Cerebrospinal fluid from Alzheimer’s disease patients as an optimal formulation for therapeutic application of mesenchymal stem cells in Alzheimer’s disease

Mesenchymal stem cells (MSCs) have emerged as one of the promising treatment options for Alzheimer’s disease (AD). Although many studies have investigated on the efficacy of MSCs in AD, how MSCs actually change following exposure to the AD environment has not been studied extensively. In this study, we investigated on the potential of AD patient-cerebrospinal fluid (CSF) samples to be used as a formulation of MSCs and its application in AD therapeutics. When Wharton’s jelly-derived mesenchymal stem cells (WJ-MSCs) were stored in the CSF of AD patients, the stemness of WJ-MSCs was preserved. Furthermore, several genes were upregulated following storage in AD CSF. This signified the therapeutic potential of CSF formulation for AD therapy. Overall, these findings suggest that CSF from AD patients can be an optimal source for MSC formulation.

SCIENTIfIC RepoRts | (2019) 9:564 | DOI: 10.1038/s41598-018-37252-9 routes of administration for MSCs is the intracerebroventricular route, where cells have higher chances to penetrate into the parenchyma of the brain 22 . To perform repeated intracerebroventricular injections, insertion of a device such as an Ommaya reservoir has to be preceded prior to the first stem cell injection, which makes the collection of CSF from AD patients easier. MSC formulation using the CSF from the patient to whom the MSCs will be injected into may therefore ensure the safety of the therapeutic agent. Second, since CSF flow is part of the Aβ clearance mechanism in AD 23,24 , CSF from AD patients might represent a microenvironment for AD. Therefore, MSCs that have been exposed to the AD patient CSF as a formulation would be preconditioned in the AD microenvironment prior to administration. This pre-exposure of MSCs to the AD microenvironment might allow MSCs to better cope with the disease environment than naïve MSCs.
In the present study, we stored human Wharton's jelly-derived MSCs (WJ-MSCs) in CSF samples obtained from four different AD patients and three normal controls under hypothermic conditions (4 °C). We then investigated whether the viability and stemness of WJ-MSCs were compromised following exposure to CSF of AD patients or normal controls. In addition, we explored changes in gene expression levels of WJ-MSCs to assess the therapeutic effects of AD CSF formulation.

CSF biomarkers of AD patients and normal controls. The four AD patients [male (M): female
(F) = 2:2, age range 50-60] who underwent 18F-florbetaben amyloid positron emission tomography (PET) scans were amyloid positive. Negative results were obtained from the three cognitively normal controls (M:F = 2:1, age range 63-81) who also underwent the same amyloid PET scans. Although some samples showed unexpected levels of CSF biomarkers (AD 1, 2: both total and phosphorylated tau lower than AD criteria; Control 1: lower Aβ 42 level, marked as * in Table 1), tau/Aβ 42 ratios (marked as ** in Table 1), which are known to be more accurate than either of tau and Aβ 42 values, were different between the normal control and AD CSF samples (Table 1). Furthermore, these ratios from all AD patients were above the normal range 25 .

Hypothermic storage of WJ-MSCs in AD and normal CSF does not compromise cell viability.
Changes in cell viability of WJ-MSCs stored in CSF samples obtained from four different AD patients and three normal controls for 72 hours were assessed by performing fluorescence-activated cell-sorting (FACS) analysis after Annexin V/7-AAD staining (Fig. 1a,b). Images of the cells were taken (Fig. 1c) and the CCK-8 cell counting kit assay was also conducted (Fig. 1d). Compared to WJ-MSCs stored in minimum essential medium alpha 1x (MEMα 1x), significant differences in viability were not observed following hypothermic storage in both AD CSF and normal CSF (Fig. 1). A small percentage of early but not late apoptotic cells was detected from MEMα 1x and CSF groups (both AD and normal) (Fig. 1a,b). Similar results were obtained from the CCK-8 assays, where differences in viability between the MEMα 1x and the two CSF groups were not detected when the assay was performed every 24 hours for up to 72 hours (Fig. 1d).

