Circulating γ/δ T lymphocytes from systemic sclerosis (SSc) patients display a T helper (Th) 1 polarization


Dr R. Giacomelli MD, Clinica Medica, Department of Internal Medicine, University of L'Aquila, School of Medicine via Vetoio, 67100, L'Aquila, Italy.  E-mail:


Systemic sclerosis (SSc) is a connective tissue disease in which immune system activation is evidenced by high levels of different cytokines in the sera and/or in the supernatants of cultured peripheral blood mononuclear cells (PBMC) and by the presence of specific autoantibodies. γ/δ T cells accumulate in the lung and the skin of SSc patients suggesting their potential role in the development and maintenance of the disease. The aim of this study was to assess cytokine production and cytotoxic activity of circulating γ/δ T lymphocytes obtained from SSc patients and to evaluate their potential role during this disorder. Our results showed that both the proportion and the absolute number of IFN-γγ/δ-producing cells (i.e. displaying a Th1 polarization) in SSc was significantly higher than either the proportion and the absolute number of IL-4 γ/δ-producing cells in SSc or the proportion and the absolute number of IFN-γγ/δ-producing cells in healthy controls (P < 0·05 for both groups). Furthermore, the cytotoxic activity of enriched γ/δ T cells was significantly increased in SSc patients compared with controls. The results concerning the Vδ1+ T cell subset paralleled those of total γ/δ T lymphocytes. In contrast, α/β T cells from SSc and control subjects displayed Th2 cytokine production. All these findings were independent of both disease subset and clinical status. Our data demonstrate that, although SSc is generally considered a Th2 autoimmune disease, Th1 polarization of γ/δ T cells and an increase in their cytotoxic activity is observed in SSc, suggesting that γ/δ T cells could have a relatively autonomous role in the pathogenesis in this disease.


Systemic sclerosis (SSc) is a connective tissue disease; there are two principal clinical subsets: limited and diffuse, which can be differentiated by the extent of skin involvement, presence of specific autoantibodies and clinical course [1]. There is growing evidence that this disorder is immunologically triggered. Perivascular mononuclear cell infiltrations via the release of soluble mediators may lead to the two main disease features: vascular tone control dysfunction (Raynaud's phenomenon) and skin sclerosis [2]. Immune system activation in SSc is further supported by the presence of high levels of interleukin (IL) 2, soluble IL2 receptor (sIL2r), IL4, IL6 and tumour necrosis factor-α (TNF-α) in the sera of SSc patients and/or in the supernatants of their cultured peripheral blood mononuclear cells (PBMC) [3–5] and by the presence of hypergammaglobulinaemia associated with the production of autoantibodies [6].

Separate groups have reported an accumulation of activated T lymphocytes, expressing a selective Vδ chain gene expansion of the γ/δ T cell receptor (TCR), in the lung [7] and skin [8] of SSc patients. These T cells display adhesion molecules [7,8] and kill endothelial cells (EC) [9], thus suggesting their potential role in the development and maintenance of the disease. γ/δ T lymphocytes have antigen receptors encoded by a separate set of genes from those used by α/β T cells; they are usually present in two different subsets, the majority expressing the Vδ2 chain and the minority expressing the Vδ1 chain. γ/δ T cells account for approximately 1–5% of total circulating T cells but are present in much larger proportions at epithelial surfaces, at sites of inflammation and following infection with certain microorganisms [10]. Despite intensive study, their exact role in the immune response is poorly understood. However, they show both major histocompatibility complex (MHC)-restricted and MHC-unrestricted cytotoxic activity [11]. Following activation, they secrete appropriate cytokines [12] and contribute to the specific milieu in which the differentiation of CD4+ T cells occurs, thus leading to primary immune responses. In fact, following antigen priming, human CD4+ T cells differentiate into three subsets, according to their cytokine secretion profiles: T helper (Th) 1 cells secreting mainly IL2, interferon-γ (IFN-γ) and linfotoxin (TNF-β); Th2 cells secreting mainly IL4, IL5, IL6, IL10 and IL13, and a third subset named Th0, which produces and secretes cytokines of both subsets (mainly IFN-γ and IL4) [13,14].

