Survivin promotes a glycolytic switch in CD4+ T cells by suppressing the transcription of PFKFB3 in rheumatoid arthritis

Summary In this study, we explore the role of nuclear survivin in maintaining the effector phenotype of IFNγ-producing T cells acting through the transcriptional control of glucose utilization. High expression of survivin in CD4+T cells was associated with IFNγ-dependent phenotype and anaerobic glycolysis. Transcriptome of CD4+ cells and sequencing of survivin-bound chromatin showed that nuclear survivin had a genome-wide and motif-specific binding to regulatory regions of the genes controlling cell metabolism. Survivin coprecipitates with transcription factors IRF1 and SMAD3, which repressed the transcription of the metabolic check-point enzyme phosphofructokinase 2 gene PFKFB3 and promoted anaerobic glycolysis. Combining transcriptome analyses of CD4+ cells and functional studies in glucose metabolism, we demonstrated that the inhibition of survivin reverted PFKFB3 production, inhibited glucose uptake, and reduces interferon effects in CD4+ cells. These results present a survivin-dependent mechanism in coordinating the metabolic adaptation of CD4+T cells and propose an attractive strategy to counteract IFNγ-dependent inflammation in autoimmunity.


INTRODUCTION
Activated CD4 + effector T cells are key players in autoimmune inflammation. These cells migrate, proliferate, and produce signal molecules at sites of inflammation to mobilize immunity. Production of IFNg, the principal coordinator of adaptive immune responses in chronic inflammation, is the major characteristic feature of the effector T cells. 1 To fuel effector responses, IFNg producing cells undergo metabolic adaptation by switching glucose metabolism from entering the tricarboxylic acid (TCA) cycle to the pentose phosphate pathway of glycolysis thereby increasing availability of nucleotides, amino acids and fatty acids. [2][3][4][5][6] The switch from TCA to a pentose phosphate-dependent utilization of glucose is an emergency act described in macrophages, T cells and neutrophils, which is maintained by through a high glucose consumption. [7][8][9][10] IFNg appeared to be particularly sensitive to cellular metabolic state and deletion of glucose transporter GLUT1 and lactate dehydrogenese (LDHA) reduced IFNg production. 11,12 IFNg dependent processes, such as inflammation and migration, initiated by the activation of IFNg receptor, production of IFN-responsive factors (IRFs) and binding of the IRF-specific regulatory elements on chromatin to trigger the production of IFN-sensitive genes. 1 Expression of the IFN-sensitive genes regulates the impact of IFNg in pro-inflammatory effector functions, and in anti-TGFb fibrotic processes that maintain autoimmunity.
Shared IFNg-dependent processes are central for the pathogenesis of autoimmune diseases. 1,7,13,14 The pleiotropy of IFNg functions provides a broad spectrum of biological effects ascribed to this cytokine in different autoimmune diseases and even in different stages of the same condition alternating between immunostimulatory and immunosuppressive effects. 13,[15][16][17] Strategies to interfere with autoimmunity by targeting concordant changes in the expression of IFN-sensitive genes in blood leukocytes and target tissues may have broad therapeutic potential for immunological disorders. Inhibition of anabolic adaptation, which fuels IFNg production, constitutes a promising approach toward mitigating the effects of IFNg in autoimmunity.  -log10p   IL10  PFKM  HK2  PFKP  BIRC5  ENO1  ACLY  LDHA  IFNG  TNF  TBX21  ALDOA  G6PD  PRF1  GZMB  IRF1  SELL  FOXP3  RORC HK2  PFKP  BIRC5  ENO1  ACLY  LDHA  IFNG  TNF  TBX21  ALDOA  G6PD  PRF1  GZMB  IRF1  SELL  FOXP3  RORC  PFKFB3   IL10  PFKM  HK2  PFKP  BIRC5  ENO1  ACLY  LDHA  IFNG  TNF  TBX21  ALDOA  G6PD  PRF1  GZMB  IRF1  SELL  FOXP3  Survivin, an oncoprotein encoded by BIRC5, is widely expressed in malignancies and during renewal of nonmalignant hematopoetic cells. 18,19 Cytosolic and mitochondrial localization of survivin is tightly linked to its anti-apoptotic function, 20 while nuclear localization of survivin has been attributed to its role in the chromosomal passenger complex 21 and in the formation of macromolecular complexes, potentially supporting gene expression. 18,[22][23][24] Shuttling of survivin between cytosol and nucleus is assisted by exportin 1. 25,26 Conditional deletion of survivin in hematopoietic progenitors 27 and in thymocytes reduces mature CD4 + and CD8 + T cell populations 28 and leads to a dysfunctional T-cell receptor and inability to mount a proper immune response to an antigen challenge. 29 Survivin expression declines in mature T cells, but re-appears during critical phases of phenotype transition, such as the effector phenotype acquisition by CD4 + or CD8 + memory T cells. 30 Accumulation of survivin in tissues and extracellular compartment is associated with severe autoimmune inflammation in rheumatoid arthritis, 31 cutaneous psoriasis, and multiple sclerosis. 32 Our earlier studies demonstrated that targeting survivin in experimental and clinical autoimmunity efficiently reduces inflammation, proliferation, and tissue damage. [33][34][35][36][37] However, despite its importance in leukocyte development and disease, the role of survivin in the basic processes of T cell homeostasis has not been investigated.
In this study, we explored the role of nuclear survivin in maintaining the effector phenotype in IFNg-producing Th1 cells acting through the transcriptional control of glucose utilization. To study this, we performed a genome-wide deep sequencing of survivin precipitated chromatin regions; identified survivin interactors on chromatin, and the biological processes regulated by survivin in cooperation with the identified interactors. Combining chromatin and transcriptome analyses with functional studies, we have searched for the genes sensitive to survivin inhibition and present a previously unknown survivin-dependent mechanism that coordinates metabolic adaptations during the activation of CD4 + T cells in autoimmunity.