WJ-MSCs maintain stemness following hypothermic storage in AD and normal CSF.
Immunophenotype characteristics of human WJ-MSCs were analyzed according to the MSC criteria proposed by the International Society for Cell Therapy (ISCT) 26 . Like WJ-MSCs stored in MEMα 1x, WJ-MSCs stored in normal and AD CSF expressed the following cell surface markers: CD90, CD73, CD105, CD166 and also did not express the following hematopoietic markers: CD14, CD11b, HLA-DR, CD34, CD45, and CD19 ( Fig. 2 and Supplementary Fig. S1). Such results verified that immunophenotypic features were not altered following exposure to both AD and normal CSF samples. WJ-MSCs were also able to differentiate into various mesenchymal linages (adipogenic, osteogenic, chondrogenic) after exposure to normal and AD CSF samples (Fig. 3). The differentiation efficiency was also similar to that of WJ-MSCs stored in MEMα 1x, although MSCs stored in AD CSF showed less tendency to differentiate into osteocytes.   WJ-MSCs exposed to AD CSF and normal CSF were compared to those of WJ-MSCs exposed to MEMα 1x. Based on the PCR array data, scatter plots and heatmaps were analyzed (Fig. 4a,b). The mRNA expression level of WJ-MSCs exposed to AD CSF was variable among the samples (Fig. 4b). Compared to the WJ-MSCs stored in normal CSF (n = 3), significant alterations in mRNA expression levels were exhibited in 47 genes when WJ-MSCs were exposed to AD CSF (n = 4) ( Supplementary Table S1). Interestingly, mRNA expression levels of WJ-MSCs exposed to AD CSF were significantly higher than those of WJ-MSCs exposed to normal CSF. Despite variability of expression levels among the patients' samples, the pattern of upregulated mRNA was consistent.  Table S2) and functional annotation clusters ( Table 2). The 10 most enriched annotation clusters of significantly changed genes included items related to angiogenesis, signal peptide, negative regulation of programmed cell death, regulation of cell proliferation, extracellular region, tube development, morphogenesis of branching structure, cellular component morphogenesis, regulation of neurogenesis, and sensory organ development.

Discussion
Recent studies have reported on the neuroprotective and neurotrophic features of MSCs 27 . Previous studies also showed that proteins secreted from MSCs induced clearance of Aβ proteins 8,9 , promoted neurogenesis and also synaptogenesis 10,13,17 . These studies indicated that a therapeutic interaction existed between the paracrine factors secreted by the MSCs and the endogenous progenitors present in the brain. While former studies were mainly focused on the therapeutic effects of MSCs, the cellular status of MSCs exposed to the AD brain microenvironment has not been fully elucidated.
Several studies have examined the effects of CSF on stem cells. Growth factors found in the CSF have been reported to affect stem cell proliferation 28,29 and regulate quiescence and activation of stem cells in the brain 30 .
Other studies have used CSF to transdifferentiate MSCs into neural cells. For example, MSCs cultured in-vitro in CSF (as a substitute of culture media) transdifferentiated into neural-like cells 31,32 . However, these studies have focused on elucidating the effects of CSF on stem cells under normal and not disease conditions.
In this study, we applied CSF samples from AD patients as a formulation of WJ-MSCs. For comparison, MEMα 1x, which is conventionally used for MSC formulation was included, and normal control CSF samples were also included to determine whether AD pathology can possibly alter the therapeutic effects of MSCs. According to the results obtained from this study, the viability and stemness of WJ-MSCs were both preserved after exposure to AD CSF under hypothermic conditions. These results confirmed the safety of AD CSF to be used as a potential source of formulation.   We further explored changes in gene expression levels of WJ-MSCs exposed to AD and normal CSF samples. Based on analysis of functional annotation clustering, WJ-MSCs stored in AD CSF expressed genes related to enhancement of extracellular transport and signal peptide, which indicates an increase in paracrine activity. These genes are known to exhibit neuroprotective and neurotrophic features such as negative regulation of apoptosis, regulation of cell proliferation, and regulation of neurogenesis. Furthermore, an increase in the expression of genes involved in cell migration or cell adhesion was also observed, which indicates potential beneficial effects on cell survival following administration. These results suggest that AD CSF may act as an optimal formulation for MSCs that will be injected back to the AD patient from whom the CSF sample was obtained from (Fig. 6).
Our study has several limitations. First, our study was based on a small number of patients. Therefore, further research involving a large number of patients is warranted. Second, although we measured the gene expressions of MSCs to investigate functional changes, we have not yet assessed the efficacy of MSCs in animal disease models (e.g., transgenic AD animal models). Nevertheless, to the best of our knowledge, this is the first study which has investigated the fate of MSCs in AD CSF, the interaction between MSCs and CSF samples, and the potential of AD CSF as a formulation source. Our approach may have several advantages. First, patients will be reinfused with their own CSF, which may contribute to minimizing the side effects of allogeneic MSC administration. Autologous MSCs are also available but not as cost-effective due to the requirement of large scale cell expansion, compared to allogenic MSCs. Moreover, the surgical process involved in isolating MSCs is difficult to perform in AD patients due to their age. Second, to achieve repeated injections of MSCs into the lateral ventricle of patients, a device such as an Ommaya reservoir has to be surgically implanted. Therefore, collection of CSF from the patient may not be as invasive as expected. In order to ensure patient safety, CSF collection must be performed under closed, sterile procedures. CSF collected under sterile conditions can be used as an optimal formulation of MSCs produced from the GMP facility. Third, using AD CSF as a formulation of MSCs allows MSCs to pre-adapt to the disease environment and to also become pre-activated prior to administration. The use of CSF from AD patients as a source of formulation may enhance the overall efficacy of AD MSC therapy. This approach can also be applied to a wide range of neurological diseases.   Disease and Related Disorders Association (NINCDS-ADRDA) criteria 33 for probable AD and also had positive 18F-florbetaben PET scans 34,35 . Patients with neurological diseases other than AD were excluded from the study. The three normal controls were recruited from orthopedic clinics in our hospital who underwent spinal anesthesia to receive knee surgeries. All of them met the criteria for normal elderly 36 , had normal Mini-mental State Examination scores defined by age/sex matched cohort 37 , and also had negative 18F-florbetaben PET scans. A total of 10-12 mL of CSF was obtained from each of the four AD patients and 3 mL from the three normal controls by lumbar puncture (between the L3/L4 or L4/L5 intervertebral space). Within two hours from collection, CSF samples were centrifuged at 4,000 g for 10 minutes at 4 °C. Aβ 1-42 , P-tau 181P , and T-tau concentrations were examined from the CSF samples by using the IINOTEST (ELISA) assay. Then the ratios of tau (both t-tau and p-tau) to Aβ 42 were obtained to validate the difference between normal controls and patients, as these ratios are known to be more accurate than the level of Aβ 42 and tau 25,38 . To analyze the viability of the cells at various time points (0, 24, 48, 72 hrs), MSCs stored in CSF were seeded onto 96-well plates (3 × 10 3 /well), and after 24 hours, CCK8 assay was performed. CCK-8 solution (10 uL) was added to each well, followed by incubation for 1 hour at 37 °C. The absorbance of CCK-8 was measured at 450 nm by using a microplate reader (x-Mark TM , Bio-Rad Laboratories, Inc, USA).

