Betaglycan (TβRIII) Is Expressed in the Thymus and Regulates T Cell Development by Protecting Thymocytes from Apoptosis

TGF-β type III receptor (TβRIII) is a coreceptor for TGFβ family members required for high-affinity binding of these ligands to their receptors, potentiating their cellular functions. TGF-β [1]–[3], bone morphogenetic proteins (BMP2/4) and inhibins regulate different checkpoints during T cell differentiation. Although TβRIII is expressed on hematopoietic cells, the role of this receptor in the immune system remains elusive. Here, we provide the first evidence that TβRIII is developmentally expressed during T cell ontogeny, and plays a crucial role in thymocyte differentiation. Blocking of endogenous TβRIII in fetal thymic organ cultures led to a delay in DN-DP transition. In addition, in vitro development of TβRIII−/− thymic lobes also showed a significant reduction in absolute thymocyte numbers, which correlated with increased thymocyte apoptosis, resembling the phenotype reported in Inhibin α −/− thymic lobes. These data suggest that Inhibins and TβRIII may function as a molecular pair regulating T cell development.


Introduction
T cell development requires the recognition of self-peptide MHC complexes by immature thymocytes, leading to the selection of a self-restricted and autotolerant T cell repertoire. In addition to the nature of TCR signals triggered by self-peptide recognition, other signals provided by thymic stromal cells, such as those triggered by members of TGF-b superfamily like TGFb, activins/ inhibin and BMP subfamilies have been shown to act as key regulators of apoptosis, survival and cell cycle progression in different cell types [1][2][3].
We and others have described that members of TGF-b superfamily are differentially expressed in the thymus and regulate specific developmental checkpoints, influencing T cell development in [4,5]. Specifically, among TGFbs, only TGFb1 and TGFb2 appear to regulate DN1-DN2 and DN-DP transitions and promote maturation of CD8SP [4]. On the other hand, BMPs and their negative regulators, chordin, noggin and twisted gastrulation (Tsg), are also expressed in the thymus. BMP2 and BMP4 were shown to negatively regulate DN1-DN2, DN3-DN4 and DN-DP transitions [4,5]. Finally, we have recently described that inhibins are abundantly expressed in the thymus by stromal cells and thymocytes [6] and that, addition of exogenous inhibins in FTOCs regulate T cell development at the DN3-DN4, DN-DP, and DP-CD8SP stages [7]. Moreover, endogenous inhibins were required to obtain normal thymocyte numbers and adequate DN-DP transition during in vitro T cell development [7].
A central coreceptor in the canonical signaling pathway of TGF-b is betaglycan, also known as the TGF-b type III receptor (TbRIII), which is a widely expressed membrane-anchored proteoglycan. Structurally, it is characterized by a large extracellular region, containing heparan and chondroitin sulphate chains, and a short cytoplasmic domain that lacks a signaling motif, which has recently been shown to regulate cell processes like apoptosis and cell migration [8,9]. TbRIII-null mice embryos die between E13.5 to E18.5 of embryonic stage by heart and liver defects, caused by an altered TGF-b2-induced mesenchymal transformation process and the incidence of apoptotic events [10]. Recently, it has been described that the absence of TbRIII also compromises normal seminiferous cord formation, Leydig cell function in testis [11] and alters kidney development [12].
Although TbRIII is broadly expressed in many tissues, its presence and the potential function of this receptor in the hematopoietic system remains poorly characterized [26][27][28][29][30][31]. Given that downstream signaling of many TGF-b ligands are regulated by TbRIII to fine tune key cellular processes, here we investigated the expression of TbRIII in the thymus and its potential role in T cell differentiation.

Materials and Methods
Mice 4 to 6 week old C57BL/6 mice were used in our experiments. TbRIII wild type, heterozygous, and null mouse embryos [10] were obtained from synchronized embryonic day 14 (E14) matings of TbRIII heterozygous mice. All animal handling and experimental procedures were done according to the Instituto de Investigaciones Biomedicas ethics guidelines. The study was approved by the ''Comité para el Cuidado y Uso de Animales de Laboratorio (CCUAL)'' of the Institute.

