1-Phosphate-Lyase-Deficient Mice Development in Sphingosine

Discontinued Postnatal Thymocyte Development in Sphingosine 1-Phosphate-Lyase-Deficient Mice

This information is current as Claudia Weber, Andreas Krueger, Anika Münk, Constantin of October 1, 2021. Bode, Paul P. Van Veldhoven and Markus H. Gräler J Immunol 2009; 183:4292-4301; Prepublished online 11 September 2009; doi: 10.4049/jimmunol.0901724 http://www.jimmunol.org/content/183/7/4292 Downloaded from

Supplementary http://www.jimmunol.org/content/suppl/2009/09/10/jimmunol.090172 Material 4.DC1 http://www.jimmunol.org/ References This article cites 48 articles, 20 of which you can access for free at: http://www.jimmunol.org/content/183/7/4292.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists by guest on October 1, 2021

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2009 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Discontinued Postnatal Thymocyte Development in Sphingosine 1-Phosphate-Lyase-Deﬁcient Mice1,2

Claudia Weber,3* Andreas Krueger,3* Anika Mu¨nk,* Constantin Bode,* Paul P. Van Veldhoven,† and Markus H. Gra¨ler4*

Circulation of lymphocytes through peripheral lymphoid tissues as well as progenitor entry into the thymus and its output of mature T cells are critical for normal immune function. Egress of lymphocytes from both peripheral lymphoid organs and thymus is dependent on sphingosine 1-phosphate (S1P) gradients. S1P-lyase 1 (SGPL1) deficiency leads to accumulation of S1P in lymphoid tissues, which blocks lymphocyte egress and induces thymus atrophy. In this study, we investigated thymocyte development in SGPL1-deficient mice (SGPL1؊/؊), which exhibited postnatal discontinuation of early thymocytopoiesis starting at 2 wk after birth. SGPL؊/؊ thymi showed a loss of developing thymocytes in the thymic cortex between 2 and 4 wk of age, whereas mature thymocytes accumulated in the medulla. Detailed analysis demonstrated a deficit in thymic early T cell progenitors (ETP) as the Downloaded from principal reason for discontinued thymocyte development. This developmental block was accompanied by accumulation of ceramides, resulting in enhanced apoptosis of developing T cells. Lack of immigration or settlement of ETP completely halted -thymocyte development. We conclude that increased ceramide levels in the thymus of SGPL1؊/؊ mice abrogate thymic develop ment postnatally by enhanced thymocyte apoptosis and depletion of thymic ETP. Our findings indicate that potentially therapeutic immunosuppression by SGPL1 inhibition should benefit from monitoring ceramides to prevent their increase to apoptosis- in- ducing levels. The Journal of Immunology, 2009, 183: 4292–4301. http://www.jimmunol.org/

elf-renewing hematopoietic stem cells are the source of well as CD127ϩCD117low common lymphoid progenitors (4) all blood cell lineages including T cells. Intrathymic T and CD117lowCD90high circulating T cell progenitors (5). cell development depends on continuous recruitment of The most potent T cell precursors inside the thymus are found S 5 ϩ Ϫ bone marrow (BM) -derived progenitors via the bloodstream. in the heterogeneous CD44 , CD25 , and CD4 and CD8 dou- The nature of thymus-seeding precursors remains yet elusive, ble-negative (DN) 1 thymocyte population and were originally but several candidate populations have been proposed such as characterized as CD4lowCD117ϩ (6, 7). These cells are able to Fms-like tyrosine kinase receptor 3 (Flt3)-positive multipotent differentiate in T, B, NK, and lymphoid dendritic cells. Further ϩ ϩ Ϫ/low progenitors, L-selectin progenitors (1), RAG-1 early lym- investigations identified LSK and IL-7R␣ early T cell pro- by guest on October 1, 2021 phoid progenitors (2), all of which share a lineage-negative genitors (ETP) in the DN1 subset with high T and limited