Paracrine Proangiogenic Function of Human Bone Marrow-Derived Mesenchymal Stem Cells Is Not Affected by Chronic Kidney Disease

Background Cell-based therapies are being developed to meet the need for curative therapy in chronic kidney disease (CKD). Bone marrow- (BM-) derived mesenchymal stromal cells (MSCs) enhance tissue repair and induce neoangiogenesis through paracrine action of secreted proteins and extracellular vesicles (EVs). Administration of allogeneic BM MSCs is less desirable in a patient population likely to require a kidney transplant, but potency of autologous MSCs should be confirmed, given previous indications that CKD-induced dysfunction is present. While the immunomodulatory capacity of CKD BM MSCs has been established, it is unknown whether CKD affects wound healing and angiogenic potential of MSC-derived CM and EVs. Methods MSCs were cultured from BM obtained from kidney transplant recipients (N = 15) or kidney donors (N = 17). Passage 3 BM MSCs and BM MSC-conditioned medium (CM) were used for experiments. EVs were isolated from CM by differential ultracentrifugation. BM MSC differentiation capacity, proliferation, and senescence-associated β-galactosidase activity was assessed. In vitro promigratory and proangiogenic capacity of BM MSC-derived CM and EVs was assessed using an in vitro scratch wound assay and Matrigel angiogenesis assay. Results Healthy and CKD BM MSCs exhibited similar differentiation capacity, proliferation, and senescence-associated β-galactosidase activity. Scratch wound migration was not significantly different between healthy and CKD MSCs (P = 0.18). Healthy and CKD BM MSC-derived CM induced similar tubule formation (P = 0.21). There was also no difference in paracrine regenerative function of EVs (scratch wound: P = 0.6; tubulogenesis: P = 0.46). Conclusions Our results indicate that MSCs have an intrinsic capacity to produce proangiogenic paracrine factors, including EVs, which is not affected by donor health status regarding CKD. This suggests that autologous MSC-based therapy is a viable option in CKD.


Background
Chronic kidney disease (CKD) affects~10% of the world population and leads to high morbidity and mortality [1]. Endothelial injury plays a key role in the development of CKD [2,3]. Once established, CKD follows a progressive course of inflammation and fibrosis, which ultimately leads to end-stage renal failure, necessitating renal replacement therapy. From a clinical perspective, kidney transplantation remains the best possible therapy for end-stage renal failure, but donor organs are scarce and long-term graft failure is still between 30 and 50% ten years posttransplant [4]. Treatments to slow or reverse CKD progression are thus urgently needed.
Bone marrow-(BM-) derived mesenchymal stromal cells (MSCs) have potent antifibrotic, proangiogenic, and immunomodulatory properties, which makes BM MSC-based therapy for CKD a viable option [5]. BM MSCs secrete paracrine factors, such as growth factors and cytokines, that enhance local angiogenesis and tissue repair [6,7]. BM MSC-secreted extracellular vesicles (EVs) also play an essential role in the proangiogenic properties of MSCs [8]. Indeed, BM MSC-based regenerative medicine approaches are being developed to treat CKD, aiming to halt progression of fibrosis by vascular regeneration, stimulate kidney regeneration, or enhance immunomodulation after renal transplantation. These therapies include administration of cells or products secreted by cells [9][10][11]. BM MSCs are an effective treatment for CKD in a range of preclinical animal models [12]. Following these promising results, clinical applications of BM MSCs are being investigated in phase I and II studies [13].
For many applications, administration of allogeneic BM MSCs is a feasible option, due to their immunomodulatory properties that inhibit rejection [14]. However, BM MSCs may not be fully immunoprivileged [14,15]. Autologous therapy should thus be considered for these patients-provided that disease-induced dysfunction is not present. Preclinical studies show a pronounced reduction in vascular regenerative effects of BM mononuclear cells of donors with cardiovascular disease (CVD), including CKD [16]. In culture-expanded MSCs, the effect of CVD is less clear, with conflicting reports depending on the cardiovascular disease (model) and little information available for CKD [17]. MSC immunomodulatory properties seem preserved in BM MSCs, though it is currently unknown whether CKD negatively influences the paracrine regenerative potential of BM MSCs and BM MSCderived EV [17][18][19].
Verification of maintained BM MSC paracrine proangiogenic function is thus an important step in the development of MSC-based regenerative medicine strategies for CKD. Here, we investigated the regenerative potential of BMderived MSCs and EVs derived from BM MSC from CKD patients versus healthy controls.

