Adipose-derived Stem Cell Conditioned Media Extends Survival time of a mouse model of Amyotrophic Lateral Sclerosis

Adipose stromal cells (ASC) secrete various trophic factors that assist in the protection of neurons in a variety of neuronal death models. In this study, we tested the effects of human ASC conditional medium (ASC-CM) in human amyotrophic lateral sclerosis (ALS) transgenic mouse model expressing mutant superoxide dismutase (SOD1G93A). Treating symptomatic SOD1G93A mice with ASC-CM significantly increased post-onset survival time and lifespan. Moreover, SOD1G93A mice given ASC-CM treatment showed high motor neuron counts, less activation of microglia and astrocytes at an early symptomatic stage in the spinal cords under immunohistochemical analysis. SOD1G93A mice treated with ASC-CM for 7 days showed reduced levels of phosphorylated p38 (pp38) in the spinal cord, a mitogen-activated protein kinase that is involved in both inflammation and neuronal death. Additionally, the levels of α-II spectrin in spinal cords were also inhibited in SOD1G93A mice treated with ASC-CM for 3 days. Interestingly, nerve growth factor (NGF), a neurotrophic factor found in ASC-CM, played a significant role in the protection of neurodegeneration inSOD1G93A mouse. These results indicate that ASC-CM has the potential to develop into a novel and effective therapeutic treatment for ALS.

microglia activation 14 . These pathogenic hallmarks are thought to play key roles in motor neuron death and ALS progression.
It has been reported that p38 mitogen activated protein kinase (MAPK) are involved in the motor neuron death [15][16][17] . Phosphorylated p38 MAPK was increased in the spinal cords of SOD1 G93A mice 15,16,18 and human ALS patients 19 . Moreover, the inhibition of p38 MAPK pathway extended motor neuron survival and also decreased the activation of microglial in ALS mouse spinal cord 17 , this suggests that the pathogenic aspect of p38 MAPK pathway play a significant role in ALS motor neuron cell death and neuroinflammation.
Activity of the Ca 2+ -activated protease, calpain, has also been observed to be increased in the SOD1 G93A mouse 20,21 due to increased levels of calcium in the cytosol caused by excitotoxicity 22,23 . Activated calpain cleaves cytoskeletal proteins, such as α -II spectrin and results in the accumulation of α -II spectrin and formation of inclusions located in motor neurons 20 . Inhibition of calpain via expression of its endogenous inhibitor, calpastatin, slowed degeneration of SOD1 G93A motor neurons 23 .
Since multiple pathways for neuronal death are involved in the development and maintenance of this disease, an effective multi-target approach may be required to treat ALS 10 . Additionally, ALS treatment can only be given after disease onset, because there are no specific predictors, clinical diagnosis, or biomarkers 4 . Lots of pharmacologic therapies used in both ALS clinical trials and stringent testing have been unsuccessful 24 . These failures at pre-clinical or clinical stage may attribute to the treatment's inability to target more than one neuronal death pathway, resulting in an incomplete or ineffective blocking of motor neuron cell loss and ALS disease progression.
Adipose-derived stem cell conditioned media (ASC-CM), a biologically-derived reagent containing a multitude of neuroprotective and neurotrophic factors, such as brain-derived neurotropic factor (BDNF), nerve growth factor (NGF), vascular endothelial growth factor (VEGF), hepatocyte growth factor (HGF), and insulin-like growth factor-1 (IGF-1) [25][26][27][28] was selected as ASC-CM has been previously shown to be neuroprotective by using both animal and cell culture models of neurodegeneration 27,29,30 . Our previous studies show that ASC-CM protected against hypoxia-ischemia-induced damage by blocking the activation of p38 MAPK 27 . NGF is a secreted growth factor in the central and peripheral nervous system. It is important in survival, growth and maintenance of specific types of neurons. The neurodegenerative disorders were due to the lack of hormones or growth factors 31 . NGF level was decreased in ALS dorsal spinal cord 32 . In this study, we evaluated the effectiveness of ASC-CM and investigated the role of NGF, a neurotrophic factor identified in the ASC-CM, in ASC-CM treatment for SOD1 G93A mouse.

