Sonic Hedgehog regulates thymic epithelial cell differentiation

Sonic Hedgehog (Shh) is expressed in the thymus, where it regulates T cell development. Here we investigated the influence of Shh on thymic epithelial cell (TEC) development. Components of the Hedgehog (Hh) signalling pathway were expressed by TEC, and use of a Gli Binding Site-green fluorescence protein (GFP) transgenic reporter mouse demonstrated active Hh-dependent transcription in TEC in the foetal and adult thymus. Analysis of Shh-deficient foetal thymus organ cultures (FTOC) showed that Shh is required for normal TEC differentiation. Shh-deficient foetal thymus contained fewer TEC than wild type (WT), the proportion of medullary TEC was reduced relative to cortical TEC, and cell surface expression of MHC Class II molecules was increased on both cortical and medullary TEC populations. In contrast, the Gli3-deficient thymus, which shows increased Hh-dependent transcription in thymic stroma, had increased numbers of TEC, but decreased cell surface expression of MHC Class II molecules on both cortical and medullary TEC. Neutralisation of endogenous Hh proteins in WT FTOC led to a reduction in TEC numbers, and in the proportion of mature Aire-expressing medullary TEC, but an increase in cell surface expression of MHC Class II molecules on medullary TEC. Likewise, conditional deletion of Shh from TEC in the adult thymus resulted in alterations in TEC differentiation and consequent changes in T cell development. TEC numbers, and the proportion of mature Aire-expressing medullary TEC were reduced, and cell surface expression of MHC Class II molecules on medullary TEC was increased. Differentiation of mature CD4 and CD8 single positive thymocytes was increased, demonstrating the regulatory role of Shh production by TEC on T cell development. Treatment of human thymus explants with recombinant Shh or neutralising anti-Shh antibody indicated that the Hedgehog pathway is also involved in regulation of differentiation from DP to mature SP T cells in the human