B as Ϫ ϩ high (Lin ), Sca-1 , and CD117 (c-kit) (LSK) phenotype (3), as well as myeloid potential (8). The molecular requirements for homing of extrathymic progenitors to the thymus are mostly unknown. Although a competitive *Institute for Immunology, Hannover Medical School, Hanover, Germany; and †Lab- oratory for Lipid Biochemistry and Protein Interactions, Department of Molecular disadvantage of ETP in seeding the thymus was reported for Ϫ Ϫ Cell Biology, Katholieke Universiteit Leuven, Leuven, Belgium CCR9 / mice as well as for mice deficient in PSGL-1, the ligand Received for publication May 29, 2009. Accepted for publication August 5, 2009. for P-selectin (9, 10), complete loss of ETP in a noncompetitive The costs of publication of this article were defrayed in part by the payment of page situation has not been observed so far, indicating that compensa- charges. This article must therefore be hereby marked advertisement in accordance tory mechanisms exist that allow thymus seeding both in the ab- with 18 U.S.C. Section 1734 solely to indicate this fact. sence of CCR9 or PSGL-1. 1 This work was supported by the Emmy Noether Program of the German Research Foundation (Deutsche Forschungsgemeinschaft), Grants GR 1943/1-4 (to M.H.G.) and Inside the thymus thymocytes pass a series of defined develop- KR 2320/2-1 (to A.K.); the Deutsche Forschungsgemeinschaft Priority Program 1267 mental stages and ultimately differentiate into mature CD4 or CD8 “Sphingolipids-Signals and Disease,” Grant GR 1943/2-1 (to M.H.G.); and the Fonds single-positive (SP) cells, which exit the thymus via a sphingosine Wetenschappelijk Onderzoek-Vlaanderen (Grants G.0405.02 and G.0581.09; to P.P.V.V., ϩ/Ϫ covering the fee for the generation of the chimeric SGPL1 mice by Lexicon Genetics). 1-phosphate (S1P) and S1P receptor type 1 (S1P1)-dependent 2 The funders had no role in study design, data collection and analysis, decision to mechanism (11, 12). S1P in blood stimulates the S1P1 receptor on publish, or preparation of this manuscript. mature thymocytes and induces their egress from thymus, where 3 C.W. and A.K. contributed equally to this work. S1P levels are typically low (13–15). Depletion of S1P from blood 4 Address correspondence and reprint requests to Dr. Markus H. Gra¨ler, Institute results in accumulation of mature thymocytes in thymus because for Immunology, Hannover Medical School, OE 9422, Building K11, Carl-Neu- of the missing egress signal (15). S1P belongs to the sphingolipid berg-Strasse 1, 30625 Hannover, Germany. E-mail address: graeler.markus@ mh-hannover.de family and is produced by N-deacylation of ceramide (Cer) and 5 Abbreviations used in this paper: BM, bone marrow; Cer, ceramide; DOP, 4-deoxy- subsequent phosphorylation of sphingosine (Sph) (16). The ratio of pyridoxine; DN, double negative; DP, double positive; ETP, early T cell progenitor; S1P, S1P and Cer was reported to be a determinant of cell fate, with S1P sphingosine 1-phosphate; SP, single positive; Sph, sphingosine; SGPL1, S1P-lyase 1; promoting survival and Cer inducing apoptosis (17). Both caspase- Flt3, Fms-like tyrosine kinase receptor 3; ETP, early T cell progenitor; S1P1, SP1 receptor type 1; SA, streptavidin; LC/MS/MS, liquid chromatography/mass spectrometry/mass dependent and -independent signaling cascades were reported for spectrometry; ESI, electrospray ionization; 2-MEE, 2-(2-methoxyethoxy)ethanol; wt, Cer-mediated apoptosis (18, 19). wild type; DL4, Delta-like 4; AxV, annexin V; LSK, LinϪSca-1ϩc-kithigh. Inhibition of the S1P-degrading enzyme SGPL1 by the vitamin

