Intravenous C16 and angiopoietin-1 improve the efficacy of placenta-derived mesenchymal stem cell therapy for EAE

The placenta has emerged as an attractive source of mesenchymal stem cells (MSCs) because of the absence of ethical issues, non-invasive access, and abundant yield. However, inflammatory cell invasion into grafts negatively impacts the survival and efficacy of transplanted cells. Previous studies have shown that synthetic C16 peptide can competitively block the transmigration of leukocytes into the central nerve system, while angiopoietin-1 (Ang-1) can inhibit inflammation-induced blood vessel leakage and inflammatory cell infiltration in rats with experimental allergic encephalomyelitis (EAE). In this study, we investigated the effects of intravenous administration of C16 and Ang-1 on the efficacy of placenta-derived MSC (PMSC) transplantation in a rat model of EAE. We found that, compared with PMSCs alone, treatment with PMSCs along with intravenously administered C16 and Ang-1 was more effective at ameliorating demyelination/neuronal loss and neurological dysfunction, reducing inflammatory cell infiltration, perivascular edema, and reactive astrogliosis (p < 0.05). Mechanistic studies revealed that intravenous C16 and Ang-1 increased PMSC engraftment in the central nervous system and promoted expression of the neurotropic proteins brain-derived neurotrophic factor, growth-associated protein 43, and p75 neurotrophin receptor as well as the neuronal-glial lineage markers neurofilament protein 200 and myelin basic protein in the engrafted PMSCs.

a rat model of EAE 14 . C16 is a synthetic peptide that selectively binds the αvβ3 and αvβ1 integrins expressed on endothelial cells, and this binding has been shown to inhibit inflammatory cell transmigration by blocking leukocyte-endothelial interaction 15 . Furthermore, C16 and Ang-1 have been reported to work synergistically to mitigate vascular leakage and inflammation and protect against demyelination and axonal loss in rats with EAE 14 .
In the present study, we examined the effects of intravenous C16 and Ang-1 on the efficacy of PMSC transplantation for treating EAE in a rat model. The neurological functions, CNS infiltration of inflammatory cells, perivascular edema, white matter demyelination, axonal loss, neuronal apoptosis, and reactive astrogliosis were evaluated. The homing of transplanted PMSCs to the CNS as well as the expression of the neurotrophic proteins brain-derived neurotrophic factor (BDNF), growth-associated protein 43 (GAP-43), p75 neurotrophin receptor (p75NTR) and the neuronal-glial lineage markers neurofilament protein 200 (NF-200) and myelin basic protein (MBP) in the engrafted PMSCs were examined.

Results
Intravenous C16 and Ang-1 enhanced the efficacy of PMSC therapy for preventing neurological dysfunctions in rats with EAE. Neurological dysfunctions in rats with EAE started 1 week post immunization (pi) (clinical scores > 2) and quickly progressed to the peak level (clinical scores ~3.7) by 2 weeks pi (Fig. 1A). After that, the rats underwent spontaneous recovery and the clinical scores returned to 2 by 8 weeks pi (Fig. 1A). Rats treated with PMSCs only also exhibited symptoms as early as 1 week pi, whereas those treated with PMSCs plus intravenous C16 and Ang-1 did not show neurological dysfunction until 2 weeks pi (Fig. 1A). The symptoms peaked at 3 weeks pi in both the PMSCs and P + C + A groups, which was 1 week later than in the vehicle-treated group (Fig. 1A). The clinical scores of rats in the PMSCs group were significantly lower than those in the vehicle-treated group from 2 to 8 weeks pi, and the scores were lower still in the P + C + A group from 1 to 3 weeks pi (the onset stage) and from 6 to 8 weeks pi (the recovery stage; Fig. 1A, P < 0.05).
