Interleukin-2 is an important T cell growth factor (1) that critically contributes to immune homeostasis by promoting development and expansion of regulatory T cells (2–7). In fact, several strategies used IL-2 to expand regulatory T cells to treat autoinflammatory and allergic diseases. Examples include IL-2/anti-IL-2 antibody complexes that extend the half-life of IL-2 and contribute to its improved targeting (8, 9), and IL-2 muteins with improved IL-2R specificity, both being the subject of preclinical (10) or clinical studies (11). IL-2 exerts its Treg-promoting activity via binding to the high-affinity IL-2 receptor (CD25) (3, 4, 7, 12). Spontaneous mutations in the human IL-2RA (CD25) gene (13) or genetic ablation of the mouse il2ra locus result in Treg deficiency (7, 12), uncontrolled T cell expansion (14), and overt autoimmunity (13, 15), a clinical picture similar to the immunodysregulation polyendocrinopathy enteropathy X-linked (IPEX) syndrome in which Foxp3 (16), the signature transcription factor of Treg cells, is dysfunctional (17, 18).

Given the importance of IL-2 (19, 20), we were interested in identifying targeted drugs that would interfere with aspects of the TCR-signaling pathway and increase secreted IL-2 levels from T cells. We hypothesized that such drugs would concomitantly limit cellular activation/proliferation of T cells. Previous reports have shown that this constellation best promotes conversion of naïve T cells to Treg cells (21–24). Accordingly, we assumed here that partial inhibition of proliferation in the presence of elevated/induced IL-2 levels could promote Treg formation and thus should be an appropriate guide for our mining strategy to identify interesting lead substances (19, 25).

Therefore, in Jurkat E6-1 cells (26, 27) we initially evaluated TCR cell signaling pathway SMI such as the Src-kinase inhibitor Saracatinib (28), the Erk-2 inhibitor Vx-11e (29), the calcineurin inhibitors Tacrolimus (30) and Cyclosporin A (31), the c-Jun-N-terminal-kinase inhibitor SP600125 (32), the mTORC1/2 inhibitor AZD8055 (33) and the PDK-1/TBK-1 inhibitor BX-795 (34) for their potential IL-2-inducing activity. Our screen revealed BX-795 as unique in its capacity to enhance IL-2 production in Jurkat E6-1 T cells upon stimulation. We went on to extensively explore the immunomodulatory properties of BX-795 in human and murine T cells using human peripheral blood mononuclear cells (PBMC) activated by CD3 antibody or superantigen from Staphylococcal enterotoxin A and allergen-specific T cells from humanized allergy mice (35). In vivo studies in these humanized allergy mice and in a house dust mite (HDM)-based mouse model of allergen-specific airway inflammation confirmed the in vivo allergy-mitigating effect of BX-795 and together with gene expression analyses established the definition of a putatively new Th cell phenotype.

Based on previous findings highlighting the importance of IL-2 for Treg expansion in vivo (8, 9), we hypothesized that SMI inducing IL-2 hypersecretion by T cells activated via their TCR would lead to the differentiation of cells with a Treg phenotype, which may have implications for the treatment of allergic inflammation. Indeed, we identified here a previously unrecognized function of the multikinase PDK-1/IKKε/TBK-1 small molecule inhibitor BX-795, namely its capability of inducing an IL-2 hypersecreting Th cell type upon TCR activation. This was intriguing because the other SMI studied, particularly the previously described TBK1 inhibitor amlexanox, failed to do so (43). Our initial screen was performed in Jurkat E6-1 cells, because these cells historically formed one of the centerpieces for the detailed investigation of T cell receptor signaling processes. While human HPB-ALL and HuT-78 and murine EL4 and LBRM-33 cells belonged to the “legendary group” of immortalized T cell lines, Jurkat E6-1 T cells stood out (26). The merits of Jurkat E6-1 T cells as in vitro test system for TCR-signaling are several-fold: i) they are excellent IL-2 producers upon TCR ligation (27); ii) they produce IL-2 especially well when two signals are provided to them for stimulation, one of them being a TCR-derived signal the second one a signal activating PKC, such as phorbol myristate acetate (57); iii) they were the basis for a number of sublines, generated by somatic mutation and lacking, for instance, TCR αβ (58), CD3 (59) and CD45 (60); iv) TCR-mutant Jurkat E6-1 cells were the basis for showing that a calcium signal, e.g., provided by a calcium ionophore, could bypass the TCR-signal-requirement and lead to full-blown T cell activation in the presence of PKC activation (57); v) Jurkat E6-1 T cells were the first T cells to be used in fluorescent calcium measurements by Imboden et al. (59).; vi) these cells were the basis for the description of protein tyrosine kinase signaling and thus spearheading the characterization of LCK, ZAP70, CARMA1, and a number of further signaling molecules involved in the TCR-signaling pathway (61–63)); vii) nowadays, Jurkat E6-1 cells form the basis for genetic and functional (reporter) screens devoted to human T cell and HIV biology, among other topics (64–67). The fact, that a currently performed data-base search for this cell line revealed 400 citations during the year 2022 in PubMed indicates that this immortalized human T cell line, since its establishment more than 40 years ago, is still a very valuable tool for immunologists world-wide.

