Life after the thymus: CD31⁺ and CD31⁻ human naive CD4⁺ T-cell subsets

The term “naive T cell” derives from the assumption that CD4⁺ T cells are antigen-inexperienced directly after their egress from the thymus until they are primed by foreign antigen and differentiate into memory/effector CD4⁺ T cells. For the latter, a plethora of surface markers are used to define subpopulations, which differ with respect to stage of differentiation and effector functions, such as cytokine expression. This has proven to be beneficial for the analysis of CD4⁺ T-cell function in physiologic circumstances as well as under pathologic conditions and helped to shape our current understanding of CD4⁺ T-cell immunology. In contrast, naive, antigen-inexperienced CD4⁺ T cells have so far mostly been regarded as an entity of uniformly behaving lookalikes differing only in T-cell receptor (TCR) specificities. With respect to a characteristic phenotype, human naive CD4⁺ T cells have been considered to express CD45RA, CD62L, CD27, CD28, and CCR7 but to lack or to be characterized by low expression of molecules, such as CD45RO, CD11a, CD44, CD95, CXCR3, and CCR4.^1,2 

In addition, naive CD4⁺ T cells display characteristic functional capabilities, such as the ability to express interleukin-2 (IL-2). In contrast, they lack expression of classic effector cytokines, such as interferon-γ and IL-4, a hallmark of antigen-experienced memory/effector T cells. Reportedly, they also display unique calcium mobilization patterns and proliferative responses on stimulation.^3,4  Because they, by definition, have not undergone clonal selection during activation with a foreign antigen, they display a highly diverse TCR repertoire.⁵ 

During recent years, a substantial number of studies have highlighted the importance of continuous contact with major histocompatibility complex (MHC) class II peptides for naive CD4⁺ T cells, contrasting the definition of naive CD4⁺ T cells as antigen-inexperienced. It was demonstrated that activation threshold, survival, and homeostatic proliferation of naive CD4⁺ T cells depend strongly on MHC class II–derived signals.^6-14  According to experimental studies, the peptides presented by MHC class II in a nonimmunogenic fashion could be derived from self-antigens, related to those displayed in the thymus. However, a role for exogenous antigens cannot be ruled out in these processes. So far, neither the origin of the peptides presented in this context (self or nonself) nor the exact circumstances of these antigen encounters are known. Hence, naive CD4⁺ T cells most certainly receive signals on TCR engagement already soon after exiting from the thymus and can therefore not be considered as true antigen-inexperienced cells any more.

Nearly uniform populations of naive CD4⁺ T cells might exist in an experimental setting using laboratory animals of a young age. However, in humans, the involution of the thymus after puberty in combination with a constant pathogen challenge creates a unique situation that is not adequately simulated in animal models. In addition, most animal studies of homeostatic proliferation have used severely lymphopenic mice, creating a setting that does not apply entirely to the situation in humans.

Interestingly, absolute numbers of naive T cells in young and elderly humans are relatively stable, as opposed to the pronounced reduction of thymic function.^15,16  The question arises how human CD4⁺ T-cell immunity is maintained in adults. Only homeostatic proliferation of naive T cells can explain this phenomenon. This implicates that, as opposed to some experimental models where memory CD4⁺ T cells are generated during homeostatic proliferation,^17-21  human naive CD4⁺ T cells can proliferate postthymically while retaining their naive phenotype and functional characteristics. The reduced output of naive CD4⁺ T cells from the thymus on ageing would be balanced by peripheral self-renewal.

Importantly, this scenario implies the existence of at least 2 distinct subsets of human naive peripheral CD4⁺ T cells differing in their proliferative history: one dormant subset highly enriched in recent thymic emigrants and a second subset comprising naive CD4⁺ T cells that have proliferated in the periphery. In case these processes depend on engagement of TCR by MHC class II–peptide complexes,^10,13,14  postthymic naive CD4⁺ T-cell proliferation could be regarded as a peripheral positive selection of CD4⁺ T cells. The repertoire of CD4⁺ T cells would be shaped and restricted to those selected to undergo peripheral postthymic proliferation. Considering this, a detailed characterization of the mechanisms involved and the responsible antigens would be of highest interest. However, until recently, no marker enabled discrimination of naive CD4⁺ T cells that have proliferated in the periphery from recent thymic emigrants (RTEs) freshly emigrated from the thymus.