In recent years, several studies have shown that an imbalance in Th1/Th2 responses may have an impact on the pathogenesis of several pathological conditions, including autoimmune diseases [15]. Recently we reported a possible involvement of Th2 responses in SSc [16] and Th2 cells seem to be implicated in several findings observed in SSc. Briefly, high levels of IL4 and low levels of IFN-γ can be observed in the disease [4,6] and these two cytokines are antagonistically involved in the modulation of connective tissue synthesis [17,18]. Furthermore, during chronic graft versus host disease (cGVHD), which is considered a natural model of SSc [19], alloreactive T cells produce IL4 and IL10 but not IL2 or IFN-γ[20].

The aim of this study was to assess, at the single cell level using two-colour flow cytometry, whether peripheral γ/δ T lymphocytes obtained from SSc patients could differentially produce cytokines known to induce Th1 and/or Th2 responses in order to evaluate their potential role during the autoimmune response of this disorder.

Patients and methods

Patients and controls

Thirteen patients (12 women and one man), aged 29–65 years (mean age 56·4 years), with SSc diagnosed according to the American College of Rheumatology (formerly the American Rheumatism Association) criteria were enrolled, after informed consent and institutional approval, in the study [21]. The disease in seven patients was classified as diffuse and in six as limited by accepted criteria [1]. Skin assessment was performed utilizing a modified Rodnan (m-Rodnan) total skin thickness score technique [22]. The patients were evaluated for involvement of the lungs (pulmunary function tests, chest HRTC), oesophagus (manometry, barium swallow radiography), kidneys (blood urea, serum creatinine levels and urinalysis) and heart (electrocardiography, 2-mode echocardiography). Eleven patients had never been treated with cytotoxic drugs or biological response-modifying agents during the 12 months preceding the study; two patients who were receving steroids (prednisone 10 mg/day or less) discontinued the treatment 30 days before immunological assays. Twelve gender-, race- and age-matched healthy volunteers were enrolled as controls.

Blood samples

Heparinized venous blood was obtained from patients and controls and peripheral blood mononuclear cells (PBMC) were isolated by centrifugation on Ficoll-Hypaque gradients. Their viability, as assessed by trypan blue dye exclusion, was routinely > 95%.

Isolation of α/β and γ/δ T cells

After depletion of plastic-adherent cells by incubation for 90 min in plastic dishes at 37°C in a humidified 5% CO2 atmosphere in RPMI 1640 with 10% FCS, specific T cell subsets were isolated from PBMC using an immunomagnetic method [23]. Briefly, different suspensions of PBMC, at 2 × 106/ml concentration, were cooled to 4°C and incubated with antitotal TCR α/β or antitotal TCR γ/δ (TCRδ1) or anti-δTCS1 (recognizing T lymphocytes bearing the Vδ1 chain of TCR) antibodies (all from T Cells Sciences, Canbridge, MA, USA) for 45 min on ice. After several washes with PBS, cells were incubated with M450 particle Dynabeads precoated with sheep anti mouse IgG (Dynal, Oslo, Norway). Rosette-forming cells were isolated by placing the tube on a flat magnet and washing the isolated cells three times using the same magnetic technique. The obtained cells were > 95% pure for the intended surface marker, as determined by flow cytometry.

Intracellular analysis of cytokine production

Intracellular cytokine staining was used to determine the cytokine production in freshly isolated α/β and γ/δ T cells at the single-cell level. To induce activation and intracellular accumulation of synthesized proteins, cells were incubated for 2 h with 25 µg phorbol myristate acetate (PMA) and 1 µg ionomycin, washed with PBS-2% FCS and then permeabilized with 0·5% saponin (Sigma Chemical Co.) in PBS for 30 min at room temperature. R-phycoeritrin (PE) conjugated anti-IL4 and fluorescein isothiocyanate (FITC) conjugated anti-IFNγ nonoclonal antibodies (used at the manufacturer's recommended concentration, Becton Dickinson, Mountain View, CA, USA) were added to the permeabilized cells and allowed to bind for 30 min. Cells were then washed with PBS-0·5% saponin and finally with PBS-2% FCS to allow membrane closure. Negative control samples were incubated with irrelevant, isotype-matched antibodies in parallel with the experimental samples. To assess possible non-specific binding of the conjugated monoclonal antibodies, some experiments were repeated after preincubation for 30 min with both anti-IL4 and anti-IFNγ unlabelled nonoclonal antibodies and, after several washes with PBS, cells were finally incubated with R-phycoeritrin (PE) conjugated anti-IL4 and fluorescein isothiocyanate (FITC) conjugated anti-IFNγ nonoclonal antibodies. All samples were analysed using a FACScan flow cytometer (Becton Dickinson, Mountain View, CA, USA).