Hypersecreted IL-2 was identified upon activation of both human and murine T cells stimulated by surrogate TCR/CD3 ligation (OKT3), superantigen from Staphylococcus aureus or nominal antigen (major mugwort pollen allergen, Art v 1). Although IL-2 hypersecreting Th-IL-2 cells lacked Foxp3 expression on the mRNA and protein level, they were found to mimick iTreg by mRNA expression phenotype. In fact, RNA-Seq analyses identified 2.816 protein-coding genes which expression was similar in Th-IL-2 and iTreg cells but exhibited differential regulation in both naive and effector CD4⁺ T cells. For comparison, only 268 genes are coregulated in Th-IL-2 cells and Teff cells, indicating the much lower degree of relatedness between these two cell types. The hypersecreted IL-2 was found to be biologically relevant, as its antibody-based neutralization, e.g., completely inhibited the activation-induced up-regulation of the Treg signature surface molecule ecto-5′-nucleotidase (Nt5e, CD73), which is well-known for its triggering of the degradation of AMP towards regulatory adenosine (68). In contrast, the increased IL-2 secretion levels were not required for the Th2 inhibitory capacity of BX-795 as shown in experiments using blocking antibodies ( Figure S10 ). The suppressive potential of BX-795-induced T cells, although somewhat lower when compared to iTregs ( Figure S9 ), is in line with previous reports showing that PDK1 activity contributes to the suppressive capacity of regulatory T cells (69).

Upon allergen-specific challenge in the presence of BX-795, T helper cells downregulated Th2 immunity in vivo and made mice less susceptible for the development of allergic airway inflammation. This was mirrored in T cells by the significant downregulation of Th2 cytokine production and the expression of the Th2 signature transcription factor GATA-3, which resulted in a reduced pathological influx of eosinophils into the lungs of allergen-sensitized and -challenged mice ( Figures 6B ; S16B ). While Th2 cell-based reduction of cytokine production certainly contributed to reduced eosinophil infiltration, we cannot exclude that direct effects on innate immune responses contributed to the observed reduction in eosinophils. For instance, it was recently shown that cGAMP triggered HDM extract-induced asthma in mice via IL-33 in a TBK-1 dependent manner which could be antagonized by the bona fide TBK-1 inhibitor amlexanox (70).

