Seventy-four healthy blood donors were recruited under NIH IRB approved protocols and all donors provided written informed consent regarding their participation in the study. Peripheral Blood Mononuclear Cells (PBMCs) were isolated from blood by Ficoll-Hypaque density gradient centrifugation ( Supplementary Table 1 ). Cells were stained for sorting with anti-CD8, anti-CD4, anti-CD62L, and anti-CD45RA. MoFlo XDP Sorter (Beckman Coulter) was used to sort CD45RA⁺ CD62L⁺ naïve cells (T_N), CD45RA^- CD62L⁺ central memory (T_CM) cells, and CD45RA^- CD62L^- effector memory (T_EM) cells. All cells were maintained at 37 °C and 5% CO₂. The human leukemic T-cell line Jurkat, subclone E6 (ATCC), was maintained in standard growth medium (RPMI 1640 medium supplemented with 10% fetal bovine serum) while 293T cell line (ATCC) was cultured in DMEM/HG media with 10% FBS. Passages 5-15 were used in this study.

The HOPX (a,b,c) sequences were synthesized by Genewiz and then converted into lentiviral transfer vector pLVX-IRES-tdTomato (Clontech) under the CMV promoter using EcoRI (5′) and BamHI (3′). We used an empty transfer vector pLVX-IRES-tdTomato as control. The overexpression efficiency of each construct was tested in vitro with fluorescence and verified by qPCR. The shRNA sequence targeted on HOPX (GACCCAGAAATGGTTTAAGCA) was synthesized by Millipore Sigma and then cloned into the vector pLKO.3G_X7 under U6 promoter using AgeI (5’) and EcoRI (3’). The vector pLKO.3G_X7 (Addgene) was used as shRNA control, and contained an EGFP reporter. The knock-down efficiency of each construct was tested in vitro with EGFP fluorescence and verified by Western blot. To produce lentiviruses, 293T cells were initially plated into a p10 dish and then transfected at around 80% confluence with 1 µg of VSVG (envelope vector), 8 µg of dR8.2 (packaging vector), and 10 µg of the transfer vector in the presence of 50 μL of FuGENE (Promega). The viral supernatant was collected after 48 hours, cleared by centrifugation with 1500 rpm for 5 min at 4°C, and then passed through a 0.45 μm pore PVDF Millex-HV filter (Millipore). Lentivirus was concentrated using PEG solution (5X) and centrifuged at 2800 rpm for 1 hour. The virus pellets were resuspended in 100 μL PBS and stored at -80°C until use. CD4⁺ T cells and CD8⁺ T cells were stimulated with anti-CD3/CD28 conjugated microbubbles one day before transduction. Lentiviral stocks were added to either primary T lymphocytes or Jurkat T cells cultured in RPMI 1640 media supplemented with 10% FBS and 5 μg/mL polybrene. After overnight incubation, lentivirus-containing supernatant was removed by centrifugation and fresh media was added. Transduced T cells were harvested for examination of the expression of tdTomato by flow cytometry (BD FACSCantoII) at the indicated time points. tdTomato⁺CD8⁺ T cells (both HOPX and control virus) were sorted by BD FACSAria Fusion on Day 3 and used for gene expression analysis by gene expression microarrays (Agilent).

The entire V regions of TCR a and b (TRAV8-6-TRAJ42) chains were synthesized (Genewiz) and then cloned into the transfer vector pHAGE (Addgene) containing the mCherry reporter. The lentivirus was produced using the same method described above. Then TCR was transduced into the NJ76 cell line. The stable NJ76-TRAV8-6-TRAJ42 cell line was sorted based on mCherry. After overexpression of HOPX or control, TCR was stimulated with anti-CD3/CD28 for 4 hours and 24 hours. The GFP signal was detected with flow cytometry and analyzed with FlowJo.

