T cells deficient in the tyrosine phosphatase SHP-1 resist suppression by regulatory T cells

Abstract

The balance between activation of T cells and their suppression by regulatory T cells (Treg) is dysregulated in autoimmune diseases and cancer. Autoimmune diseases feature T cells that are resistant to suppression by Tregs, whereas in cancer, T cells are unable to mount anti-tumor responses due to the Treg-enriched suppressive microenvironment. Here, we observe that loss of the tyrosine phosphatase SHP-1, a negative regulator of T cell receptor (TCR) signaling, renders naïve CD4+ and CD8+ T cells resistant to Treg-mediated suppression in a T cell-intrinsic manner. At the intracellular level, SHP-1 controlled the extent of Akt activation, which has been linked to the induction of T cell resistance to Treg suppression. Finally, under conditions of homeostatic expansion, SHP-1-deficient CD4+ T cells resisted Treg suppression in vivo. Collectively, these data establish SHP-1 as a critical player in setting the threshold downstream of TCR signaling and identify a novel function of SHP-1 as a regulator of T cell susceptibility to Treg-mediated suppression in vitro and in vivo. Thus, SHP-1 could represent a potential novel immunotherapeutic target to modulate susceptibility of T cells to Treg suppression.

Introduction

Regulatory T cells (Treg) play an essential role in shaping T cell responses and maintaining immune homeostasis¹. Deficits in Treg function or number allow T cell responses to go unchecked, leading to the development of autoimmunity and chronic inflammatory diseases². Dysregulation of the balance between activation and suppression of T cells can also occur when T cells become resistant to Treg-mediated suppression². Many autoimmune diseases, including type 1 diabetes, multiple sclerosis, rheumatoid arthritis, systemic lupus erythematosus, and inflammatory bowel disease, feature not only impaired Tregs but also T cells that are resistant to suppression³. However, the potential mechanism(s) by which T cells might acquire resistance to Treg-mediated suppression remain unclear. While several extracellular factors have been linked to inducing resistance in T cells³, the intracellular signaling mechanisms that can render T cells resistant to Treg suppression are poorly defined. Further, strong activation through the T cell receptor (TCR) and/or costimulatory receptors can cause T cells to become refractory to Treg suppression^4–8, but the specific pathways allowing this resistance remain elusive. Similarly, while resistance to suppression occurs in both CD4+ and CD8+ T cells³, whether resistance is induced by the same mechanism in both subsets is not known.

SHP-1 is a cytoplasmic protein tyrosine phosphatase expressed in all hematopoietic cells, which has been implicated in the regulation of TCR-mediated signaling in T cells⁹, including the PI3K/Akt pathway¹⁰. We¹¹ and others^12,13 have previously shown that SHP-1-deficient T cells are hyper-responsive to TCR stimulation. This was done using the motheaten (me/me) mouse model, in which all hematopoietic cells lack SHP-1 due to a splicing mutation¹⁴, as well as cell lines expressing dominant negative mutant forms of SHP-1¹². However, one recent study¹⁵, using conditional T cell deletion of SHP-1 via CD4-Cre, challenged the role of SHP-1 in regulating T cell development, while another report using the same mouse model confirmed the role of SHP-1 during T cell development¹⁶. Here, we generated a conditional knockout mouse model wherein SHP-1 deletion is driven by the distal Lck promoter¹⁷, resulting in abrogation of SHP-1 expression in post-selection thymocytes. This model allows largely normal T cell development such that any phenotypic and/or functional alterations observed due to SHP-1 deficiency can be directly ascribed to its role in mature T cells^11,12,18. Using this approach, we show that SHP-1 negatively regulates the activation and proliferation of CD4+ and CD8+ T cells in response to TCR stimulation, and that in the absence of SHP-1, T cells become resistant to Treg-mediated suppression. Such resistance is T cell-intrinsic, as SHP-1^−/− T cells could not induce “bystander resistance” when co-cultured with wild type T cells. Our data also suggest a role for the PI3K/Akt pathway in mediating both CD4+ and CD8+ T cells to resist suppression. This resistance of CD4+ SHP-1^−/− T cells to Treg-mediated suppression was also observed during homeostatic expansion in vivo. Collectively, these data identify a novel function of SHP-1 in regulating the susceptibility of T cells to Treg-mediated suppression in vitro and in vivo, through controlling the strength of signal received via the TCR and attenuating subsequent activation of the downstream PI3K/Akt pathway.

Materials and Methods

Mice

SHP-1^flox/flox (SHP-1^f/f) mice¹⁹ (generously provided by B. Neel) were crossed to distal Lck-Cre (dLck-Cre) mice¹⁷ purchased from The Jackson Laboratory (Bar Harbor, ME). Genotyping of all mice was done by PCR as described previously for the SHP-1^f/f allele¹⁹ and dLck-Cre alelle¹⁷. For all experiments, 6- to 10-week old female and male mice were used, and control mice were dLck-Cre- SHP-1^f/f or dLck-Cre+ SHP-1^+/+ littermates of dLck-Cre+ SHP-1^f/f mice. Rag1^−/− mice were purchased from The Jackson Laboratory. CD45.1 wild type C57BL/6 mice were purchased from Charles River. All mice were bred and maintained in accordance with the policies of the Institutional Animal Care and Use Committee at the University of Virginia. All experiments involving mice were conducted with the approval of Institutional Animal Care and Use Committee.

