degree of sequelae, and the disability rate is as high as 3 out of 4, which seriously affects the patients' quality of life [3]. Currently, stroke treatments are limited to anticoagulation [4], thrombolysis [5], antiplatelet [6], and intravascular therapies [7]. An increasing number of scientists and clinicians are focusing on stem cell therapy because of the ability of stem cells to differentiate and proliferate. To elucidate the role of stem cells in regenerative medicine, scientists have isolated and cultured stem cells from different tissues or organs, and a nonsurgical approach for the treatment of neurological diseases, including nerve defects, has been proposed [8]. Transplanted stem cells may secrete some neuroprotective or neuroreparative neurotrophic factors [9], however many problems must be resolved before these cells can be used in the clinic. As a regenerative organ, hair follicles have the characteristics of periodic growth and continuous self-renewal [10]. Hair follicle mesenchymal stem cells (HF-MSCs) can be harvested from hair follicles [11]. Due to the abundance of hair follicle tissue, HF-MSCs are easily obtainable, and the feasibility of autotransplantation and avoidance of immune rejection have been demonstrated [12]. Studies have shown that transplanted HF-MSCs participate in not only skin reconstruction but also angiogenesis and Page 4/23 nerve repair [13]. In addition, HF-MSCs have great potential for myoblast differentiation and can provide functional smooth muscle for tissue regeneration and cell therapy [14]. Since the neural potential of HF-MSCs was rst reported in 2004 [15], several studies have shown that transplanted HF-MSCs can differentiate in peripheral nerve injuries both in vitro and in vivo [16] including in spinal nerve injury [17][18][19] and Alzheimer's disease animal models [20]. Moreover, HF-MSCs can induce a rich population of glial and neuronal lineages after incubation with exogenous growth factors [21].
The polypeptide, insulin-like growth factor 1 (IGF-1) has neuroprotective effects in the central nervous system on the migration, growth and proliferation of cells. IGF-1 has been shown to be an effective neuroprotective compound in rat ischemic stroke models [22][23][24]. Studies have shown that injection of IGF-1 can reduce infarct volume, increase cell survival, and improve sensorimotor function [25] via the inhibition of the mRNA expression of iNOS and interleukin (IL)-1β in the ischemic hemisphere. It has been suggested that IGF-1 can induce neuroprotective effects by regulating neuroin ammation [26].
In the present study, we investigated whether intravenous administration of HF-MSCs can improve neurological function in a rat model of MCAO, and whether HF-MSCs and IGF-1 in combination are more effective than either is alone in improving neurological function in MCAO rats.

Experimental design
Thirty rats were randomly divided into 6 groups: healthy control group (n = 5), sham operation group (n = 5), MCAO group (n = 5), HF-MSCs group (n = 5), IGF-1 group (n = 5), and HF-MSCs + IGF-1 group (n = 5). Animals in the healthy control group were not treated, and those in the sham operation and MCAO groups were treated with 1 ml of saline. After the MCAO model was established (Fig. 1), the rats in the HF-MSCs transplantation group were injected with 1 ml of HF-MSCs (1×10 ) through the tail vein, and those in the IGF-1 transplantation group were injected with 1 ml of IGF-1 (R&D Systems, MI, USA, 50 µg/ml) via the tail vein. Both HF-MSCs (1×10 ) and IGF-1 (50 µg/ml) were injected into the tail veins of rats in the HF-MSCs + IGF-1 group.
On the 1st, 7th, 14th, 21st and 28th days after model establishment, the rats were assigned modi ed neurological severity scores (mNSS) and subjected to the beam-walking test. The samples were assayed for immuno uorescence on the 21st day, and immunohistochemistry analysis, western blot analysis and rt-PCR assays were performed on the 28th day, followed by statistical analysis (Fig. 1).

Isolation, identi cation, and in vitro uorescence labeling of HF-MSCs
The skin of the whiskers of 1-week-old male Sprague-Dawley (SD) rats was excised after being disinfected with alcohol and digested with 3 mg/ml collagenase type I. The hair follicle bulge area was then separated under an anatomical microscope (Olympus, Japan), and rat HF-MSCs were cultured by tissue block attachment in F12 DMEM (HyClone) containing 10% foetal bovine serum (FBS, ScienCell Research Laboratories). Upon reaching 80% con uence, the cells were digested, subcultured at 37℃ and 5% CO 2 and puri ed based on their adhesion to type I collagen, and cell density was adjusted to 1×10 cells/ml. The expression levels of stem cell surface markers, including FITC-CD29, PE-CD90, PerCP-CD45 and APC-CD31, were measured by a ow cytometry (BD Biosciences) [11,12,27], and HF-MSCs were labeled with the green uorescent dye PKH67 (SigmaAldrich, St. Louis, MO, USA). The labeled cells were observed under a uorescence microscope (Olympus, Japan). Cell isotype control experiments were performed simultaneously to identify false positives.

