Analysis of long term CD4+CD25highCD127- T-reg cells kinetics in peripheral blood of lung transplant recipients

Lung transplant (Tx) is the only therapy for patients with end-stage respiratory failure, but despite the use of potent immunosuppressive protocols, the long-term survival of lung graft is hampered by the occurrence of chronic lung allograft dysfunction that occurs in nearly 50% of patients by the 5th post-Tx year [1]. According to recent classification, there are two major clinical phenotypes of chronic irreversible allograft dysfunction (CLAD): the so called obstructive form: bronchiolitis obliterans syndrome (BOS), (nearly 70% of cases) [2], and the newly described restrictive allograft syndrome (RAS) [3, 4]. CLAD is considered a complex multifactorial process [5], which leads to chronic airway inflammation and tissue injuries and ultimates in a fibro-reparative response. This can involve either the airway lumen in BOS, or the interstitial and airways in RAS [6]. Although recent evidence ascribes an important pathogenic role in CLAD to aspecific inflammatory mechanisms [7‐10], the importance of specific immunity can’t be denied. Several previous studies support a crucial role for allo and auto-reactive T cells in BOS pathogenesis [11‐14]. T cell response as well as the production of allo or auto-antibodies have been described, and are dependent on the migration of antigen-presenting cells in secondary lymphoid organs and on the direct stimulation of T cells within graft [15, 16]. All of this experimental and clinical evidence strongly points out the difficulties in achieving immunologic tolerance, either central or peripheral, of lung graft [17‐19].

Since 1995, when Sakaguchi described a CD4⁺CD25⁺T cell subset displaying regulatory properties, several subsets of T-reg cells have been described and classified as natural T-reg, thymus-derived and naturally committed to immunoregulation, and inducible T-reg, generated in the periphery during immune response. So far the best characterised T-reg population is the thymus-derived CD4 subset, constitutively expressing CD25 (at a high rate) and FoxP3, a transcription regulator which controls the maturation and function of T-reg cells. Although a specific and exclusive marker of T-reg activity has not yet been identified, the diminished surface expression of CD127, as well as the expression of CD39 or CD152, are considered characteristic of CD4⁺CD25⁺ T-reg cells. [20, 21]These molecules can be expressed, even if transiently, as Foxp3, by other cell types including recently activated effector T cells. However, experimental evidence suggests that most of the Foxp3+ T-reg cells are within the CD4⁺CD25^highCD127⁻ cell population, so that sorting of this subset is actually considered the best method for the isolation and in vitro amplification of these cells [22].

The role of CD4⁺CD25^highCD127⁻ T-reg cells in Tx acceptance has been studied in experimental models and also in clinical settings. Animal studies demonstrated that CD4⁺CD25⁺ Foxp3+ T-reg cells are expanded in tolerized animals, that tolerance can be infectiously transferred to naïve animals, and appears to be a specific and localized phenomenon. These data have been partially confirmed in kidney and liver Tx recipients, showing a positive correlation between graft survival and the number of circulating CD4⁺CD25⁺ T-reg cells, as well as a correlation between their peripheral fluctuation and the occurrence of acute and chronic rejection [18, 23, 24].

As for lung Tx, evidence is limited and somewhat contradictory [25]. In previous cross- sectional studies, we showed that lung Tx recipients with BOS had significantly lower peripheral CD4⁺CD25^high T-reg cells than clinically stable lung recipients, and demonstrated their functional regulatory profile, in vitro [26, 27]but subsequesnt studies failed to demonstrate a correlation between T-reg cell counts and long term lung Tx outcome [28‐30].

To date, however, little is known about the long term evolution of peripheral CD4⁺CD25^highCD127⁻ regulatory T cells in lung Tx. The present study aims at analyzing their long-term kinetic on a cohort of lung recipients and at testing its association with several clinical and pharmacological variables (CLAD, treatment, infections, kidney dysfunction or neoplasia).

On the basis of our results we can confirm and extend our previous observation on the role of CD4 + CD25high T-reg cells in lung graft acceptance/rejection [27]. To our knowledge, this is the first longitudinal study reporting the long-term kinetics of peripheral CD4⁺CD25^highCD127⁻ T-reg cells in lung recipients, and the first observation, that variation of peripheral T-reg cell counts can predict CLAD onset/progression.

Although, during the study period, a methodological evolution in the identification of T-reg cells subsets occurred, we clearly demonstrated that CD4⁺CD25^highCD127⁻ subset is significantly enriched with FOXP3+ cells, thus inferring that it is endowed with regulatory functions. Based on these observations, we have referred to these cells as CD4⁺CD25^highCD127⁻ T-reg cells.