Isolation and Purification of Primary Cells

CD4+ T cells were isolated from peripheral lymph nodes (combined inguinal, axillary, brachial, cervical, sacral, and renal nodes) or spleens by negative selection using the CD4+ T cell isolation kit (Miltenyi Biotec, Auburn, CA) according to manufacturer’s protocol. CD8+ T cells were isolated from spleen by negative selection using the CD8α+ T cell isolation kit (Miltenyi Biotec). For naïve CD4 or CD8 T cell experiments, CD4+CD44^lo or CD8+CD44^lo T cells were isolated from spleens by negative selection using the naïve CD4 T cell isolation kit or the CD8 T cell isolation kit, respectively (Miltenyi Biotec). For splenic T cell isolation, red blood cells were lysed using BD Pharm lyse buffer (BD Biosciences, San Jose, CA) before T cell isolation. For Treg isolation, CD4+ T cells were subsequently labeled with CD25-PE in order to separate conventional T cells (Tcon defined as CD4⁺CD25⁻) and Tregs (CD4⁺CD25⁺). Labeled CD4+ cells were run on an AutoMACS Pro separator (Miltenyi Biotec) using the posseld2 program in order to obtain Treg cells with >85% purity as assessed by Foxp3+CD25+ staining. CD4+ T cell-depleted splenocytes were irradiated (2000rad) and used as APCs in culture where indicated.

Flow cytometry

Cells were stained directly after isolation or harvested after 24, 72, or 96 hours of culture as indicated. Cells were surface stained with anti-CD4, anti-CD25 (eBiosciences, San Diego, CA), anti-CD8, anti-CD44, anti-CD62L, anti-CD69, anti-CD45.1, anti-CD45.2 (BD Biosciences) in PBS supplemented with 1% BSA and 0.1% sodium azide. Staining for live cells was done following surface staining and washing, using Live/Dead Fixable Dye (Life Technologies, Carlsbad, CA). Cells were then fixed with BD Fix/Lyse (BD Biosciences) and washed. For intracellular Foxp3 staining, cells were fixed and permeabilized using the Foxp3 staining buffer set (eBiosciences) according to the manufacturer’s protocol and stained with anti-Foxp3 (eBiosciences). For phospho-Akt intracellular staining, cells were fixed and permeabilized using BD Cytofix/Cytoperm (BD Biosciences) according to manufacturer’s protocol and stained with anti-pAkt T308 (Cell Signaling Technology, Danvers, MA). For caspase-3 staining, cells were stained with the CaspGlow kit (eBiosciences) according to manufacturer’s protocol for the last 60 minutes of in vitro culture. Stained cells were collected on a BD FacsCanto I or II, using FACSDiva version 8 software (BD Biosciences), or using a Beckman Coulter CytoFlex and CytExpert Software (Beckman Coulter, Brea, CA) and subsequent analyses were done using FlowJo Software version 9.9 or version 10.1 (FlowJo, LLC, Ashland, OR). Analyses were performed on singlet-gated cells as defined by FSC-W vs. FSC-A, and live cells as defined by Live/Dead dye negative. Gates were set based on FMO controls.

Proliferation and suppression assays

Assessment via CellTrace Violet dilution

To assess proliferation, isolated T cells [CD4+CD25- (Tcon cells), CD4+CD44^lo (naïve CD4+ T cells), CD8+, or CD8+CD44^lo (naïve CD8+ T cells)] were stained with 5μM CellTrace Violet for 20min at 37°C followed by quenching with pre-warmed complete RPMI for 5min at 37°C (Life Technologies). Stained cells were washed, and 2.5×10⁴ T cells were plated (in quadruplicate, pooled at time of harvest) in a total volume of 200μL RPMI 1640 complete medium (supplemented with 10% FBS, 50μM 2-ME, 2mM _L-glutamine, 10mM HEPES, MEM non-essential amino acids, 1mM sodium pyruvate, and 100U/mL pen/strep) in round-bottom 96-well plates. Irradiated (2000rad), CD4+ T cell-depleted splenocytes were added at 5×10⁴ cells/well along with anti-CD3 Ab (2C11; CedarLane Laboratories, Burlington, NC) at 10-1000ng/mL as indicated. For suppression assays, CD4+CD25+ Treg cells were plated with responder T cells at indicated ratios. For proliferation assays, cells were cultured for 72 or 96 hours, and for suppression assays cells were cultured for 96hrs followed by flow cytometric analyses.

Analysis of Proliferation Assay

CellTrace Violet dilution was assessed by flow cytometry, and subsequently analyzed using FlowJo v 9.9 Software Proliferation Wizard Platform (FlowJo, LLC.) Briefly, after sequentially gating on Singlets, Live cells, CD4-positive cells, and CellTrace Violet-positive cells, the percent of responding (dividing) cells relative to the input was obtained using the provided software algorithm.

Analysis of Suppression Assay

To compensate for the increased baseline responsiveness of SHP-1^−/− T cells, the percentage of responding cells in the no Treg condition was set to 100% (maximum responsiveness) for each genotype. The percentage of responding cells was calculated as described above for the proliferation analyses for all Treg:T cell ratios and normalized to the maximum responsiveness for their own genotype (no Treg condition). Percent suppression equals 100 minus percent responding cells.

24 hour T cell activation

CD4+CD25− Tcon cells or naïve (CD44^lo) CD8+ T cells were isolated from spleens of indicated mice and 2.5×10⁴ cells were cultured per well in a 96-well round bottom plate with 5×10⁴ irradiated (2000rad) CD4+ T cell-depleted splenocytes and indicated doses of anti-CD3 Ab (2C11, CedarLane Laboratories). After 24 hours, cells were harvested and stained for flow cytometric analysis of CD25 and pAkt T308 expression.

