T Cell Development Critically Depends on Prethymic Stromal Patched Expression Free

Hedgehog (Hh) signaling is involved in the regulation of a variety of differentiation processes in the embryo and in the adult. Patched (Ptch), the receptor for Hh, inhibits its signaling partner Smoothened (Smo). Binding of Hh to Ptch or inactivating Ptch mutations suspend the inhibition of Smo. The subsequent nuclear translocation and activation of the Gli family of transcription factors cause activation of signaling and the expression of downstream target genes, including Gli1 and Ptch itself (reviewed in Ref. 1).

Recently, it was shown that all major components of the Hh-signaling cascade (e.g., Hh, Ptch, Smo, and Gli) are important for T cell development (2–9). In the thymus, Hh is produced by the thymic stroma, where Ptch and its signaling partner Smo are expressed (2, 6). Ptch and Smo are also expressed in CD4⁻CD8⁻ double-negative (DN) thymocytes (5) and are less pronounced in CD4⁺CD8⁺ double-positive (DP) thymocytes (2, 5).

Using tPtch^−/− mice, in which Ptch was ubiquitously deleted in 6–8-wk-old animals by tamoxifen-induced activation of CreERT2 (ERT2) in composite Ptch^floxflox ERT2^+/− mice, we recently showed that Hh signaling is required for the development of the T cell lineage as early as at the level of the common lymphoid progenitor (CLP) in the bone marrow (BM). In addition, T cell progenitors did not differentiate beyond the DN1 stage within the thymus of these mice (6). However, a cell-autonomous function of Ptch in developing T cells in this model was not excluded.

To answer the question whether Ptch fulfills cell-autonomous functions in T cell development, we pursued several experimental approaches using conditional Ptch^floxflox mice. The role of Ptch in the stromal cell compartment was addressed using tPtch^−/− mice reconstituted with wild-type (wt) BM. Transplantation of glucocorticoid receptor (GR)-deficient BM into tPtch^−/− mice allowed us to exclude any involvement of increased levels of the stress hormone corticosterone from our model. To study T cell-intrinsic Ptch functions during intrathymic T cell development, we deleted Ptch in T cells at the DN3 stage. Furthermore, a putative function of Ptch in thymic stromal cells was investigated by transplantation of Ptch-deficient thymi under the kidney capsule of wt mice. In a fourth experiment, we analyzed T cell development in the thymus upon Ptch deletion in T cells and in the thymic stroma using fetal thymus organ culture (FTOC). Finally, we studied the role of Ptch for the appearance of earliest thymic progenitors (ETPs) in the thymus. Collectively, we concluded that only Ptch expressed in stromal cells at the prethymic stage impacts T cell development.

All animal experiments were performed with consideration of the necessary legal requirements.

Ptch^flox/flox ERT2^+/− and Ptch^flox/flox ERT2^−/− mice were derived from a cross between Ptch^flox/flox (6) and Rosa26CreERT2 (ERT2) (10) mice. tPtch^−/− mice were generated by induction of ERT2 activity in 6–8-wk-old Ptch^flox/flox ERT2^+/− mice by i.p. injection of 1 mg tamoxifen (solvent was sterile ethanol/sunflower oil, 1:10) on five consecutive days (cumulative dose, 5 mg), as described (6).

To obtain Ptch^flox/flox CD4Cre^+/− mice, Ptch^flox/flox mice were crossed with CD4Cre mice, which express the Cre-recombinase under the CD4Cre promoter as early as in the DN3 stage of T cell development (11).

CD4Cre-recombinase was detected using the primer combination 5′-CCA GGC TAA GTG CCT TCT CTA CA-3′/5′-AAT GCT TCT GTC CGT TTG CCG GT-3′. Genotyping of Ptch^flox/flox and ERT2 mice and detection of the Cre-mediated excision of exon 8 and 9 of the Ptch^flox allele (Ptch^del mutation) were performed as described (6, 12).

GR^flox/flox and GR^flox/flox lckCre^+/− mice were described previously (13) and were genotyped using primers that distinguish the 225-, 275-, and 390-bp fragments of the wt GR, GR^flox, and GR^del alleles, respectively (14).