We initially hypothesized that Th-IL-2 cells may be generated by modulation of Signal-1 strength. Indeed, when looking at TCR-induced proliferation in allergen-specific T cells, we observed that BX-795 moderately inhibited T cell proliferation ( Figure S5 ). Also, TCR-induced intracellular Ca²⁺-fluxing in Jurkat E6-1 and primary human T cells as determined by the calcium indicator Indo-1 was slightly reduced at biologically relevant BX-795 concentrations ( Figures S2A, B ). These findings may be related to early BX-795-attenuated tyrosine-phosphorylation (as determined with the anti-pan tyrosine-phosphorylation antibody 4G10) ( Figure S2 ). However, BX-795 sustained overall ( Figure S2C ), Lck- ( Figure S2D ) and CD3 ζ-chain-specific ( Figure S2E ) tyrosine-phosphorylation at later time points, while it left Erk tyrosine-phosphorylation unaffected ( Figure S2F ) when compared to stimulation with solvent conditions (DMSO). These findings let us speculate that BX-795 may enhance IL-2 expression by disinhibiting a TCR-distal feedback loop, which may explain how, in fact, attenuated Signal-1 strength increases IL-2 production. Such a feedback loop had indeed been posited for IKKε, suggesting that the BX-795 target IKKε provided a negative TCR-signaling feedback (45). In that study, knockout of IKKε resulted in decreased phosphorylation of the transcription factor NFATc1 and enhanced its nuclear translocation thereby enhancing IL-2 expression (45). However, when we investigated NFATc1 translocation by imaging flow cytometry in the presence of BX-795, we did not find evidence for enhanced NFATc1 translocation. Rather, in our studies, the increase in IL-2 secretion clearly preceded the expression and translocation of NFATc1 ( Figure S4 ), ruling out direct involvement of the IKKε-NFATc1 axis in the IL-2 hypersecretion phenotype, and suggesting a hitherto undisclosed mechanism. That the activation induced shut-off of IL-2 may depend on a suppressor, which is sensitive to cycloheximide has been shown previously and allows to speculate that this yet unidentified suppressor could be a target of BX-795 (71). The authors cannot completely exclude the possibility of decreased consumption of secreted IL-2 as a possible contributor to the elevated IL-2 levels detectable in culture supernatants, e.g., due to decreased CD25 expression. However, it seems rather unlikely that this is the only factor for the increased IL-2 levels in culture supernatants, because IL-2 mRNA expression levels were also significantly increased in human ( Figure 1 ) and murine T cells ( Figure 3 ). In addition, increased IL-2 levels were also observed at the single cell level in flow cytometry ( Figure S8 ), suggesting that increased production of IL-2 by Th-IL-2 cells is certainly an important, if not the main, factor for the observed increase in IL-2 levels in supernatants.