Immunoblotting

SHP-1 protein level in CD4+ T cells was assessed by lysing 5×10⁵ cells in NP40 lysis buffer (1% Nonidet P-40, 150mM sodium chloride, 50mM Tris, 4mM sodium pyrophosphate, 5mM sodium fluoride, 10ug/mL sodium vanadate, 50ug/mL antipain, 40ug/mL PMSF, 1X Protease Inhibitor Cocktail [Sigma Aldrich, St. Louis, MO]) and resolving lysates on an Any KD TGX Tris-glycine-SDS gel (BioRad, Hercules, CA). Blots were probed with monoclonal anti-SHP-1 (clone 1SH01, Neomarkers/ThermoFisher Scientific, Fremont, CA) and re-probed for β-actin as a loading control (anti-β actin-HRP, clone AC-15, Sigma-Aldrich). Blots were imaged using the ChemiDoc Touch gel imaging system (BioRad). Bands densities were quantified using ImageLab (BioRad) software after normalization to loading control.

In vivo T cell transfer

Tcon (CD4+CD25−) cells were isolated by MACS as described above from spleens of CD45.2 SHP-1^+/+ or SHP-1^−/− mice and CD45.1 wild type mice. Tcon cells were labeled with 5μM CellTrace Violet, and Treg (CD4+CD25+) cells were isolated from SHP-1^+/+ mice and pooled. Tcon cells were resuspended at a 1:1 ratio of either CD45.2 SHP-1^+/+: CD45.1 wild type Tcon cells or CD45.2 SHP-1^−/−: CD45.1 wild type Tcon cells, and a total of 3×10⁶ Tcon cells from either mix were injected i.v. in 200uL sterile PBS via the tail vein into Rag1^−/− recipients. Additionally, half the recipient mice also received 7.5×10⁵ SHP-1^+/+ Tregs along with Tcon cells (1:4 Treg:Tcon ratio). After 10 days, spleens were harvested from recipient mice and stained for flow cytometric analysis. Donor and recipient mice were age-matched.

Statistical analysis

T cell proliferation, CD25 upregulation, proliferation index, and suppression assays using CD4+CD25− or total CD8+ T cells were analyzed using a three-way ANOVA with a 95% confidence interval. A Student’s t test was used to analyze the comparison of percentage and absolute number of CD44+ T cells from SHP-1^+/+ or SHP-1^−/− mice. A Student’s t test was used to analyze naive CD4+ and naive CD8+ T cell suppression assay data for each Treg:T cell ratio. A two-way ANOVA with a Sidak’s multiple comparison post-test was used to analyze cell death and apoptosis data. A one-way ANOVA with a Tukey’s multiple comparison post-test was performed to analyze the absolute number of LN and splenic T cells in dLck-Cre SHP-1^f/f mice, the percentage of responding cells in co-culture experiments, and the percent suppression of in vivo T cell transfer experiments. For pAkt T308 flow cytometric data, a one-column t test was applied to the fold change pAkt MFI values of SHP-1^−/− T cells compared to the pAkt MFI of SHP-1^+/+ cells for each anti-CD3 dose, with a null hypothesis of 1 (if no change from control, fold change = 1). p values ≤0.05 were considered significant.

Results

SHP-1 sets threshold for activation and proliferation of T cells in response to TCR stimulation

While several previous studies have suggested that SHP-1 is a negative regulator of signaling downstream of the TCR, based on in vitro cell culture studies and primary T cells from total body knockout of SHP-1 (reviewed in ⁹), a recent study performed using conditional deletion of SHP-1 in T cells has disagreed with this notion¹⁵. One potential reason for this discrepancy might have been the type of Cre line that was used to delete SHP-1 by Johnson et al., as the CD4-Cre used gets expressed from earlier stages of T cell development. To test this possibility, we crossed mice carrying floxed alleles of Ptpn6 (Shp1)¹⁹ with mice that express the Cre recombinase under the control of the distal promoter of Lck¹⁷. The distal Lck promoter drives Cre expression at late stages of T cell development, allowing TCR-dependent selection to occur under conditions of SHP-1 sufficiency^12,16,18,20. We confirmed that SHP-1 was deleted in peripheral CD4+ (Supplementary Fig. 1A) and CD8+ T cells (Supplementary Fig. 1B) from the lymph nodes and spleen of dLck-Cre+ SHP-1^f/f (referred to here as SHP-1^−/−) mice. Importantly, we observed no changes in the composition of the thymic or peripheral T cell compartments with respect to absolute numbers (Supplementary Fig. 1C), or percentages of CD4+ or CD8+ T cells or Treg cells in the lymph nodes or spleens (Supplementary Figs. 1 D and E). Expression of dLck-Cre alone did not affect the peripheral T cell compartment, consistent with previous reports²¹.

To assess the role of SHP-1 during T cell activation, we first compared the proliferation capacities of SHP-1^+/+ and SHP-1^−/− CD4+CD25− T cells, from hereon referred to as conventional T (Tcon) cells to differentiate them from CD4+ Treg cells. We found that a greater percentage of SHP-1^−/− Tcon cells proliferated compared to Tcon cells from SHP-1^+/+ mice. Enhanced proliferation in SHP-1^−/−Tcon cells was especially apparent at sub-optimal concentrations of anti-CD3 stimulation (Fig. 1A). We considered three reasons for SHP-1^−/− Tcon cells to display the observed increase in proliferation, which are not mutually exclusive: (1) an increase in the percentage of cells that initially become activated and go on to proliferate, (2) a decreased cell cycle time, and/or (3) an increased survival of cells in the culture. We first determined whether a greater proportion of SHP-1^−/− Tcon cells responded to TCR stimulation by using the FlowJo Proliferation Platform algorithm, which takes into account the number of cells in each round of division relative to the input cells²², and thereby estimates the fraction of T cells that initially responded to the stimulation. Based on this metric, we found a significant increase in the percentage of responding SHP-1^−/− Tcon cells compared to SHP-1^+/+ Tcon cells, with the largest difference at the lowest stimulation dose (Fig. 1B). To complement this finding, we assessed the upregulation of CD25 (IL-2Rα) as a measure of early Tcon cell activation, and found that a significantly greater percentage of SHP-1^−/− Tcon cells were CD25+ after 24 hours of stimulation compared to SHP-1^+/+ Tcon cells (Fig. 1C). Importantly, we observed no CD25 up-regulation on Tcon cells of either genotype in the absence anti-CD3 stimulation, indicating that any observed T cell activation was TCR/CD3 stimulation-dependent. Interestingly, SHP-1^+/+ Tcon cells reached a maximum percentage of CD25+ cells at 150ng/mL anti-CD3, with no further increase at 1000ng/mL anti-CD3, whereas the subpopulation of responding SHP-1^−/− Tcon cells increased further at 1000ng/mL anti-CD3 compared to 150ng/mL. Up-regulation of CD69, another marker of activation, followed the same pattern (data not shown). These data suggest that there is a greater percentage of SHP-1^−/− Tcon cells that respond and are activated by any given stimulation.

