Selective modulation of gene expression in activated normal human peripheral blood mononuclear cells by store-operated calcium entry blocker BTP2

Calcium is a critical signaling molecule in many cell types including immune cells. The calcium-release activated calcium channels (CRAC) responsible for store-operated calcium entry (SOCE) in immune cells are gated by STIM family members functioning as sensors of Ca2+ store content in the endoplasmic reticulum. We investigated the effect of SOCE blocker BTP2 on human peripheral blood mononuclear cells (PBMC) stimulated with the mitogen phytohemagglutinin (PHA). We performed RNA sequencing (RNA-seq) to query gene expression at the whole transcriptome level and identified genes differentially expressed between PBMC activated with PHA and PBMC activated with PHA in the presence of BTP2. Among the differentially expressed genes, we prioritized genes encoding immunoregulatory proteins for validation using preamplification enhanced real time quantitative PCR assays. We performed multiparameter flow cytometry and validated by single cell analysis that BTP2 inhibits cell surface expression CD25 at the protein level. BTP2 reduced significantly PHA-induced increase in the abundance of mRNAs encoding proinflammatory proteins. Surprisingly, BTP2 did not reduce significantly PHA-induced increase in the abundance of mRNAs encoding anti-inflammatory proteins. Collectively, the molecular signature elicited by BTP2 in activated normal human PBMC appears to be tipped towards tolerance and away from inflammation.


Introduction
Calcium (Ca 2+ ) is a critical signaling molecule in many cell types including T cells and cytosolic Ca 2+ oscillations impact gene expression [1,2]. We previously found that the increase in intracellular Ca 2+ in activated normal human T cells is biphasic; an immediate increase when T cells are signaled via the CD3/T cell receptor complex, and a slower and sustained increase when the T cells are signaled via the costimulatory CD2 protein [3,4]. Importantly, the biphasic Ca 2+ mobilization was essential for the induction of transcription factors, gene expression, and proliferation of normal human T cells signaled with a synergistic combination of Anti-CD3E monoclonal antibody (mAb) and anti-CD2 mAb [3,4]. It is established that the immediate increase in cytosolic Ca 2+ concentration is mediated by the activation of phospholipase C, phosphatidylinositol hydrolysis, generation of inositol triphosphate 3 (IP 3 ), and IP 3 binding to its receptor on the endoplasmic reticulum (ER) and triggering the release of Ca 2+ from the ER into the cytosol [5][6][7][8]. Depletion of Ca 2+ from the ER stores induces the stromal interaction family members (STIM) to undergo conformational changes and cluster at the ER and plasma membrane (PM) junctions where they recruit members of the Orai1 channel family by diffusional trapping and gating them open to mediate Store Operated Calcium Entry (SOCE) [8][9][10][11][12]. Ca 2+ in ux through SOCE activates calcineurin, a serine-threonine phosphatase that dephosphorylates the transcription factor NFAT-1 leading to its nuclear localization and transcription of genes for T cell activation and function [13].
Emerging data also suggests that SOCE plays a signaling role in NF-kB activation [14].
A nonredundant role for the CRAC channels in-vivo is suggested by the observations that mutations in the CRAC channels are associated with both immunode ciency and autoimmunity [15][16][17][18]. Using a oxed STIM1 hypomorph mouse model, we identi ed that a functional de ciency of STIM1 is associated with reduced SOCE, impaired NFAT-1 activation, and reduced in ammation [19]. In our study, a cell permeable pyrazole derivative BTP2, a potent inhibitor of CRAC channels and SOCE [20,21], mimicked several of the effects of STIM1 de ciency in the genetic mouse model.
Existing data and emerging knowledge inspired the current study investigating the effect of BTP2 on the activation of normal human peripheral blood mononuclear cells (PBMC). We performed RNA sequencing (RNA-seq) to examine the effect of BTP2 on genome-wide gene expression in PBMC signaled with PHA. We prioritized and validated differential gene expression of a select panel of mRNAs using a preampli cation enhanced real-time quantitative PCR (RT-qPCR) assay developed in our laboratory [22].
