The human kidney capsule contains a functionally distinct mesenchymal stromal cell population

We recently demonstrated that the adult human kidney cortex contains a perivascular stromal cell (kPSC) that shows organotypic properties and is important for repair and stabilisation of kidney function. Not only the kidney cortex but also the kidney capsule contains stromal cells that are important for the three dimensional organisation of the kidney during nephrogenesis. They provide the barrier function of the capsule which is critical for homeostatic processes such as pressure natriuresis. We postulated that stromal cells derived from the kidney capsule may therefore also have specific properties and functions. To this end, we isolated these capsule mesenchymal stromal cells (cMSC) from human cadaveric kidneys that were not suitable for transplantation. There were several similarities between cMSCs and kPSCs including support of vascular plexus formation, phenotypic marker expression and resistance against myofibroblast transformation. However, compared to kPSCs, cMSCs showed distinct mRNA and miRNA expression profiles, showed increased immunosuppressive capacity, and displayed strongly reduced HGF production, contributing to the inability to enhance kidney epithelial repair. Therefore cMSCs are a distinct, novel human kidney-derived MSC-population and these data underpin the large functional diversity of phenotypic similar stromal cells in relation to their anatomic site, even within one organ.


Introduction
The kidney capsule consists of a layer of stromal cells surrounded by a sheet of connective tissue and is of major importance during kidney development and adult kidney homeostasis. During kidney development 3 different locations of stromal cells can be observed; capsular, cortical and medullary stromal cells. Nephrogenesis takes place in the outer nephrogenic zone, the region just below the kidney capsule [1][2][3][4]. Stromal cells within the kidney capsule are of major importance for nephrogenesis. Normally the kidney capsule consists of a continuous layer of Foxd1 and Hox10 positive stromal cells. However, in Foxd1-null embryos, the capsule contains several different cell types including endothelial cells and Bmp4-expressing cells resulting in delayed and disorganized nephrogenesis. Moreover, the defects in capsule formation resulted in adherence to the body wall and failure of kidney ascension which leads to fused kidneys on a pelvic location [5]. Comparable to Foxd1-null mice, triple mutants of the renal stromal marker Hox10 also show a failure of capsule formation with similar effects on nephrogenesis [6].
In the adult kidneys, the kidney capsule is important for kidney homeostasis. The strong layer of connective tissue is of importance for preserving the renal interstitial hydrostatic pressure (RHIP) which is necessary for pressure natriuresis. Decapsulation of the kidney in rats resulted in lower RIHP and reduced pressure-natriuretic response [7]. Moreover, in adult mice, stromal cells from the kidney capsule could be isolated and showed mesenchymal stromal cell (MSC) features such as the ability to adhere to plastic, the presence of MSC-markers such as Sca-1, trilineage differentiation capacity and clonogenicity. Moreover, these murine capsule MSCs were able to migrate into the injured kidney after ischemic injury and seemed to protect against kidney injury as decapsulation of the kidneys resulted in deceleration of recovery of kidney function [8]. However, little is known about the human kidney capsule.
We recently isolated and characterised perivascular stromal cells from the human adult kidney cortex (kPSC) and compared them to mesenchymal stromal cells (MSC) from the bone marrow (bmMSC) [9]. bmMSCs are immunomodulatory cells with reparative properties and have shown beneficial effects in kidney diseases and transplantation [10][11][12]. We showed that, although there are similarities between bmMSCs and kPSCs, there are also major differences. kPSCs show a different, organotypic gene expression, most notably the renal homeobox genes, Hox10 and Hox11, as well as different function. For example, in contrast to bmMSCs, they had the capacity to reintegrate into the renal cortex and stimulated kidney tubular epithelial repair [9]. These data suggest that kPSCs may have retained organotypic properties. In this respect, we were interested to see whether stromal cells isolated from the human renal capsule would show distinct functionalities compared to the cortical kPSCs based on their site-specific role during organogenesis.
Here we show that within the human kidney stromal cells of the kidney capsule share common general characteristics of MSCs, but yet are functionally distinct from the cortical kidney perivascular stromal cells (kPSCs) we previously described.

