Human FOXN1-Deficiency Is Associated with αβ Double-Negative and FoxP3+ T-Cell Expansions That Are Distinctly Modulated upon Thymic Transplantation

Forkhead box N1 (FOXN1) is a transcription factor crucial for thymic epithelium development and prevention of its involution. Investigation of a patient with a rare homozygous FOXN1 mutation (R255X), leading to alopecia universalis and thymus aplasia, unexpectedly revealed non-maternal circulating T-cells, and, strikingly, large numbers of aberrant double-negative αβ T-cells (CD4negCD8neg, DN) and regulatory-like T-cells. These data raise the possibility that a thymic rudiment persisted, allowing T-cell development, albeit with disturbances in positive/negative selection, as suggested by DN and FoxP3+ cell expansions. Although regulatory-like T-cell numbers normalized following HLA-mismatched thymic transplantation, the αβDN subset persisted 5 years post-transplantation. Involution of thymus allograft likely occurred 3 years post-transplantation based on sj/βTREC ratio, which estimates intrathymic precursor T-cell divisions and, consequently, thymic explant output. Nevertheless, functional immune-competence was sustained, providing new insights for the design of immunological reconstitution strategies based on thymic transplantation, with potential applications in other clinical settings.


Introduction
The thymus is a primary lymphoid organ essential for normal T-cell development. The unique ability of the thymic microenvironment to generate and select T-cells requires a specialized epithelium that is regulated by forkhead box N1 (FOXN1) [1][2][3][4]. This transcription factor is expressed by the thymic anlage that emanates from the epithelium of the pharyngeal pouch, and is required for the differentiation of thymic epithelial cells [1,2]. Thymic epithelium comprises two main compartments defined by their position and functional characteristics: the cortex, and the medulla, the latter being thought to be fundamental for the negative selection of auto-reactive thymocytes and the generation of central tolerance. The cortical and medullary epithelia have been shown, in mice, to differentiate from a common epithelial progenitor that expresses FOXN1 [4]. Additionally, continuous FOXN1 expression was shown to be essential for both the maintenance of thymopoiesis in the murine postnatal period [4], and the prevention of thymic involution during adulthood [3,5], highlighting FOXN1 as an important target for immune reconstitution strategies. Mutations in FOXN1 lead to athymia together with total alopecia, due to the additional role of FOXN1 in hair follicle differentiation [1,2,6].
Human FOXN1-deficiency was first reported by Pignata et al. in two sisters from Campania, Italy [7,8]. Notwithstanding the evidence of athymia, a significant number of circulating T-cells was observed in these children with close to normal numbers of CD8 T-cells [9]. We identified the same homozygous R255X mutation [7,9,10], in a Portuguese child, who presented at 5 months with total alopecia and Bacillus Calmette-Guérin (BCG) dissemination, following routine neonatal vaccination with this live-attenuated mycobacterium, that also presented with a significant pool of circulating non-maternal T-cells [11]. This child underwent HLA mismatched thymic transplantation with evidence of functional immunologic reconstitution, as we recently reported [11].
The aim of this work was to investigate the T-cell compartment of this FOXN1-deficient child and the mechanisms underlying the immunological reconstitution upon thymic transplantation. Our findings suggest that T-cell development can occur in a putative FOXN1-deficient thymus rudiment, albeit with altered positive and negative selection as suggested by the marked expansion of T-cells expressing the alpha-beta T-cell receptor (TCR) in the absence of both CD4 and CD8 expression (double-negative DNab) together with an over-representation of T-cells of a regulatory-like phenotype expressing high levels of FoxP3 in conjunction with other regulatory markers. Notwithstanding, an extrathymic origin of the altered T-cell populations cannot be discarded. We also showed that the transplantation of HLA-mismatched FOXN1 competent thymic epithelium led to the achievement of sustained immune-competence despite evidence of involution of the allogeneic thymus 3 years post-transplantation. We believe that these data on the follow-up of this unique clinical case contribute significantly to the debate on the mechanisms underlying T-cell development and immune reconstitution that may help in the design of new therapeutic approaches in other clinical settings.