Considering the positive signaling role of cytosolic Ca 2+ , we anticipated that the blockade of SOCE with BTP2 would result in the inhibition of PHA-induced gene expression. Indeed, we found that BTP2 blocked activation induced higher expression of multiple genes including genes encoding proin ammatory proteins. Very surprisingly, we found that BTP2 did not inhibit the activation-dependent increase in the expression of mRNAs encoding anti-in ammatory proteins such as immunosuppressive cytokines TGFB1 and IL10, and negative regulators of immunity such as FOXP3 and CTLA4. The newly identi ed differential effects of BTP2 on activation induced gene expression patterns in normal human PBMC form the basis for this report.

Results
RNA-seq of human PBMC. We isolated PBMC from peripheral venous blood collected from healthy volunteers and incubated PBMC without PHA or BTP2 (control PBMC), with BTP2, with PHA, or with PHA + BTP2 for 16 hours at 37 0 C in 95% air and 5% CO 2 atmosphere. Because DMSO was used as the solvent for BTP2, DMSO was also added to control PBMC and PBMC stimulated with PHA.
We isolated total RNA from the PBMC and performed RNA-seq to identify gene expression patterns at the whole transcriptome level (Fig. 1a). Prior to differential gene expression analysis, RNA-seq data was processed to obtain the gene expression count data for individual samples (Supplementary Fig. 1a and 1b). These counts were then normalized to total counts, corrected for batch effects, and used for hierarchical clustering based on the Pearson correlation between samples ( Supplementary Fig. 1c).
Cluster analysis of RNA sequencing data revealed three distinct groups of samples. PBMC incubated alone and PBMC incubated with BTP2 formed one cluster, PBMC incubated with PHA alone formed another cluster, and the third cluster was formed by PBMC incubated with both PHA and BTP2 (Fig. 1b). Notably, samples obtained from the four different healthy volunteers formed clusters based on experimental conditions and there were no clusters based on individual volunteers indicating that the source of PBMC has a minimal, if any, impact. We employed Principal Component Analysis (PCA) as a means of dimensionality reduction and visualized the data on the rst two principal components. PC1 explained 89% of the variance and PC2 5% of the variance, and we observed three groups consistent with the hierarchical clustering results (Fig. 1c).
It is evident that PCA analysis shows clear separation based on experimental conditions and not based on individual healthy volunteers. PBMC incubated alone (control) and PBMC incubated with BTP2 showed minimal variation and clustered together suggesting that BTP2 alone has minimal impact on unstimulated PBMC. On the other hand, PBMC stimulated with PHA clustered separately from resting PBMC and from PBMC activated with PHA in the presence of BTP2.
Differential gene expression analysis. To comprehensively identify gene expression changes induced by PHA, we compared PHA-treated PBMC to control PBMC using differential gene expression analysis (DGEA). DGEA revealed a signi cant increase in the expression of 2883 genes (two-tailed Wald test, log fold-change > 1.0 and adjusted p-value < 0.01) and a signi cant decrease in the expression of 3172 genes following PHA treatment (two-sided Wald test, log fold-change < -1.0 and adjusted pvalue < 0.01) (Fig. 1d). Analysis of the DGEA results revealed that the largest increase in expression was observed for genes involved in T cell proliferation, T and B cell differentiation, and genes encoding cytokines and chemokines involved in chemoattraction of T and B cells. A gene ontology enrichment analysis on the panel of 81 genes most upregulated (two-sided Wald test, log2 fold-change > 5.0 and adjusted p-value < 0.01) in PHA-activated PBMC revealed an enrichment of ontology terms related to the positive regulation of receptor signaling pathway via the JAK-STAT pathway, the STAT pathway including positive regulation of tyrosine phosphorylation of STAT protein, positive regulation of cytokine expression, myeloid leukocyte differentiation, and alpha-beta T cell activation and differentiation in immune responses (Fig. 1e).
Supplementary Table S1 lists base mean counts of all genes from RNA-seq data and log2 fold changes based on comparisons of gene expression in PBMC stimulated with PHA vs. gene expression in control PBMC, and adjusted P values.
All mRNAs speci ed in the Results section are highlighted yellow in Supplementary Table S1.