We have recently demonstrated that the surface molecule CD31 (PECAM-1) can be used to distinguish CD31^{+ thymic}naive and CD31^{− central}naive CD4⁺ T cells in the peripheral blood of healthy humans.²²  We and others have demonstrated a significantly higher signal joint T-cell receptor excision circle (sjTREC) content in CD31^{+ thymic}naive compared with CD31^{− central}naive CD4⁺ T cells.^22-25  During T-cell development, first sjTREC are generated through excision of the TCR D locus during TCR-α rearrangement. Because TRECs are not replicated during mitosis, the TREC content per cell in the progeny cells is diminished with each cell division. Thus, the quantification of TREC allows an assessment of the proliferative history of T cells at the population level.²⁶  As the high sjTREC content of CD31^{+ thymic}naive CD4⁺ T cells was only slightly reduced compared with thymocytes, it strongly supports the assumption that this subset contains RTE. On the other hand, the drastically diminished sjTREC content of CD31^{− central}naive CD4⁺ T cells strongly implies peripheral proliferation. Indirectly, this is also confirmed by other data. CD31^{− central}naive CD4⁺ T cells are preferentially infected with HIV compared with CD31^{+ thymic}naive CD4⁺ T cells.²³  Apparently, the proliferative active CD31^{− central}naive CD4⁺ T cells are more susceptible to infection and thereby provide a reservoir for HIV in humans.

Both, CD31^{+ thymic}naive and CD31^{− central}naive CD4⁺ T cells equally fulfill aforementioned phenotypic and functional criteria for naive CD4⁺ T cells, namely, expression of surface markers, such as CD45RA, CD27, and CD62L and secretion of IL-2 in the absence of significant effector cytokine production after polyclonal stimulation.²²  Moreover, both subsets do not comprise memory/effector Th cells specific for recall antigens, such as tetanus toxoid or CMVpp65, and can equally be differentiated into TH1 and TH2 cells (S. Kimmig, unpublished data, August 2001). However, frequencies of CD31^{+ thymic}naive and CD31^{− central}naive CD4⁺ T cells change during human life. In accordance with their phenotypic proximity to thymocytes, which also express CD31,^27,28  the frequency of CD31^{+ thymic}naive CD4⁺ T cells among CD45RA⁺ CD4⁺ T cells decreases dramatically with age. Although in cord blood 90% to 95% of CD45RA⁺ CD4⁺ T cells express CD31, the majority lack CD31 expression in the elderly.^{22,24,25,29,30}  In addition, absolute numbers of CD31^{+ thymic}naive CD4⁺ T cells in peripheral blood decrease during ageing and correlate well with the decline in TREC.^25,31  Absolute numbers of CD31^{− central}naive CD4⁺ T cells in the peripheral blood, however, remain rather constant independent of the reduced thymic function in the elderly.

What signals are involved in the generation of CD31^{− central}naive CD4⁺ T cells from the pool of dormant CD31^{+ thymic}naive CD4⁺ T cells? In agreement with studies in mice, demonstrating the importance of MHC molecules in homeostatic proliferation and/or survival of naive CD4⁺ T cells, we detected signs of recent TCR triggering in postthymically proliferated human CD31^{− central}naive CD4⁺ T cells.³¹  This is further supported by the notion that TCR-induced activation down-modulates the regulation of CD31 expression on human CD31^{+ thymic}naive CD4⁺ T cells³²  (S.K., unpublished data, January 2005). In contrast, during cytokine-induced in vitro proliferation, CD31 is expressed in a stable manner by CD31^{+ thymic}naive CD4⁺ T cells.³¹  The coexistence of 2 distinct subsets of naive CD4⁺ T cells in human adult peripheral blood implicates that current models of naive CD4⁺ T cells homeostasis are in-sufficient. Accordingly, we have developed a new, refined model that comprises a peripheral positive selection process for CD31^{+ thymic}naive CD4⁺ T cells (Figure 1).