Induced differentiation of HF-MSCs
To test the multidirectional differentiation function of the HF-MSCs, passage 3 cells were seeded in a 6well culture plate. A lipogenic differentiation medium kit (Cyagen, RASMX-90031) and an osteogenic differentiation medium kit (Cyagen, RASMX-90021) for SD rat bone marrow mesenchymal stem cells (BMSCs) were used to induce osteogenesis and adipogenesis, respectively. After induction for 2 weeks to 1 month, the HF-MSCs were stained with Alizarin Red (Sigma-Aldrich) and Oil Red O (Sigma-Aldrich), and images were acquired by microscope (Olympus, Japan). Currently, there are no osteogenic or adipogenic media for inducing differentiation of HF-MSCs commercially available. As HF-MSCs and BMSCs are both mesenchymal stem cells, BMSC differentiation media were needed in this study.

Middle cerebral artery occlusion model
Adult male clean-grade SD rats weighing 240-260 g were obtained from the Animal Experiment Center of the Second A liated Hospital of Harbin Medical University. All relevant procedures were approved by the Ethics Committee of Harbin Medical University and were in compliance with National Institutes of Health (NIH) guidelines for the use and care of laboratory animals. After 24 hours of fasting and 2 hours of water deprivation, the rats were anesthetized with ketamine (60 mg/kg) and xylazine (10 mg/kg) and placed in the supine position for surgery, with their limbs and heads being xed. The proximal common carotid artery (CCA) and the external carotid artery (ECA) of each rat were ligated, and the internal carotid artery (ICA) was clamped with arteriolar forceps. A silicon-coated mono lament was slowly inserted into the ICA through the CCA. After loosening the arterial forceps, the mono lament nally reached the middle cerebral artery [28], and insertion of the mono lament was halted as soon as resistance was felt. During the surgery, the body temperature of the rat was maintained at 37 ± 1℃ with an electric blanket. After one hour, part of the mono lament was removed, but it was retained in the ICA. Finally, the incision was sutured, and the exposed portion of the mono lament was cut to prevent the rat from removing it on its own. Rats in the sham-operated group underwent only vascular dissection. The occurrence of hemiplegia was indicative of successful model establishment.
2,3,5-triphenyltetrazolium chloride (TTC) staining Ischemic brain injury was assessed by TTC staining at 24 hours after model establishment. The brain tissue was quickly removed from ice, ash frozen at -80℃ for 1 minute, sliced into an average of ve consecutive coronal sections and incubated in 2% TTC (St Louis, Missouri). A digital camera was used to image the brain tissue sections.
Hematoxylin-eosin (HE) staining HE staining can be used to distinguish normal cells from necrotic cells. Brain tissues were para nized and then cut into successive coronal sections. According to a standard protocol, the specimens were stained with HE, and observed under a light microscope.
Beam-walking test and modi ed neurological severity scores (mNSS) Motor coordination in rats can be evaluated by the beam-walking test and mNSS. For the beam-walking test, rats were placed on a high, narrow beam and allowed to walk, and hemiplegia and even falling were observed. The mNSS test was used to examine the behaviors of the model rats, which received scores ranging between 0 and 18 [29,30]. Rats with mild or severe symptoms and even those that died were excluded from the analysis. On days 1, 7, 14, 21 and 28 after the operation, the beam-walking test and mNSS test were performed to evaluate the improvement of motor function.