RNA-Seq experiments showed that BX-795-induced Th cells appeared to be way more similar to iTregs than to naïve T cells or Teff ( Figure 3C ). This indicated that BX-795 does not inhibit the transition from naïve T cells towards a differentiated phenotype but rather directly induces a distinct Th cell phenotype which is highly similar to iTregs which we refer to here as Th-IL-2 cells. Although there was a high degree of similarity between Th-IL-2 cells and iTreg, Th-IL-2 cells completely lacked Foxp3 expression. Moreover, under iTreg differentiating conditions BX-795 reduced numbers of CD25⁺Foxp3⁺ Tregs in vitro from 50.4 ± 7.5% mean ± SEM to 19.2 ± 4.0% mean ± SEM, corresponding to a 62% decrease ( Figures S12 , S13 ). In contrast, the ecto-5’-nucleotidase CD73 and the ectonucleoside triphosphate diphosphohydrolase-1(CD39), both important adenosine-generating enzymes of Treg cells (48, 72), remained highly expressed (98.2 ± 0.7 mean ± SEM versus 96.5 ± 1.6% mean ± SEM) or were only modestly downregulated (17.56 ± 4,2% mean ± SEM versus 11.99 ± 4.8% mean ± SEM) ( Figures S12 , S13 ). However, another transcription factor closely related to the development of Treg cells, Ikzf2 (Helios) remained stably expressed in iTregs even in the presence of BX-795. This was in concordance with our finding that Helios expression was significantly induced in the larger part of Th-IL-2 cells ( Figures 3D ; S8D ), and may indicate that the BX-795 targeted pathway represents a switch for the differentiation of Helios⁺/Foxp3^- T helper cells (73). Although differences between Foxp3⁺/Helios⁺ T cells and Foxp3⁺/Helios^- have been associated with different types of regulatory T cells the phenotypic characteristics of Helios⁺/Foxp3^- T cells have not been described in detail so far (46, 47). Helios was proposed as a specific marker for thymic-derived Treg cells (47) and the co-expression of Helios and Foxp3 was reported to increase the regulatory activity of human engineered regulatory T cells (74). Moreover, Helios expression has been linked to the control of several regulatory functions of Treg cells (46) and has been described as an exclusive marker of activated Treg cells expressing the markers GARP/LAP (75). We could confirm our findings obtained in the RNA-Seq experiments on the protein level in flow cytometric analyses demonstrating that BX-795-induced expression of the transcription factor Helios in a fraction of Th-IL-2 cells ( Figure S8 ). Notably, Helios expression was independent of soluble IL-2 and TGF-β levels as determined with blocking antibodies (data not shown). Whether Helios is responsible for the induction of the here presented T cell phenotype remains to be shown. The BX-795-induced increase in Th-IL-2 cells amounted to 36.1 ± 6.2% (BX 1.2 µM, % IL-2⁺ of CD3⁺CD4⁺, mean ± SEM) while only 12.8 ± 1.2% (% of CD3⁺CD4⁺IL-2⁺, mean ± SEM) of Th-IL-2 cells were co-expressing Helios on a single cell level. This may indicate that Helios positivity is either a salient feature of a subset of Th-IL-2 cells. Alternatively, parallel immunological detection of Helios and IL-2, which is technically difficult because of gross differences in the optimal composition of the fixation reagents used, could considerably underestimate the number of Th-IL-2 cells co-expressing Helios as determined by flow cytometry. A regulatory cell type lacking Foxp3 expression is represented by IL-10⁺Foxp3^- Treg cells (76). Since treatment with BX-795 decreased the number of IL-10⁺ T cells, it seems rather unlikely that Th-IL-2 cells resemble IL-10⁺Foxp3^- Treg cells. However, the authors cannot exclude the possibility that Th-IL-2 cells share a common progenitor with IL-10⁺Foxp3^- cells.

Apart from the induction of Th-IL-2 cells resembling iTregs, the inhibition by BX-795 of type 2 immunity was the second major finding of this study. In fact, BX-795-dependent inhibition of type 2 cytokine secretion was a robust trait observed in different species and model systems in vitro and in vivo applying, among others, double transgenic allergy mice ( Figures 6 , 7 ) and wild type C57BL/6 mice for detailed preclinical analyses ( Figure S16 ), which was accompanied by a decrease in GATA-3 expression in lung T cells of allergen-sensitized and challenged mice ( Figures 6 ; S16 ). Moreover, the influx of eosinophils into the lungs of allergen-exposed mice was reduced to background levels in the presence of BX-795, as was the formation of goblet cells, indicating the far-reaching effects of blockade of type 2 immunity by BX-795. Interestingly, GM-CSF which is a cytokine also secreted by Th2 cells was not inhibited throughout our studies in different model systems (77). This raises the question whether Th-IL-2 cells have similarities with the recently described T_GM cells (78). The transcription factors c-Maf and Fli-1 have been described as regulators of GM-CSF expression (79, 80). Previously, the transcription factor c-Maf has been shown to be exclusively involved in positive regulation of IL-4 transcription but not other Th2 cytokines (81). In our transcriptomic analyses, up-regulation of c-Maf was clearly inhibited by BX-795 whereas Fli-1 was increased although non-significantly ( Figure S17 ). This may explain the discrepancy observed with regards to IL-4 inhibition but GM-CSF maintenance. Somewhat similar to GM-CSF, also IL-17A was not inhibited throughout our studies by BX-795. Th17 cell differentiation has been shown to rely on the cell cycle-dependent function of CARMA1, which suggests that the changes in cytokine expression seen in Th-IL-2 cells are not explainable by mere inhibition of cell cycle progression/proliferation as inhibition of IL-17 expression would then also be an expected outcome (82).