GR^flox/flox lckCre^+/− and GR^flox/flox mice were pure C57BL/6J, whereas all other mice were maintained on a mixed C57BL/6N/129Sv genetic background, unless indicated otherwise.

To obtain fetal thymi from embryos at embryonic day 15.5 (E15.5) postcoitum, Ptch^flox/flox mice were mated with Ptch^flox/flox ERT2^+/− mice. Ptch^flox/flox and Ptch^flox/flox ERT2^+/− embryos were genotyped by PCR using genomic DNA isolated from embryonic paw. Each lobe of the fetal thymus was cultured separately at 37°C and 5% CO₂ on 0.8-μm-pore filters (Nucleopore membrane, Whatman) at the liquid/air interface while floating on RPMI 1640 medium (10% FCS, 1% penicillin/streptomycin) supplemented with 10 μM tamoxifen (first lobe) or vehicle (DMSO; second lobe). Culture medium was refreshed every second day and changed completely after 6 d. Four, eight, and twelve days after dissection, lobes were harvested, and genomic DNA was analyzed for recombination efficiency at the Ptch^flox locus. Lobes were harvested after 12 d for RNA isolation or flow cytometric analyses of the T cell populations. T cells were depleted by incubation with 2′-desoxyguanosine (dG) to selectively analyze the stromal compartment in FTOCs (15). To this end, FTOCs were cultured for 5 d in media supplemented with 1.35 mM dG and 10 μM tamoxifen, followed by incubation for 3 d with 10 μM tamoxifen alone. T cell depletion was confirmed by FACS analysis using anti–CD4-PerCP, anti–CD8α-PE-Cy7, anti–CD44-FITC, and anti–CD25-allophycocyanin-Cy7 Abs. Recombination efficiency at the Ptch^flox locus and Ptch^del expression were analyzed as described below.

Genomic DNA was isolated from FTOC or from thymocytes separated with anti-Thy1.2 microbeads of 10-wk-old mice, according to the manufacturer’s instructions (Miltenyi Biotec, Bergisch Gladbach, Germany). Recombination efficiency at the Ptch^flox locus was quantified on an ABI PRISM 7900 sequence detection system (Applied Biosystems) by real-time PCR, as described recently (12). Total RNA was extracted from the same cell populations using TRIzol reagent (Invitrogen, Carlsbad, CA). RNA was reverse transcribed using random hexamers and SuperScriptII reverse transcriptase (Invitrogen). SYBR Green-based assays were performed to detect wt Ptch (5′-AAA GCC GAA GTT GGC CAT GGG TAC-3′/5′-TGC TTG GGA GTC ATT AAC TGG A-3′) or Ptch^del (5′-AAA GCC GAA GTT GGC CAT GGG TAC-3′/5′-TTA AAC AGG CAT AGG CAA GCT GAC-3′) transcripts. The primer pair 5′-GGT CAT CTA CGA GAC CAA CTG C-3′/5′-GTG TCT TCA GGT TCT CCA GGC-3′ was used to quantify Gli2 transcripts. Quantification of Gli1 transcripts was analyzed as described (12). Amplification of 18S rRNA served to normalize for the amount of sample cDNA. All individual DNA or cDNA samples were measured in triplicate. Ptch and mGapd transcripts were detected as described elsewhere (6).

Blood smears were prepared from EDTA blood samples and stained according to the Pappenheim method. For immunohistochemical analyses, paraffin sections of thymi were boiled in 10 mM citric acid (pH 6). Sections were stained with anti-CD3 Ab (1:50; Serotec) or anti-B220 Ab (1:200; BD Biosciences), a secondary biotinylated rabbit anti-rat Ab (1:50; Dako) and alkaline phosphatase-conjugated streptavidin (1:300; Dako). The staining was visualized by Neufuchsin as chromogene.