Introduction
The thymus provides a specialised environment for the production of T cells. Thymic epithelial cells (TECs) are an essential component of the thymic stroma, and are required to support T cell development. Two broad categories of TEC, which are believed to arise from a common progenitor, have been defined by their localisation, function and cell surface markers [1,2]. Cortical(c)TEC provide Dl4 for T cell fate specification, and present MHC þ peptide ligands for positive selection. They are defined as EpCam1 þ , CD40 þ , CD205 þ , Ly51 þ and MHCII þ , and express genes for antigen presentation, including Cathepsin-L, Prss16 and b5t. Medullary (m)TEC are specialised for negative selection, and are defined as surface EpCam þ , CD40 þ , CD205 À , Ly51 À and MHCII þ cells that react with the lectin UEA-1. Some mTEC express the Aire gene and Cathepsin-S, facilitating expression and presentation of Tissue Restricted Antigens for induction of tolerance. While TEC provide multiple essential signals for T cell development, they also require signals from thymocytes for their maturation.
Aire function in mTEC is essential for the induction of tolerance to self in both humans and mice, and Aire mutation leads to profound multi-organ autoimmunity [3,4]. Other factors which regulate mTEC differentiation and function are also likely to influence self-tolerance, but currently TEC differentiation is not well understood. To date, only a few factors have been identified that are required for TEC differentiation, such as the transcription factor FoxN1, which when expressed ectopically can programme other lineages to a TEC fate [5].
During foetal thymus ontogeny, TEC differentiation has been defined in terms of cell surface expression of CD40 and CD205. The TEC progenitor population, which is bipotential for cTEC and mTEC is contained within the CD40 low CD205 low population [6,7]. Further gradual acquisition of CD40 and increase in CD205 expression gives rise to a transitional progenitor, able to differentiate into either functional cTEC, or into mature mTEC that lose CD205 expression and acquire the expression of mTEC characteristic markers, such as Aire [8,9]. These observations are consistent with another study which showed that Aire-expressing mTEC originate from b5t þ precursors, a molecule expressed in mature cTECs and not in other cell types [10]. Although these markers have proved useful for investigating TEC development, the lineage relationship between mTEC and cTEC populations and the factors that drive the progression from bipotent progenitors, through transitional intermediates, to mature TEC are not well understood in foetal or post-natal thymus [1].
Interactions between TEC and thymocytes have been shown to promote the terminal differentiation of TEC lineages, particularly mTEC, but fate specification to either lineage is believed to occur independently of signals from thymocytes [11]. Relatively few secreted factors or cellecell interactions have been identified that regulate TEC differentiation, although members of the tumour necrosis factor receptors super family (TNFRSF), including RANK (TNFRST11a), and CD40 and TGF-b are required for normal thymus medulla development, growth and function [12e14].
Here, we investigate the role of Sonic Hedgehog (Shh) in the regulation of TEC development. Shh is one of three mammalian Hedgehog proteins (Shh, Indian hedgehog (Ihh) and Desert Hedgehog (Dhh)) which share a common signalling pathway. Hedgehog proteins signal by binding to their cell surface receptor Patched1 (Ptch1), and this binding releases Ptch1's repression of Smoothened (Smo), allowing Smo to transduce the Hh signal. At the end of the signalling pathway are the Hh-responsive transcription factors, Gli1, Gli2 and Gli3. Gli1 is itself an Hh-target gene, and encodes an activator of transcription, whereas Gli2 and Gli3 can be processed to function as transcriptional activators (in the presence of Hh pathway activation) or transcriptional repressors (in the absence of Hh pathway activation). Gli2 is required to intiate the Hh signal, and functions largely as a transcriptional activator in vivo, whereas Gli3 functions predominantly as a transcriptional repressor in vivo, and can act to repress Shh transcription (by repression of an intermediate transcriptional activator) [15]. In fact, in many tissues, Shh and Gli3 have opposing functions, with Shhdeficiency and Gli3-deficiency giving opposing phenotypes [15].
Hedgehog proteins are expressed in the thymus [16e18], and signal to developing T cells to promote differentiation and proliferation of early thymocyte progenitors [19,20]. In both mouse and human studies, Hh signalling has been shown to negatively regulate pre-TCR induced differentiation from CD4 À CD8 À double negative (DN) to CD4 þ CD8 þ (DP) cell [17,21e23]. In addition, in mouse studies, Shh has been shown to inhibit TCR-induced differentiation from DP to mature CD4 and CD8 single positive (SP) thymocytes [24e27]. Sonic hedgehog (Shh) is expressed by thymic stromal cells, and immunofluorescence has located these cells  primarily to the cortical medullary junction in mouse and human [17,18,20]. Dhh has also been shown to be expressed primarily by stromal cells [18], whereas Ihh is expressed by both stromal cells and thymocytes, with highest expression detected in the DP thymocytes [22].
Given the pivotol role of the thymus in preventing autoimmunity, it is of interest that Hedgehog pathway genes and Hedgehogpathway targets have been identified in genome wide association studies (GWAS) for the human autoimmune disorder primary biliary cirrhosis, and in primary open-angle glaucoma (POAG), which may be associated with autoimmunity [28,29]. In addition, components of the Hedgehog-pathway have been detected in sera from patients with Rheumatoid arthritis, Lupus and ankylosing spondylitis [50]. In the human thymus, Hedgehog proteins are expressed by TEC and have been shown to signal to developing thymocyte progenitors, and aberrant Hedgehog pathway activation has been observed in both T acute lymphoblastic leukaemia (T-ALL) and thymoma, which is associated with the neuromuscular autoimmune disorder Myasthenia gravis [18,23,30e34].
Here we analyse the function of Hedgehog signalling in TEC. We show that developing TEC also express components of the Hh signalling pathway and transduce Hh signals in the foetal and adult thymus. We show that Shh is required for normal TEC development, and that it particularly influences the mTEC lineage.
ShhF/F were crossed with FoxN1-Cre transgenic mice [37]. Mice were bred and maintained at UCL. All mice studies were reviewed and approved by the British Home Office.