Experimental evidence in animals have strongly suggested that lung graft acceptance is associated to an intra-graft [16, 38] and peripheral [39] T-reg cell expansion and that adoptive transfer of “in vitro” generated and expanded T-reg cells attenuates airway obliteration in a rat OB model [40].

Conversely, human studies have provided, to date, sub-optimal evidence on the protective role of T-reg cells towards chronic graft rejection. Most studies, in fact, have explored peripheral CD4 + CD25+ T-reg cell counts early after Tx, suggesting that a low number of these effectors might predict subclinical or clinical acute kidney rejection [41, 42], and can be significantly affected by the type of immunosuppressive regimen [43, 44]. A few studies are available on lung Tx recipients with conflicting results. Neujahr and colleagues found higher BAL CD4 + CD25high cell counts in patients with biopsy-proven AR, but failed to find a correlation with peripheral blood findings [45]. On the contrary, lower CD4 + CD25high T cell counts were detected both at peripheral and BAL level of BOS patients with respect to stable lung recipients [27, 46].Moreover, an increase in BAL CCR7(+) T-reg cell percentage was found to correlate to a reduced risk of BOS by Gregson and colleagues [47]. Krustrup and colleagues, analyzing CD4 + CD25 + FOXP3+ cell at tissue level, found higher frequencies during acute rejection episodes [29], but failed to demonstrate an association between these cells and the risk of CLAD occurrence [30]. We believe that prediction studies on peripheral T-reg in Tx recipients are invalidated by the high fluctuation of blood-T-reg cell in time, possibly in relation to a number of confounding variables including drugs, infectious complications, neoplasia, kidney failure and ECP treatment.

Thanks to the large number of determinations that we included in this long term study, we took into account this variability in the analysis, thus we succeeded in demonstrating that: 1) peripheral counts of CD4⁺CD25^highCD127⁻ T-reg cells significantly decrease in CLAD patients; 2) the degree of their decrease is associated with the severity of BOS and, most noteworthy; 3) CLAD onset or GRADE is significantly associated to mean T-reg cell counts in the previous trimester.

Finally, we also tested the association of CD4⁺CD25^highCD127⁻ T-reg cell counts with a number of clinical variables, including also infections and different immunosuppressive drugs. Interestingly, only azathioprine and ECP were found to be significantly associated to them at the univariate analysis, while, in the multivariate model, only the ECP effect was significant. However, the ECP effect was confounded on CLAD and CLAD severity, and is therefore to be considered irrelevant from the clinical point of view.

Interestingly, unlike previous evidence in literature [26, 48]we could not detect any significant variation of T-reg cell counts with respect to other specific immunosuppressive drugs such as cyclosporine A, tacrolimus or everolimus. The effect of maintenance drugs on T-reg cell counts has been poorly evaluated in humans, the role of CsA and tacrolimus has been studied in mice with controversial results [44] while everolimus has been shown to enhance the number of peripheral T-reg cells in liver recipients following a tacrolimus to everolimus conversion [48]. However, it was recently observed that sirolimus, another mTor inhibitor, did not expand peripheral T-reg cells in a cohort of de novo kidney recipients [49], in analogy to our present observation. Finally, in the present study we could not assess the possible role of lympho-depleting strategies, since induction treatment is not performed at our center.

We must acknowledge a number of limitations. First, although larger than other cohorts, our sample is not sufficiently large to identify small associations, especially in some subgroups. Second, patients started FU at different times from Tx; although this was taken into account in the analysis, we cannot be certain that bias would not impact on results. Third, although data were collected prospectively, this study was designed retrospectively as an exploratory analysis; therefore, patients were sampled at irregularly spaced times, and sometimes upon clinical basis, so there might be unobserved variables that might have affected results. In addition, we performed all the analysis using both the absolute number and the percentage of T-reg in peripheral blood, but, although both had the same trends over time and were significantly correlated with CLAD, we decided to show only the results concerning absolute number analysis because they had the best correlation with the analyzed variables. Finally, a relevant issue, from the biological point of view, is that graft tolerance is a results of a balance between regulatory and effector clones. In this study we couldn’t take into account the weight of allospecific effectors at different time points.

In conclusion, given all the above limitations, we could find an association between T-reg cell counts and CLAD onset/progression, including both BOS and RAS phenotype, thus confirming animal observations on the protective role of T-reg cells with respect to CLAD.