RESULTS
Survivin is an essential marker of the IFNg-producing cell phenotype Survivin expressing CD4 + T cells were identified by flow cytometry of the mononuclear leukocytes from the peripheral blood of 22 (16 female, 6 male) patients with rheumatoid arthritis (RA) ( Table S1). The gating strategy of T cell subsets is shown in Figure 1A. We found that the effector cells (T EFF ) defined as CD62L neg CD45RA +/À CD27 neg had higher levels of survivin than memory cells. On average, 9.2% (range 5.4-16.4%) of T EFF cells contained survivin and had highest amount of survivin per cell ( Figure 1A). A different set of CD4 + T cells isolated from 24 patients with RA (all female) was used to investigate the phenotype of survivin-producing CD4 + T cells by RNA-seq analysis. Unsupervised clustering of the RNAseq datasets by the core genes characteristic of T-helper subsets 38 identified the accumulation of survivin/BIRC5 in the T EFF cluster marked by expression of Th1 signature genes (e.g., TBX21, EOMES, IL2RA, and IFNG) (Figures 1B and S1A) and cytokines (IFNg, IL9 and IL10) ( Figure 1C), which correlated with BIRC5 expression ( Figure S1B). Comparison of BIRC5 hi and BIRC5 lo CD4 + cells revealed the complete Th1 signature to be enriched in the BIRC5 hi cells ( Figure 1D).
Availability and efficient metabolism of glucose are required for IFNg production and effector T cell function. 8,11 Expression of the main glucose metabolism regulator HIF-1a differed between BIRC5 hi and BIRC5 lo CD4 + cells, but their expression of MYC and MTOR was similar ( Figure 1E). Since HIF1A expression is controlled by hypoxia, the selective enrichment for HIF1A in BIRC5 hi cells prompted us to evaluate other genes of the hypoxia signature. 39 We found that BIRC5 hi cells overexpress the canonical HIF-1a target genes, including lactate dehydrogenase (LDHA), enolase (ENO1), phosphoglycerate kinase 1 (PGK1), and aldolase A (ALDOA), associated with glucose metabolism ( Figure S1C). Specifically, BIRC5 hi cells had a reduction in the key regulator of glucose processing the phospho-fructokinase 2, encoded by PFKFB3 ( Figure 1E) suggesting it's deficiency. As a result, glucose was shunted to the pentose phosphate pathway, as reflected by increased expression of glucose-6-phosphate dehydrogenase (G6PD) and ATP citrate lyase (ACLY), favoring active fatty acid metabolism. The correlation matrix of the core Th1 genes iScience Article and glycolysis markers revealed clear divergence in glucose utilization between BIRC5 hi and BIRC5 lo cells ( Figure 1F). The tight interactions in BIRC5 hi cells suggested that survivin expression is functionally connected to these processes.