Figure 1. SHP-1 limits the number of T cells responding to TCR stimulation.

(A) 72-hour proliferation of splenic CellTrace Violet-labeled CD4+CD25- T cells isolated from SHP-1^+/+ or SHP-1^−/− mice. Proliferation was measured in response to indicated concentrations of anti-CD3 and irradiated CD4+ T cell-depleted splenocytes as APCs. Histograms shown are representative of 3 independent experiments, n=5-7 per genotype. Note that the number of cells on the y axis for histograms is greater for SHP-1^−/− T cells than SHP-1^+/+ T cells. (B) Percent of CD4+ T cells within each culture initially responding to the indicated stimulation. Data were obtained from the proliferation assays presented in A. Percent of T cell responders was calculated using the precursor frequency algorithm of the FlowJo Proliferation Platform, which takes into account the number of cells in each round of division relative to the input cells, and thereby estimates the fraction of T cells that initially responded to the stimulation. (C) Proliferation assays were set up as described in A, but cells were harvested after 24 hours and assessed for CD25 surface expression. Data represents 3 independent experiments, n=5-9 per genotype. (D) Proliferation index, which corresponds to the average rounds of division of T cells, was obtained from the proliferation assays presented in A using the FlowJo Proliferation Platform. (E) Proliferation assays were set up as described in A. Cells were stained for activated caspase-3 (Casp3) with Fitc-DEVD-FMK for the last hour of culture before harvest and flow cytometric analyses. Data represent percent of Casp3+ cells within CD4+ T cell population, n=3 per genotype. A standard regression ANOVA was performed for B-D, a two-way ANOVA with Sidak’s multiple comparison post-test was used for statistical analysis of E. Error bars indicate ±SEM; * p≤0.05, **p≤0.01, ***p≤0.001.

Second, when we calculated the proliferation index of each sample (using the FlowJo Proliferation Platform that provides the average number of cellular divisions of the cells that divided in culture), SHP-1^−/− and SHP-1^+/+ Tcon cells underwent comparable rounds of divisions at any given dose of stimulation, with a slight increase in divisions at higher concentrations of stimulation (Fig. 1 D). These data indicate that SHP-1 deficiency did not affect cell cycle time. To assess whether SHP-1^−/− Tcon cells had an in vitro survival advantage over SHP-1^+/+ Tcon cells, we stained cells for activated caspase-3, a marker of apoptosis. After 24 hours in culture, we observed very little apoptosis among Tcon cells (<1%) with no significant difference between SHP-1^+/+ or SHP-1^−/− Tcon cells in terms of apoptosis or cell death (Supplementary Fig. 2A, B). By 72 hours, we still observed very low levels of apoptosis (≤ 2%) with no statistically significant differences between SHP-1^−/− Tcon cells and SHP-1^+/+ Tcon cells (Fig. 1E). Taken together, these data demonstrate that SHP-1 controls the extent of TCR/CD3-driven proliferation by setting the threshold that determines the subpopulation of T cells responding to a given TCR stimulation.

CD4+ T cells lacking SHP-1 resist in vitro Treg suppression

Since our data indicated that SHP-1 lowered the threshold for Tcon cell activation and proliferation, we asked whether SHP-1 also regulated the susceptibility of Tcon cells to Treg-mediated suppression. Using an in vitro suppression assay, Tcon cells from SHP-1^−/− and SHP-1^+/+ mice were assessed for their susceptibility to wild type Treg cell-mediated suppression (Fig. 2A). Strikingly, SHP-1^−/− Tcon cells displayed ~3-fold greater responsiveness compared to SHP-1^+/+ Tcon cells, even at the maximally suppressive condition (Fig. 2A). Even after normalization to account for the increased baseline proliferation (no Treg condition in Fig. 2A), SHP-1^−/− Tcon cells were significantly less suppressed by Tregs than SHP-1^+/+ Tcon cells (Fig. 2B), indicating that SHP-1 can influence the level of susceptibility to in vitro Treg-mediated suppression.

Figure 2. SHP-1^−/− CD4+ T cells resist Treg-mediated suppression.