Materials and methods
Isolation and expansion of human kidney capsule-derived MSCs (cMSC) Cells were isolated from kidney capsules from human transplant grade kidneys not suitable for transplantation due to surgical reasons. Specific research consent was given for all kidneys by either the donor, confirmed by the next of kin or by the next of kin directly according to Dutch legislation. None of the transplant donors were from a vulnerable population. The study was approved by the local medical ethical committee of the Leiden University Medical Centre (p13.054) and the ethical advisory board of the European Union consortium STELLAR. Kidney capsules were carefully removed from the kidney and pericapsular fat was removed. Slices of 1 cm 3 were made and cultured with the kidney side down in a small layer of alpha-MEM (Lonza, Verviers, Belgium) containing 5% platelet lysates (PL), 1 mmol/L glutamine (Lonza) and penicillin/streptomyozine (Lonza) in a T75 cell culture flask until outgrowth culture was obtained. Cultures were maintained at 37˚C and 5% carbon dioxide. Half of the medium was refreshed twice a week. After confluency was reached, cells were collected using trypsin (Lonza) and replated at 4x10 3 cells/cm 2 . Experiments were performed with cMSCs between passage 4-8. The population doubling was calculated with the following formula: Cumulative PD = Slog(N1/N0). PD is population doubling, N1 the accumulated cell number at the end of incubation and N0 the accumulated cell number at the beginning of incubation.

Isolation and expansion of human kidney cortex perivascular stromal cells (kPSC)
kPSCs were isolated from human transplant grade kidneys as described previously [9]. Shortly, after removal of the kidney capsule for cMSC isolation, kidneys were perfused via the renal artery with collagenase (2500 units, NB1, Serva) and DNAse (2,5ml Pulmozyme, Genetech) at 37˚C with a flow of 100ml/min. After approximately 30 minutes, the tissue was digested and the resulting cell suspension was washed and collected. Cells were either directly cultured or frozen in liquid nitrogen. Kidney cell suspensions were cultured in alphaMEM (Lonza) containing 5% platelet lysates, glutamine (Lonza) and penicillin/streptomyzine (lonza). At passage 1 NG2 cell enrichment was performed using MACS according to manufacturer's protocol (Miltenyi Biotech, Gladbach, Germany) and afterwards cells were cultured in alphaMEM containing 5% platelet lysates. Experiments were performed between passage 4-8. bmMSCs were isolated as described previously [9].

Trilineage differentiation potential
cMSCs and kPSCs were cultured in adipogenic, osteogenic and chondrogenic medium according to manufacturer's protocol (Lonza). Calcium deposits were shown with Alizarin Red in the osteogenic differentiation assay. In the adipogenic differentiation assay, lipid droplets were stained using Oil Red O. In the chondrogenic differentiation assay, cell pellets were formalinfixed (4% PFA o/n) and embedded in paraffin. 5 μm sections were de-parafinnized, rehydrated and incubated with 1% toluidine blue for 20 minutes. All differentiations were analysed with an inverted bright-field microscope (Leica DFC 295).

Microarray sample preparation and data analysis
RNA was isolated from biological triplicates of both cMSCs and kPSCs of the same donor using Trizol reagent (Life Technologies, Bleiswijk, the Netherlands) and the RNeasy kit (Qiagen, Heidelberg, Germany) according to the manufacturer's protocols and as described previously [9]. RNA quality and quantity were assessed using the Bioanalyzer (Agilent Technologies, Santa Clara USA) and Nanodrop spectrophotometer (Nanodrop Technologies, Wesington, USA). Gene expression profiling was performed by Aros Applied Biotechnology (Aarhus, Denmark). cDNA and cRNA synthesis, labelling and hybridization on the Human HT12 V4 Gene Expression Beadchips (Illumina Inc., San Diego, USA) were performed according to manufacturer's instructions. The beadchips were scanned using iSCAN system (Illumina Inc., San Diego, USA) and fluorescence intensities were uploaded into GenomeStudio Software (Illumina Inc., San Diego, USA). Genes with a detection p-value of >0.05 for all samples were excluded. Subsequent data was quantile normalized and the Pearson's correlation coefficient was calculated (r 2 ). Average signals>200 in either the kPSCs or cMSCs were considered above background levels. Differential expression was analysed using the gene expression module of Genome Studio (Illumina). For the table of the top differential expressed genes, genes were sorted on highest DiffScore (Illumina). For the heatmap the delta average signal of cMSCs vs. kPSCs was set to 200 (which is the threshold of the measurement) and the resulting 3034 genes selected were analysed for clustering in R software.