In this scenario of homeostatic proliferation in humans, the function of the CD31 molecule itself could play a role. CD31 (PECAM-1) is a member of the Ig superfamily consisting of 6 extracellular C2-type Ig domains, a transmembrane region, and an intracellular domain with potential sites for intracellular signal transduction.^33-35  It is expressed on a variety of cell types, including T cells, mast cells, NK cells monocytes, granulocytes, endothelial cells, and platelets,³⁶  and several binding partners have been described. Homophilic binding of CD31 by CD31 can mediate cellular adhesive contacts. But also heterophilic interactions of PECAM-1 with several ligands, including CD38, integrin αvβ3, a 120-kDa ligand, and glycosaminoglycans, have been described.^37-40 

Therefore, depending on cell type and the binding partner involved, CD31 can operate in various manners (eg, as a sensor of shear stress and mechanical force, as well as regulator of adhesion, migration, and activation in T cells and a variety of other cell types).⁴¹  CD31 expression by T cells has also been described to confer suppressive capacities via a so far unidentified ligand.^39,42,43 

Furthermore, CD31 signaling in T and B cells may as well be a critical regulator of antigen-induced cell activation. Accordingly, stimulation of the TCR complex on Jurkat T cells leads to tyrosine phosphorylation of CD31 and recruitment of PTPs, SHP-2, and possibly SHP-1.⁴⁴  Additional coligation of the immunoreceptor tyrosine-based inhibitory motif-bearing CD31 led to a transient block of calcium mobilization.⁴⁵  Similar data have been reported for B cells with respect to coligation of the BCR complex with FcγRIIB1-CD31, resulting in the inhibition of downstream effector responses.⁴⁶  Collectively, these observations suggest that CD31 may serve as a negative regulator of immunoreceptor-mediated signaling events in T and B cells. Thus, it could also be involved in modulating the TCR-driven postthymic proliferation of CD31^{+ thymic}naive CD4⁺ T cells.

A concept of true homeostatic proliferation of naive CD4⁺ T cells implicates the self-renewal of naive T cells retaining all properties of their progenitors. However, we and others have obtained evidence that this is not the case in human naive CD4⁺ T-cell homeostasis. At first sight, the generation of CD31^{− central}naive CD4⁺ T cells preserves immunity because it compensates for a reduced thymic output. Even in the elderly, the number of naive CD4⁺ T cells is almost stable. This process of human postthymic naive CD4⁺ T-cell proliferation provides functional naive CD4⁺ T cells that should mount efficient adaptive immune responses. However, the restricted access to homeostatic signals, growth factors, or niches leads to selection pressure: some CD4⁺ T cells are expanded and others are lost during the process. The net effect has to be a loss of TCR specificities in the proliferated naive CD4⁺ T-cell pool (Figure 2).

In agreement with this theory, we have shown that, in healthy adults, CD31^{+ thymic}naive CD4⁺ T cells display a polyclonal TCR repertoire, whereas CD31^{− central}naive CD4⁺ T cells are characterized by striking TCR repertoire restrictions.³¹  As the frequency of CD31^{− central}naive among naive CD4⁺ T cells increases with ageing, the altered TCR repertoire of CD31^{− central}naive CD4⁺ T cells dominates naive Th cell immunity in the elderly. Indeed, repertoire restrictions can be detected already in the complete pool of naive CD4⁺ T cells.⁴⁷  This coincides with the reduced ability of the immune system in the elderly to respond to vaccination and various infections.^48,49  Many studies have demonstrated that immunity against neoantigens depends on CD4⁺ T cells recognizing a broad repertoire.^50-53  Thus, the loss of CD31^{+ thymic}naive CD4⁺ T cells could be detrimental.

