It could be argued that

It could be argued that the lack of association between pre-diagnostic Treg/tTL ratios and CVD risk in the present study is due to the fact that plaques represent a different immunological compartment than blood (Grivel et al., 2011), and it remains to be determined whether Treg levels in blood adequately reflect sustained immune responses in the arterial wall. Actually, a correlation of circulating Tregs with the carotid intima-media thickness, a proxy of subclinical atherosclerosis, could not be found in both healthy populations (Wigren et al., 2012; Ammirati et al., 2010) and ACS patients (Ammirati et al., 2010). Nevertheless, it is increasingly acknowledged that antigen-specific atherogenic T cell responses are initiated in secondary lymphoid organs by presentation of plaque cytokine receptor (Ammirati et al., 2015) and a systemic nature of immune reactions related to atherosclerosis is further supported by the fact that both local and systemic inflammation affect early steps of atherogenesis and CVD (Hansson, 2005). In a previous study, we have also shown that higher pre-diagnostic Treg/tTL ratios in blood are associated with the risk of lung, colorectal and ER-Negative breast cancer (Barth et al., 2015). These observations support the relevance of Tregs in the periphery to local immune homeostasis and tolerance. Another potentially important result of this study is that CVD risk across quartiles of Treg to tTL ratios was attenuated on adjustment for smoking, which is in line with previous data demonstrating that both acute and cumulative smoking exposure positively correlates with the Treg/tTL ratio (Wiencke et al., 2012). Although largely unknown, a smoking-associated Treg increase in blood may either reflect a more general state of suppressed immune and inflammatory responses (Stampfli and Anderson, 2009; Shiels et al., 2014) or relates to facilitated mobilization and migration of Tregs to tissue sites affected by smoke exposure to dampen local inflammation (Ritter et al., 2005). In fact, our subgroup analyses by smoking status did point to some heterogeneity in the associations between Treg frequencies and MI but not stroke risk. With respect to MI, a significant inverse association with Treg frequencies was only observed in former smokers, while there were no significant associations in current and never smokers. Admittedly, our subgroup samples were rather small and possible smoking-related influences on the relationship between peripheral Treg variability and MI risk may require further study in a larger sample. The strengths of our study include its prospective design, the large sample size for our main analyses, and the use of epigenetic assays, which enable the quantification of immune cells in buffy coat samples after long-term storage. There are also some limitations to this study that have to be considered. First, the generalizability of the association between cellular immune markers and CVD development is limited to some extent because the EPIC-Heidelberg cohort represents a population with a higher socio-economic status and more favorable lifestyle factor profile compared to populations from other regions in Germany (Boeing et al., 1999). This overrepresentation of health-conscious individuals could have become even more pronounced by restriction of our investigation to non-diabetics. With regard to the quantification of overall Tregs, it must be noted that we could not address the issue of heterogeneity within the Treg compartment in our study and that different T cell subsets may have distinct properties in CVD development. Our study was set up to analyze Treg-mediated immune tolerance as a global phenomenon, but studies on more specific functions of immune cell subsets are clearly needed. Another general limitation of long-term observational studies is that exposures are measured only at a single point in time. However, in a reproducibility study of 100 healthy EPIC-Heidelberg participants with repeated measurements, we have previously shown a moderate-high correlation between individual\'s peripheral Treg/tTL ratio over time periods of one year (r=0.73) and up to fifteen years (r=0.53) (Barth et al., 2015). Thus, a single quantification may provide a good proxy for long-term FOXP3+ Treg values and supports the notion that relative numbers of these cells in the periphery are kept in a homeostatic state throughout life (Liston and Gray, 2014; Fessler et al., 2013).