Early T-bet promotes LFA1 upregulation required for CD8⁺ effector and memory T cell development

Understanding the transcriptional regulation of polarized T cell responses and their link to disease outcomes remains a major theme in immunology. The identification of T-bet, GATA3, and RORγt as lineage-specifying transcription factors that influence the development of CD4 T helper (Th) 1-, Th2-, and Th17-type responses, respectively, has provided key milestones in understanding T cell function and plasticity (Szabo et al., 2002; Ivanov et al., 2006; Zheng and Flavell, 1997). Consistent with a key role in cell-mediated immunity and CD4⁺ T cell production of IFN-γ, T-bet has been shown to be essential for protective T cell responses to many intracellular pathogens as well as having a pathological role in other diseases (Lazarevic and Glimcher, 2011; Pritchard et al., 2019). However, in CD8⁺ T cells, there are T-bet–independent pathways to IFN-γ, mediated in part through the expression of a related T-box transcription factor, Eomesodermin (Eomes; Intlekofer et al., 2008; Pearce et al., 2003). These findings have led to the idea that in an infectious setting, T-bet and Eomes are functionally redundant in CD8⁺ T cells (Lazarevic and Glimcher, 2011), but these transcription factors do have different mechanisms of induction, unique functions, and even antagonistic activities (Buggert et al., 2014; Paley et al., 2012). Indeed, while less studied, T-bet also has a role in T cell trafficking and chemokine production, findings that highlight a broader role for T-bet in T cell biology (Harms Pritchard et al., 2015; Kao et al., 2011; Koch et al., 2009; Lord et al., 2005; Oakley et al., 2013). In current models, naive CD4⁺ and CD8⁺ T cells do not express T-bet, but activation of CD4⁺ T cells through the TCR induces an early transient wave of T-bet expression (Schulz et al., 2009; Szabo et al., 2000), and in the context of sustained TCR engagement, signals from IFN-γ and IL-12 result in a second prolonged phase of T-bet expression. While the second wave of T-bet is associated with the acquisition of effector functions (Afkarian et al., 2002; Schulz et al., 2009), the role of early T-bet expression is uncertain. Similarly, T-bet is expressed in activated CD8⁺ T cells prior to the first round of cell division (Chang et al., 2011), and the significance of this prompt T-bet expression, perhaps prior to fate determination, may be distinct from its function during the effector phase.

Many of the studies to understand the function of T-bet in the context of infection have focused on its role during the effector phase of the immune response (Harms Pritchard et al., 2015; Oakley et al., 2013; Ravindran et al., 2005; Sullivan et al., 2005; Szabo et al., 2002; Way and Wilson, 2004). For example, in germline T-bet knockout (T-bet^−/−) mice infected with Toxoplasma gondii, the generation, expansion of parasite-specific T cells, and their production of IFN-γ appear largely intact, but the effector T cells have reduced expression of CXC motif chemokine receptor 3 (CXCR3) and lymphocyte function-associated antigen 1 (LFA1), which is associated with decreased T cell trafficking to secondary sites of infection and results in a failure to control parasite growth (Harms Pritchard et al., 2015). While these types of studies illustrate the importance of T-bet for T cell–dependent control of infections, the increased susceptibility of the T-bet^−/− mice to primary challenges makes it difficult to assess how T-bet influences memory formation or its role in a secondary response. Moreover, in some infected T-bet deficient mice, the elevated antigen loads and systemic inflammation that promote T cell activation confound the interpretation that T cell responses appear intact (Harms Pritchard et al., 2015; Paley et al., 2012; Ravindran et al., 2005; Svensson et al., 2005). Thus, in mixed chimeras in which WT and T-bet^−/− T cells are both present, the WT T cells dominate, but it is unclear whether T-bet^−/− T cells are at a competitive disadvantage or whether their inability to access sites of infection results in reduced antigen stimulation that compromises their expansion (Harms Pritchard et al., 2015). One strategy to overcome these limitations is the use of mutant strains of T. gondii or irradiated parasites that are unable to replicate in vivo and that do not generate systemic inflammation but induce protective pathogen-specific effector and memory CD4⁺ and CD8⁺ T cell responses (Suzuki and Remington, 1988; Gazzinelli et al., 1991; Denkers et al., 1993; Sukhumavasi et al., 2008; Wilson et al., 2010; Dupont et al., 2014; Shah et al., 2015; Chu et al., 2016). The data presented here provide evidence that the early activation of T-bet has an unanticipated cell-intrinsic function in CD8⁺ T cells that promotes the increased expression of LFA1 that is required for priming at early time points necessary for normal expansion of pathogen-specific T cells.