RNA Isolation and RT-PCR
Total RNA was obtained from total fetal (E14-E15 gestation) and adult thymi, E15 thymic stromal cells, sorted adult thymocyte subpopulations, L6E9 and BG22 myoblast cell lines. Testis and brain were used as control tissues. E15 thymic stromal cells were purified as previously described [6]. Sorted thymocyte subsets (DN, DP, CD4 + and CD8 + ) were obtained with FACS Aria cell sorter (BD Biosciences) with a purity of .95%. RNA was isolated using RNA-STAT60 reagent (Tel-Test Inc, Friendslaw, TX) according to manufacturer's protocol. 5-10 mg of total RNA was treated with ''DNA-Free'' reagent (Ambion Inc. Austin TX). cDNA was synthetized using M-MLV RT and oligo dT (both from Invitrogen Inc, Carlsbad, CA) according to manufacturer's recommendations.

Real Time PCR Analysis
The following primers were used: TbRIII Forward 59 -GCCAGACGGCTACGAAGATTT-39, TbRIII reverse 59-AACACTACCACTCCAGCACGG-39, b-Actin Forward 59 -TGGAATCCTGTGGCATCCATGAAAC-39 and b-Actin Reverse 59-TAAAACGCAGCTCAGTAACAGTCCG-39. Measurement of gene expression was performed amplifying cDNA with SYBR Green PCR Core Kit (Applied Biosystems) and analyzed with the ABI PRISM 7000 Sequence Detection System software (Applied Biosystems, Foster City, CA). Amplification conditions for TbRIII and b-actin were an initial step of 5 min at 95uC, followed by 40 cycles of 95uC for 30 sec, 58uC for 1 min and 61.3uC for 1 min. Expression specific gene was calculated using the formula 2 2(DCt) . The calculation of TbRIII expression was normalized to b-actin and performed as previously described [6]. The PCR products were analyzed on a 2% (w/v) agarose gel to confirm purity and size of products.

Flow Cytometry
A total of 0.5-1610 6 cells were treated with Fc block for 30 minutes at 4uC following by staining with the indicated antibodies: FITC-coupled anti-CD4 (RM4-5), PE-coupled anti-CD44 (Pgp-1, obtained from BD Biosciences and Biolegend. For TbRIII detection a polyclonal antibody was used [13]. As secondary reagents, FITC-conjugated goat anti-rabbit IgG (Invitrogen), PEconjugated anti-rabbit IgG (Invitrogen) and APC-conjugated streptavidin (BD Biosciences) were used. Dead cells were gated out depending on forward scattering (FSC) and side scattering (SSC). All samples were captured in a FACsCalibur and FACSAria (BD) and data were analyzed with FlowJoß Tree Star software.

FTOCs
Fetal thymi of TbRIII wild type, heterozygous, and null mouse embryos were obtained from timed matings of heterozygote mice (C57BL/6 background) and genotyped as described above. Fetal thymi were cultured as previously described [7]. Fetal thymi were cultured with or without anti-TbRIII blocking Ab (1/100 dilution) for 3-7 days, and medium was refreshed every third day. On days 0, 3 and 7 of culture, thymic lobes were disaggregated and thymocytes counted and stained for flow cytometric analysis.

Apoptosis Assays
Thymocytes were initially stained with CD4/CD8 antibodies followed by Annexin-V, according to the manufacturer's protocol (BD Biosciences). For the analysis of active caspase 3, cells were permeabilized with Fixation/Permeabilization solution (BD Biosciences) for 1 hr at 4uC and then incubated with anti-active caspase 3-PE (eBioscience, San Diego, CA) for 30 minutes at 4uC.

Statistical Analysis
For FTOCs using the antibody anti-TbRIII, a paired two-tailed Student's test was used to compare lobes from each fetus (control lobe versus treated lobe). For the rest of experiments, an unpaired Students test was used. Asterisks indicate p#0.05(*) and p#0.01(**). P values #0.05 were considered as statistically significant.