Flow-cytometric analysis for cell surface markers of WJ-MSCs.
In order to confirm the stemness of WJ-MSCs, cell surface marker analysis was performed. Harvested MSCs were washed in PBS supplemented with 2% FBS in order to block for non-specific binding sites. Immunophenotypic analysis of MSCs was carried out using flow cytometry for the following markers: CD73, CD90, CD105, CD166, CD14, CD11b, HLA-DR (MHCII), CD34, CD45 and CD19 (BD Biosciences, USA). At least 10,000 events were acquired by using the BD FACS Verse flow cytometer, and the results were analyzed by using the BD FACSuite software version 10. Flow cytometry for appropriate isotype controls were also performed.

Mesoderm differentiation assays.
Another way to test stemness is to examine the differentiation capabilities of MSCs. WJ-MSCs stored under hypothermic conditions for 72 hours were seeded onto 6-well culture plates at a density of 5000 cells/cm 2 and expanded until cells reached 80-90% confluency. For osteogenic and adipogenic differentiation and the respective immunostaining experiments, cells were incubated in differentiation media according to manufacturer's instructions (Gibco, USA). Differentiation medium was replaced every 3 days. After 2 weeks, differentiated cells were stained using the following staining methods: osteogenic; Alizarin Red S, adipogenic; Oil Red O. Osteocytes were fixed with 4% paraformaldehyde (PFA) for one hour and then washed with PBS not including both calcium and magnesium (Gibco, USA). Mineralization of the extracellular matrix was visualized by staining with 40 mM Alizarin Red S (Sigma-Aldrich, USA), pH 4.2, for 5 minutes. Adipocytes were fixed with 4% PFA, washed in 60% isopropanol, and subsequently incubated for 10 minutes with Oil-Red O (Sigma-Aldrich, USA) to visualize the lipid droplets. Cells were then washed in isopropanol and counterstained with hematoxylin.
Previously described methods were used to quantify the differentiation potential between each samples [39][40][41] . was evaluated using the RT2 RNA QC PCR Arrays (Qiagen) and cDNA was synthesized from 500 ng of the total RNA using the RT2 First Strand Kit (Qiagen, USA). The samples were analyzed using the RT2 Profiler PCR Array. The reaction mix was prepared from 2x RT2 qPCR Master Mix and 102 μL of sample cDNA. 10 μL of this mixture was added into each well of the PCR Array. Altogether, 84 different genes were simultaneously amplified in the sample. A melting curve analysis was performed to verify that the product consisted of a single amplicon. PCR arrays were performed in 96-well plates by using a QuantStudio 6 flex real-time PCR system (Applied Biosystems, Thermo Fisher, USA). Data were analyzed via the QuantStudio software and the Ct values were determined for each gene. The thresholds and baselines were set according to the manufacturer's instructions (Qiagen, USA). The fold change in gene expression (compared to the positive control: WJ-MSCs) was calculated using the ΔΔCt method. Compared to the control (MEMα 1x), gene expressions with a fold change ≥2 was only considered. Clustering analysis of altered gene expressions was performed by using the MeV software (Ver. 4.9.0). Statistical analyses. The results are an average of three independent experiments. Data are presented as mean ± standard error of the mean (SEM). Statistical comparisons of each samples between groups were performed using a one-way ANOVA test (both between groups and within groups). Differences were considered statistically significant when P < 0.05. All the statistical analyses were performed using SigmaPlot, version 12.5 and SPSS software, ver 19.0 for Windows.