Eosinophils affect functions of in vitro-activated human CD3-CD4+ T cells

Hypereosinophilic syndromes (HES) are characterized by eosinophil-mediated tissue and organ damage, occurring in the setting of persistent hypereosinophilia [1]. Although several HES variants have been defined on the basis of pathogenic disease mechanisms, the majority of patients are still classified as “idiopathic” HES. In lymphocytic variant HES (L-HES), hypereosinophilia develops in response to marked over-production of eosinophilopoietic cytokines [most notably interleukin (IL)-5] by clonally expanded T-cells that often bear a CD3^-CD4⁺ phenotype [2]. A small subset of patients with L-HES may develop T-cell lymphoma years after diagnosis, possibly in relation with occurrence of specific cytogenetic abnormalities [3].

For patients with both idiopathic HES and L-HES, corticosteroids (CS) represent first-line therapy [4]. Interferon (IFN)-α is an alternative for those who are CS-dependent or CS-resistant. Both treatment options target T-cell functions and have the potential to reduce CD3^-CD4⁺ T-cell counts [2]. Treatment of HES is not curative and must therefore be maintained long-term to prevent the development of irreversible eosinophil-mediated organ damage. As a result, significant treatment-related toxicity and development of drug-resistance are common.

The recent development of targeted eosinophil-specific therapies has aroused enthusiasm for HES management. Interleukin-5 is a key mediator involved in differentiation, expansion, chemotaxis, activation, and survival of eosinophils. Because of restricted expression of the human IL-5 receptor α chain (IL-5Rα), monoclonal antibodies directed against IL-5 or its receptor are expected to have a selective impact on eosinophils [5]. The anti-IL-5 antibody, mepolizumab, was recently shown to be an effective CS-sparing and eosinophil-lowering agent in CS-responsive HES patients with both lymphocytic and idiopathic disease variants [6‐8]. Roughly half of mepolizumab-treated patients can durably be tapered off CS. Although this represents a major step forward for patients with HES, withdrawal of therapeutic agent(s) active on T-cells is a subject of concern for patients with L-HES because of the pre-malignant potential of CD3^-CD4⁺ T-cells.

The perspective of eosinophil-targeted therapy for HES has therefore raised new questions, one of which pertains to the fate of pathogenic CD4 T-cells. The present study was undertaken to explore this issue, with a special focus on clonal CD3^-CD4⁺ T-cells associated with L-HES. The effects of eosinophils on CD4 T-cell activation were investigated in vitro as an indirect means of exploring potential consequences of treatment-induced eosinophil depletion.

Eosinophil-specific therapeutic agents targeting IL-5 or its receptor would not be expected to directly affect CD4 T-cells, which are known not to express the IL-5 receptor [12]. We first established that this is also true for CD3^-CD4⁺ T-cells, showing that aberrant cells from 3 patients with L-HES stain negatively for the IL-5Rα chain (Figure 1).

Therapeutic eosinophil depletion may however affect CD4 T-cells indirectly, by interrupting cross-talk between these two cell-types. To explore this eventuality, we first investigated whether eosinophils affect in vitro activation of CD4 T-cells from healthy subjects. CD4 T-cells were cultured in presence of anti-CD3/CD28 antibodies for 48 hours, in absence and in presence of autologous IgA/anti-IgA pre-activated eosinophils. Eosinophils consistently and significantly increased T-cell proliferation (Figure 2a), as assessed by H³-thmidine incorporation (median 15722 versus 174332 cpm, in absence and in presence of eosinophils respectively, p = 0.008). This effect was dose-dependent, with proliferation increasing in parallel with eosinophil-T cell ratios (Additional file 1: Figure S1). Eosinophils also further increased membrane CD25 expression by CD4 T-cells (Figure 2b) (median proportion of CD25⁺ CD4 T-cells 51.4 versus 86.1% in absence and in presence of eosinophils respectively, p = 0.03). In contrast, we did not observe a statistically significant effect on the proportion of IL-2 (Figure 2c), or IL-5 (data not shown)-expressing T-cells as assessed by flow cytometry, and measurement of IL-2 in culture supernatants did not yield reproducible results (not shown).

To more specifically address the fate of clonal T-cells in L-HES patients treated with eosinophil-depleting therapy, we set up cultures with CD3^-CD4⁺ T-cells from L-HES patients, in absence and in presence of eosinophils. In these experiments, T-cell activation was achieved using allogeneic mature DC, which provide costimulatory, TCR-independent signals to CD3^-CD4⁺ cells [11]. Consistent with our previous work, DC induced proliferation of CD3^-CD4⁺ cells (Figure 3a) and secretion of IL-5 (Figure 3b). Eosinophils consistently and significantly reduced DC-induced proliferation of the aberrant T-cells (median reduction 58%), as well as IL-5 production (median reduction 93%). We previously reported that DC-induced CD3^-CD4⁺ T-cell proliferation and Th2 cytokine production involve IL-2/IL-2Rα interactions [11]. We did not observe any reproducible eosinophil-induced changes in IL-2 concentrations in culture supernatants (not shown), which were low or undetectable in all culture conditions, presumably due to recapture by activated T-cells. Attempts to re-stimulate the cells at the end of co-cultures for detection of intracytoplasmic cytokine expression were unrewarding in the majority of experiments; however, we did observe a clear-cut decrease in the percentage of IL-2-positive and IL-5-positive T-cells from patient L-HES₁ when eosinophils were added (Additional file 2: Figure S2). As for IL-2Rα, CD3^-CD4⁺ T-cells from patient L-HES₁ showed increased membrane expression in response to DC, and eosinophils consistently lowered the percentage of positively stained cells (Figure 3c).