Copyright © 2009 by The American Association of Immunologists, Inc. 0022-1767/09/$2.00 B6 antagonist 4-deoxypyridoxine (DOP) causes increased S1P www.jimmunol.org/cgi/doi/10.4049/jimmunol.0901724 The Journal of Immunology 4293 levels in lymphoid organs which block lymphocyte egress and in- overnight. Debris was removed by centrifugation, DNA was precipitated duce lymphopenia (14). Accumulated S1P in thymus and lymph with isopropanol, washed once with 75% ethanol, and resuspended in wa- nodes annihilates a proposed gradient between lymphoid tissues on ter after drying. The genotype was determined by a duplex PCR performed with reagents from Fermentas in the MasterCycler Epgradient (Eppendorf) one side and blood and lymph on the other side (13–15). It induces using the following primers (Operon) and optimized conditions (P. P. down-regulation of S1P1 receptor surface expression on lympho- Van Veldhoven, unpublished data): Gal-2s: 5Ј-CGAATACCTGTTCCGT cytes, which, in turn, renders them unresponsive to S1P (13, 14), CATAGC, Gal-2r: 5Ј-ACCACTACCATCATCAATCCGGTAG, MmSPL- and it entails the closure of suggested portals in the sinus-lining Trap-s: 5Ј-TGATAGGG CTGAAAACCACTG, and MmSPL-Trap-r: 5Ј- TCAGAAGCAAAACTGCCTTG; conditions were: 94°C for 2 min for 1 endothelium of lymph nodes (20–22). As a consequence, SP thy- cycle, 94°C for 30 s, 60°C for 30 s, 72°C for 1 min for 35 cycles, and 72°C mocytes and lymph node lymphocytes do not exit into blood and for 7 min for 1 cycle. PCR with these primers yielded a single 667-bp band lymph, but remain in lymph nodes and accumulate in the thymus for SGPL1Ϫ/Ϫ (derived from the ␤-geo sequence in the gene trap), a single ϩ ϩ (11, 12, 14). SGPL1 inhibitors were therefore proposed as poten- 499-bp band for SGPL1 / (amplicon covering the trap insertion site), and ϩ/Ϫ tial immunosuppressive drugs (14) and the SGPL1 inhibitor both bands for SGPL1 mice, when PCR products were separated on an ethidium bromide-agarose gel. LX2931 showed positive results in phase 1 clinical trials for treatment of the autoimmune disorder rheumatoid arthritis (23). Thus, Thymic RNA isolation, cDNA synthesis, and real-time PCR the function of SGPL1 and the consequences of increased S1P Thymi were carefully minced through a nylon mesh until thymocytes were levels in lymphoid organs warrant extensive investigation. removed and only the capsule and stroma cells were left over. Capsules Ϫ/Ϫ To this end, we analyzed SGPL1 mice (24). In addition to were vortexed twice with fresh PBS to get rid of the residual immune cells, severe T cell lymphopenia in blood similar to mice treated with the were disintegrated into small pieces by continuous pipetting, resuspended Ϫ/Ϫ in 250 ␮l of PBS, and mixed thoroughly with 750 ␮l of TRIzol LS reagent vitamin B6 antagonist and SGPL1 inhibitor DOP (14), SGPL1 mice also manifest a 10-fold decrease in thymus size and weight, (Invitrogen). RNA was extracted and cDNA was synthesized with a Downloaded from RevertAid First Strand cDNA Synthesis Kit (Fermentas) after a DNase I caused particularly by severe reduction of double-positive (DP) digest (Invitrogen) according to the manufacturer’s protocols. Real-time thymocytes. Investigation of thymocyte development in these mice PCR was performed with the TaqMan Universal PCR Master Mix (no revealed a progressing phenotype with increased apoptosis and AmpErase UNG) and MM 00444619_M1 (Delta-like 4 (DL4)) (20 x) Taq- ϫ loss of thymic ETP at 4 wk of age, probably due to a defect in Man Gene Expression Assays and Mouse GAPDH ( 20) (Probe-Dyes FAM-MGB) primer from Applied Biosystems using the Applied Biosys- progenitor entry. Several lines of evidence point to increased Cer tems Prism 7700 Sequence Detector and the following conditions: 94°C for http://www.jimmunol.org/ levels as a major inducer of the observed thymus atrophy: First, 2 min for 1 cycle, 94°C for 2 min, and 60°C for 1 min for 40 cycles. Cer levels were increased mainly in thymus. Second, no major Cer Relative expression of DL4 was calculated with the ⌬⌬Ct