Allogeneic stimulation and expansion of γ/δ T cells

Briefly, γ/δ T cells (5 × 105/ml) of six consecutive patients (four diffuse and two limited) and six controls were co-cultured with irradiated (3000 rad) RPMI 8866 cells (1 × 105/ml) at 37°C in a humified 5% CO2 atmosphere for 10 days as described previously [24]. On day 10 the population was routinely composed of 90% TCRδ1+ (total γ/δ T cells), CD16+, CD14cells, as determined by flow cytometry.

Cytotoxic activity

Cytotoxic assays were performed by incubating serial dilutions of effector cells with 5 × 10351Cr-labelled (Na251CrO4; New England Nuclear Dreiech, Germany) K562 target cells, in triplicate wells of round-bottomed microtitre plates (Sterilin Teddington, Middlesex, UK) in a final volume of 0·2 ml. Stimulated and expanded γ/δ T cells were used as effector cells at different effector/target (E/T) ratios. After 4 h of incubation, the plates were centrifuged and 0·1 ml supernatant was removed and counted. The percentage of specific 51Cr release was calculated as follows: 100 × (experimental release – spontaneous release/maximum release – spontaneous release).

Statistical analysis

The non-parametric Wilcoxon two-tailed test (normal approximation) and the rank correlation test were used when appropriate for the statistical analysis of the data.


Unless stated otherwise, results are expressed as (median, range).

Intracellular T cell cytokines

Our results for circulating γ/δ T cells of SSc patients showed that the proportion (49·2%, 5·7–65%) and the absolute number (75, 11–92) of IFN-γ-producing cells were significantly higher than both the corresponding proportion (16·4%, 8·1–27·5%) and the absolute number (20, 9–35) of IL-4 producing cells (P < 0·05). Furthermore, the numbers of IFN-γ producing cells in SSc patients were significantly increased when compared to the proportion (23·4%, 4–41%) and the absolute number (31, 8–68) of IFN-γ-producing cells in normal controls (P < 0·05) (Fig. 1). A trend toward increased amounts of γ/δ T cells expressing a Th0 phenoype was observed in SSc patients (proportion 6·2%, 2–13%; absolute number 8, 3–18) compared with controls (proportion 2·2%, 0–5·6%; absolute number 1, 0–12), although this difference did not reach statistical significance. The results concerning the Vδ1+ T cell subset parallels those observed among total γ/δ T lymphocytes (Fig. 2). In this subset, the IFN-γ-producing cells of SSc patients were the large majority (proportion 72·2%, 28–96·9%; absolute number 65, 24–73) compared with the IL-4-producing cells of the same patients (proportion 20%, 5–38,4%; absolute number 15, 4–28) and the IFN-γ-producing cells in normal controls (proportion 29·9%, 15–56%; absolute number 5, 2–9) (P < 0·05 for both groups).

Figure 1.

 Percentages of α/β and γ/δ T cells in patients (P) and controls (C) producing IFN-γ and IL4. Bars represent medians (°P < 0·05, *P < 0·001 Wilcoxon's two-tailed test). ○ IL4; ● IFN-γ.

Figure 2.

 Percentages of Vδ1+ T cells in patients (P) and controls (C) producing IFN-γ and IL4. Bars represent medians (°P < 0·05, Wilcoxon's two-tailed test). ○, IL4; ●, IFN-γ.

