Impaired interferon-γ production in response to live bacteria and Toll-like receptor agonists in patients with ataxia telangiectasia

Abstract

Ataxia telangiectasia (AT) is a pleiotropic autosomal recessive neurodegenerative disorder with associated immunodeficiency and cancer predisposition, caused by mutational inactivation of the ATM gene. Early death usually results from lymphoreticular malignancy or recurrent, chronic respiratory infections. Immune deficiency of AT patients is heterogeneous and involves both humoral and cellular responses. Reports on the number and integrity of immunocompetent cells in AT are conflicting. In the early phase of infection, the interleukin (IL)-12/interferon (IFN)-γ axis plays a crucial role in first-line defence against pathogens. In a whole blood assay we studied the IL-12/IFN-γ axis in the immune response of AT cells to the Toll-like receptor agonists lipopolysaccharide and heat-killed Staphylococcus aureus, as well as whole live M. bovis bacille Calmette–Guérin (BCG). The function of AT antigen-presenting cells was normal in terms of IL-12 production, while IFN-γ production by T and natural killer (NK) cells was severely impaired, even in the presence of adequate co-stimulation by exogenous IL-12.

Introduction

Ataxia telangiectasia (AT, OMIM 208900) is a human autosomal recessive disease characterized by neurodegeneration, immunodeficiency and cancer predisposition, caused by mutational inactivation of the ATM gene. Clinical features of the disease are cerebellar ataxia, oculocutaneous telangiectasias, growth retardation, gonadal dysgenesis, premature ageing and hypersensitivity to ionizing radiation [1]. Early death usually results from lymphoreticular malignancy or recurrent, chronic respiratory infections. Immune deficiency of AT patients is heterogeneous and involves both humoral and cellular responses, with a reduction of IgG2, IgA and IgE, thymic hypoplasia, relative increase of lymphocytes expressing the γ/δ form of the T cell receptor and selective reduction of circulating CD45RA⁺ T cells [2,3]. The proportion of total B cells in the peripheral blood of AT patients is usually normal [1]. Although lymphocyte subsets and T cell function have been studied intensively in AT, there are still conflicting reports about the number and integrity of immunocompetent cells.

Members of the human Toll-like receptor (TLR) family play a key role in innate immunity with regard to the induction of immune and inflammatory responses in the early phase of infection. Stimulation of these receptors leads to activation of transcription factors of the nuclear factor-κB (NF-κB) family and subsequent expression of a variety of cytokines and co-stimulatory molecules crucial to adaptive immune responses [4,5]. One of the central cytokines produced after TLR engagement is interferon (IFN)-γ. It plays diverse important roles in the maintenance of immunological homeostasis, in particular in the activation of a wide range of T helper (Th)-1-associated cellular immune functions that are central to host defence against viral and bacterial infections [6]. The main stimulus for IFN-γ production is interleukin (IL)-12, which is secreted mainly by monocytes/macrophages in response to bacterial stimuli. Upon stimulation with IL-12, IFN-γ is produced by T and natural killer (NK) cells [6–9].

To determine the immunocompetence of AT cells, we investigated the IL-12/IFN-γ axis in a ‘close to nature’ whole blood assay in 10 patients with AT upon stimulation by live bacille Calmette–Guérin (BCG), the TLR agonists lipopolysaccharide (LPS), and Staphylococcus aureus with or without addition of IL-12 and IFN-γ, respectively.

Materials and methods

Subjects and patients

We compared 10 patients (ratio male/female: 6 : 4) with adult local (n = 50) healthy subjects as well as travel controls (n = 10). As we have shown earlier, age and gender had no significant effect on the production of IFN-γ or IL-12p40 by controls, regardless of the type of stimulation [8]. Cytokine production is known to be in the range of adult values from age 3 years onwards [10–12]. We therefore did not analyse healthy age-matched controls in parallel with the present experiments. Mean age [standard deviation (s.d.)] was 34 years (s.d.) for local controls, 16 years (s.d.) for travel controls and 12 years (s.d.) for patients. For group descriptions see Table 1. All patients examined were clinically stable and free of infection at the time of analysis. None of them was on prophylactic antibiotic treatment or received immunoglobulin replacement therapy. All patients or patients’ guardians supplied written informed consent prior to the study. Human experimentation guidelines of Good Clinical Practice, the German Drug Act and the declaration of Helsinki/Hong Kong were followed in the conduct of clinical research. Local ethical committee approval was obtained.

Table 1.