Patient
Female child, born at term to consanguineous Portuguese parents, admitted at day 157 of life with respiratory failure due to Bacillus Calmette-Guérin (BCG) dissemination, following routine neonatal BCG vaccination. The FOXN1 mutation identified is a homozygous C-to-T transition at nucleotide position 792 (Gen-Bank accession no. Y11739) leading to a nonsense mutation at residue 255 (R255X) in exon 4, formerly exon 5 [10]. Maternal chimerism was assessed using AmpFISTR Identifiler PCR Amplification Kit (Applied Biosystems, detection limit 1/100). The patient's clinical data were the focus of another manuscript [11]. Failure to thrive and progressive nutritional status deterioration were observed despite antibiotic/tuberculostatic therapy and intravenous immunoglobulin G. Thymus transplantation was performed (day 424 of life), under protocols approved by the Duke Institutional Review Board (IRB) and reviewed by the Food and Drug Administration under an Investigational New Drug (IND) application, as described [11]. Unrelated allogeneic thymus tissue, routinely discarded from infants less than 9 months of age undergoing cardiac surgery, was used for transplantation after informed consent [12][13][14].

Cell isolation and cell sorting
Peripheral blood mononuclear cells (PBMC) were isolated immediately after collection by Ficoll-Hypaque. Naïve CD4+ Tcells were sorted using the EasySep CD4+ naïve T-cell enrichment magnetic kit (StemCell Technologies). DNab and CD8 T-cells were isolated using the BD FACSAria High Speed Cell Sorter (BD Biosciences) after surface staining for CD3, CD4, CD8, TCRcd and TCRab. Population purity after sorting was greater than 98%.

Flow Cytometry
Lymphocyte subsets were characterized using fresh whole blood after acquiring at least 100,000 events within a lymphogate using a FACSCalibur flow cytometer (BD Biosciences). TCR Vb family frequency was quantified in whole blood using IOTest Beta Mark (Beckman Coulter). PBMC were stained intracellularly for CTLA-4 (clone BNI3) and/or Ki-67 (clone MOPC-21) both from BD Biosciences, and FoxP3 (clone PCH101) using eBiosciences's kit after surface staining, as described [15]. FoxP3 expression posttransplant was assessed in fresh PBMC, whereas cryopreserved PBMC were used for time-points prior to thymic transplantation. Apoptosis was assessed in fresh PBMC using Annexin V-FITC Apoptosis Detection Kit (BD Biosciences) and propidium iodide (PI) staining. Analysis was done using FlowJo software (TreeStar). Results were expressed as median intensity of fluorescence (MFI) of a molecule or percentage of positive cells, and absolute numbers calculated by multiplying their percentage by the absolute lymphocyte count.

TCR -chain CDR3 spectratyping
Total RNA was extracted from 10 5 to 10 6 cells with RNeasy kit (Qiagen) and first strand cDNA synthesized from 1-2 mg of RNA with the Superscript III kit (Invitrogen) using an equivolume mixture of random hexamers and oligo (dT). Amplification of the TCRVb CDR3 was performed using primers specific for each Vb family [18] except for Vb6 and Vb21 [19] and a common Cb reverse primer [18]; followed by a run-off reaction that extends each different PCR product with a constant Cb FAM labeled primer; and a third step, in which each different Vb PCR labeled fragment was separated using a capillary electrophoresis based DNA automated sequencer. Data were collected and analyzed with GeneMapper v4.0 (Applied Biosystems) for size and fluorescence intensity determination.

