The acidic protein rich in leucines Anp32b is an immunomodulator of inflammation in mice

ANP32B belongs to a family of evolutionary conserved acidic nuclear phosphoproteins (ANP32A-H). Family members have been described as multifunctional regulatory proteins and proto-oncogenic factors affecting embryonic development, cell proliferation, apoptosis, and gene expression at various levels. Involvement of ANP32B in multiple processes of cellular life is reflected by the previous finding that systemic gene knockout (KO) of Anp32b leads to embryonic lethality in mice. Here, we demonstrate that a conditional KO of Anp32b is well tolerated in adult animals. However, after immune activation splenocytes isolated from Anp32b KO mice showed a strong commitment towards Th17 immune responses. Therefore, we further analyzed the respective animals in vivo using an experimental autoimmune encephalomyelitis (EAE) model. Interestingly, an exacerbated clinical score was observed in the Anp32b KO mice. This was accompanied by the finding that animal-derived T lymphocytes were in a more activated state, and RNA sequencing analyses revealed hyperactivation of several T lymphocyte-associated immune modulatory pathways, attended by significant upregulation of Tfh cell numbers that altogether might explain the observed strong autoreactive processes. Therefore, Anp32b appears to fulfill a role in regulating adequate adaptive immune responses and, hence, may be involved in dysregulation of pathways leading to autoimmune disorders and/or immune deficiencies.

ANP32 proteins belong to a relatively new protein family of evolutionary conserved acidic nuclear phosphoproteins (ANP32A-H). Members of this family are characterized by an amino-terminal domain, harboring several leucine-rich repeats (LRR) and an acidic carboxyterminus. The LRR structures -generating hydrophobic regions -supply bona fide surfaces for protein-protein interactions that may participate in intracellular signal transduction 1 . In silico studies identified seven human family members located on chromosomes 1, 4, 9, 12 and 15 2,3 . ANP32 proteins are not expressed in archaea or eubacteria, but are expressed in all eukaryotes. Several studies showed a high conservation of these proteins, not only between the individual family members, but also between the various eukaryotic species [4][5][6] . These studies imply that ANP32 proteins originate from one single ancestor gene.
Interestingly, ANP32A, B and E proteins harbor classical basic nuclear localization signals (NLS) for importin-dependent nuclear import and one or two nuclear export signals (NES) for CRM1-dependent nuclear exit [7][8][9][10] . These ANP32 family members are expressed at varying levels in several tissues and cell types. For example, in murine and human samples strong expression was detected in brain, lung, heart, kidney, muscles, intestine, stomach, liver, pancreas, leucocytes, and prostate tissue, as well as in highly proliferative organs such as the spleen, thymus and placenta 1,[11][12][13][14] .
ANP32E was originally described in Xenopus laevis oocytes, and was subsequently also identified in murine tissues 11,14,15 . The localization and expression pattern of this protein in the cerebellum suggested a function during neuronal differentiation. Furthermore, its inhibitory potential for protein phosphatase PP2A suggested a possible  www.nature.com/scientificreports www.nature.com/scientificreports/ to the general apoptosis rate in splenocytes (Fig. S4), the suppressive activity of regulatory T cells (Fig. S5), and, upon in vitro activation, the overall proliferation of lymphocytes and the differentiation of T cell subsets (Fig. S6).

Lowered Th2 Cell Numbers and Increased Th17
Cell Differentiation in Anp32b KO Mice. So far, under unstressed, and naïve circumstances no obvious effect on tissue pathology, lymphocyte biology, apoptosis or proliferation of hematopoietic cells was detectable in the Anp32b KO animals. Therefore, to analyze Anp32b KO under immune stimulatory conditions, isolated splenocyte derived CD4 + T cells were activated in vitro using PMA and ionomycin. The stimulated cells were subsequently incubated with Brefeldin A (BD GolgiPlug) to trap cytokines in the endoplasmic reticulum. Moreover, IFN-γ, IL-4, IL17-A and Foxp3 expression levels were determined to detect Th1, Th2, TH17 or regulatory T cells. The data obtained demonstrated that CD4 + T cells derived from Anp32b KO animals were characterized by a significant increase towards Th17-specific immune responses, while at the same time, Th2-specific immune responses were reduced (Fig. 4). However, analysis of the capability of naïve CD4 + CD62L + splenic cells to proliferate into TH2 specific cells by adding polarizing cytokines, showed no significant impact onto this development per se (data not shown). Furthermore, Anp32b knockout mice displayed normal numbers of Th2 cells within the CD4 + T cell population. This indicated that Anp32b deficiency may influence the number of Th17 cells in naïve Anp32b KO animals.