Vaccination with a replication-deficient strain of T. gondii, CPS, results in the development of a strong effector and memory T cell response that protects from challenges with the lethal RH strain (Fox and Bzik, 2002; Jordan et al., 2009). Indeed, when WT or T-bet^−/− mice were injected intraperitoneally with CPS and challenged via the same route with the RH strain 12 d later, WT mice controlled parasite replication whereas T-bet^−/− mice developed a high parasite burden and succumbed to infection (Fig. 1 A, data not shown). Our previous studies showed that T-bet^−/− mice were challenged intraperitoneally (i.p.) with a replication-competent strain of T. gondii–generated parasite-specific effector or memory precursor CD8⁺ T cell populations and could control parasite replication in the peritoneum. Therefore, the failure of the vaccinated T-bet^−/− mice to control parasite replication in the peritoneum was unexpected. To understand the basis for this phenotype, WT and T-bet^−/− mice were immunized with CPS, and 9–11 d later an MHC Class I tetramer was used to identify, enumerate, and phenotype parasite-specific CD8⁺ T cells in the spleen. A distinct population of parasite-specific CD8⁺ T cells was detected in WT and T-bet^−/− mice at this time point (Fig. 1 B). However, in contrast to what was observed with replication-sufficient parasites (Harms Pritchard et al., 2015), this population was reduced in the T-bet^−/− mice by frequency and absolute number (Fig. 1 B). Recent studies have utilized the expression of LFA1, killer cell lectin-like receptor G1 (KLRG1), and CXCR3 as phenotypic markers to identify effector (T_eff), memory (T_mem), and intermediate (T_int) effector CD8⁺ T cell populations after infection with T. gondii (Chu et al., 2016; Shah et al., 2015). In WT mice, parasite-specific CD8⁺ T cells are characterized by the upregulation of LFA1 (Fig. 1 C), and based on the nomenclature proposed by Robey and colleagues (Chu et al., 2016), multiple subpopulations of parasite-specific CD8⁺ T cells could be detected: CXCR3^-KLRG1⁺ T_eff, CXCR3⁺KLRG1⁺ T_int, and CXCR3⁺KLRG1⁻ T_mem cells (Fig. 1 D). However, in the T-bet^−/− mice, the parasite-specific CD8⁺ T cells were characterized by lower levels of LFA1 (Fig. 1 C), and the subsets of CXCR3^-KLRG1⁺ T_eff and the CXCR3⁺KLRG1⁺ T_int were significantly reduced by both frequency and number (Fig. 1 D). Although CXCR3 is considered a T-bet target, the CXCR3⁺KLRG1⁻ T_mem population was intact in the T-bet^−/− mice (Fig. 1 D). In WT mice infected with T. gondii, the CXCR3⁺KLRG1⁺ T_int population gives rise to the T_eff population (Chu et al., 2016), and so the absence of this population in the T-bet^−/− mice may underlie the reduced CD8⁺ effector T cell population after vaccination.

To determine whether T-bet was intrinsically required in CD8⁺ T cells for their optimal expansion and differentiation into T_eff, WT and T-bet^−/− TCR transgenic OT-I CD8⁺ T cells (WT and T-bet^−/− OT-I T cells) were adoptively transferred into WT congenic mice that were then immunized with an ovalbumin-expressing strain of CPS (CPS-Ova). In this system, the WT OT-I T cells expanded and were KLRG1^hi and LFA1^hi (Fig. 1, E and F). However, similar to the endogenous parasite-specific population, there was a reduced expansion of T-bet^−/− OT-I T cells that was accompanied by the limited acquisition of KLRG1 expression by the small population of T-bet^−/− OT-I T cells that were present. These cells also expressed lower levels of LFA1 (Fig. 1, E and F). The reduced expansion of antigen-specific CD8⁺ T cells following vaccination in T-bet^−/− mice and reduced differentiation into T_eff CD8⁺ T cells corresponded with a reduced proportion and the number of T. gondii–specific CD8⁺ T cells in the spleens of immunized T-bet^−/− mice at memory time points (>day 30 after immunization; Fig. S1 A). These studies indicate that T-bet^−/− mice vaccinated with a replication-deficient strain of T. gondii can develop a T_mem KLRG1^- population, but this is not sufficient for protection to a secondary challenge. Thus, T-bet is intrinsically required for parasite-specific CD8⁺ T cells to expand and acquire a T_eff phenotype.