This study is not without certain limitations. For instance, the epigenetic changes of IL-4, IL-5, IL-13, and IL-2 loci in Th-IL-2 cells and the stability of their phenotype over time remain to be determined in future studies using epigenetic analysis tools, e.g., by ATAC-Seq. Moreover, transcriptomic analyses of Th-IL-2 cells at the single cell level will help to better understand the transition processes that drive naïve T cells to develop into Th-IL-2 cells and will be necessary to identify and compare phenotypic clusters of in vivo-induced with in vitro-generated Th-IL-2 cells.

Furthermore, we have not yet answered the question why the IL-2 hypersecreting phenotype, which was so robustly reproducible in vitro with primary ( Figures 2 ; S3 ) but also with Th2-differentiated ( Figure S6 ) cells, was not more clearly apparent in the allergic airway inflammation models in vivo. Our findings in that respect may be attributable to variables such as available drug concentration over time, complexity of the tissue/cellular composition, different pharmacokinetic and pharmacodynamic parameters of BX-795 or IL-2 itself, which are less well controllable and assessable in vivo when compared to in vitro experiments. Nevertheless, frequencies of IL-2⁺/IL-4⁺ and IL-2⁺/IL-13⁺ T cells were clearly reduced in vivo (2.6 ± 0.4% and 1.5 ± 0.3% to 1.4 ± 0.1% and 0.7 ± 0.1% mean ± SEM, respectively) while overall IL-2 producing cells remained the same ( Figure 6 ). It is tempting to speculate that the cell fractions turning into IL-2 single positive cell types could be identical to the cell type observed in vitro upon BX-795 stimulation. In addition, our study is limited by the fact that we focused primarily on the effects of BX-795 on the differentiation and function of CD4⁺ T helper cells. Thus, we can only speculate how BX-795 may act on other cell types such as, e.g., CD8⁺ T cells, dendritic cells or type 2 innate lymphoid cells (ILCs), the latter representing recently discovered key players in the initial phase of allergic diseases (83).

Also, the role and capability of Th-IL-2 cells to regulate levels of immunosuppressive adenosine by virtue of their CD73 and CD39 up-regulation requires to be determined in future studies.

Taken together, we have provided evidence in this study of the potential of BX-795 to limit Th2 inflammation and inhibit T cell activation and proliferation in a dose-dependent manner in multiple species and model systems. If applied at a constant (low) dose over time during differentiation of naïve T cells, BX-795 induced a phenotype displaying characteristics of iTregs with hallmark differences between Th-IL-2 cells and iTregs seen with regards to IL-2 secretion and Foxp3 expression. In addition to other studies showing anti-inflammatory capacities of BX-795 (84) and anti-viral effects (85) we here extend the potential field of application for this SMI compound towards Th2 inflammatory diseases such allergic asthma.

Since the transcriptional regulation of type 2 cytokine expression in ILC2 cells has been found to be very similar to that in Th2 cells, it can be speculated that BX-795 treatment would produce similar effects, i.e., inhibition of type 2 inflammatory cytokine secretion, in this cell type (86).

The RNA-seq data presented in the study are deposited in the GEO repository, accession number GSE224271 and are accessible under the following link: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE224271.

The studies involving human participants were reviewed and approved by Ethics Commission of the Medical University of Vienna, Borschkegasse 8b/E06, 1090 Vienna, Austria (EK Nr: 203 1565/2017). The patients/participants provided their written informed consent to participate in this study. The animal study was reviewed and approved by Institutional Review Board of the Medical University of Vienna and the Federal Ministry of Science, Research and Economy, Minoritenplatz 5, 1010 Vienna, Austria (GZ : BMBWF-66.009/0288-V/3b/2018).

PT and WP conceived the research; PT designed, performed and analyzed most of the experiments with the help of BK, US, CK, PSc, DT, and MZ; LR, JK, and TB helped performing and analyzing Imaging flow-cytometry experiments. TB provided resources and infrastructure. AN generated double transgenic allergy mice. SJ and PSt provided reagents and materials. GG provided recombinant Art v 1 protein. HS provided resources and ideas. SD performed bioinformatic analyses. PT and WP wrote the paper. All authors contributed to the article and approved the submitted version.