Survivin-bound chromatin is annotated to regulators of glucose metabolism
Since survivin has been previously reported to bind to genomic DNA elements that regulate gene transcription, [22][23][24] we performed the chromatin immunoprecipitation sequencing (ChIP-seq) analysis of 12 CD4 + cell cultures pooled in 4 replicates, which revealed 13704 nonredundant survivin-ChIP peaks (enrichment against input, adjusted p < 10 À5 ) ( Figure 2A). The peaks were unevenly distributed across the genome and were specifically accumulated in the chromatin areas within 10-100 kb distance from the cis-regulatory elements (RE) occupied by promoters, enhancers, chromatin insulator regions, and CTCF binding sites ( Figures 2B and 2C).
To characterize the TF landscape of the survivin-ChIP peaks, we used the global ChIP-seq dataset for 1034 human transcriptional regulators in the ReMap database. 40     iScience Article overlapping peaks 10%) and 100-kb flanking regions as compared with the regions within 1 Mb around the peaks ( Figure 2D). The q significance of association with survivin was higher for TFs in the regions of 0-100 kb and lower for TFs within 1Mb.
To identify TF binding sites in open chromatin of CD4 + cells, we used the ATAC-seq dataset (GSE138767 40 ) to annotate unique nonredundant survivin-ChIP peaks. Analysis of the survivin-ChIP peaks within the open chormatin demostrates that survivin is tightly associated with a subset of TFs identified by the wholegenome analysis ( Figure 2D, inset). The strength of the association defined by q-significance, did not differ between the chromatin regions accessible at 2 and 4 h. The top TFs identified by both analyses were those regulators of glucose and insulin metabolism, including CREBBP, 41 KDM5B, 42 FOXK2, 43,44 CTBP1, 45 and IKZF1. 46 To identify biological processes regulated by the survivin-bound chromatin, we analyzed biological functions of the 146 TFs that co-localized with survivin peaks and annotated them to 2749 protein-coding genes expressed in CD4 + cells (RNA-seq, normalized raw counts >0.5) and to the Gene Ontology terms. This approach identified functional groups that regulate chromatin remodeling, protein modification, and metabolism ( Figure 2E). Other functional groups regulated the response to hypoxia and organic substances, including glucose and cytokines (Table S2). These sets of analyses demonstrated that survivin was frequently located near the cis-REs and was functionally linked to the regulation of protein and carbohydrate metabolism.

Survivin restricts PFKFB3 expression and changes the metabolic requirements of CD4 + cells
To investigate the role of survivin in the predicted biological processes, we used YM155 to inhibit survivin function 23,47 in freshly isolated CD4 + T cells. Cells were polarized with IFNg for the final 2 h. Comparison of differentially expressed genes (DE-Gs) identified by RNA-seq analysis of YM155-treated (0 and 10 nM) CD4 + cells (nominal p < 0.05, DESeq2) with those annotated to survivin peaks showed that 11.8% (24 h) and 4.5% (72 h) of the protein-coding genes expressed in CD4 + cells were sensitive to survivin inhibition ( Figure 3A). To identify TFs controlling the transcription of the DE-Gs, we used the curated TRRUST database. We found that the central metabolism regulators HIF-1a, c-MYC, and SP1 were among the transcriptional supervisors of the DE-Gs after 24 and 72 h of survivin inhibition. Other effects were attributed to the activity of SMAD4, JUN, NF-kB, RELA, ETS1 TFs at 24 h and to interferon regulatory factor 1 (IRF1) and the MHC class II transactivator at 72 h ( Figure 3B).
To study in-depth engagement of survivin in the process of cellular glucose utilization, we analyzed YM155treated and IFNg-polarized CD4 + cells by RNA-seq. We found that mRNA levels of PFKFB3 and LDHA increased rapidly in the YM155-treated cells ( Figure 3C), indicating conversion of pyruvate into lactate, while PGLS and ACLY mRNA levels were decreased, indicating the downregulation of the pentose phosphate pathway and fatty acid metabolism ( Figures 3C and 3D). These results show that survivin prevented the alterations in carbohydrate metabolism seen in the BIRC5 hi CD4 + cells from patients with RA ( Figure 1F) but did not alter the mRNA levels of HIF1A or its metabolic targets HK2, ALDOA, ENO1, and GAPDH ( Figure S2A).
The TGFb/SMAD pathway often counteracts the pro-inflammatory properties of IFNg, and SMAD4 is a predicted upstream regulator of the DE-Gs ( Figure 2B). We, therefore, investigated the effects of survivin inhibition on this pathway ( Figures 4A and 4B) and found the upregulation of (1) the E3 ubiquitin ligases SMURF2, SPSB1, SIAH3, LDLRAD4, and PMEPA1, which facilitate proteolysis required for T cell reprogramming; (2) SMAD7 and its co-repressors SKI and SKIL, which physically interact with the receptor-activated SMADs; and (3) the chromatin-binding SMAD3 co-factors JUN, FOXO1, and BACH1 ( Figure 4B). Notably, all those genes were among the most sensitive DE-Gs after survivin inhibition (nominal p < 0.005; Figure S3).
In agreement with the increased glycolytic activity of PFKFB3 and LDHA, which control the NOTCH1 and FOXO1 pathways, 51-53 survivin inhibition increased mRNA levels of FOXO1 and NOTCH1 ( Figure 4B). Consequently, CD4 + cells expressed higher levels of the surface receptors CD44, IL21R, ITGA5, and CXCR3 downstream of NOTCH1 and the FOXO1 target genes IL2RB, CCR5, CCR7, and CXCR4 ( Figure 3H), which enabled the phenotype transition of CD4 + cells.