To identify the impact of BTP2 on PBMC activation with PHA, we performed DGEA of PBMC incubated with PHA + BTP2 and PBMC incubated with PHA. DGEA revealed that the impact of BTP2 on PHA-induced gene expression was not monotonic and that both a signi cant increase in the expression of 745 genes (two-tailed Wald test, log2 fold-change > 1.0 and adjusted pvalue < 0.01) and a signi cant decrease in the expression of 723 genes in the PBMC treated with both PHA and BTP2 (twosided Wald test, log2 fold-change < -1.0 and adjusted p-value < 0.01) were identi ed (Fig. 1f). Analysis of the DGEA results revealed that out of 2883 genes which were signi cantly upregulated in PHA activated cells and 623 genes were signi cantly downregulated in cells treated with both PHA and BTP2 (two-sided Wald test, log fold-change < -1.0 and adjusted p-value < 0.01). The downregulated genes include the genes involved in T cell proliferation and differentiation such as IL2 and IL2RA (CD25). A gene ontology enrichment analysis on the genes most upregulated (log fold-change > 5.0) in PHA-activated cells but downregulated (log fold-change < -1.0) on BTP2 treatment of PHA-activated PBMC revealed an enrichment of ontology terms related to the regulation of regulatory T cell differentiation and down regulation of interferon gamma and JAK-STAT signaling pathways upregulated in PBMC stimulated with PHA (Fig. 1g). Collectively, these results suggest that BTP2 can positively impact regulatory T cell differentiation and suppress T cell proliferation. Importantly, our analysis identi ed that the impact of BTP2 on gene expression in stimulated PBMC is not unidirectional, and an almost equal number of genes were upregulated and downregulated by BTP2 in PBMC signaled with PHA.
Supplementary Table S2 lists base mean counts of all genes from RNA-seq data and log2 fold changes based on comparisons of gene expression in PBMC stimulated with PHA + BTP2 vs. gene expression in PBMC + PHA, and adjusted P values. All mRNAs speci ed in the Results section are highlighted yellow in Supplementary Table S2.
Impact of BTP2 on mRNAs encoding CRAC channel proteins. We investigated the effect of PHA on the expression of mRNAs encoding CRAC proteins in PBMC and the impact of BTP2 on PHA-induced alterations in mRNA counts (Fig. 1h).
Altogether, RNA-seq of PBMC identi ed that BTP2 treatment signi cantly suppresses mRNA encoding proteins responsible for T cell expansion, activation, and cytotoxicity without signi cantly reducing the expression of mRNA encoding negative regulators of immunity.
Validation of differential gene expression using preampli cation enhanced RT-qPCR assays. We used an orthogonal platform, RT-qPCR assay developed in our laboratory [22] to measure absolute copy numbers of transcripts; we prioritized for measurement mRNAs encoding immunoregulatory proteins implicated in autoimmunity and organ transplantation [23,24]. Table 1 shows absolute copy numbers of mRNAs measured in four consecutive experiments, the median and 25th and 75th percentile values, multiple group comparison P values calculated using Kruskal Wallis test, and pairwise comparisons P values calculated using Mann Whitney U tests. The abundance of 18S rRNA is relatively stable across cell types and experimental conditions and therefore was used as a housekeeping (reference) gene in our study. As shown in Table 1, 18S rRNA copy numbers were not signi cantly different across the four experimental conditions (P = 0.82, Kruskal Wallis multiple comparison test). Pairwise comparisons of PBMC vs. PBMC + BTP2 (P = 0.89, Mann Whitney test), PBMC vs. PBMC + PHA (P = 0.89), and PBMC + PHA vs. PBMC + PHA + BTP2 (P = 0.51) showed no signi cant differences in the abundance of 18S rRNA across the 4 experimental conditions. The lack of a signi cant difference in the abundance of 18S rRNA copy number suggests that signi cant differences observed with a mRNA of interest are due to differences in expression due to the experimental condition rather than due to technical artifacts.
mRNAs for T cell surface proteins. Measurement of absolute copy number of mRNAs using the RT-qPCR assays showed excellent agreement with the alterations in mRNA counts identi ed by RNA-seq. As found by RNA-seq, expression levels of CD3E mRNA or CD8A mRNA did not vary across all four experimental conditions (Table 1). CD4 mRNA copy number was numerically lower following PHA activation and further reduced by BTP2 ( Table 1).