Changes in cortical somatosensory evoked potential (CSEP) and cortical motor evoked potential (CMEP) have been commonly used to assess the level of neural damage in MS patients [16][17][18][19][20] . Compared with normal rats, rats in which EAE was induced displayed prolonged latency to waveform initiation and lower peak amplitudes for both CSEP and CMEP at 3 and 8 weeks pi (Table 1, Fig. S1), indicating a slower speed of conduction and loss of functioning nerve fibers, respectively, in these rats. These EAE-associated electrophysiological disturbances were significantly attenuated by treatment with PMSCs and PMSCs plus intravenous C16 and Ang-1 (Table 1, Fig. S1; P < 0.05).

Intravenous C16 and Ang-1 enhanced the efficacy of PMSC therapy for inhibiting inflammatory cell infiltration in rats with EAE.
Immunostaining of tissues collected 3 weeks pi showed diffuse CNS infiltration of CD68 + cells around blood vessels, throughout brain tissue and spinal cord parenchyma, and below the meninges (Fig. 1C,H-I,L,O). The perivascular and parenchymal infiltrates at 3 and 8 weeks pi were significantly reduced by PMSC transplantation (Fig. 1D,E,J,M,P) and further decreased by treatment with PMSCs plus intravenous C16 and Ang-1 (Fig. 1F,K,N,Q). These results were confirmed by inflammatory scores at 3 and 8 weeks pi (Fig. 1V,W; P < 0.05). CD68 was not detected in the engrafted PMSCs; however, significant CD68-positive cell infiltrates were detected in the areas surrounding the grafts and in the central region of the graft cell mass. Cell invasion into the grafts was less severe in the P + C + A group than in the PMSCs group ( Fig. 1R-T).
Astrocyte proliferation in response to neuronal injury in EAE (reactive astrogliosis) drives CNS inflammation and can lead to the formation of glial scars at the lesion site, which in turn inhibits axonal regeneration 21 . In this study, we assessed the expression and distribution of the specific astrocyte marker GFAP in the brain cortex and spinal cord by immunofluorescence staining. The results revealed astrocyte proliferation from 3 weeks pi (Fig. 4B,F), with noticeable glial scar formation at 8 weeks pi in rats with EAE ( Fig. 4I,L) but not in normal rats (Fig. 4A,E). Astrocyte proliferation was significantly inhibited by treatment with PMSCs alone or in combination with intravenous C16 and Ang-1 at both 3 ( Fig. 4C,D,G,H) and 8 weeks (Fig. 4J,K,M,N) PI. Moreover, P + C + A rats exhibited significantly reduced astrocyte proliferation compared with PMSCs rats at 8 weeks pi (Fig. 4S,T; P < 0.05). These results were confirmed by western blot (Fig. 4G-I) and RT-PCR (Fig. SP2) analysis of GFAP expression in the brain cortex.
GFAP was not detected in the engrafted PMSCs; however, astrocyte proliferation was observed in areas circumjacent to the grafts, which, in turn, obstructed the transmigration of PMSCs (Fig. 4O-R). Consequently, the P + C + A group exhibited more efficient PMSC homing to the CNS compared with the PMSCs group ( Fig. 4O  perivascular edema, and neuronal cell death in the CNS. In this study, demyelination was assessed by western blotting and immunofluorescence staining for MBP, a specific marker of myelination. In addition, morphological changes in the myelin sheath, neurons, and blood vessels in the CNS were examined by TEM. Western blotting revealed significantly decreased MBP expression in the brain cortex of rats with EAE at 3 and 8 weeks pi ( Fig. 5A-C, P < 0.05). Immunofluorescence staining for MBP showed myelin disruption and demyelination in the brain cortex and spinal cords of rats with EAE ( Fig. 6B,E,I,L). EAE-associated MBP loss and demyelination were ameliorated by treatment with PMSCs, either alone or in combination with intravenous C16 and Ang-1 ( Fig. 6A-U, I-III; P < 0.05), with the P + C + A group showing more effective prevention of MBP loss and demyelination than PMSC treatment alone ( Fig. 6I-III, P < 0.05). Engrafted PMSCs (green) were detected in the subarachnoid space, with infiltration into the parenchyma (Fig. 6V-Y). MBP, a marker of the oligodendrocyte lineage, was detected in the engrafted PMSCs in both the PMSCs and P + C + A groups, with relatively higher levels detected in the P + C + A group ( Fig. 6V-Y, I; P < 0.05). LFB staining for myelin revealed loss of myelin where inflammatory cell infiltration was observed in the brain cortex and spinal cords of rats with EAE, and these EAE-associated changes were reversed by treatment with PMSCs, alone or in combination with intravenous C16 and Ang-1 (Fig. S6).
