Clusterin Is a Potential Lymphotoxin Beta Receptor Target That Is Upregulated and Accumulates in Germinal Centers of Mouse Spleen during Immune Response

Clusterin is a multifunctional protein that participates in tissue remodeling, apoptosis, lipid transport, complement-mediated cell lysis and serves as an extracellular chaperone. The role of clusterin in cancer and neurodegeneration has been extensively studied, however little is known about its functions in the immune system. Using expression profiling we found that clusterin mRNA is considerably down-regulated in mouse spleen stroma upon knock-out of lymphotoxin β receptor which plays pivotal role in secondary lymphoid organ development, maintenance and function. Using immunohistochemistry and western blot we studied clusterin protein level and distribution in mouse spleen and mesenteric lymph nodes in steady state and upon immunization with sheep red blood cells. We showed that clusterin protein, represented mainly by the secreted heterodimeric form, is present in all stromal compartments of secondary lymphoid organs except for marginal reticular cells. Clusterin protein level rose after immunization and accumulated in light zones of germinal centers in spleen - the effect that was not observed in lymph nodes. Regulation of clusterin expression by the lymphotoxin beta signaling pathway and its protein dynamics during immune response suggest a specific role of this enigmatic protein in the immune system that needs further study.


Introduction
Lymphotoxin beta receptor (LTbR) signaling plays a crucial role in development of secondary lymphoid organs (SLO). Surface lymphotoxin (LT) is a transmembrane heterotrimeric protein that belongs to the tumor necrosis factor (TNF) family and is expressed by lymphoid tissue inducer cells during early phases of SLO formation [1]. Acting through LTbR on lymphoid tissue organizer cells and earlier on their mesenchimal precursors, it activates synthesis of chemokines, adhesion molecules and lymphangiogenic factors through classical and alternative NFkB pathways, leading to maturation of stroma and lymphocyte homing [2,3,4]. In postnatal period, LTbR signaling is required for follicular dendritic cell (FDC) maintenance and germinal center (GC) formation in lymph nodes [5]. And it is even more important in spleen, where postnatal LTbR-Ig treatment leads to disruption of follicles and marginal zone, as well as GC failure [6].
Clusterin (also known as testosterone repressed prostate message-2, sulfated glycoprotein-2, apolipoprotein J (ApoJ), and X-ray-inducible transcript 8) was first described as the major glycoprotein in ram rete testis fluid with the capacity to elicit clustering of cells in an in vitro assay [7]. It is a multifunctional protein, which is mainly studied for its role in neurodegeneration and cancer [8]. Its mRNA is present at relatively high levels in brain, ovary, testes, liver, heart and adrenal gland; at moderate levels in spleen, lung, breast, kidney, seminal vesicle, prostate, and uterus; at low levels in skin, bone, thymus and digestive tract; and is absent in T-lymphocytes [9,10]. Clusterin participates in tissue remodeling, apoptosis, lipid transport, complement-mediated cell lysis, and serves as an extracellular chaperone [8,11,12,13,14,15,16].
At the protein level, clusterin was found in non-lymphoid cells of many SLO: gut-associated lymphoid tissue, Waldeyer's ring [17,18], reactive tonsils, lymph nodes and spleen [5,19], but virtually nothing is known about its function in these organs. Clusterin is also present in medullary epithelial stromal cells of the primary lymphoid organ -thymus, but its precise function there is also not clear [20,21].
In the present work we used expression profiling to identify new potential target genes of LTbR signaling pathway by comparing transcriptomes of spleen stromal cells derived from wild type and LTbR knock-out (LTbR-KO) mice. Since LTbR signaling drives morphogenesis and functional maturation of SLO, we expected to find new immunity-relevant genes among its targets. After filtration of the microarray results we focused on clusterin as it was significantly downregulated in LTbR-deficient spleen at both mRNA and protein level and its function in the immune system was poorly studied. We demonstrated activation of clusterin gene transcription upon interaction of mouse embryonic fibroblasts (MEF) with lymphoid cells bearing LT and significant changes in clusterin protein level and tissue distribution during primary immune response to T-dependent antigen.