Demographic data and lymphocyte subsets of patients with ataxia telangiectasia; major lymphocyte subsets analysed by fluorescence activated cell sorter; values are given as cell counts/µl. Normal values are expressed as ranges, according to values obtained by Shearer et al. [42].

Patient no.	Age/sex (years)	Leucocytes (cells/μl)	Lymphocytes (cells/μl)	CD19+	CD3– CD16+CD56+	CD3+	CD3+/CD4+	CD3+/CD8+	CD4+/ CD45RA+	CD4+/ CD45RO+	CD8+/ CD45RA+	CD8+/ CD45RO+
1	2/m	4 220	945	153	159	575	302	192	n.d.	n.d.	n.d.	n.d.
2	7/f	6 560	1692	256	664	810	552	249	148	485	94	180
3	11/m	4 430	1418	116	554	748	343	311	58	920	78	273
4	11/m	6 400	1184	83	322	750	423	350	37	411	240	239
5	13/f	7 200	1275	64	812	1058	676	412	280	1249	151	261
6	14/f	7 800	2652	203	865	1585	671	724	133	552	151	644
7	14/f	5 300	1590	63	135	994	472	483	116	381	n.d.	n.d.
8	14/m	5 200	1825	66	895	845	264	515	29	252	162	444
9	19/m	1 080	2700	60	550	1110	671	1036	50	636	93	389
10	20/m	16 200	1490	139	253	987	304	630	5	301	n.d.	n.d.
Normal	Under 6	5 500–15 000	> 1200	390–1400	130–720	1400–3700	700–2200	490–1300	430–1500	220–660	380–1100	90–440
values	6–12	4 500–12 000	> 1200	270–860	100–480	1200–2600	650–1500	370–1100	320–1000	230–630	310–900	70–390
	13–20	4 800–10 800	> 1200	110–570	70–480	1000–2200	530–1300	330–920	230–770	240–600	240–710	60–310

Whole blood cultures and activation by live BCG

Venous blood samples were collected into heparinized tubes. They were diluted 1 : 2 in RPMI-1640 (GibcoBRL, Gaithersburg, MD, USA) supplemented with 100 U/ml penicillin and 100 µg/ml streptomycin (GibcoBRL). We dispensed 6 ml of the diluted blood sample into four wells (1·5 ml/well) of a 24-well plate (Nunc, Roskilde, Denmark). It was then incubated in a two-stage procedure during 18 and 48 h at 37°C in an atmosphere containing 5% CO₂/95% air, and under four different conditions of activation: with medium alone, with live BCG (Mycobacterium bovis BCG, Pasteur substrain) at a multiplicity of infection (MOI) of 20 BCG/leucocyte, with BCG plus IFN-γ (5000 IU/ml) (Imukin®; Boehringer Ingelheim, Ingelheim, Germany), and with BCG plus recombinant IL-12p70 (20 ng/ml) (R&D Systems, Minneapolis, MN, USA). The first incubation stage was completed after 18 h of culture, 450 µl supernatant was collected from each culture well and frozen at −80°C. After 48 h, by the end of the second incubation stage, the whole remaining volume of each well was recovered, centrifuged at 1800 g for 10 min, and the supernatant was stored frozen at −80°C until analysis. For patients whose blood samples were transported from elsewhere, we also analysed a ‘travel’ control in parallel.

TLR stimulation

For activation with TLR ligands, whole blood was diluted 1 : 5 in RPMI-1640 (Gibco) and incubated with polymyxin B (10 µg/ml, Sigma, Lyon, France) (except for stimulation by LPS) at 37°C for 30 min. Cells were then activated with LPS from Salmonella minnesota R595 (10 μg/ml, List Biologicals, Campbell, CA, USA), phytohaemagglutinin (PHA) (final dilution 1/700, Difco), tumour necrosis factor (TNF)-α (20 ng/ml, R&D Systems) or heat-killed Staph. aureus [SAC, 5 × 10⁶ particles/ml, American Type Culture collection (ATCC)]. Supernatants were collected from each culture well after 18 h and 48 h and frozen at –80°C until analysis.

Cytokine enzyme-linked immunosorbent assay (ELISA)

Cell culture supernatants were taken at 18, 24 and 48 h and their cytokine concentrations were analysed by ELISA, using the human Quantikine IL-12p40 and IL-12p70 kits from R&D Systems, the human Pelikin or Pelipair IFN-γ kit from Sanquin (Amsterdam, Netherlands) and the human IL-10 kit from Sanquin, according to the manufacturers’ guidelines. These kits were applied using matched antibody pairs. The production of IL-12p40 and IL-12p70 was evaluated in 18 h supernatants; IFN-γ production was analysed in 48 h supernatants after activation with BCG, LPS and SAC and IL-10 production was evaluated in 48 h supernatants. Optical density was determined using an automated MR5000 ELISA reader (Dynatech, Denkendorf, Germany).