As found by RNA-seq, BTP2 reduced the PHA-induced increase in CD27 mRNA copy number; BTP2 also reduced the PHAinduced increase in CD28 mRNA copy number (Table 1). RT-qPCR assays results were also concordant with RNA-seq data regarding CD103 mRNA -a non-signi cant decrease following activation with PHA and partial reversal by BTP2 (Table 1).
mRNAs for IL2 and IL2RA (CD25). As observed with RNA-seq, PHA-activation of PBMC caused a signi cant increase in IL2 mRNA copy number and BTP2 signi cantly reduced PHA-induced increase ( Table 1). PHA-activation of PBMC mediated a signi cant increase in IL2RA copy number and BTP2 reduced signi cantly PHA-induced increase (Table 1).
mRNAs for IFNG, CXCL9 and CXCL10 and CXCR3. As found by RNA-seq, PHA-induced a signi cant increase in IFNG mRNA copy number and BTP2 partially reduced IFNG mRNA copy number ( Table 1). Measurement of absolute copy numbers of CXCL9 mRNA and CXCL10 mRNA using RT-qPCR assays validated both the increase in the abundance of CXCL9 mRNA and CXCL10 mRNA following PHA activation, and the lack of inhibition by BTP2 of the PHA induced increase in the abundance of CXCL9 mRNA or CXCL10 mRNA (Table 1).
CXCR3 is a receptor for both CXCL9 and CXCL10 and its abundance was not altered in PBMC by PHA stimulation or by BTP2 (Table 1).
mRNAs for PRF1, GZMB and CD96. RT-qPCR assay measurements of PRF1 and GZMB copy numbers validated RNA-seq ndings that PHA stimulation increases PRF1 mRNA copy number and GZMB mRNA copy number in PBMC and BTP2 mediates a non-signi cant decrease in PRF1 mRNA copy number and a signi cant decrease in GZMB mRNA copy number (Table 1). mRNAs for TGFB1, IL10, FOXP3 and CTLA4. Measurement of mRNAs encoding immunosuppressive cytokines TGFB1 and IL10 and the negative immune regulators FOXP3 and CTLA4 con rmed the RNA-seq ndings that BTP2 does not signi cantly reduce the expression of mRNAs encoding these anti-in ammatory mediators. PHA increased the abundance of mRNAs for TGFB1(P = 0.06) and IL10 (P = 0.03) in PBMC and BTP2 did not reduce the PHA-induced increase in TGFB1 mRNA (P > 0.99) or the increase in IL10 mRNA (P = 0.69) ( Table 1). PHA also increased the abundance of mRNAs for FOXP3 (P = 0.03) and CTLA4 (P = 0.03) and BTP2 did not reduce the PHA-induced increase in FOXP3 mRNA (P = 0.11) or the increase in CTLA4 mRNA(P = 0.20) (Table 1).
Single cell analysis for cell surface expression of IL2 receptor alpha (CD25). We performed multiparameter ow cytometry analysis to determine the effect of BTP2 on PHA-induced cell surface expression of IL2RA (CD25). Figure 3 is representative ow cytometry data from all four experimental conditions and shows that PHA increases the percentage of CD25 + cells and BTP2 decreases cell surface expression of CD25.
Results from four consecutive experiments are shown in Table 2. Table 2a shows data from each of the four experiments and Table 2b is a summary of statistics from the four consecutive experiments. Pairwise comparisons showed that PHA stimulation signi cantly increases CD25 display on the cell surface and BTP2 signi cantly decreases the cell surface expression of CD25 (Table 2b). An analogous pattern of an increase with PHA and a decrease with BTP2 was also observed with the MCF of CD25 positive cells (Table 2b). Altogether, multiparameter ow cytometry analysis of CD25 expression at the protein level con rmed and extended our ndings that BTP2 inhibits PHA-induced expression of CD25 at the pre-translational level.

Discussion
We identi ed that the SOCE blocker BTP2 has a gene speci c effect on PHA-induced alterations in the abundance of mRNAs in normal human PBMC. Intriguingly, the activation induced increases in the abundance of mRNAs encoding anti-in ammatory proteins were not reduced signi cantly by BTP2 whereas activation induced increases in the abundance of mRNAs encoding proin ammatory proteins were signi cantly reduced.