(A) Splenic CD4+CD25- T (Tcon) cells were isolated from SHP-1^+/+ or SHP-1^−/− mice, labeled with CellTrace Violet and cultured either alone or with wild type Tregs at the indicated ratios in the presence of 150ng/mL anti-CD3 and irradiated CD4+ T cell-depleted splenocytes as APCs, and proliferation was measured after 4 days. Histograms shown are representative of 5 independent experiments, n=8–10 mice per genotype. Bold numbers indicate most significant differences observed. (B) Suppression was calculated by normalizing each data point to the corresponding baseline proliferation (no Tregs, maximal response = 100% proliferation), which was then subtracted from 100 % proliferation. Note that as described in Fig. 1, proliferation was computed using FlowJo Proliferation Platform, which takes into account the number of cells in each round of division relative to the input cells, and thereby estimates the fraction of T cells that initially responded to the stimulation. A three-way ANOVA was performed. (C) Naïve CD4+CD44^lo T cells were purified from spleens of SHP-1^+/+ or SHP-1^−/− mice and suppression assays were performed as described in A in the presence of 30 ng/mL anti-CD3 and CD4+ T cell-depleted splenocytes as APCs, n=3 mice per genotype. (D) Suppression was calculated as in B. Student’s t tests were performed for each Treg:T cell ratio. Error bars indicate ±SEM; * p≤0.05, **p≤0.01, ***p≤0.001.

To determine whether the observed resistance to suppression in Tcon cells could be attributable to an expanded memory T cell population in mice with SHP-1^−/− T cells ^23,24, we first assessed whether there were any differences in the memory T cell compartment of dLck-Cre SHP-1^f/f mice compared to SHP-1^+/+ mice, as has been described for me/me mice²⁵ and CD4-Cre SHP-1^f/f mice^15,16. However, we did not observe an increase in percentage (Supplementary Fig. 3A, B) or absolute number (Supplementary Fig. 3C) of Ag-experienced/memory-like CD44^hi CD4+CD25-Foxp3- Tcon cells in the lymph nodes or spleens of dLck-Cre+ SHP-1^f/f mice compared to control SHP-1-sufficient mice. As further indication that the composition of the CD4+ T cell compartment in the dLck-Cre+ SHP-1^f/f mice was not altered, we detected no differences in the percentage of cells expressing activation markers CD69 or CD25 (data not shown). Furthermore, SHP-1^−/− CD4+ T cells depleted of the CD44^hi subpopulation (referred to here as naïve CD44^lo T cells, Supplementary Fig. 3D) retained a greater responsiveness to TCR stimulation (Supplementary Fig. 3E, F), without any changes in cell cycle time. These data are consistent with SHP-1 regulating signaling downstream of the TCR in naïve T cell subsets. Moreover, SHP-1^−/− naïve CD4+ T cells were resistant to Treg-mediated suppression in vitro (Figs. 2 C and D), confirming what we observed in the CD4+CD25− Tcon population. Together, these data suggest that SHP-1 regulates the susceptibility of CD4+ T cells to Treg-mediated suppression in vitro.

CD8+ T cells lacking SHP-1 also resist in vitro Treg suppression

Resistance of T cells to Treg-mediated suppression has not only been observed in CD4+ T cells, but also in CD8+ T cells^26–29, which has important clinical implications for cancer immunotherapy and chronic viral infection therapies. Like SHP-1^−/− CD4+ T cells, SHP-1^−/− CD8+ T cells exhibited greater responsiveness to TCR stimulation compared to SHP-1^+/+ CD8 T cells (Fig. 3A) without any detectable changes in cell cycle time (data not shown). SHP-1^−/− CD8+ T cells also resisted Treg-mediated suppression (Fig. 3A, B). However, in contrast to the CD4+ T cell compartment, we did observe a substantial increase in the percentage and number of CD8+CD44^hi T cells in the lymph nodes and spleens of SHP-1^−/− mice compared to SHP-1^+/+ mice (Supplementary Figs. 4 A–C). To determine if SHP-1 deficiency also conferred naïve CD8+ T cells with resistance to Treg suppression, we isolated naïve CD8+ (CD44^lo) T cells and measured their suppression in vitro (Fig. 3C). A greater proportion of SHP-1^−/− naïve CD8+ T cells responded to TCR stimulation in the absence of Treg cells (Fig. 3C). Similar to what we observed for CD4+ T cells, there were no significant differences in cell death or apoptosis between SHP-1^+/+ and SHP-1^−/− naïve CD8+ T cells after 24 hours of stimulation across a range of anti-CD3 stimulation (Supplementary Fig. 2C, D). There was also no observed survival advantage in SHP-1^−/− naïve CD8+ T cells after 3 days of stimulation, and in fact at the highest dose of stimulation, SHP-1^−/− naïve CD8+ T cells displayed enhanced apoptosis compared to SHP-1^+/+ cells, possibly due to an increase in activation-induced cell death (Supplementary Fig. 2E). Furthermore, SHP-1^−/− naïve CD8+ T cells exhibited resistance to suppression, similar to the total CD8+ T cell population (Fig. 3D), indicating that the phenotype was independent of the expanded antigen-experienced/memory-like CD8+ T cell subpopulation. These data demonstrate a role for SHP-1 in regulating susceptibility of not only CD4+ T cells, but also CD8+ T cells, to Treg-mediated suppression, likely by a similar mechanism in both T cell subsets.

Figure 3. SHP-1^−/− CD8+ T cells resist Treg-mediated suppression.

(A) Splenic CD8+ T cells were isolated from SHP-1^+/+ or SHP-1^−/− mice, labeled with CellTrace Violet and cultured either alone or with wild type Tregs at the indicated ratios in the presence of 10ng/mL anti-CD3 and CD4+ T cell-depleted splenocytes as APCs, and proliferation was measured after 3 days. Data are representative of 2 independent experiments; n=4 mice per genotype. (B) The percent responding cells was obtained using FlowJo Proliferation Platform as described in Figs. 1 and 2. Suppression was calculated by normalizing each data point to the corresponding baseline proliferation (no Tregs, maximal response = 100% proliferation), which was then subtracted from 100 % proliferation. A three-way ANOVA was performed. (C) Naïve CD8 T cells (CD8+CD44^lo) were isolated, labeled with CellTrace Violet, and cultured with Tregs as described in A. n=3 mice per genotype. (D) Percent suppression was obtained as in B. Student’s t tests were performed for each Treg:T cell ratio. Error bars indicate ±SEM. * p≤0.05, **p≤0.01.

