The Functions of Clusterin Expression in Renal Mesenchymal Stromal Cells: Regulation of Cell Proliferation and Macrophage Activation

The expression of clusterin (CLU) in mice increases resistance to renal ischemia-reperfusion injury and promotes renal tissue repair. However, the mechanisms underlying of the renal protection of CLU remain largely unknown. Mesenchymal stromal cells (MSCs), found in different compartments of the kidney, may contribute to kidney cell turnover and injury repair. This study investigated the in vitro functions of CLU in kidney mesenchymal stromal cells (KMSCs). KMSCs were isolated by digestion of kidney tissues with collagenase type 1 and growth in plastic culture plates. Cell surface markers, apoptosis and phagocytosis were determined by ow cytometry, and CLU protein by Western blot.


Isolation of microvesicles
The monolayers (80% con uence) of KMSCs were grown in the DMEM medium containing 5% of pooled serum from CLU KO mice at 37°C in the culture incubator with 5% CO 2 for 4 days, followed by harvesting the conditioned medium. A two-step different centrifugation was used for the isolation of the microvesicles from the conditional medium. The rst step was to pellet dead cells and large cell debris by the centrifugations of the samples at 2,000 ×g for 20 min, and the resultant supernatants were collected and transferred to new tubes. Then the supernatants were ultracentrifuged at 20,000 ×g for 30 min at 4°C to pellet the microvesicles. The microvesicle pellets were washed with ice-cold PBS to remove proteins. The protein fraction in the supernatants from both centrifugation steps was precipitated by 80% saturated ammonium sulfate solution. All the pellets and protein fractions of the supernatants were collected for Western blot analysis of CLU protein.
Trilineage differentiation to chondrocytes, osteocytes or adipocytes KMSCs were harvested from cultures after P5, and their potential for differentiation to chondrocytes, osteocytes or adipocytes was examined as previously described by Liu et al. [19] . The chondrogenic differentiation of KMSCs was con rmed by positive staining of cartilage formation in the cells with 1% of acidic Alcian blue in 80% methanol (v/v) (pH 2.5), the osteogenic differentiation was con rmed by positive staining of Ca2 + matrix mineralization with 2% Alizarin red S in 0.5% NH 4 OH (pH 4.2), and the adipogenic differentiation was examined by the presence of intracellular lipid droplets stained with Oil red O.

Hypoxia and Western blot assay
KMSCs were incubated in the O 2 -Control-InVitro-Glove-Box (Coy Laboratory, Michigan, USA) under hypoxic condition (1% O 2 , 5% CO 2 and 94% N 2 ) for different periods of time (day 0 to day3). The total cellular protein was extracted from the KMSCs at these time points, and CLU expression was examined by Western blot. Brie y, protein samples (50 µg/sample) were fractionated by 10% SDS-PAGE and then transferred onto nitrocellulose membrane. The CLU protein in the blot was identi ed with primary goat polyclonal anti-CLU antibody (Santa Cruz Biotechnology, Santa Cruz, CA) and secondary donkey anti-goat IgG H&L-Alexa Fluor® 680 (Abcam). The CLU protein-antibodies was viewed by using Odyssey (LI-COR, Inc., Lincoln, Nebraska, USA). Blots were reprobed using β-actin (Sigma-Aldrich Canada) for con rmation of loaded protein in each sample.

Determination of apoptosis
Apoptosis of KMSCs under the hypoxia for three days was measured by FACS analysis with Annexin-V-PE for staining of early apoptosis and 7-Aminoactinomycin D (7-AAD for late apoptosis as described before [17] . Brie y, KMSCs were cultured under the hypoxia as mentioned above for three days, and were harvested by a brief incubation with trypsin-EDTA solution (Sigma-Aldrich). The apoptotic cells were stained with Annexin-V-PE and 7-ADD in 1×binding buffer for 15 min. The intensity of staining uorescence was measured by a FACS and analyzed as compared with background controls using FlowJo software (Tree Star Inc.).

MTT ASSAY
The MTT assay is a colorimetric method of speci cally measuring the activity of NAD Inc.), respectively.