TEM images of the brain cortex and spinal cord of normal rats showed neurons with an intact myelin sheath as well as well-defined nuclei (Fig. 7A,B). However, neurons in rats with EAE began to display noticeable myelin   (Figs 6V-Y and 8O-R) but not the astrocyte marker GFAP (Fig. 4O-R) in PMSCs homing to the CNS of rats with EAE. Intravenous C16 and Ang-1 increased the expression of NF-200 and MBP in these cells (Figs 6I and 8S), suggesting that the combinatorial treatment can promote differentiation of engrafted PMSCs along neuronalglial lineages. We next examined the expression of the neurotropic proteins GAP-43, p75NTR, and BDNF; the proinflammatory factors NF-κB and COX-2; and the apoptosis marker caspase-3 in engrafted PMSCs by immunofluorescence staining. NF-κB and COX-2 were not detected (Fig. 10A,C,E,G), but caspase-3 was detected at low levels at 8 weeks pi (Fig. 10I,K). GAP-43, p75NTR, and BDNF were all detected in the engrafted PMSCs (Fig. 10M,O,Q,S,U,W). Intravenous C16 and Ang-1 upregulated GAP-43, p75NTR, and BDNF ( Fig. 10II-IV, P < 0.05) and downregulated caspase-3 ( Fig. 10I, P < 0.05) in the engrafted cells.

Discussion
In a recent side-by-side comparison of PMSC versus EMSC therapy for EAE, PMSCs demonstrated similar efficacy to EMSCs at inhibiting inflammatory cell infiltration, demyelination, axonal loss, and neurological dysfunction 12 . Moreover, both PMSCs and EMSCs demonstrated the ability to migrate into the inflamed CNS tissues as well as the potential to differentiate along the neural-glial lineage 12 . These findings supported PMSCs as an ideal cell-based therapy for MS. However, significant inflammatory cell infiltration and inflammation of the CNS remained after PMSC transplantation, which limited the therapeutic potential of this treatment modality 12 . C16 and Ang-1 have been reported to block inflammatory cell migration by promoting vascular integrity and interfering with leukocyte-endothelial interactions, respectively [13][14][15] . In addition, because these two molecules act through different mechanisms, they have been shown to work synergistically to mitigate inflammatory responses  in EAE 14 . In the present study, intravenous administration of C16 and Ang-1 significantly increased the efficacy of PMSC therapy for reducing CNS inflammation, neuronal injury, and neurological dysfunction in a rat model of EAE.
Similar to our previous findings 12 , transplantation of PMSCs downregulated the proinflammatory molecules TNF-α, IFN-γ, IL-17, NF-κB, and COX-2 and upregulated the anti-inflammatory cytokine TGF-β. The perivascular/parenchymal infiltration of CD68 + microglia/macrophages was also attenuated by treatment with PMSCs. These anti-inflammatory effects of PMSCs were enhanced by intravenous administration of C16 and Ang-1.
Our mechanistic studies suggested that C16 and Ang-1 enhanced the efficacy of PMSC therapy by promoting homing of these cells to the inflamed tissues in the CNS as well as enhancing their differentiation/regenerative capacity. Reactive astrogliosis in rats with EAE, which obstructed PMSC transmigration, was attenuated by combinatorial treatment with C16 and Ang-1, leading to more effective PMSC engraftment in the CNS. Transplanted stem cells can produce neurotrophins to stimulate neuronal cell growth in animals with EAE. They can also differentiate along the neuronal-glial lineage to replace damaged neurons [27][28][29][30] . C16 and Ang-1 boosted the ability of the engrafted PMSCs to differentiate down the neuronal-glial lineage as assessed by the expression of the neuronal-glial markers NF-200 and MBP in these cells. Moreover, C16 and Ang-1 upregulated BDNF (a key player in neuron survival and differentiation 22,23 ), GAP-43 (a driving factor of axonal sprouting and restoration 24,25 ), and p75NTR (a regulator of neuron proliferation and maturation 31,32 ) and downregulated the apoptosis marker caspase-3 in the engrafted PMSCs. Invasion of inflammatory cells into MSC grafts is considered to negatively impact the survival and function of the grafted cells 12 . In this study, we detected infiltration of CD68 + microglia/macrophages into the PMSC grafts in the CNS, mainly localized in the central part of the cellular mass as described in our previous reports 12 . Combinatorial treatment with the C16 and Ang-1 markedly reduced inflammatory cell infiltration into the grafts, which, we believe, contributed to the improved retention, differentiation, and neuroprotective function of the transplanted PMSCs.