method. increase was detected in DOP-treated mice, with no developing thymus atrophy. And third, Cer were potent inducers of apoptosis Thymic DNA isolation ex vivo in contrast to Sph and S1P as the other major sphingolipids Genomic DNA was precipitated from the interphase and the organic phase accumulating in SGPL1Ϫ/Ϫ mice. Side effects of potentially ther- of TRIzol-lysed thymocytes (Invitrogen), washed twice with a 0.1 M so- apeutic immunosuppression through SGPL1 inhibition may there- dium citrate solution containing 10% ethanol, and dissolved in water after drying. The genomic DNA content was determined by PCR using the same fore be prevented by monitoring Cer levels during treatment with primers as for the genotyping and the following conditions: ␤-galactosi- SGPL1 inhibitors to avoid their increase to apoptosis-inducing dase-PCR: 94°C for 4 min for 1 cycle, 94°C for 1 min, 58°C for 1 min, by guest on October 1, 2021 levels. 72°C for 1 min for 35 cycles, and 72°C for 4 min for 1 cycle and S1P- lyase-PCR: 94°C for 4 min for 1 cycle, 94°C for 1 min, 58°C for 1 min, Materials and Methods 72°C for 1 min for 24, 27, and 30 cycles, and 72°C for 4 min for 1 cycle. Reagents Cell preparation and in vitro incubation Corticosterone and 1,3-dihydroxy-2-amino-4-octadecene-1-phosphate (S1P) Mice were sacrificed, thymi were removed, weighted, and photographs were obtained from Sigma-Aldrich. 1,3-Dihydroxy-2-amino-4-heptadecene-1- were taken with a Camedia C-5060 Wide Zoom Digital Camera (Olym- phosphate (a 17-carbon analog of S1P, C17-S1P) and 1,3-dihydroxy-2-amino- pus). Thymi were minced through a nylon mesh to obtain single-cell sus- 4-heptadecene (a 17-carbon analog of Sph, C17-Sph) were purchased from pensions. For T cell lineage depletion, thymocytes were incubated with Otto Nordwald. 1,3-Dihydroxy-2-amino-4-octadecene (Sph) was acquired RL1.72 (anti-CD4) and 31M (anti-CD8) hybridoma supernatant, Low from Tocris. N-octanoyl-D-erythro-sphingosine (C8-Cer), N-hexadecanoyl-D- Tox-M Rabbit Complement (Cedarlane Laboratories), and DNase I (Roche erythro-sphingosine (C16-Cer), and N-octadecanoyl-D-erythro-sphingosine Diagnostics) for 30 min at 37°C, followed by a Lympholyte M (Cedarlane (C18-Cer) were ordered from Matreya. FTY720 was provided by Dr. V. Laboratories) gradient centrifugation for 30 min at room temperature. BM Brinkmann and DOP was from Sigma-Aldrich. Compounds were dissolved in was obtained by flushing out femoral bones with ice-cold PBS. Blood was acidified drinking water supplemented with 8 g/L sugar. Corticosterone, S1P, withdrawn by cardiac puncture, and cells were separated from plasma by and Sph were dissolved in methanol and C8-Cer, C16-Cer, and C18-Cer in 1:1 centrifugation and subsequent washing steps. Erythrocyte lysis was per- (v/v) methanol/chloroform. formed with spleen, blood, and BM for 5 min at room temperature in 10

volumes of 139 mM NH4Cl, 1 mM KHCO3, and 0.1 mM EDTA. Spleno- Mice cytes and thymocytes without complement lysis were incubated with 60 S1P-lyase-deficient mice (SGPL1Ϫ/Ϫ) were generated starting from ␮M C8-Cer, 10 ␮M Sph, or carrier alone in RPMI 1640 (PAA Laborato- Ͼ OST58278 (25) gene-trapped embryonic stem cells by Lexicon Genetics ries) supplemented with 10% FCS for 0–4 h at 5% CO2 and 95% hu- ϩ/Ϫ on a fee basis and SGPL1 mice were further bred to a Swiss Webster midity in a CO2 incubator MCO-20AIC (Sanyo). (SW) background in the animal facility of the Katholieke Universiteit Leu- ven as previously described (24). Heterozygous mice were crossed to ob- Flow cytometry tain Ϫ/Ϫ pups. IL-7R ␣-deficient (IL-7RϪ/Ϫ; The Jackson Laboratory), Thymocytes were stained before or after T cell lineage depletion by com- nu/nu NMRI, and SW mice were bred and maintained under specific patho- plement lysis with fluorescently labeled Abs for1hat4°C. The