Immunohistochemistry
After heart perfusion, the brain tissue was embedded in para n and sectioned (10 µm) before being dewaxed and hydrated in xylene and ethanol solutions. The sections were heated at 95℃ for 10 minutes, antibodies at 4℃, and then reacted with the biotinylated secondary antibodies for 20 minutes at room temperature. After the sections were incubated in 0.05% diaminobenzidine (DAB) and PBS, the product of this reaction turned brown, at which point the sections were counterstained with hematoxylin. After dehydration in an ethanol solution, the sections were sealed on glass slides and analyzed under an optical microscope, and semiquantitative analysis was conducted with ImageJ (National Institutes of Health, Bethesda, USA).
Western blot Fresh brain tissue was taken, and the corresponding volume of lysate was added according to the quality and volume of each sample and then centrifuged at 12000 rpm and 4℃ for 10 minutes. The protein concentration was determined using a bicinchoninic acid (BCA) protein assay kit. The electrophoretic device was assembled, and the electrophoretic protein sample (every 40 µg) was transferred to a polyvinylidene uoride (PVDF) membrane, which was incubated with TBST buffer and 5% m/V skimmed milk powder under gentle shaking and then incubated overnight at 4℃ with anti-NeuN (1:500), anti-NF-H (1:1000), anti-DCX (1:500) and anti-GFAP (1:500) antibodies overnight. After the membrane was incubated with the secondary antibody (1:5000) coupled to horseradish peroxidase at 37℃ for 45 minutes, the substrates were detected with an ECL chemiluminescence detection kit. The lm was scanned, and the optical density of the target band was analyzed.

Real-time polymerase chain reaction (rt-PCR)
Total RNA was harvested from frozen brain tissue with the TRIzol (Invitrogen, Carlsbad, CA, USA) extraction kit in accordance following the manufacturer's instructions, and the RNA concentrations in each sample were determined using a NANO 2000 photometer. The RNA samples were reversetranscribed by PCR to obtain the corresponding cDNA, and β-actin served as an internal reference. Fluorescence quantitative analysis was performed on an ExicyclerTM 96 uorescence quantizer made by Bioneer (Korea) under the following conditions: pre-denaturation at 94 for 5 minutes, followed by 40 cycles of 94℃ for 10 seconds, and an annealing step at 60℃ for 20 seconds. The gene expression levels of neural-speci c markers in each group were compared using the 2 − ΔΔ CT method. All PCRs were repeated three times, and the existence of a single peak was veri ed by the melting curve. The primers are shown in Table I.

Statistical analysis
All data are expressed as the mean ± standard error of the mean (SEM) from three repeated experiments and were analyzed by Student's t-test, Bonferroni's multiple comparison test, a two-factor factorial design and ANOVA in a fully randomized manner. All statistical analyses were performed with SPSS 24.0 (SPSS Inc, Chicago, IL, USA) and Prism 8.0 (GraphPad), and differences between the means were considered statistically signi cant for P < 0.05.

Evaluation of the MCAO rat model
An experimental model was established with MCAO. Focal cerebral infarction in the temporal, parietal, frontal, occipital and striatum regions was induced by this model. TTC staining showed white ischemic brain tissue in the MCAO group (Fig. 2B), and red in the healthy control group (Fig. 2A). HE staining showed that noninfarcted brain tissue contained normal nerve cells with round or oval shapes (Fig. 2C), whereas a large number of nerve cells disappeared in the infarcted area, and the remaining neurons were sparsely distributed. In addition, we observed a substantial amount of in ammatory cell in ltration (Fig. 2D, E). Under normal conditions, the rats showed signi cantly different mNSS and beam-walking test performances, and rats of the acute cerebral ischemia model showed obvious signs of neurologic de cits.

Characteristics of HF-MSCs
After reaching approximately 80% con uence, HF-MSCs were passaged, and they began to exhibit a paving-stone shape after the third passage (Fig. 3A, B). Calcium precipitation was observed in HF-MSCs cultured in osteoblast differentiation medium for 1 month (Fig. 3D). After HF-MSCs were cultured in adipogenic differentiation medium for two weeks, lipid droplets were formed in HF-MSCs stained with Oil Red O (Fig. 3C). Fluorescence-activated cell sorting (FACS) analyses revealed that CD29 (Fig. 3E, I, J) and CD90 (Fig. 3F, I) were highly expressed in HF-MSCs, while the expression of endothelial cell surface marker CD31 (Fig. 3H, J, L) and hematopoietic cell surface markers CD45 (Fig. 3G, K, L) was lower.