Strong Enhancement of Experimental Autoimmune Encephalitis in Anp32b KO Mice.
Taking into account that Th17/Th1 T cells play a crucial role in autoimmunity, and to unravel a possible role of Anp32b in a more complex and inflammatory immunological environment, we next decided to investigate Anp32b KO animals using the experimental autoimmune encephalitis (EAE) model 49 . Interestingly, immunization of KO mice with myelin oligodendrocyte glycoprotein (MOG)-peptide led to faster appearance of disease symptoms (i.e. paralyses) and to a clearly higher clinical score with ranking significance (Fig. 5a). In line with increased EAE symptoms Anp32b KO mice show more CD45 + -infiltrates in the brain (Fig. S10).
Furthermore, also the recovery from EAE associated symptoms is inhibited in Anp32b KO mice, as compared to control animals. These data indicated that Anp32b indeed plays a crucial role during the onset, severity and duration of inflammation. The activation status of splenic CD4 + or (d) CD8 + peripheral T cells as well as the activation status of (e) B220 + B cells was further analyzed with respect to naive and effector memory lymphocyte subsets of the spleen. (f) Splenic cells were further analyzed for percentages of CD4 + CD25 + Foxp3 + T regs and CD4 + CD25 + Foxp3 − within the CD4 + population. Data are mean ± SD. Anp32b WT (n = 4) and Anp32b KO (n = 4). *p < 0.05; Mann-Whitney U test. ( ther investigate the effects in Anp32b KO animals, splenocytes were isolated from EAE-mice (wild type and KO) at day 30 post EAE induction and then restimulated with MOG-peptide in vitro. When compared to control cells, stimulation led to stronger proliferation of Anp32b KO splenocytes (Fig. 5b). Furthermore, secretion of the proinflammatory cytokines TNF-α, INF-γ, IL-17A and IL-21 was strongly enhanced in these cells (Fig. 5c). Subsequent flow cytometry analyses of cells revealed minimally reduced CD3 + lymphocyte abundance, but reduced levels of naïve CD4 + T cells (CD4 + /CD62L + ) and somewhat enhanced numbers of activated CD4 + CD44 + splenocytes (data not shown). However, long-term monitoring of splenic lymphocyte populations for 60 days (Fig. 6) was accompanied by a significant decline in CD4 + T lymphocytes, which could be explained by overall reduced CD3 + cell numbers. In contrast, CD8 + cell numbers were increased in Anp32b KO mice. In addition, CD62L positive naïve CD4 + and CD8 + T lymphocytes were strongly reduced, whereas activated T lymphocytes were rather increased ( Fig. 6a-d). CD4 + CD25 + Foxp3 + T reg cell numbers are not significantly changed (Fig. 6e). However, it became obvious, that the splenic cells in general showed a more exhausted phenotype indicated by PD-1 and CTLA-4 expression (Fig. 6f). Strikingly, a higher prevalence of follicular T helper cells was observed (Fig. 6g,h), a cell type that have been reported to be involved in autoimmune disorders, as depletion of these cells reduced the pathogenic phenotype 50-52 . Enhanced Transcription of Proinflammatory Genes in Anp32b KO Animals. Next, transcriptome analyses were performed to investigate the impact of Anp32b KO during EAE at the transcriptional level by isolating mRNA from splenocytes of KO and wild type animals (at day 6 after MOG-immunization) and performing next generation sequencing. Results showed, that a large number of transcripts were dysregulated at day 6 post MOG injection in Anp32b KO mice-derived cells (see Fig. 7a). Interestingly, several mRNAs affecting immune responses, particularly transcripts enhancing the activation of effector T cells, were significantly and strongly upregulated in Anp32b KO cells after immunological stress (Fig. 7b). These included key players for T cell activation/expansion, such as IL2, IL2Ra/b, IL6R, CD4, CD5, CD37, TNFr, TLR9, Akt3/1, JAk3, and GRB7. In contrast, messages responsible for the induction of TH2 immune responses, for example IL-4, were clearly reduced. Bioinformatics of the affected transcripts revealed an impact of Anp32B knock out on huge numbers of pathways. Significantly downregulated gene onthologies (GO) (126 GOs, FDR < 0.01) included mitochondrial . Increased Th17 CD4 + T cell population in Anp32b KO mice. Splenic CD4 + T cells of Anp32b KO mice and controls were incubated for 4-6 h with 1 mg/ml ionomycin, 20 ng/ml PMA, and 1 mg/ml Brefeldin A (BD GolgiPlug). Cells were stained to determine the percentage of Th1, Th2 and Th17 cells among the CD4 + cell population. Anp32b wild type (WT; n = 6) and Anp32b KO (n = 7). Pooled data (mean ± SEM) from two independent experiments are shown. *p < 0.05, **p < 0.01, Mann-Whitney U test was used.