Survivin colocalizes with IRF1 and SMAD3 on chromatin
After establishing survivin binding to chromatin, we sought to infer and validate associated protein partners through motif enrichment analysis in the region covered by survivin peaks. Using the JASPAR database of human TF, we discovered enrichment in IRF-binding motifs in all 4 independent ChIP-seq replicates. Predominant among the IRF motifs were IRF1 and IRF8, both containing the conserved GAAA repeat ( Figure 4C). The survivin peaks were also enriched in the composite motifs AP1:IRF (AICE motif, GAAAnnnTGAc/gTCA) and SPI1:IRF (EICE motif, GGAAnnGAAA). Multiple binding sites for each motif were frequently present in a single survivin peak. The ISRE motif (GRAASTGAAAST), which bound two IRFs, was also enriched compared to the whole genome, yet infrequent within the survivin peaks ( Figure 4C).
To connect survivin peaks with transcription, we annotated the whole set of unique survivin-ChIP peaks to open chromatin in human CD4 + cells activated with anti-CD3 and anti-CD28 antibodies, using ATAC-seq data (GSE138767 40 ). We found that 12.3% (2 h) and 21.5% (4 h of cell stimulation) of the peaks were located within 0-10 kb of open chromatin regions ( Figure S4A). An independent de novo motif search in those survivin peaks revealed up to 4.88-fold enrichment in the binding motifs of IRF1 and the SMAD3/SMAD4 complex, against the randomized background of all open chromatin ( Figure 4C). No enrichment in JUND and JUN motifs was found. These findings confirmed the functional specificity of survivin binding.
To validate the colocalization of survivin with the anticipated TF partners identified by the bioinformatic analysis, we utilized the human monocytic cell line THP1 and found 50-to 150-fold higher spontaneous expression of survivin compared to the conA activated primary CD4 + cells ( Figure 4E). Using immunocytochemistry with antibodies against survivin we found that survivin immunoreactivity almost exclusively in the    iScience Article nucleus ( Figure 4G). Consistent with the findings made in primary CD4 + T cells, inhibition of survivin with YM155 resulted in the upregulation of the PFKFB3 mRNA in THP1 cell culture ( Figure 4F). We immunoprecipitated survivin from the total cell lysate and nuclear extract of THP1 cells using monoclonal rabbit-antihuman survivin antibodies and total rabbit IgG for control IP. Survivin-bound proteins were affinity isolated, heat denatured, and separated by electrophoresis. Western blotting of the nuclear extract showed that IRF1 and SMAD3 co-precipitated with survivin in three independent experiments ( Figures 4H and S4D). Monoclonal antibodies to IRF1 identified a band of approximate size of 45 kDa corresponding to IRF1 with the calculated molecular weight of 37 kDa, which was not present in control IP. Antibodies targeting SMAD3 identified a band of approximate size of 50 kDa corresponding to SMAD3 with the calculated molecular weight of 48 kDa. Both, IRF1 and SMAD3 targeting antibodies revealed several additional bands, which varied in molecular weight and could be presumably explained by the presence of multiprotein complexes not resolved by disintegration step. No bands were identified in the material precipitated with control IgG ( Figure 4H). Neither IRF8 nor c-MYC, JUN, or JUND ( Figure S4D) was identified in the survivin IP material from those experiments.
To confirm reciprocally the observed co-precipitation of survivin with IRF1 and SMAD3 proteins, we performed an independent IP of THP1 nuclear material using antibodies to IRF1 and to SMAD3. Western blot of the IRF1-IP and SMAD3-IP with survivin antibodies revealed a band of approximately 20 kDa (Figure 4I), which corresponded to survivin protein monomer with molecular weight of 16.5 kDa.
Thus, survivin recruitment to open chromatin occurs through its interaction with IRF1 and SMAD3 in the regions containing sequence-specific motifs of those TFs ( Figure 4D). These results provide molecular evidence that IRF1 and SMAD3 assist and coordinate the survivin-dependent transcriptional control which is described in the functional experiments.