Overall, these in vitro experiments suggest that eosinophils display differential effects on healthy activated CD4 T-cells, and on abnormal clonal CD3^-CD4⁺ T-cells found in L-HES.

The data presented herein with normal CD3/CD28-activated CD4 T-cells, supports previous studies indicating that eosinophils potentiate CD3-mediated T-cell activation. Eosinophils have indeed been shown to support antigen and super-antigen induced activation and proliferation of normal CD4 T-cells in a HLA-DR dependent fashion [13‐15]. It is therefore conceivable that eosinophil depletion in clinical situations typically associated with antigen-driven T-cell activation and (hyper)eosinophilia, including allergic disorders, may ultimately blunt these T-cell responses to some degree in vivo. Our observation that CD4 T cell proliferation progressively decreases in parallel with eosinophil concentrations indicates that tighter therapeutic control of eosinophil levels in patients with reactive eosinophilia may result in inhibition of other down-stream consequences of pathogenic T cell activation.

The relevance of these observations for humans receiving eosinophil-targeted therapy remains elusive. Others have explored short-term (maximum 3 monthly infusions) effects of anti-IL-5 treatment on normal CD4 T-cells in vivo, in patients with asthma, idiopathic HES, and eosinophilic esophagitis [16‐18]. Although these studies suggest overall that therapeutic eosinophil depletion has no significant effects on the biology of normal CD4 T-cells, the observation period is short compared to the prolonged requirement for treatment in patients with allergic disorders and HES. Long-term in vivo evaluation of CD4 T-cell biology in patients treated with eosinophil-depleting therapy, together with longitudinal in vitro quantification of pertinent antigen-specific responses are warranted.

In contrast to normal CD4 T-cells, CD3^-CD4⁺ Th2-cell functions were consistently inhibited in vitro by eosinophils in the present study. This may be related to intrinsic properties of the CD3^-CD4⁺ T-cells and/or their reliance on TCR/CD3-independent, rather than TCR-dependent, activation pathways. Indeed, studies reported so far have shown a central role of HLA-DR, and thus TCR engagement, in eosinophil-mediated activation of T-cells [13‐15]. Because L-HES-associated CD3^-CD4⁺ cells lack membrane TCR/CD3 expression, they are unable to interact with MHC molecules, and our recent study on their gene expression profile has shown reduced expression of numerous CD3-related intracellular signaling molecules [10]. In contrast, these cells have been shown to respond to costimulatory ligands [11], some of which are expressed by eosinophils [14, 19]. Eosinophil-derived signals may therefore affect downstream signaling pathways differently, depending on whether the T-cells are activated through the TCR/CD3 complex or not. Besides reduced expression of many CD3-associated genes, significant alterations in the expression level of various other membrane-expressed receptors and signaling pathways were observed in CD3^-CD4⁺ cells [10], some of which may be relevant with regard to interactions with eosinophils.

Our findings suggest that eosinophils may actually play a regulatory role on the global activation status of CD3^-CD4⁺ T-cells. Should this be the case in vivo, eosinophil depletion with eosinophil-targeted treatment may actually favor clonal T-cell outgrowth in some patients with L-HES. Several patients with L-HES have been treated with anti-IL-5 for several years in the setting of an open-label study, but none have been reported to develop T-cell lymphoma [8]. Although two patients with idiopathic HES enrolled in the study did develop lymphoma during the observation period, neither of them had detectable CD3^-CD4⁺ T-cells. Long-term, longitudinal, in vivo data on CD4 T-cells in patients with L-HES receiving eosinophil-targeted therapy is currently unavailable. Absolute counts of abnormal T-cells in patients with L-HES were not systematically assessed throughout the above-mentioned open-label study to assess for clonal expansion, nor were TCR gene rearrangement analyses performed on the entire cohort to check for appearance of new T-cell clones. Our in vitro data emphasize the need for punctilious in vivo assessment of aberrant T-cell clones in the setting of long-term eosinophil-targeting treatment protocols. Although the risk of developing lymphoma in patients with L-HES is low, careful clinical follow-up is warranted, regardless of the chosen treatment strategy.

In conclusion, our data provide further evidence that activated eosinophils interact directly with CD4 T-cells and may affect their functions in diseases associated with (hyper)eosinophilia. Additionally, in vivo, tissue-dwelling eosinophils have been shown to interact with numerous other immune and non-immune cell types thereby indirectly affecting T cell functions [20]. With increasing opportunities to treat various eosinophilic conditions using eosinophil-targeted therapy, special attention should be paid to potential consequences of eosinophil depletion on T-cell biology. Specifically, the present findings suggest that eosinophils may restrain expansion of clonal CD3^-CD4⁺ T-cells associated with L-HES, perhaps contributing to the indolent nature of this pre-malignant disease. Future investigations should focus on the eosinophil-derived mediators and surface molecules affecting CD3^-CD4⁺ T-cell proliferation and activation.