Quantitative analysis was carried out using the non-linear four-parameter logistic calibration model developed by O’Connell [13]. An in-house software based on Microsoft Excel application matrix was developed for this purpose: model adequacy was always satisfactory. This software was validated by comparing model coefficients estimates obtained with those given by S-Plus 6·2 (Insightful, Seattle, WA, USA) using the same four-parameter model. Intermediate results for each cytokine are expressed in pg/ml. However, peripheral blood mononuclear cell (PBMC) counts vary interindividually and in particular are dependent on age. We therefore standardized the final results by expressing them per million PBMC in the following units: pg/ml/10⁶ PBMC. The number of PBMC was determined from blood cell counts carried out on day 0.

Distribution of IFN-γ producing cells by intracellular flow cytometry

Experiments for the analysis of intracellular IFN-γ staining by fluorescence activated cell sorter (FACS) were performed after 24-h culture of PMBC of healthy controls and AT patients with LPS at 10 ng/ml (LPS from S. minnesota R595; Sigma). Cell culture and intracellular staining of IFN-γ were performed as described by Schubert et al. [3]. In analogy to stimulation by PBu₂ and ionomycin, the distribution of IFN-γ producing cells was two-thirds T cells to one-third NK cells in healthy controls as well as in patients with AT. We therefore decided to standardize results for cytokine production with respect to the number and distribution of IFN-γ-producing T and NK cells.

Results

Production of IL-12 and IFN-γ in whole blood from healthy controls

We compared the production of IL-12 or IFN-γ after stimulation with BCG alone, BCG plus IFN-γ, BCG plus IL-12 and after stimulation with LPS or SAC alone, or in combination with IL-12 in diluted whole blood. We added IL-12 or IFN-γ to BCG, LPS or SAC, as they are known to be potent inducers of IFN-γ and IL-12. We decided to use whole blood for the study, as this method was more likely to accord with the purpose of this assay, being more reliable, and taking into account the reciprocal interactions of all the blood cells. In vitro depletion of human cells would result in difficulties inherent to the depletion techniques. Antibody-mediated depletion would cause cytokine release, whereas column depletion would cause a mechanical stress.

PBMC counting is known to vary with age, but age and gender had no significant effect on the production of IFN-γ or IL-12p40 by controls, regardless of the type of stimulation [8]. Among the 50 healthy controls there was no significant correlation between the levels of blood monocytes and IL-12p40 production (not shown). Results for cytokine production were standardized with respect to the number of PBMC and are expressed as pg/ml/10⁶ PBMC.

We also analysed the IL-12p40, IL-12p70 and IFN-γ production of healthy controls who had not been vaccinated with BCG (n = 8), five of whom had been activated with a delay due to the shipment. We observed a similar range of variation to that observed for the BCG-vaccinated healthy controls. Similar responses were found for the subgroup of non-vaccinated healthy travel controls, with slightly lower values (not shown). Thus, these results for a limited cohort of non-BCG-vaccinated healthy subjects suggest that previous BCG vaccination has no effect on the results of the assay.

Intact mitogen-triggered T cell activation

Direct stimulation of T cells via PHA showed a significantly reduced secretion of IFN-γ after 48 h in AT whole blood compared to controls. Addition of IL-12 resulted in an increase of IFN-γ secretion, but patients still produced significantly less IFN-γ than controls (Fig. 1a). As we have shown earlier, purified CD45RO⁺ T cells from patients with AT proliferated normally and secreted normal amounts of cytokines [3]. To exclude bias induced by disease-associated reduced numbers of CD45RA⁺ T cells and increased numbers of CD3^–CD16⁺CD56⁺ NK cells in AT, we normalized the amount of IFN-γ produced for the number of IFN-γ-producing T and NK cells in patients and controls. In accordance with previous results, we thereby found normal activation of AT cells with regard to IFN-γ production (Fig. 1b).

Fig. 1.

Mitogen-triggered T cell activation. Whole blood of ataxia telangiecasia patients and controls was stimulated with phytohaemagglutinin (PHA) (final dilution 1/700), or with PHA plus recombinant interleukin (IL)-12p70 (20 ng/ml). Secretion of interferon (IFN)-γ (pg/ml) was analysed after 48 h of culture. (a) IFN-γ secretion (pg/ml) after stimulation of whole blood; (b) IFN-γ secretion (pg/ml) after stimulation of whole blood per IFN-γ-producing cells × 10³.