IRF1 and SMAD3 partner with survivin to regulate gene transcription
Since survivin-ChIP peaks accumulated in regulatory chromatin that was occupied by enhancers ( Figure 2C), we analyzed the presence of survivin peaks in the cis-REs connected to the top protein-coding DE-Gs (Figure S3). Using the likelihood score for the enhancer-gene pairing, 54 we identified 117 REs that were both connected to DE-Gs and associated with survivin peaks within 0-10 kb, and 852 REs with no survivin peaks ( Figures 5A and 5B). These two groups of REs were similar in GeneHancer (GH) score, length/size of REs, and distance to the transcription start site (TSS) ( Figure S5A).
Among the TF ChIP-seq peaks that co-localized with the survivin peaks (10% overlap, 0-kb flanks) in ChIP-seq datasets (Figures 2D), 58 TFs were expressed in CD4 + cells and were more abundant in survivin-containing REs compared to the whole genome and to the remaining REs (all p < 10 À3 ) ( Figure 5C). IRF1 and SMAD3 were among the most frequent and abundant survivin partners in REs connected to DE-Gs, as shown by density distribution analysis ( Figure 5D). Principal component analysis of the enriched TF distribution across the REs, followed by unsupervised clustering of the components ( Figure 5E) revealed that the REs clustered by the total density of TFs (TF-poor and TF-rich) rather than by gene association and further by the association of TFs around IRF1 or SMAD3 ( Figure 5E). Thus, the immunoprecipitation of survivin with IRF1 and SMAD3 suggests its participation in TF complexes with distinct functions and diverse protein compositions. Using the BioGrid database to analyze protein-protein interactions, we identified histone acetyltransferase EP300 and glycogen synthase kinase 3B as the only common interactors of IRF1 and SMAD3 ( Figure 5F). EP300, a protein that recruits TFs to distant enhancers, was enriched in survivin-containing REs and physically interacted with several other enriched TFs ( Figures 5E and 5F), providing a broad platform for building multiprotein complexes. This prediction of multiprotein interactions also finds indirect support in the differential expression of the known IRF1 and SMAD3 interactors in BIRC5 hi CD4 + cells of patients with RA ( Figures 5G, S5B, and S5C). iScience Article Survivin has a specific pattern of transcriptional regulation To explore the mode of survivin-specific transcriptional regulation, we analyzed chromatin regions containing genes highly sensitive to survivin inhibition. Several common features emerged, including (1) long-range interactions between survivin-containing REs and the promoters of target genes, (2) the location of survivin-containing REs among REs clustered into regulatory modules, and (3) the location of survivin-containing REs on repressed/poised chromatin. These features are clearly seen in three genes critical for survivin-dependent metabolism in CD4 + cells: PFKFB3 ( Figure 6A), BIRC2, and SMURF2 ( Figures S6A and S6B), all of which were transcriptionally activated by survivin inhibition.
Four survivin-ChIP peaks were associated with 5 high-scored REs paired to PFKFB3 ( Figure 6A). These REs covered a region extending from $20 kb upstream to 250 kb downstream of PFKFB3. According to the ReMap database, both the upstream and the downstream REs contained ChIP peaks for IRF1 and SMAD3 grouped together with the survivin peaks ( Figure 5E). Three survivin-ChIP peaks were annotated to the REs connected to BIRC2 and located $100 and $400 kb downstream of the TSS according to the GeneHancer database ( Figure S6A). Despite their distant location, both REs were strongly linked to BIRC2 (GH scores of 1.56 and 10.95, respectively) and according to the ReMap, contained multiple IRF1 and SMAD3 ChIP-seq peaks. Additionally, both RE were located within the repressed/poised chromatin according to the functional chromatin segmentation in CD4 + cells. Five survivin-ChIP peaks were found in the genomic region adjustent to the SMURF2 gene ( Figure S6B). Four of those peaks were annotated to the REs that formed a dense cluster spanning the region of $100 kb upstream of the TSS and built a higher-order regulatory unit at that site. Thus, the inhibition of survivin could trigger simultaneous activation of the clustered REs as predicted by the RoadMap data, which could explain the pronounced upregulation of SMURF2 expression observed in the functional experiment ( Figure 4B).  To further investigate survivin binding to the chromatin regions containing RE connected to PFKFB3, we performed a ChIP experiment targeting survivin in THP1 cells. To amplify the survivin-bound chromatin fraction in the genomic loci that overlap REs and survivin peaks, we designed a set of specific primers within the enhancers connected to the PFKFB3 gene ( Figure 6A. Table S3A). Using conventional qPCR, survivin ChIP of THP1 cells was amplified in four independent regions within REs GH10J006129, GH10J006398, and GH10J006199. After adjustment to the PCR results in ChIP with total rabbit IgG, we found that survivin IP was significantly enriched within the central part of RE GH10J006398 (site 2), while the regions in REs GH10J006129 and GH10J006199 had no such enrichment ( Figure 6B). The TF motif analysis of the amplified regions in RE GH10J006398 identified multiple binding sites of IRF1, and accumulation of SMAD3 binding motifs in RE GH10J006129 ( Figure 6C). In contrast, the amplified region in RE GH10J006199 contained no binding motifs of these TFs, supporting the observations from bioinformatic and physical interaction experiments. These findings confirm the specificity of survivin binding to the chromatin of THP1 cells and reproduced the results obtained in the survivin ChIP-Seq experiments on CD4 + T cells. Thus, survivin binding to these REs may control glucose utilization in T cells through the regulation of PFKFB3 expression.