The injection of antigens into the anterior chamber of the eye induces a systemic suppression of cell-mediated and humoral immune responses to the antigen. This so called "anterior chamber associated-immune deviation (ACAID)" is largely mediated by TGF-β2 in the aqueous humor acting on ocular antigen-presenting cells (APCs), eventually leading to activation of antigen-specific T regulatory cells (Tregs) 33 . Previous studies have shown that intravenous injection of in vitro-generated APCs specific to the encephalitogenic antigens myelin oligodendrocyte glycoprotein (MOG) and/or MBP induces antigen-specific tolerance [33][34][35] . These encephalitogenic antigen-specific APCs might also augment the efficacy of PMSC and P + C + A therapy for EAE, especially in a MOG/MBP-induced chronic EAE model with pathological features typical of chronic, progressive MS 36 .
Granulocyte macrophage colony-stimulating factor (GM-CSF) has emerged as a putative therapeutic target in MS 37 . Macrophage infiltration is considered a major contributing factor to demyelination in both clinical MS and animal models of EAE. GM-CSF stimulates proliferation and activation of macrophages, monocytes, neutrophils, eosinophils, dendritic cells, and microglia with subsequent induction of pro-inflammatory mediators, and evidence suggests that this cytokine may be involved in the inflammatory processes related to MS [38][39][40] . Indeed, GM-CSF −/− mice are resistant to EAE and immune cell infiltration in the CNS 37 . However, GM-CSF has also been known to suppress autoimmune diseases such as Crohn's disease, type-1 diabetes, Myasthenia gravis, and experimental autoimmune thyroiditis by promoting Treg expansion and/or modulating phenotype-specific differentiation of precursor immune cells 41,42 . Thus, GM-CSF could function as a double-edged sword in EAE development. The effects of GM-CSF on PMSC and P + C + A therapy for EAE warrant further investigation.

Conclusion
Intravenous C16 and Ang-1 increased the efficacy of PMSC therapy for preventing demyelination/neuronal loss and ameliorating neurological dysfunction in a rat model of EAE by inhibiting inflammatory cell infiltration and enhancing PMSC engraftment, survival, differentiation, and neurotrophin production. Further pharmacokinetic and pharmacodynamic studies, preferably in primates, are needed to evaluate the therapeutic potential of this treatment regimen for MS.