www.nature.com/scientificreports www.nature.com/scientificreports/ pathways involved in energy metabolism, pathways involved in ribosomal protein synthesis and overall DNA replication (data not shown). Significantly upregulated gene onthologies (GO) (540 GOs, FDR < 0.01) (Fig. 8a) included pathways for cell growth, tissue and neuronal development and regulation of cell death. In addition, further pathways frequently affect cellular signaling and are involved in various aspects of the regulation of immune responses, including lymphocyte differentiation and activation (Fig. 8b).
Thus, the observed altered transcription of proinflammatory mRNAs as well as transcripts involved in immune signaling and lymphocyte biology correlated well with the enhanced pathogenic phenotype observed in MOG-immunized Anp32b KO animals. www.nature.com/scientificreports www.nature.com/scientificreports/

Discussion
Recently, several in vitro studies demonstrated the involvement of the acidic protein rich in leucines ANP32B in the regulation of cell proliferation, apoptosis, cell cycle progression and gene expression at the epigenetic and posttranscriptional level 7,9,10,13,24,31,32,41 . In addition, involvement of ANP32B in viral infection, acting either as a critical cellular cofactor or as restriction factor, has been reported for various viruses [36][37][38][39][40] . However, most of these data were derived from in vitro studies. A systemic KO of Anp32b in mice has been previously shown to result in www.nature.com/scientificreports www.nature.com/scientificreports/ embryonic lethality 44 . Here, we succeeded in generating a conditional systemic Anp32b KO in adult animals and, hence, were able to show that Anp32b is a modulator of adaptive immune responses.
By targeting the murine Anp32b gene with loxP sites for Cre-dependent recombination, and using a tamoxifen inducible recombinase expression cassette (CAG-Cre-ER(T)) 48 all cells should be affected in Anp32b KO animals upon feeding with tamoxifen. This experimental strategy led to no obvious adverse effects in adult and immunologically unchallenged Anp32b KO animals, such as overall viability and fitness, breeding behavior, organ development, hematopoiesis, or clinical chemistry parameters. This rather wild type-like phenotype of Anp32b KO largely reflected the previously reported observations in mice, where the Anp32 family members a and e were inactivated 42,43 .
Since ANP32B has been previously correlated with expression of immunologically relevant proteins in lymphocytes and dendritic cells 7,41 , we compared the stimulatory capacity of splenocytes derived from Anp32b KO mice to control animals. Our data revealed an impact of Anp32b KO on the ratio of Th1/Th17 versus Th2 immune responses. In vitro differentiation analyses showed normal development of these cell types, thus Th2 immune responses per se are not affected, but this has to be invested in a further study. However, because of the strong Th17 response in the Anp32b KO cells the respective mice were further analyzed using an EAE model, as these cells have been correlated with autoimmune reactions. Interestingly, knockout of Anp32b led to earlier induction, stronger progression and longer extension of the autoimmune pathology. This could be correlated with hyperactivation of splenic lymphocytes at mRNA levels, protein expression, activation and exhaustion status as well as cellular differentiation levels.