Study Participants.
This study was approved by the local institutional review board of the University Medical Center Utrecht (METC number  and complies with the Declaration of Helsinki. Written and verbal informed consent was provided by all participants prior to inclusion. Kidney recipients and kidney donors participating in the living donor program and who were above 18 years of age were eligible for participation. Exclusion criteria for CKD patients (kidney recipients) were stem cell transplantation in the past, and general exclusion criteria for renal transplantation are active infection: hepatitis B and C, tuberculosis, human immunodeficiency virus; life expectancy of <2 years; and malignancy not curatively treated. Exclusion criteria for healthy controls (kidney donors) were stem cell transplanta-tion in the past and present kidney disease. One week prior to bone marrow collection, all CKD patients started an oral immunosuppressive regimen to prepare for kidney transplantation, which consisted of mycophenolate mofetil (MMF) 750 mg twice daily and prednisone 7.5 mg once daily.
2.2. BM Aspiration. BM aspiration took place after induction of anesthesia for the kidney donation or transplantation procedure. BM aspiration was conducted at the iliac crest in accordance with the standard operating procedures at the UMC Utrecht Haematology Department. A T-Lok bone marrow biopsy device (Argon Medical Devices, Frisco, TX) was used. 20 mL of aspirated bone marrow was collected in heparin-coated vacuum tubes.
2.3. Materials and Reagents. All reagents were obtained from ThermoFisher unless otherwise specified.

MSC Isolation and Expansion
. BM MNCs were obtained by gradient density centrifugation using Ficoll-Paque (Sigma-Aldrich, Zwijndrecht, NL). 10 × 10 6 BM MNCs were plated in one 6-well plate in 2 mL α-MEM+10% FCS (Gibco), 0.1 ng/mL basic fibroblast growth factor, 100 μM ascorbic acid, and 1% penicillin/streptomycin. After 24 hours, the medium was changed, with subsequent medium changes thrice a week. When the cells reached 80% confluence, cells were reseeded on 25 cm 2 , then 75 cm 2 , and then 4 × 75 cm 2 . At third passage, cells were frozen in α-MEM containing 10% DMSO and 20% FCS. A Mr. Frosty was used to ensure freezing at a controlled rate. All experiments were conducted with passage 3+1 cells (washed and adhered overnight). Experiments were started 48 h after thawing and seeding of the cryopreserved cells.
2.6. EV Isolation and Sucrose Gradient Analysis. EVs were isolated from conditioned medium of 13 MSC lines (6 controls, 7 CKD). For EV isolation, BM MSCs were cultured in EV-free medium (prepared using FCS centrifuged for at least 1 hr at 200,000×g, followed by 0.2 μm filter sterilization). Per participant, MSCs were grown to 80% confluence in 8x T175 flasks. CM collection was initiated after washing cells. CM was collected for a period of 24 h. EVs were then isolated from the CM by differential ultracentrifugation as previously described [21].
2.8. Nanosight Nanoparticle Tracking Analysis. Vesicle size was determined using a Nanosight NS500 (Marvern, Worcestershire, UK) with the following capture settings: camera level: 13, slider shutter: 1232, slider gain: 219, and a detection of 5, and autoblur mode for the analysis.
2.10. Osteogenic Differentiation. 40,000 cells (were incubated with Stem Pro Osteogenic Differentiation medium for 14 days. Osteogenic differentiation was assessed by measuring alkaline phosphatase activity using p-nitrophenyl phosphate (pNPP). Cells were lysed with a Tris/SDS buffer and the conversion of pNPP substrate as a proxy for alkaline phosphatase activity was measured by measuring absorption at 405 nm with a spectrofluorometer (Fluoroskan, Thermo-fisher). Values were corrected for cell viability assessed using PrestoBlue. For microscopy confirmation of bone formation, cells were stained with alizarin red, as previously described [24]. MSCs were fixed with 4% (w/v) paraformaldehyde for 15 min at room temperature. Cells were then stained with 40 mM alizarin red staining solution for 20 min. Cells were washed 4 times with PBS, and pictures were taken using an inverted microscope (Olympus, Zoeterwoude, NLD).
2.11. Chondrogenic Differentiation. 100,000 cells were washed with Stem Pro Chrondrogenic Differentiation medium and pelleted at the bottom of a 15 mL conical tube by centrifugation at 500g for 15 minutes. Cells were incubated with Stem Pro Chrondrogenic Differentiation medium for 28 days (medium changes thrice a week). Glycosaminoglycan content was measured with the BlyScan kit (NBiocolor, Carrickfergus, UK).

