Hspa13 Promotes Plasma Cell Production and Antibody Secretion

The generation of large numbers of plasma cells (PCs) is a main factor in systemic lupus erythematosus (SLE). We hypothesize that Hspa13, a member of the heat shock protein family, plays a critical role in the control of PC differentiation. To test the hypothesis, we used lipopolysaccharide (LPS)-activated B cells and a newly established mouse line with a CD19cre-mediated, B cell–specific deletion of Hspa13: Hspa13 cKO mice. We found that Hspa13 mRNA was increased in PCs from atacicept-treated lupus-prone mice and in LPS-stimulated plasmablasts (PBs) and PCs. A critical finding was that PBs and PCs [but not naïve B cells and germinal center (GC) B cells] expressed high levels of Hspa13. In contrast, the Hspa13 cKO mice had a reduction in BPs, PCs, and antibodies induced in vitro by LPS and in vivo by sheep red blood cells (SRCs)- or 4-hydroxy-3-nitrophenylacetyl (NP)-immunization. Accordingly, the Hspa13 cKO mice had reduced class-switched and somatically hypermutated antibodies with defective affinity maturation. Our work also showed that Hspa13 interacts with proteins (e.g., Bcap31) in the endoplasmic reticulum (ER) to positively regulate protein transport from the ER to the cytosol. Importantly, Hspa13 mRNA was increased in B220+ cells from patients with multiple myeloma (MM) or SLE, whereas Hspa13 cKO led to reduced autoantibodies and proteinuria in both pristane-induced lupus and lupus-prone MRL/lpr mouse models. Collectively, our data suggest that Hspa13 is critical for PC development and may be a new target for eliminating pathologic PCs.


INTRODUCTION
Plasma cells (PCs) play a critical role in the immune response by producing antibody (1,2). B cells arise in the bone marrow and mature and differentiate into germinal center (GC) B cells, plasmablasts (PBs), and terminally differentiated PCs in the peripheral secondary lymphoid tissues such as the spleen and lymph nodes (LNs) (3)(4)(5). The process of the differentiation of B cells into PCs is regulated by some important transcriptional factors, including the PC-inhibitory factors Pax5 and Bcl6 and the PC-promoting factors Prdm1 (Blimp1) and Xbp1 (3)(4)(5)(6)(7).
The abnormal production of PCs is involved in the pathology of both multiple myeloma (MM) and systemic lupus erythematosus (SLE). MM is a malignancy of PCs; patients typically present with an infiltration of the bone marrow with clonal PCs and monoclonal protein in the serum and/or urine (8)(9)(10). Most cases of MM, especially in relapsed patients, are incurable (11). Autoreactive PCs and pathogenic autoantibodies are critical factors involved in SLE pathology (12,13). Belimumab, a human anti-BAFF (B-cell activation factor) antibody that selectively depletes mature and activated B cells and PBs and results in an increase in the number of PCs, has been used to treat patients with SLE (14, 15). The drug atacicept (TACI-IgG) is a recombinant fusion protein containing the extracellular ligandbinding protion of human TACI (transmembrane activator and calcium modulator and cyclophilin-ligand interactor, one of the BAFF receptors) linked to the Fc fragment of human IgG; its effects are similar to those of belimumab (16,17). These results suggest there is no effective curative treatment for MM or SLE that targets PCs.
The binding immunoglobulin protein (BiP), also known as GRP-78, heat shock 70-kDa protein 5 (HSPA5), or (Byun1), is the first chaperone discovered that non-covalently binds to free IgH but not to IgH associated with IgL (18). Heat shock proteins (HSPs) (e.g., Hsp90) ensure correct protein folding (e.g., antibody) in PCs and cell survival. Stressful conditions often stimulate cells to produce HSPs (19,20). HSPs interact with cellular proteins to ensure proper protein folding and transport from the endoplasmic reticulum (ER) into the cytoplasm or secretory pathway (21)(22)(23). HSPs also contribute to protein (e.g., HSP) misfolding that mediates amyloid β oligomer accumulation (24,25). HSP (e.g., Hsp90) inhibitors induce the unfolded protein response (UPR) to reduce abnormal immunoglobulin production and cell death (26). These results suggest that targeting HSPs may represent a novel therapeutic strategy for controlling abnormal PCs.
We hypothesize that Hspa13, a member of the heat shock protein family, plays a critical role in the control of PC differentiation. To test the hypothesis, we used lipopolysaccharide (LPS)-activated B cells and a newly Three lupus-prone MRL/lpr mice per group were injected intraperitoneally (i.p.) with 5 mg/kg atacicept (TACI-IgG) and control (IgG) at 1, 2, 3, and 4 weeks (two times per week) after mice reached 6 months of age. On day 4 after therapy, mice were euthanized and B cells were separated from the spleens by B220 microbeads. The transcripts in B cells were determined by Affymetrix Microarrays. The fold change of germinal center (GC) B cell-associated genes including Pax5 and Bcl6, PB/PC-promoting genes including Prdm1 (Blimp1) and Xbp1, and the interested gene Hspa13 (Stch) in atacicept-treated group to those in IgG group is shown. 1st, 2nd, and 3rd represent 3 independent experiments.
established mouse line with a CD19 cre -mediated, B cellspecific deletion of Hspa13 (Hspa13 cKO). We found that PBs and PCs (but not naïve B cells and GC B cells) expressed high levels of Hspa13. In contrast, the Hspa13 cKO mice had a reduction in PBs, PCs, and antibodies induced in vitro by LPS and in vivo by sheep red cells (SRCs) or 4-hydroxy-3nitrophenylacetyl (NP)-immunization, and there were reduced numbers of autoantibodies and levels of proteinuria in both pristane-induced lupus and lupus-prone MRL/lpr mouse models. Collectively, our data suggest that Hspa13 is critical for PC development and may be a new target for eliminating pathologic PCs.