Phagocytosis Assay
The phagocytic activity of RAW264.7 cells was measured by using uorescence-labeled latex beads (Sigma-Aldrich Canada, L-3030) as described previously [20] . In brief, RAW264.7 cells in 24-well plates were treated with the supernatant from KMSCs cultures for 24 h as described above, followed by incubation with 2.0 μL of uorescent latex beads per well for 2 h. Phagocytosis was determined by the number of cells engul ng beads, which was measured by using the FACS and was presented as the increased intensity of uorescence (%) in RAW264.7 cells as compared to background controls (no beads) using FlowJo software (Tree Star Inc.).

Statistical analysis
Data were collected from at least three separate experiments in each study, and were presented as mean ± standard deviation (SD) for each group in the text. The difference between groups was analyzed using t-tests or analysis of variance (ANOVA) as appropriate with GraphPad Prism 7 (GraphPad Software, Inc., La Jolla, CA, USA). A p value of ≤0.05 was considered signi cant.

CLU expression in KMSCs from WT mice
The expression of CLU, both intracellular and secreted CLU proteins, in cultured KMSCs was examined by Western blot. As shown in Fig. 1A and B, the KMSCs isolated from WT mice (WT KMSCs) expressed the intracellular CLU in the cellular protein extracts and secreted mature CLU (mCLU) in the culture supernatants, whereas these protein bands were absent in KMSCs from CLU KO mice. In addition, the intracellular CLU -CLU precursor (Pre-CLU) in WT KMSCs was upregulated in response to hypoxia (Fig.   1A).
To verify if mCLU presented in the microvesicles secreted from the KMSCs, the monolayer of KMSCs was grown in the DMEM medium containing 5% pooled serum from CLU KO mice in a 5% CO 2 incubator at 37℃ for 4 days, and the secreted microvesicles in the culture medium were isolated by centrifugation at 20,000 ×g. As shown in Fig. 1C, the presence of CLU protein was seen in cellular protein extracts, dead cells or cell debris and supernatants, but not in microvesicle fractions. The sizes and morphology of the isolated microvesicles from WT KMSCs were not different from those from KO MSCs (Fig. 1D). These data might suggest that mCLU was not associated with the secreted microvesicles but a soluble factor secreted from the KMSCs.
The effect of CLU expression on the morphology of KMSCs in plastic culture dishes The minimal criteria for de ning MSCs include physical adherence to plastic culture dishes, broblast-like morphology in culture, positive and negative expression of speci c cell surface markers and multipotential for differentiation to chondrocytes, osteocytes or adipocytes [21,22] . The morphology and adherent property of KMSCs from CLU KO mice as compared to WT mice in the plastic culture dishes was examined by using a microscope. As shown in Fig. 2, there was little difference of morphologic changes and physical adherence to the plastic culture dishes between WT and KO KMSCs in cultures from P1 to P15. In P1 cultures after 2 weeks of incubation, there were some cells in both groups that exhibited broblast-like spindle shape, and after P4 the number of cells with spindle-shaped broblast-like morphology was gradually increased, and these cells were propagated well in the plastic culture dishes and reached a signi cant level after the P5. These data suggested that the KMSCs from CLU KO mice were not different from WT mice in the morphology and adherent property.