Intact monocyte activation but impaired IFN-γ production in response to live BCG

Given the essential role of monocytes/macrophages as one of the main sources of the IFN-γ inducing cytokine IL-12, we next analysed AT monocyte function after activation with whole live bacteria (BCG). IL-12p40 (Fig. 2) and IL-12p70 (not shown) production after stimulation with BCG, and BCG plus IFN-γ was intact in patients with AT, suggesting normal function of AT monocytes/macrophages, as well as intact T cell help for IL-12 production by these cells. In spite of intact IL-12 production, indirect stimulation of T cells with BCG resulted in significantly impaired IFN-γ production after 48 h in whole blood of AT patients compared to controls. However, IFN-γ production is not completely abolished, because the addition of IL-12p70 to cultures induced a small amount of IFN-γ production in patients with AT (Fig. 3a).

Fig. 2.

Intact activation of monocytes/macrophages in response to live bacille Calmette–Guérin (BCG). Whole blood was stimulated with live BCG at a multiplicity of infection (MOI) of 20 BCG/leucocytes, or with BCG plus interferon (IFN)-γ (5000 IU/ml). Production of interleukin (IL)-12p40 (pg/ml) was analysed after 18 h of culture.

Fig. 3.

Interferon (IFN)-γ secretion in response to activation with live bacille Calmette–Guérin (BCG) and Toll-like receptor (TLR) agonists. Results are expressed as IFN-γ secretion (pg/ml) per IFN-γ-producing cells × 10³. (a) Monocyte/macrophage and T cell interaction. Whole blood was stimulated with live BCG at a multiplicity of infection (MOI) of 20 BCG/leucocytes, or with BCG plus interleukin (IL)-12p70 (20 ng/ml). Production of IFN-γ (pg/ml) was analysed after 48 h of culture. (b) Lipopolysaccharide (LPS) activation. Whole blood was stimulated with LPS from Salmonella minnesota R595 at 10 μg/ml or with LPS plus IL-12p70 (20 ng/ml). Production of IFN-γ (pg/ml) was analysed after 48 h of culture. (c) Heat-killed Staphylococcus aureus (SAC) activation. Whole blood was stimulated with SAC at 5 × 10⁶ particles/ml, or with SAC plus IL-12p70 (20 ng/ml). Production of IFN-γ (pg/ml) was analysed after 48 h of culture.

Impaired IFN-γ production in response to LPS and Staph. aureus

We then tested the response of patient’s cells to TLR ligands. Whole blood was stimulated by LPS (n = 7), which is the major component of the outer surface of Gram-negative bacteria. LPS is a potent activator of cells of the immune and inflammatory system, including macrophages, monocytes and endothelial cells [4]. Patients’ cells responded very poorly to LPS in terms of IFN-γ production (Fig. 3b). Recognition of whole microorganisms, i.e. heat-killed Staph. aureus, was also severely impaired. Levels of production of IFN-γ were considerably lower than those of healthy controls (Fig. 3c). The addition of IL-12 resulted in an increase of IFN-γ secretion, but patients still produced significantly less IFN-γ than controls. Analogous to BCG stimulation, IL-12p40 and IL-12p70 production were intact in the cells of AT patients after TLR stimulation (not shown), suggesting a functionally normal Toll-like/IL-1 receptor superfamily pathway in monocytes and macrophages.

Normal IL-10 production in response to TNF-α

Because AT cells are known to be vulnerable to diverse sources of oxidative stress [14–16], we needed to exclude bias induced by shipment when interpreting our results. To analyse functionality of cells after shipment, IL-10 production was measured. After stimulation via TNF-α, normal to slightly increased amounts of IL-10 were produced in AT cells compared to controls, ruling out an effect of delayed shipment, and a possible higher susceptibility of AT cells to shipment-associated stress (Fig. 4). To analyse the increased production of IL-10 in more detail, we measured cytokine production in AT PBMC and in isolated CD45RO⁺ T cells after stimulation via TCR/CD28 and PHA in patients with AT. We found a different pattern of cytokine expression after mitogen activation, and more markedly after TCR engagement in patients with AT (not shown, manuscript in preparation).

Fig. 4.

Normal to increased interleukin (IL)-10 secretion in response to tumour necrosis factor (TNF)-α. Whole blood was stimulated with TNF-α at 20 ng/ml. IL-10 production was measured after 48 h of culture.