FTOC and culture of human thymus explants
E14.5 Foetal thymi were cultured as described [38] for seven days before analysis. In some experiments, recombinant rHhip (SigmaeAldrich) was used at 1 mg/ml. For culture of human thymus explants, human thymus tissue was obtained after surgical removal in children undergoing corrective cardiac surgery. The study had full ethical approval from the Local Research Ethics Committee; in accordance with the Declaration of Helsinki, fully informed consent was obtained from parents of all child donors. The human thymus tissue was cut intõ 1 mm cubed fragments using a scalpel, and cultured on 0.8 mm Millipore filters (Millipore, Massachusetts, US) on 1 ml AIM-V serum free medium (Invitrogen, US) in 24-well plates for five days before analysis. Recombinant Shh protein (at 0.5, 0.25 or 0.125 mg/ml) (R&D Systems) or the anti-Shh neutralising antibody 5E1 (5 mg/ml), as described previously [17], were added where stated.

Isolation of epithelial cells from embryonic and adult thymus
For isolation of adult TEC, thymic lobes were dissected and cut into small pieces (~10 mm Â 10 mm), excess thymocytes removed by sequential washes in HBSS containing FCS 10% (v/v) and incubated for 30 min at 37 C in HBSS containing FCS, 3.3 mg/ml Liberase and 0.5 mg/ml DNaseI (Roche Diagnostics UK). After incubation, cells suspensions were prepared by gentle mechanical dissociation.
For isolation of embryonic TEC, thymic lobes were dissected and washed once in PBS followed by an incubation of 15 min in PBS containing 0.25% trypsin and 0.02% EDTA and gentle mechanical dissociation to obtain cell suspensions.

PCR analysis for genotyping
DNA for PCR analysis was extracted from tissue as described [39]. Approximately 1 mg of DNA was used as a template in each PCR reaction, using primers as described in Refs. [37] and [19].

Antibodies and Flow cytometry
Cells were stained as described [40,41,9] using directly conjugated antibodies against the markers stated (eBioscience; BD PharMingen; and Biolegend, US). Identification of mTEC populations was done using the lectin UEA-1 (Vector Labs, UK). For intracellular Aire staining cells were fixed and permeabilised using the IC fixation/permeabilisation kit (eBioscience).
Experiments were acquired on a Becton Dickinson LSR-II and data representative of >3 experiments analysed using Flowjo 10.6 (Tree Star, US).

Microarray data normalisation
Publicly available gene-expression microarray datasets performed on RNA from wild type highly purified TEC [41] were analysed for the expression of several components of the Hh signalling pathway. Packages "affy" and R were used for quality control and normalisation of the datasets by log-scale Robust Multi-Array analysis (RMA).

Hedgehog signalling is active in thymic epithelial cells (TEC) of adult mice
We have previously shown that the components of the Hedgehog signalling pathway are expressed by foetal thymic stroma, [42]. To test if components of the signalling pathway are expressed specifically by TEC, we analysed microarray datasets performed on highly pure sorted TEC populations [41] to assess the expression of Shh, the Shh-receptor Ptch1, the signal transducer Smo and the three Hh-responsive transcription factors (Gli1, Gli2 and Gli3). Transcripts for Shh and the signal transducers Gli1-3 were detected both in cortical and medullary TEC ( Fig. 1A and B). Transcripts for the transcriptional activator Gli1 were low in both TEC subpopulations whereas transcripts for Gli2 and Gli3, whose products can act either as transcriptional activators or repressors, were detected at significantly higher concentrations in cTEC when compared to mTEC (Fig. 1B). In addition, Smo and Ptch1 were highly expressed in both TEC subpopulations (Fig. 1C). Taken together, this transcriptional analysis suggested that (autocrine) Shh signalling is likely to occur in TEC.
We next wished to demonstrate active Hedgehog signalling in TEC using a Gli binding site (GBS)-GFP reporter transgenic [35] and determine the proportion of epithelia that display Gli-mediated transcription. Using this transgenic system to measure Hh signaling activity in the adult thymus, we observed that a higher frequency of mTEC showed active signalling (11.3%) when compared to cTEC (3.9%; Fig 1D). This difference is consistent with the higher expression of the transcriptional repressor Gli3 in cTEC ( Fig 1B). Most of the mTEC undergoing Hedgehog signalling (GFPþ) displayed a mature phenotype, as measured by the expression of high levels of cell surface MHCII, and approximately half of MHCII high mTEC were GFPþ (Fig. 1E). These data thus indicate that adult TEC are undergoing active Hh-signalling, particularly in the mature MHCII high mTEC subset.