Abnormal expansion of DNab and FoxP3 bright T-cells in human FOXN1-deficiency
Circulating T-cells of non-maternal origin were documented at close to normal numbers (2219 cells/ml) at 5 months of age in a child with a homozygous R255X mutation in the FOXN1 gene, as we recently reported [11]. The T-cell compartment exhibited similar proportions of CD4+, CD8+ and, ab-cells that expressed neither CD4 nor CD8 (double-negative, DNab), which usually represent less than 1% of circulating T-cells. Athymia was diagnosed based on absent thymus-shadow on x-ray, lack of naïve T-cells, and undetectable levels of products of TCR rearrangements in the thymus, specifically signal-joint TCR excision circles (sjTREC) [11]. Here, we aimed to further investigate the phenotype and the function of these T-cells developed in the context of a putative athymia. We found that CD4+ T-cells exhibited an activated memory-effector phenotype with preserved IL-2, IFN-c and IL-4 production ( Figure 1A). CD8+ T-cells featured a similar activated phenotype, with no terminal-effector differentiation, as illustrated by the maintenance of CD45RO/ CD28/CD27 expression ( Figure 1B) and the low frequency (3%) of perforin-producing cells ( Figure 1B). The aberrantly expanded DNab T-cells (676 cells/ml) also expressed CD45RO, most were CD27+, and produced IL-2 ( Figure 1C), in agreement with a lack of terminal-effector differentiation [17]. A significant proportion of T-cells were cycling ( Figure 1D). It is likely that IL-7 played a role in T-cell maintenance/expansion given their preserved expression of the IL-7 receptor a-chain (IL-7Ra), as illustrated for CD4 T cells in Fig. 1A, and lack of elevated serum IL-7 levels (6.3 pg/ml, 209 days pre-transplantation). IL-7 serum levels typically increase in lymphopenic settings [17]. In agreement with the evidence of IL-7 use, IL-7 serum levels were not increased before transplantation, and increased transitorily to 44.3 pg/ml (133 days posttransplantation) during T-cell depletion in the peri-transplantation period.
The thymus is known to produce a regulatory CD4+ T-cell subset (Treg), fundamental for preventing autoimmunity, currently best identified by expression of the forkhead box P3 transcription factor (FoxP3) [22][23][24][25]. We found that up to 40% of the CD4 subset (328 cells/ml) expressed high levels of FoxP3, and observed atypical populations of FoxP3+ DNab and double-positive T-cells ( Figure 1E). FoxP3 can also be up-regulated in non-Treg T-cells upon activation [24,26]. Nonetheless, several findings support these cells being bona fide Treg. In agreement with human Treg phenotype [24], they expressed FoxP3 at high intensity concomitantly with other Treg-associated markers, namely CTLA-4 and CD39 ( Figure 1E). Moreover, in contrast to activated T-cells, they did not produce IL-2 or IFN-c ( Figure 1E). The limited amount of blood samples precluded suppressive assays as well as the evaluation of Treg-like cell repertoire by CDR3 sequence analysis. Of note, comparison of the relative representation of different Vb families within the FoxP3+ and FoxP32 CD4 sub-populations revealed a skewed repertoire in both populations but with distinct oligoclonal expansions in each, providing further support that these FoxP3+ cells represented a separate CD4 lineage ( Figure 1F).
Thus, despite the presence of FOXN1-deficiency, we observed a reasonable number of circulating non-maternal T-cells associated with oligoclonal expansions of DNab and FoxP3+ T-cells. The Sustained immune-competence was achieved upon HLA mismatched thymic transplantation despite involution of thymus allograft As FOXN1 mutations impact on thymic epithelium rather than on hematopoietic precursors, we predicted that thymic transplantation, although never previously performed in this setting, could provide a curative strategy. As shown in our previous clinical report [11], the efficacy of the thymic transplantation performed was best demonstrated by the temporal association between the clearance of ongoing BCG adenitis and the development of PPDspecific proliferative responses. We continued to follow the child and confirmed that she remained free of significant infections 5 years after cessation of all prophylactic therapies. Here, we aimed to further investigate the kinetics of T-cell recovery upon HLAmismatched thymic transplantation and its relationship with output of the thymic explants.
A slow progressive increase in the proportion of circulating naïve CD4+ T-cells, mainly expressing CD31, a marker associated with recent-thymic emigrants [17], was observed (Figure 2A), accompanied, as expected, by an increase in sjTREC levels within total PBMC ( Figure 2B). Despite the HLA-mismatch of the thymic epithelia, naïve CD4+ T-cells showed a fully diverse TCR repertoire 5 years post-transplantation ( Figure 2C), confirming the trend shown in our previous report within total CD4 T-cells [11].