To analyze developmental stages of T cells, single-cell suspensions of thymi of adult mice or of FTOCs were prepared. Thymocytes of adult mice were stained with anti–CD4-biotin, anti–CD8α-PE, anti–CD25-FITC, or anti–CD44-PE-Cy5 Abs in combination with streptavidin-PE-Cy7 (BD Biosciences). Thymocytes of FTOC were stained with anti–CD3-PerCp, anti–CD8α-PE-Cy7, anti–CD4-allophycocyanin, anti–CD44-biotin, anti–CD25-allophycocyanin-Cy7, and streptavidin-FITC. All Abs were obtained from BD Biosciences. A BD FACSCantoII and a BD LSR II cytometer (Becton Dickinson) were used for data collection, and analysis was performed using FlowJo (Treestar) or BD FACSDiva software, respectively.

For analyses of ETPs, thymi of tPtch^−/− mice were harvested 19 d after the first tamoxifen injection and carefully minced in RPMI 1640. The cell suspension was centrifuged at 300 × g for 4 min at 4°C, washed in PBS, and centrifuged again, and the cell pellet was resuspended in 50 μl PBS. To block nonantigenic-specific binding of Ig to FcγIII and FcγII receptors, 25 μg rat-anti-mouse-CD16/CD32 (Mouse BD Fc Block; BD Biosciences) was added, and the suspension was incubated for 15 min on ice. Cells were washed again, centrifuged, and resuspended in 50 μl PBS. Next, cells were incubated with 15 μl anti–Lin-FITC Ab mixture (anti-CD3ε, anti-CD8α, anti-CD8β, anti-Ter119, anti-CD19, anti-B220, anti-CD11c, anti-NK1.1; but not anti-CD4, because some ETPs express CD4), anti–CD25-PE, and anti-CD117(c-Kit)-allophycocyanin for 45 min on ice. After washing, the stained cells were resuspended in PBS. All cells isolated from the thymi were analyzed on a FACSCalibur (Becton Dickinson). BM of a 6-wk-old C57BL/6N mouse served as a positive control.

A total of 2 × 10⁶ BM cells was isolated from wt B6.SJL-PtprcaPepcb/BoyJ (CD45.1-congenic C57BL/6J) mice, GR^floxflox lckCre^+/− mice, or GR^flox/flox mice and engrafted in 9-wk-old Ptch^flox/flox ERT2^+/− and Ptch^flox/flox mice after irradiation (11 Gy).

Reconstitution of the wt BM engraftment was analyzed by monitoring the peripheral blood chimerism 6 wk after transplantation using the CD45.1 (B6.SJL-PtprcaPepcb/BoyJ donor)/CD45.2 (mixed 129sv/C57BL/6N host) isogenic system. The following Abs were used for flow cytometric analyses: anti–CD45.1-PE, anti–CD45.2-allophycocyanin, anti–CD3-PerCp, and anti–CD19-FITC (BD Biosciences).

The peripheral blood chimerism of Ptch^flox/flox mice transplanted with GR^flox/flox lckCre^+/− and GR^flox/flox BM was monitored by genotyping for wt Ptch, Ptch^flox, Ptch^del, GR^flox, and GR^del alleles on genomic DNA isolated from peripheral blood before and 6 wk after transplantation and by flow cytometry using anti–CD4-PerCp, anti–CD8α-allophycocyanin, and B220-PE-Cy7 Abs (BD Biosciences).

The Ptch^del mutation in all reconstituted Ptch^flox/flox ERT2^+/− mice was induced by five consecutive i.p. tamoxifen injections, as described (6). All tamoxifen-treated Ptch^flox/flox ERT2^+/− mice are named tPtch^−/−; the origin and genotype of the respective transplanted BM are indicated after the strain name using a double slash (e.g., tPtch^−/−//wt BM).

Individual thymic lobes were dissected from E15.5 Ptch^flox/flox ERT2^+/− and Ptch^flox/flox littermate mice (CD45.2-congenic). One lobe of each genotype was transplanted under the kidney capsules of the same adult wt B6.SJL-PtprcaPepcb/BoyJ (CD45.1-congenic) recipient mouse (16). Seven days after transplantation, tamoxifen was injected i.p., as described above, to induce recombination of the Ptch^flox locus in the grafted thymi. This resulted in one lobe carrying the Ptch^del mutation and one wt lobe within the same recipient mouse. Recombination efficiency of the Ptch^flox locus and development of T cells in the grafted fetal thymi were analyzed 14 or 21 d after the first tamoxifen injection. The following Abs were used for flow cytometric analyses: anti–CD45.1-PE, anti–CD45.2-APC, anti–CD4-PerCp, anti–CD8α-PE-Cy7, anti–CD44-FITC, and anti–CD25-allophycocyanin-Cy7 (BD Biosciences).

