Mesenchymal Stromal Cells Promote Axonal Outgrowth Alone and Synergistically with Astrocytes via tPA

We reported that mesenchymal stromal cells (MSCs) enhance neurological recovery from experimental stroke and increase tissue plasminogen activator (tPA) expression in astrocytes. Here, we investigate mechanisms by which tPA mediates MSC enhanced axonal outgrowth. Primary murine neurons and astrocytes were isolated from wild-type (WT) and tPA-knockout (KO) cortices of embryos. Mouse MSCs (WT) were purchased from Cognate Inc. Neurons (WT or KO) were seeded in soma side of Xona microfluidic chambers, and astrocytes (WT or KO) and/or MSCs in axon side. The chambers were cultured as usual (normoxia) or subjected to oxygen deprivation. Primary neurons (seeded in plates) were co-cultured with astrocytes and/or MSCs (in inserts) for Western blot. In chambers, WT axons grew significantly longer than KO axons and exogenous tPA enhanced axonal outgrowth. MSCs increased WT axonal outgrowth alone and synergistically with WT astrocytes at both normoxia and oxygen deprivation conditions. The synergistic effect was inhibited by U0126, an ERK inhibitor, and receptor associated protein (RAP), a low density lipoprotein receptor related protein 1 (LRP1) ligand antagonist. However, MSCs exerted neither individual nor synergistic effects on KO axonal outgrowth. Western blot showed that MSCs promoted astrocytic tPA expression and increased neuronal tPA alone and synergistically with astrocytes. Also, MSCs activated neuronal ERK alone and synergistically with astrocytes, which was inhibited by RAP. We conclude: (1) MSCs promote axonal outgrowth via neuronal tPA and synergistically with astrocytic tPA; (2) neuronal tPA is critical to observe the synergistic effect of MSC and astrocytes on axonal outgrowth; and (3) tPA mediates MSC treatment-induced axonal outgrowth through the LRP1 receptor and ERK.


Introduction
Stroke is one of the leading causes of death and disability worldwide. Currently, one of a few evidence-based acute stroke treatments is thrombolysis induced by intravenous administration of recombinant tissue plasminogen activator (tPA). Unfortunately, only a small percentage of patients benefit from this treatment primarily due to a narrow therapeutic time window of 4.5 hours [1][2][3].
Restorative therapy for stroke may provide a complementary and an alternative therapeutic approach [4][5][6][7], and functional recovery is key to ameliorate post-stroke deficits and improve life quality of stroke patients [8]. Among potential restorative treatments, exogenous cellbased therapies have been extensively studied, and multipotent mesenchymal stromal cell (MSC) has emerged as a promising therapeutic candidate [9][10][11][12][13]. We and others reported that exogenous administration of MSCs after experimental stroke facilitates neurite outgrowth, accelerates axonal sprouting and regeneration, enhances intercortical and intracortical axonal connections and improves neurologic recovery after stroke [14][15][16][17][18][19]. In vivo data revealed that reactive astrocytes promote brain plasticity and recovery from stroke, and astrocytes are involved in MSC mediated neurological recovery [20,21].
Astrocytes are a major constituent of the central nervous system, with versatile functions [22]. In the developing brain, astrocytes support and direct neurite extension through their synthesis of cell surface and extracellular matrix (ECM) molecules [13,23]. In the adult animals after stroke, axons may also acquire their potential for outgrowth from neighboring astrocytes and help establish contacts with existing circuits in the CNS [24]. MSCs stimulate neurotrophins and growth factors, including vascular endothelial growth factor (VEGF) [25,26], basic fibroblast growth factor (bFGF) [19,27] and brain derived neurotrophic factor (BDNF) [28,29] within reactive astrocytes in response to the ischemic brain environment [30,31].
In response to MSC treatment, white matter changes are mediated by astrocytes via increased tPA activity [18,32]. In vitro data suggest that the MSC induced activation of tPA in astrocytes promotes neurite outgrowth after ischemia [32,33], and MSCs significantly increase tPA expression and concomitantly decrease PAI-1 expression in astrocytes [33]. Therefore, exogenously administered MSCs may promote neurite remodeling in the CNS via astrocytic tPA and thereby improve neurological recovery. Recently, Mantuano et al reported that in PC12 and N2a neuron-like cells, tPA binds low density lipoprotein receptor related protein 1 (LRP1) and activates its downstream signals, including ERK in a ligand specific manner [34]. Here we test the hypothesis that MSCs stimulate tPA expression in astrocytes and activate neuronal LRP1 and ERK, which thereby enhances axonal outgrowth.