The expanded Treg-like cells disappeared during the peritransplant period and a parallel reconstitution of the Treg pool and CD4 subset was observed, leading to stable frequencies within the normal range ( Figure 2D) as well as absolute counts (44 cells/ ml, 36 months post transplantation). Of note, distribution of Vbfamilies within FoxP3+ and FoxP32 CD4+ T-cells was very similar despite the HLA-mismatch between thymic epithelia and host ( Figure 2E).
CD8+ T-cell recovery was disproportionally low compared to CD4 T-cells (92 cells/ml, 9% of T-cells, 5 years post-transplantation), confirming our previous report [11]. Nevertheless, we show here that the kinetics of naïve cell expansion within CD8+ subset paralleled those observed for their CD4+ counterparts (Figure 2A). In agreement, we found similar sjTREC levels in purified CD4+ and CD8+ T-cells (7155 and 7540 sjTREC/10 5 cells; respectively, 30 months post-transplantation). A poor CD8 recovery has been described following HLA-mismatched thymic transplantation in DiGeorge syndrome [12,13], possibly related to the full HLA class I mismatch between host hematopoietic precursors and allogeneic thymic epithelia or to alterations in the thymic graft associated with transplantation procedures. Nevertheless, the transplant-derived CD8+ T-cells were apparently functional in this child [11], as documented by the transitory expansion of terminally differentiated CD8 T cells during Varicella Zoster virus infection ( Figure 3).
Additionally, in order to estimate the functionality of the allogeneic thymic graft we quantified, for the first time in the context of thymic transplantation, the sj/bTREC; a ratio between early and late products of TCR rearrangements providing an indirect measurement of thymocyte division-rate and a direct correlate of thymic output [20,27,28]. Whilst very low during the peri-transplant period, the sj/bTREC ratio increased progressively ( Figure 2B), reaching levels comparable to those observed in healthy children, 2.5 years post-transplant. Of note, a sharp decline in sj/bTREC was observed 4 years post-transplantation ( Figure 2B). This was accompanied by a modest decrease in sjTREC levels within total PBMC and in the proportion of naïve cells, supporting the decline in thymic allograft output (Figure 2A-B). These values plateaued thereafter (Figure 2A-B), suggesting that a steady-state equilibrium was established after replenishment of the immune system.
These data provide novel evidence that immune-competence can be achieved in the absence of long-term sustainability of thymic output from the allogeneic tissue, with implications for other clinical settings aimed at immunological reconstitution.

Persistence of circulating DNab T-cells despite the immunological recovery upon thymic transplantation
We also investigated the fate of the DNab T-cell population that was markedly expanded pre-transplantation [11]. In contrast with the recovery of all T-cell subsets, a significant population of circulating DNab persisted, at relatively stable numbers (range 125-548 cells/ml throughout the follow-up) [11], being 188 cells/ ml at 6-years post-transplantation ( Figure 4A).
DNab cells maintained a similar memory phenotype, being all CD45RO+ ( Figure 4B), and skewed repertoire, as assessed by spectratyping ( Figure 4C), throughout the follow-up period. They also displayed no evidence of terminal-effector differentiation as indicated by their preserved ability to produce IL-2 in the absence of significant amounts of effector cytokines such as IFNc, IL-4, IL-10 or IL-17, following short-term PMA/Ionomycin stimulation ( Figure 4D), as well by their preserved expression of CD27, in the absence of CD57 and perforin ( Figure 4E). A potential ability for mucosal homing was suggested by significant expression of CCR6 and CD103 ( Figure 4B), which was of even more interest given that DNab cells neither produced IL-17 ( Figure 4D) nor expressed the Treg markers FoxP3 or CTLA-4 ( Figure 4F). Their levels of CD38 and HLA-DR expression were low ( Figure 4F), suggesting that they were not in an activated state.