SHP-1 regulates TCR signaling and susceptibility to Treg suppression in a cell-intrinsic manner

To further understand how SHP-1 regulates signaling downstream of TCR/CD3 stimulation and susceptibility to Treg suppression, we asked whether these phenotypes occurred in a cell-intrinsic and/or extrinsic manner. To investigate whether SHP-1^−/− CD4+CD25− Tcon cells could transfer their enhanced TCR responsiveness to neighboring SHP-1^+/+ Tcon cells via a soluble mediator, we set up co-cultures (Fig. 4 A); we labeled either SHP-1^+/+ or SHP-1^−/− Tcon cells with CellTrace proliferation dye, mixed them at a 1:1 ratio with SHP-1^+/+ or SHP-1^−/− Tcon cells, respectively, and assessed the proliferation of labeled cells (Fig. 4B). We found that adding SHP-1^−/− Tcon cells to SHP-1^+/+ Tcon cells did not enhance the response of the SHP-1^+/+ Tcon cells (Fig. 4B), indicating that the enhanced responsiveness to TCR stimulation cannot be transmitted to neighboring T cells. We next asked whether SHP-1^−/− Tcon cells could render their local environment resistance-promoting for Treg mediated suppression, perhaps via the production of cytokines or other factors that could directly influence APCs. If this were the case, one would expect that SHP-1^−/− Tcon cells would be capable of inducing “bystander resistance” in SHP-1^+/+ Tcon cells exposed to the same environment. Using the same experimental co-culture set-up as above, but in the presence of wild type Tregs, the addition of SHP-1^−/− Tcon cells to SHP-1^+/+ Tcon cells did not induce any resistance to suppression in the SHP-1^+/+ Tcon cell population (Fig. 4C). These data suggest that SHP-1^−/− Tcon cells resist Treg suppression by a cell-intrinsic mechanism, which does not affect neighboring cells or induce bystander resistance.

Figure 4. SHP-1-mediated T cell phenotypes are cell intrinsic.

(A) Schematic representation of experimental setup. Splenic CD4+CD25- T cells (Tcon cells) were isolated from SHP-1^+/+ or SHP-1^−/− mice and labeled with CellTrace dyes. Differently-labeled Tcon cells of indicated genotypes were co-cultured at 1:1 ratio in the presence of 30ng/mL anti-CD3 and irradiated CD4+ T cell-depleted splenocytes as APCs. (B) After 72 hours, the proliferation of the CellTrace Violet-labeled cells was measured and assessed using the FlowJo Proliferation platform as in Fig. 1. Graph shows percent responding cells of indicated genotype, compiled from two independent experiments; n= 6 mice per genotype. (C) Using the same setup as in B, with the addition of wild type Tregs at a ratio of 1:4 of Treg:total Tcon cells. After 96 hours, proliferation of CellTrace Violet-labeled cells was measured and analyzed as in B. Graph shows percent of suppression (calculated as in Figs. 2 and 3). A one-way ANOVA was performed on data in B and C. Error bars indicate ±SEM; * p≤0.05.

SHP-1 deficiency enhances activation of the Akt pathway

Our data suggested that SHP-1 deficiency led to intracellular changes in the signaling pathways downstream of the TCR, which might ultimately mediate T cell resistance to suppression. A number of reports have implicated enhanced PI3K/Akt in conferring T cells with resistance to suppression^{24,27,29–33}. Moreover, it was previously demonstrated that SHP-1 negatively regulates the PI3K/Akt pathway in me/me thymocytes^10,34. We therefore assessed the phosphorylation of Akt at T308 as a measure of Akt activation³⁵ in response to TCR/CD3 stimulation. At 24 hours post-stimulation, SHP-1^−/− CD4+ T cells displayed enhanced Akt phosphorylation over a range of TCR stimulation conditions compared to SHP-1^+/+ CD4+ T cells (Figs. 5 A and B). Additionally, there was also a slightly higher baseline activation of Akt in SHP-1^−/− CD4+ T cells that received no TCR/CD3 stimulation. We observed the same enhanced Akt activation in SHP-1^−/− CD8+ T cells compared to SHP-1^+/+ CD8+ T cells, both at baseline as well as following TCR/CD3 stimulation. (Figs. 5 C and D). Taken together, these data suggest that enhanced activation of the PI3K/Akt pathway in the SHP-1^−/− T cells may provide one component of resistance to Treg-mediated suppression, similar to what has been described for T cells isolated from patients with lupus³², multiple sclerosis²⁹, and juvenile idiopathic arthritis^24,33.

Figure 5. SHP-1^−/− T cells exhibit enhanced activation of Akt.

(A) Splenic CD4+CD25- T cells were isolated from SHP-1^+/+ or SHP-1^−/− mice and cultured in the presence of indicated concentrations of anti-CD3 and irradiated CD4+ T cell-depleted splenocytes as APCs. After 24 hours, intracellular levels of phospho-Akt (T308) were assessed by flow cytometry. Histograms represent phospho-Akt (T308) levels within live CD4+Foxp3- SHP-1^+/+ or SHP-1^−/− T cells compared to FMO control. Data represent 3 independent experiments, n=6-9 mice per genotype. (B) Bar graph represents compiled relative increase in phospho-Akt MFI compared to baseline (unstimulated SHP-1^+/+ CD4+ T cells). (C) Splenic naïve (CD44^lo) CD8+ T cells were isolated from SHP-1^+/+ or SHP-1^−/− mice and cultured in the presence of indicated concentrations of anti-CD3 and irradiated CD4+ T cell-depleted splenocytes as APCs. After 24 hours, intracellular levels of phospho-Akt (T308) were assessed by flow cytometry. Histograms represent phospho-Akt (T308) levels within live CD8+ SHP-1^+/+ or SHP-1^−/− T cells compared to FMO control. n=3 mice per genotype. (D) Bar graph represents relative increase in phospho-Akt MFI compared to baseline (unstimulated SHP-1^+/+ CD8+ T cells). A one-column t test with a null hypothesis of 1 was applied to fold change MFI values in B and D, obtained by comparing MFI of SHP-1^−/− cells to the MFI of the SHP-1^+/+ cells at each dose. Error bars indicate ±SEM; * p≤0.05, **p≤0.01,***p≤0.00.1