Total RNA was isolated with RNeasy Mini Kit (QIAGEN) according to the manufacturer’s instructions. One microgram of total RNA was reverse transcribed to cDNA using the SuperScript™ IV First-Strand Synthesis System kit (Invitrogen). cDNA was subsequently subjected to SYBR Green-based real-time PCR using an ABI 7900 Real-time PCR System (Applied Biosystems). Primers used in real-time PCR are shown in Supplementary Table 2 . All results were obtained from at least three independent experiments. ACOX1 was used as an internal control.

Total RNA was extracted from tdTomato⁺CD8⁺ T cells (n=4 donors) sorted by BD FACSAria Fusion on Day 3 using the RNeasy Mini Kit (QIAGEN) according to the manufacturer’s instructions and stored at -80°C. RNA integrity and concentration were evaluated with the Bioanalyzer 2100 and RNA 6000 Nano chip (Agilent Technologies). Cyanine-3 (Cy3) labeled cRNA was prepared from 200 ng total RNA using the Low-Input Quick Amp Labeling Kit (Agilent, Santa Clara, CA) according to the manufacturer’s instructions, followed by RNeasy column purification (QIAGEN, Valencia, CA). Dye incorporation and cRNA yield were evaluated with the NanoDrop ND-1000 Spectrophotometer (Wilmington, DE). A total of 300 ng Cy3-labeled cRNA and 300 ng of a Cy5-labeled universal standard was fragmented at 60°C for 30 minutes in a reaction volume of 25 μL containing 0.5x Agilent fragmentation buffer and 2x Agilent gene expression blocking agent. After fragmentation, 25 μL of 2× Hi-RPM Hybridization Buffer was added to the fragmentation mixture and hybridized to Agilent SurePrint G3 Human GE v3 8x60K Microarray gene expression microarrays for 17 hours at 65°C in a rotating Agilent hybridization oven. After hybridization, microarrays were rinsed one minute at room temperature in GE Wash Buffer 1 (Agilent) and one minute in 37°C GE Wash buffer 2 (Agilent) and dried by slowly removing from Wash buffer 2. Following post-hybridization rinses, arrays were scanned using a dual-intensity Agilent SureScan microarray Scanner at 5 micron, and hybridization intensity data extracted from the scanned images using Agilent’s Feature Extraction Software 9.1 (Agilent) using default parameters to obtain background subtracted and spatially detrended Processed Signal intensities.

Extraction software with default settings was used to obtain normalized expression values from the raw scans. Normalization modified the raw intensity values to compensate for the different dye efficiency in two-channel microarray experiments using Cy3-labeled samples and universal Cy5-standard. A typical volcano plot shows the Z transformed data (Z ratio) on the x-axis and minus log10 of the p-value on the y-axis. Significant differentially expressed genes were defined as Z ratio > 1.5 or < -1.5 and p < 0.01. Functional profiling was performed as Reactome pathway analysis (17, 18). We applied Normalized Enrichment Score (NES) > 1.5 or < -1.5 and p < 0.01 to identify significantly changed pathways. Raw and normalized data are deposited to the Gene Expression Omnibus (GEO) repository accession number GSE234697.

Approximately 10⁷ of fresh and activated CD8⁺ T cells were pooled. Cells were crosslinked in 1% formaldehyde at room temperature for 10 minutes. After quenching crosslinking for 5 minutes with 125 mM glycine and washing the cells in cold PBS, cells were lysed in 50 mM Tris-Hcl, pH 8.0, 1% SDS, 10mM EDTA. Cells were sheared to 300-500 bp fragments (Covaris) and immunoprecipitated with indicated amount of anti-HOPX (sc-398703, SANTA CRUZ) antibody or IgG1 (401401, Biolegend) (1 ug antibody for ChIP-qPCR) overnight. Protein-G magnetic beads (Thermo Scientific) were added to each reaction for 1 hour at 4°C the following day. The magnetic beads were then washed in a series of buffers including a low salt buffer (20 mM Tris-HCl, pH 8, 150 mM NaCl, 2mM EDTA, 0.1% SDS, 1% Triton X-100), a high salt buffer (20 mM Tris-HCl, pH 8.1, 500 mM NaCl, 2 mM EDTA, 0.1% SDS, 1% Triton X-100), a LiCl buffer (10 mM Tris-HCl, pH 8.1, 0.25M LiCl, 1 mM EDTA, 1% NP-40, 1% deoxycholic acid), and finally with TE buffer, with each wash lasting for 5 minutes. Complexes were eluted in 1% SDS, 0.1 M NaHCO3. DNA was decrosslinked and qPCR was performed for regions of interest (Primer sequences are shown in Supplementary Table 2 ).