DISCUSSION
This study demonstrates a survivin-dependent mechanism of metabolic adaptation existing in the IFNgproducing CD4 + cells. We show that nuclear survivin exhibits genome-wide and motif-specific binding to chromatin with unappreciated function in gene transcription control. The exact position of survivin binding is defined here by its physical interaction with the sequences of cis-RE and the TFs IRF1 and SMAD3/ SMAD4. We show that survivin accompanied by IRF1 and SMAD3 keeps control of PFKFB3, the major point of metabolic adaptation for autoreactive T cells, and other genes responsible for glycolysis and sugar transport. Thus, survivin binding to chromatin acts as an epigenetic check-point coordinating a metabolic switch required for the effector function of the IFNg-producing CD4 + cells.
This study demonstrated a solid reciprocal connection between survivin and IFNg expression in the clinical material of patients with RA and in healthy CD4 + T cell cultures. Previous studies reported that the activation of IFNg signaling induced survivin transcription through STAT1 binding to the survivin/BIRC5 gene promoter. 55 Stimulation of human T cells with survivin peptides induced IFNg production. 56 Concordant with our results, inhibition of survivin with YM155 in human T cells led to a significant reduction in IFNg production. 57 We did not find any evidence for transcriptional control of the IFNG gene by survivin. Instead, survivin mediates IFNg effects and regulates metabolic genes acting as an IRF1 partner.
We show that the repression of the key glycolytic enzyme PFKFB3 is central to the survivin-dependent metabolic effects in CD4 + cells. It leads to the activation of LDHA and aerobic glycolysis and a cessation of the pentose phosphate pathway. Expression of PFKFB3 is altered in response to growth factors, inflammation, and ischemia, all of which activate estrogen receptor-, hypoxia-, or progesterone response elements on its promoter. 58 Maintenance of PFKFB3 repression requires energy. Integrative analysis of ChIP-seq and protein binding data identified IRF1/survivin/SMAD3 complex as a potent repressor of the REs connected to PFKFB3. Inhibition of survivin increased PFKFB3 expression and restored conventional aerobic glycolysis through the TCA cycle, which reduced the glucose uptake and IFNg production. This survivin-dependent change in the mode of glucose utilization is consistent with the logical connection between survivin, and IRF1-dependent effector function of CD4 + cells. Survivin is frequently bound to chromatin sequences containing IRF motifs and directly binds IRF1, the lineage-specific TF that mediates IFNg signaling, enabling the transcriptional control of IRF1 target genes. Inhibition of survivin significantly impaired both IFNg production and the sensitivity of CD4 + cells to IFNg stimulation, which is required to maintain their effector phenotype and chronic inflammation.
Our findings showed that survivin represses TGFb/SMAD-dependent processes in CD4 + cells. iScience Article SMAD-dependent increase in chromatin accessibility. 61 Hypothetically, formation of the survivin/SMAD3 complex might anchor SMAD3 to inactive/poised chromatin, creating a predisposition for rapid changes in transcriptional activity, as observed in our study. EP300 and CREB1 were the only common interactors for IRF1 and SMAD3. The binding sites for both TFs were significantly enriched in REs connected to the DE-Gs upregulated after survivin inhibition. In this scenario, survivin acts as a guardian of the functional chromatin state by preventing the EP300/CREB1 complex interaction with SMAD3. Remarkably, the activity of EP300/CREB1 is mediated by glucose 62 and integrates the immune processes initiated by IFNg 63,64 and TGFb-signaling, potentially by patronizing the transcriptional activity of the IRF1/survivin and survivin/ SMAD3 complexes.
In agreement with our findings, repression of PFKFB3, which switched glucose processing to the pentose phosphate pathway, has been suggested as the major point of metabolic adaptation for T cells in RA contributing to autoimmunity 65 and the invasive phenotype of synovial fibroblasts. 66 In contrast to RA, autoimmune conditions such as type 1 diabetes, multiple sclerosis, and systemic lupus erythematosus utilize the pyruvate kinase-dependent hyperproduction of lactate to meet their energy demands and could experimentally be improved with the inhibition of PFKFB3. [67][68][69] Analyses of publicly available datasets (e.g. HiC, eQTL) indicated a strong internal connection between the REs with survivin ChIP peaks and the PFKFB3 promoter region. Multiple critical SNPs associated with autoimmune diabetes, RA, and celiac disease were discovered by GWAS in the PFKFB3 genomic region close to the survivin binding sites. This strongly linked the region to metabolic and autoimmune conditions through variation in T cell transcription [70][71][72] and lends relevance to the transcriptional control in the PFKFB3 gene region for the development of autoimmune conditions.
In summary, our study identifies a previously unknown epigenetic mechanism that connects oncoprotein survivin with the effector phenotype of IFNg-producing CD4 + T cells. This occurs through the regulation of glucose utilization and transcriptional control of the PFKFB3 locus. The tight interaction with the IRF1/survivin or survivin/SMAD3 complexes maintains expression of IFN-sensitive genes that are clinically relevant in several autoimmune diseases, including RA, 48,64 systemic lupus erythematosus, 49 and Sjö gren's syndrome. 50 Our findings provide an insight in the fundamental role of survivin in bridging the transcriptional programs governed by IRF1 and SMAD3 in the regulation of the balance between IFNg-and TGFb-dependent processes. This knowledge could have direct practical application for patients. Mapping of metabolic state in CD4 + T cells could be used to personalize treatment choice and reduce drug resistance. Pharmacological interventions that selectively target the molecular interactions of survivin could be an attractive approach to improve control of IFNg-dependent autoimmunity and treatment of RA.