Proliferation.
To measure proliferation, the xCElligence RTCA DP platform was used (Acea Biosciences, San Diego, CA). 500 cells (in quadruplicate) were plated in an E16 plate. Growth curves were constructed using electrical impedance. Results were validated using the PrestoBlue viability assay.
2.13. Collection of Conditioned Medium. 80,000 cells per donor were seeded in a 6-well plate, and cells were left to adhere for 24 hours. Cells were washed with serum-free α-MEM and fresh serum-free α-MEM was added. After 24 hours, conditioned medium (CM) was then collected in serum-free α-MEM for 24 hours. CM was centrifuged at 2,000g for 10 minutes prior to freezing and storage at -80°C.
2.14. Tubule Forming Assay. Tubule forming capacity was measured using HMEC-1 cells (passage 16). CM was layered on top of 10 μl solidified growth factor reduced Matrigel (BD, Vianen, NL) on an IBIDI angiogenesis μ-Slide (München, DE). 10,000 HMEC-1 cells were resuspended in serum-free α-MEM and seeded into the CM. The assay was performed in triplicate. After 16 hours, microscopic pictures were taken at 4x magnification. Tubule networks were manually traced in Adobe Photoshop and subsequently analyzed using the AngioAnalyzer ImageJ plugin, which provided the junction-to-tubule length ratio as a measure for network maturity. Values are presented as relative to the positive control and normalized per experiment.

Endothelial Scratch Wound
Assay. HMEC-1 cells were seeded in a 24-well plate and grown until confluence. A scratch was created using a p100 pipette tip. The cells were washed once with PBS to remove detached cells. Conditioned medium or control medium was then applied. Pictures at specifically marked locations were taken at baseline and after 6 hours. The assay was performed in quadruplicate. Scratch delta width in pixels was manually measured in Adobe Photoshop, and % migration was calculated by dividing the scratch width at T = 6 h by the width at T = 0 h. Values are presented as relative to the positive control and normalized per experiment.
2.16. Senescence. 40,000 cells were seeded in a 12-well plate. Senescence-associated β-galactosidase activity was assessed as previously described [25]. Briefly, after overnight incubation, cells were incubated for 1 hr with 100 nM bafilomycin A1 (Sigma) to induce lysosomal alkalinization. Subsequently, 5-dodecanoylaminofluorescein diβ-D-galactopyranoside (C12FDG, Invitrogen) was added to a final concentration of 30 μM, and cells were incubated for an additional hour. Cells were trypsinized and washed with PBS. Median cellular fluorescence in the FITC channel was quantified using Flow Cytometry. SYTOX ® Blue (Invitrogen) was added immediately prior to measurement to identify dead cells, which were excluded from analysis.
2.17. Statistical Analysis. Statistical analyses were performed in IBM SPSS version 24 as well as Prism Graphpad version 6.01. Values are presented as relative to the healthy control group to allow a comparison between CKD and control MSCs. Comparisons of continuous variables are made using the Student's t test for normally distributed data or the Mann-Whitney test for nonnormally distributed data. Ratios were generated using the Fieller method [26]. If necessary, for nonnormally distributed parameters, means and SD were generated using bootstrapping. A value of P = <0:05 is considered statistically significant.

Isolation and Characterization of MSCs.
MSCs were obtained from bone marrow of 32 study participants (17 healthy kidney donors and 15 kidney recipients) participating in the living donor kidney transplantation program at the University Medical Center Utrecht. Bone marrow was aspirated at the iliac crest after induction of anesthesia for the donor or transplantation procedure. Table 1 lists the participant characteristics. MSCs were successfully expanded from all BM MNC cultures. MSCs displayed CD73, CD90, CD105, and CD140b and did not display CD14, CD19, CD34, or CD45 (Supplementary Figure S1). The time of expansion was similar for all samples (data not shown).

Proliferation.
Proliferation was assessed in real-time using the xCELLigence platform (Figure 2(a)). There was no difference between cell indexes of healthy and CKD MSCs nor was there a difference in extrapolated population doubling time (Figure 2 3.4. Senescence. Senescence was assessed by measuring senescence-associated β-galactosidase activity using a fluorescent substrate (C12FDG). There was a significant difference between CKD and healthy MSCs; β-galactosidase activity was higher in healthy MSCs than in CKD MSCs (Figure 2(c), ratio CKD : healthy 0.79, 95% CI 0.64-0.99, P = 0:05). Senescence was not associated with age ( Figure 2

Scratch Wound Migration
Assay. The effect of MSC CM on HMEC-1 migration was studied by using the scratch wound assay, which models endothelial cell migration as it occurs in injury to the vascular lumen (Figure 3(a)). The degree of cell migration in response to MSC CM addition is a measure for paracrine stimulation of endothelial cells. Healthy and CKD MSC CM induced similar migration of HMEC-1 cells. (ratio CKD : healthy 1.28, 95% CI: 0.89 to 1.77, P = 0:18) (Figure 3(b)). Scratch wound migration did not correlate with senescence or statin use, but scratch wound migratory capacity was increased in females (ratio female : male 1.4, 95% CI: 1.02 to 1.96, P = 0:04). Two-way ANOVA showed an association between sex and paracrine function (P = <0:001) (Supplementary Figure S4).