Tcon cells lacking SHP-1 resist Treg-mediated suppression in vivo

To assess whether SHP-1 regulates the susceptibility to Treg mediated suppression in vivo, we used a murine model of Treg-mediated control of homeostatic expansion³⁶. We intravenously injected SHP-1^+/+ or SHP-1^−/− CD4+CD25− T cells (CD45.2) at a 1:1 ratio with CD45.1 wild type CD4+CD25− T cells into Rag1^−/− mice, with or without wild type Tregs. After 10 days, we assessed the expansion of CD4+Foxp3- T cells, in the spleens of Rag1^−/− recipient mice (Fig. 6A). In the absence of Tregs, we observed no significant differences in the expansion of SHP-1^+/+ or SHP-1^−/− CD4+ T cells, when compared to the percentages (Fig. 6B) and absolute numbers of co-injected wild type CD45.1 CD4+ T cells, suggesting that SHP-1 does not regulate homeostatic expansion. In the presence of Tregs, we observed a substantial reduction in absolute number of T cells recovered, indicating Treg-mediated suppression of homeostatic expansion (Fig. 6C). There was no difference in the extent of suppression between SHP-1^+/+ T cells and wild type CD45.1 T cells co-injected with SHP-1^+/+ or SHP-1^−/− T cells. However, SHP-1^−/− T cells exhibited significantly greater homeostatic expansion (~2.5fold) in the presence of Tregs compared to SHP-1^+/+ T cells or co-injected wild type CD45.1 T cells, indicating a resistance to Treg-mediated suppression (Figs. 6B and C). Taken together, these data strongly suggest that SHP-1 regulates the susceptibility of CD4+ T cells to Treg-mediated suppression in vitro as well as in vivo.

Figure 6. SHP-1^−/− CD4+ T cells resist Treg suppression in vivo.

(A) Schematic representation of experimental setup. Splenic CD4+CD25- Tcon cells were isolated from wild type CD45.1 mice or SHP-1^+/+ or SHP-1^−/− CD45.2 mice and labeled with CellTrace Violet. Wild type Tregs (CD4+CD25+) were isolated from spleens of SHP-1^+/+ mice. 3×10⁶ total Tcon cells were injected i.v. via the tail vein into Rag1^−/− recipient mice, at a 1:1 ratio of either CD45.2 SHP-1^+/+:CD45.1 wild type Tcon cells or CD45.2 SHP-1^−/−:CD45.1 wild type Tcon cells. Half the recipients received Tcon cells only, and the other half received Tcon cells along with 7.5×10⁵ SHP-1^+/+ Tregs (1:4 Treg:Tcon ratio). After 10 days, spleens of recipient mice were harvested and stained for analysis by flow cytometry. (B) (Left) Representative flow plots of CD4+CD25- input Tcon cells. (Top) Input for conditions i and ii: CD45.2 SHP-1^+/+ with CD45.1 wild type Tcon cells; (Bottom) input for conditions iii and iv: CD45.2 SHP-1^−/− with CD45.1 wild type Tcon cells. (Right) Plots show percentages of splenic CD45.1+ and CD45.2+ CD4+Foxp3- T cells recovered 10 days post-injection; experimental conditions (with or without Tregs) as indicated. (C) Percent suppression was computed by subtracting the percent relative expansion for each indicated genotype from 100%. Percent relative expansion was calculated by dividing the absolute number of CD4+Foxp3- T cells recovered in the presence of co-injected Tregs over the absolute number of CD4+Foxp3- T cells recovered in the absence of Tregs (maximal expansion), multiplied by 100. n=3-4 recipient mice per donor condition. Error bars indicate ±SEM; * p≤0.05.

Discussion

For T cells to mount a productive response against a pathogen, they must be able to transiently overcome constraints imposed by Treg cells. Environmental factors as well as strong antigenic signals through the TCR in the presence of co-stimulation have been shown to allow T cells to become refractory to Treg suppression^4–8. However, the intracellular signaling pathways that result in resistance to suppression are not well-defined. Here, we identify the tyrosine phosphatase SHP-1 as one of the intracellular regulators of conventional T cells that influence their susceptibility to Treg suppression. Both SHP-1^−/− CD4+ and CD8+ T cells resisted Treg suppression of proliferation in vitro, and SHP-1^−/− CD4+ T cells resisted Treg suppression of homeostatic expansion in vivo. Moreover, SHP-1^−/− T cells resisted Treg suppression in a T cell-intrinsic manner, as co-culture (Fig. 4) or co-injection (Fig. 6) of SHP-1^+/+ (wild type) and SHP-1^−/− CD4+ T cells could not induce SHP-1^+/+ (wild type) CD4+ T cells to become resistant to suppression. SHP-1^−/− T cells have been reported to produce increased amounts of IL-4 when stimulated in vitro, and SHP-1 additionally negatively regulates the subsequent downstream phosphorylation of STAT6, suggesting that SHP-1^−/− T cells are hyper-responsive to IL-4 signaling¹⁵. Since IL-4 has been shown to induce resistance to Treg suppression in vitro³⁷, it raised the possibility that IL-4 might play a role in mediating the observed resistance to suppression in SHP-1^−/− T cells. However, we found that neither IL-4 neutralizing antibodies nor antibody blockade of IL-4Rα-mediated signaling altered the resistance of SHP-1^−/− T cells to Treg suppression (data not shown), indicating that the resistance reported here is IL-4-independent. This is consistent with a T cell-intrinsic mechanism and likely mediated by alterations in intracellular signaling events.