Migration of transplanted HF-MSCs and expression of neural-speci c markers
The expression of four neural-speci c markers was examined, including NeuN, NF-H, DCX and GFAP, was examined to evaluate the therapeutic effect of HF-MSCs and IGF-1. The expression levels of the neuralspeci c markers in each group were analyzed and compared.
Immuno uorescence analysis revealed that the transplanted HF-MSCs exhibited green uorescence in the cell membrane (Fig. 4B, F, J and N), red uorescence in the cytoplasm (Fig. 4E, I and M) and blue uorescence in the nucleus (Fig. 4C, G, K and O) after proliferation and differentiation as determined by laser scanning confocal microscopy. After cosynthesis, HF-MSCs exhibited a yellow-green uorescence, and the nucleus was blue (Fig. 4H, L and P). Since NeuN is a nuclear protein (Fig. 4A), the nucleus showed both blue and red (Fig. 4D). We can clearly see the color of the cell membrane, cytoplasm and nucleus from the enlarged picture (Fig. 4d, h, l and p). These ndings indicated that HF-MSCs had homing abilities, allowing their migration to ischemic brain tissue areas and their differentiation into related nerve cells to promote the expression of neural-speci c proteins (Fig. 4). In addition, our data suggest that PKH67, as an ideal marker of living cells, can help track the implantation, proliferation and differentiation of transplanted HF-MSCs in brain tissue.
Immunohistochemistry analysis showed that the percentage of positive cells of NeuN, NF-H, DCX and GFAP of the infarct zone in the HF-MSCs, IGF-1 and HF-MSCs + IGF-1 groups was higher than that in the MCAO group. In addition, the percentage of positive cells in the HF-MSCs + IGF-1 group was signi cantly higher than that in the HF-MSCs and IGF-1 groups (Fig. 5). The expression of NeuN, DCX and NF-H in Western blot (Fig. 6) and rt-PCR (Fig. 7) analyses con rmed the above results at the protein and mRNA levels respectively. While GFAP expression was markedly different among the groups, as the highest GFAP expression was observed in the MCAO group, while the healthy control and sham operation groups exhibited the lowest GFAP expression.

Treatment enhances functional recovery
The beam-walking test and mNSS were used to evaluate the neural function of the rat MCAO model on days 1, 7, 14, 21 and 28 after surgery. After the surgery, the rats showed symptoms of neurological de cits on the rst day. The treated animals showed improved neurological function in the beam-walking test (Fig. 8B) and higher mNSSs (Fig. 8A) than rats in the MCAO group at 1 week. Rats in the HF-MSCs + IGF-1 group exhibited the greatest improvement in neural function. These data indicate that intravenous injection of HF-MSCs and IGF-1 signi cantly improves the neural function of rats after MCAO.