following gen-free conditions at the animal facility of the Hannover Medical School. Abs and proteins were used: anti-CD3␧-FITC, anti-CD4-FITC, anti-CD4- All experimental procedures were approved by ethical committees of the PE, and annexin V-FITC (Immunotools), anti-CD8-Cy5, anti-CD11b-FITC, universities and of the state Lower Saxony and performed in accordance anti-CD117-allophycocyanin, anti-CD44-Pacific Blue, anti-Sca-1-PE-Cy7, with the guidelines of the Hannover Medical School animal facility. and anti-Flt3-PE (eBioscience), anti-CD25-PerCP-Cy5.5, anti-TCR␤- Genotyping of mice PE, streptavidin (SA)-PE-Cy7, SA-PerCP (BD Biosciences), anti-F4/80- PE, and SA-Cy5 (Caltag Laboratories). To exclude residual linage-positive Tips of mouse tails were taken at the age of 3 wk or postmortem and lysed thymocytes evading complement lysis, thymocytes, BM, and blood cells in 100 ␮g/ml proteinase K (Roth) containing lysis buffer (100 mM Tris- were incubated with biotinylated Abs against CD4, CD8␣, GR1, CD11c, HCl (pH 8.5), 5 mM EDTA, 200 mM NaCl, and 0.2% SDS) at 55°C DX5, and NK1.1 (BD Biosciences), CD11b (Caltag Laboratories), TCR␣␤ 4294 THYMOCYTE DEVELOPMENT IN S1P-LYASE-DEFICIENT MICE

(Caltag Laboratories), TCR␥␦, and CD19 (Southern Biotechnology Asso- Lipid extraction ciates) after a 20-min blocking step (PBS, 3% FCS, 10% rat serum) for 1 h ␮ ␮ on ice. Biotinylated primary Abs were detected with fluorescently labeled Biological samples (50 l of erythrocyte cell pellet, 100 l of plasma, in SA. Subsequently, after staining with anti-CD25, anti-CD44, anti-c-Kit vitro-incubated cell pellet) were adjusted to a 1-ml sample volume with (CD117), anti-Sca-1, and anti-Flt3 (CD135) followed by intracellular stain- 1 M NaCl and transferred into a glass centrifuge tube. Organs were ho- ing of TCR␤ according to the manufacturer’s protocol (BD Biosciences), mogenized after weight determination in a Stomacher 80 homogenizer expression of these markers and CFSE labeling of adoptively transferred (Seward) in 5 ml of PBS for 2 min and in case of lymph nodes in 1 ml of ϫ BM cells was detected in lineage-negative gated fractions. Annexin V PBS for 2 2 min and further processed as described previously (26). staining was performed after the other staining and lysis procedures ac- Sample preparation for liquid chromatography/mass cording to the manufacturer’s protocol (apoptosis detection kit; BD Bio- sciences). Thymic macrophages were identified by staining with anti-F4/ spectrometry/mass spectrometry (LC/MS/MS) analysis 80-PE and anti-CD11b-FITC Abs. Flow cytometry measurements were The extracted lipids were dissolved in a methanol/chloroform mixture acquired with the LSR II and FACSCalibur flow cytometers (BD Bio- (4:1, v/v) and stored at Ϫ20°C. C8-Cer (400 pmol) and C17-Sph (300 sciences) and analyzed with FlowJo (Tree Star) and Winlist software (Ver- pmol), used as the internal standards, were added at the beginning of ity Software House). lipid extraction. Immunohistochemistry Electrospray ionization (ESI) LC/MS/MS analysis Thymi were embedded in Tissue-Tek, frozen on dry ice, and stored at A QTrap triple-quadruple mass spectrometer (Applied Biosystems) inter- Ϫ80°C. Cryosections (8 ␮m) were prepared, air dried, and fixed in Ϫ20°C faced with the GPM-2 gradient pump (Dionex), the autosampler AS3500 acetone. Sections were blocked with 10% BSA and 1/100 anti-Fc␥II/III (Thermo Fisher Scientific), the degasser Uniflow DU 8001 (Sanwa (BD Biosciences) Ab-containing PBS for1hinahumidified chamber, Tsusho), and the Gynkotek High Precision Pump M300 (Dionex) for de- stained with anti-CD4-Cy5 and anti-CD8-Cy3 for 1 h, and counterstained livery of the postcolumn modifier 2-(2-methoxyethoxy)ethanol (2-MEE;