To confirm resistance of GR^flox/flox lckCre^+/− thymocytes to glucocorticoid-induced apoptosis, thymocytes were isolated from whole thymi of tPtch^−/−//GR^flox/flox BM, tPtch^−/−//GR^flox/flox lckCre^+/− BM, Ptch^flox/flox//GR^flox/flox lckCre^+/− BM, and C57BL/6N wt mice. A total of 3 × 10⁵ cells/well was incubated in RPMI 1640 supplemented with or without 10⁻⁸ M dexamethasone for 20 h. Subsequently, cells were harvested and stained with Annexin V-FITC, 7-aminoactinomycin D (7-AAD), anti–CD8α-PE-Cy7, and anti–CD4-allophycocyanin (BD Biosciences). Because CD4⁺CD8⁻ DP T cells are particularly sensitive to dexamethasone-mediated apoptosis (17), this subpopulation was analyzed for viable, early apoptotic, late apoptotic, and dead cells, which are Annexin V⁻/7-AAD⁻, Annexin V⁺/7-AAD⁻, Annexin V⁺/7-AAD⁺, and Annexin V⁻/7-AAD⁺, respectively.

Blood samples of tPtch^−/−//wt BM and Ptch^flox/flox//wt BM mice were collected from the retro-orbital plexus before and 10 and 15 d after the first tamoxifen injection. After coagulation for 2 h at 4°C, samples were centrifuged at 1500 × g for 30 min, and serum was collected. Serum was stored at −80°C until analysis. Serum corticosterone levels were measured using a corticosterone-specific RIA, according to the manufacturer’s instructions (MP Biomedicals, Eschwege, Germany).

The Mann–Whitney U test was performed to determine the significance of the results from quantitative RT-PCR analyses and analyses of ETP cell numbers; p values < 0.05 were considered significant and were estimated when sample number was n ≥ 3.

Previous experiments suggested that Ptch function in CLPs itself is dispensable for proper lymphocyte development, whereas the inductive environment of stromal components is required for their differentiation (6). To support this notion, we analyzed thymocyte development in mice, in which Ptch was solely ablated in stromal nonhematopoietic cell compartments (thymus and BM). For this purpose, whole BM of CD45.1⁺ wt mice (B6.SJL-PtprcaPepcb/BoyJ) was transplanted into lethally irradiated Ptch^flox/flox and Ptch^flox/flox ERT2^+/− mice. After establishment of chimerism (Fig. 1A, Supplemental Fig. 1A), all mice were treated for five consecutive days with tamoxifen to induce the Ptch deletion in Ptch^flox/flox ERT2^+/− mice. Hereafter, the respective mice are named Ptch^flox/flox//wt BM and tPtch^−/−//wt BM.

Fifteen days after induction of the Ptch deletion, tPtch^−/−//wt BM mice showed a strong involution of the thymus, with a concomitant reduction in the absolute numbers of lymphocytes and DP cells, compared with Ptch^flox/flox//wt BM mice (Fig. 1B). In addition to the DP fraction, the DN2 and DN3 T cell populations in tPtch^−/−//wt BM mice were lost (Fig. 1C, Supplemental Fig. 1B). This is very similar to the T cell phenotype of untransplanted tPtch^−/− mice 19 d after induction of the mutation (6), confirming a crucial role of Ptch expression in nonhematopoietic stromal cells for proper thymocyte development.

Mice lacking Ptch ubiquitously or selectively in the stromal compartment are not solely characterized by massive T cell depletion; they also develop tumor precursor lesions in a variety of tissues and various organ failures (6, 12). This results in a general diseased condition and is accompanied by strongly elevated levels of the stress hormone corticosterone, as exemplified by the analysis of blood samples obtained from Ptch-mutant and healthy control mice (Fig. 1D).