METHODS AND MATERIALS Ethics Committee Approval
Care, use, and treatment of mice in this study were in strict agreement with international guidelines for the care and use of laboratory animals. This study was approved by the Animal Ethics Committee of the Beijing Institute of Basic Medical Sciences.  mice to generate CD19 cre Hspa13 fl/fl (Hspa13 cKO) mice. Wild type (WT), Hspa13 fl/fl , and heterologous CD19 cre mice were used as the control for Hspa13 cKO mice. Three lupus-prone MRL/lpr mice per group were injected intraperitoneally (i.p.) with 5 mg/kg atacicept (TACI-IgG) and control (IgG) at 1, 2, 3, and 4 weeks (two times per week) after mice reached 6 months of age based on a previous protocol (28).
To explore the role of Hspa13 in lupus, the floxed Hspa13 (Hspa13 fl/fl ) mice in lupus-prone MRL/lpr mice background were generated and crossed with CD19 cre mice to generate CD19 cre Hspa13 fl/fl (Hspa13 cKO) mice.

Affymetrix Microarrays
Affymetrix microarrays were done based on a previous method (31). Total RNA was extracted from B cells with Trizol and purified over Qiagen RNeasy columns (Qiagen). Synthesis and labeling of RNA and hybridization of arrays were conducted. Stained arrays (430 2.0) were scanned on an Agilent Gene Array Scanner (Affymetrix).

RNA-Sequencing
The transcripts in cells were determined by RNA-sequencing using previous methods (32)(33)(34). Briefly, RNeasy Mini Kit (Qiagen, Venlo, Netherlands) was used to isolate and purify total RNA from cells. NanoDrop R ND-1000 spectrophotometer and Agilent 2100 Bioanalyzer and RNA 6000 NanoChips (Agilent, Palo Alto, CA, USA) were used to determine RNA concentration and quality, respectively. TruSeq Stranded Total RNA Library Prep Kit with Ribo-Zero Gold (Illumina) was used to prepare Libraries. Transcripts were analyzed by RNAsequencing (Genewiz Corp., Suzhou, China).