Proteins were extracted from CD8⁺ T cells (approximately 10⁶) with lysis buffer (50 mM Tris-HCl, pH 7.5, 15 mM EGTA, 100 mM NaCl, 0.1% Triton X-100 and the protease inhibitors). Protein samples were separated on 12% SDS-PAGE and blotted onto PVDF membranes. Blots were incubated with primary antibodies HOPX (sc-398703) and HRP-conjugated secondary antibodies (7076s, CST), and then visualized by the ECL chemiluminescence system (Cytiva).

Sorted T cell subsets were stimulated with anti-CD3 and anti-CD28 antibodies conjugated to microbubbles and placed into either 12-well or 96-well plates at 0.5 × 10⁶ cells/ml for nine days. The transduction with lentivirus (either overexpression or knock-down) was carried out 24 hours post-stimulation. Cells were collected every three days for cell counting and cell proliferation was measured by flow cytometry with Carboxyfluorescein succinimidyl ester (CFSE) dye (Molecular Probes, Eugene, Oregon). Proliferation analysis by CFSE dilution was performed in FlowJo_V10 to calculate replication index. To avoid inter-counter variation, all counting was performed by the same operator.

Sorted T cell subsets were washed once in phosphate-buffered saline (PBS) and diluted to 1 million per 500 μl PBS and 500 μl of PBS with 10 μM CFSE was added to a final cell suspension volume of 1 ml and final concentration of CFSE 5 μM). This was incubated for 10 min at RT, and protected from light, and these labeled T cells were stimulated with anti-CD3/CD28 and analyzed at indicated time. For surface staining, cells were stained with antibodies (CD4, CD8, CD45RA, CD62L from BD Biosciences, dilution 1:200) for 30 min in FACS buffer (HBSS, 0.01% NaN₃, 0.2%BSA). FACS data collection was done by using a FACScalibur flow cytometer (BD Biosciences). Routinely, 10,000 cells were collected for analysis. Flow cytometry data were analyzed using FlowJo software (Tree Star Inc., Ashland, OR).

Results are shown as the mean ± SEM for at least 3 repeated independent experiments for each group. Student’s t-test was performed to determine statistical differences between 2 groups and two-way analysis of variances (ANOVA) for multiple comparisons using GraphPad. p < 0.05 were considered to indicate statistical significance. All results were representative of at least three independent experiments.

While repeated T cell activation induces inhibitory receptor expression (5), it is currently unknown if a single round of activation is capable of inducing the negative regulators of T cell activation and proliferation. To investigate this, we analyzed the expression of HOPX during activation of human naïve CD8⁺ and CD4⁺ T cells (T_N, CD62L⁺CD45RA⁺) isolated from peripheral blood of healthy adults by cell sorting and stimulated with anti-CD3 and anti-CD28 antibodies in vitro ( Figure 1A ; Supplementary Figure 1A ). HOPX mRNA was present in freshly isolated T_N and increased significantly post-activation ( Figure 1B and Supplementary Figure S1B , p=0.021 and 0.036, for CD8⁺ and CD4⁺ T_N cells, respectively). It has been reported that human HOPX has three major isoforms (11) ( Supplementary Figure 1C ), and we found that all three isoforms were present in different CD8⁺ T_N cell subsets with HOPXb as the dominant isoform expressed in both freshly isolated (accounting for ~60% of total HOPX mRNA) and activated (accounting for ~70% of total HOPX mRNA) T cells ( Figure 1C ).