Limitations of the study
Our study has some limitations. We show that high levels of survivin in CD4 + cells result in low expression levels of PFKFB3 in the IFNg producing cells. The connection between the expression, protein, and functional levels of the phospho-fructokinase 2, coded by PFKFB3 needs more detailed studies. Immunoprecipitation experiments demonstrate the presence of survivin, IRF1, and SMAD3 in the pulled-down material. This finding calls for structural studies to confirm the direct interaction between those proteins in a complex and to deduct the nature of this interaction. Finally, the study does not address the specificity of the YM155 inhibitory effect. Sepantronium bromide (YM155), a small-molecule that specifically suppresses survivin expression but not the expression of cIAP2, XIAP, Bcl-2, Bcl-XL, Bad, 47 cIAP1, p53 or Stat3. 73 Later studies indicated side effects and secondary targets of YM155 [74][75][76] . In the light of our findings, unexpected translocation of TF SP1, ILF/NF110, and NF-kB heterodimers during YM155 treatment 74,75 could be explained by previously unappreciated survivin binding to chromatin as a part of large TF complexes described by our studies. Concordant with Sim et al., 76 we observed a significant upregulation of FOXO1 and CYLD in YM155 treated CD4 + cells, which could be explained by survivin binding to RE connected to these genes. Thus, the results of this study support survivin-targeting specificity of YM155, but this assumption needs further investigations.

STAR+METHODS
Detailed methods are provided in the online version of this paper and include the following:

ACKNOWLEDGMENTS
We would like to thank the research nurses Anneli Lund and Marie-Louise Andersson at the Rheumatology Clinic, Sahlgrenska University Hospital, Gothenburg, for their help with blood sampling. We also thank all patients with RA, who participated in this study.

DECLARATION OF INTERESTS
The authors declare no competing interests.

Conventional qPCR
To validate results of survivin ChIP-seq, primers were designed to cover the peak region in the REs connected to the PFKFB3 gene ( Figure 6A). The RE with no survivin peak was used as a negative control.

ChIP-seq analysis
The fastq sequencing files were mapped to the human reference genome (hg38) with the STAR aligner 79 ; the alignIntronMax flag was set to 1 for end-to-end mapping. The quality of sequenced material was assessed with the FastQC tool and MultiQC (v.0.9dev0) (Babraham Institute, Cambridge, UK). Peaks were called with MACS2 algorythm for narrow peaks and default parameters. Peaks were filtered for the survivin antibody IP fraction (IP) and unprocessed DNA (Input), which is a generally accepted normalization approach to identify protein-specific enrichment of DNA interaction areas. 80 A set of peaks with enrichment versus surrounding region and Input (adjusted p < 10 À5 ) was identified and quantified separately for each sample. Peaks that overlapped by at least 1 nucleotide in several samples were merged as survivin-ChIP peaks. Peaks in all samples were scored by the number of tags of difference between IP and Input (average of these differences between samples). Peaks were annotated with HOMER software 81 in standard mode to the closest TSS with no distance restriction. HOMER (findMotifsGenome.pl) and the homer2 engine were used for de novo motif discovery and motif scanning. The most common de novo motifs were identified separately for each IP sample and examined for detected motifs in the JASPAR database of human TF binding sites. 82 The Input bed regions were compared with random global controls generated by the service to match the input dataset. For genomic interval datasets, including survivin-ChIP peaks and REs, the Table Browser for the hg38 human genome assembly (http://genome.ucsc.edu/cgi-bin/hgTables) and Galaxy suite tools (https:// usegalaxy.org/) were used for estimating distances between nearest intervals, merging, overlapping, calculating genomic coverage, and other standard procedures. The genome-wide distribution of survivin-ChIP peaks was initially screened with the cis-regulatory annotation system (CEAS v0.9.8; accessed November 1, 2020 with Cistrome Galaxy, http://cistrome.org/ap/root). For enrichment analysis, we used the list of all survivin-ChIP peaks and the fraction of them located within 100 kb of the known genes. To estimate pairwise distances and statistical significance of pairwise interval overlaps for survivin-ChIP peaks with genome elements defined above, we used Bedtools suite (https://github.com/arq5x/bedtools2; accessed 01feb2021-15apr 2021). For each comparison, a pairwise, two-tailed Fisher's exact test was used.
Comparison was based on initial survivin-ChIP peak positions as intervals and extended regions with 1-kb, 10-kb and 50-kb flanks.