Paracrine Effects of BM MSC-Derived Extracellular Vesicles
3.6.1. Extracellular Vesicles. EVs were isolated by differential ultracentrifugation of conditioned medium from 13 MSC cultures derived from male participants and included 7 CKD patient-and 6 healthy control-derived MSCs. EVs showed a typical size and profile for MSCs [27], and there was no difference in number or size of the EVs (Supplementary Figure S5). Absence of contamination with nuclei and mitochondria, which can be derived from dead cells, was verified by immunoblotting for the mitochondrial proteins ATP5a and TOM20, and the nuclear envelope protein observed in the isolated vesicles. (Figure 4(a)). Sucrose density gradient analysis and subsequent immunoblotting for β-actin and Flotillin-1 demonstrated that the isolated EVs had a density of 1.10 g/mL (Figure 4(c)). Nominal size of these vesicles, as verified by Nanosight Nanoparticle Tracking Analysis, was 159 nm (Figure 4(b)).

Scratch Wound Migration and Tubule Formation
Assay. To specifically investigate whether the regenerative capacity of MSC-derived exosomes (or small EV) was affected by disease status, the stimulatory function of the isolated vesicles in scratch wound migration and Matrigel angiogenesis assays was assessed. Both CKD and healthy MSC-derived EVs induced migration in the endothelial scratch wound migration assay. There were no statistical differences ( Figure 5(a), ratio CKD : healthy 0.86, 95% CI: 0.34-1.42, P = 0:6). Matrigel angiogenesis analysis showed similar tubule-length-to-junction ratios for both healthy and CKD MSC-derived EVs ( Figure 5(b)), ratio CKD : healthy 1.33, 95% CI: 0.52-2.27, P = 0:46). EV and CM angiogenic potential were not correlated (P = 0:6).