To exclude that stress-induced lymphocyte death significantly contributes to the compromised development of the T cell lineage in tPtch^−/− mice, we reconstituted lethally irradiated Ptch^flox/flox ERT2^+/− mice with BM of GR^flox/flox lckCre^+/− or GR^flox/flox control mice. Because of the deletion of the GR in the T cell lineage, thymocytes from GR^flox/flox lckCre^+/− mice are protected from apoptosis induced by increased levels of corticosterone (13, 18). If stress accounts for the massive reduction in T cell numbers after disruption of the Ptch gene, such an effect should not be observed in tPtch^−/−//GR^flox/flox lckCre^+/− BM mice because of the lack of the GR. In contrast, tPtch^−/−//GR^flox/flox BM control mice would normally succumb to lymphocyte apoptosis and, thus, impaired T cell development. In contrast, if corticosterone-induced T cell depletion does not play a role in T cell depletion, reduced T cell numbers should be observed in tPtch^−/−//GR^flox/flox lckCre^+/− mice, as well as in tPtch^−/−//GR^flox/flox BM mice.

The establishment of successful chimerism was confirmed by genotyping genomic DNA isolated from peripheral blood samples 6 wk after transplantation of GR^flox/flox lckCre^+/− BM or GR^flox/flox BM in Ptch^flox/flox ERT2^+/− or Ptch^flox/flox control mice (Fig. 2A). Although Ptch^flox and wt GR alleles were detected in blood DNA before transplantation, wt Ptch and GR^del or GR^flox alleles predominated after reconstitution of the mice with donor GR^flox/flox lckCre^+/− or GR^flox/flox BM, respectively (Fig. 2A).

Fifteen days after induction of the Ptch mutation, flow cytometric analyses of thymocytes derived from tPtch^−/−//GR^flox/flox lckCre^+/− BM mice revealed a largely similar T cell phenotype as for tPtch^−/−//GR^flox/flox BM mice, characterized by a decrease in absolute lymphocyte and DP thymocyte numbers in comparison with control Ptch^flox/flox//GR^flox/flox lckCre^+/− BM mice (Fig. 2B, 2C). Thus, all tPtch^−/− mice, regardless of whether they had been transplanted with GR^flox/flox BM, GR^flox/flox lckCre^+/− BM, or wt BM (Figs. 1B, 1C, 2B, 2C), developed a global thymocyte deficiency, as previously described for untransplanted tPtch^−/− mice (6). Although we cannot rule out a minor contribution of stress-induced corticosterone release to the reduction in thymocyte numbers, our results demonstrated that Ptch expression itself primarily accounts for the disturbed T cell development in our model.

To prove that thymocytes of mice transplanted with GR^flox/flox lckCre^+/− BM were insensitive to the proapoptotic effects of glucocorticoids, thymocytes from the same mice as in Fig. 2A–C were isolated and treated ex vivo with or without the synthetic glucocorticoid dexamethasone. Twenty-hours later, cell death was determined by staining for CD4 and CD8 in combination with 7-AAD and Annexin V. DP thymocytes, the population most sensitive to glucocorticoid-induced cell death, were largely resistant to dexamethasone-induced early and late apoptosis in tPtch^−/−//GR^flox/flox lckCre^+/− BM (and Ptch^flox/flox//GR^flox/flox lckCre^+/− BM) mice, whereas DP thymocytes in control mice (tPtch^−/−//GR^flox/flox BM or C57BL/6N wt) were not (Fig. 2D, data not shown). Thus, our data allowed for the conclusion that the loss of T cells after deletion of Ptch is not a consequence of stress.