Hedgehog signalling is active in developing TEC
To assess whether Hh signalling is active in TEC during foetal thymus development, we used the Gli Binding Site (GBS)-GFP reporter transgenic to analyse GFP expression in populations of developing TEC, defined by the changing expression of the developmental markers CD40 and CD205 [8,9]. We analysed TEC populations on E14.5, E16.5, E18.5 and E21.5 (neonate). We detected GFP expression in all TEC populations, defined by CD40 and CD205 expression (Fig. 2). GFP expression was overall highest in terms of mean fluorescence intensity (MFI) and percentage of positive cells early during development (on E14.5 and E16.5, Fig. 2A and B) with intensity declining through out ontogeny. On E16.5 (Fig. 2B) we found that both CD40 int CD205 high and CD40 high CD205 low TEC populations expressed similar GFP levels but at the later E18.5 stage, CD40 high CD205 low TEC expressed higher GFP levels, indicating greater Hh signalling (Fig. 2C). At this time in development, these are cells that are likely to have downregulated CD205 and specified to the mTEC lineage. In the neonatal thymus, the intensity of GFP expression had declined, but GFP was still detectable in both cTEC and mTEC lineages (Fig. 2D). Thus, developing TEC have Gli-mediated transcriptional activity in the foetal thymus, suggesting a role for the Hh pathway during their development.

Foetal TEC development requires Shh
To test whether Shh is required for foetal TEC development, we analysed TEC development in Shh À/À embryos. Shh-deficiency is embryonic lethal, and Shh is expressed by TEC in the foetal and adult thymus (Fig. 1A) [17,18]. Previous studies have shown that the Shh À/À thymus contains fewer thymocytes than wild type (WT), with inefficient expansion and differentiation of early CD4 À CD8 À (DN) thymocytes [19,22], but accelerated pre-TCR dependent differentiation from DN to CD4 þ CD8 þ double positive (DP) thymocyte, and increased efficiency of differentiation from DP to single positive (SP) thymocyte in vitro [21,43]. These defects in thymocyte development could be the consequence of loss of direct Shh-signalling from TEC to thymocytes, or alternatively could result indirectly due to changes in the thymopoietic capacity of TEC, leading to an indirect effect on thymocytes.
As Shh deletion is embryonic lethal, and most embryos die before E16, we first examined TEC development in the E15.5 foetal thymus. As described previously, the E15.5 Shh À/À thymus was smaller than WT [19]. The number of CD45 À EpCam1 þ (epithelial lineage cells) was reduced (Fig 3A), and cell surface expression of CD205, and the percentage of CD205 þ cells within the EpCam1 þ population were also reduced, relative to WT (Fig. 3B). In order to analyse TEC development in the absence of Shh at later stages of their differentiation, we used foetal thymus organ culture (FTOC) to compare development of Shh À/À and WT TEC (Fig. 3CeG). Shh À/À and WT littermate thymus lobes were cultured for 7 days (Fig. 3CeG). The number of cells recovered from Shh À/À thymi was decreased (Fig. 3C) and there was an overall decrease in the number of TEC (CD45 À EpCam1 þ cells), relative to WT (Fig. 3D). When we stained against markers of mature cTEC and mTEC populations, Ly51 and UEA-1, we found a significant change in the frequencies of cTEC and mTEC in the Shh-deficient thymus, with a marked reduction in mTEC (Fig. 3EeF). The percentage of mTEC was reduced from 27.4% in the WT FTOC to 13.2% in Shh À/À FTOC, with a concomitant increase in the proportion of cTEC (Fig. 3E). Overall, both the number of cTEC and the number of mTEC were significantly reduced in Shh À/À FTOC compared to WT FTOC, although the mTEC population was affected more severely than cTEC (Fig. 3F). Cell-surface expression of MHCII is essential for both the role of cTEC in positive selection, and the role of mTEC in negative selection, and high cell surface MHCII expression is a marker of mature mTEC. Interestingly, cell surface expression of MHCII was higher in Shh À/À FTOC compared to WT, in both cTEC and mTEC populations (Fig. 3G).
Gli3 acts as a suppressor of Shh mediated signalling [15] and the Gli3-deficient thymus shows increased Hh-mediated transcription in the thymic stroma [42]. We therefore tested if constitutive loss of Gli3 expression caused changes in TEC cellularity, differentiation and phenotype that are opposite to those observed in Shh-deficient FTOC. We observed an increase in TEC numbers in Gli3 À/À FTOC relative to WT littermates (Fig. 4A). In addition, TEC from the Gli3deficient thymus had lower cell surface MHCII expression in both cTEC and mTEC populations (Fig. 4B).