Quantitative PCR (qPCR) Analysis
Total RNA was extracted from cells with Trizol (Invitrogen Life Technologies). The final RNA pellets were dissolved in 0.1 mM EDTA (2 µl/mg original wet weight). Reverse transcription reactions were carried out on 22 µl of sample using superscript II RNA H-Reverse Transcriptase (Invitrogen Life Technologies) in a reaction volume of 40 µl. All samples were diluted in 160 µl nuclease-free water. qPCR was employed to quantify mouse gene expression from the cDNA samples. Mouse gene expression was normalized to the levels of the β-actin gene. For sheep red cell (SRC) immunization, three 9-week-old C57BL/6 mice were i.p. injected with 1 x 10 9 SRC and were sacrificed on day 14. Splenic B cells were sorted using B220 microbeads. The transcripts in B cells were determined by 10x Genomics and single-cell RNA-sequencing. 14 immature B cells ( were chosen and Hspa13 expression was compared in these cells.

Plasmid Constructs and Transfection
The recombinant plasmids expressing Hspa13-V5 and Bcap31-Flag were constructed by PCR-based amplification of Hspa13 and Bcap31 cDNA from SP 2/0 cells (ATCC R CRL-1581, Rockville, MD, USA), which was then subcloned into the pcDNA3.1 eukaryotic expression vector. The recombinant plasmids were transiently co-transfected into 293T cells with jetPEI (Polyplus Transfection) according to the manufacturer's instructions.

Assessment of Proteinuria
Urine was manually expressed from each mouse on a weekly basis, collected into a sterile container, and assayed for the presence of protein (specifically albumin) using a colorimetric method (Albustix Reagent Strips, Bayer Corporation, Elkhart, IN).

Statistics
Statistics were generated using t-test in GraphPad Prism (version 5.0, GraphPad Software Inc., USA) and values are represented as mean ± SEM. Results were considered statistically significant at p < 0.05.

PBs and PCs Expressed High Levels of Hspa13
Previous studies have shown that atacicept (TACI-IgG) reduced the number of mature and activated B cells but resulted in an increase in the number of terminally differentiated PCs (16,17). In this work, we observed that atacicept reduced the expression of PC-inhibitory genes, including Pax5 and Bcl6, and  Hspa13 cKO did not affect LPS-stimulated B-cell activation. On day 3 following LPS stimulation, cells were stained with isotype control antibodies, anti-mouse B220 and GL7 antibodies, and analyzed by FACS. The percentages (B) and the absolute numbers (C) of B220 + GL7 + B cells are shown. (D,E) Hspa13 cKO reduced LPS-induced PBs, early PCs, and mature PCs. On day 3 following LPS stimulation, cells were stained with isotype control antibodies, anti-mouse TACI, CD19, B220, and CD138 antibodies, and were then analyzed by FACS. The percentages (D) and the absolute numbers (E) of TACI + CD138 + B220 int CD19 int PBs, TACI + CD138 + B220 − CD19 int early PCs, and TACI + CD138 + B220 − CD19 − mature PCs are shown. (F) Hspa13 cKO reduced LPS-induced antibody secretion. On day 3 following LPS stimulation, culture supernatants were collected and the total IgM, IgG1, IgG2b, IgG2c, and IgG3 antibody levels were analyzed by ELISA. (A-F) Data represent three independent experiments, with six mice per group per experiment. Data were analyzed by two-way (A) and one-way (C,E,F) ANOVA plus the Bonferroni test: compare selected pairs of columns and show as mean ± s.e.m (N = 6 for all groups). ***P < 0.001.
up-regulated PC-promoting genes, including Prdm1 (Blimp1) and Xbp1 (Table 1). Interestingly, Hspa13 expression was much increased in response to atacicept treatment ( Table 1). These results suggest that Hspa13 mRNA levels were increased in atacicept-induced PBs and PCs.
To confirm that PCs expressed high levels of Hspa13 mRNA, LPS was used to induce PB and PC production in vitro. As expected, LPS stimulation reduced the expression of B cellassociated genes, including CD19 and Ms4a1 (CD20), and PCinhibitory genes, including Pax5, Bcl6, and Aicda (Aid), and it up-regulated PC-promoting genes, including Prdm1 (Blimp1), Xbp1, and Hspa13 ( Table 2). These results suggest that Hspa13 mRNA levels were increased in LPS-induced PBs and PCs. Furthermore, qPCR (Figure 1A), RT-PCR/agarose (Figure 1B), and western blot (Figure 1C) analysis demonstrated that LPS up-regulated Hspa13 mRNA and protein expression in a timedependent manner.
To further confirm that PBs and PCs expressed high levels of Hspa13 mRNA, we used single-cell RNA-sequencing to evaluate Hspa13 expression in 14 immature B cells (14.29% were positive for Hspa13 expression), 16 mature B cells (0% were positive for Hspa13 expression), 13 memory B cells (0% were positive for Hspa13 expression), 34 GC B cells (11.76% were positive for Hspa13 expression), and 10 PBs (100% were positive for Hspa13 expression) ( Table 3). Only PBs expressed high levels of Hspa13, relative to immature, mature, memory, and GC B cells ( Table 3). These results suggest that PBs but not naïve, memory, or GC B cells expressed high levels of Hspa13 mRNA.