for nuclei with 1/2000 4Ј,6-diamidino-2-phenylindole (Roche Diagnostics) Sigma-Aldrich) was used for lipid detection. Positive ion ESI LC/MS/MS Downloaded from for 5 min. Images were acquired following sealing of sections with Mowiol analysis was used for detection of all analytes. The ion source conditions using an Olympus X81 microscope with an F-View II camera and CellP and gas settings for positive ESI LC/MS/MS analysis were as follows: ion software (Olympus). spray voltage ϭ 5500, ion source heater temperature ϭ 450°C, collision gas setting ϭ medium, ion source gases 1 and 2 settings ϭ 30 and 40 psi, Caspase activity assay and curtain gas setting ϭ 40 psi. The multiple reaction monitoring transi- tions for the detection were as follows: S1P m/z 380/264, DH-S1P m/z Thymocytes (5 ϫ 105/96-well in a volume of 50 ␮l of PBS) incubated with 382/284, DH-Sph m/z 302/284, Sph m/z 300/282, C17-Sph m/z 286/268, sphingolipids in vitro for 0–4 h were mixed with 50 ␮l of caspase 3/7, http://www.jimmunol.org/ C17-S1P m/z 366/250, C8-Cer m/z 426/264, C16-Cer m/z 538/264, and caspase 8, and caspase 9 substrates in the corresponding reaction buffer of C18-Cer m/z 566/264. Liquid chromatographic resolution of all analytes the LuminescentGlo Caspase assays (Promega). Plates were incubated at was achieved using a MultoHigh-C column (2 ϫ 60 mm, 3 ␮m; CS- room temperature for 2 min while shaking at 600 rpm, followed by a 1-h 18 Chromatographie Service). The elution protocol was composed of a 10-min incubation at ambient temperature. In the case of caspase 8 and caspase 9, column equilibration with 10% solvent A (methanol), 40% solvent B the proteasome inhibitor MG-132 was added to the reaction. Caspase ac- (H O), and 50% solvent C (100 mM HCl) followed by sample injection tivities were measured for 3 min in the luminescence modus of the No- 2 and an 18-min period with 100% solvent A. Samples were infused into the vostar Microplate Reader (BMG Labtech). electrospray ion source (ESI) through an electrode tube at a rate of 300 ␮ ␮ Generation of BM chimeras l/min along with 2-MEE (9–14 min) at a rate of 100 l/min. Standard curves were generated by adding increasing concentrations of the analytes

BM was removed by flushing out femoral bones from donor mice with 5 ml to 300 pmol of C17-Sph and 400 pmol C8-Cer (internal standards). Cor- by guest on October 1, 2021 of ice-cold PBS under sterile conditions. Cells were washed once with ticosterone was quantified accordingly without 2-MEE infusion using H2O PBS, collected by centrifugation, resuspended in 1 ml of freezing medium as solvent C and atmospheric pressure chemical ionization with the fol- (90% FCS and 10% DMSO), and stored in liquid nitrogen until use. Cells lowing settings: nebulizer current ϭ 2 ␮A, ion source heater temperature ϭ were slowly thawed, washed once with 3% FCS in PBS, carefully trans- 450°C, collision gas setting ϭ medium, ion source gases 1 and 2 settings ϭ ferred on top of 10 ml of Lympholyte M solution, and centrifuged at 60 and 40 psi, and curtain gas setting ϭ 40 psi. The corticosterone-specific 850 ϫ g for 20 min. Purified cells were removed, washed once with 3% multiple reaction monitoring transition m/z 347/329 was monitored. Lin- FCS in PBS, and kept on ice in 3% FCS in PBS. Before injection into the earity of the standard curves and correlation coefficients were obtained by tail vein of SGPL1ϩ/ϩ recipient mice, these cells were centrifuged, resus- linear regression analyses. All mass spectrometry analyses were performed pended in PBS without FCS, and counted. Before injection, recipient mice using Analyst 1.4 (Applied Biosystems). were irradiated twice with 4.5 Gy in an interval of 5 h. Eight weeks later, chimeric mice were killed and organs and blood were removed for further Graphics and statistical analysis analysis. Graphs were either generated by GraphPad Prism 4.0 or Microsoft Excel 