Results and Discussion
Transcriptome of splenic stromal cells in normal and LTbR-deficient mice LTbR-KO mice lack all lymph nodes and Payer's patches. Their spleen is completely disorganized with no proper segregation into red and white pulp. Nevertheless, lymphocytes are present in the organ, but they are not organized in conventional T-and B-zones. GCs also fail to develop [22].
We compared transcriptomes of freshly isolated LTbR-KO and WT splenic stroma (see Materials and Methods for details) and identified 505 genes (Table S1) with predominantly stromal expression, which showed .1.5-fold reduction of mRNA level upon LTbR knock-out (further designated as ''potential LTbR target'', or ''PLT'' genes). Microarray results were confirmed by quantitative PCR and Northern blot analyses ( fig. 1 and data not shown). Among PLT genes are a number of known targets of LTbR: Madcam1 [23], Cxcl13(Blc), Ccl21a(Slc), Vegfa [24], Prnp [25], which confirms the validity of our approach. Some of the known LTbR targets such as Ccl19, Baff and Sdf1, could not be assessed with the array used for technical reasons. Il7 appeared to be downregulated in LTbR-KO spleen in agreement with its expression by gp38-positive stromal cells [26], which mature under LT control [27]. Somewhat surprisingly, Vcam1 expression in splenic stroma was not affected by LTbR knock-out, even though this gene was previously shown to be directly activated by LTbR stimulation via canonical NFkB pathway [28,29]. 24 genes showed .1.5-fold elevated expression in LTbR-KO spleen as compared to wild type (Table S2), indicating that these genes may be negatively regulated by LTbR signaling pathway.
Strikingly, spleen stroma cultured for 3-5 weeks in the absence of lymphocytes dramatically changed its expression profile. Cluster analysis revealed that transcriptome of cultured splenic stroma is the most distant one among all the studied groups of samples ( Figure S1).
Endothelial and smooth muscle cells of white pulp express LTbR, and its activation is required to maintain proper marginal sinus vascular structure and function [30]. In accordance with that, Gap Junction Protein Alpha 4, Endothelin 1 and other genes of vasculature development cluster were highly enriched among PLT genes (Enrichment Score: 6.43, p-values for individual annotation terms #3.4N10 26 ). Interestingly, nervous system-related genes also appeared to be highly enriched in PLT gene group (Enrichment Score: 3.2, p-values#0.006 for neurogenesis cluster; Enrichment Score: 1.94, p-values#0.026 for regulation of neurotransmitter secretion cluster). That may be of physiological significance, since innervation of SLOs was shown to be important for their function [31]. Distribution of nerve fibers of different types in spleen of LTbR-KO mice has not been studied yet, but represents an interesting question.

Clusterin gene expression is dependent on LTbR signaling
LTbR knock-out results in a significant decrease in Clu gene expression in splenic stroma (by the factor of 3 according to microarray data and by almost an order of magnitude according to quantitative real-time PCR) (Table S1 and fig. 1). This is in accordance with previously reported Clu downregulation in mouse spleen upon combined lymphotoxin-a and TNF knock-out [32] as well as with the fact that Clu transcripts are significantly overrepresented in FDC-enriched cell fraction of mouse spleen and are consistently down-regulated in soluble LTbR-Ig-treated mesenteric lymph nodes [5]. Among studied organs (bone marrow, spleen, liver, lung and thymus), dependence of Clu mRNA level on LTbR expression was seen only in spleen, where it also depended on the presence of TNFR1 but to a lesser extent. Interestingly, relationship between splenic Clu mRNA levels in WT, LTbR-KO and TNFR1-KO mice was very similar to that of two well-studied LTbR targets Blc and Slc (Fig. 1).
In order to demonstrate more directly that Clu expression can be activated by LT, we incubated MEF with Reh human Blymphocytic leukemia cells. Reh cells were shown to constitutively express high amounts of LT heterotrimer on their surface without expressing TNFa [33], and human LT was shown to effectively interact with the murine LTbR receptor [34]. Blc and Vcam1 mRNAs, previously shown to peak in response to LTbR crosslinking in MEF at 24 h and 3 h, respectively [2,35], were used as controls for proper activation. We used Jurkat human T-cell line as a negative control, since flow cytometry showed the absence of surface LT epitopes on these cells [33].
Cocultivation of MEF with Reh but not Jurkat cells for 24 h significantly induced Clu mRNA ( fig. 2). Temporal dynamics of Clu upregulation resembled that published for Blc, but not Vcam1 [2,28], indicating involvement of alternative rather than classical NFkB pathway in Clu activation via LTbR.