Importantly, DNab cells expressed high levels of CD25, the IL-2 receptor a-chain, in conjunction with CD127, the IL-7 receptor a-chain ( Figure 5A). Moreover, they showed higher levels of Bcl-2 expression than CD4 and CD8 T-cells ( Figure 5B), suggesting in vivo responsiveness to IL-7. To test this hypothesis we quantified the levels of STAT-5 phosphorylation (p-STAT-5) within DNab, CD4 and CD8 T-cells upon short-term exposure to IL-7, IL-2 or IL-15 in vitro. As shown in Figure 5C, DNab cells exhibited the highest levels of p-STAT-5 upon IL-2 or IL-7 stimulation and were not responsive to IL-15. Additionally, freshly isolated PBMC were cultured in the presence and in the absence of these cytokines as well with anti-CD3 plus anti-CD28 mAb as a positive control for T-cell stimulation. We found that, as for CD4 T-cells, DNab cells up-regulated CD25 expression upon IL-7 stimulation in vitro ( Figure 5D), but showed no distinct proliferative abilities, as assessed by the expression of the cell-cycling marker Ki67, in response to these cytokines in comparison with CD4 or CD8 Tcells ( Figure 5E). Of note, they respond to TCR stimulation with both up-regulation of CD25 ( Figure 5D) and increased frequency of cycling cells ( Figure 5E). Notably, the frequency of cycling DNab cells ex vivo was very low ( Figure 5F). Thus, our findings supported a role of IL-7 and IL-2 in DNab cell maintenance, given their high levels of expression of IL-7 and IL-2 receptors and ability to respond to these cytokines in vitro. Moreover, DNab cells showed increased Bcl-2 levels in the absence of cell cycling markers in vitro and ex vivo, suggesting that IL-7 and/or IL-2 mainly impacted on DNab cell survival rather than their turnover.
In order to evaluate the thymic contribution for the maintenance of DNab cells, we quantified the sjTREC levels in sorted DNab cells from the peripheral blood 6 years post thymus transplantation. The sjTREC levels were found to be very low (3.2 sjTREC/10 5 DNab cells as compared with 3783 sjTREC/10 5 sorted CD4 T cells and with 2212 sjTREC/10 5 sorted CD8 T cells), supporting the possibility that DNab were long-lived cells originated before the thymus transplantation.
Overall, we found that the DNab cells abnormally expanded before thymic transplantation were maintained for up to six years post-thymic transplantation and appeared to rely mainly on cytokine-driven survival.

Discussion
The transcription factor FOXN1 is considered essential for the development of the thymic epithelia and the prevention of its involution during adulthood. Nevertheless, we demonstrated here that human homozygous R255X FOXN1-deficiency may be associated with significant numbers of circulating T-cells of nonmaternal origin with major expansions of DNab and FoxP3+ cells through the study of a rare clinical case.
Our data raise important questions regarding T-cell origin in the context of human athymia. One plausible explanation is that a thymic rudiment may persist, facilitating a limited production of T-cells that subsequently expanded in the periphery. In mice, the Foxn1 gene was shown to be dispensable for the initial formation of the thymic primordium [4,29]. There is also evidence of functional T-cells in nude mice [30], although, at least some of these T-cells seem to be generated extra-thymically [31]. CD4+ and CD8+ ab T-cells accumulate with ageing in nude mice [32], but characterization of a putative Treg compartment has not been conducted. We have shown that FoxP3 induction can occur in early stages of both murine and human normal T-cell differentiation [15,25]. Of note, significant DNab as well as FoxP3+ cells were found in a mouse model of extra-thymic lymphopoiesis induced by Oncostatin M, a cytokine that induces thymic atrophy and lymph node alterations that support T-cell differentiation [33].
Importantly, circulating T cells of non-maternal origin were found in all the cases described with R255X FOXN1 mutation [7][8][9]11], which leads to a short N-terminal FOXN1 protein without DNA binding domain [34]. In contrast no circulating T-cells were found in a patient with a second mutation identified in the human FOXN1 gene (R320W), a missense mutation of the DNA binding domain [11]. These discrepant findings raise new hypotheses about the role of FOXN1 in the development of the thymic epithelium that deserve further exploration using mouse models and comparative structural studies [1,35,36].
As all patients reported with R255X FOXN1 mutation presented circulating T-cells [7][8][9]11], it is plausible that they retained a dysplastic thymic rudiment capable of supporting T-cell differentiation, albeit with a narrow TCR repertoire and impaired T-cell selection, allowing the emergence of atypical DNab and Treg. Alternatively, this thymic rudiment could allow T-cell commitment of precursors for subsequent extra-thymic development. In support of this possibility, progenitor T-cell commitment was shown to occur in the thymus prior to their extra-thymic development in mouse models [37].