Moreover, assuming that postthymic proliferation of CD31^{− central}naive CD4⁺ T cells is driven by self-peptides, a positive selection and expansion of potentially autoreactive Th cells might occur. As a result, together with cofactors such as predisposing HLA haplotypes, the accumulation of autoreactive naive CD4⁺ T cells could facilitate the initiation of autoimmune diseases. This harmful potential of lymphocyte homeostasis has been proven in mice where lymphopenia with consecutively increased peripheral homeostatic proliferation of T cells leads to severe autoimmunity.⁵⁴  In addition, thymic dysfunction resulting in increased T-cell turnover is associated with rheumatoid arthritis and multiple sclerosis.^55,56 

Nevertheless, under different circumstances, the selection and expansion of specific naive T cells could also be advantageous. As discussed earlier, the origin of the selecting peptides in the periphery is not known, and they could also be derived from foreign antigens. In this case, the postthymic proliferation of naive CD4⁺ T cells could be beneficial because the selected CD31^{− central}naive CD4⁺ T cells might build a potent alternative reservoir to support immune responses on reencounter with the foreign antigen. However, this would be helpful only in the case of pathogen reencounters. The efficiency of neoantigen-specific immune reactions would still be diminished because of the aforementioned loss of naive TCR spectrum diversity.

Can assessment of CD31^{+ thymic}naive CD4⁺ T cells be used for the analysis of RTE and thereby be used as a cellular correlate for human thymic function? RTEs are naive peripheral T cells, which have only recently exited the thymus and have not undergone further peripheral proliferation and antigen selection.⁵⁷  Accordingly, it can be postulated that they have high TREC content, that they dominate the peripheral naive T-cell pool in neonates, and that their numbers in peripheral blood depend on the magnitude of thymic export. All these characteristics apply to human CD31^{+ thymic}naive CD4⁺ T cells: similar to thymocytes they express CD31,^27,28  they do have a high TREC content, and they show an age-dependent decrease in absolute counts and frequencies among naive CD4⁺ T cells.^22,24,25,31  Thus, RTEs should express CD31.

But do all CD31^{+ thymic}naive CD4⁺ T cells represent RTEs? Interestingly, TREC levels of CD31^{+ thymic}naive CD4⁺ T cells tend to slightly decline with ageing,²⁴  implicating a certain degree of in vivo turnover. Furthermore, CD31^{+ thymic}naive CD4⁺ T cells decrease only approximately 5-fold during ageing and their absolute numbers in the peripheral blood in aged persons are higher than expected from studies determining thymic function by TREC levels.²⁶  In addition, the low deuterium labeling of naive CD4⁺ T cells in a recent study suggested that only a minor fraction of the peripheral blood naive CD4⁺ T-cell pool in adult healthy humans might be RTEs.⁵⁸  However, both TREC and deuterium labeling studies are based on assays that do not operate on the single-cell level. Thus, both analyses may underestimate thymic output, particularly in situations of altered peripheral T-cell activation, turnover, or even in healthy aging.⁵⁹ 

Reduced TREC levels among CD31^{+ thymic}naive CD4⁺ T cells in the elderly could also be a result of increased intrathymic proliferation of thymocytes or be caused by peripheral proliferation of CD31^{+ thymic}naive CD4⁺ T cells that maintain expression of CD31. Because TCR engagement would induce down-modulation of CD31 expression, this second mechanism of peripheral turnover of CD31^{+ thymic}naive CD4⁺ T cells is probably driven by homeostatic cytokines and does not require signals via the TCR. This could mean that homeostatic proliferation of human naive CD4⁺ T cells operates at 2 different levels. The transition from CD31^{+ thymic}naive CD4⁺ T cells into CD31^{− central}naive CD4⁺ T cells apparently necessitates engagement of TCR and involves intensive cell division, whereas cytokine-driven homeostatic proliferation might result in low-level autoreplication of CD31^{+ thymic}naive CD4⁺ T cells. Taken together, this leads to the conclusion that RTEs are highly enriched in CD31^{+ thymic}naive CD4⁺ T cells, although not all CD31^{+ thymic}naive CD4⁺ T cells necessarily represent RTEs. Thus, the assessment of absolute numbers of CD31^{+ thymic}naive CD4⁺ T cells should be a suitable method to evaluate human thymic function.