miRNA profiling
For miRNA-profiling, reverse transcription of total RNA of cMSCs and kPSCs of 3 donors was performed using the miRNA reverse transcription kit (Thermo Biosystems, Foster City, CA). cDNA was preamplified using Megaplex PreAmp primers pools A V2.1 (Thermo Biosystems) according to manufacturer's protocol. 384 miRNAs including six controls (RNU44, RNU48, 4 Ã U6) were profiled using TaqMan 1 Array MicroRNA Human Card A V2.0 (Thermo Biosystems). Samples were normalized to RNU48 expression. The heatmap of 2 40-CT values are shown for the top 250 differentially expressed microRNAs.

Peripheral blood mononuclear cell isolation, proliferation assay and cytokine analysis
Peripheral blood mononuclear cells (PBMCs) were isolated from buffy coats of healthy blood donors obtained from the local blood bank by density gradient centrifugation using Ficoll-isopaque and frozen in liquid nitrogen until use. cMSCs of 3 donors and kPSCs of 2 donors were plated in triplo in flat-bottom 96-well plates (Costar, Sigma-Aldrich) and allowed to attach overnight in DMEM-F12 with 10% normal human serum (NHS) as described previously [9]. cMSCs and kPSCs were seeded in increasing cell numbers to eventually get a ratio of 1:128 (MSC:PBMC) (766 cells) to 1:4 (25.000 cells). PBMCs were stimulated with anti-CD3/ antiCD28 Dynabeads (Invitrogen) and added to the kPSCs/cMSCs in a concentration of 1x10 5 cells/well. After 5 days the supernatant was collected and 3 H-thymidine (0.5 mCi) was added. After 16 hours 3 H-thymidine incorporation was determined as a measure of proliferation. Cytokine excretion was analysed in the supernatant with the Bio-Plex Human Cytokine 17-Plex assay following manufacturer's protocols (Bio-Rad Laboratories, Veenendaal, the Netherlands).
Kidney epithelial scratch assay cMSCs and kPSCs of 3 different donors were plated in a density of 200.000 cells/well in a 6 well culture plate (Costar, Sigma-Aldrich) and cultured for 48 hours [9]. Subsequently, the supernatant was collected for the use as conditioned medium. Growth factors in the conditioned medium of cMSCs and kPSCs were measured using a custom made growth factor multiplex-elisa panel following manufacturer's instruction. (R&D systems, Minneapolis, USA). Proximal tubular epithelial cells (PTEC), HK-2 [14] were seeded in PTEC medium consisting of a Dulbecco's modified Eagle's medium (DMEM) F-12 (Lonza) supplemented with insulin (5μg/ml), transferrin (5μg/ml), selenium (5ng/ml), hydrocortisone (36 ng/ml), tr-iodothyrinine (40 pg/ml) and epidermal growth factor (10 ng/ml) (Sigma-Aldrich) in a density of 500.000 cells/well in a 6 wells cell culture plate (Costar) and cultured until confluent. A scratch wound was created in the monolayer of HK-2 cells using a 200μl pipette tip. After the scratch, cells were washed with PBS and provided with either fresh medium (alphaMEM+5%PL) or with complete conditioned medium from either cMSCs or kPSCs. Scratches were imaged at 3, 6, 12 and 24 hours at the same position in duplicates with an inverted bright-field microscope (Leica DFC 295). The scratch area was measured at each time point using ImageJ software and the percentage wound closure was calculated.

Statistical analysis
Statistical analysis was performed with Graph Pad Prism (Graph Pad Prism Software Incl. San Diego, USA). Differences between kPSCs and cMSCs were analysed using either an unpaired two-tailed t-test or, when more variables are analysed a two way ANOVA with Bonferroni's posthoc comparison analysis. Differences were considered statistically significant when p<0.05.