To further investigate the role of HOPX in T_N cell activation and proliferation, we used lentiviral transduction to express HOPX in a Jurkat T cell reporter cell line, which lacks endogenous HOPX expression ( Figure 1A ). After comparing the three isoforms (HOPXa, b, and c) individually, we found that HOPXb exhibited the strongest inhibitory effects on proliferation compared to the control and other isoforms ( Figure 1D , and Supplementary Figure 1D ). To investigate if HOPXb negatively regulates proliferation in primary T cells, we expressed HOPXb via lentiviral transduction in sorted CD8⁺ and CD4⁺ T_N cells and stimulated them in vitro with anti-CD3/CD28. Again, we found a significant reduction of cell growth in HOPXb-transduced T_N cells compared to control T_N cells for both CD8⁺ and CD4⁺ T_N cells ( Figure 1E and Supplementary Figure 1E ).

To determine if the reduced cell expansion observed upon overexpression of HOPXb was due to decreased proliferation or to increased cell death, we utilized a cell division tracking dye (CFSE) and found a reduction in the replication index (RI) of HOPXb-transduced CD8⁺ T_N cells (RI=5.18) compared to control CD8⁺ T_N cells (RI=7.39, p=0.055) on Day 6 and RI=8.86 vs 11.61, (p=0.022) on Day 9 ( Figures 1F, G ). Cell viability analysis showed no difference in the percentage of viable cells between HOPXb-overexpressing and control CD8⁺ and CD4⁺ T_N cells ( Figure 1H ; Supplementary Figure 1F ). Taken together, we demonstrated that HOPXb negatively regulated activation-induced proliferation of both CD8⁺ and CD4⁺ T_N cells by reducing cell cycle progression without affecting their survival.

To gain insight into the mechanism underlying HOPXb-mediated T_N cell proliferation inhibition, we profiled the transcriptome of HOPXb-transduced and control CD8⁺ T_N cells using microarray analysis. Three days after transduction with HOPXb-expressing or control lentivirus, transduced cell expressing a fluorescent reporter (tdTomato⁺) were isolated by cell sorting, and RNA was extracted for microarray analysis ( Supplementary Figure 2A ). We found that 229 genes were upregulated, while 824 genes were downregulated in HOPXb-transduced CD8⁺ T cells (including MYC,and NR4A1) compared to the control cells ( Figure 2A ). Gene expressions were reduced in HOPXb transduced Gene set enrichment analysis revealed that the downregulated genes to be predominantly associated with DNA replication pathway ( Figure 2A ), including key genes such as MYC, CCNA1, CCNA2, CDC20, and CDK2 ( Figures 2B, C ), which was further confirmed by qPCR analysis ( Supplementary Figure 2B ).

HOPX has been shown to interact with serum response factor (SRF) to modulate gene expression (19) and SRF binding sites, also known as CArG boxes, are involved in this process (20). To investigate whether HOPX is directly involved in regulating the expression of MYC gene in CD8⁺ T cells, we employed the anti-HOPX chromatin immunoprecipitation (ChIP) method and targeted a site that was indicated to be occupied by SRF (21, 22), located near the transcription start site (TSS) of MYC ( Supplementary Figure 2C ). We found a significant enrichment of the predicted DNA fragment from the MYC gene in the pull-down of HOPX antibody compared to control (IgG) in both activated (7.3-fold, p=0.03) and freshly isolated CD8⁺ T cells (3.1-fold enrichment) ( Figure 2D ). Taken together, these results support that HOPXb negatively regulates T_N cell proliferation by directly repressing MYC expression.

