Peripheral Circulation and Astrocytes Contribute to the MSC-Mediated Increase in IGF-1 Levels in the Infarct Cortex in a dMCAO Rat Model

Materials and Methods Ischemic brain injury was induced by dMCAO in Sprague-Dawley rats. The transplantation group received MSC infusion 1 h after dMCAO. Expression of IGF-1 in GFAP+ astrocytes, Iba-1+ microglia/macrophages, CD3+ lymphocytes, Ly6C+ monocytes/macrophages, and neutrophil elastase (NE)+ neutrophils was examined to determine the contribution of these cells to the increase of IGF-1. ELISA was performed to examine IGF-1 levels in blood plasma at days 2, 4, and 7 after ischemia onset. Results In total, only 5-6% of Iba-1+ microglia were colabeled with IGF-1 in the infarct cortex, corpus callosum, and striatum at day 2 post-dMCAO. MSC transplantation did not lead to a higher proportion of Iba-1+ cells that coexpressed IGF-1. In the infarct cortex, all Iba-1+/IGF-1+ double-positive cells were also positive for CD68. In the infarct, corpus callosum, and striatum, the majority (50-80%) of GFAP+ cells were colabeled with ramified IGF-1 signals. The number of GFAP+/IGF-1+ cells was further increased following MSC treatment. In the infarct cortex, approximately 15% of IGF-1+ cells were double-positive for CD3. MSC treatment reduced the number of infiltrated CD3+/IGF-1+ cells by 70%. In the infarct, few Ly6C+ monocytes/macrophages or NE+ neutrophils expressed IGF-1, and MSC treatment did not induce a higher percentage of these cells that coexpressed IGF-1. The IGF-1 level in peripheral blood plasma was significantly higher in the MSC group than in the ischemia control group. Conclusion The MSC-mediated increase in IGF-1 levels in the infarct cortex mainly derives from two sources, astrocytes in brain and blood plasma in periphery. Manipulating the IGF-1 level in the peripheral circulation may lead to a higher level of IGF-1 in brain, which could be conducive to recovery at the early stage of dMCAO.


Introduction
Insulin-like growth factor-1 (IGF-1) is a member of the insulin gene family [1]. In addition to regulating cerebral development, neurogenesis, cognition, and memory function [2], IGF-1 is also an important player during the damage and recovery processes in ischemic stroke [3,4].
It has been widely recognized that neuroinflammation plays a critical role in brain injuries and neurodegeneration. The role of IGF-1 in the central nervous system (CNS) is, to a large extent, due to its ability to regulate immune cells in brain, such as microglia and infiltrated macrophages.
Microglia are important players in both innate immunity and adaptive immunity. The polarization of microglia is associated with the pathogenesis of a number of inflammatory disorders, such as the acute and chronic damage after stroke. Several in vitro studies revealed a direct antiinflammatory effect of IGF-1 on microglia [5,6]. Accumulating evidence suggests that IGF-1 may also modulate microglial phenotypes; for example, an increase in IGF-1 levels promotes the switch to the M2 phenotype [7]. Macrophages can also be regulated by IGF-1. In peripheral tissues, IGF-1 impacts macrophagic functions and leads to downregulation of proinflammatory cytokines and a change in disease progression [8,9].
Astrocytes can also produce IGF-1 and are positive for IGF-1 receptors [10,11]. Addition of IGF-1 to the in vitro culture of astrocytes promotes astrocyte growth and formation of glycogen [12]. Overexpression of IGF-1 by astrocytes through an AAV-mediated delivery improves outcome in a rat stroke model [13]. Astrocyte-derived IGF-1 can also protect neurons from kainic acid-(KA-) induced excitotoxicity in an astrocyte-neuron coculture system, and the rescue effect is abrogated by adding IGF-1R inhibitor [14].
By using ELISA in a previous study, we reported an increased level of IGF-1 in the ischemic core and periinfarct striatum in dMCAO rats at 48 h after intravenous (i.v.) infusion of rat bone marrow-derived MSCs [10]. MSC treatment leads to a higher level of IGF-1 compared to dMCAO rats without MSC infusion. By using immunostaining, we found that IGF-1 signals are mainly located in the infarct area. A minority of IGF-1 signals colocalize with NeuN+ neurons and CD68+-activated microglia in infarcts; nonetheless, quantitative analysis showed that these cells cannot account for all of the IGF-1-positive signals [15]. Other contributors in the brain and periphery (IGF-1 can cross the blood-brain barrier (BBB) [16][17][18]) to the increased IGF-1 signals in the brain warrant further investigation. In this study, we surveyed a wide spectrum of cell types that included Iba-1+ microglia, GFAP+ astrocytes, infiltrated immune cells such as CD3+ lymphocytes, neutrophil elastase (NE)+ neutrophils, and Ly6C+ monocytes/microphages, as well as the peripheral circulation, to determine their contribution to the increased IGF-1 level in the brain.