Although our previous experiments were in favor of an important role for Ptch expression in stromal cells, an additional T cell-intrinsic function of Ptch during intrathymic T cell development could not be excluded. Therefore, we crossed Ptch^flox/flox mice with CD4Cre mice. The CD4Cre deleter expresses the Cre-recombinase under the control of the CD4 promoter, which is active as early as the DN3 stage of T cell development (11). As shown in Fig. 3A, anti–Thy1.2-purified thymocytes of 10-wk-old Ptch^flox/flox CD4Cre^+/− mice showed an almost complete recombination at the floxed Ptch locus (resulting in the Ptch^del allele; Fig. 3A). This was also demonstrated by loss of wt Ptch and a concomitant increase in Ptch^del allele expression using transcript-specific primers (Fig. 3B). Despite effective Ptch deletion, this did not result in significant activation of Hh signaling in anti-Thy1.2 purified thymocytes, as revealed by unchanged Gli1 and Gli2 expression levels (Fig. 3B). In addition, thymocyte development in Ptch^flox/flox CD4Cre^+/− mice was normal and indistinguishable from Ptch^flox/flox control mice, as revealed by flow cytometric analysis of CD4 and CD8 surface levels, as well as for the different DN stages on the basis of CD25/CD44 expression (Fig. 3C, Supplemental Table I). Finally, no abnormal cells were detected in blood smears or in H&E, anti–CD3-Ab, or anti–B220-Ab stainings of thymi obtained from Ptch^flox/flox CD4Cre^+/− mice (data not shown). Thus, T cell-autonomous Ptch expression from the DN3 stage on is not required for proper T cell development.

To investigate whether Ptch in the thymus stroma was required for intrathymic T cell development in vivo, we transplanted thymic lobes derived from E15.5 Ptch^flox/flox ERT2^+/− and Ptch^flox/flox embryos under the kidney capsule of CD45.1⁺ congenic wt mice. One week later, the Ptch mutation in the transplants was induced by tamoxifen, which corresponds to the time when the grafted Ptch-deficient and wt thymi, each on one side of the recipient mouse, started to become populated with wt lymphoid progenitor cells. This strategy allows for the investigation of T cell development in the presence or absence of Ptch in the thymic stroma.

Host- and graft-derived intrathymic T cell populations were distinguished by FACS using anti-CD45.1 (host) and anti-CD45.2 (graft) Abs. Fourteen (data not shown) and 21 d after the first tamoxifen injection, the transplanted lobes had been fully populated with host-derived T cells (Fig. 4A). Flow cytometric analysis of CD4 and CD8 did not show any difference between tamoxifen-treated Ptch^flox/flox ERT2^+/− and Ptch^flox/flox control lobes (Fig. 4A). Of note, Ptch ablation in tamoxifen-treated Ptch^flox/flox ERT2^+/− lobes was nearly complete, as demonstrated by FACS analysis of the cell debris remaining after T cell isolation, which corresponds to the thymic stroma (Fig. 4B). Thus, our data indicated that Ptch expression in the thymic stroma is dispensable for proper T cell development in vivo.

We next studied T cell-intrinsic and -extrinsic Ptch functions in the thymus. For this purpose, we analyzed thymocyte development in FTOCs derived from E15.5 Ptch^flox/flox ERT2^+/− embryos after tamoxifen-induced Ptch deletion ex vivo.

In FTOCs, T cell development is independent of extrathymic ETP supply and solely depends on development of ETPs already immigrated into the thymus. After 4, 8, and 12 d of culture in tamoxifen-supplemented medium, the recombination efficiency at the Ptch^flox locus in Ptch^flox/flox ERT2^+/− lobes was 81, 89, and 100%, respectively (Fig. 5A). The deletion of Ptch in FTOC was accompanied by the loss of wt Ptch and the simultaneous increase in Ptch^del transcripts, and it resulted in significant activation of Hh signaling, as revealed by upregulation of Gli1 and Gli2 transcripts (Fig. 5B).

To verify that Ptch was deleted in the stromal compartment of the FTOCs, E15.5 thymi derived from Ptch^flox/flox ERT2^+/− and Ptch^flox/flox embryos were cultured for 8 d with tamoxifen and concomitantly treated with dG for 5 d. This resulted in a complete loss of thymocytes (Fig. 5C, Supplemental Fig. 2), as well as efficient deletion of Ptch in the remaining thymic stroma (Fig. 5D). Again, this resulted in the loss of wt Ptch and an increase in Ptch^del transcripts, respectively (Fig. 5E).

Microscopic analysis revealed a normal development of Ptch-deficient lobes (data not shown). In addition, flow cytometric analyses showed no significant differences between tamoxifen-treated Ptch^flox/flox ERT2^+/− and Ptch^flox/flox control FTOCs with regard to the relative numbers of single-positive (SP), DP, and DN thymocytes, as well as the four stages of DN thymocytes (Fig. 5F, Supplemental Fig. 3, Supplemental Table II).