Distribution of IFN-γ producing cells

We have performed intracellular IFN-γ staining previously in patients with AT and in healthy age-matched controls, and could demonstrate that the distribution of IFN-γ-producing cells in AT patients and in healthy age-matched controls is two-thirds T cells to one-third NK cells [3]. This distribution, observed after PBu₂ and ionomycin stimulation, was also confirmed by intracellular cytokine measurement after LPS stimulation of healthy controls and AT patients (Fig. 5). For analysis of IFN-γ production in all previously detailed experiments, we therefore chose to standardize the results for cytokine production with respect to the number of IFN-γ-producing T and NK cells, as patients with AT are known to have lower numbers of CD45RA⁺ T cells but increased numbers of CD3^–CD16⁺CD56⁺ NK cells (Table 1). Results are expressed as pg/ml/10⁶ IFN-γ producing cells.

Fig. 5.

Distribution of interferon (IFN)-γ producing cells. Experiments for the analysis of intracellular IFN-γ staining by fluorescence activated cell sorter (FACS) were performed after 24 h culture of peripheral blood mononuclear cells (PMBC) from healthy controls and ataxia telangiecasia patients (AT) with lipopolysaccharide at 10 μg/ml. Cell culture and intracellular staining of IFN-γ were perfomed as described in Schubert et al. [3]. Control cells (b, e) and cells from AT patients (c, f) were stained with anti-CD3 monoclonal antibody (mAb), anti-CD8 mAb and anti-IFN-γ mAb. Distribution of IFN-γ production between natural killer and T cells was analysed by gating whole lymphocyte population (a), distribution of CD4⁺ and CD8⁺ cells was analysed by gating on CD3 positive cells (d).

Discussion

Host organisms have developed a set of evolutionarily conserved receptors that can recognize pathogen-associated molecular patterns specifically, which are shared by a multitude of viral, bacterial, fungal and parasitic pathogens. The human pattern recognition TLRs are part of the TIR (Toll/IL-1 receptor) superfamily, which is involved in innate immune recognition and cellular activation in response to bacterial and viral cell components [5]. Activation of TLRs initiates multiple common signalling events, including the stimulation of pathways that lead to activation of the transcription factor NF-κB, and induction of cytokines and co-stimulatory molecules required for the activation of the adaptive immune response [4]. Production of IFN-γ is pivotal in determining the effectiveness of the immune response to pathogens, bridging both innate and adaptive immunity. IFN-γ is produced not only by CD4⁺ and CD8⁺ T cells but also by γ/δ T cells, NK cells and NK T cells, allowing rapid responses in the absence of specific T cell recognition of pathogens, i.e. in the initial stages of infection [7,17,18]. IFN-γ enhances the antigen-presenting capabilities of dendritic cells and macrophages, and promotes the killing of intracellular pathogens in macrophages. Antigen-presenting cells (APCs) regulate this process in a potent positive feedback loop by producing IL-12, which in turn induces the production of IFN-γ by NK cells and directs naive CD4⁺ T cells to differentiate into IFN-γ-producing T helper 1 (Th1) cells [19]. In addition, IFN-γ is an important part of the cancer immunosurveillance network, protecting the host against tumour growth [20,21].

However, available data in the literature are conflicting with regard to the major source of IFN-γ in response to bacterial and viral stimuli, and seem to be highly dependent on the set of experimental conditions [7]: 48 h co-cultures of T cells, antigen-pulsed antigen-presenting cells (APCs) and rIL-12 without NK cells were shown to produce no IFN-γ protein in some studies [22], whereas others [23] have shown a predominant role of T cells in IFN-γ production in response to both mycobacteria and Leishmania. Human NK cells comprise approximately 10% of peripheral blood lymphocytes. Up to 90% of human NK cells are CD56^dim, whereas a minority are CD56^bright. CD56^bright NK cells were identified recently as the major NK cytokine producer [24]. Optimal IFN-γ production by CD56^bright NK cells requires co-stimulation by IL-2 and IL-12 [22,25,26], leading to considerable IFN-γ gene expression and protein secretion at 24 h. In addition, intracellular staining for IFN-γ of CD56^dim and CD56^bright NK cells after direct activation with live BCG has been shown to lead to higher percentages of IFN-γ-producing cells among the CD56^bright NK cell subsets, peaking after 24 h [27], suggesting an ability of NK cells to interact directly with microorganisms. In our experience, distribution of IFN-γ-producing cells in AT patients and in healthy age-matched controls is two-thirds T cells to one-third NK cells [3], which prompted us to normalize results for cytokine production with respect to the number of IFN-γ- producing T and NK cells.