TbRIII mRNA is Expressed in Murine Thymocytes and Stromal Cells
We first investigated the expression of TbRIII in the thymus at the mRNA level. Although we expected to find expression on stromal lineage cells (which include epithelial cells, macrophages and dendritic cells), surprisingly, we found significant levels of TbRIII in thymocytes ( Figure 1). This expression was comparable to levels observed in the brain and testis, which were used as control tissues [32]. TGFb superfamily members have an essential function in early T cell development that initiates with the arrival of lymphoid progenitors to the thymic anlage from 13.5 days of gestation [4,5]. In order to determine if TbRIII signaling might have a role in fetal thymocyte development, we analyzed the gene expression of the receptor in E14, E15 and E16 fetal thymi and compared it to that in adult thymi. We observed that TbRIII was significantly expressed in all fetal stages at levels equivalent to adult thymus. Although the expression of TbRIII was slightly elevated E15 fetal thymi, it was not significantly different from other timepoints during development (Figure 1). To directly determine the expression of TbRIII in thymocytes at different stages of development from that in stromal cells, we sorted DN, DP, SP CD4 + and SP CD8 + thymocyte subsets from adult thymi and E15 thymic stromal cells. As expected, we observed that TbRIII mRNA was highly expressed in stromal cells (Figure 1). Remarkably, thymocyte subpopulations at all stages of development expressed significant levels of this proteoglycan with the CD4 + SP subset expressing the highest levels ( Figure 1). Western blot analysis revealed confirmed a 100 kDa protein that corresponds to the core protein of mouse TbRIII in total adult thymocytes, lymph node cells and splenocytes ( Figure S1A).
Thus, this is the first evidence that TbRIII is expressed in thymic stromal cells and also in thymocytes, suggesting its potential role during T cell development.

TbRIII Expression is Regulated during T Cell Ontogeny
Since type I and type II TGFb receptors are differentially expressed in thymocyte subsets and their expression is associated with distinct thymocyte responsiveness to TGF-b superfamily ligands [4,5,33], we investigated whether TbRIII is differentially expressed on the cell surface of developing thymocytes. For this purpose we used a polyclonal antibody directed against the TbRIII ectodomain. Its specificity was confirmed by staining of TbRIII deficient fetal thymocytes and after competition with a soluble form of TbRIII ( Figure S1B and C). Consistent with the mRNA expression data (Figure 1), all thymocyte subsets expressed TbRIII protein on the cell surface, showing an increased percentage of TbRIII + cells within the SP subset ( Figure 2A). Based on mean fluorescence intensity values (MFI), we observed that the DN subset expressed the highest levels of TbRIII, which diminished at DP stage and moderately increased at the SP stages ( Figure 2A).
It is well known that TGFbs, activins/inhibins and BMPs regulate early stages of T cell differentiation [7], we therefore analyzed the expression of TbRIII in immature DN thymocytes by using anti-CD44 and anti-CD25 antibodies to identify DN1, DN2, DN3 and DN4 stages. As shown in Figure 2B, the percentage of DN cells expressing TbRIII cells increased from DN1 to DN2 reaching a peak at the DN3 stage followed by a significant decrease at the DN4 stage. Analysis of MFI confirmed the downregulation of TbRIII at DN4 stage, suggesting that pre-TCR signaling may regulate TbRIII expression in thymocytes. Consistent with these findings, we suggest that TbRIII expression at DN3 stage may be important to integrate the signaling output of BMPs and inhibins influencing the transition of thymocytes to the DP stage. In fact, we recently demonstrated that inhibin A, a ligand with high affinity for TbRIII, promoted DN3 to DN4 transition in FTOCs, suggesting their positive role during TCR b-selection Figure 3. The blocking of TbRIII in FTOCs alters T cell development. E14 thymic lobes were cultured in the presence of anti-TbRIII antibody or in the presence of pre-immune serum (control lobe). At day 3 and 7 of culture thymic lobes were disaggregated, counted and stained with antibodies to CD4, CD8. (A) Representative CD4 versus CD8 staining dot plots. (B). Comparative graphs represent the percentages of DN, DP, CD4SP and CD8SP thymocytes obtained after 3 and 7 days of culture between both treatments. (C) Analysis of cell numbers in non-treated and anti-TbRIII treated FTOCs at day 3 and 7. Data are representative of two independent experiments. Mean values 6 SEM are shown (n = 7 per group for day 3, and n = 9 per group for day 7). Asterisks indicate *p#0.05. doi:10.1371/journal.pone.0044217.g003 process [7]. Moreover, an interesting mechanism was described for BMPs, where the expression of twisted gastrulation (Tsg) is induced after pre-TCR signaling to reverse the BMP2/4dependent arrest and to promote T cell differentiation [34]. However, future experiments will be necessary to discriminate the contribution of TbRIII in each TGFb ligand-mediated functions at this stage.
Since TbRIII is expressed in SP thymocytes, we next investigated whether upregulation of this receptor might accompany the process of positive selection. We used TCRb, CD5 and CD69 as markers to analyze TbRIII expression in DP cells that have undergone positive selection. No correlation was observed between the expression of TbRIII and the levels of CD5, CD69 and TCRb in DP thymocytes, suggesting that TbRIII expression in thymocytes is not modified by TCR signaling (data not shown). However, analysis of mature SP thymocytes indicated that TbRIII is upregulated in CD4 + SP CD69 2 compared to CD4 + SP CD69 + thymocytes ( Figure 2C, left panels), as well as in CD62L hi CD4 + SP and CD62L + CD8 + SP compared to CD62L 2 cells ( Figure 2C, right panels). Altogether, these data show that TbRIII expression may be associated with the terminal differentiation of thymocytes and is preferentially expressed in the most functionally mature CD4 + and CD8 + SP cells (CD69 2 and CD62L +/hi , respectively).