Because Ptch ablation was almost complete in all cellular compartments of tamoxifen-treated Ptch^flox/flox ERT2^+/− lobes, our data showed that Ptch, regardless of whether it was expressed in developing thymocytes or the thymic stroma, is dispensable for intrathymic T cell development throughout all DN and later stages. These results also indicated that Ptch deletion in intrathymic ETPs does not impact T cell development.

Having excluded a crucial role for Ptch expression in the thymus, we addressed its function in developmental stages before T cell progenitors settle into the thymus. To this end, we analyzed thymic ETPs in tPtch^−/− mice at day 19 after the first injection of tamoxifen.

In the thymus, ETPs reside in the DN1 fraction of thymocytes, which is developmentally heterogeneous and includes B lymphocytes, granulocytes, macrophages, NK cells, and the respective progenitors (19). ETPs are the only cell population in the thymus known to contribute significant numbers of precursors to T cell development (19, 20) and to give rise directly to DN2 and DN3 thymocytes that express CD25 (20, 21). ETPs are characterized by the lack of lineage marker expression (CD3ε, CD8α, CD8β, Ter119, CD19, B220, CD11c, NK1.1) combined with a CD44^hiCD25⁻cKit^hi expression profile; thus, they can be studied in the thymus by FACS analysis (Fig. 6A) (21).

Interestingly, the absolute number of ETPs was dramatically reduced (154-fold) in tPtch^−/− thymi compared with control thymi (Fig. 6B). Thus, in combination with the aforementioned notion that Ptch deletion in intrathymic ETPs does not impact T cell development, these data showed that Ptch function is necessary for intrathymic settlement or for the development of thymic-homing progenitors that give rise to ETPs.

Using tPtch^−/− mice, we previously demonstrated that Ptch acts as a master switch for diversification of hematopoietic stem cells (HSCs) and identified the stromal cell compartment as a Ptch-dependent critical inducer in this process (6). In this study, we investigated the role of Ptch in T cell development by an in-depth analysis of its T cell-extrinsic and -intrinsic function.

It is well known that stress results in elevated corticosterone levels, which corrupt T cell development and survival. Although mature thymocytes are protected from glucocorticoid-induced cell death in vivo, immature DP thymocytes readily undergo apoptosis after administration of the synthetic glucocorticoid dexamethasone or as a result of stress (17). Because Ptch-deficient mice develop tumor precursor lesions and are in a general diseased condition (6), we analyzed whether stress and, thus, elevated corticosterone levels, contributed to the compromised T cell phenotype in these mice. In fact, lack of the GR in the T cell lineage did not prevent the massive reduction in thymocyte numbers after disruption of the Ptch gene in nonhematopoietic cells, thereby excluding any major effects of corticosterone and stress on T cells in tPtch^−/− mice.

Our data also excluded a T cell-intrinsic function of Ptch in T cell development. Initially, this interpretation was based on the fact that T cells develop normally in Ptch^flox/flox CD4Cre^+/− mice, as well as on our previous finding showing normal T cell development in mice lacking Ptch exclusively in the hematopoietic compartment (6). Intrinsic inactivation of Ptch in thymocytes of Ptch^flox/flox CD4Cre^+/− mice did not result in activation of Hh signaling in whole thymus (data not shown) or in isolated T cells. Lack of induction of Hh signaling was demonstrated in another model in which the Ptch gene was disrupted using lckCre transgenic mice, which express Cre at an earlier T cell stage than do CD4Cre mice (7). The finding that T cell-intrinsic Ptch deletion does not enhance Hh signaling can be explained by the lack of cilia on T cells (22, 23). Because cilia are important for Ptch/Smo-mediated Hh signaling, as well as for related tumorigenesis (24), their absence on T cells may explain the lack of any abnormalities with respect to T cell development, infections, or T cell-based cancers in Ptch^flox/flox CD4Cre^+/− mice up to an age of 500 d (A. Uhmann and H. Hahn, unpublished observations).