Splenic CLU isoform and its distribution among stromal cell subpopulations
There are several CLU protein isoforms encoded by two CLU gene transcripts. The main and longer gene transcript encodes glycosylated presecretory form psCLU with apparent molecular weight of about 60 kDa. Cleavage into aand b-chains and further extensive glycosylation produces a mature, secreted heterodimeric 70-80 kDa protein referred to as sCLU. Under reducing conditions both aand b-subunits of sCLU run at about 40 kDa at SDS-PAGE. The second transcript lacks the endoplasmic reticulum-targeting sequence at exon 2 and its product is detected as 49 kDa non-glycosylated pnCLU precursor in the cytosol and a 55-kDa glycosylated nCLU protein in the nucleus [36].
Secretory and nuclear forms of clusterin are considered to have somewhat opposing functions, with sCLU being a cell-protective, anti-apoptotic protein, and nCLU acting as a pro-death signal, inhibiting cell growth and survival [37]. As it is important for understanding the clusterin functions in SLO, we assessed CLU protein isoform in the splenic stroma using Western blot. Clusterin immunopositive band ran around 70 kDa in non-reducing conditions, and around 40 kDa in reducing conditions, which corresponds to sCLU and its two co-migrating subunits, respectively ( fig. 3a). The pattern was similar for WT and KO mice, however the intensity of CLU bands in KO mice was significantly reduced ( fig. 3b).
In order to assess cellular distribution of sCLU in splenic stroma, we used immunohistochemical staining of frozen spleen sections with commercial polyclonal anti-CLU antibodies (R&D AF2747) raised against recombinant mouse CLU Glu22-Glu448. Polyclonality and usage of almost full-length protein as immunogen ensured that this antibody would recognize different CLU isoforms in different applications. AF2747 specificity was confirmed by specific staining of HEK293 cells transiently transfected with full-length CLU (data not shown).
Multi-color immunostaining with B-220 (recognizing B-cells) and ER-TR7 (recognizing fibroblastic reticular cells, marginal reticular cells and red pulp fibroblasts) showed that clusterin was expressed by all subsets of stromal cells in spleen and mesenteric lymph nodes (MLN) except for marginal reticular cells (MRC) ( fig. 4, 5). This expression pattern is broader than previously reported [5,19,38], though the brightest staining was still observed in B-cell areas, especially in GCs after immunization, and is attributed to FDC for which clusterin is used as one of differential markers [38,39]. An important difference with the previous observations consists in the clear absence of marginal zone staining in spleen ( fig. 4d). Diffuse staining was observed in spleen red pulp, MLN medulla and lumen of high endothelial venules, which can be explained by the high amount of sCLU in blood [13]. GC staining also had a diffuse appearance, not resembling stromal cell contours, which may be indicative of active secretion of sCLU in this area. Previously, sCLU secretion by FDC was shown by Verbrugghe et al. who detected clusterin immunoreactivity in the endoplasmic reticulum, Golgi apparatus, and on the plasma membrane of FDC in human Payer's patches by electron microscopy [18].
In contrast to the wild type pattern, only faint staining of few stromal cells could be seen in disorganized white pulp of the spleens of LTbR-KO mice (data not shown). Diffuse staining of red pulp was not affected. This may reflect not only the absence of FDC, which contribute to the bright staining of B-cell follicles in WT mice spleen, but also downregulation of CLU in other stromal cell types in the absence of LTbR signal. sCLU dynamics during immune response CLU was previously shown to be induced during tissue remodeling in mammary gland [40]. Its mRNA and protein levels  substantially rise during pregnancy, when mammary tissue undergoes structural changes, and at the early stages of postweaning involution accompanied by high rates of apoptotic death [11]. This data and also the fact that CLU may serve as a survival factor for GC B-cells [5] prompted us to study CLU protein dynamics in spleen and MLN during immune response by immunohistochemistry and Western blot. The development of GC reaction was confirmed by peanut agglutinin (PNA) staining. Only in spleen, but not in MLN, CLU protein level was elevated at the peak of GC formation ( fig. 6). This overall increase was accompanied by the accumulation of CLU immunoreactivity in the light zones of GCs ( fig. 7a). That was not the case in MLN, where bright focuses of CLU could be seen in centers of B-cell follicles independently of immunization and GC presence ( fig. 7b). No parallel increase in Clu mRNA level could be detected in spleen by quantitative PCR (data not shown) indicating that either Clu mRNA peaks before the 8th day post immunization, or CLU protein synthesis is controlled at the post-transcriptional level.