In contrast, for total α/β T cells, the number of IL-4-producing cells (Th2 phenotype) (proportion 41·7%, 7·9–61·4%; absolute number 700, 60–870) was significantly higher than both the number of IFN-γ-producing cells (proportion 29·4%, 11·6–37·1%; absolute number 300, 125–359) in SSc patients and IL-4-producing cells in normal controls (proportion 17·2%, 3·3–29·6%; absolute number 350, 62–635) (P < 0·05 and P < 0·001, respectively) (Fig. 1). Lower levels of α/β T cells expressing a Th0 phenotype were observed in SSc patients (proportion 3·7%, 1·3–7·4%; absolute number 4, 1–9) than in controls (proportion 5·1%, 0–14·3%, absolute number 5, 0–15), although this difference did not reach statistical significance.

γ/δ T cell cyotoxicity assay

As shown in Fig. 3, the cytotoxic activity of enriched γ/δ T cells of six consecutive SSc patients was significantly increased compared with the controls at every E : T ratio studied. These findings were independent of both the disease subset and the clinical status of the patients. No correlation between the percentage of Th1+ γ/δ T cells and the levels of cytotoxic activity was observed.

Figure 3.

Figure 3.

 Cytotoxic activity of serial dilutions of highly enriched γ/δ T cells as effector cells (E), from six SSc patients (●) and six control subjects (○) measured in vitro against 5 × 103 K562 cells (human erythroleukaemia) as target cells, at different effector : target (E : T) ratios, in a 4-H chromium release test. The points represent the means of triplicate determinations (s.d. < 5%). The differences between the two groups were significant (P < 0·01) at all E : T ratios as evaluated by the Wilcoxon two-tailed test.

Clinical findings

All patients classified as limited disease were anticentromere-positive. Lung involvement was detected in four of six patients and oesophageal involvement in five of six patients. One patient displayed intestinal involvement (steatorrhoea with increased faecal fat excretion and weight loss).

In the group of patients with diffuse disease, six were antitopoisomerase I-positive, all had lung and oesophageal involvement, two had heart involvement and two had kidney involvement.

No significant correlations were detected between the clinical features of our SSc patients and the values of circulating Th1 or Th2 γ/δ T cells.


Abundant evidence suggests a critical role for activated T cells in the pathogenesis of SSc. In fact, several cytokines secreted by T cells may contribute to modulate fibrosis and promote vascular damage [2–6]. Although the initial trigger of immune activation is unknown, the prominent perivascular location of immune cells suggests a vascular origin for this trigger and the endothelium as the potential target of the immune aggression. We reported recently a selected expansion of activated, adhesive Vδ1+ T cells in the skin and blood of SSc patients [8], suggesting the involvement of this subset in the disease. In this work we assessed, in vitro, some functional activities of γ/δ T cells, such as their cytokine profile and cytotoxic activity. Furthermore, in the same patients we also evaluated α/β T cell polarization in order to define whether these two separate subsets display a similar cytokine repertoire.

To our knowledge, no study has reported on the cytokine pattern of the γ/δ T cell subset in SSc and similar studies in humans are limited. γ/δ T cells in SSc patients exhibit a Th1 phenotype following stimulation with PMA and ionomycin. A slight increase in γ/δ T cell numbers with an intermediate phenotype, representing Th0 cells, was observed. The increase in IFN-γ-producing cells observed in our patients is in agreement with results obtained in patients with tuberculosis [25] and rheumatoid arthritis (RA) [26], and mirrors recent data showing that γ/δ T cells of animal models constitutively display a non-random pattern of Th1 cytokine production [27]. Furthermore, activation of human γ/δ T cells by non-peptide antigens induces an early Th1 response which is not reversed by the presence of IL4 [28]. In our study, the major subset of γ/δ T cells displayed the Vδ1 chain and clearly showed Th1 polarization. Although in normal donors the Th1 polarized subset seems to display the Vδ2 chain [29], Vδ1+ T cells with a Th1 phenotype are increased in some clinical conditions, such as HIV infections [30] and RA [26]. Taken together, these data suggest that, although the reason for this highly conserved type-1 response remains unclear, the possibility that an innate property of γ/δ T cells or, alternatively, that their priming and activation in vivo may be triggered by some viral infections [31] should be considered.