To gain insight into the mechanism underlying HOPXb-mediated T_N cell proliferation inhibition, we profiled the transcriptome of HOPXb-transduced and control CD8⁺ T_N cells using microarray analysis. Three days after transduction with HOPXb-expressing or control lentivirus, transduced cells expressed a fluorescent reporter (tdTomato⁺) were sorted, and RNA was extracted for microarray analysis. We found that 229 genes were upregulated, while 824 genes were downregulated in HOPXb-transduced cells compared to the control cells ( Supplementary Figure 2A ). Gene set enrichment analysis revealed that the downregulated genes to be predominantly associated with DNA replication pathway ( Figure 2A ), including key genes such as MYC, CCNA1, CCNA2, CDC20, and CDK2 ( Figures 2B, C ), which was further confirmed by qPCR analysis ( Supplementary Figure 2B ).

HOPX has been shown to interact with serum response factor (SRF) to modulate gene expression (19) and SRF binding sites, also known as CArG boxes, are involved in this process (20). To investigate whether HOPX is directly involved in regulating the expression of MYC gene in CD8⁺ T cells, we employed the anti-HOPX chromatin immunoprecipitation (ChIP) method and targeted a site that was indicated to be occupied by SRF (21, 22), located near the transcription start site (TSS) of MYC ( Figure 2E ). We found a significant enrichment of the predicted DNA fragment from the MYC gene in the pull-down of HOPX antibody compared to control (IgG) in both activated (7.3-fold, p=0.03) and freshly isolated CD8⁺ T cells (3.1-fold enrichment) ( Figure 2F ). Taken together, these results support that HOPXb negatively regulates T_N cell proliferation by directly repressing MYC expression.

The proliferation of T cells induced by TCR activation is a tightly regulated process that requires the activation and re-localization of multiple transcription factors (23). NR4A1 is an immediate-early gene following T cell receptor ligation (24) and a key regulator for T cell functions (25, 26). We found a significant reduction of NR4A1 expression in the HOPXb-transduced CD8⁺ T cells compared to the control ( Figure 2A and Supplementary Figure 2D ) and identified a distinct site within the promoter region of NR4A1 from analyzing the published SRF CHIP-seq dataset (22) ( Supplementary Figure 2E ). To confirm whether HOPX/SRF directly binds to the NR4A1 promoter, we performed ChIP-qPCR using anti-HOPX antibody and showed a significant enrichment (9.6-fold, p=0.05) of the DNA fragment containing the reported HOPX/SRF-binding site in the NR4A1 promoter region in activated CD8⁺ T cells ( Figure 2E ), providing evidence that HOPX/SRF directly regulates NR4A1 expression in CD8⁺ T cells. To further determine the regulatory role of HOPXb on NR4A1 during TCR activation, we generated a TCR-signaling reporter system (NJ76-TRAV8-6-TRAJ42) by introducing a functional TCR in tandem with an mCherry reporter into NR4A1-GFP (NJ76) cell line (27) ( Figure 2F ). The HOPXb-tdTomato⁺ and control constructs were delivered into these cells by viral transduction. We compared the strength of GFP signal between HOPXb and control transduced NJ76-TRAV8-6-TRAJ42 cells after anti-CD3/CD28 stimulation and found a significant reduction of GFP level in HOPXb-transduced cells compared to the control cells ( Figure 2G ). This suggests that HOPXb represses both cell cycle-related genes (MYC) and an immediate-early TCR activation-induced gene and TCR signaling (NR4A1).

It is reported that HOPX expression increases from T_N to memory (T_EM) T cells (13), but its role in naïve and memory T cells has not been fully characterized. We hypothesize that HOPX regulates the proliferation behavior of different T cell subsets. To directly test this, we analyzed the proliferation of CD8⁺ T_N (CD28⁺/CD45RA⁺/CD62L⁺) and T_EM (CD28⁺/CD45RA^-/CD62L^-) cells following anti-CD3/CD28 antibody stimulation in vitro and found that HOPX mRNA levels were four-five times higher in T_EM cells than in T_N cells (0 - 6 days post-stimulation ( Figure 3A ). Correspondingly, CD8⁺ T_EM cells exhibited a significantly slower growth than CD8⁺ T_N cells over 9-day culture ( Figure 3B ).