2002. Statistical analysis was performed with Microsoft Excel 2002 and Generation of neonatal BM chimeras results were presented as means Ϯ SEM. To compare groups for significant BM was removed by flushing out femoral bones from donor mice with 5 differences ( p values), two-tailed unpaired Student’s t test was applied. ml of ice-cold PBS under sterile conditions. Erythrocytes were lysed, and cells were washed once with PBS supplemented with 3% FCS, collected by Results centrifugation, resuspended in 2 ml of blocking buffer for 20 min (PBS, 3% Depletion of cortical DP thymocytes and accumulation of FCS, and 10% rat serum), incubated with biotinylated Abs against GR1, medullary SP T cells B220, CD11c, DX5, CD11b, TCR␣␤, and Ter119 (BD Biosciences), and depleted by MACS for positive cells to enrich for reconstitution relevant Lymphopenia induced by DOP has been linked to increased S1P cells. Ab-incubated cells were washed once with MACS buffer, incubated levels in lymphoid organs through SGPL1 inhibition (14). To as- ␮ ␮ with 200 l of magnetic SA beads (Miltenyi Biotec) in 1800 l of MACS sess whether or not SGPL1 inhibition is sufficient for the DOP- buffer on ice for 20 min, and separated with an AutoMACS (Miltenyi Biotec) after two washing steps with MACS buffer. The negative cell frac- induced phenotype, peripheral blood T cells were analyzed in un- Ϫ/Ϫ tion was collected by centrifugation, washed twice with sterile PBS, and treated and DOP-treated SGPL1 and wild-type (wt) control resuspended in 250 ␮l of sterile PBS. Twenty-five microliters of this cell mice. Whereas DOP treatment induced a severe T cell lymphope- suspension was injected into the liver of each 12-h-old mouse, which were nia in 4-wk-old wt control mice, age-matched SGPLϪ/Ϫ mice were irradiated twice with 2.3 Gy in an interval of 5 h. Chimeric mice were already lymphopenic in their naive state (Fig. 1A). Treatment with killed 4 wk later and thymi were removed for further analysis. DOP did not significantly increase their inherent lymphopenia, in- Thymus transplantation dicating that SGPL1 inhibition is indeed sufficient for peripheral T cell lymphopenia induced by DOP. However, when comparing Thymi of 4-wk-old IL-7R and S1P-lyase-deficient mice were removed, the Ϫ/Ϫ two lobes carefully separated, and one lobe per animal was transplanted thymi of 4-wk-old SGPL mice with those of wt and heterozy- under the kidney capsule of 10-wk-old anesthetized nu/nu recipient mice. gous littermates, we observed a 10-fold decrease in size and The Journal of Immunology 4295

first analyzed the composition of immature DN1–4 thymocyte subsets, which can be characterized by differential expression of the surface markers CD44, CD25, and CD117. ETP are CD44ϩCD25Ϫ, express high levels of CD117, and constitute a subset of the DN1 population (8, 27). Flow cytometric analysis revealed the absence of thymocytes in DN2 (CD44ϩCD25ϩ) and DN3 (CD44ϪCD25ϩ) stages and a reduction in DN1 and DN4 cells (Fig. 2B). Analysis of intracellular (ic) TCR␤ expression in DN4 cells showed that the majority of DN4 cells lacked icTCR␤ expression, indicating that these cells do not represent conven- tional DN4 thymocytes (Fig. 2C). Analysis of the DN1 subset for expression of CD117 revealed a virtually complete lack of ETP in SGPL1Ϫ/Ϫ thymi at 4 wk of age (Fig. 2D). Thus, these data indicate that the reduction in DP thymocyte numbers may be a consequence of the absence of DN thymocytes. To assess whether the lack of ETP was due to a defect in BM-derived and circulating progenitors, we analyzed BM and blood for the presence of LinϪ, Sca-1ϩ, and c-kitϩ LSK cells, which comprise CD135ϩ multipotent progenitors as well as CD135Ϫ hematopoietic stem cells. LSK cells were slightly re- Downloaded from duced in BM of SGPL1Ϫ/Ϫ mice (supplemental Fig. S1, A