Materials and Methods
2.1. dMCAO Model. The performance of allogeneic bone marrow MSC (BMSC) culture, infusion, dMCAO model establishment, and behavioral tests have been described in our previous study [19].
In brief, primary cultures of bone marrow stromal cells were obtained from donor young adult male rats, and BMSCs were isolated as previously described [15,19]. Animals were anesthetized with 3.5% isoflurane and then maintained with 1.0-2.0% isoflurane in N 2 O : O 2 (2 : 1). One-hour postischemia, randomly selected rats received BMSC infusion. Approximately 1 × 10 6 BMSCs in 1 mL of vehicle (1 mL of saline) were slowly injected over a 5 min period into each rat via the tail vein.
In total, 100 wild-type Sprague-Dawley rats were purchased from Vital River Laboratory Animal Technology (Beijing, China). All animal protocols were in accordance with the Guidelines for the Care and Use of Experimental Animals and approved by the Institutional Animal Care and Use Committee of Capital Medical University. All experimental animals were housed in a specific pathogen-free rodent barrier facility at the Xuanwu Hospital Capital Medical University, on a 12 h light : 12 h darkness cycle with food and water ad libitum.
Ten Sprague-Dawley rats were used for harvesting bone marrow MSCs. Ninety Sprague-Dawley rats were divided into three groups: "sham", "ischemia," and "ischemia+MSCs" with 30 rats in each group. And for each group, 3 time points-days 2, 4, and 7 postischemia-were chosen with 10 rats used for each time point.

IGF-1 Measurement in Blood
Plasma. Blood was collected in heparinized tubes from the abdominal aorta. Subsequently, the blood samples were centrifuged at 1000×g for 10 min at room temperature, and the resulting plasma was obtained. The plasma was apportioned into 0.5 mL aliquots and stored at -80°C for further use.
IGF-1 levels were determined by ELISA using the R&D systems quantitative rat IGF-1 immunoassay kit (Minneapolis, MN, USA), which demonstrates high cross-reactivity with rat IGF-1. Plasma samples were diluted in the calibrator diluent provided in the kit at 1 : 10. The results were presented as ng/mL in undiluted plasma.

Brain Immunohistochemistry and Counting under
Confocal Microscopy. Immunohistochemistry and cell counting were performed as previously described [15,19,20].
Confocal images were acquired using a Leica TCS SP5 II AOBS laser scanning confocal microscope. The section thickness was 40 μm to minimize the possibility of cell body overlap in the z-axis. FITC or Cy3 fluorescence was acquired with an excitation wavelength of 633 nm and detection at 648-712 nm. Cy5 was detected with an excitation wavelength of 488 nm and detection at 501-542 nm.
2.4. Quantification and Statistical Analysis. As described in our previous study [15,19], the selection of sections for each rat and the demarcation of the counting area in the infarct cortex were in accordance with the work published by Gelosa et al. [21]. In brief, we outlined the counting area in the dorsal infarct cortex which was 2 mm adjacent to the boundary line between the normal and infarct areas 2 Stem Cells International (Supplementary Figure 1). The counting area in the striatum and corpus callosum was also demonstrated in the Supplementary Figure 1. IGF-1-positive cells in the brain were counted in 4 coronal sections from each rat (40 μm thickness, 480 μm interval, located between -2.0 mm and 2.0 mm to bregma); for each slide, 2 squares of images were captured under a microscope with a view field set as 800 × 800 μm (200x).
Eight fields of view were digitalized under a 20x objective, and the number of positive cells was summed for IGF-1positive cells that coexpressed Iba-1, GFAP, CD3, Ly6C, or neutrophil elastase (NE). The proportion of all IGF-1+ cells that were double-positive for each lineage marker was calculated and averaged.
Data were presented as the mean ± standard error of the mean (SEM). The comparisons were analyzed by one-way analysis of variance (one-way ANOVA) and Bonferroni-Dunnett corrections using SPSS 19.0. The level of significance of all comparisons was set at p < 0:05.  (Figure 1), and both activated and resting stage microglia are positive for Iba-1. In our previous studies, we reported that CD68+ microglia express IGF-1 in the brain of a stroke model [15,19]. In the present study, we continued to look into a wider range of microglia population-Iba-1+ cells, with regard to IGF-1 expression.