MSC-like cells can be isolated from the human kidney capsule
In order to analyse the presence and phenotype of stromal cells in the adult human kidney capsule, we stained the kidney capsule for the stromal markers CD73, CD105, PDGFR-β and α-SMA. All these markers are expressed in both the kidney capsule and in the perivascular regions of the cortex (Fig 1A). Via outgrowth culture of the kidney capsule on cell culture plastic, spindle shaped cells could be obtained (cMSC) (Fig 1B). Maker expression of cMSCs isolated from three donors were compared to the kidney perivascular stromal cells (kPSC) sorted from the same kidneys based on NG2-expression as we described previously [9]. cMSCs were, like kPSCs positive for the surface markers CD73, CD90, CD105 and HLA-ABC while being negative for CD31, CD34, CD45 and HLA-DR (Fig 2A). This marker expression was robust and similar to the marker expression of kPSCs isolated from the same donors (Fig 2B). cMSCs showed chondrogenic, osteogenic and adipogenic differentiation capacity, which is different from kPSCs which lack adipogenic differentiation (Fig 2C). In contrast to kPSCs, there was no NG2-enrichment step needed to isolate cMSCs. As shown in Fig 2D, there is, as observed with other MSC populations, donor variation in growth characteristics in both population doubling and in the moment of reaching senescence. Senescence was observed between 50-70 days of culture.

cMSCs show a distinct mRNA and miRNA expression profile
In order to evaluate the differences between cMSCs and kPSCs in more detail, Illumina microarray expression profiling was performed on biological triplicates. kPSCs and cMSCs were derived from the same donors. 35000 transcripts were analysed and cMSCs and kPSCs had a similar expression profile of most genes with a Pearson correlation coefficient of 0,9594 ( Fig  3A), suggesting high similarities between the cell types. However, still approximately 3000 genes were differentially expressed. Hierarchical clustering showed clustering between cMSCs and kPSCs indicating more similarities between cMSCs of different donors, than between cMSCs and kPSCs of the same donor (Fig 3B). This suggests that cMSC and kPSC are distinct cell populations. Table 1 shows the top 5 up and down regulated genes comparing the two cell types. The top 50 differentially expressed genes can be found in the supporting information (S1 Table) Expression profiles of cMSCs were also compared to those of bmMSCs and also here cMSCs and bmMSCs clustered according to cell source (S1A Fig). Similar results were observed for the expression profile of miRNAs. Most miRNAs showed a similar expression profile with a Pearson correlation coefficient of 0.8614. However, 114 out of 315 miRNAs showed a fold increase of 2 or more (S2 Table) and again hierarchical clustering was observed according to cell source and not according to donor as depicted in the heatmap (Fig 3D) The heatmap of cMSCs compared to bmMSCs can be found in the supporting information (S1B Fig). cMSCs did not transform into myofibroblasts after stimulation with TGF-β As the kidney capsule contains a strong fibrous layer we evaluated whether cMSCs have the capacity to transform into myofibroblasts. However, comparable to kPSCs, cMSCs did not show an increased a-SMA expression after stimulation with TGF-β (Fig 4A). This indicates that cMSCs do not transform into myofibroblasts upon in vitro stimulation. This is in contrast to bmMSCs for which we previously showed that they do increase α-SMA expression upon stimulation [9].

Vascular network formation was stabilized by coculture of human glomerular endothelial cells with cMSCs and kPSCs
Due to their perivascular location, we hypothesized that kPSCs are able to stabilize endothelial cells to a greater extent than cMSCs. When looking at angiogenic growth factors, kPSCs indeed excrete vascular endothelial growth factor (VEGF) in higher levels than cMSCs. No differences were observed in angiopoietin 1 (ang1) excretion (Fig 4B). To assess the functional consequences, cMSCs and kPSCs were cocultured with human glomerular micro vascular endothelial cells (hgMVECs). hgMVECs alone were not able to form vascular networks, however, coculture with either cMSCs or kPSCs gave rise to the formation of vascular tubular structures in a vascular network (Fig 4C). There were no differences in the amount of vascular tubular structures as quantified in Fig 4D. cMSCs are more immunosuppressive compared to kPSCs To evaluate the immunosuppressive capacity of cMSCs a peripheral blood mononuclear cell (PBMC) suppression assay was performed. In this assay, PBMCs were activated with polyclonal CD3/CD28 activation in the absence or presence of increasing amounts of cMSCs or kPSCs. While kPSCs were only able to inhibit PBMC proliferation at high MSC:PBMC ratios (1:8 and 1:4), cMSCs already showed an effect at lower cell numbers and showed significantly more inhibition than kPSCs (Fig 5A). This is to a similar extent compared to the inhibition by bmMSCs (S2A Fig). Moreover, the excretion of the pro-inflammatory cytokine IFN-y by the activated PBMCs was lower in coculture with cMSCs than in coculture with kPSCs indicating more immunosuppression by the cMSCs (Fig 5B). When looking at general cytokine excretion   profiles, unstimulated cMSCs and kPSCs excrete similar levels of major cytokines (Fig 5C). However, the secretome of unstimulated MSCs does not reflect their functionality while the decrease in IFN-y excretion by PBMCs does.
cMSCs were not able to enhance human kidney epithelial repair We previously showed that the conditioned medium of kPSCs was, in contrast to the conditioned medium of bmMSCs, able to accelerate kidney tubular epithelial repair in a kidney epithelial wound scratch assay and that this effect was HGF depended [9]. When looking at major growth factors, most growth factors showed similar excretion profiles, with the exception of HGF. HGF was excreted in a more than 200 fold lower concentration by cMSCs compared to kPSCs. This also resulted in a decreased epithelial wound healing by the conditioned medium of cMSCs as shown in Fig 5E and