Materials and Methods
Animals. SOD1 G93A heterozygous mice that overexpressing human SOD1-G93A mutation and wild type mice (WT) (all with B6SJL genetic background) were purchased form the Jackson Laboratory (Bar Harbor, ME, USA) and bred in the Animal Center of Indiana University School of Medicine. All animal procedures were performed in accordance with the protocols approved and authorized by the Institutional Animal Care and Use Committee at Indiana University School of Medicine. SOD1 G93A mice were identified by PCR performance with DNA derived from tail tissue using a protocol provided by The Jackson Laboratory 33 .
Animal behavioral Assessment. As described in a previous paper 33 , behavioral assessment was performed beginning at 90 days of age. SOD1 G93A mice were randomly assigned to 5 different treatments of "ASC-CM", "ASC-CM-NGF Ab", "NGF Ab", "NGF" or "vehicle". SOD1 G93A mice were tested twice a week on a Rotarod apparatus at a speed of 15 rpm (ENV-575M; Med Associates, Inc., St. Albans, VT, USA) and up to 3 trials per day. Mice that were unable to remain on the Rotarod for 10 minutes were determined as disease onset, this performance was further tested the following day under same parameters to further verify disease onset. Animal that could not right itself in 20 seconds when gently rolled on its side indicated the end stage (surrogate death time point). Mice were checked every morning for mortality and morbidity, and every afternoon for the righting reaction.

Poteomic analysis of the presence of neurotrophic factors in human ACS-CM. Human ASC
secrete over 400 proteins into the medium during culture 34 . We detected interesting factors in ASC-CM using an antibody array (RayBio Human Growth Factor Array I, RayBiotech, GA) Quantification of NGF by ELISA. NGF level in ASC-CM was measured using Human NGF ELISA kits (Abcam, Cambridge, MA, USA), an enzyme-linked immunosorbent assay (ELISA) kit, according to the manufacturer's instructions 35 . Animal Treatment. Human adipose-derived stem cell conditioned media (ASC-CM) was made as previously described 29,36 and injected intraperitoneally (i.p.) at a volume of 200 μ l once daily. To neutralize the activity of NGF, a polyclonal anti-NGF antibody (Abcam, Cambridge, MA, USA) was introduced and tested for its specificity. No cross-reactivity has been detected after using it against VEGF, GDNF, IGF1 and BDNF.1 μ g/μ l of the antibody was incubated with ASC-CM at 5 μ g/ml overnight at 4 °C 28,37 . NGF antibody neutralized ASC-CM treated mice group (ASC-CM-NGF Ab) received 200 ul ASC-CM-NGF antibody once daily. Additionally, as controls, the NGF Ab group daily received 5 μ g/ml NGF antibody in 200 μ l Basal Media Eagle (BME), the NGF group daily had 200 pg/ml human NGF (R&D systems, Minneapolis, MN, USA) in 200 μ l BME, and the vehicle group mice were injected with 200 μ l BME. One daily dose of ASC-CM, NGF antibody neutralized ASC-CM, NGF antibody, NGF or BME was given until a humane death endpoint for survival studies, and 3 or 7 days after onset for biochemical and immunohistochemical studies.

Immunohistochemistry.
Mice were anesthetized and perfusion-fixed with 4% paraformaldehyde (PFA) after treatments with ASC-CM or BME. PFA-fixed and paraffin embedded spinal cords were sectioned (15 μ m) serially via the lumber enlargement and immunostained with anti-microtubule-associated protein 2 antibody (1:1000, MAP2; Millipore), then followed by a biotinylated anti-mouse IgG antibody (Sigma-Aldrich Corp.). Sections were visualized under the microscope after applying 3,3′ -diaminobenzidine (DAB) substrate solution. A 40x objective light microscope (Nikon, Japan) was used to count large motor neurons in the ventral horn of the lumbar spinal cord 39,40 . 6-8 sections of each lumbar were counted. Data was reported as number of motor neurons per section as previously described 33,41 . Statistical Analysis. One-way analysis of variance (ANOVA) was used for statistical analyses and comparisons for the differences between groups. All data are expressed as mean ± standard error of the mean (SEM). Differences between two means were considered significant when p was equal or less than 0.05.