To distinguish whether T-bet is required during initial CD8⁺ T cell activation or only during the effector phase to generate T_int or T_eff populations, studies were first performed to characterize where initial priming of CD8⁺ T cells occurred after CPS immunization. Because the omentum is the site of T cell priming after i.p. injection of antigens (Carlow et al., 2009; Rangel-Moreno et al., 2009; Christian et al., 2020), we examined the T cell response in this tissue using CD69 as a marker of TCR engagement combined with Ki67 as a marker of cells entering cell cycle (Eickhoff et al., 2015; Feng et al., 2005; Gerner et al., 2017; Testi et al., 1989). In unimmunized mice, 5–10% of CD8⁺ T cells in the omentum displayed low basal levels of CD69 (Fig. 2 A). By 48 h after immunization, two distinct populations (CD69⁺Ki67⁻ and CD69⁺Ki67⁺) of activated CD8⁺ T cells could be distinguished in the omentum (Fig. 2 A). To distinguish between antigen-induced and bystander activation of the CD8⁺ T cells, purified OT-I T cells were adoptively transferred into congenic hosts that were then immunized with CPS or CPS-Ova. Following CPS, the OT-I T cells showed a modest increase in CD69 expression, but in response to CPS-Ova there was a CD69^hi population (Fig. 2 B). Analysis of bulk CD8⁺ T cells stimulated with cytokines alone or through the TCR highlights that while IL-12 alone could promote a small (4–5%) bystander population of T cells that express CD69 and Ki67, TCR stimulation resulted in >75% of these cells expressing CD69 and Ki67 (Fig. S1 B). These data indicate that in this experimental system, the majority of CD69 upregulation is not due to bystander activation or cytokine signaling. Furthermore, WT OT-I T cells that were transferred into naive mice that were not subsequently challenged with CPS-Ova did not express Ki67 and had lower levels of CD69 compared with WT OT-I T cells that were transferred into naive mice that were challenged with CPS-Ova (Fig. 2 C). Thus, similar to other studies (Eickhoff et al., 2015; Gerner et al., 2017), the majority of CD69 upregulation and entry into the cell cycle is dependent on TCR engagement. When WT OT-I T cells labeled with CFSE were transferred into mice challenged with CPS-Ova and then examined 2 d later, the CD69⁺Ki67⁻ OT-I T cells did not dilute CFSE (Fig. 2 D), whereas the CD69^hiKi67⁺ had undergone multiple rounds of cell division (Fig. 2 D) and the CD69⁻Ki67⁺ population had proliferated most extensively (Fig. 2 D). Together, these data indicate that CPS challenge leads to TCR-dependent T cell priming in the omentum, and as T cells divide, they downregulate CD69 (a mediator of tissue retention; Shiow et al., 2006), which would be a prerequisite to exit the site of priming. To assess the impact of T-bet loss on the behavior of CD8⁺ T cells in the omentum early after immunization, we performed intravital imaging of WT and T-bet^−/− OT-I T cells in the same host omentum 2 d after immunization. 48 h after immunization, the WT OT-I T cells showed an increase in mean speed and a concurrent decrease in arrest coefficient, consistent with successful priming and increased movement away from APCs following activation. In the same setting, T-bet^−/− OT-I T cells had a significantly reduced mean speed and a significantly increased arrest coefficient, suggesting that priming or activation is altered in the absence of T-bet (Fig. S1, C–E). Together, these data sets identify a cell-intrinsic role for T-bet in the events involved in early CD8⁺ T cell priming.