Materials and Methods
All experimental procedures were carried out in accordance with the NIH Guide for the Care and Use of Laboratory Animals and approved by the Institutional Animal Care and Use Committee of Henry Ford Hospital. Animals were maintained on a 12/12 hour Light/Dark cycle with food and water available ad libitum.

Isolation and culture of primary neurons and astrocytes
Wild type (WT, B57BL/6J) and tPA knockout (KO, with C57BL/6J background) mice (2-3 month-old, purchased from Jackson Laboratories, Bar Harbor, ME) were paired, respectively, for the first two hours of the light period and plug positive mice were separated. The day of plug detection was considered to be embryonic day 0 (E0). Cortical cells were dissected from day 17-18 (E17-18) mouse embryos according to our established procedure with some modifications [35,36]. Briefly, embryos were removed under deep Ketamine anesthesia, and the cerebral cortex dissected, stripped of meninges, and dissociated by a combination of Ca 2+ and Mg 2+ free Hanks balance salt solution (HBSS, Thermo Fisher Scientific Inc. Wayne, MI) containing 0.125% trypsin (Thermo Fisher Scientific Inc.) at 37˚C for 20 min, then mechanically triturated for~20 times. The triturated cells were passed through a 40 μm cell strainer (BD Falcon) and counted. For neuron isolation, cells were cultured in neurobasal growth medium (Thermo Fisher Scientific Inc.) containing 2% B-27 (Thermo Fisher Scientific Inc.), 2 mM GlutaMax (Thermo Fisher Scientific Inc.), and 1% antibiotic-antimycotic (Thermo Fisher Scientific Inc.) (Neurobasal/B27/Glu/AA) in a moist incubator at 37˚C/5% CO 2 . For astrocyte isolation, cortical cells were cultured and purified in high glucose Dulbecco's modified eagle medium (DMEM, Thermo Fisher Scientific Inc.) containing 20% FBS (Thermo Fisher Scientific Inc.), 2 mM glutamin and 1% antibioticantimycotic (DMEM/Glu/AA/20% FBS) in T-75 tissue culture flasks (Corning St. Louis, MO) in a moist incubator at 37˚C/5% CO 2 as reported [32]. Astrocytes at passage 1 (P1) were stored in liquid nitrogen and P2-3 astrocytes were used for studies. The growth media was changed every other day thereafter.

Primary MSC culture
Wild type mouse MSCs (P9) were purchased from Cognate Inc. (Fremont, CA), and cultured in Complete Stem Cell Medium (Stem Cell Technology, Vancouver, Canada) and P11-13 cells were used for co-culture experiment.

Oxygen deprivation of neurons
Three days after seeded in microfluidic chamber, neurons in Neurobal/B27/Glu/AA were cultured in an enclosed anaerobic chamber (Model 1025, Forma Scientific, Marietta OH) at 37˚C for 2 hrs. Then the neurons were exposed to normal culture conditions for the following experiments until sample collection [32,38]. This chamber maintains strict anaerobiosis to less than 10 μg/mL O 2 (according to the specifications provided by the manufacturer). The oxygen level within the chamber was routinely measured with a BD Disposable Anaerobic Indicator (Becton, Dickinson and Company, Sparks, MD), which confirmed that the oxygen level remained below 0.2% [32,38].

Axon Immunostaining and quantification
Five days after primary cortical cells were seeded, neurons in the microfluidic chamber were fixed by 4% paraformaldehyde, incubated with anti-Tuj-1 antibody (1:500, Covance Princeton, NJ) overnight and followed with Cy3 labeled secondary antibody. Nuclei were counter-stained by DAPI (1:10000, Thermo Fisher Scientific Inc). Axons were recognized by Tuj-1 positive fibers and quantified with ImageJ software 1.34 to show total axon length (μM) in a field. To rule out possible dendrite contamination in microgrooves, we selected cross-line between the ends of microgrooves and axonal compartment as the reference line and do binning through the entire axonal compartment. At least fifty randomly selected axon fields in more than 8 microfluidic chambers from 3 different experiments per group were quantified by experimenters blinded to each culture condition.