Results
Neurotrophic factors in ASC-CM. We tested interesting neurotrophic factors in ASC-CM by using a Human Growth Factor Antibody Array (containing 41 growth factor antibodies). Image showed the presence of NGF, VEGF, GDNF, IGF-1 in human ASC-CM. BDNF was detected in human ASC-CM by using enzyme-linked immunosorbent assay (ELISA) kit (Chemicon, Temecula, CA, USA) 27 .

ASC-CM increased the number of motor neurons in the lumbar spinal cord.
A positive correlation was observed between extended survival time of ASC-CM treated SOD1 G93A mice and the prevention of motor neuron loss. Changes in motor neuron number at 7 days following onset were analyzed in cross-sections of the lumbar spinal cords of SOD1 G93A mice ( Fig. 2A). BME-treated control mice showed a marked (***p < 0.005) loss of motor neurons in the ventral horn of the lumbar spinal cord versus WT control (7.8 ± 0.84 and 21.9 ± 0.79, respectively; Fig. 2B), but this loss was inhibited by ASC-CM treatment (15.9 ± 2.0, p < 0.01 vs. vehicle). Although there was still significant motor neuron loss (p < 0.05) compared to age-matched and wild type mice, this loss was not as severe in the vehicle group (Fig. 2B).
Administration of ASC-CM had no effect on mouse SOD1 wild type and mutant human SOD1 G93A expression in SOD1 G93A mice spinal cords. Studies have shown that the transgene copy number (copies of the transgene with the G → A at position 93, human mutant form of SOD1) affects survival of the SOD1 G93A mouse 9 . In order to determine if ASC-CM has an effect on SOD1 gene expression, immunoblot analyses were performed on spinal cord homogenates after 7 days of ASC-CM treatment  With ASC-CM treatment, a higher number of motor neurons was observed in SOD1 G93A mice with disease onset (n = 5) versus vehicle (n = 5). Wild type control data were generated from 4 mice (B). Experimental group data are presented as averages (± SEM) and statistical analyses were performed using one-way ANOVA. *p < 0.05; ***p < 0.005. (Fig. 3A). Band density measurements showed that levels of both wild type endogenous mouse SOD1 (mSOD1 WT ) and mutated human SOD1 (hSOD1 G93A ) exhibited no differences with or without 7-day ASC-CM treatment (Fig. 3B). "Wild type/WT" mice express wild type endogenous mSOD1 WT but does not carry the mutated hSOD1 G93A transgene; "vehicle" and "ASC-CM" mice are the experimentally-treated SOD1 G93A transgenic mice that express wild type endogenous mouse SOD1 (mSOD1 WT ) in addition to the hSOD1 G93A transgene. Actin levels remain unchanged and assayed as loading controls for endogenous mSOD1 WT . There were differences in endogenous mSOD1 WT levels, which then served as an internal control for mutated hSOD1 G93A expression.
GFAP expression was reduced in the spinal cords of symptomatic SOD1 G93A mice treated with ASC-CM. It has been demonstrated that motor neuron loss occurs partly through activated inflammatory cells in the surrounding area 42 . Because treatment with ASC-CM showed significant neuroprotection of motor neurons possibly via anti-inflammatory mechanisms thus far, glial activation after treatment with ASC-CM was examined to further explore this possibility. Expression of GFAP was measured in the spinal cords of SOD1 G93A mice after 3 days of treatment with ASC-CM or vehicle. Representative immunoblots demonstrated that there is significantly reduced GFAP expression with ASC-CM treatment (2.71 ± 0.08) versus vehicle (4.12 ± 0.07; ***p < 0.005; Fig. 5). The GFAP levels observed in the ASC-CM treated group were comparable to that of the wild type controls (2.20 ± 0.08).