Despite the individual wide range of Th1 γ/δ-values, no differences in disease subset or other clinical properties such as autoantibody pattern, skin score or organ involvement have been observed in our patients and no correlation can be shown between clinical data and Th1 γ/δ T cell numbers. This finding confirms our previous data [8] on γ/δ T cells in SSc and suggests that these cells might play a pathogenetic role which is strongly retained and independent of disease status.

In contrast, α/β T cells displayed a Th2 phenotype, confirming previous reports by ourselves and other groups showing, in SSc patients: increased IL4 production by both activated PBMC and cells of the bronchoalveolar lavage fluid; increased IL4 serum levels; and increased expression of CD30 (a Th2 marker) in blood and skin [3,4,16,32,33]. The central role of IL4 in the pathophysiology of SSc is supported by the well-known observations that this cytokine induces: the collapse of vimentin intermediate filaments in endothelial cells [34], fibroblast activation and extracellular matrix production [35], as well as polyclonal B cell activation and autoantibody production [36].

Our results showed that γ/δ T cells of SSc patients displayed potent cytotoxic activity toward K562 cells at every E : T ratio. Human γ/δ T cells are capable of mediating non-MHC-restricted cytotoxicity toward a variety of target cell lines, although the precise steps concerning antigen interaction with the γ/δ receptor are not known [37], and these cells share many features in common with natural killer (NK) cells, including high levels of spontaneous cytotoxicity and a bias toward IFN-γ secretion [38]. Several studies have been performed in order to assess the role of cytoxicity in SSc [39–41]. NK activity and NK cell numbers are increased in SSc patients [39], the Vδ1+ T cell subset of SSc patients display the accessory molecule CD16 [8], and it has been shown recently that γ/δ T cells recognize endothelial cells, proliferate and display a cytotoxic activity toward these cells, which can be inhibited by antiγ/δ TCR antibodies [9].

In this study, the increase in γ/δ T cell cytotoxic activity was observed after stimulation with RPMI 8866 cells for 10 days. One possible disadvantage of this method is a generalized allogeneic stimulation of γ/δ T cells, with the selection of specific cytotoxic clones; none the less, in view of the low number of circulating γ/δ T cells, an in vitro enrichment procedure is essential to obtain sufficient numbers of cells to execute the studies. Furthermore, this method was selected after considering recent reports which suggest that an allogeneic process (the transfer of allo-immune competent cells occurring across the decidua/trophoblast interface during pregnancy) rather than a breakdown in self-tolerance may play a pathogenetic role in SSc [42]. It has been shown that the numbers of Vδ1 T cells increase significantly in the early pregnancy decidua, recognize trophoblast allo-antigens, secrete Th2 cytokines and suppress the maternal antifetal imune response [43]. To support pregnancy, the decidual microenvironment might induce a Th2 activity of host Vδ1+ T cells. In contrast, the presence of fetal cells in the skin of SSc patients or alternatively an activating stimulus, such as some viral infections [31], might induce a shift toward Th1 functions of Vδ1+ T cells, activating their cytotoxic machinery and subsequently activating a cytotoxic acquired response.

In conclusion, our data demonstrate that although SSc is generally considered a Th2 autoimmune disease, a Th1 polarization of γ/δ T cells and an increase in their cytotoxic activity, under the experimental conditions selected for the current study, can be observed. It must be pointed out that γ/δ T cells seem to use a variant of the basic Th1 cytokine secretion programme, in which the production of IL3 and colony-stimulating factors is disfavoured [27]. Taken together, these findings suggest that these cells, rather than playing an early modulating role in the immune response in which the production of recruiting factors would be important, might have a relatively autonomous role and their ultimate activities can be differentially modulated by several mechanisms, such as the nature of the antigen and the organs involved. Further experiments are in progress in our laboratory in order to verify this hypothesis.


The authors wish to thank Dr E. C. LeRoy and Dr B. Kahaleh for helpful advice and review of the manuscript.