Results
By immunostaining, we found that IGF-1 signals were almost absent in the brain cortex in the sham group (data not shown). In the ischemia control group at day 2 after ischemia, Iba-1 and IGF-1 signals were mainly localized in the cortical infarct area, specifically the inner infarct border zone (Figures 1(a)-1(d)).
The quantities of Iba-1+/IGF-1+ double-positive cells in the brain infarct area, both before and after MSC infusion, were very low, indicating that Iba-1+ cells may not be the major source of IGF-1 in the ischemic cortex.
3.2. Iba-1+ Microglia/Macrophages Are Not the Main Source of IGF-1 in the Striatum and Corpus Callosum. In the sham group, no significant IGF-1+ signals were detected in the striatum or corpus callosum (data not shown).
Two days after ischemia onset, IGF-1+ signals were found scattered in the striatum and corpus callosum ( Taken together, Iba-1+/IGF-1+ cells constituted a small part of the IGF-1+ cells in the infarct area, striatum, and corpus callosum, suggesting that the Iba-1+ cell population is not the main source of IGF-1.

GFAP+ Astrocytes
Are the Main Cell Source of IGF-1 in Infarct Area. Previously, we found that in the ischemia brain cortex of dMCAO model, IGF-1 was partially expressed by NeuN+ neurons and CD68+ microglia/macrophages. On top of that, a significant number of other IGF-1-expressing cells existed, which included cells with a ramified morphology, reminiscent of GFAP+ astrocytes.
Due to the similarity in the morphology and distribution pattern of IGF-1+ signals and GFAP+ astrocytes in the ischemic hemisphere, double fluorescent immunostaining was employed to examine the spatial relationship of these two markers.
The above data suggested that GFAP+ astrocytes present in the infarct area of dMCAO rats were the main source of   (Figure 3(o)).

CD3+/IGF-1+ Cells Increase Significantly after Ischemia
and Are Decreased with MSC Treatment. In the brain cortex of the sham group, few CD3+ lymphocytes were detected (Figures 5(a)-5(d)). In the ischemia control group, a markedly  In the ischemia group, the percentage of CD3+/IGF-1+ cells among IGF-1+ cells was 26:46 ± 8:91%, which was

Limited Contribution from Ly6C+
Monocytes/ Macrophages to IGF-1 Expression. In our previous study, we reported that both Ly6C+ infiltrated monocytes/macrophages and brain-derived neurotrophic factor (BDNF) immunostaining are mainly located in the infarct boundary zone. The Ly6C+ cells are the main source of BDNF [19]. In the present study, we continued to investigate the contribution of infiltrated Ly6C+ cells to IGF-1 expression.
Taken together, these results indicated that only a small portion of Ly6C+ monocytes/macrophages were capable of expressing IGF-1, and Ly6C+ cells in this experimental setting were probably not the main cellular source of IGF-1 expression.

NE+ Neutrophils Do Not
Express IGF-1. Few NE+ neutrophils were detected in the brains of the sham group (data not shown). In the dMCAO model, NE+ neutrophils existed in both the ischemia control group (Figures 7(a)-7(d)) and the MSC treatment group (Figures 7(e)-7(h)). However, no NE+ neutrophils coexpressing IGF-1 were observed, suggesting that NE+ neutrophils were not the source of IGF-1 in this experimental setting.

