Cortactin is expressed in human and murine T cells and interacts with the CD3 complex and F-actin at the immunological synapse
Given the controversial results regarding cortactin expression in different T cells [21, 23, 24], we first analyzed cortactin expression in different human and murine T cells. In resting human T cells, we detected low expression of the 70 kDa SV1 isoform of cortactin (Fig. 1A). By contrast, the human T-ALL cell lines Jurkat and Molt-3 expressed high amounts of cortactin, whereas CEM expressed low levels. Of note, HS1 expression in all leukemic T cell lines was significantly reduced in comparison to normal T cells. T cells derived from Cttn+/+ mice also expressed the 70 kDa cortactin isoform that was absent in T cells derived from Cttn−/− mice (Fig. 1B). By contrast, HS1 expression was similar in both Cttn+/+ and Cttn−/− CD3+ cells.
We next analyzed whether cortactin is recruited to the IS. We found that cortactin is polarized to the IS formed by Jurkat and Raji cells, in colocalization with F-actin and CD3 (Fig. 1C). Similar results were obtained by analyzing the IS between primary mouse CD3+ T cells and B-cell lymphoma A20 cells (Fig. 1D) and in primary human CD3+ T cells and monocytes (Figure S1A). Since APCs also express cortactin [24–26], we asked whether cortactin is specifically recruited to the IS in T cells. Jurkat cells incubated with anti-CD3/28 beads were able to bind multiple beads, with the contact zones being enriched in F-actin, CD3 and CTTN (Fig. 1E), while primary human CD3+ T cells in addition to the contact zone formed a cortactin- and F-actin-enriched zone at the opposite pole, i.e., the uropod (Figure S1B). Therefore, our results prove that cortactin is expressed in human and mouse T cells, where it is recruited to the contact point during IS formation together with the TCR complex and actin filaments. Thus, cortactin seems to be an important regulator of T cell activation potentially by controlling actin cytoskeleton dynamics.
Cttn−/− T cells display impaired F-actin dynamics at the IS
Actin cytoskeleton dynamics are crucial for several T cell functions including signaling, differentiation, and migration [27]. Considering that cortactin localizes at sites of dynamic actin assembly such as the IS [21], we next analyzed whether cortactin is required for the formation of the IS. To answer this, we induced IS formation by allogeneic stimulation of T cells from Cttn+/+ or Cttn−/− mice (H-2kb haplotype) using A20 cells (H-2kd haplotype) and found that Cttn−/− T cells formed significantly less conjugates with A20 cells via an IS than Cttn+/+ T cells (Fig. 2A). Given that IS formation depends on LFA-1 binding to ICAM-1 on APCs, we next evaluated whether the reduced numbers of IS-forming Cttn−/− T cells were due to impaired TCR-mediated LFA-1 activation. However, we observed that similar numbers of Cttn+/+ and Cttn−/− T cells adhered to ICAM-1 in response to PMA or anti-CD3 stimulation (Figure S2A). Additionally, we discarded that IS formation was impaired due to aberrant expression of different molecules needed for proper IS formation in resting and activated cells with no differences in the levels of LFA-1, the VLA-4 subunit CD49d and CXCR4 in Cttn+/+ and Cttn−/− CD3+ T cells (Figure S2B-C). Given that cortactin tunes actin dynamics, we next analyzed the quality of the IS by measuring F-actin at the site of contact. First, we observed that Cttn+/+ and Cttn−/− CD3+ cells displayed similar amounts and distribution of F-actin in the cell periphery under basal conditions (Fig. 2B). Of note, the flat interface at the site of cell-cell contact, characteristic of a normal IS, was observed only at the IS of Cttn+/+ but not Cttn−/− CD3+ cells (Fig. 2C). Using CD3 as control of polarization, we detected that the F-actin signal at the site of cell-cell contact was more enriched in Cttn+/+ compared to Cttn−/− CD3+ cells (Fig. 2C line scans). Of note, the signal intensity of F-actin in Cttn+/+ CD3+ cells at the IS contact site with A20 cells was significantly stronger in comparison to Cttn−/− CD3+ cells (Fig. 2D).