Methods
Isolation and culture of GFP-expressing rat PMSCs. Green fluorescent protein (GFP)-expressing PMSCs were isolated from rats as previously described 12 . The cells were maintained at 37 °C in a humidified incubator with 5% CO 2 for 4-5 weeks prior to experimental studies. The medium was exchanged every 3-4 days. edema and leakage as well as extravasated inflammatory cells were still present (M). Some neurons exhibited signs of necrosis such as large vacuoles and degenerated organelles in the perikaryon, ruptured cytoplasmic membranes, and oncolytic chromatins (arrow in N). (O-T) PMSC-(O-Q) and P + C + A-(R-T) treated EAE rats at 8 weeks pi. Newly formed myelin sheaths were detected surrounding intact axons (arrow in O). The morphology of neuron nuclei became relatively normal, especially in the P + C + A-treated group. Perivascular edema and leakage were evidently alleviated. K, M, S, scale bar = 5 µm; B,D,E,F,G,I,L,N Animal model of EAE. A total of 77 adult male Lewis rats (9-10 weeks of age, 200-250 g) were obtained from the Center of Laboratory Animal Services at Zhejiang University. The rats were randomly assigned to four groups: normal control group (normal, n = 11), vehicle-treated EAE group (vehicle, n = 22), PMSC-treated EAE group (PMSCs, n = 22), and EAE group treated with PMSCs plus intravenous C16 and Ang-1 (P + C + A, n = 22). EAE was induced as previously described 12 by subcutaneous injection of guinea pig spinal cord homogenate (GPSCH) emulsified at a 1:1 ratio with complete Freund adjuvant (CFA) containing heat killed Mycobacterium tuberculosis. Rats in the normal group received an injection of CFA emulsified at a 1:1 ratio in 0.9% saline. Each rat received an intraperitoneal injection of 300 ng Pertussis toxin (Sigma-Aldrich, St. Louis, MO, USA) in 0.1 ml distilled water immediately after the subcutaneous injection and again 48 h later. Rats in the P + C + A group also received once daily intravenous administration of C16 (2 mg) and Ang-1 (400 µg) until the time of sacrifice, starting immediately after EAE induction. C16 and Ang-1 were purchased from Shanghai Science Peptide Biological Technology Co., Ltd. (Shanghai, China).
The clinical manifestations of EAE were assessed daily until the time of sacrifice. Disease severity was scored on a 5-point scale: 0 = no signs, 1 = partial loss of tail tonicity, 2 = loss of tail tonicity, 3 = unsteady gait and mild paralysis, 4 = hind limb paralysis and incontinence, and 5 = moribund or death 15 . Disease scoring was performed by pathologists blinded to treatment conditions.
All animal studies were approved by the Animal Ethics Committee of Zhejiang University and carried out in accordance with the National Institutes of Health (NIH) Guidelines for the Care and Use of Laboratory Animals. PMSC transplantation. One week after EAE induction, rats in the PMSCs and P + C + A groups received intrathecal infusion of 1 × 10 6 PMSCs into the subarachnoid space as previously described 12 . Rats in the vehicle group received injections of phosphate-buffered saline (PBS) instead.
Neurophysiological testing. Cortical somatosensory evoked potentials (CSEPs) were recorded as previously described 16-18 at 3 weeks (peak stage of clinical manifestations) and 8 weeks (recovery stage) pi (n = 5  per group). For measurement of CSEPs, rats were fixed to a stereotaxic frame. A constant current stimulator (Digitimer, Welwyn Garden City, UK) was used to deliver positive current pulses (15 V, 40 ms duration) to produce a maximum SEP (averaged over 30 stimuli). The SEPs from three series of stimulations were amplified, filtered, digitally converted, and stored for post-hoc analysis. Peak positive and negative values were recorded. The results are presented as the mean ± standard deviation (SD) of voltage amplitude (µV) and latency (ms).
Cortical motor evoked potentials (CMEPs) were recorded following previously reported procedures 19,20 at the same time points (n t 5 per group). Following anesthesia, a midline incision was made on the scalp. The tissues underneath were cleaned, and the cranium exposed. Screw electrodes were implanted to a depth of 0.75 mm over the primary somatomotor cortical areas, gently touching the dura mater. A needle electrode was inserted into the muscle of the hindlimb, and an inactive reference electrode was inserted under the skin, 2 mm from the screw electrode. The somatomotor cortex was stimulated with a train of 10-25 pulses at 10 Hz that evoked visible contralateral hindlimb movement. The signals were recorded and CEMPs were calculated over three independent experiments.