Next, experiments were performed to determine whether this early CPS-induced T cell activation was associated with increased expression of T-bet or LFA1, a T-bet-dependent target associated with T cell activation. Compared with the “naive” CD69⁻Ki67⁻ population, both T-bet and LFA1 were upregulated in each of the activated populations and were most highly expressed in the cells that were CD69^hiKi67⁺ (Fig. 3 A). However, as these populations downregulated CD69, levels of LFA1 and T-bet were reduced (Fig. 3 A). Because T-bet localization in the nucleus and cytoplasm can vary with activation state (Chang et al., 2011; Harms Pritchard et al., 2015; McLane et al., 2013), CD8⁺ T cells isolated from the omenta of mice 2 d after CPS immunization were analyzed using an Amnis ImageStream to visualize T-bet localization. Interestingly, in each of the T-bet expressing populations, T-bet was present in the nucleus and cytoplasm; however, the CD69⁺Ki67⁺ CD8⁺ T cell population had the highest frequency of cells with exclusively nuclear T-bet (Fig. 3 B). Thus, in vivo early TCR ligation is associated with the upregulation of T-bet and LFA1 and is reminiscent of early kinetics of T-bet expression observed in vitro using CD4⁺ T cells (Schulz et al., 2009).

To determine if the CPS-induced expression of T-bet impacts early T cell activation or expansion, WT and T-bet^−/− mice were immunized with CPS, and the CD8⁺ T cell responses in the omenta were analyzed 1–2 d later. In WT and T-bet^−/− mice, there was no difference in the expression of CD69 or Ki67 basally (Fig. 4 A) or 1 d after immunization (Fig. 4 B). However, at 2 d after immunization, WT CD8⁺ T cells continued to upregulate CD69, whereas for the T-bet^−/− CD8⁺ T cells this was deficient. Additionally, in WT mice the appearance of a CD69^hiKi67⁺ population of CD8⁺ T cells was apparent as early as day 2 after immunization, yet this was reduced in the T-bet^−/− mice (Fig. 4 C). Moreover, when 1 × 10⁶ WT and T-bet deficient OT-I T cells were transferred into hosts that were then immunized with CPS-Ova, by day 2 the majority of WT OT-I cells in the omentum coexpressed CD69 and Ki67, whereas this population was diminished among T-bet–deficient OT-I cells (Fig. 4 D). In a co-transfer system, the ability to directly compare the kinetics of WT and T-bet^−/− OT-I T cells in a T-bet sufficient environment at 24 and 48 h after vaccination highlighted that the absence of T-bet resulted in reduced levels of CD11a (the α-subunit of LFA1; Fig. 2, A and B). This reduced expression of CD11a was associated with decreased proliferation 48 h after immunization when compared with the proportion of WT versus T-bet^−/− OT-I T cells recovered from the omentum and the percent that diluted CellTraceViolet (Fig. 2 C). Our previous studies found that in T-bet^−/− mice infected with a replicating strain of T. gondii, the magnitude of the T cell response at day 10 was largely intact (Harms Pritchard et al., 2015), but analysis at day 2 after infection revealed that the T-bet^−/− mice also had an early defect in T cell priming (Fig. 2 D). Thus, regardless of vaccination or infection, T-bet is not required for initial T cell activation (based on upregulation of CD69) but does promote progression into the cell cycle.

During T cell priming, the integrin LFA1 is part of the peripheral supramolecular activation cluster and has a key role in the T–dendritic cell (DC) interactions required for T cell activation (Dustin and Springer, 1989; Gérard et al., 2013). Because early CD8⁺ T cell expression of T-bet correlated with the upregulation of LFA1 (Fig. 3), experiments were performed to determine if the absence of T-bet affected early upregulation of LFA1 during T cell priming. Naive T cells rapidly expose an intracellular pool of LFA-1 following activation to facilitate interactions necessary for activation (Capece et al., 2017). To determine if T-bet impacted this intracellular pool, we analyzed levels of CD11a mRNA, protein expression, and localization in WT and T-bet^−/− CD8⁺ T cells (Fig. 2, E–H). No defect in CD11a mRNA, protein levels, or localization was observed in naive T-bet^−/− CD8⁺ T cells corresponding with the lack of T-bet expression in CD8⁺ T cells prior to activation. However, at day 2 after CPS immunization, the CD69⁺ CD8⁺ T cells from WT mice could be divided into two populations (CD69^int and CD69^hi; Fig. 5 A), and the highest levels of LFA1 were apparent in the CD69^hi population (Fig. 5 B). In contrast, the CD69⁺CD8⁺ T cells from the T-bet^−/− mice were almost exclusively CD69^int and the CD69^hiLFA1^hi population was largely absent (Fig. 5 A). Furthermore, although naive WT and T-bet^−/− CD8⁺ T cells expressed similar basal levels of LFA1, activation of the WT T cells with αCD3 resulted in an upregulation of LFA1 that was reduced in the absence of T-bet (Fig. S2 H). When the ability of these populations of activated T cells to adhere to ICAM1 (the binding partner of LFA1) was tested, the activated T-bet^−/− T cells showed decreased binding to ICAM1-coated surfaces (Fig. 5 C).