We next generated cortactin-depleted Jurkat cells using the CRISPR/Cas9 system and 2 different cortactin-specific gRNAs, CTTN#2 and CTTN#4, which induced depletion of around 80% in both cases (Fig. 3A). Cortactin-depleted Jurkat cells showed similar levels of different surface molecules needed for proper IS formation such as CD3, LFA-1, CXCR4, and migration such as VLA-4, PSGL-1 and CD62L (Figure S3). However, we found that depletion of cortactin in Jurkat cells caused significantly reduced levels of F-actin under basal conditions that remained reduced over time upon TCR activation with soluble anti-CD3/CD28 crosslinking with goat anti-mouse (GαM) antibody, whereas control Jurkat cells induced robust actin polymerization after TCR engagement (Fig. 3B).
It is well established that actin remodeling contributes to the signaling events downstream of TCR leading to the transcription of IL-2 mRNA [27]. Indeed, we found that cortactin-depleted Jurkat cells expressed significantly less IL-2 mRNA compared to control Jurkat cells upon TCR stimulation (Fig. 3C). IL-2 expression is regulated by multiple transcription factors such as AP1, NF-κB and NFAT, all of which are activated upon TCR engagement [28]. ERK regulates the activation of the transcription factor AP-1, and since cortactin has been shown to interact with ERK in other types of cells [21, 29], we analyzed whether the phosphorylation of ERK upon TCR engagement is affected by the absence of cortactin. However, ERK phosphorylation dynamics in cortactin-depleted Jurkat cells were similar to control cells (Figure S4). Together, these results demonstrate that cortactin is necessary for proper F-actin dynamics and IS formation, but not for ERK phosphorylation.
Cttn−/− T cells display defects in TCR-mediated proliferation and CXCR4-mediated TCR-costimulation
Since cortactin-deficient T cells displayed impaired actin dynamics upon TCR engagement, we hypothesized that this also affects T cell functions upon activation. First, we evaluated whether Cttn−/− T cells presented defects in proliferation upon TCR engagement. Using a model of mixed lymphocyte reaction (MLR) with A20 as stimulating cells and Cttn+/+ or Cttn−/− T cells as responder cells, we observed a higher percentage of proliferating Cttn+/+ cells compared to Cttn−/− T cells (Fig. 4A and B). Interestingly, upon activation, murine T cells responded by switching to the expression of the 80 kDa full-length isoform of cortactin (Fig. 4C&D). However, overall levels of cortactin and HS1, the cortactin hematopoietic homolog, remained unchanged (Fig. 4C&E). By contrast, activated primary human and Jurkat T cells showed a significant increase in the expression of the 70 KDa cortactin SV1 isoform that was accompanied by a significant down-regulation of HS1 expression in primary T cells, but not in Jurkat cells (Figure S5). These results indicate a functional relevance for cortactin in activated human and mouse T cells.
Next, we analyzed the production of cytokines by the effector cells generated in the MLR assay and observed that Cttn+/+ and Cttn−/− CD4+ T cells produced similar amounts of IFNγ and TNFα (Figure S6). Hence, cortactin regulates T cell proliferation, but is dispensable for the expression of effector molecules during T cell differentiation in this model.
Given that CXCR4 is able to costimulate TCR signaling in T cells [30], and that cortactin can control CXCR4 surface levels in T cells [14], we next analyzed whether the costimulatory activity of CXCR4 is impaired in the absence of cortactin. However, we found no significant differences in the production of IL-2, CD25 or CD69 in Cttn+/+ and Cttn−/− CD4+ T cells upon TCR engagement with or without CXCL12 stimulation (Figure S7). Using plate-bound anti-CD3/CD28 antibodies to induce proliferation, we also did not detect any significant difference in the percentage of proliferating CD4+ T cells (Fig. 4F&G), suggesting that this kind of stimulus is perhaps not strong enough to detect the defect in proliferation as we observed using the MLR assay. Importantly, costimulation with CXCL12 induced increased proliferation upon TCR engagement only in Cttn+/+, but not in Cttn−/− CD4+ T cells (Fig. 4F&G). Together, these results highlight a novel role for cortactin in T cell proliferation induced by TCR/CXCR4-costimulation.