Discussion
Previously it has been shown that from most organs MSC-like cells can be isolated. These organ-derived MSC like cells are mainly isolated from the perivascular compartment and exhibit tissue specific properties [15,16]. From the human kidney cortex perivascular stromal cells could also be isolated and these cells showed kidney-specific regeneration and integration properties when compared to bmMSCs [9,17].
Here we show that within the human kidney, not only in the cortical perivascular compartment of the kidney but also in the kidney capsule, a MSC-like cell population is present. Although there were similarities between the cell types in for example marker expression and vascular stabilisation properties, there were also major differences. Though isolated from the same donors, cMSCs showed a distinct mRNA and miRNA expression profile with clustering according to cell type and not to donor. Moreover, cMSCs showed increased immunosuppressive capacities while lacking the properties to accelerate kidney epithelial wound healing. Thus, cMSCs are a distinct kidney-derived MSC-like cell population.
Only few studies have been performed comparing MSC-like cells isolated from one tissue. Functional differences have been described between different isolation sides of adipose tissue [18,19]. Naftali-Shani et al., for example, showed that MSCs isolated from the pericardial fat, epicardial fat and subcutaneous fat showed distinct cytokine excretion profiles, mRNA expression profiles and reparative properties. Interestingly and unexpectedly, cardiac dysfunction after myocardial infarction was worst after transplantation of MSCs from the epicardial fat compared to subcutaneous fat MSC transplantation and control [18].
MSCs are usually studied as candidate cells for cellular therapies. Although the in-organ differential functionality of MSCs are of interest to our understanding of the structural biology of the kidney, cMSCs are less likely to be used as candidate for clinical therapies. In clinical trials usually 2 cell infusions of 1-2 million cells per kilogram body weight are given [12]. To obtain such cells numbers, cMSCs should be isolated from several different donors for infusion into one patient, which makes the use of cMSCs less feasible compared to other MSC sources. This is the first description of the functional characteristics of human kidney capsulederived MSCs. Moreover, this is the first extensive functional comparison between kidney cortex derived and kidney capsule-derived MSC-like cells. The differences observed between the cells types in gene expression profile and functionality may reflect site-specific differentiation during organogenesis.

Conclusion
cMSCs are a distinct, novel human kidney-derived MSC-population which underpins the large functional diversity of phenotypic similar stromal cells in relation to their anatomic site, even within one organ.  Fig. A) Within a PBMC proliferation assay cMSCs were able to supress PBMC proliferation significantly more than kPSCs and to a similar extent as bmMSCs. B) Both bmMSCs and cMSCs were not able to accelerate epithelial wound healing while kPSCs did show this potential. Abbreviations: cMSC: human kidney capsule-derived mesenchymal stromal cell, kPSC: kidney cortex derived perivascular stromal cell, act.: activated, PBMC: peripheral blood mononuclear cells, n.s non significant, Ã p<0.05, ÃÃ p<0.01, ÃÃÃ p<0.001. (TIF) S1 Table. Top 50 differentially expressed genes between kPSCs and cMSCs. Quantile normalized average signals are shown. Abbreviations: kPSC: kidney cortex perivascular stromal cell, cMSC: kidney capsule-derived mesenchymal stromal cell. Reinders, Hetty C. de Boer, Cees van Kooten, Ton J. Rabelink.