Ethics Statement
This study was carried out in strict accordance with the recommendations in the Guide for the Care and However the cultured splenic stroma has dramatically changed the expression profile, thereby couldn't be utilized for further LTbR regulated genes/proteins studies in vitro. The use of primary MEFs co-culture with lymphoid cell lines Reh and Jurkat can only partially replace in vivo experiments, and won't be suitable for histological studies. To perform ex vivo organ examination and primary cultures all efforts were made to ameliorate animal suffering. Animal sacrifice was performed by CO 2 asphyxiation, checked for responsiveness and followed by cervical dislocation for adult mice; b) reduction: animal studies are preceded by multiple experiments ex vivo, in vitro or in primary cell cultures in vitro. Minimum number of animals was used in each experiment to achieve statistically significant data (n = 3 or 4). The isolation of spleens and lymph nodes was combined from each animal and parts of the same organs were used for histology, Western blot and real-time PCR to keep the number of required animals to an absolute minimum; c) refinement: Animal studies were only used at advanced stages of investigations when few, specific and highly relevant questions -such as clusterin bio-distribution upon immunization -were addressed by a limited number of experiments.

Animals and Housing Conditions
We obtained 8-10-week-old inbred mice (weight 20-25 g) C57BL/6, LTbR-KO [22] (mice deficient for lymphotoxin b receptor), and TNFR1-KO [41] (mice deficient for tumor necrosis factor receptor type 1) mice from the stock of the specific pathogen free Laboratory Animal Breeding Facility ''Pushchino'' (http:// spf-animals.ru/). The knockout mice were backcrossed to C57BL/ 6 more than 10 times. All experimental animals were housed according to ''Recommendations for the health monitoring of mouse, rat, hamster, gunea pig and rabbit breeding colonies'' FELASA (Federation of European Laboratory Animal Science Associations) recommendations, June, 2001. All mice were housed (for 4-6 weeks) in single-sex groups, of 3 or 5 in a cage. Male and female groups were maintained in standard plastic cages (3562169 cm) with sawdust (wooden flakes) as nesting material. Food pellets (BioPro, Novosibirsk) and water were provided ad libitum. Mice were maintained at 22-24uC under 12:12 light-dark cycle with lights off at 9.00 p.m.. Every 5 days, animals were placed into a clean cage with fresh sawdust. In each cage, mice were individually marked by ear cuts. When C57/BL6 mice were used as controls they were age and sex matched to the KO mice and co-housed for a minimum of two weeks, in the most cases littermates were used as controls.

Mice immunizations
For immunization, 8 week-old mice were injected intraperitoneally in their home cages with 2N10 8 SRBC in 300 ml of sterile PBS. Control littermate animals were left intact. Mice were sacrificed 8 days after injection on the pick of germinal center reaction, and mesenteric lymph nodes and spleens were immediately isolated and put on ice in RPMI medium until further manipulations. 3 mice of each genotype: C57Bl6 and LTbR-KO (overall 2 experimental groups), were immunized with SRBCs; organs from 3 untreated mice of each genotype were used as a respective controls for histological studies or Western blot. Immunization and analysis were repeated at least 2 times. Thereby, the total mice number was N = 24, where C57BL/6 n = 12 and LTbR-KO n = 12. We utilized 3 mice of each genotype or group per experiment as minimum sample size sufficient for statistical analysis, however the statistical analysis of histological score and/or counts of certain stained objects on several fields of microscopic view are not presented, since the observed clusterin distribution was clearly reproduced in several independent experiments.

