The main findings from the present study are that human Treg cells exhibit markedly increased sensitivity to H2O2-induced ROS production and cell death compared to Tconv cells in both CCS patients and healthy controls. Treg cells were also more sensitive to spontaneous cell death when cultured in medium overnight and to cell death induced by STS, a potent inducer of apoptosis. Hitherto, the impact of oxidative stress on human Treg cells has been far from clarified. The existing literature is both sparse and inconsistent. Mor et al  reported that the number of Treg cells was reduced to a considerably larger extent than the number of Tconv cells after in vitro incubation with oxidized LDL. They further showed that this effect was attenuated in the presence of caspase inhibitor suggesting that apoptosis contributed to the loss of cells. Others demonstrated that Treg cells were more prone to apoptosis than Tconv cells when they were cultured in medium only [23, 24]. On the other hand, Mougiakakos et al  used a protocol that was similar to ours, i.e. incubation with 5, 10 or 20 µM H2O2 for 18 h, but obtained completely opposite results. They found that Treg cells were significantly more resistant to H2O2-induced cell death compared to Tconv cells. Moreover, they found that naïve Treg cells were more resistant than memory Treg cells. One possible explanation for the contradictory results may be the choice of study subjects. Mougiakakos et al  used cells from a few healthy donors of unknown age while we used cells from sick and elderly subjects. Ageing and cardiovascular disease are both linked to elevated oxidative stress. We previously showed that naïve Treg cells constituted only 15% of all Treg cells in elderly CAD patients as opposed to 25% in age-matched controls, and also that the function of naïve Treg cells was impaired in patients . Differences in naïve Treg cell pool size and function as well as individual variations in oxidative status may thus contribute to the disparity in results between our study and Mougiakakos et al .
In an attempt to assess intrinsic ability of these cells to counteract oxidative stress, we measured the mRNA expression of several main antioxidant genes, which all contribute to inactivate ROS, from freshly isolated Treg and Tconv cells from both patients and controls. Collectively, the expression of catalase was significantly higher in Treg cells than in Tconv cells and so was the expression of OXR1, a protein that has been described as a cellular oxidative stress sensor regulating the expression of several antioxidant enzymes . Also, mRNA levels of the two main proteins, Trx and TrxR1, in the antioxidant Trx system were markedly upregulated in Treg cells. Our findings partly agree with the previous study by Mougiakakos et al , who reported that Trx expression was higher in Treg cells compared with Tconv cells while, on the other hand, catalase expression did not differ in their small study group of volunteers. An earlier genomic and proteomic screening study examined the H2O2-induced gene and protein expression in human skin fibroblasts and reported that TrxR1 was the only oxidation-related candidate with elevated levels at both the mRNA and protein level, the latter measured in cell lysates . Another study by Söderberg et al  provided evidence that TrxR1 was secreted by human PBMCs upon inflammatory stimulation. Here, we were able to show that TrxR1 was secreted into the cell supernatant to a greater extent by Treg cells than by Tconv cells upon H2O2 treatment. Altogether, our findings indicate that Treg cells have a higher endogenous antioxidant capacity than Tconv cells in both CCS patients and controls.
Several studies have reported reduced levels of Treg cells in peripheral blood of patients with ACS [10, 16–18]. It has also been proposed that oxidative stress is involved in Treg cell depletion. Mor et al  showed that Treg cells from ACS patients were more sensitive to oxidized LDL than Treg cells from CCS patients or patients with normal coronary angiograms. A later study by Zhang et al  showed that the spontaneous apoptosis of Treg cells was pronounced in ACS patients and further demonstrated that oxidized LDL induced apoptosis of human Treg cells in vitro, yet without comparing patients and controls. Recently, we found that the numerical and functional Treg cell deficit in ACS patients was not merely a transient phenomenon but that it remained after clinical stabilization . In the present study, we therefore focused on the potential role of oxidative stress in the depletion of Treg cells in CCS patients. However, there was no proof that Treg cells from patients were more prone to H2O2-induced cell death than Treg cells from controls, when assessed ex vivo. Moreover, the expression of antioxidant genes in freshly isolated Treg cells from patients and controls indicated that Treg cells in CCS patients and controls had similar levels of endogenous antioxidant enzymes in vivo. However, the significantly lower intrinsic TAC levels in PBMC from CCS patients pointed toward an oxidant-antioxidant imbalance implicating the possible impact of exogenous environmental factors, such as sedentary lifestyle and unhealthy diet. The TAC assay is a copper-based assay which measures a large range of lipophilic or thiol-based antioxidants and antioxidant macromolecules but not antioxidant enzymes discussed above. [27, 28] Therefore, given their high susceptibility to oxidative stress, it is reasonable to assume that Treg cell numbers are affected by the prooxidant state in CCS patients.
Interestingly, our results indicate that Tconv cells from CCS patients were affected to a greater extent by oxidative stress than Tconv cells from controls. Tconv cells from CCS patients were more prone to undergo cell death when treated with H2O2 or STS and they also expressed significantly higher levels of TrxR1. It is well documented that the proatherogenic T cell response in atherosclerotic lesions is reflected in peripheral blood . In ACS patients, systemic T cell activation is associated with plaque instability and thrombus formation but there is also consistent evidence that the T cell activation persists after clinical stabilization [9–11]. T cell activation is accompanied by the release of ROS which in turn leads to an upregulation of endogenous antioxidants . Moreover, the induction of Trx and TrxR1 in T cells upon activation was recently suggested to be a critical pathway controlling T cell activation and expansion . We believe that the increased susceptibility to oxidative stress in Tconv cells from patients reflects a CD4+ T cell activation that remains in CCS patients despite clinical stability and medical treatment.
A couple of limitations should be considered in our study. One limitation is the limited number of patients and controls per experimental set-up, permitting only cautious conclusions about differences between the two groups. The yield levels of Treg cells in the subjects, particularly in the CCS patients, were not high enough to allow ex vivo experiments and gene expression analyses to be carried out in the same individuals, nor to perform functional assays. Gene expression analyses were therefore performed in a separate group of subjects, 10 patients and 10 controls, though with similar characteristics as the other subjects. Another potential limitation is that the ex vivo model, where isolated fractions of T cell subsets are exposed to increasing doses of H2O2 for a relatively short period of time, may not reflect chronic inflammation.