Previous reports dealing with IFN-γ production in patients with AT are conflicting. Defective IFN-γ production in AT PBMC was demonstrated as early as 1984 [28]. Defective macrophage–T cell co-operation as well as a selective deficiency of helper T cells was ruled out in these studies. The authors conclude that the level of the defect lies in the inappropriate transcription or translation of the IFN-γ gene [29]. In contrast, in a recent series of patients, intracellular IFN-γ was found to be normal in PMA–ionomycin-activated AT memory T cells [30]. In the present report we demonstrate intact IL-12p70 and IL-12p40 production in response to live BCG and TLR agonists LPS, which is mediated predominantly through TLR4, and SAC, which is mediated predominantly through TLR2 and TLR6 [4,5]. Thus, the initial steps of innate immunity, i.e. pattern recognition and activation of APCs, seem to be intact in AT, at least in terms of IL-12 production and subsequent priming of T cells. Nevertheless, AT cells fail to mount a significant IFN-γ response to all tested stimuli, while T cell help for IL-12 production, as well as the function of AT monocytes/macrophages in terms of IL-12 production, are intact. This finding has to be interpreted in the light of previous results from our group, reporting diminished numbers of CD45RA⁺CD4⁺ and CD45RA⁺CD8⁺ T cells, normal numbers of CD45RO⁺ T cells and increased numbers of (CD3^–CD56⁺CD16⁺)CD45RA⁺ NK cells in patients with AT [3].

As T cells and NK cells are the main sources of IFN-γ [6,8,9], the increased number of NK cells in AT patients does not seem to compensate for the lower amount of IFN-γ produced by T cells, even in the presence of adequate stimuli such as IL-12 [9]. This finding could be explained by several hypotheses: (1) CD45RA⁺ T cells are the main producers of IFN-γ. This point is difficult to evaluate, due to small numbers of CD45RA⁺ T cells in AT, and the resulting technical difficulty of cell isolation. However, this hypothesis is contrasted by reports of other groups, who found that high IFN-γ production occurs mainly in CD45RO⁺ T cells [31–33], and only a minor subpopulation of CD4⁺CD45RA⁺CD45RO^– cells, which are CD31^–, is able to produce IFN-γ [9,34]. (2) IFN-γ production is normal in CD45RO⁺ cells in AT, and reduced amounts of IFN-γ is due solely to reduced numbers of CD45RA⁺ T cells. This view is contradicted by earlier findings of our group, demonstrating reduced IFN-γ production even in isolated CD45RO⁺ T cells, and by normalizing results for the number of IFN-γ-producing cells [3]. (3) NK cells are not fully functional in patients with AT, as despite increased numbers of NK cells in AT, levels of IFN-γ produced are still inferior to controls. This hypothesis is unlikely, as we have demonstrated previously that intracellular IFN-γ after 5 h activation is reduced in CD4⁺ and CD8⁺ T cells, but normal in NK cells [3].

We therefore favour a fourth hypothesis of a different pattern of activation and cytokine production, linked to increased levels of oxidative stress and possibly impaired NF-κB activation in AT cells [35,36]. A subtle activation defect of AT lymphocytes seems likely, because we could demonstrate intact mitogen activation, contrasting reduced IFN-γ production after TLR activation and indirect stimulation with live BCG. Due to oxidative DNA damage, epigenetic changes in chromatin organization and subsequent gene silencing or activation might be impaired in AT [37]. In this regard, the finding of increased IL-10 and IL-4 production in AT cells is interesting. It might point to a different, more general, pattern of activation and cytokine production in AT regulatory T cells (Schubert et al., manuscript in preparation). Reduced production of the Th1 cytokine IFN-γ and increased production of the Th2 cytokines IL-4 and IL-10 have been described in the elderly who, as patients with AT, are more susceptible to bacterial and virus infections and neoplasias [38]. Because patients with AT display several features of accelerated ageing, such as telomere length shortening, increased sensitivity to oxidative stress, higher levels of oxidative DNA damage, as well as up-regulated apoptosis [14,15,35,39–41], reduced IFN-γ and increased IL-10 production in AT might, as in the elderly, be a marker for immunosenescence. Clearly, further experiments need to be conducted in order to dissect the cellular and molecular basis of the production of, and response to, the IL-12/IFN-γ axis in patients with AT.