Preparation of samples and microarray hybridization
For microarray hybridization, four types of samples were prepared: uncultured splenic stroma of C57BL/6 mice, uncultured splenic stroma of LTbR-KO mice, splenocytes of C57BL/6 mice and cultured splenic stroma of C57BL/6 mice. To obtain splenocyte-and stroma-enriched fractions, freshly isolated spleens were rubbed over 70 mm mesh and washed with PBS on ice. The wash, containing splenocytes, was collected, centrifuged, and the pellet was homogenized in TRIzol (Life Technologies). Remaining stroma was collected from the mesh, cut with scissors and homogenized in TRIzol. Total RNA isolation, amplification and hybridization on Illumina chip were performed by ZAO ''Genoanalytica''. Briefly, total RNA was extracted from the samples according to TRIzol manufacturer's instruction. RNA was quantified using Nanodrop and its quality was assessed by Agilent Total RNA 6000 chip. 400 ng of total RNA was amplified by Illumina TotalPrep RNA Amplification Kit (Ambion). Amplified RNA was hybridized with MouseRef-8 v1.1 Expression BeadChips (Illumina) according to Illumina protocol.

Microarray data acquisition and analysis
Microarray data acquisition and analysis were done with GenomeStudio Gene Expression Module v1.0 (Illumina) (accession number in the NCBI GEO database is GSE52294). After normalization between samples and replicates using «cubic spline» method [42] and background substitution, 1542 genes were selected on the basis of differential p value ,0.05 when comparing wild type with knock-out stroma (p values were calculated using ''Illumina Custom'' algorithm for multiple comparisons with Benjamini-Hochberg procedure controlling the false discovery rate). To avoid artifacts resulting from splenocyte contamination of stromal preparations, only genes with expression levels in C57BL/ 6 stroma greater than in C57BL/6 splenocytes were considered for further analysis. Functional annotation clustering of differentially expressed genes was performed using DAVID analytic tools [43]. Customized parameters were: Annotation categories -GO-TERM_BP_4, Classification stringency -medium, Enrichment Thresholds -EASE = 0.05.  Quantitative real-time PCR Entire organs or isolated stroma were first cut with scissors on ice. Resulting material, bone marrow aspirates or cultured cells were homogenized in TRIzol reagent by pipetting. Total RNA was isolated according to TRIzol manufacturer's instructions. The RNA quality was assessed by spectrophotometry and agarose gel electrophoresis. 4 mg of the total RNA was taken for cDNA synthesis using RevertAid First Strand cDNA Synthesis Kit (Thermo Scientific) according to manufacturer's protocol. Quantitative real-time PCR was performed using EVA Green Real-Time PCR kit (Syntol, Moscow, Russia) and 7500 Real-Time PCR System (Applied Biosystems). Primer sequences were: Clu, 59 -CTGTCCACTCAAGGGAGTAGG and 59 -GTGTCCTCC-AGAGCATCCTC; Blc, 59 -CATAGATCGGATTCAAGT-TACG and 59 -TCTTGGTCCAGATCACAACTTC; Clec4g, 59 -TACTGTCCAGTGCCTCCAGCAAG and 59 -TGTC-ACGGAGCAGCAATTCCTG; Gja4, 59 -GCAAGCAGGCGA-GAGAGG and 59-AGATGAAGAGCACCGTTAACCAG; Hmgcs2, 59 -GGTGTCCCGTCTAATGGAGA and 59 -ACAC-CCAGGATTCACAGAGG; Slc, 59 -ATGGCTCAGATGAT-GACTCTG and 59 -TAGCCTCGGACAATACTGTAGG. For positive control and normalization, b-actin primers discriminating mouse from human were used: 59 -CCGCGAGCACAGCT-TCTTTG and 59 -CCATCACACCCTGGTGCCTA. Forward and reverse primers were complimentary to different exons so that they did not give any products using genomic DNA as a template.