Next, ATAC-seq was performed to determine whether TCR activation led to an association of T-bet with the itgal locus. Analysis of naive and activated T cells revealed two distinct regions of open chromatin 5′ of the itgal locus. One region, previously identified by Kaech and colleagues (Dominguez et al., 2015) as a T-bet binding site at the itgal transcriptional start site (TSS), was open in naive and activated T cells (Fig. 5 D). In addition, a new region was identified 5′ to the TSS that was inaccessible in naive cells but accessible after activation. To determine if T-bet was associated with either of these regions, we performed anti-T-bet chromatin immunoprecipitation assay (ChIP) followed by qRT-PCR using region-specific primers. PCR specific to the newly identified region revealed substantial enrichment after T-bet IP early (day 2), but not late (day 6), after T cell activation (Fig. 5 E). Surprisingly, despite the open nature of the region at the TSS even in naive T cells, T-bet association with this region was only enriched at a later time point after activation (Fig. 5 F). Together, these data indicate that T-bet associates with the itgal locus early after T cell activation and at a previously unidentified site 5′ of the TSS that becomes accessible only after T cell activation.

To determine whether LFA1 is important for the CPS-induced activation of CD8⁺ T cells in vivo, WT mice were treated with an isotype control or a blocking antibody against CD11a prior to CPS injection. At day 2 after immunization, the frequency of CD69⁺Ki67⁻ CD8⁺ T cells was equal between the control and the treatment groups (Fig. 5 G). However, the blockade of CD11a resulted in a significant reduction in the frequency of cells that co-expressed CD69 and Ki67 (Fig. 5 G). Furthermore, this early CD11a blockade resulted in a reduction in parasite-specific CD8⁺ T cells at 10 d after immunization (Fig. 5 H), and this population expressed reduced levels of KLRG1 (Fig. 5 I). Similar effects of CD11a blockade were seen when mice were reconstituted with a low number of OT-I T cells and immunized with CPS-Ova (Fig. 5, J and K). Importantly, when CD11a blockade was performed on day 5 and the polyclonal T cell response was assessed on day 10, the T. gondii–specific CD8⁺ T cell response in the peritoneum appeared normal, while the antigen-specific CD8⁺ T cell response in the spleen was partially restored (Fig. S3). Thus, early CD11a blockade during this vaccination does not impact the initial upregulation of CD69, but it does affect entry into the cell cycle and the expansion of CD8⁺ effector T cells. These data sets need to be interpreted with care, but the decreased CD8⁺ T cell response and reduced expression of KLRG1 observed with CD11a blockade or in the absence of T-bet suggest that lower LFA1 expression on the T-bet^−/− CD8⁺ T cells would contribute to the reduced expansion and absence of effector populations.

In several models of infection, the absence of T-bet does not affect the overall magnitude of the CD8⁺ T cell responses (Harms Pritchard et al., 2015; Intlekofer et al., 2007; Way and Wilson, 2004). However, the absence of T-bet during infection with LCMV or T. gondii does result in an altered phenotype of pathogen-specific CD8⁺ T cells characterized by reduced expression of LFA1, Ly6C, and KLRG1 (Harms Pritchard et al., 2015; Intlekofer et al., 2007; Intlekofer et al., 2008). In those previous studies, the impact of T-bet on early events involved in T cell activation was not examined but the data presented here indicate that a role for T-bet in priming is not restricted to vaccination. A likely explanation for this apparent disparity for T-bet in T cell priming but not expansion following infection is that during acute infection it is associated with inflammation and an increasing antigen load, which provides an environment that overrides the impact of T-bet on T cell priming.