The immune pro le elicited by BTP2 appears conducive to antigen speci c tolerance (Fig. 4). The nding that BTP2 does not alter the expression of mRNAs for T cell surface molecules CD3E, CD8A, and CD103 suggests that T cell antigen education via physical contact with the cognate antigen and transmembrane signaling should proceed uninterrupted. On the other hand, the expected outcomes of BTP-associated reduction in the mRNAs for CD4, CD27, and CD28 include impaired T cell helper function, sub-optimum T cell co-stimulation, and T cell anergy. CXCR3 mRNA and mRNAs for the chemokines CXCL9 and CXCL10 were not reduced by BTP2, and a predicted outcome is unhindered tra cking of antigen reactive cells to their target cells. Such homing is unlikely to have an adverse consequence since neither clonal expansion nor cytolytic activity would be realized since BTP2 reduced mRNAs for T cell growth factor IL2 and its receptor IL2RA and mRNA for the cytolytic effector granzyme B. In striking contrast, mRNA for TGFB1, a potent immunosuppressive cytokine, and mRNA for the cytokine synthesis inhibitory factor IL10 were not reduced by BTP2. Moreover, mRNA for FOXP3, a speci cation factor T regulatory cells, and mRNA for the negative regulator CTLA4 were not reduced by BTP2. Collectively, the molecular signature elicited by BTP2 in activated normal human PBMC appears to be tipped towards tolerance and away from in ammation.
Several features of our research design enabled the identi cation of selective modulation of gene expression by BTP2. First, we used normal human PBMC as indicator cells rather than the malignant Jurkat CD4 + T cells or T cell clones to assess the effects of BTP2. The PBMC activation model is also physiological with respect to T cell co-stimulation. PHA was used to signal PBMC in our study and PHA is considered a superior stimulator of lymphocytes compared to several other mitogens including the widely used lectin concanavalin A [25,26]. Importantly, we performed unbiassed RNA-seq to query gene expression at the whole transcriptome level and identi ed differential gene expression in an unbiased fashion. We prioritized mRNAs encoding proteins implicated in autoimmunity, allograft rejection, and allograft tolerance for validation using an orthogonal platform and con rmed differential expression by absolute quanti cation of mRNA copy number. The measurement of absolute copy number of mRNAs in the RT-qPCR assays [22] obviated the inherent limitations of relative quanti cation of mRNAs in RT-qPCR assays.

The CD3 proteins are an integral component of the T cell receptor (TCR) antigen recognition complex and the CD4 and CD8
proteins are expressed by helper T cells and cytotoxic T cells, respectively [27][28][29]. BTP2 had a minimal impact on mRNAs encoding CD3E or CD8A. On the other hand, BTP2 reduced mRNAs for CD4, CD27, and CD28 in PBMC. Full T cell activation is dependent upon co-stimulatory signals and the T cell CD27 and CD28 are prominent T cell co-stimulatory proteins [30,31]. An expected consequence of BTP2 reducing their expression is impaired T cell co-stimulation and T cell anergy [32]. Altogether, the T cell pro le elicited by BTP2 is permissive of T cell antigen recognition and transmembrane signaling, impaired T cell helper function, and defective T cell co-stimulation and anergic T cells (Fig. 4).
IL2 is the primary trophic factor for T cell clonal expansion and the acquisition of IL2RA by activated T cells is necessary for the formation of high-a nity IL2R trimeric complex [33,34]. BTP2 reduced activation induced increased expression of mRNAs for IL2 and IL2RA. A predicted outcome is impaired T cell clonal expansion (Fig. 4).
IFNG was rst identi ed as secretory product of PHA-stimulated leucocytes and is a prototypic proin ammatory cytokine [35].
BTP2 reduced activation induced increase in IFNG mRNA in PBMC. CXCL9 and CXCL10 are IFNG induced chemokines that signal T cells via CXCR3 displayed their cell surface [36,37]. PHA mediated a signi cant increase in the abundance of mRNA for both chemokines and intriguingly, BTP2 did not reduce signi cantly the induced higher abundance. These surprising ndings and the additional observation that BTP2 does not affect CXCR3 mRNA expression indicate that cellular tra c to target sites such as an allograft will not be jeopardized by BTP2 and antigen-experienced cells, a prerequisite for tolerance induction, could emerge (Fig. 4).
Perforin is a pore forming protein and granzymes enter cells via the pores in the plasma membrane and mediate cell death by apoptosis. PHA-activation signi cantly increased the expression levels of both perforin and granzyme B. Heightened intragraft expression of perforin and granzymes have been repeatedly identi ed during an episode of acute allograft rejection [23,34].