Reduction of PBs, PCs, and Antibodies in Hspa13 cKO Mice
To explore the role of Hspa13 in PBs and PCs, CD19 cre Hspa13 fl/fl (B-cell specific knock-out of Hspa13, cKO) mice were developed (Figure 2A). PCs (TACI + CD138 + B220 − CD19 − ) were sorted from the spleens and bone marrows (BMs) of 7-to 9-week-old, heterologous CD19 cre , Hspa13 fl/fl , and CD19 cre Hspa13 fl/fl mice by FACS and subjected to PCR ( Figure 2B) and western blot (Figure 2C) analysis. The data demonstrated that Hspa13 was knocked out in PCs from the CD19 cre Hspa13 fl/fl mouse. As compared with the control group that included WT, Hspa13 fl/fl , and CD19 cre mice, the Hspa13 cKO mice contained reduced levels of TACI + CD138 + B220 int CD19 int PBs, TACI + CD138 + B220 − CD19 int early PCs, and TACI + CD138 + B220 − CD19 − mature PCs in the spleens, lymph nodes (LNs), and BMs, but this was not the case in naïve B220 + CD19 + B cells or CD38 lo GL7 hi B220 + CD19 + GC B cells (Figures 3A,B). Accordingly, the total IgM, IgG, IgG1, IgG2b, IgG2c, IgG3, IgA, and IgE antibody levels were also reduced in Hspa13 cKO mice ( Figure 3C). Collectively, these data suggest that the specific knock-out of Hspa13 in B-cells reduced the number of PBs and PCs and the levels of antibodies in mice.