Further insight into the role of Ptch in the thymus was obtained by in vivo transplantation and ex vivo FTOC experiments. T cell development of wt cells proceeded normally in Ptch-deficient thymi that were engrafted into wt mice and populated by lymphoid progenitor cells from the host. Similarly, thymocytes developed normally in FTOCs when Ptch was efficiently deleted in lymphocytes and stromal components of the thymus. Collectively, our experiments indicated that Ptch expression in both cellular compartments of the thymus is dispensable for T cell development. The data also allowed for the conclusion that Ptch does not impact on intrathymic ETPs. Although the Ptch deletion in ETPs was not formally determined in the FTOC model, the fact that the recombination efficiency at the Ptch locus was 100% in whole thymi (Fig. 5A, day 12) supports the assumption that Ptch was deleted in ETPs as well. Being a cell population with an enormous proliferative capacity (25), any defects in the differentiation or proliferation of ETPs would have led to a block in T cell development and, thus, reduced numbers of DN and DP thymocytes in our FTOC experiments, which was not the case. This suggested that the developmental block between the DN1 and DN2 stage described in tPtch^−/− mice (6) must result from a defect in the thymic-homing progenitors from which ETPs originate.

Our data showed that ETPs are dramatically missing in tPtch^−/− thymi. The loss of ETPs was not a consequence of tamoxifen application, because ETP numbers in tamoxifen-treated Ptch^flox/flox control mice were the same as in vehicle-treated Ptch^flox/flox ERT2^+/− mice (see controls in Fig. 6B). Because ubiquitous deletion of Ptch results in a severe developmental defect of the lymphoid lineage accompanied by the loss of CLPs (6) from which ETPs mainly develop (26), the missing ETPs in tPtch^−/− thymi reflect a general loss of thymic-homing progenitors rather than a homing defect of these cells into the thymus. Therefore, the function of Ptch seems to differ from the one of its immediate signaling partner Smo, for which a cell-autonomous function in the expansion and differentiation of ETPs has been shown (5). It may also differ from the function of other components involved in canonical Hh signaling. Examples are Hh, which was reported to activate or inhibit intrathymic T cell proliferation and differentiation in a dose-dependent manner (2, 3), and Gli2 and Gli3, which were shown to influence, for example, TCR repertoire selection (8, 9).

We previously hypothesized that Ptch expression in the stromal BM cell compartment plays a prominent role in the correct diversification of the lymphoid lineage (6). This was based on the lack of any T cell phenotype in Rag2^−/−cγc^−/− mice reconstituted with tPtch^−/− BM, which carry a hematopoietic cell-intrinsic Ptch deletion (6). This hypothesis is now strengthened by the fact that T cell development is normal in T cell-specific Ptch knockout mice, Ptch-deficient FTOCs, and in thymi with Ptch-deficient stroma, whereas it is compromised in Ptch knockout mice reconstituted with wt BM (T cell-extrinsic Ptch deletion).

Recently, the immediate signaling partner of Ptch, Smo, was shown to be dispensable for adult HSC differentiation and specification of the T cell lineage (27). The discrepancy between the role of Smo and its immediate signaling partner Ptch in T cell development can be explained by the fact that Ptch also acts on molecular processes without the involvement of the Smo/Gli axis (28–30). For example, it was demonstrated that Ptch functions as a dependence receptor, which induces apoptosis in the absence of Hh, whereas its proapoptotic activity is blocked the presence of Hh (29). In addition, it was demonstrated that Ptch deletion leads to elevated levels of circulating thymic stromal lymphopoietin, a cytokine important in the regulation of T cell numbers (7). Therefore, we propose that the specific effects of Ptch on T cell development may involve the above-mentioned, or thus far unknown, molecular processes necessary to ensure the cell-extrinsic, prethymic homeostasis required for proper T cell specification of HSCs. The identification of these Ptch-dependent factors will be of great importance for a better understanding of the role that Ptch plays in T cell development. This knowledge will also be important for current considerations of Hh inhibitors upstream of Ptch that may impact on Ptch-mediated prethymic homeostasis necessary for proper T cell development, in cancer treatment in the clinics.

We thank Stefan Wolf and Susan Peter for excellent animal care.

The authors have no financial conflicts of interest.