The effect of CLU expression on the expression of cell surface markers of KMSCs
The MSC-speci c cell surface markers of KMSCs at P4 from both WT and CLU KO mice were examined by using FACS analysis. As shown in Fig. 3, the KMSCs from both groups were strong positive in the expression of CD133, Sca-1, CD44, and CD117, and negative in CD45, CD163, CD41, CD276, CD138, CD79a ( Fig. 3; Table 1). As compared to the background staining, there was a small proportion (5% -10% percentage) of the cells from both WT and KO KMSCs expressing CD34, suggesting that a small population of KMSC might weakly express this hematopoietic stem cell marker or the cells from hematopoietic system. These data suggested that the KMSCs from CLU KO mice were not different from WT mice in the expression of MSC-speci c cell surface markers.
The effect of CLU on the multipotential of KMSCs for trilineage differentiation The multipotential of KMSCs from CLU KO mice as compared to WT mice for their differentiation to chondrocytes, osteocytes or adipocytes was examined by using standard methods as described previously [19] . As shown in Fig. 4, the differentiation of KMSCs of both groups to adipocytes was seen after incubation with adipocygenic differentiation medium for 4 weeks, in which the lipid droplets in differentiated cells (adipocytes) were con rmed by oil red O-staining (right column). For osteogenic differentiation, a con uent monolayer with some "clustering" of osteocytes was observed after incubation with osteogenic differentiation medium for 4 weeks, indicated by alizarin red S staining of the extracellular calcium deposits (middle column). Similarly, the KMSCs from WT and CLU KO mice were differentiated to the phenotype of chondrocytes, indicated by the presence of Alcian blue-stained cartilage matrix (right column). Further, the chondrocytes in WT group were clustering but not in KO group. These data indicated that KMSCs from CLU KO mice had equal capability as WT groups of differentiating to the chondrocytes, osteocytes or adipocytes.
The effect of CLU on KMSCs survival in response to hypoxia Hypoxic injury is an initial insult of IRI, and expression of CLU increases the survival of kidney tubular epithelial cells in hypoxia [17,23] Here, the impact of CLU expression on KMSCs survival in hypoxia was investigated. The KMSCs from WT and CLU KO mice were incubated in hypoxia for three days, and cell survival or death was examined by both microscopy and FACS analyses. Microscopic examination revealed that there were barely dead cells in the KMSCs cultures of both group (Figs. 5b, 5d). Further FACS analysis con rmed the microscopic observation, indicated only (9.179 ± 0.6757, n = 8) and KO (10.02 ± 1.102, n = 8), which were not signi cantly different (Fig. 5e). These data suggest that a lack of CLU expression did not affect the survival of KMSCs in the hypoxic environment.
The effect of CLU on KMSCs proliferation in vitro The difference of the cellular proliferation between WT and KO KMSCs was examined using MTT assay. As shown in Fig 6, as compared to WT group, the growth rates of KO group were signi cantly slower (WT vs. KO: p = 0.0174, two-way ANOVA), especially at day 4 (p = 0.0189, two-way ANOVA multiple comparisons), day 5 (p = 0.0118, two-way ANOVA multiple comparisons), and day 6 (<0.0001, two-way ANOVA multiple comparisons). Taken together, these data imply that the lack of CLU expression signi cantly reduced the proliferation of KMSCs.
The effect of KMSCs-secreted CLU on regulation of phenotypes and phagocytosis of macrophages MSC-sourced secretome consists of MSC-derived bioactive factors (soluble proteins, nucleic acids, lipids and extracellular microvesicles) secreted to the extracellular space, and mediates the biological functions of MSCs [24,25] . Here, the different effects of the secretome from KMSCs from CLU KO mice compared to WT mice on the phenotype of macrophage was examined by using FACS (Fig. 7). As shown in Fig. 7b The phagocytosis of macrophages affected by the supernatant or secretome from KO KMSCs as compared to that from WT KMSCs was examined by using FACS (Fig. 8). As shown in Fig. 8b, the supernatant or secretome from WT KMSC enhanced more phagocytosis of the macrophages than that from KO KMSCs, as indicated by 18.98 ± 3.379% in WT compared to 8.043 ± 1.052% in KO group (p = 0.005), and there was no signi cant difference between KO and normal control group (7.228 ± 0.314 %). Taken together, these data suggest that the supernatant or secretome from KMSCs of WT but not from CLU KO increased the phagocytosis function of macrophage.