TbRIII Blocking Results in Delayed T Cell Development
TbRIII ligands TGF-bs, inhibins and some BMPs, have been reported to regulate specific checkpoints during T cell differentiation [4]. To directly determine if TbRIII has a role in T cell development, we evaluated the requirement of this co-receptor by performing blocking assays in FTOCs with specific antisera directed against its extracellular domain. This strategy has been previously employed to evaluate the requirement of Tb RIII for epithelial-mesenchymal transition [17] and to restore the T cell stimulatory function of DCs suppressed by TGFb [29]. No differences were observed in absolute cell numbers or in percentages of DN1-DN4 thymocytes at day 3 of culture under blocking conditions ( Figure S2). In addition, when we examined T cell development at day 7 of culture, a significative decrease of DN2 subset and a slight reduction was observed in anti-TbRIII treated lobes ( Figure S2C and D). Interestingly, blocking of TbRIII signaling significantly reduced the proportion of DP thymocytes that was associated with a corresponding increase in the DN subset ( Figure 3A and B) indicating that TbRIII may act by regulating pre-TCR mediated signals, as has been reported for its ligands, BMPs and inhibins [7,34]. This data correlate with the downregulation of TbRIII expression observed between DN3-DN4 and between DN and DP (Figure 2A and 2B). As shown in Figure 3C, analysis of thymocyte cell numbers showed a slight decrease, although not significant, in DP and CD4SP cell numbers at day 3 of culture in fetal lobes treated with anti-TbRIII antiserum.