Computational analysis
To identify transcription regulators near survivin-ChIP peaks, we used the ReMap database (http://remap. univ-amu.fr/; accessed November 15, 2020) for colocalization analysis of aggregated cell-and tissueagnostic human ChIP-seq datasets of 1034 transcriptional regulator. ReMapEnrich R-script (https:// github.com/remap-cisreg/ReMapEnrich; accessed November 15, 2020) was used for colocalization enrichment analysis. The hg38 human genome assembly was used for all comparisons. Two-tailed p values were estimated and normalized with the Benjamini-Yekutielli test, using the maximal allowed value of shuffled genomic regions for each dataset (n = 15), kept on the same chromosome (shuffling genomic regions iScience Article parameter byChrom = TRUE). The default fraction of minimal overlap for input and catalogue intervals was set to 10%. Bed interval files of survivin-ChIP peaks with 0-and 100-kb flanks were prepared. The dataset with 0-kb flanks was compared with the Universe sets of genomic regions, defined as within 1 Mb of the same ChIP-seq peaks. For analysis of the regulatory chromatin paired with DEGs, input bedfiles were selected according to their distance from the genome region containing REs paired to DEGs; bedfiles for individual TFs were downloaded from ReMap2020.
TFs with statistically significant enrichment of overlaps (q value < 0.05, n > 100) were selected. TFs that were enriched with respect to the genomic background were identified within each RE by using ReMap database, as described above. A subset of TFs enriched within the survivin-associated REs was identified by chi-square test (chisq.test, R-studio) and false-discovery rate correction (R-studio). To explore the involvement of these TFs in regulating DEGs, we prepared the presence matrix (1/0 type), excluded regions with 0 overlaps with top TFs, and did a principal component analysis with singular value decomposition imputation. Hierarchical clustering of TFs was done with Canberra or Euclidean distances (prcomp, hclust, R-studio). False discovery rate-adjusted p values and the ratio between the survivin-associated and survivin-independent REs per 1 Mb was estimated.
DEGs were compared to all protein-coding human genes (by default) by gene set enrichment analysis (https://www.gsea-msigdb.org/gsea/index.jsp; accessed November 15, 2020). Transcriptional regulators with significant overlap between ChIP-seq and survivin-ChIP peaks were analyzed in comparison to all 1034 transcriptional regulators in ReMap2020. The list of genes corresponding to these regulators was used for functional enrichment analysis in Gene Ontology Biological Processes (GO:BP) using the total list of ReMap transcription regulators as a custom background. Enriched GO:BP categories were grouped together and visualized on the 2D map based on their semantic similarity to show the functional preferences of potential survivin-associated transcriptional regulators. Grouping was carried out by medium term similarity of 0.7 and using ReViGo service (http://revigo.irb.hr/, accessed 01dec2020).
Known genetic associations of the analyzed regulatory regions of DEGs were examined with NHGRI's collection of GWAS (http://genome.ucsc.edu/). All published GWAS SNPs were included without p value or ancestry filtering. A subset of relevant SNPs was selected by keyword searches for traits of individual autoimmune disorders in Table Browser. 83 Primary functional chromatin segmentation was accessed by using NIH Roadmap Epigenomic Project data for intact CD4 + CD25 À Th Primary cells (E043 PrimaryHMM; accessed with the Washington University Epigenomic Browser http://epigenomegateway.wustl.edu/browser/roadmap/on March 22, 2022). The default color scheme was applied to chromatin segments of active enhancers, transcribed regions, and repressed and poised loci.

ll
OPEN ACCESS