Western blot
Entire spleens and MLNs were cut with scissors on ice, homogenized in ice-cold RIPA buffer (150 mM NaCl, 1% NP-40, 0.1% SDS, 50 mM Tris buffer, pH 8.0) by gentle sonication, and clarified by centrifugation at 12000 g for 10 min at +4uC. Total protein concentration was measured in supernatants by Bradford protein assay. 56 Laemmli buffer with or without bmercaptoethanol was then added 1:4 (v/v) and samples were boiled for 5 min. After separation by SDS-PAGE electrophoresis (20 mg of total protein per well) proteins were transferred to Hybond-C Extra nitrocellulose membrane (Amersham Biosciences) using Mini Trans-Blot system (Bio-Rad). Uniform loading and transfer was confirmed by Ponceau Red staining. Membranes were blocked with 3% BSA in TBST for 30 min, stained with anticlusterin antibody (1:3000; AF2747 R&D Systems) overnight at + 4uC, washed with TBST, and incubated with secondary HRPconjugated anti-goat antibody (1:50000; Pierce) in 5% NFDM TBST solution for 1 h at room temperature. After membranes were thoroughly washed in TBST, specific bands were visualized by SuperSignal West Dura Extended Duration Substrate (Thermo Scientific) and x-ray film. Densitometry was performed using ImageJ software (NIH, USA).

MEF activation
MEFs were prepared from 13.5-day-old fetuses of C57BL/6 wild type mice. All procedures were approved by the Scientific Council of the Engelhard Institute of molecular biology, according to ''The Guidelines for Manipulations with Experimental Animals'' (the decree of the Presidium of the Russian Academy of Sciences of April 02, 1980, no. 12000-496). To obtain embryos all efforts were made to ameliorate animal suffering. Adult mice sacrifice was performed by CO2 asphyxiation, checked for responsiveness and followed by cervical dislocation. Embryos were separated from the placenta and surrounding membranes. The brain and dark-red organs were dissected and discarded. Remaining body parts were rinsed with DMEM, dissected into small pieces and treated with 1 mL of trypsin/EDTA at 37uC for 20 min to separate cells. After addition of 6 mL of fresh medium, cells were further separated by pipetting and put to CO 2 incubator overnight. The next day non-adherent cells were removed during medium replacement. Adherent cells were let to reach confluence for 1-3 days. After consequent trypsinization and centrifugation only single-cell fraction was used for further replating. MEF were cultured in DMEM (Paneco, Moscow, Russia) supplemented with 10% FBS (Biological Industries, Kibbutz, Israel), 2 mM Lglutamine (HyClone), MEM Non-Essential Amino Acids Solution (HyClone), 0.1 mM Sodium Piruvate (HyClone), 10 mM HEPES (HyClone), 100 U/ml Penicillin 100 ug/ml Streptomycin (Gibco). At the second passage MEF were plated after trypsinization onto 6 cm dishes, 5610 5 cells per dish. Next morning, the medium was removed and 8 ml of fresh RPMI containing 5610 6 Reh or Jurkat cells was added for LTbR activation. After 3 and 24 hours, MEF were washed twice with PBS and homogenized in 1 ml of TRIzol reagent for further RNA isolation. Vcam1 and Blc expression levels were assessed by semi-quantitative PCR (primer sequences were: Blc, 59-CAGAATGAGGCTCAGCACAGC and 59-TGCAACGGAGCTTGAGCATTCC; Vcam1, 59-AATGCC-TGTGAAGATGGTCG and 59-GAACAGGTCATTGTCA-CAGC) as a control for proper activation.

Statistical analysis
Statistical significance of differences between groups was tested with the Mann-Whitney U test using Statistica 7.0 software (StatSoft, Inc.). All data are expressed as mean 6SD. Figure S1 Cluster analysis of the microarray data. (DOCX)

Table S1
The list of genes which mRNA levels were more than 1.5-fold higher in wild type spleen stroma comparing to LTbR-KO spleen stroma. (DOCX)

Table S2
The list of genes which mRNA levels were more than 1.5-fold lower in wild type spleen stroma comparing to LTbR-KO spleen stroma. (DOCX) Author Contributions