Thus, BTP2 mediated decrease in GZMB mRNA is likely to bene t the transplanted organ. We found that the abundance of mRNA for CD96, expressed on the surface of T cells and NK cells and a negative regulator NK/T cell cytolytic activity [38,39] is increased by BTP2. This data suggests that BTP2 would help constrain cytolytic effector mechanisms implicated in autoimmunity and allograft rejection (Fig. 4).
Genome wide pro ling using RNA-seq identifying that inhibition of SOCE results in diminished expression of genes in activated PBMC is predictable considering the positive signaling role of SOCE in the activation of immune cells. The unanticipated nding from our investigation is that genes encoding anti-in ammatory proteins were largely unaffected by BTP2. In this study, the abundance of mRNAs encoding potent immunosuppressive cytokines TGFB1 and IL10 were not reduced signi cantly by BTP2. Furthermore, BTP2 did not reduce signi cantly the activation induced increase in mRNA for FOXP3 protein, a speci cation factor for regulatory T cells, and activation induced increase in mRNA for CTLA4, a master negative regulator of immunity [40,41]. Intragraft over-expression of FOXP3 + regulatory cells is a feature of tolerant allografts [42,43] and CTLA4 mRNA is over expressed in tolerant kidney allograft recipients [44] (Fig. 4).
Our investigation has limitations. BTP2 has effects on TRPM4, TRPC3, and TRPC5 channels [21] in addition to CRAC channels, and some of our results might be related to the effect of BTP2 on these other channels. We have not deciphered whether the differential effect of BTP2 on gene expression is cell autonomous [45] or due to differential sensitivity of cell subtypes. Real time simultaneous imaging of calcium in the ER lumen, cytosol, and mitochondria [45] and single cell RNA-seq may help resolve this issue. It is important to investigate whether our ndings are translatable to in-vivo settings. BTP2 (also known as YM58483) has been reported to inhibit T cell responses in-vivo [46] and is e cacious as an anti-in ammatory agent [47][48][49][50]. The BTP2 associated tolerogenic signature identi ed in this study suggest a mechanistic basis for the e cacy observed in these preclinical models. Additional studies including transplantation across strong histocompatibility barriers are needed to evaluate the utility of BTP2 in the transplantation setting.
In conclusion, BTP2 has differential effects on activation induced gene expression in normal human PBMC signaled with PHA.
Intriguingly, BTP2 did not signi cantly reduce the abundance of mRNAs encoding anti-in ammatory proteins whereas the mRNAs encoding proin ammatory proteins were signi cantly reduced by this pyrazole derivative. The tipping of the immune repertoire towards tolerance and away from in ammation pro le by BTP2 may be of value for the management of autoimmune disease states and for protecting lifesaving organ transplants.

Methods
Reagents. Ficoll-Paque™ Premium solution was purchased from GE Healthcare Bio-Science AB, Pittsburgh, PA. for 16 hours at 37 0 C in 95% air and 5% CO 2 atmosphere. Because DMSO was used to dissolve BTP2, DMSO was added to control PBMC and to PBMC stimulated with PHA. After overnight incubation, PBMC were retrieved and separate aliquots of PBMC suspended in complete medium aliquots were used for total RNA isolation and single cell analysis by ow cytometry.
Isolation of total RNA from PBMC. PBMCs were washed with PBS (1000 µl) and centrifuged at 13,400 g for 2 minutes in room temperature. The supernatant was discarded, and RNAlater (50 µl) and Buffer RLT with β-mercaptoethanol (350 µl) were added to the cell pellet. The pellet was mechanically lysed with a 1ml 25G 5/8 syringe. The lysate was transferred to a QIAshredder spin column and centrifuged at maximum speed for 2 minutes at room temperature. Total RNA was isolated from the lysate using RNeasy mini kit, according to the manufacturer's protocol (Qiagen). The quantity (absorbance at 260nm) and purity (ratio of the absorbance at 260 and 280 nm) of the total RNA isolated from the lysate were measured using the NanoDrop® ND-1000 UV-Vis spectrophotometer (Thermo Scienti c).
RNA library preparation for RNA-seq. The total RNA samples were treated with DNase using RNase-Free DNase kit (Qiagen) according to the manufacturer's protocol to remove DNA prior to RNA sample library preparation and sequencing. The RNA library preparation and RNA sequencing were performed by the Genomics Core Laboratory at Weill Cornell Medicine. The RNA libraries were prepared using the TruSeq RNA Library Prep Kit (Illumina) and sequenced using NovaSeq 6000 System (Illumina) as 100 x 2 paired-end reads.