The Treg compartment has recently been investigated in other clinical settings associated with peripheral oligoclonal T-cell proliferation in patients with thymic impairment either due to hypomorphic mutations in hematopoietic precursors (Omenn syndrome) [38][39][40], or to developmental defects associated with variable degrees of thymic hypoplasia (DiGeorge syndrome) [41,42]. In these settings, Treg frequencies appeared to be unaltered or reduced [38][39][40][41][42], emphasizing the distinctiveness of the pattern of high levels of Treg seen before transplantation in our case of R255X FOXN1-deficiency. Remarkably, the peripheral pool of FoxP3+ T cells normalized upon transplantation of FOXN1 competent thymic epithelia. Thymic Treg development is currently thought to be dependent on a small developmental niche that tightly controls Treg output [43,44]. It is thus plausible to speculate that FOXN1 may play a role in such niches, contributing to the thymic regulation of Treg numbers.
The numbers of circulating DNab T-cells remained relatively constant up to 6 years after thymic transplantation. Of note, these cells have not been shown to expand following HLA-mismatched thymus transplantation [11,12], and were reported to progressively decline after effective naïve reconstitution in complete DiGeorge patients that presented with this atypical phenotype before thymic transplantation [12].
Regarding the origin of DNab T-cells, their reduced sjTREC levels suggest that they were not produced after transplantation. Moreover, it is unlikely that they are activated terminallydifferentiated CD8+ T-cells that lost CD8 expression, as suggested in other clinical settings associated with abnormal expansions of circulating DNab cells, such as autoimmune lymphoproliferative syndromes (ALPS) [45], given the lack of expression of markers of terminal differentiation. The high levels of expression of mucosalhoming markers suggest an extra-thymic origin of DNab T-cells. They may have been generated pre-transplant or result from continuous de novo production. In the latter case their persistence 5 years post-transplantation favors a thymic rather than an extrathymic origin since any putative extra-thymic lymphopoiesis is likely to be shut-down upon thymic transplantation [31]. We showed that DNab T-cells expressed high levels of IL-2 and IL-7 receptor a-chain and of Bcl-2 suggesting in vivo responsiveness to IL-7. Our data showing their ability to phosphorylate STAT-5 upon IL-7 or IL-2 stimulation further support this hypothesis. On the other hand, the low ex vivo frequency of cycling cells, and their reduced proliferative response to these cytokines, suggest that the DNab cell maintenance may be largely dependent on cytokineinduced survival, rather than cytokine-driven expansion.
In order to estimate the functionality of the allogeneic thymic graft, we took advantage sj/bTREC ratio quantification for the first time in this context. We found a progressive sj/bTREC increase in our clinical case, reaching levels comparable to those observed in healthy children. Importantly, a sharp decline of sj/ bTREC, accompanied by a decrease in sjTREC levels and the proportion of naïve cells was observed 4 yrs post-transplantation. These values plateaued thereafter, illustrating the contribution of peripheral homeostasis to T cell pool maintenance following its replenishment, even in the absence of sustained thymic production. It is possible that the apparent lack of sustained thymocytedivision rate (as evidenced by sj/bTREC) resulted from an intrinsic reduced longevity of the thymus allograft or that the shutdown of the thymocyte production being a consequence of the replenishment of the peripheral pool. Given the rarity of FOXN1 deficiency, the underlying factors contributing for the sustainability of the thymus allograft could be investigated in future studies on immunological reconstitution upon thymus transplantation in cases of DiGeorge syndrome with athymia. The data generated will be important to appraise the promising use of thymus transplantation in other clinical contexts.
Overall, our finding of significant numbers of oligoclonal T-cells in this case of FOXN1-deficiency due to R255X mutation suggest that, to a certain extent, T-cell development still occurs, albeit with altered positive/negative selection, as illustrated by the aberrant expansion of FoxP3+ and DN subsets. Importantly, we showed that immune-competence can be achieved through HLA-mismatched thymic transplantation, in spite of the apparent lack of long-term functionality of the allogeneic thymic tissue, which has potential implications for the design of immunological reconstitution strategies in other clinical settings.