The impressive and unexpected dysregulation of immunologically relevant transcript expression resulted in stronger activation of Th17 immune responses, as well as weaker Th2 responses. Furthermore, the strong upregulation of follicular Th cells in Anp32b KO mice suffering from an EAE furthermore indicate a substantial dysregulation of immunologic pathways in these animals and an involvement of the Anp32b protein in fine tuning of this complex pathways. The hyperactivated immune status, the strengthened Th17 response, especially the reduction of naïve lymphocytes and the strong exhaustion status of the T cells as well as the increased presence of Thf cells in Anp32b KO mice suffering from EAE furthermore would explain the strong progression of the autoimmune process in these animals as reflected by literature 53,54 .
In summary, we provide evidence that Anp32b is necessary to find a balance in the interplay of m-RNA expression, cellular differentiation and exhaustion status of T lymphocytes and the regulation of adequate immune responses. These findings may have an important impact on the development of autoimmune disorders, and unraveling the underlying mechanisms could lead to a more detailed insight into this complex dysregulated network and may help to find new drugable targets to treat autoimmune disorders in the future.

Methods
Animal studies. All animal procedures were approved by the Hamburg State Authority according to the German Animal Welfare Act (permission number 132/13). Mice were maintained in specific pathogen-free conditions at the Heinrich Pette Institute -Leibniz Institute for Experimental Virology Hamburg, or at the Department of Immune Modulation, Universitätsklinikum Erlangen.
Generation of conditional knockout mice. Embryonic stem cell clones derived from C57BL/6 mice for conditional Anp32b knockout (exon 4 of Anp32b gene was flanked with loxP sites, JM8.N4 #HEPD0503_1_ H01) were obtained from the EUCOMM -International Knockout Mouse Consortium 47 . Briefly, stem cells were characterized by Southern blotting and PCR methods and microinjected into BALBc blastocysts. Speckled chimeric offspring were back-bred into the C57BL/6 background. Deletion of the selection cassettes was performed by breeding the animals with Flp-deleters (C57BL/6). The resulting mouse strain was crossed with the final Cre-deleter strain (C57BL/6) to insert a heterozygous tamoxifen-inducible Cre expression cassette (CAG-Cre-ER(T)) 48 and then again bred homozygous for the "floxed" Anp32b gene. Maintenance breeding was done with male homozygous Anp32b "floxed"/heterozygous Cre animals and female homozygous Anp32b "floxed" animals. According to Mendelian distribution, half of the offspring, homozygous for "floxed" Anp32b gene, were positive for Cre and therefore could be used for Anp32b knockout induction. Tamoxifen (4-OHT) induced knockout was performed by administering 4-OHT containing nutrients (400 mg/kg, LASCRdiet CreActive TAM400, Genobios) for up to 4 weeks. For phenotyping, Cre positive and Cre negative sister animals were fed in parallele and Cre negative animals served as negative controls (wild type).
Paraffin sections and immunohistochemical staining. Immunohistochemical analyses were performed as described previously 56,57 . Briefly, organs were fixed with 2% buffered paraformaldehyde for at least 2 d at 4 °C. Femora were decalcified in 10% (w/v) EDTA in PBS for 4 d. Organs were dehydrated through serial ethanol exposures, embedded in paraffin wax and sections of 4 µm thickness were prepared using a Thermo www.nature.com/scientificreports www.nature.com/scientificreports/ Fisher microtome. Sections were rehydrated and treated for five min with Fast Enzyme (Cytomed) at room temperature for antigen retrieval. After three washes (5 min each) in TBS buffer, the primary anti-ANP32B antibody (ABIN872630; Antibodies Online) diluted 1:100 in Dako antibody diluent (Dako) was applied at room temperature in a moist chamber. Rabbit negative control antibodies (Dako X0903) served as an isotype control. After three careful washes (3 × 5 min) in TBS, antibody binding sites were located with the Dako REAL detection system alkaline Phosphatase (K5005). A light hemalumn stain was used for counterstaining. Sections were dehydrated and mounted in Eukitt (Kindler).