Neutralisation of Hh signalling influences TEC differentiation in WT FTOC
Given that Shh is required for normal TEC development in FTOC, we tested if we could manipulate TEC differentiation in WT FTOC by neutralisation of Hh proteins by treatment with soluble recombinant (r)Hhip. Hhip binds to and sequesters Hh proteins [44], and therefore reduces the endogenous Hh signal in FTOC. Treatment of WT FTOC for 7 days with rHhip reduced the proportion of TEC in the CD45 À population from 78.2 to 62.5% (Fig. 4D). Within the EpCam1 þ population, the mTEC lineage (UEA-1 þ ) was reduced from 25.2% in untreated FTOC to 16.9% in rHhip-treated FTOC, with a corresponding decrease in mTEC cell number (Fig. 4D). We examined cell surface MHCII expression and found that intensity of staining was significantly increased on the mTEC population ( Fig. 5E and F), consistent with the phenotype of Shh À/À FTOC ( Fig  3G). The MHCII high mTEC population are mature and this population contains Aire-expressing mTEC. Interestingly, although MHCII expression was increased in the rHhip treated cultures, the proportion of mTEC that expressed intracellular Aire protein, declined (Fig. 4F), from 13.6% in control cultures to 7.45% in treated cultures.
Neutralisation of endogenous Hedgehog proteins in FTOC thus had a similar effect to Shh-deficiency, as TEC numbers, and the proportion of mTEC were decreased, whereas cell surface MHCII was increased, highlighting the importance of Hh signals for TEC differentiation.

TEC differentiation is reduced in the Shh-deficient adult thymus
To extend our analysis of the role of Shh in TEC differentiation and maintenance, we generated mice with a TEC-specific loss of Shh expression. For this purpose we crossed mice transgenic for the expression of the Cre recombinase under the Foxn1 promoter to animals with a conditional Shh allele to conditionally delete Shh from TEC lineage cells (Shh coKO ). The Shh coKO thymus contained fewer cells overall (Fig. 5A), with a profound impact on TEC. There were fewer TEC than WT (Fig. 5A), and numbers of both cTEC and mTEC lineage cells were significantly reduced, although there was an increase in the proportion of cTEC relative to mTEC in the Shh coKO (Fig. 5BeC). Comparable to the findings with foetal Shh À/À FTOC, cell surface expression of MHCII was increased on mTEC, but the proportion of mTEC expressing Aire protein was reduced (Fig. 5DeE).