Discussion
IGF-1 has been implicated as a potential neuroprotective agent in hypoxia-ischemia-induced damage [20,22]. Previously, by using ELISA on nonperfused brain tissues, we reported that IGF-1 concentrations increase in ischemia core and striatum in dMCAO model at 48 h after MSC infusion; yet levels of other neurotrophic factors, such as glial cell line-derived neurotrophic factor (GDNF) and vascular endothelial grown factor (VEGF) are not changed compared to those of the ischemia vehicle group at 2, 4, and 7 days after ischemia onset and MSC infusion [15].
In the previous study, we reported that IGF-1 signals are mainly (∽60%) located in the cortex infarct area. A smaller proportion (∽40%) of IGF-1 signals are detected in the striatum and corpus callosum. In the infarct area, IGF-1 signals are colocalized with NeuN and CD68, the markers of neurons and activated microglia/macrophages, respectively; but all of these cells together only contribute to a small part of the entire IGF-1+ signals [15]. These results strongly suggest that there are other cell types as the source(s) of IGF-1 in the ischemia ipsilateral hemisphere. In the present study, we examined additional possible sources in the brain and periphery and found that astrocytes and peripheral circulation may be the main contributors to the increased level of IGF-1.
Due to the autofluorescence problem caused by blood cells, brain section staining is usually performed on transcardially perfused animals. IGF-1 is a factor that can readily cross the BBB. The consequence of perfusion is that IGF-1 diffused in the CSF and extracellular matrix would have been washed away, and only intracellular IGF-1 could be detected by staining on brain sections. This caveat may lead to an underrepresentation of IGF-1+ signals on perfused brain sections. Upon brain injury, astrocytes are activated, which leads to phagocytosis, antigen processing and presentation, and the production of both neurotrophic and cytotoxic factors. Astrocytes are also a heterogeneous population, and reactions of astrocytes may depend on the nature of the activating stimulus and the microenvironment.
During ischemic injury, astrocytes were activated and upregulated expression of IGF-1 (Figures 3 and 4). MSC treatment induced astrocytes to further enhance IGF-1 expression without increasing the number of activated astrocytes (Figures 3 and 4). The molecular cues that promoted astrocytic expression of IGF-1 remained to be addressed.
Astrocytes are considered to play dual functions in neuroinflammation. On the one hand, astrocytes can produce  neurotrophic factors including nerve growth factor (NGF), BDNF, and IGF-1 [23]. On the other hand, activated astrocytes may release cytotoxic factors, such as reactive oxygen and nitrogen species (superoxide anion, nitric oxide) and toxic amino acids [24]. It was proposed by some researchers that astrocytes could be categorized into polarized A1 and alternatively A2 status [25]. Yet how MSC treatment may impact this A1 vs. A2 balance remains unknown and warrants further investigation. In the present study, a majority (50-80%) of GFAP+ astrocytes were IGF-1+. Upregulation of IGF-1 in the brain may further facilitate astrocytes to generate IGF-1. IGF-1  11 Stem Cells International gene therapy can reduce the reactivity of astrocytes in response to proinflammatory stimuli in vitro [26] and exert neuroprotective and neuroreparative actions in experimental animal models of hypoxia and stroke [27][28][29]. Our results suggest that IGF-1 may mediate a positive feedback loop in the regulation of astrocytes. Further studies are needed to investigate the detailed mechanisms underlying the interaction between astrocytes and IGF-1 in dMCAO models.
IGF-1 is able to cross the BBB; therefore, it is important to examine the peripheral levels of IGF-1 as well. The primary site of IGF-1 production in peripheral organs is the liver, but IGF-1 is not required for postnatal body growth in mice [30]. Blood-derived IGF-1 was increased by MSC treatment at days 2, 4, and day 7 in this study. The results suggest that MSC treatment-mediated changes of IGF-1 levels in blood may be an important mechanism leading to an increased level of IGF-1 in the brain. In some studies where MSC was infused through an intravenous route, no MSCs are observed in brain but the therapeutic and neuroprotective effects remain obvious, implying an indirect function through peripheral mediators [31].
It is unclear how MSC infusion induces upregulation of IGF-1 in the periphery and whether this increase of IGF-1 results from a direct effect of MSCs on hepatocytes. If manipulation of circulating IGF-1 levels at the acute phase of dMCAO can affect disease outcome, future therapeutic strategies may be contemplated in this direction for treatment of stroke.

Data Availability
The data used in the manuscript supporting the conclusions of the study can be accessed by email request.