Discussion
Multipotent kidney progenitor cells or KMSCs have been identi ed in different anatomical locations of adult kidneys of both humans and rodents [26,27] , including nontubular MSCs in renal interstitial space of mouse adult kidneys [4,28] and progenitor-like cells in S3 segment of nephron in rat adult kidneys [29] . These KMSCs have been found to contribute to the regeneration or repair of renal endothelium and tubules after renal IRI [28,30] . However, the molecular mechanisms by which KMSCs repair the kidney injury are not fully understood. We have previously demonstrated that as compared to CLU KO mice, CLU expression in mice is required for tissue repair and suppresses macrophages in ltration and proin ammatory M1 polarization in the kidney after IRI [16,31] . In this study, we isolated KMSCs from the kidneys of both WT and CLU KO mice. There were not differences in the trilineage differentiation and survival under hypoxic condition between WT and KO KMSCs. The WT KMSCs secreted CLU protein as a soluble factor that was not associated with secreted microvesicles. As compared to KO KMSCs, CLUexpressing KMSCs proliferated faster in culture, and the CLU-containing conditional medium from the KMSCs promoted anti-in ammatory M2 polarization and phagocytosis of macrophages.
MSCs are a heterogeneous population of broblast-like cells in culture and can be isolated from all different tissues in the body [32] . The essential standards for phenotypic and functional characterization of these cells have been initially recommended by the International Society for Cellular Therapy (ISCT); MSCs are plastic-adherent, have trilineage differentiation capacity (adipogenic, osteogenic, and chondrogenic) and express a cell surface pro le of CD73 + , CD90 + , CD105 + , CD11b − /CD14 − , CD19 − /CD79a − , CD34 − , CD45 − , and HLA-DR − [22] In addition, the most common positive markers of CD29, CD44, CD90, and CD105 are recommended especially for adult MSCs in a recent systemic review [33] and the most prominent negative markers are CD34, CD45, and CD14 [33] .. Similarly, the mouse kidney-derived MSCs or KMSCs have reported to be positive for CD29, CD44, Sca-1, CD73, CD105, and CD117 and negative for hematopoietic markers (e.g. CD11b, CD19, and CD45) [34][35][36] . In the present study, the plasticadherent, spindle-shaped and broblast-like cells were isolated from the kidneys of both WT and CLU KO mice by culturing in plastic culture plates (Fig.1). These cells from either WT or KO mice were highly expressed CD133 + (Prom1), Sca-1(Ly6a) + , CD44 + and CD117 + (c-kit)( Fig. 2 and Table 1), which are also de ned as prostate cancer stem cell phenotype [37] . In addition, these cells were almost negative in the expression of CD45, CD163, CD41, CD276, CD138 and CD79a, and the low expression of CD34, indicating that a small fraction of the isolated KMSCs were hematopoietic cells which represented "contamination" from intrarenal blood cells in our cultures. However, the ''universal'' stem cell markers such as CD133, CD24, Sca-1, and CD117 are also positively expressed in differentiated epithelia, including renal epithelia [38,39] . CD133 was the most common marker for the phenotype of stem cells which provide protection in acute tubular and glomerular damage [40,41] , and it is also expressed in many different cell types, suggesting that a single marker is probably insu cient to identify a particular stem cell type [42] . In fact, one study has documented that both CD133 + and CD133 − cells improve renal function and promote renal regeneration to a similar degree [43] . The other studies indicate that the CD133 + /CD24 + cells form both multicellular spheroids (nephrospheres) and branched tubule-like structures and undergo neurogenic, adipogenic, osteogenic differentiation, whereas the CD133 − /CD24 + cells are unable to form these structures and unable to grow in some of the specialized differentiation media [44] . Indeed, there is no consensus among researchers regarding speci c markers (alone or in combination) for identifying or isolating MSCs from the kidney.
In humans, there are two isoforms of CLU from the initial protein precursor, secretory CLU (sCLU, 75-80 kDa) and nuclear CLU (nCLU, 55 kDa), that differs in their functions [45,46] . The sCLU is a major cytoprotective factor, while the nCLU is a cell death protein [45,46] . However, only secretory CLU has been identi ed in mice [47] . Therefore, the biological functions of CLU in mice may only re ect those of sCLU in humans. Microvesicles secreted by MSCs are an important vehicle to transfer the proteins, mRNA and miRNA from the MSCs to target cells, by which the target cell proliferation and differentiation are regulated [48] . In this study, we con rmed that CLU secreted by KMSCs was a soluble factor but not with the microvesicles (Fig. 1), suggesting that the functions of extracellular CLU outside the KMSCs are not mediated by the microvesicles. Further, a lack of CLU expression did not signi cantly affect the differentiation capacity of KMSCs to adipocytes, osteocytes, and chondrocytes in vitro (Fig. 4). However, recombinant sCLU protein treatment of mouse MSCs