Influence of conditional Shh-deficiency on thymocyte populations
Constitutive Shh-deficiency has previously been shown to influence T-cell development at the transition from DP to SP in FTOC, increasing the proportion of mature SP thymocytes [24]. However, it has not been possible to use the constitutive Shh À/À mutants to investigate postnatal T cell development given that a constitutive loss of Shh expression results in embryonic lethality, and the influence of conditional deletion of Shh from TEC on thymocyte development is not known. We therefore examined the influence of conditional deletion of Shh from TEC in the adult Shh coKO thymus on SP thymocyte differentiation. The proportions of CD4SP and CD8SP cells were significantly increased in the Shh coKO thymus, compared to WT (Cre-) littermates, with concommitant significant decrease in the DP population (Fig 5F). The percentage of CD3high cells in the DP, CD4SP and CD8SP populations were higher in the Shh coKO compared to WT thymus ( Fig  5G). CD3 is upregulated on DP cells undergoing positive selection, and its high expression in SP populations indicates maturity of the cells, so the higher proportion or CD3high cells suggests that differentiation from DP to mature SP is heightened in the Shh coKO . Consistent with this, expression of the maturation marker CD24 (HSA), which is downregulated in more mature SP cells, was lower in the Shh coKO populations, with fewer cells falling in the CD24 high marker (Fig. 5G). In addition, we measured cell surface CD5 expression. Levels of cell surface CD5 expression on thymocytes are believed to correspond to TCR signal strength [45], and treatment of WT FTOC with rShh has been shown to decrease cell surface CD5 on thymocytes in FTOC [24]. Interestingly, a greater proportion of both CD4SP and CD8SP cells had high levels of cell surface CD5 in the Shh coKO thymus than in WT (Fig. 5G). Overall, the Shh coKO thymus showed an increase in both mature SP populations. Thus, secretion of Shh by TEC is important both for TEC maturation and T-cell development.

Shh regulates differentiation from DP to SP in the human thymus
Shh is expressed by human TEC [18]. Therefore, in order to test the effect of Shh signalling in human thymocyte differentiation we treated human thymus explants with rShh or neutralising anti-Shh antibody for 5 days, and assessed the SP and DP populations by flow cytometry. Single positive thymocytes accumulated in the thymus explants, and to analyse the mature SP populations, and exclude TCR-negative SP cells that occur before the DP population, we gated on CD3 high cells (Fig. 6A) and analysed the percentage of CD4SP, CD8SP and DP within the CD3 high gate (Fig. 6B). Treatment with rShh decreased the proportion of CD3 high CD4SP and CD3 high CD8SP and increased the proportion of CD3 high DP cells, with 68.7% of the CD3 high gate being CD4SP cells in control-untreated cultures, compared to 49.3% in cultures treated with 0.25 mg/ml rShh. There was a concomitant rise in the DP population, from 13.2% in the control to 29.1% in the rShh-treated cultures (Fig. 6BeC). The proportion of CD4SP cells was more affected by rShh-treatment than proportion of CD8SP cells, so that overall the ratio of CD4:CD8 SP was decreased on rShh treatment, in a dose-dependant manner, as were the ratios of CD4:DP and SP:DP (Fig. 6D). In contrast, when we treated the thymus explants with a neutralising monoclonal antibody against Shh, to neutralise the endogenous Hh proteins in the culture, we found the opposing effect. The accumulation of CD3 high CD4SP cells was increased compared to untreated cultures, and the ratios of CD4SP:DP, CD4SP:CD8SP and SP:DP were all increased (Fig. 6BeD). Thus, the action of rShh-treatment or neutralisation of endogenous Hh proteins in human thymus explants, mirror the influence of the Hedgehog pathway in the mouse thymus, where increased Hh signalling reduces differentiation from DP to SP cell, and reduction in Hh signalling increases differentiation from DP to SP cell [24e27,42].