Hspa13 cKO-Mediated Reduction of the Production of Antigen-Induced PCs and Antibodies
To explore the effect of the Hspa13 cKO on the production of antigen-induced PBs, PCs, and antibodies, a T-cellindependent antigen, LPS, was used to induce the production of PBs, PCs, and antibodies in vitro. The results showed On day 21 following SRC stimulation, splenic lymphocytes were stained with isotype control antibodies, anti-mouse CD19, B220, CD38, and GL7 antibodies, and were then analyzed by FACS. The percentages (A) and the absolute numbers (B) of CD38 lo GL7 hi GC cells gated on CD19 + B220 + are shown. (C,D) Hspa13 cKO did not affect the SRC-induced dark zone (DZ) and light zone (LZ) GC B-cell production. On day 21 following SRC stimulation, splenic lymphocytes were stained with anti-mouse CD19, B220, CD38, GL7, CXCR4, and CD86 antibodies, and were then analyzed by FACS. The percentages (C) and the absolute numbers (D) of CXCR4 hi CD86 lo DZ and CXCR4 lo CD86 hi LZ GC B cells gated on CD19 + B220 + CD38 lo GL7 hi GC cells are shown. (E,F) Hspa13 cKO reduced SRC-induced IgG1-, IgG2b-, IgG2c-, and IgG3-expressing PBs/PCs. On day 21 following SRC stimulation, splenic lymphocytes were collected and intracellular staining was performed with isotype control antibodies, anti-mouse B220, IgG1, IgG2b, IgG2c, and IgG3 antibodies. The percentages (E) and the absolute numbers (F) of IgG1-, IgG2b-, IgG2c-, and IgG3-expressing B220 + PBs and B220 − PCs are shown. (G) Hspa13 cKO reduced SRC-induced antibody secretion. On day 21 following SRC stimulation, sera were collected and the total IgM, IgG, IgG1, IgG2b, IgG2c, and IgG3 antibody levels were analyzed by ELISA. (A-G) Data represent three independent experiments, with six mice per group per experiment. Data were analyzed by two-way (D) and one-way (B,F,G) ANOVA plus the Bonferroni test: compare selected pairs of columns and show as mean ± s.e.m (N = 6 for all groups). **P < 0.01, ***P < 0.001. that the Hspa13 cKO did not affect LPS-stimulated B-cell proliferation (Figure 4A) or the production of activated B220 + GL7 + B cells (Figures 4B,C). These data suggest that the Hspa13 cKO did not affect LPS-stimulated B-cell activation. An RNA-sequencing assay showed that the Hspa13 cKO led to reduced levels of LPS-induced Prdm1 and Xbp1 mRNA ( Table 4). Accordingly, the Hspa13 cKO reduced the number of LPS-induced TACI + CD138 + B220 int CD19 int PBs, TACI + CD138 + B220 − CD19 int early PCs, and TACI + CD138 + B220 − CD19 − mature PCs (Figures 4D,E). Thus, the Hspa13 cKO reduced the levels of LPSinduced IgM, IgG1, IgG2b, IgG2c, and IgG3 antibodies ( Figure 4F). These data suggest that the Hspa13 cKO reduced the production of LPS-induced PBs, PCs, and antibodies.