Disscussion
Acute cerebral ischemia is a kind of focal ischemic necrosis caused by disturbance of blood circulation in the brain, which can lead to different degrees of neurological impairment and sequelae. Stem cells have homing ability and can migrate to the border area of infarcted tissue and colonize, proliferate and differentiate. In the past, our team has studied the therapeutic effects of transplanted BMSCs [31] and dental pulp stem cells [32] on ischemic stroke models. However, the methods of obtaining stem cells are harmful to the individual, and HF-MSCs can improve this shortcoming, as they are easy to not only obtain but also expand, culture and cryopreserve. Moreover, no controversies are known in regards to immune rejection and mortality for HF-MSCs, and they could thus be considered an ideal source of stem cells for transplantation in neurological disorders. IGF-1 is involved in many central nervous system processes, including cell growth, proliferation, neurogenesis and migration, and has neuroprotective effects.
Treatment of the acute cerebral blood model with IGF-1 is associated with decreased expression of in ammatory factors in hemispheres. IGF-1 could induce microglial changes and reduce neuroin ammation [26]. Experimental studies show that transplanted HF-MSCs can migrate to the area around the infarcted tissue. In addition, HF-MSCs and IGF-1 transfuse intravenously, which is not only nonsurgical but also safe and effective.
The MCAO model was established, and TTC and HE staining showed the ischemic regions in the brain tissues of the model rats. In addition, successful rat models were characterized by neurological dysfunction, and rats in the sham operation and healthy control groups showed no differences. HF-MSCs is a type of MSC. CD29 and CD90 are mesenchymal stem cells indicating antigens. The two antigens were positive in ow cytometry, con rming that HF-MSCs belongs to MSC, which are cells with multipotential differentiation potential. Moreover, it has been proved that HF-MSCs can differentiate into osteogenic and adipogenic cells.
DCX plays an important role in the nal localization of new neurons, which are biomarkers of immature neurons [31,33,34]. NeuN is a common marker of mature peripheral and central nervous system neurons and is widely expressed in rats. Moreover, the NeuN protein was expressed only in nerve tissue and mainly in the nucleus [35]. NF-H is a mature neuronal marker [36]. GFAP is a hallmark intermediate lament protein mainly expressed in astrocytes [37,38]. GFAP increases dramatically in the periinfarct area after infarction, which may be associated with swelling of the neuroglia, increased exposure to postdisintegrating sites, and reactive astrocyte. Furthermore, the GFAP of infarct area was reduced due to necrosis [39][40][41][42]. This is the reason for the highest GFAP expression in the model group. The reason that GFAP expression trends differed among the three treatment groups may be due to a reduction in GFAP expression by intravenous injection of cells and factors that reduce swelling of the neuroglia and proliferation of reactive glial cells in the periinfarct area, the ability of stem cells to differentiate increases the expression of GFAP by increasing the neuroglia of the infarct zone. In WB and PCR our brain tissue samples included both infarct and periinfarct regions. So it was precisely because of the dissynchronization of GFAP expression in the infarcted area and periinfarct regions that there was no signi cant difference in GFAP expression between the HF-MSCs + IGF-1 group and the IGF-1 group. The HF-MSCs group mainly decreased the expression of GFAP in the periinfarct regions, but had little effect on increasing the expression in the infarct area, So its GFAP expression was minimal compared to the other two treatment groups. Although the amount of GFAP expression does not account for the e cacy of combination therapy or monotherapy, it does at least account for the e cacy of either therapy in comparison with the model group.
In this study, HF-MSCs labeled with green PKH67 dye were tracked in vivo and colocalized with neuralspeci c markers and DAPI. The transplanted HF-MSCs expressed the neural-speci c markers NF-H, DCX, NeuN and GFAP, which showed that they could differentiate into neuron-like cells. Semi-and quantitative data analysis revealed that the positive expression levels of NeuN, NF-H and DCX in the IGF-1, HF-MSCs and HF-MSCs + IGF-1 groups were much higher than those in the MCAO group. In addition, the expression levels of NeuN, NF-H and DCX were highest in the HF-MSCs + IGF-1 group. The nding suggest that IGF-1 can improve the migration and differentiation of HF-MSCs after transplantation, thus enhancing the neuroprotective effect. Simultaneous transplantation of IGF-1 and HF-MSCs has a synergistic protective effect in improving ischemic stroke, thus representing an effective method to restore neural function.
However, the limitation of this study was the short-term follow-up of HF-MSCs that survived in vivo transplantation. Therefore, further studies are needed to elucidate the molecular mechanism underlying the effects of HF-MSCs in combination with IGF-1 in the treatment of acute cerebral ischemia.
In summary, this study is the rst to demonstrate that intravenously transplanted HF-MSCs and IGF-1 can effectively improve the neurological function of an animal model of focal cerebral ischemia without signi cant side effects, despite lacking mechanisms to establish the therapeutic effect of stem cells. These results indicate that intravenous injection of HF-MSCs and IGF-1 may be a promising treatment for focal cerebral ischemia.  Experimental ow chart. HF-MSCs were isolated and cultured for about 21 days before the MCAO model was established. Rats were divided into six groups randomly. On the 1st, 7th, 14th and 21st days after the model established, the rats were given behavioral tests. The samples were assayed for immuno uorescence on the 21st day, and the immunohistochemistry, western blot and rt-PCR assays were performed on the 28th day prior to statistical analysis.   Immunohistochemical expression of transplanted HF-MSCs in the brain tissue. A-D) NeuN-, DCX-, GFAP-, and NF-H-immmunoreactive cells show neuronal morphology in the infarcted area. E-H) Semi-quantitative data. The percentage of positive cells in the IGF-1, HF-MSCs + IGF-1, and HF-MSCs groups was signi cantly higher than that in the MCAO groups. In addition, the percentage of positive cells in the HF-MSCs + IGF-1 group was signi cantly higher than that in the IGF-1 and HF-MSCs groups. (*P <0.05, **P 0.01, ***P 0.001).  Modi ed Neurological Severity Score (mNSS) and the Beam-Walking test. Compared with Healthy group and Sham group, the other groups all have neurological de cit symptoms. The score of mNSS in MCAO group was the highest, with the score in HF-MSCs + IGF-1 group was lower than that of IGF-1 group and HF-MSCs Group. While the score of the beam-walking test was highest in HF-MSCs + IGF-1 group, with lowest in MCAO group. (*P <0.05, **P 0.01, ***P 0.001).