Preprocessing of RNA-seq data. Raw data from RNA sequencing experiments was aligned to the human genome (Assembly: GRCh37/hg19) using STAR short read aligner software (v-2.6) [51] for individual samples. The gene expression counts for individual samples were estimated by setting the --quantMode argument to GeneCounts. The gene expression values for individual samples were compiled into a single table for downstream analysis.
Analysis of RNA-seq data. Count matrix containing gene expression counts and the corresponding metadata and alignment statistics for all samples were loaded into R (v-4.0.3). The matrix was converted into a DESeqDataSet object for downstream analysis using the DESeq2 package (v-1.30.1) [52]. SizeFactors and Dispersion values were calculated for individual genes using the DESeq2 package and genes with less than ten copies across all samples were removed from the analysis. The data was normalized using variance stabilizing and rlog transformations (VST) method implemented in the DESeq2 package. The normalized data was batch-corrected across the samples using linear models implemented in the limma package (v-3.46.0) [53]. The batch corrected data was transferred back into the DESeqDataSet object and the control PBMC sample was set as the reference. Pairwise correlations were calculated for individual samples and visualized on a Heatmap. Dimensional reduction was performed on batch corrected data using Principal Component Analysis, which was visualized using the DESeq2 package. Differential Gene Expression Analysis (DGEA) was performed using the DESeq2 method. Effect size shrinkage was used on DGEA results using apeglm (Approximate posterior estimation for GLM coe cients) method [54] implemented in lfcShrink function in the DESeq2 package. The DGEA results containing log fold-changes and adjusted pvalues were exported in tables and visualized on volcano plots.
We designed gene speci c oligonucleotide primers and TaqMan uorogenic probes and measured a custom panel of mRNAs and 18S rRNA using pre-ampli cation enhanced real-time quantitative PCR (RT-qPCR) assays. Supplementary Table S5 lists the oligonucleotide primers and gene speci c TaqMan probes used to measure absolute copy number of transcripts.
The RT-qPCR assay consisted of two sequential steps [22]. In the rst step, the cDNA was pre-ampli ed with gene-speci c primer pairs in a nal reaction volume of 10.0 µl in a 0.2 ml PCR tube and each sample contained cDNA (3.0 µl), Platinum® Multiplex PCR Master Mix (5.0 µl), 1.68 µl primer mix (50 µM sense, 50 µM antisense primer per gene), and water (0.32 µl) to a nal volume of 10 µl. PCR ampli cation was set up in a Veriti thermal cycler (Applied Biosystems) and the PCR pro le consisted of an initial hold at 95oC for 2 min, 11 cycles of denaturing at 95oC for 30 seconds, primer annealing at 60oC for 90 seconds and extension at 72oC for 1 min, and nal extension at 72oC for 10 min and nal hold at 4oC. Following 11 cycles of ampli cation, we diluted the PCR amplicons by adding 290 µl of TE buffer to the 10 µl PCR reaction. In the second step, absolute copy number of mRNAs were quanti ed using the Bak standard curve in the RT-qPCR assay on the Applied Biosystems QuantStudio™ 6 Real-Time PCR system (ThermoFisher) using diluted PCR amplicons (2.5 µl), TaqMan™ Fast Universal PCR Master Mix (2X), no AmpErase™ UNG (Applied Biosystems), sense and antisense primers (50 µM each), TaqMan™ probes (100 µM), and water (7.1 µl) in a nal reaction volume of 20 µl. We established the Bak standard curve with 6 log 10 concentrations of Bak amplicon (starting copy number 2.5 x 10 6 million copies of Bak mRNA) as the template and PCR ampli cation with Bak-speci c primer pair and detection using Bak-speci c TaqMan probe. The ampli cation e ciencies of Bak and the mRNAs quanti ed in this study were greater than 90% in the RT-qPCR assays. Absolute copy number of mRNA was reported as mRNA copies per microgram of total RNA. Flow cytometry analysis of PBMC. PBMC, 1x10 6 cells/ml of RPMI 1640 + 5% heat inactivated fetal bovine serum were incubated without PHA