Brain histology. Brains of mice were harvested at day 60 of EAE, fixed in liquid nitrogen, and stored at −80 °C. Organs were embedded in Tissue-Tek (Sakura), and 5-μm sections were made using a cryotome (Kryocut CM 2000; Leica). Immunohistological staining was performed using an immunoperoxidase detection system in a humid incubation chamber. Acetone-fixed sections were incubated in PBS, and endogenous peroxidase activity was blocked by incubating sections in 3% H 2 O 2 . Sections were incubated with a primary anti-CD45 Ab (clone 30G12; provided by L. Sorokin, Lund University, Lund, Sweden). The primary Ab was detected by a streptavidin HRP-coupled goat ant-rat IgG Ab (Biocare Medical). To visualize Ags, sections were incubated with AEC Chromogen Substrate (Vector Laboratories). Sections were stained with hematoxylin (Sigma-Aldrich), mounted in Aquatex (Merck), covered with a coverslip, and examined by light microscopy (Leica).
Bone marrow histology. For bone marrow histology mouse femurs were isolated and placed in 10% buffered formalin. After decalcification the tissues were embedded in paraffin, sectioned, and stained using standard Wright-Giemsa protocol. Histology images were taken with a Zeiss Axioplan 2 microscope (Carl Zeiss).
Isolation of mouse splenocytes. Isolated spleens were disrupted in ice cold buffer (RoboSep buffer, Stemcell Technologies). Cell suspensions were prepared by percolation of the spleen fragments using Cell Strainers (BD Falcon). Immediately, cells were harvested by centrifugation (150 × g, 5 min, 4 °C), suspended in suitable medium and cultured at ambient conditions or used directly for FACS analysis.
Protein analyses. Western blot analysis was performed as described previously 41,58 . Briefly, crude extracts were prepared by lysis of the cells with lysisbuffer A (0.1% NP40, 150 mM NaCl, 50 mM Hepes at pH 7.3, protease inhibitors) followed by denaturation in SDS-buffer (65 mM Tris-HCl pH 6.8, 2% SDS, 10% glycerol, 1.5% DTT). SDS-PAGE was carried out using 13% polyacrylate-gels and Western blot analysis was performed with polyclonal rabbit antiserum specific for ANP32B (described in 7 ). Transcriptome analysis. Total cellular RNA was isolated according to the manufacturer's protocol using TRIzol reagent (Invitrogen). cDNA libraries were prepared with NEBNext Ultra II DNA Library Prep Kit for Illumina and NEBNext Multiplex Oligonucleotides for Illumina (New England Biolabs) according to manufacturer's instructions. Sequencing was performed on the Illumina NextSeq system. Sequencing reads were aligned against the mouse reference assembly (GRCm38.p5) and quantified using STAR (v2.5.2b) 59 . Based on these counts, statistical analysis of differential expression was carried out with DESeq. 2 (v1.14.0) 60 . Mixed lymphocyte reaction. Bone marrow-derived DCs were generated from BALB/c mice, as described elsewhere 61 . At day 9, bone marrow-DC cultures were matured with LPS overnight or were left unstimulated. A total of 4 × 10 6 allogeneic splenic cells, from Anp32b wild type or Anp32b KO mice were incubated with differentially stimulated DCs in a 96-well flat-bottom plate (Falcon) for 72 h at 37 °C and 5% CO2. Then, 1 μCi/well [ 3 H]-methyl-thymidine (Amersham Biosciences) was added, and cells were incubated for an additional 16 h. The plates were harvested with an IH-110 harvester (InnoTech), and filters were counted in a 1450 Microplate Counter (Wallac) to detect incorporated radioactivity as readout for proliferation.