from the bone marrow inhibits the osteoblast differentiation while stimulates adipocyte differentiation [12] . The difference between our observations and this study may be due to the fact that the addition of recombinant sCLU and endogenous complete CLU gene knockout or only endogenous CLU may have different effects on trilineage differentiation. In this study, we also noticed that the lack of CLU expression obviously inhibited the "clustering" formation of the differentiated chondrocytes (Fig. 4). This nding is consistent with the ability of CLU to elicit clustering of Sertoli cells in an early study [49] and to suppress tubular epithelial cell migration [16,50] , suggesting that the differentiated chondrocytes secret CLU protein that is required for cluster formation via an autocrine manner, which however remains further investigation.
CLU expression is up-regulated in normal human diploid broblasts exposed to hypoxic conditions (1.5% O 2 or CoCl 2 ) [51] . Similar results have been seen in human kidney tubular epithelial cells (HKC-8) [23] and WT KMSCs (Fig. 1). The expression of CLU protein protects kidney epithelial cells from hypoxia-induced cells [23,50] ; however, it has no effect on KMSCs viability or death under prolonged hypoxia exposure (3 days) as no difference of cell death/viability between WT and KO KMSCs was seen (Fig. 5). The reasons why CLU cytoprotective activity against hypoxia is different between tubular epithelial cells and KMSCs remain unknown. It has been well known that hypoxia is a critical component for the maintenance of undifferentiated and slow-cycling proliferation states of a broader spectrum of stem cell niche [52,53] , and in the renal stem cell niche of papilla of adult kidneys, the oxygen tensions are less than 1% (4-10 mmHg) [54] . These ndings simply imply that KMSCs may have unique anti-hypoxia feature for living in hypoxic microenvironment, as a result, the up-regulated CLU did not provide additional protection against hypoxia. These observations are supported by the fact that the renal progenitor cells have higher resistance to injury in comparison to all other differentiated cells of the kidney [52] .
Recombinant sCLU protein stimulates the proliferation of mouse bone marrow-derived MSC [12] . Our previous studies demonstrated that the role of CLU in renal repair and tubular cell proliferation in culture is associated with up-regulation of a panel of genes that positively regulate cell cycle progression and DNA damage repair, which may promote cell proliferation but not involves epithelial cell migration [16,50] .
Here, the proliferation of KMSCs from WT and KO also been investigated and the results showed that CLU de ciency signi cantly inhibited the proliferation capacity of KMSCs (Fig. 6). Further, we observed the CLU gradually up-regulated in hypoxia (1% O 2 ) at day1 and day2 (Fig. 1), which is consistent with the faster proliferation of WT KMSCs under hypoxic conditions (Supplement data: Fig. 1). All of these data imply that CLU participates in the regulation of proliferation of KMSCs, which may contribute to kidney repair -more proliferating cells in WT than in CLU KO kidneys [16] .
The renoprotective activity of MSCs is associated with their secreted cytokines, growth factors and other molecules that inhibit in ammation and brosis, and promote endogenous repair processes [42] . In the present study, the CLU-containing conditional medium from WT KMSCs signi cantly upregulated the M2 marker CD206 expression and not affected the M1 marker CD80 expression in macrophages (Fig.7) and increased the phagocytosis of macrophages (Fig. 8). These results are consistent with our observations in vivo -the kidney repair in WT is associated with more in ltrating M2 macrophages [31] and are supported by that MSCs ameliorate acute kidney injury via the activation of macrophages to a trophic M2 phenotype [55] . It has been known that the M2 macrophages have higher phagocytosis capacity for clearance of apoptotic cells, produce extracellular matrix, mitigate in ammatory response, and promote wound healing [56,57] , suggesting that KMSCs is immunosuppressive mediators by upregulated the M2 polarization but not the M1 polarization through a paracrine action.
One has to acknowledge the limitations of this study. This study only focussed on the role of CLU in some characteristics (e.g. differentiation, survival, proliferation, and macrophage regulation) of KMSCs in culture. The biological functions of CLU in KMSCs in vivo were not investigated. We also did not investigate the molecular pathways by which CLU functions in KMSCs.

Conclusions
CLU KO results in more severe renal tissue damage and impairs renal repair after IRI [16,17] and accelerates renal brosis in response to IRI and UUO [13,14] . The present study indicates that CLU de ciency inhibits the proliferation of KMSCs and hinders the polarization-inducing ability of M2 macrophages and the ability to induce macrophage phagocytosis. Our data imply that CLU may play an important regulatory role in the involvement of KMSCs in various diseases and in ammatory injuries.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.