or BTP2 (control), with BTP2 (1000 nM), with PHA (2 µg/ml), or with PHA (2 µg/ml) + BTP2 (1000 nM) for 16 hours at 37 0 C in 95% air and 5% CO 2 atmosphere. At the end of incubation, the cells were washed and labeled with FITC mouse IgG2a,k isotype control mAb (G155-178, BD Pharmingen) and PE Mouse IgG1, k isotype control mAb (MOPC-21, BD Pharmingen) or FITC mouse IgG2a, k anti-human CD3E mAb (HIT3a, BD Pharmingen) and PE mouse IgG1, k anti-human CD25 mAb (M-A251, BD Pharmingen). Each isotype mAb (10 µl) and each antibody marker (10 µl) for CD3E and CD25 was added to 100 µl of cell suspension in 5ml falcon tube. The sample tubes were vortexed and incubated in dark at 4 C for 30 min in refrigerator. After incubation, cells were washed by adding 3 ml of cold working wash buffer (PBS containing 1% FBS and 0.1% wt/vol sodium azide). Cells were then resuspended in 400 µL of working xative reagent (PBS + 0.1% sodium azide + 1% fetal bovine serum + 37% formalin) and ow cytometry analysis was performed. Flow cytometry data was acquired on a   Inhibition of cell surface expression of IL2 receptor alpha (CD25) by BTP2.
Peripheral blood mononuclear cells were incubated either alone (control), with 1000 nM BTP2, with PHA (2µg/ml), or with PHA (2µg/ml) + BTP2 (1000nM) for 16 hours at 37 0 C in 95% air and 5% CO 2 atmosphere. At the end of incubation, the cells were washed and labeled with FITC mouse IgG2a isotype control mAb and PE IgG1 isotype control mouse mAb or FITC IgG2a mouse anti-human CD3 mAb and PE IgG1 mouse anti-human CD25 mAb. Flow cytometry data was acquired on a FACSCanto II using FACSDiva software (8.0.1) and the FCS les were analyzed using FlowJo 10.8.1 software. Fig. 3a, d, g, and j are contour plots based on cell size (forward scatter) and granularity (side scatter) of control PBMC (Fig. 3a), PBMC treated with BTP2 (Fig. 3d), PBMC+ PHA (Fig. 3G) or PBMC+PHA+BTP2 (Fig. 3j). Each contour plot shows the gates and the percentage of cells analyzed. Fig. 3b, f, h, and k are bivariate plots of gated PBMC from all four experimental and the cells segregated based on FITC CD3E orescence signal and PE CD25 orescence signal. The vertical line in the bivariate plots is based on corresponding isotype mAb labeling discriminating CD3E+ cells from CD3E-cells and the horizontal line is based on corresponding isotype mAb discriminating CD25+ cells from CD25-cells. Fig. 3c, f, i, and l are single color histograms of CD25+ cells from all four experiments conditions. The horizontal line indicates the percentage of CD25+ cells.

Figure 4
Potential consequences of BTP2 mediated selective modulation of gene expression in normal human PBMC. The lack of inhibitory effect of BTP2 on the expression of mRNAs for CD3E, CD8A, and CD103 should be permissive of physical contact, antigen recognition, and T cell signaling. BTP2 does not reverse activation induced decrease in mRNA for CD4 and this could lead to impaired helper cell function including T follicular helper cell function required for antibody production. Down regulation of mRNA for CD27 and CD28 by BTP2 should impair T cell co-stimulation and promote T cell anergy. T cell homing needed for antigen contact and education should proceed unhindered since BTP2 does not reduce activation induced up regulation of CXCL9 mRNA and CXCL10 mRNA. BTP2 mediated inhibition of activation induced mRNA for IL2 and mRNA for IL2RA is predicted to prevent T cell clonal expansion, a prerequisite for destructive immunity. BTP2 mediated reduction in mRNA for GZMB and the increase in CD96 are expected to impair cytolytic activity of T cells and NK cells. The lack of reduction by BTP2 of mRNAs for the immunosuppressive cytokines TGFB1 and IL10, and the lack of reduction by BTP2 of activation induced increase in mRNAs for the negative regulators of immunity, FOXP3 and CTLA4, are postulated to promotes tolerance. Collectively, BTP2 mediated selective modulation of gene expression in normal human PBMC is hypothesized to tip the immune balance towards tolerance and away from in ammation.

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