Statistics
Data are expressed as Means±SE. The differences between mean values were evaluated with the two tailed Student's t-test (for 2 groups) and the analysis of variance (ANOVA, for >2 groups). All calculations and statistical tests employed Microsoft Excel 2013 (Microsoft, Redmond, WA) or SPSS 11.5 (SPSS, Chicago, IL). P<0.05 was considered significant for all analyses.

Both endogenous and exogenous tPA promote axonal outgrowth
To investigate the role of neuronal endogenous tPA on axonal outgrowth, both WT and tPA-KO neurons were, respectively, cultured under normoxia and oxygen deprivation conditions in microfluidic chambers. As shown in Fig 3A, total axonal length from WT neurons was significantly longer than the length from KO neurons at both normoxia (6650±112 μM/field vs 6020±105 μM/field, n = 10-11/group, P<0.05) and oxygen deprivation (6020±119 μM/field vs 5390±168 μM/field, n = 10/group, P<0.05) conditions, suggesting that neuronal endogenous tPA contributes to axonal outgrowth even for hypoxic neurons. To test the effect of exogenous tPA, neurons were seeded and tPA (10, 30 and 100 nM) was added to the medium of Neurobasal/B27/AA/Glu medium in microfluidic chambers throughout the 5-day culture period. tPA increased total axonal length dosedependently (data not shown). As shown in Fig 3B, tPA (100 nM) increased total axonal length significantly from 11900±385 μM/field to 14560±476 μM/field (n = 5-6/group, P<0.05).

MSCs increase tPA expression in both astrocytes and neurons
To test whether tPA expression in astrocytes and neurons is altered by co-culture with MSCs, Western blot was carried out. MSCs and WT-astrocytes were co-cultured in dishes for 3 days and then Western blot was performed to detect intracellular tPA. As shown in Fig 4A, co-culture with MSCs significantly increased astrocytic tPA expression approximately 3 fold (n = 6/ group, P<0.05), in agreement with our previous publication that MSCs increase astrocytic tPA [32].
Next, WT-neurons were plated for 3 days and then co-cultured with MSCs and/or WT astrocytes seeded in cell culture inserts. Three days later, neurons were harvested and tPA was quantified by Western blot. As seen in Fig 4B,  and MSCs plus astrocytes increased tPA nearly 3 fold (P<0.05), indicating that MSCs increase neuronal tPA expression directly and synergistically with astrocytes.

MSCs activate neuronal ERK via tPA binding to LRP1
To investigate neuronal ERK activation following MSC treatment, WT neurons were cultured in plates for 3 days. Then, MSCs and/or WT astrocytes seeded in cell culture inserts were cocultured with the neurons in Neurobasal/B27/AA/Glu/2% FBS media and this media containing RAP (50 nM), respectively. Three days later, neurons were harvested, and p-ERK and ERK were quantified by Western blot. As shown in Fig 6, MSCs activated neuronal ERK 1 fold higher (n = 6/group, P<0.05), and MSCs plus WT astrocytes elevated p-ERK more than 2 fold (P<0.05). However, the combination of MSCs plus WT astrocytes mediated increase was inhibited by RAP (P<0.05), suggesting that MSC induced ERK activation is dependent on upstream tPA binding to LRP1 receptor.