There are multiple variables that influence the earliest T–DC interactions required to initiate T cell activation, which affect the magnitude of the effector and memory pools. Relevant T cell–extrinsic factors include local chemokine production (Castellino et al., 2006) and the ability of DCs to provide costimulation and cytokine signals (Benvenuti, 2016; Chen and Flies, 2013). MHC–peptide affinity also impacts T cell activation, levels of proliferation, and migration kinetics within secondary lymphoid organs (Zehn et al., 2009). Other studies have demonstrated that MHC–peptide affinity also regulates intranodal localization of proliferating T cells in a CXCR3-dependent manner (Ozga et al., 2016), and activated CD8⁺ T cells are able to recruit additional DC subsets to create an “optimal priming microenvironment” (Brewitz et al., 2017). Interestingly, while T-bet can promote the T cell expression of CXCR3, multiple studies have demonstrated T-bet–independent CXCR3 expression (Beima et al., 2006; Dominguez et al., 2015; Harms Pritchard et al., 2015; Intlekofer et al., 2008; Zhu et al., 2010). Less is known about the T cell–intrinsic factors that influence these events, but LFA1–ICAM1 interactions are essential for the formation of the immunological synapse and sustained cellular adhesion needed for optimal T cell priming (Gérard et al., 2013; Kandula and Abraham, 2004). Studies by Krummel and colleagues using a Listeria-OVA system have highlighted that primary T–DC and secondary homotypic T–T clustering synaptic interactions are important for effector CD8⁺ T cell responses, and that LFA1-blockade during the “critical differentiation period” leads to a lower frequency of KLRG1^hi short-lived effector cells and a higher frequency of KLRG1^lo memory precursor effector cells (Gérard et al., 2013). This loss of KLRG1^hi T_eff cells is similar to the phenotype observed here when CPS parasites are used to immunize T-bet^−/− mice or when LFA1 is blocked, as well as in CD11a knockout mice challenged with Listeria or Plasmodium (Bose et al., 2013; McNamara et al., 2017). These observations motivate the need to understand how initial T–DC interactions and T-bet impact the ability of CD8⁺ T cells to alter their metabolism and generate building blocks that are required for the proliferative burst, which is a hallmark of activated CD8⁺ T cells (Buck et al., 2015; Christian et al., 2020).

Despite the interest in understanding the events that allow the development of protective effector and memory CD8⁺ T cell responses (Harty and Badovinac, 2008; Im et al., 2016; Utzschneider et al., 2016), there is a paucity of studies on the role of T-bet in the setting of vaccination (Pritchard et al., 2019). While the data presented here identify an essential role for T-bet in the ability to generate a long-lived memory response and resistance to rechallenge, the finding that this transcription factor influences early events required for CD8⁺ T cell activation, entry into the cell cycle, and the expansion of effector cells was unanticipated. These findings also appear relevant to the early CD8⁺ T cell responses following immunization strategies that use adjuvants comprised of defined TLR agonists, singly or in combination with other molecular adjuvants (Klarquist et al., 2018). Studies from Reiner and colleagues (Lin et al., 2016) showed that in mice infected with Listeria, the amplifying population of effector-like CD8⁺ T cells (analogous to the CXCR3⁺KLRG1⁺ T_int population induced in response to T. gondii) divided more and had elevated levels of T-bet while “quiescent” cells (analogous to the CXCR3⁺KLRG1^- T_mem population shown here) divided less and had lower levels of T-bet. Since T-bet contributes to the generation of the T_int population, this lack of T_int cells would likely explain the reduced population of CXCR3^-KLRG1⁺ T_eff populations observed in the immunized T-bet^−/− mice. The T-bet–dependent alteration of CD8⁺ T cell behavior within 2 d after immunization suggests that T-bet impacts a specific behavioral program, such as cell–cell interactions or T cell motility during priming, likely through LFA1-dependent interactions that promote the ability of T cells to adhere to APCs or other T cells. Nevertheless, it is likely that T-bet will affect additional pathways that influence CD8⁺ T cell activation and differentiation, and understanding these activities of T-bet should inform the development of vaccination strategies that promote the generation of protective memory CD8⁺ T cell responses.

T-bet–deficient (T-bet^−/−) mice were purchased from the Jackson Laboratory. WT C57BL/6 mice were purchased from Taconic. CD45.1⁺C57BL/6 mice were purchased from Charles River. All mice were housed in a specific pathogen–free environment at the University of Pennsylvania School of Veterinary Medicine in accordance with the federal guidelines and with the approval of the Institutional Animal Care and Use Committee. CPS-Ova parasites were derived from the RHΔcpsII (CPS) clone, which was provided as a generous gift by Dr. David Bzik (Geisel School of Medicine at Dartmouth, Lebanon, NH; Fox and Bzik, 2002). Parasites were cultured and maintained by serial passage on human foreskin fibroblast cells in the presence of a parasite culture media (DMEM [Invitrogen], 20% medium M199 [Invitrogen], 10% FBS [Serum Source International], 1% penicillin–streptomycin [Invitrogen], and 25 μg/ml gentamycin [Gibco]) which was supplemented with uracil (Sigma-Aldrich; final concentration of 0.2 mM uracil). Tachyzoites of each strain were prepared for infection by serial needle passage and filtered through a 5-μm-pore-size filter. Mice were infected i.p. with 10⁵ live parasites.