www.nature.com/scientificreports www.nature.com/scientificreports/ Additionally, 200 ng of pertussis toxin (List/Quadratec) was injected i.p. at days 0 and 2 of EAE. Scoring of EAE symptoms was performed as follows: 0, no disease; 1, tail weakness; 2, paraparesis; 3, paraplegia; 4, paraplegia with forelimb weakness; and 5, moribund or dead animals. MOG restimulation. At day 30 or day 60 after MOG injection, splenic cells from mice were harvested, and 4 × 10 5 total spleen cells or purified CD4 + splenic cells were incubated with different concentrations of MOG peptide and/or 50 U/ml IL-2 (Proleukin) in 200 μl RPMI medium supplemented with 10% FCS, 100 U/ml penicillin, 100 μg/ml streptomycin, 2 mM L-glutamine, and 50 μM 2-ME (all additives were from Sigma-Aldrich) per well in a 96-well tissue culture plate for 72 h at 37 °C and 5% CO2. Then, 0.1 μCi/well [3H] methyl-thymidine (Hartmann Analytic) was added and cells were incubated for an additional 12 h. Subsequently, the plates were harvested with an IH-110 harvester (InnoTech), and filters were counted in a 1450 Microplate Counter (Wallac) to detect incorporated radioactivity as readout for proliferation. Supernatants from all restimulated cultures were harvested to determine ex vivo cytokine production.
Cytokine quantification/ELISA. To determine cytokine production in supernatants of MOG-restimulated DCs in T cell co-cultures, a commercially Cytometric Bead Array (BD Biosciences) was used to detect IL-17A, IFN-γ, IL-21, and TNF-α, according to the manufacturer's instructions.
Stimulation of CD4 + CD62L + T Cells. CD4 + CD62L + naïve T cells were obtained from spleen and lymph node of naive mice using a CD4 + CD62L + T Cell Isolation Kit II (Miltenyi Biotec), according to the manufacturer's instructions. Cells were stimulated with 1 μg/ml ionomycin and 20 ng/ml PMA, as well as 1 μg/ml Brefeldin A (BD GolgiPlug), and incubated for 6 h at 37 °C and 5% CO 2 and analyzed for the expression of intracellular cytokines IL-17A, IFN-γ, IL-4 and Foxp3.
Suppression assay. For the suppression assay, T eff and T regs were cultured on feeder cells derived from RAG1 −/− mice. These cells were treated with Low-Tox ® -M Rabbit Complement (Cedarlane) and an anti-mouse CD90.2 antibody (30-H12; BioLegend). Afterwards, 5 × 10 4 feeder cells were seeded into 96-well roundbottom plates (BD Falcon). CD4 + CD25 − Teff and CD4 + CD25 + Tregs were isolated from naive Anp32b KO and wild type mice or EAE treated KO an wild type animals (day 60) using the mouse CD4 + CD25 + Regulatory T cell Isolation Kit (Miltenyi Biotec) according to the manufacturer's instructions. Purity of sorted cells was assessed by FACS directly after isolation and revealed ≥92%. T eff were labeled with 0.75 µM CellTraceTM Violett (CellTraceTM Violett Cell Proliferation Kit; molecular probes by life technologiesTM) per 10 6 cells and 5 × 10 4 labeled T eff per well were added to the RAG1 −/− feeder cells. Finally, T regs were plated onto these cells in titrated numbers and co-cultures were stimulated using 0.25 mg/ml of an anti-mouse CD3ε antibody (145-2C11; BioLegend) and incubated for 72 h before percentage of proliferating cells was determined by FACS. Results were analyzed using FCS Express Flow Cytometry Software (De Novo).

Statistics.
Results are expressed as mean and SEM, unless stated otherwise in the figure legends. A two-tailed Student t test or and two-way ANOVA with the Bonferroni posttest (for multiple groups) were used to determine statistical significance, unless stated otherwise in the figure legends. A nonparametric Mann-Whitney U test was used to analyze clinical EAE scoring. All statistical tests were two-sided and differences with p values of *p < 0.05, **p < 0.01 and ***p < 0.001 were considered statistically significant. Statistical analyses were performed with Excel or Prism 4.03 or 5.0 (GraphPad).

Data Availability
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.