Hspa13 cKO Reduced Class-Switched and Somatically Hypermutated Antibody With Defective Affinity Maturation
To assess the effect of the Hspa13 cKO on antibodies, single-cell RNA-sequencing was used to examine single PBs isolated from SRC-immunized Hspa13 cKO and control (Hspa13 fl/fl ) mice. In 774 single B220 + cells from Hspa13 cKO mice, we observed that 3.49% had single-cell PBs, and in 1,025 single B220 + cells from the control mice, we observed that 1.07% had single-cell PBs. These data suggest that the Hspa13 cKO reduced the number of SRC-induced PBs (Figure 7A). Further analysis of the IgD, IgM, FIGURE 7 | Hspa13 cKO reduced class switch recombination (CSR), somatic hypermutation (SHM), and affinity maturation of antibodies. Nine-week-old female Hspa13 fl/fl (control) and CD19 cre Hspa13 fl/fl (Hspa13 cKO) mice (three mice per group) were injected i.p. with 1 × 10 9 SRCs (A-C) or NP-KLH (D,E) on days 0 and 7. On day 21 following SRC stimulation, splenocytes were stained with PerCP-conjugated anti-mouse B220 antibodies and sorted by FACS. Single cells were captured using the 10 X Genomics Full Chromium platform and subjected to RNA-and VDJ-sequencing. (A) Hspa13 cKO reduced SRC-induced PBs. Of the single PBs, 27 (3.49%) and 11 (1.07%) (Ighm + , Ighg1 + , Ighg2b + , Ighg2c + , Ighg3 + , Igha + , or IgG1, IgG2b, IgG2c, IgG3, IgA, and IgE isotypes showed that the Hspa13 cKO reduced levels of SRC-induced class-switched antibodies (e.g., IgG1, IgG2b, IgG2c, and IgG3) ( Figure 7B). Furthermore, single-cell VDJ-sequencing was used to assess the somatic hypermutation (SHM) in the CDR (complementaritydetermining region) of the heavy (H) and light (L) chains of 734 and 382 antibody genes from SRC-induced Hspa13 cKO and control mice, respectively. Extensive somatic mutation occurred in the CDR of the H and L chains in the controls, but it was clearly reduced in the Hspa13 cKO group (Figure 7C). These results suggest that the Hspa13 cKO reduced the SRC-induced SHM. In addition, we analyzed the NP-specific SHM in unique FIGURE 8 | Hspa13 interacts with endoplasmic reticulum (ER) proteins involved in positive regulation of protein transport from the ER to the cytosol. Splenic B220 + B cells from 9-week-old C57BL/6 mice (three mice per group) were sorted using B220 microbeads and were then stimulated for 3 days with 10 µg/ml LPS. LPS-stimulated B cells were collected for anti-Hspa13 antibody co-immunoprecipitation (IP) experiments. Co-immunoprecipitated proteins were identified by mass spectrometry. (A) SDS-PAGE and silver staining results showing affinity captured interacting proteins from whole cell extracts. Putative interacting protein bands marked as 1, 2, 3, and 4 were excised for mass spectrometry analysis. (B) Anti-Hspa13 antibody specifically co-immunoprecipitated 56 proteins. We identified 393, 56, and 148 proteins that were co-immunoprecipitated by control IgG antibodies only, anti-Hspa13 antibodies only, and the two antibodies together, respectively. (C) Bar plot ranking of the top 10 cellular components (CC), based on enrichment score. Gene ontology (GO)-analysis was performed using a gene ontology website (http://www.geneontology.org/). (D) Bar plot ranking of the top 10 biologic processes (BP) based on enrichment score. GO-analysis was performed based on the gene ontology website (http://www.geneontology.org/). (E) A list of the 10 best hits of Hspa13 interacting partners. Detailed interacting protein data are shown in Supplementary Table 1. (F) Interaction of Hspa13 and Bcap31. The recombinant plasmids expressing Hspa13-V5 and Bcap31-Flag were transiently transfected into 293T cells. At 48 hrs after transfection, cells were lysed and anti-V5 antibody was used to immunoprecipitate proteins probed with anti-Flag antibody. Data are shown for one representative experiment from three independent experiments with similar results. (G) Hspa13 cKO reduced Bcap31 mRNA expression in PBs induced by SRC. Bcap31 mRNA expression was analyzed from 10 and 8 PBs from the single-cell RNA sequencing data of SRC-primed Hspa13 fl/fl and CD19 cre Hspa13 fl/fl mice, respectively, described in Figure 7. Student's t-test (two tailed). Error bars represent s.e.m. *P < 0.05. clones (VH186.2 segment) that had been induced by NP-KLH. Compared with the control group, Hspa13 cKO group had a lower mutational load in the V186.2 regions (Figure 7D). These results suggest that the Hspa13 cKO reduced the NPspecific SHM induced by the NP-KLH. Finally, we analyzed the NP-specific high-affinity clones from purified GC B cells that contained the W33L mutation in CDR1. The data demonstrated that the Hspa13 cKO reduced the number of NP-specific highaffinity clones induced by the NP-KLH ( Figure 7E). Collectively, these data suggest that the Hspa13 cKO reduced the number of class-switched and somatically hypermutated antibodies with defective affinity maturation.

Hspa13 Interacts With Proteins in the ER to Positively Regulate Protein Transport From the ER to the Cytosol
To explore the mechanisms underlying the production of Hspa13-regulated PBs and PCs, and antibodies, we used anti-Hspa13 antibodies to co-immunoprecipitate (IP) proteins that interact with Hspa13 in PCs induced by LPS. The results of SDS-PAGE and silver staining show affinity-captured interacting proteins from whole cell extracts ( Figure 8A). Putative interacting protein bands, marked as 1, 2, 3, and 4, were excised for mass spectrometry analysis ( Figure 8A). We identified  393, 56, and 148 proteins that were co-immunoprecipitated by the control IgG antibody, the anti-Hspa13 antibody, and both antibodies, respectively ( Figure 8B). Gene ontology (GO)analysis was performed based on the gene ontology website (http://www.geneontology.org/). Using the enrichment score, we determined the top 10 cellular components and biologic processes, which are listed in Figures 8C,D, respectively. GOanalysis suggests that Hspa13 interacts with proteins in the ER to positively regulate protein transport from the ER to the cytosol. The 10 best hits of the Hspa13 interacting partners are listed in Figure 8E. Detailed interacting protein data are shown in Supplementary Table 1. The expression of the most interesting target Bcap31 has been confirmed at the protein level by western blotting and co-IP experiments with a tagged, transfected target gene ( Figure 8F). Finally, the Bcap31 mRNA expression was analyzed from the single-cell RNA sequencing data of the SRC-primed Hspa13 fl/fl mice (10 PBs) and CD19 cre Hspa13 fl/fl mice (8 PBs) (Figure 7). The results suggest that the Hspa13 cKO reduced the Bcap31 mRNA expression ( Figure 8G). Collectively, our data suggest that Hspa13 interacts with proteins (e.g., Bcap31) in the ER to positively regulate protein transport from the ER to the cytosol.