Discussion
We show that CKD does not affect the intrinsic capacity of BM MSCs to produce proangiogenic paracrine factors, including EVs. We demonstrate that CM and EVs derived from CKD patients' BM MSCs display similar in vitro angiogenic potential and propensity to induce endothelial scratch wound migration. Additionally, we found that CKD does not lead to increased senescence nor altered differentiation capacity and growth rate of MSCs.
There is a continued debate on whether MSC-based therapy should be performed with autologous or allogeneic cells. In many cases, allogeneic MSCs are preferred, as they are available "off the shelf" and can therefore be used in acute conditions such as myocardial infarction. Lastly, many diseases have been shown to affect the stem cell niche, potentially reducing the effectiveness of patient-derived progenitor cells [28]. MSCs have been shown to be less affected by this than other cell types considered for cell therapy, but whether disease-related dysfunction carries over in the cell product seems to depend on the disease [17].
There are, however, also potential limitations to allogeneic therapy, mainly related to the development of alloimmunity and rejection. There is increasing evidence that MSCs will elicit an immune response after a period of engraftment and are ultimately rejected [29]. Whether this affects their clinical utility depends on the intended application. Clinical studies have shown that single administrations of allogeneic MSCs are safe and potentially effective in ischemic heart disease [30]. Preclinical studies have shown that in CKD, repeated injections are often required to ameliorate disease progression [31,32], which may increase the likelihood of sensitization which is less desirable in patients likely to require a kidney transplant. Autologous MSCs may thus be preferable for these patients, provided there is limited disease-related dysfunction as we have demonstrated in the present study.
In the present study, we have focused on an MSC application earlier in the disease course of CKD, with an emphasis on paracrine angiogenic functions. For many regenerative medicine applications, it is paramount to elucidate this, given that secretion of paracrine proangiogenic factors is an  essential mechanism by which MSCs exert their effects [7,[33][34][35]. There were preliminary indications that the paracrine function might not be preserved in CKD. Progenitor cells in experimental CKD display an altered paracrine profile, including reduced expression of vascular endothelial growth factor (VEGF) and reduced migration towards chemotactic stimuli [16,36]. Furthermore, in vitro studies have shown that a uremic environment negatively affects progenitor cell function [37,38]. While cell dysfunction might be resolved through ex vivo culture, in many cardiovascular diseases, MSC dysfunction remains present [17]. Studies on MSCs from CKD patients yielded conflicting results; Yamanaka and colleagues showed that adipose tissue-(AT-) derived MSC displayed reduced proangiogenic effects after implantation of MSC in vivo [39], contrasted by studies by Reinders and colleagues that showed that CKD and healthy MSCs derived from AT and BM produce similar amounts of cytokines essential for the proangiogenic properties of MSCs in vitro [18,19]. Importantly, no functional testing of BM MSC paracrine abilities was performed. Our finding that CKD does not affect the paracrine function of BM MSC CM regarding stimulation of angiogenesis and migration is thus an important addition to the current body of knowledge of MSC function in diseased individuals. These results also highlight that the interindividual variation between MSC function is not necessarily dependent on disease state, similar to our findings in critical limb ischemia. Performance in the tubule formation assay varied between individuals independent of the presence of disease [20]. Given that "healthy" MSCs may not necessarily display sufficient paracrine effector potential, donor screening may be warranted. EVs secreted by MSCs are essential for the paracrine effects of MSCs [40,41]. In preclinical studies, MSCderived EVs indeed contribute to recovery from renal injury and provide protection against chronic and acute kidney injury [42][43][44]. In a small placebo-controlled clinical study, administration of EV from umbilical cord-derived MSCs to CKD patients had a favorable effect on kidney function [13]. It was hitherto unknown whether CKD affects the functional characteristics of MSC-secreted EVs, but it was conceivable, given that stress conditions or disease can alter the content of EVs secreted by cells, which may negatively affect target cells [21,45,46]. However, here we show that CKD MSC-derived EVs display similar effects on migration and tubule formation of endothelial cells as healthy MSCs, which suggests that in CKD patients, the paracrine proangiogenic potential of BM MSC EVs is preserved.
CKD is associated with increased cellular senescence, at the kidney level, but also systemically through uremiainduced toxicity [47,48]. In animal studies, increased senescence was detected in CKD BM MSCs [49]. In contrast, it was previously reported that human healthy and CKD MSCs did not have differences in cytochemical β-galactosidase staining, a marker for senescence [39]. Our study also did not show a difference in senescence between CKD and healthy MSCs, assessed with the more sensitive flow cytometric C12FDG SA-β-gal activity assay, after correction for biological sex. In our study, MSCs from female patients display a higher rate of senescence-associated β-galactosidase activity; and MSCs from female patients also displayed a higher scratch wound migratory capacity. However, this did not affect proliferation, which was similar between females and males. We observed no relations between senescence and any of the studied parameters, most notably age.
Our study has some limitations. Further in vivo studies are needed to confirm our in vitro results. Additionally, there is a biological sex-mismatch in our cohort, which could have influenced the results. However, previous studies did not find sex-specific differences in secretion of proangiogenic and anti-inflammatory cytokines [50,51]. Concomitant medication use also differed between CKD patients and controls, which may have affected MSC function. CKD patients used short-term pretransplant immunosuppression (mycophenolate mofetil (MMF) and prednisone), and statin use was more prevalent in the CKD group. In vitro addition of these drugs inhibits MSC proliferation [52,53], but we did not observe such an effect in CKD MSCs, which makes it less likely that negative effects of CKD on cell proliferation were obscured. Our results are corroborated by a study that assessed growth and other cell characteristics in BM MSCS from patients with graft-versus-host disease (GHVD) versus healthy controls and found no differences, despite the extensive use of immunosuppressive medication in the GVHD group [54]. In subgroup analyses of our data, there was also no correlation between statin use and MSC parameters, though our study does not possess sufficient power to reliably detect such differences with statistical validity. The ex vivo expansion protocol of MSCs used in this study requires over three weeks, in which the cells undergo 8-10 population doublings. It is therefore unlikely that residual medication is still present at the time of the functional experiments described here. As little is known about the effect of comedication of MSC donors on MSC regenerative potential, this warrants further study.

Conclusions
We have shown that CKD does not influence the proangiogenic potential of BM MSC-derived CM and EVs, which is an important step towards autologous BM MSC and BM MSC-derived EV-based regenerative strategies for CKD patients.

Data Availability
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. 116003005). We gratefully thank the study participants. We thank the Department of Vascular Surgery for facilitating this study. Denise Buter, MPA; Laura Michielsen, MD, PhD; Merle Krebber, MSc; and Merel Janssen, BSc, are gratefully acknowledged for technical and/or logistical assistance.