To quantify parasite burden in the peritoneal exudate, 100,000 cells were used to prepare cytospins. Cells were methanol fixed and then stained with the Protocol Hema-3 Stain Set, and the ratio of infected cells to total cells in a field of vision was calculated. All images were obtained on a Nikon E600 microscope using a 40× objective and NIS Elements Imaging software.

For antibody treatment experiments, mice were treated with 150 μg αCD11a (M17/4; BioXcell) or isotype control (BioXcell) at the time of immunization.

Splenocytes were obtained by grinding spleens over a 70-µm filter (VWR) and washing them in complete media. Red blood cells were then lysed by incubating for 5 min at room temperature in 5 ml of lysis buffer (0.864% ammonium chloride [Sigma-Aldrich] diluted in sterile de-ionized H₂O), followed by washing with complete media. Cells from the omentum were obtained by incubating omenta for 1 h in 0.2 U/ml liberase TL (Roche) diluted in RPMI at 37°C, followed by grinding over a 70-µm filter (VWR) and washing in complete media.

The following antibodies were purchased from BD: Ki67 AF700, CD4 Pacific Blue, CD8α PE-CF594, CD3 PE-CF594, CD4 BV650, and CD11a BUV805. The following antibodies were purchased from eBioscience: KLRG-1 FITC, Ly6C (clone HK1.4) PerCP-Cy5.5, CD69 eFluor450, CD69 FITC, CD69 PE-Cy7, T-bet eFluor660, CD45.2 APC-eFluor780, and CD3 APC-eFluor780. The following antibodies were purchased from BioLegend: LFA-1 PerCP-Cy5.5, CD3 Pacific Blue, CD3 BV785, CD11a PE, and CD11a PerCP-Cy5.5. Invitrogen live/dead Aqua stain or Tonbo Ghostdye510 was used to determine viability. PE-conjugated Tgd-057 MHC-I tetramers and PE-conjugated AS-15 MHC-II tetramers were provided by NIH Tetramer Core. All samples were run on an LSRFortessa (BD), Canto (BD), or FACSymphony A3 and analyzed using FlowJo software (Tree Star). Images were obtained using the ImageStream (Amnis) and analyzed using IDEAS software (Amnis). To determine T-bet localization, nuclear and cytoplasmic masking functions were made using DAPI staining; these masks were then applied to T-bet expression.

In vitro–activated WT and T-bet^−/− CD8⁺ T cells were labeled with cell tracker green (CMFDA) or cell tracker red (CMRA; Life Technologies), and equal numbers of each cell type were mixed and plated on ICAM-1 (R&D Systems) coated surfaces for 20 min. Cells were then fixed and imaged. For each condition, three random fields were taken in triplicate wells. The average number of T cells per well is displayed.

Primary CD8⁺ (OT-I) T cells were isolated from the spleen using the MojoSort Mouse CD8⁺ T Cell Isolation Kit (Biolegend) according to the kit protocol. Antigen-presenting cells were isolated from the spleens of C57BL/6J mice via negative selection using an antibody cocktail consisting of αTer119-biotin, αCD8-biotin, and αCD4-biotin (Tonbo Biosciences, Biolegend, and Biolegend, respectively) followed by magnetic separation using MojoSort Streptavidin Nanobeads (Biolegend). Following isolation, the APCs were peptide-pulsed with 6 μg/ml of SIINFEKL peptide for 2 h at 37°C. After the peptide pulse, the cells were washed extensively with PBS and resuspended in RPMI + 10% FBS complete T cell media. The cells were then cocultured in RPMI + 10% FBS complete T cell media at a ratio of 2:1 CD8⁺ T cells (OT-I) to APCs for 24 h (day 2) or 144 h (day 6). For the day 6 samples, the cells were split every 48 h until harvest.