CD31^{+ thymic}naive CD4⁺ T cells represent an excellent cellular marker for human thymic activity for several reasons. TREC measurement on the DNA level has the disadvantage that the analysis is restricted to cell populations. Moreover, for TREC determination, different techniques and normalization standards exist. Accordingly, published results for the median TREC content in peripheral CD4⁺ T cells of normal healthy persons display great variations.^60-63  In contrast, the cytometric analysis of CD31^{+ thymic}naive CD4⁺ T-cell absolute counts is easily standardized and enables a direct assessment on a single-cell level. In addition, only minute amounts of blood are needed, and the simple measurement is performed within less than 1 hour.

To date, it has not only been demonstrated that the frequency and absolute numbers of TREC-enriched CD31^{+ thymic}naive CD4⁺ T cells correlate with age in healthy persons,^{22,24,25,29,31}  implying a link to thymic function. Furthermore, strong correlations of the TREC content in CD4⁺ T cells and frequencies of CD31^{+ thymic}naive CD4⁺ T cells have been described in common variable immunodeficiency patients⁶⁴  and in multiple sclerosis.²⁹  And, confirming its use as direct marker of thymic output, virtually all newly generated naive CD4⁺ T cells after autologous stem cell transplantation express CD31.⁶⁵  CD31 measurements have been successfully used in a variety of conditions, such as in the course of hematopoietic stem cell^25,65,66  and organ transplantation,⁶⁷  in autoimmune disease,²⁹  and HIV patients⁶⁸  and T-cell abnormalities, such as common variable immunodeficiency⁶⁴  or NF-κB essential modulator.⁶⁹  Interestingly, increases in the frequencies of CD31^{+ thymic}naive CD4⁺ T cells after autologous stem cell transplantation were statistically significant, whereas changes in TREC levels were not.⁶⁶  Although this might be the result of a higher sensitivity or less variability of the cytometric analysis compared with reverse-transcribed polymerase chain reaction, it could imply a superiority of CD31 assessment at least under certain conditions.

Although the determination of CD31^{+ thymic}naive CD4⁺ T cells might even be able to substitute TREC analysis in conditions of restricted sample availability and difficult laboratory settings, especially the combined assessment could be valuable. In general, the interpretation of TREC data is hampered by the fact that homeostatic proliferation resulting in changes in the composition of the analyzed T-cell population greatly influences the results. This has to be taken into account, especially in pathologic conditions. Under these circumstances, the parallel measurement of CD31^{+ thymic}naive and CD31^{− central}naive CD4⁺ T cells could help to interpret the results, as homeostatic proliferation would be reflected in increased CD31^{− central}naive CD4⁺ T cells.

In conclusion, CD31^{+ thymic}naive and CD31^{− central}naive CD4⁺ T cells are 2 new distinct subsets of human naive CD4⁺ T cells. So far various studies have used the measurement of CD31^{+ thymic}naive CD4⁺ T cells as a cellular correlate of thymic function particularly to monitor reactivation of thymic activity after immune ablative therapy or to study the pathogenesis of a variety of diseases. However, assessment of CD31^{+ thymic}naive CD4⁺ T cells could also be used as a prognostic marker for immune competence in general. Particularly, the analysis of CD31^{+ thymic}naive CD4⁺ T cells could enable to identify potential low responders or nonresponders with respect to primary immune responses against pathogens or in the course of vaccination. Conversely, as discussed, increased numbers of CD31⁻ naive CD4⁺ T cells might indicate a stronger disposition to autoimmunity. In the future, a further characterization of CD31^{+ thymic}naive and CD31^{− central}naive CD4⁺ T cells should deepen our knowledge of the mechanisms and implications of homeostatic proliferation of naive CD4⁺ T cells in humans and help to gain new insights into the role of thymic function in health and disease.

The authors thank S.N. Stehr and M. Dziubianau for critical reading of the manuscript.

This work was supported in part by the Berlin collaborative research centres SFB650 and TR36 and by a DC-Thera fellowship (S.K.).

Contribution: S.K. and A.T. wrote the paper.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Andreas Thiel, Clinical Immunology Group, Deutsches Rheuma-Forschungszentrum Berlin, Charitéplatz 1, 10117 Berlin, Germany; e-mail: thiel@drfz.de.