Previous studies demonstrated that deficiency of Cbl-b³⁸ and TRAF6³¹, two other negative regulators of T cell activation, also resulted in T cells that resist Treg suppression. A recent study suggested that SHP-1 regulates the degradation of Cbl-b, such that SHP-1-deficient T cells have decreased levels of Cbl-b protein after TCR stimulation alone³⁹. While there are striking similarities between SHP-1^−/− and Cbl-b^−/− T cells, our proliferation and suppression assays included costimulatory signals from irradiated APCs, which lead to Cbl-b degradation⁴⁰ in both SHP-1^+/+ and SHP-1^−/− T cells, and therefore would not account for the observed resistance to Treg suppression. Moreover, we did not detect any differences in Cbl-b protein expression between SHP-1^−/− T cells and SHP-1^+/+ T cells (data not shown). We did, however, observe enhanced activation of the Akt pathway in SHP-1^−/− CD4+ T cells and naïve CD8+ T cells, both basally and upon TCR stimulation. The PI3K/Akt pathway is primarily activated downstream of the TCR and CD28 costimulatory signaling, and the resultant signaling cascade allows T cells to proliferate by increasing cell size and glucose metabolism, inactivating cell cycle inhibitors, and enhancing cellular survival⁴¹. An important mechanism of Treg suppression is depriving T cells of costimulatory signals via downregulation of costimulatory molecules CD80/CD86 (B7.1/B7.2) on APCs and upregulation of inhibitory molecules like CTLA-4 and LAG3⁴². CTLA-4 can outcompete CD28 for binding of B7 molecules on APCs, and LAG3 can prevent maturation of APCs to adequately engage T cells⁴³. Previous work suggested that SHP-1-deficient T cells have a reduced requirement for costimulation¹³. Since SHP-1^−/− T cells show enhanced Akt activation upon TCR stimulation, they likely resist Treg suppressive mechanisms that specifically inhibit costimulation, as their need for costimulation is reduced by the enhancement in Akt activation. Interestingly, many of the environmental factors shown to induce suppression-refractory T cells have been linked to enhancing activation of the PI3K/Akt pathway³.

Our work also helps to clarify recent discrepancies reported on SHP-1 function in negative regulation of TCR signaling due to the use of CD4-Cre mediated deletion. Using the distal Lck-Cre line, in which SHP-1 deletion is temporally distinct from early stages of thymic selection, minimized developmental or potential repertoire changes to the T cell compartment. Importantly, dLck-Cre+ SHP-1^f/f mice did not display any detectable differences in the composition of the thymic or peripheral T cell compartments compared to SHP-1-sufficient control mice, nor the expansion of CD4+ memory (CD44^hi) T cells. However, consistent with data published by others and us (reviewed in ⁹), we observed increased responsiveness to TCR stimulation in SHP-1^−/− CD4+ and CD8+ T cells, which was directly attributable to loss of SHP-1 within the T cells rather than an expansion of antigen-experienced T cells.

Aside from gaining insight into the molecular mechanisms of Treg resistance, which has been linked to the pathophysiology of autoimmune diseases, our findings might also be applicable toward tumor immunotherapy. Tumors actively recruit and generate Tregs to maintain a suppressive microenvironment⁴⁴. Thus, the goal of current adoptive cell transfer and/or chimeric antigen receptor (CAR)-T cell therapies is to modify or create CD8+ T cells with enhanced responsiveness toward tumor antigen⁴⁵. Many of the signaling components being incorporated into CAR-T cells are from costimulatory molecules, which have also been found to induce resistance to Treg suppression. For example, both 4-1BB and OX40 signaling in T cells has been found to induce Treg-resistance^46–51, and components of both have been used in second generation CAR-T cells⁵². Along these lines, adoptive transfer of SHP-1^−/− or SHP-1 knockdown (via siRNA) CD8+ T cells improved tumor control in a mouse model of disseminated leukemia⁵³. However, whether CD8+ T cell resistance to Treg suppression played a role in tumor control was not examined. Our findings suggest that incorporating SHP-1 ablation could be useful in current CAR-T cell or adoptive cell transfer therapies to allow CD8+ cytotoxic T lymphocytes to overcome Treg suppression and better control tumor outgrowth. Signaling through many of the costimulatory molecules being used currently in CAR-T cell trials also enhance Akt activation. Directly enforced constitutive Akt activation induced human CD8+ T cells to resist Treg suppression and led to enhanced cytotoxicity toward a neuroblastoma cell line⁵⁴. Not only are these findings translatable to tumor immunotherapy, but also for treatment of chronic viral infections. It has been shown that chronic viral infection induces Tregs to suppress the function of CD8+ T cells, preventing viral clearance⁵⁵. Stimulation of CD8+ T cells with the costimulatory molecule 4-1BB rendered T cells resistant to Treg suppression and able to clear a chronic viral infection in mice⁴⁸. Therefore, our data reveal SHP-1 as a possible target to modulate the activation and function of T cells for tumor and chronic viral infection immunotherapies, and provide more evidence pointing to the critical nature of the PI3K/Akt pathway in regulating the balance between T cells and Treg cells.