Defective migration of Cttn−/− T cells depends on the Arp2/3 complex
Since cortactin has been associated with CXCR4 regulation [14, 31], we wanted to know whether cortactin deficiency affects migration. To this end, we performed chemotaxis assays using transwell filters and CXCL12, the CXCR4 ligand, as chemoattractant. We observed that random migration of resting Cttn+/+ and Cttn−/− CD3+ cells was equally low (~ 5% of total cells), whereas in response to CXCL12 a significantly higher number of resting Cttn+/+ T cells migrated compared to Cttn−/− T cells (Fig. 5A). Activated T cells showed a significantly higher rate of random migration than resting cells and activated Cttn+/+ CD3+ cells displayed a significantly higher motility than activated Cttn−/− CD3+ suggesting that cortactin expression, specifically the full-length 80 kDa isoform in activated mouse T cells induces a more motile phenotype in T cells. Using CXCL12 as chemoattractant for activated CD3+ cells, we also observed significantly more migration in Cttn+/+ when compared to Cttn−/− CD3+ cells (Fig. 5A).
Using cortactin-depleted Jurkat cells, we observed that random migration in transwell filters was similar to control cells (~ 4% of total cells). With CXCL12 as chemoattractant, significantly more control Jurkat cells migrated compared to cortactin-depleted Jurkat cells (Fig. 5B).
To further investigate a potential role of cortactin in actin remodeling during T cell migration, we analyzed F-actin contents by flow cytometry, and found that F-actin levels in control and cortactin-depleted Jurkat cells were similarly up-regulated upon CXCR4 stimulation; however, F-actin in cortactin-depleted Jurkat cells was significantly lower from the beginning compared to control Jurkat cells (Fig. 5C).
Since cortactin stabilizes Arp2/3-dependent branched actin networks [16, 32], we next evaluated whether cortactin-depletion affects Arp2/3-mediated actin polymerization and migration. To this end, we pharmacologically inhibited the Arp2/3 complex using CK-666 and employed cytochalasin E as control that more generally interferes with all types of actin polymerization. We found that cytochalasin E prevented migration of both control and cortactin-depleted Jurkat cells (Fig. 5D). Interestingly, inhibition of Arp2/3 with CK-666 specifically reduced the migration of control Jurkat cells to levels observed in cortactin-depleted Jurkat cells and did not further reduce the migration of cortactin-depleted Jurkat cells (Fig. 5D). Considering this similar effect of Arp2/3 inhibition and loss of cortactin on the migration of Jurkat cells, we assume that cortactin regulates the migration of Jurkat cells in an Arp2/3-dependent manner.
Homing of CD4+ cells to the lymph nodes is defective in Cttn−/− mice
Next, we characterized T cell subsets and APC contents in the thymus, peripheral blood (PB), lymph nodes (LN) and spleens of Cttn+/+ and Cttn−/− mice. We detected no significant differences in the proportion of double negative, double positive or single positive cells in the thymus (Fig S8A). Additionally, no differences were observed in the frequencies of TCRβ+ cells, CD19+ B cells or CD14+ monocytes in the spleen, LN or PB (Fig S8B-D). However, we found significantly reduced numbers of CD4+, but not CD8+ T cells, in the LNs of Cttn−/− mice in comparison to Cttn+/+ mice (Fig. 6A), whereas the numbers of both CD4+ and CD8+ T cells in the spleen and PB were similar (Fig. 6B-C) suggesting that cortactin is specifically required for the migration of CD4+ T cells to lymph nodes.
Cortactin-depleted T cells fail to home to the BM and infiltrate the CNS
Finally, we tested the pathological relevance of cortactin in a disease model. It is well established that leukemic T cells display hyperactive TCR signaling [7, 33]. Given that leukemic T cells express higher levels of cortactin (Fig. 1A), and cortactin is important for leukemic infiltration and relapse in B-ALL [20], we tested whether cortactin is also important for T-ALL pathology. To this end, we performed xenotransplantations in immune-compromised NSG mice using cortactin-depleted and control Jurkat cells and analyzed organ infiltration. First, we confirmed that control Jurkat cells were detectable in mouse organs after transplantation with Jurkat cells present in BM and CNS after five weeks, but not in liver, lung, spleen or testis (Fig S9). Of note, we observed that the leukemic burden in the PB was reduced with cortactin-depleted Jurkat cells (Fig. 7A), and that they colonized the BM less efficiently compared to control Jurkat cells (Fig. 7B). Most importantly, cortactin-depleted Jurkat T cells virtually failed to infiltrate the CNS (Fig. 7C). Thus, cortactin appears to be relevant for one of the most important complications of T-ALL, which is CNS infiltration by T-ALL cells.