T Cell Development is Attenuated in TbRIII Null Embryos
To further confirm the role of TbRIII during thymocyte development, we performed FTOCs of TbRIII null embryos. TbRIII null mice show intrauterine lethality due to proliferative defects in heart and apoptosis in liver, beginning at E13.5 and with higher mortality by E16.5 [10]. Therefore to analyze the impact of TbRIII deficiency in thymocyte development and to reduce deleterious systemic defects, we isolated E14 fetal thymi obtained from TbRIII heterozygous pregnant females and analyzed T cell development in FTOC. We observed no differences in cell numbers at day 0 between all of the genotypes ( Figure 4A). In addition, analysis of DN immature subsets showed no significant differences in the absence of TbRIII ( Figure S3). However by day 3 and significantly by day 7, thymocyte numbers were greatly reduced in TbRIII null fetal thymi compared to wild type thymi. Consistently, analysis at day 7 of TbRIII 2/2 FTOCs showed a significant reduction in the percentage of DP thymocytes, which was accompanied by an increase in DN thymocytes ( Figure 4B), which correlates with the reduced cellularity observed in the absence of TbRIII ( Figure 4A). In fact, analysis of cell numbers in TbRIII 2/2 FTOCs at day 7 clearly show a significant decrease of DP and consequently in CD4SP and CD8SP thymocytes ( Figure 4C). No differences in cell numbers between TbRIII +/+ and TbRIII 2/2 fetal thymi were observed at day 3 of culture. (Figure 4C). This reduction in thymus cellularity is more pronounced in TbRIII 2/2 FTOCs than under anti-TbRIII blocking conditions, suggesting that the antibody might not completely block all the available TbRIII (membrane bound and/or soluble) or alternatively, it may only affect membrane bound TbRIII-mediaded actions, while in TbRIII deficient thymocytes, also ligand-independent functions are absent, resulting in a more marked phenotype.
Although the kinetics are different, our results show that DN-DP transition is decreased when TbRIII signaling is compromised by use of blocking antibodies or by genetic deletion of the receptor. These data support the notion that TbRIII regulates DN-DP transition during thymocyte differentiation. In line with these findings, despite to the difficulty to discriminate the contribution of TbRIII in the actions mediated by each ligand during thymocyte development, we have recently demonstrated that inhibin a 2/2 thymic lobes [7] also show reduced thymocyte numbers and a delayed DN-DP transition. We propose that TbRIII and inhibins may function as a molecular pair to regulate T cell differentiation, a functional association observed in other cell types [35].

Increased Apoptosis in the Absence of TbRIII
Enhanced apoptosis of thymocytes in TbRIII 2/2 thymic lobes might be the mechanism responsible for the impaired DN-DP transition and the reduced cellularity observed during thymocyte differentiation. This prediction arises from the study of fetal liver of TbRIII null embryos, which showed increased apoptosis associated with a significant reduction in cellularity in liver [10]. To directly test this, we analyzed the extent of ongoing apoptosis at day 7 of culture by measuring the levels of the active form of caspase 3 in all thymocyte subpopulations ( Figure 5A). Interestingly, we observed a significant increase in the percentage of active caspase 3 + cells in the DN, DP and CD4 + SP subpopulations from TbRIII 2/2 fetal thymi compared to wild type thymi ( Figure 5A). In accordance, MFI for active caspase 3 also showed an increase in this apoptotic mediator in DP and DN subpopulations, although  the latter did not reach statistical significance, suggesting that DP thymocytes are more susceptible to die in TbRIII 2/2 fetal thymi ( Figure 5A). In addition, annexin V staining confirmed the increased apoptosis of DP and CD4 + SP thymocytes in the absence of TbRIII ( Figure 5B).
In agreement with our data, several findings highlight a crucial role for TbRIII as a regulator of apoptosis including the higher levels of caspase 3 and downregulation of the prosurvival factor Akt, which impact the integrity and function of liver [10]. In addition, overexpression of TbRIII leads to protection of cardiac fibroblasts from hypoxia-induced apoptosis by reversing caspase 3 activation, Bax upregulation and inducing Bcl-2 downregulation [36].
Although most of the functions mediated by TbRIII may involve the interaction with TGFb members, we cannot rule out additional functions mediated by the intracellular domain of this co-receptor, as it has been recently involved in cell migration and apoptosis in a ligand independent fashion [37,38]. Indeed, after shedding of the TbRIII ectodomain, the resulting transmembranecytoplasmic domain is cleaved by c-secretase, influencing the TGFb signaling response [39]. Thus, we may argue that this TbRIII cytoplasmic domain could trigger transduction signals in a similar fashion as Notch [40], acting as novel regulator of early T cell differentiation.
In summary, our data provide the first evidence that TbRIII plays a key role during T cell development: it is differentially expressed during T cell ontogeny and regulates DN-DP transition by protecting thymocytes from apoptosis. However, many questions remain unanswered concerning the role of TbRIII in T cell immunobiology. Indeed, since TGFb is a critical regulator of T cell-mediated immunity, it is feasible to propose that the described effects of TGFb members in immune cells may be regulated by TbRIII. In this sense, the generation of conditional knockout models for TbRIII will allow us to elucidate the function of this receptor in specific cell types, including mature T cell populations.