Increased Hspa13 Expression in B220 + Cells From Patients With MM or SLE
To explore the Hspa13 expression in PC-related diseases, two such diseases (MM and SLE) were studied in this work. CD19 + cells from peripheral blood monocytes of healthy donors and patients with MM or SLE were sorted using CD19 microbeads, and an RNA-sequencing assay was used to determine the transcript sequences. The results demonstrated that PC-promoting genes including Xbp1 and Sdc1 (Cd138) were increased in patients with MM ( Table 5) or SLE ( Table 6). These data suggest that there were more PCs in MM patients. Accordingly, we found that Hspa13 mRNA was also increased in B220 + cells from patients with MM (Table 5) or SLE ( Table 6).
To explore the role of Hspa13 in SLE, two lupus mouse models were used. The hydrocarbon oil 2, 6, 10, 14tetramethylpentadecane (TMPD; also known as pristane)induced experimental lupus mice displayed some important immunologic and clinical features that are similar to those in human SLE (38)(39)(40). We found that Hspa13 cKO reduced the number of autoantibodies ( Figure 9A) and level of proteinuria ( Figure 9B) in the mouse model with pristane-induced lupus. MRL/lpr mice are considered a good spontaneous model of human SLE diseases (41,42). We found that Hspa13 cKO reduced the number of autoantibodies ( Figure 9C) and the level of proteinuria (Figure 9D) in the lupus-prone MRL/lpr mouse model. Collectively, our data suggest that Hspa13 was increased in PC-associated diseases (e.g., MM and SLE), whereas the B-cell-specific KO of Hspa13 reduced the production of autoantibodies and proteinuria in the lupus mouse model.