Discussion
Here we show that Hedgehog signalling regulates the development of TEC during foetal thymus ontogeny, and also influences TEC production and differentiation in the adult steadystate thymus. The use of GBS-GFP reporter mice, and Shh À/À , Shh coKO and Gli3 À/À mouse mutants confirmed that Shh is required for normal TEC differentiation, that it is produced by TEC and that developing TEC transduce Hh signals.
Interestingly, reduction in Hh signalling to TEC, by Shh-deficiency, conditional deletion of Shh from TEC, or neutralisation of endogenous Hh proteins in WT FTOC all led to an increase in cell surface MHCII expression on TEC. Thus, although there are overall fewer TEC in the Shh-deficient thymus, on each individual TEC more selecting MHCII plus peptide complexes for positive and negative selection will be available to developing thymocytes, potentially influencing the outcome of TCR repertoire selection by altering MCHII-restricted TCR antigen dwell time or avidity [46]. Analysis of T cell development in the foetal Shh À/À thymus showed increased differentiation from DP to SP cell, whereas treatment of WT FTOC with rShh decreased differentiation from DP to SP, and the CD4:CD8 ratio [24]. In addition, inhibition of Gli2mediated transcription by transgenic expression of a repressor form of Gli2 in T-lineage cells increased differentiation from DP to SP thymocyte [25,26], whereas increased Hh-mediated transcription by transgenic expression of an activator form of Gli2 in Tlineage cells, or by conditional deletion of Ptch1 from T cells reduced differentiation from DP to SP thymocyte [24,25,47]. Activation of Hh-mediated transcription in T lineage cells has been shown to decrease TCR signal strength [24,48], providing an explanation for the decrease in differentiation from DP to SP on Hh pathway activation and relatively greater impact on the CD4SP lineage on treatment of WT FTOC with rShh. However, our data here suggest that another way in which Shh might impact on differentiation from DP to SP cell is to modulate cell surface expression of MHCII on TEC, influencing signals for positive selection to the CD4 lineage.
Treatment of human thymus explants with rShh or neutralising anti-Shh antibody indicated that Shh also influences differentiation from DP to SP cell, thus underscoring the relevance of these mouse models to human immunity and disease. In both human GWAS studies, patients and experimental animal models, the Hedgehog signalling pathway and its gene targets have been implicated in autoimmunity and allergy [28,49e53], but genetic susceptibilities to different autoimmune diseases are complex [54e57], so in the future it will be interesting to assess the specific contribution of Hedgehog signalling in TEC differentiation and function to the induction and severity of human autoimmune diseases. Interestingly, in primary biliary cirrhosis patients, GWAS have identified both the Hedgehog pathway genes and MHCII genes [54], supporting the idea that the influence of Shh on MHCII expression in the thymus might be important for central tolerance of the CD4 population.
It is important to highlight the influence of Shh on TEC development. Normal differentiation and cellularity of medullary TEC, including the mature MHCIIhigh Aire þ mTEC population, is altered in the absence of Shh production by TEC or by neutralization of endogenous Hh signals. These observations are consistent with the role of other morphogens such as TGF-b, which also negatively regulates mTEC maturation and growth, with the adult TGF-b conditional knockout showing a similar proportional change in the mTEC population as we observed in the adult Shh coKO thymus [14]. Gradient signalling by morphogens such as Shh could fine-tune the correct induction of central T-cell tolerance.

Conclusions
Few factors that influence TEC differentiation have been identified, and here we show that morphogens, such as Sonic Hedgehog, play an important role in this process. Alterations in the Shh signalling pathway have an impact on mTEC development and MHCII surface expression with potential to alter T cell central tolerance. It will in the future be important to investigate the impact of Shh signalling to TEC on other known signals required for mTEC development and function such as TNFRST11a and TGF-b, and the way in which Hh pathway activation integrates with RANK or TGF-b signalling.