Figure 3. Treatment with ASC-CM did not affect expression of endogenous mouse wild type SOD1 and mutated human SOD1 transgene in spinal cord.
Wild type control (WT) and SOD1 G93A mice (vehicle and ASC-CM) spinal cords were prepared as detailed in materials and methods. Levels of actin were unchanged and used as an internal loading control for endogenous mouse wild type SOD1 (mSOD1 WT ). There was no difference in mSOD1 WT levels, which was the loading control for mutated human SOD1 (hSOD1 G93A ). After 7 days of ASC-CM treatment, there was no change in expression of either endogenous mSOD1 WT or hSOD1 G93A mice in symptomatic SOD1 G93A mice. n = 3 per group. Unchanged total p38 levels were used as internal loading controls. Values are shown as mean density ± SEM and statistical analyses were performed using one-way ANOVA. *p < 0.05; **p < 0.01; ***p < 0.005; n = 3 per group. Administration of ASC-CM to SOD1 G93A mice with disease onset decreased CD11b expression in spinal cords. In order to further investigate glial activation after ASC-CM treatment, spinal cords were prepared for Western blot and examined for expression of CD11b, a glial marker found to be induced by neuroinflammation 14 . CD11b expression was observed at 3 days (Fig. 6A) and 7 days (Fig. 6B) after ASC-CM or vehicle treatment. 3 days of ASC-CM treatment demonstrated no differences in CD11b expression in all groups (group averages ± SEM: 0.073 ± 0.018 for wild type; 0.068 ± 0.012 for vehicle; 0.066 ± 0.018 for ASC-CM, Fig. 6A). However, 7 days of ASC-CM treatment demonstrated a significant decrease in CD11b expression in the spinal cord of ASC-CM-treated animals (0.109 ± 0.017, Fig. 6B)when compared to vehicle mice (0.251 ± 0.015 for vehicle, ***p < 0.005 versus wild type and ASC-CM), whose CD11b expression level was at a similar level to wild type mice (0.100 ± 0.005) after 7 days.
ASC-CM treatment of symptomatic SOD1 G93A mice resulted in decreased expression of α-II spectrin in spinal cords. ASC-CM's effect on the calpain system in SOD1 G93A mice was determined by measuring expression of α -II spectrin, a known cleavage target of calpain 43 . After 3 days following disease onset, activity of the Ca 2+ -activated protease, calpain, was increased in SOD1 G93A mice spinal cords given vehicle(0.81 ± 0.002, n = 4, ***p < 0.005) as compared to wild type controls (0.10 ± 0.006, n = 3), consistent to a previous report 20 . Cleaved α -II spectrin was significantly decreased in SOD1 G93A mice receiving ASC-CM treatment as compared to mice given vehicle (0.15 ± 0.026, n = 3, ***p < 0.005, Fig. 7). Levels of actin remained unchanged for all groups and were used as internal loading controls.