DISCUSSION
PCs play an important role in both MM and SLE. However, there is still not an effective way to control PCs. In this work, we showed that Hspa13 was increased in B220 + B cells from patients with MM or SLE. PBs and PCs (but not naïve B cells or GC B cells) expressed high levels of Hspa13. The B-cell-specific KO of Hspa13 reduced the production of PBs and PCs, and the secretion of antibodies. These results suggest that patients with PC-associated diseases (e.g., MM and SLE) may benefit from treatments based on Hspa13.
Published reports regarding the role of HSPs in PCs are limited. One previously published paper showed that Hsp90 inhibitors induced the UPR to reduce abnormal immunoglobulin production and finally resulted in myeloma cell death (26). In this work, we found that the B-cell specific knock-out of Hspa13 reduced the numbers of PBs and PCs and the levels of antibodies in mice (Figure 3). In addition, the Hspa13 cKO reduced the PB, PC, and antibody production induced by LPS (Figure 4), SRCs (Figure 5), and NP-Ficoll and NP-KLH (Figure 6). Collectively, these data demonstrated that the Hspa13 cKO reduced the production of PBs, PCs, and antibodies, suggesting that Hspa13 may be an effective target for PCs.
We found that 100% of PBs expressed Hspa13, whereas only a few immature, mature, memory, and GC B cells expressed low levels of Hspa13 (Table 3). In addition, Hspa13 mRNA and protein were highly expressed in PBs and PCs but not in naïve B cells or GC B cells (Figures 1D,E). These results suggest that PBs and PCs expressed high levels of Hspa13 mRNA. Since the discovery of Hspa13 (Stch) in 1994 (43), only PBs and PCs have been shown to express Hspa13. This suggests that targeting Hspa13 in PBs and PCs may not result in serious side effects.
Our present study and previous studies demonstrated that atacicept (TACI-IgG) (16,17) and LPS (29,30,40) resulted in an increase of terminally differentiated PCs. In this work, we found that the expression of PC-promoting genes, including Prdm1 (Blimp1) and Xbp1, and of Hspa13, was up-regulated by atacicept ( Table 1) and LPS ( Table 2), whereas the Hspa13 cKO reduced the levels of LPS-induced Prdm1 and Xbp1 mRNA ( Table 4). Because PBs and PCs express high levels of Hspa13, Hspa13 expression may be positively associated with the PCpromoting genes Prdm1 and Xbp1. In fact, when the numbers of PBs and PCs are reduced, the Hspa13 is also reduced, and in turn, when the Hspa13 is reduced, the number of PCs is reduced.
High-affinity antibody production is a critical step in longterm immune responses (44). The mediation of SHM and CSR by activation-induced cytidine deaminase (Aicda, AID) is an important step in the generation of high-affinity responses (45). Following B-cell activation, CSR of IgM into IgG, IgE, or IgA occurs rapidly (46,47). All of the SHM, CSR, and affinity maturation occurs in the germinal center (48). We found that Hspa13 was not expressed in GC B cells (Tables 1-3, and Figure 1) and that the Hspa13 cKO did not affect Bcell activation or GC B-cell production (Figures 3-5). However, the Hspa13 cKO reduced the class-switched ( Figure 7B) and somatically hypermutated (Figures 7C,D) antibodies with defective affinity maturation ( Figure 7E). This may have been because the Hspa13 cKO reduced the production of PBs and PCs (Figures 3-5).
A co-IP assay was used to identify 56 proteins that interact with Hspa13 (Figure 8 and Supplementary Table 1). GO-analysis and co-IP experiments (Figure 8) suggest that Hspa13 interacts with proteins (e.g., Bcap31) in the ER to positively regulate protein transport from the ER to the cytosol. Our results are consistent with previous studies that suggested that Hspa13 (Stch) belongs to the HSP70 family with ATPase that aids in the production of cytosolic and secretory proteins (43). Importantly, our experiments also demonstrated that Hspa13 regulates PB, PC, and antibody production (Figures 3-6). These results suggest that Hspa13 regulates the production of PBs, PCs, and antibodies by regulating the protein (e.g., antibody) transport from the ER to the cytosol.
To explore the role of Hspa13 in PC-related diseases, we used an RNA-sequencing assay to evaluate Hspa13 expression. The results demonstrated that the expression of PC-promoting genes, including Xbp1 and Sdc1 (Cd138), was increased in patients with MM ( Table 5) or SLE (Table 6). Accordingly, we also found that the Hspa13 mRNA was increased in B220 + cells from patients with MM (Table 5) or SLE ( Table 6). Critically, the Hspa13 cKO reduced the number of autoantibodies and the level of proteinuria in the pristane-induced lupus and lupusprone MRL/lpr mouse models (Figure 9). Autoantibodies have been shown to be related to pathologic features of SLE such as lymphopenia and proteinuria (49). This suggests that targeting of the Hspa13 may be a productive strategy for treating patients with PC-related diseases (e.g., SLE).
In conclusion, PBs and PCs expressed high levels of Hspa13, which was increased in MM and SLE, whereas the B-cell-specific KO of Hspa13 reduced the production of PBs and PCs and the secretion of antibodies. Thus, the targeting of Hspa13 may be a productive strategy for treating patients with PC-related diseases (e.g., SLE).

ETHICS STATEMENT
The animal study was reviewed and approved by the Animal Ethics Committee of the Beijing Institute of Basic Medical Sciences.

AUTHOR CONTRIBUTIONS
YH, RX, BZ, YF, CH, and CX performed the experiments. HX, GC, XW, NM, and GH contributed essential reagents and materials for the experiments. RW conceived of and designed the studies. RX and RW contributed to data analysis and manuscript preparation. All authors have read and approved the final manuscript.

FUNDING
This study was supported by grants from National Nature and Science Funds (31770956) and Beijing Natural Science Foundation (7182121).