To investigate whether the elevated expression of HOPX causes reduced the proliferative capacity of CD8⁺ T_EM cells, we first compared expression of HOPXb in transduced and control T_N cells and control T_EM cells. In comparison to the control-virus-transduced CD8⁺ T_N cells, we observed approximately a 2-fold increase in HOPXb expression in HOPXb-transduced CD8⁺ T_N cells and a 3-fold increase in activated CD8⁺ T_EM cells post 3- and 6-day stimulations ( Figure 3C ). We found that HOPXb level correlates well with activation-induced proliferation of control T_N cells, HOPXb-transduced T_N cells, and T_EM cells ( Figure 3D ). We then reduced expression of HOPX in T cells via shRNA targeting exon 2 of HOPX ( Supplementary Figure 3 ). To verify the efficiency of HOPX knockdown in T cells, we transduced the knockdown vector into HOPX-expressing Jurkat T cells and observed a 20-30% reduction in the protein level of HOPX 3 days post-transduction ( Figure 3E ). When we used shRNA to knockdown HOPX expression in CD8⁺ T_EM, we observed a 1.4-fold increase in cell number in shHOPX-transduced CD8⁺ T_EM cells compared to control cells ( Figure 3F ). Importantly, knockdown of HOPX eliminated the reduced proliferation of CD8⁺ T_EM cells compared to that of CD8⁺ T_N cells ( Figure 3F ). Collectively, these results show that the level of HOPXb determines the degree of activation induced proliferation of CD8⁺ T_N and T_EM cells.

Age-related changes in CD8⁺ T cell responses are, in part, caused by the reduced activation-induced proliferation of CD8⁺ T_N cells (28–30). Our recent study has identified over 1,500 genes whose expressions significantly increased in human CD8⁺ T cell subsets over 70 years of adult life. Many of these genes are involved in both immune functions (antigen processing and presentation, exocytosis, etc.) and general cellular functions (oxidant detoxification, oxidative phosphorylation, etc.) (31). Interestingly, we observed a significant increase in HOPX mRNA levels with age in both CD8⁺ T_N cells (22,603 cells from 22 adults, p=0.024) and T_EM cells (27,709 cells from 22 adults, p<0.001) in this published dataset (31) ( Figure 4A ). To determine if such age-related changes remain post-activation, we isolated and stimulated (anti-CD3/CD28) CD8⁺ T cell subsets from young (age<40 years old, n=8) and old donors (age≥70 years old, n=8) ( Supplementary Figure 4A ). In stimulated CD8⁺ T_N cells ( Figure 4B ), we found significantly higher levels of HOPX mRNA in old than in young donors, whiles this was not the case in stimulated memory CD8⁺ subsets (T_CM and T_EM cells) ( Supplementary Figures 4B, C ). Again, HOPXb was the most abundant HOPX isoform in CD8⁺ T cell subsets of both young and old donors ( Supplementary Figures 4D–F ). We further confirmed the age-associated increase in mRNA level resulting in increased protein in activated CD8⁺ T_N cells from old compared to those from young donors ( Figure 4C ). Functionally, we found that CD28⁺CD8⁺ T_N cells from old donors were significantly less expanded after in vitro stimulation than those from young donors ( Figure 4D ). To further determine if activation-induced proliferation of T cells from old adults could be improved by reduced HOPX expression, we applied the shRNA mediated HOPX knockdown in CD8⁺ T_N and T_EM cells isolated from healthy old adults (Age≥70 years old). shRNA and control virus transduction yielded comparable percentage of reporter (EGFP⁺) ( Supplementary Figure 4G ). We found that reduced HOPX expression resulted in improved proliferation in both CD8⁺ T_N (1.2-fold increase, p=0.013, n=9) and CD8⁺ T_EM (1.5-fold increase, p=0.009, n=10) cells compared to controls ( Figure 4E ). The proliferations of shRNA treated CD8⁺ T_N and T_EM from old adults were comparable to the control treated CD8⁺ T_N and T_EM from young adults. Taken together, increased HOPX expression with age contributed to the reduced activation-induced expansion of CD8⁺ T_N and T_EM cells.