Tissue collection and processing. Rats in each group were sacrificed at 3 and 8 weeks after immunization (n = 5 per group at each time point). Rats were anesthetized with an intraperitoneal injection of 1% Nembutal (40 mg/kg) and perfused intracardially with cold saline followed by 4% paraformaldehyde in 0.1 M phosphate buffer (PBS, pH 7.4). The spinal cord and brain tissues were carefully dissected. One centimeter of the lumbar spinal cord and half of the brain of each animal were fixed in the same fixative for 4 h and then transferred to 30% sucrose in PBS until the tissue sunk to the bottom of the container. Twenty-micrometer-thick sections were cut on a freezing microtome through the coronal plane of the brain and transverse plane of the spinal cord using a Leica cryostat and then mounted onto 0.02% poly-L-lysine-coated slides. All sections were collected for histological assessment and immunohistological and immunofluorescent staining. The remains of the CNS tissue were fixed in 2.5% glutaraldehyde solution and examined by transmission electron microscopy (TEM).
Histological assessment. Extravasated macrophages were detected by immunofluorescence staining for CD68, a specific macrophage marker. In digital photomicrographs taken at 200x magnification in three fields per tissue section, the severity of inflammatory cell infiltration was scored from 0-4 43 : 0 = no infiltration, 1 = infiltration only around blood vessels and meninges, 2 = light infiltration in the parenchyma (1-10 cells per section), 3 = moderate infiltration in the parenchyma (11-100 cells per section), and 4 = severe infiltration in the parenchyma (100 + cells per section). Neuronal loss was assessed by Cresyl Violet staining (Nissl staining). Neuronal counts were restricted to cells that displayed a well-defined nucleolus as well as adequate amounts of endoplasmic reticulum.
The severity of axon demyelination was assessed by Luxol fast blue (LFB) staining and immunofluorescence staining for MBP. Demyelination was scored from 0-5 43 : 0 = no demyelination, 1 = rare foci of demyelination, 2 = light demyelination, 3 = confluent perivascular or subpial demyelination, 4 = substantial perivascular and subpial demyelination in at least one half of the spinal cord together with inflammatory cell infiltration in the CNS parenchyma, and 5 = massive perivascular and subpial demyelination across the entire spinal cord along with inflammatory cell infiltration in the CNS parenchyma.
Axonal loss was assessed by Bielschowsky silver staining and scored from 0-3 44 : 0 = no loss, 1 = superficial loss in less than 25% of tissue, 2 = deep loss in over 25% of tissue, and 3 = deep loss encompassing the entire tissue. TEM analysis. Sections of the brain cortex and lumbar spinal cord were examined by TEM as described previously [12][13][14][15] . Both low and high magnification images were recorded. The extracellular space surrounding the vessels were calculated with NIH Image software.
Enzyme-linked immunosorbent assay (ELISA). Peripheral blood samples were collected at 3 and 8 weeks pi (n = 5 per group at each time point). Concentrations of TNF-α, IL-17, and TGF-β were determined using ELISA kits from Abcam. Concentrations of IFN-γ were determined using an ELISA kit from BioLegend Inc. (San Diego, CA, USA). Optical density at 450 nm was recorded, and the data were tabulated using GraphPad Prism 4.0 (GraphPad Software, Inc., San Diego, CA, USA).

RT-PCR.
Rats were sacrificed by decapitation at 3 and 8 weeks pi (n = 5 per group at each time point). Total RNA was extracted from brain tissues using TRIzol reagents (Invitrogen, CA, USA) according to the manufacturer's instructions. cDNA was synthesized from 2 μl RNA using a cDNA Reverse Transcription Kit (Thermo Fisher Scientific, CA). The expression of NF-200, GAP-43, caspase-3, NF-κB p65, p75NTR, MBP, GFAP, COX-2 and BDNF was determined by PCR amplification followed by agarose gel electrophoresis. The results were normalized to those for GAPDH. All experiments were performed in triplicate.
Western blotting. Rats were sacrificed by decapitation at 3 and 8 weeks pi (n = 5 per group at each time point). Brain cortex tissues and 10-mm lumbar spinal cord segments were homogenized and analyzed by western blotting as previously described 12-15 . Statistical analysis. Data were analyzed using SPSS 13.0 software (SPSS, Inc., Chicago, IL, USA). Differences between groups were identified by two-way analysis of variance (ANOVA) followed by post-hoc Tukey t-tests. A P value of <0.05 was considered statistically significant. All statistical graphs were created using GraphPad Prism Version 4.0.