Discussion
This study was designed to confirm the effects of a biologically-derived reagent on an ALS animal model that is clinically relevant such that treatment was applied after disease onset was established in the widely used SOD1 G93A mouse model 8 . The data collected from this study demonstrate the effectiveness of ASC-CM treatment in prolonging post-onset survival and extending the lifespan of symptomatic SOD1 G93A mouse. These results are in agreement with previous experiments from our laboratory, showing that the treatment with ASC-CM is beneficial and protective of neurons in vivo during neuronal injury or challenge, such as in the hypoxic-ischemic rat neonatal brain 27 . The findings from this study Although the etiology of sporadic ALS is currently unknown, through the use of animal models and correlation with neuropathological findings from human ALS patients, several key mechanisms of neuronal death have been proposed to disrupt normal motor neuron function and lead to their degeneration. Current studies implicate glutamate-induced excitotoxicity, mitochondrial dysfunction, oxidative stress, neuroinflammation, aberrant axonal transport systems, endoplasmic reticulum (ER) stress, inhibition of proteosomal function, compromised blood-brain barrier, and abnormal protein aggregation 44 . There is much debate as to which processes contribute towards disease etiology and which events are secondary to the established disease state. This presents ALS drug development with difficulties as neuropathology-specific targeting of treatment may be ineffective.
Despite the mechanistic uncertainties underlying events leading to ALS neuropathologic changes, treatments like ASC-CM may prove to be beneficial as they are able to block or modify several neuronal death pathways due to the heterogeneity in their composition. ASC-CM has been found to contain a multitude of neurotrophic factors that could be beneficial for the diseased motor neuron 27 . It was shown that neurotrophic factors may play a critical role in protecting or delaying motor neuronal death in ALS mice 45 . In this study, neurotrophic factors in ASC-CM could access affected neurons in CNS by either directly crossing the blood brain barrier 46,47 or going through the retrograde transport mechanism 45,46 . Additionally, since motor neuronal death in ALS mice showed a distal-proximal gradient, it is also possible some neurotrophic factors in ASC-CM increased lifespan of ALS mice by directly protecting the neuromuscular junctions in peripheral 45 . Furthermore, in order to overcome the short half-life of polypeptides in ASC-CM peripherally, we i.p. injected ASC-CM daily until the end of experiments 48,49 . Here, after post-onset treatment with ASC-CM peripherally, symptomatic SOD1 G93A mice have significant prolongation of post-onset survival times, which translated into an overall extension of lifespan. However, since many neurotrophic factors existing in ASC-CM may exert neuroprotective effects together as a neuroprotective mixture body 27 , deletion of single factor may fully remove its neuroprotective effects. In this study, we test this hypothesis by targeting the NGF contained in ASC-CM 50 . As predicted, NGF neutralized ASC-CM did not elongate the post-onset survival times and lifespan of SOD1 G93A mice.
Additionally, the lengthening of post-onset lifespan following ASC-CM administration was correlated with markedly increased numbers of motor neuron survival in the spinal cord lumbar area. Further examinations at the cellular level determined that the increase in motor neuron survival after ASC-CM treatment was accompanied by decreased expression of glial activation markers and phosphorylated p38 MAP kinase, which are important components of the neuroinflammation pathway 51 contributing to neuronal death in the spinal cords of SOD1 G93A mice 17 . Other data from this study shows lower levels of cleaved α -II spectrin, which is a known substrate of both calpain and caspase-3 52,53 . Decreased cleavage of α -II spectrin in the spinal cords of ASC-CM-treated SOD1 G93A mice may indicate that calpain has protective effects against glutamate-induced neuronal apoptosis 20,21,23 . Growing evidence has shown that NGF induces neuronal survival with low-affinity binding to the p75 neurotrophin receptor (p75NTR) 54 . Additionally, it was shown that NGF-induced neuroprotection against OGD insult by inhibiting OGD-induced p38 activation 55 and against UV neurotoxicity by inhibiting calpain activity 56 . Furthermore, increased production of NGF in central nervous system (CNS) during diseases is able to suppress inflammation by switching the immune response to an anti-inflammatory 57 . Therefore, it is quite possible that NGF induced neuroprotection in ASC-CM-treated SOD1G93A mice via neuroprotection, inhibition of p38 and calpain, as well as anti-inflammation. It would be interesting to investigate by using appropriate models how NGF protects motor neurons in ASC-CM-treated SOD1 G93A mice and whether glial activation or inflammatory processes play important roles in this neuroprotective process.
Taken together, this study demonstrated that ASC-CM provides significant neuroprotection in the SOD1 G93A mouse model of ALS and extendsSOD1 G93A mice life span. NGF in ASC-CM plays a significant role in this neuroprotection. This protective ability may associate with the preservation of motor neurons and inflammatory pathway inhibition in the spinal cord. This study establishes the therapeutic potential of this agent for treating ALS. Additionally, ASC-CM also holds great promise as potential therapies for many other neurodegenerative diseases due to the validated beneficial effects in many experimental models.