Discussion
Axons play a central role in both the injury and repair phases after stroke [39]. The microfluidic chamber provides an excellent system for separating soma and dendrites from axons [35][36][37]. Axons in the chamber can be individually studied without interference from soma or dendrites. We are the first to co-culture MSCs, astrocytes and axons in a microfluidic chamber and investigate the effects of MSCs and astrocytes exclusively on axonal outgrowth. Our data demonstrate that MSCs promote WT axonal outgrowth alone and synergistically with WT astrocytes, but not with tPA-KO astrocytes ( Fig 1A); however, MSCs showed no effect on tPA-KO axonal outgrowth even with astrocyte (WT and KO) co-culture (Fig 1B), indicating that MSCs increase axonal outgrowth by both neuron and astrocyte tPA, and the expression of tPA in neurons is critical to observe the synergistic effect of MSCs and astrocytes on axonal outgrowth. When the same experiments were performed under oxygen deprivation conditions (an in vitro condition mimic to brain ischemia) [32,38,40], the MSC promoting effect on axonal outgrowth remained intact (Fig 2), suggesting that MSCs may provide benefit for patients of ischemic stroke via tPA. In addition, we compared axonal outgrowth from WT and tPA-KO neurons, and found that WT axons exhibited increased total axonal length compared to tPA-KO axons. This further confirmed that neuronal tPA contributes to axonal outgrowth ( Fig 3A). Then we cultured neurons with tPA in the medium and observed that exogenous non-neuronal tPA also promoted axonal outgrowth, consistent with the data of Fig  1, that astrocytic tPA mediates MSC-induced axonal outgrowth. Co-culture of MSCs with astrocytes significantly increased tPA expression in astrocytes compared with astrocyte cell culture alone (Fig 4A), which is in agreement with prior studies that MSCs increase tPA expression and secretion from astrocytes in vitro [32].
MSCs modulate endogenous tPA level and activity in the ischemic boundary zone in mice subjected to middle cerebral artery occlusion and tPA plays a pivotal role in neurite outgrowth [18,32]. The sonic hedgehog (Shh) pathway mediates the overexpression of tPA in neurons and astrocytes of ischemic boundary zone after MSC transplantation [41]. In the central nervous system, tPA is the major plasminogen activator and tPA activity is mainly inhibited by plasminogen activator inhibitor -1 (PAI-1) [42,43]. tPA promotes brain plasticity via its proteolytic and non-proteolytic pathways [44][45][46]. As the proteolytic function, tPA cleaves the precursor forms of neurotrophins, for example, pro-BDNF and pro-NGF, to the active forms of BDNF and NGF, respectively [47,48]. These active neurotrophins promote neurite remodeling [49][50][51][52][53]. Thus, it is reasonable to expect that the individually promoting effect of MSCs on axonal outgrowth may be mediated through tPA proteolytic function. tPA also acts by the non-proteolytic pathways. Recently reported, tPA combines with the LRP1 receptor in PC12 and N2a neuron-like cells to initiate downstream signaling of ERK in a biphasic manner [34]. However, combination of myelin-associated glycoprotein (MAG) with LRP1 did not activate the ERK pathway, suggesting a mechanism of tPA-ligand-specific coreceptor recruitment of LRP1 [34,54].
In the present study, we concentrated on non-proteolytic tPA pathways. To investigate the synergistic promoting effects of tPA on axonal outgrowth, the ERK inhibitor, U0126 and the LRP1 inhibitor, RAP, were included in the microfluidics chambers. As shown in Fig 5, MSCs enhanced axonal outgrowth alone. Co-culture of MSCs, astrocytes and axons synergistically further enhanced axonal outgrowth, which was inhibited by RAP and U0126, indicating that the synergistic effect was mediated by LRP1 receptor and was ERK-dependent. We have reported that MSCs increase tPA expression and secretion from astrocytes in vitro [32,33]. We therefore speculate, that MSCs enhance secretion of astrocytic tPA which is transported to axonal membranes, possibly via exosomes, and which combines with LRP1 resulting in neuronal ERK activation and axonal outgrowth [34,64].
Technically, it is difficult to test ERK activation in axons with microfluidic model, since axons were co-cultured with astrocytes and/or MSCs in axonal compartments. Instead, neuronal ERK activation was studied by Western blot. The Western data were not specifically axonal, but are consistent with our data of axonal growth and support our hypothesis. In Fig 6, MSCs activated neuronal ERK alone and synergistically with astrocytes. The activation was inhibited by RAP, indicating that the ERK activation is dependent on tPA binding to LRP1. Significant differences in ERK activation were detected between WT neurons and tPA-KO neurons before co-culture with other cells (Fig 6). Possibly, neuronal tPA from WT-neurons comes out of cytoplasm (e.g. exocytosis via exosomes) and then combines LRP1, resulting in activation of neuronal ERK [34,64,69,70]. Further investigation is warranted.

Conclusions
In summary, our data indicate that MSCs stimulate axonal outgrowth alone by neuronal tPA and synergistically with astrocytic tPA. Neuronal tPA is critical to observe the synergistic effect of MSC and astrocytes on axonal outgrowth. The increased tPA expression from astrocytes (and neurons) may bind to the neuronal membrane LRP1 receptor, which activates downstream ERK and thereby promotes axonal outgrowth.