The ChIP assay was performed using the SimpleChIP “R” + Sonication Kit (Cell Signaling Technologies) according to the manufacturer’s protocol. The precipitated DNA was then subject to qPCR analysis using primers listed below that were designed for the promoter region of Itgal or primers designed for an upstream putative enhancer region. To calculate fold enrichment, the mock IgG IP Ct was subtracted from T-bet IP Ct to control for nonspecific binding, followed by the ΔΔCt method (2^−ΔΔCt). Primers used are as follows: Itgal promoter region (IDT): forward: 5′-TGC​TGT​CTG​CTC​TTG​GAT​GAT-3′; reverse: 3′-CCG​GGT​TGT​GAA​ACC​TCT​CT-5′. Itgal putative enhancer region (IDT): forward: 5′-AAA​CTT​GAA​CTC​CTC​AGC​AGC-3′; reverse: 3′-AGC​CTG​AAG​AGA​TGA​AGC​GTG-5′. For day 2 samples, two mice were pooled for each n with n = 2. For day 6 samples, two mice were pooled for n = 1.

Mice were anesthetized with isofluorane and maintained at a core temperature of 37°C. The peritoneum was opened minimally to access omentum. A small portion of visceral adipose tissue was resected to assist in immobilizing the omentum via a soft-tissue vacuum apparatus (VueBio). No greater than 40 kPa of pressure was applied to steady the omentum for imaging. Image acquisition was performed on a Leica SP8 multiphoton confocal with a 20× water immersion objective (1.0 NA) with a resonant scanner (8,000 kHz) and four external HyD detectors. The excitation wavelength of Chameleon Vision II Ti:Sapphire laser was tuned for optimal detection of CellTrace dye–labeled CD8⁺ T cells in each experiment, typically 880 nm. Images collected x-, y-, and z-plane data over time with a step size of 2 μm and a z-thickness that allowed for a complete z-series every 22 s. This was carried out for ∼30 min for each region imaged. The resulting images were segmented in Imaris (v9.7.2; Bitplane) with spot-specific position data exported and analyzed in R with the CellTrackR package (Wortel et al., 2021).

Statistical analyses were performed using PRISM software (Graphpad Software). Significance was calculated using an unpaired two-tailed Student’s t test except when otherwise noted.

Fig. S1 contains data regarding the impact of T-bet loss on antigen-specific memory CD8⁺ T cell development following immunization with CPS, the regulation of CD69 and Ki67 expression by T cell activation and cytokines, and summarized analysis of WT and T-bet KO CD8⁺ T cell behavior before and after immunization. Fig. S2 consists of data from experiments comparing the expression of CD11a and the proliferation of WT and T-bet KO OT-I T cells in an adoptive transfer system, WT and T-bet KO cells following infection with T. gondii, examination of intracellular and surface expression of CD11a on WT and T-bet KO CD8⁺ T cells, and T-bet’s impact on CD11a expression over time following immunization. Fig. S3 includes data from experiments inhibiting CD11a activity 5 d after immunization and assessing the impact on the proliferation of antigen-specific CD8⁺ T cells.

The authors thank Janis Burkhardt and Steven Reiner for helpful discussions. We would also like to thank the animal care staff.

This study was supported by NIH grants R01AI125563, RO1 AI1268, 1R21AI126042-01, AI126899-01 and AI-42334 (C.A. Hunter), AI-007532 (G.H. Pritchard), and T32 TCA009140 (N.H. Roy). N.H. Roy and A.T. Phan are Cancer Research Institute Irvington Fellows supported by the Cancer Research Institute, and A.T. Phan is also a Robertson Foundation Fellow supported by the Robertson Foundation.

Author contributions: Conceptualization: G.H. Pritchard, A.T. Phan, and C.A. Hunter; Investigation: G.H. Pritchard, A.T. Phan, D.A. Christian, T.J. Blain, Q. Fang, J. Johnson, N.H. Roy, and L. Shallberg; Formal analysis: G.H. Pritchard, A.T. Phan, D.A. Christian, N.H. Roy, and T.J. Blain; Writing—original draft: G.H. Pritchard, A.T. Phan, and C.A. Hunter; Writing—review and editing: G.H. Pritchard, A.T. Phan, D.A. Christian, R.M. Kedl, and C.A. Hunter; Supervision: R.M. Kedl and C.A. Hunter.