and B).6 But obvious differences regarding their occurrence in blood were not detected (supplemental Fig. S1, A and B). The lack of ETP is therefore not derived from missing progenitor cells in blood, but likely due to a thymic defect, probably a defect in progenitor entry (3, 28). However, it cannot be ultimately ex- http://www.jimmunol.org/ cluded that, despite being abundant, the LSK cells analyzed FIGURE 1. T cell lymphopenia and abnormal weight and cellularity of display a developmental defect, or that other candidate T cell thymi in SGPLϪ/Ϫ mice. A, Blood T cell concentration of 4-wk-old wt precursors with a different phenotype, such as circulating T cell control (ϩ/ϩ) and SGPLϪ/Ϫ (Ϫ/Ϫ) mice untreated (Cont) and treated with progenitors, might be absent from circulation (5). Increased 100 mg/l DOP in the drinking water (DOP). B, Weight of thymi from plasma corticosterone levels that have been proposed to cause a Ϫ/Ϫ 4-wk-old wt (ϩ/ϩ), heterozygous (ϩ/Ϫ), and SGPL (Ϫ/Ϫ) mice. C, similar phenotype in latent TGF␤-binding protein 3-deficient ϩ ϩ ϩ Ϫ Tissue sections of thymi from 4-wk-old wt ( / ), heterozygous ( / ), mice (29) were not detected in SGPL1Ϫ/Ϫ mice (supplemental and SGPLϪ/Ϫ (Ϫ/Ϫ) mice were stained with Abs against CD4 (red fluo- Fig. S1C), excluding a dominant role of this stress hormone for rescence) and CD8 (green fluorescence) and counterstained with 4Ј,6-dia- by guest on October 1, 2021 midino-2-phenylindole (blue). Cortical regions (blue boxes) and medullary the observed thymus atrophy. regions (red boxes) are enlarged (bar, 20 ␮m; enlarged 5 ␮m). DP thymo- Progressive age-dependent loss of developing thymocytes cytes are mostly missing in the cortical region of SGPLϪ/Ϫ thymi and SP CD4 and CD8 thymocytes accumulate in their medullary region. Considering the fact that no ETP were found in the thymus of 4-wk-old SGPL1Ϫ/Ϫ mice, it was surprising to see the high amount of fully developed SP CD4 and CD8 mature thymocytes weight, which has not been observed in DOP-treated mice (Fig. (Fig. 2A). These SP thymocytes must have developed at an 1B). Fluorescence microscopy of tissue sections stained with Abs earlier point in time and did not leave the thymus due to the against CD4 and CD8 revealed virtually empty cortical regions described block of thymocyte egress (14). To test this hypoth- harboring only very few DP thymocytes, whereas the amount of esis, thymocyte development of 1- to 4-wk-old SGPL1Ϫ/Ϫ mice CD4 and CD8 SP thymocytes in the medullary regions was in- was analyzed. Similar to thymi of control mice, SGPL1Ϫ/Ϫ creased (Fig. 1C). This phenotype indicated that thymocyte devel- thymi grew in size between weeks 1 and 2 (Fig. 3A). But opment was dramatically impaired and prompted us to analyze T whereas control thymi continued to grow, SGPL1Ϫ/Ϫ thymi cell development in SGPLϪ/Ϫ mice in greater detail. started to shrink in weeks 3 and 4 (Fig. 3A). Contemporane- ously DP thymocyte numbers were decreasing, and cell num- Halted thymocytopoiesis by loss of thymic ETPs bers of ETP along with thymocytes of DN2, DN3, and DN4 Flow cytometric analysis of thymocytes from 4-wk-old SGPLϪ/Ϫ stages were reduced (Fig. 3B and supplemental Fig. S2, A–D). mice showed a significant increase in the frequency of SP CD4 and The loss of ETP and subsequent developing T cells in the thy- Ϫ Ϫ CD8 T cells, which is in line with the accumulation of SP CD4 and mus of SGPL1 / mice therefore started postnatally after 2 wk CD8 thymocytes due to elevated S1P levels and consequent li- of age. On the contrary, mature SP thymocytes accumulated gand-induced down-regulation of the egress-mediating G protein- over time due to the reported block of egress (Fig. 3C and A coupled receptor S1P1 in DOP-treated mice (11, 12, 14). However, supplemental Fig. S2 ) (14). Taken together the data shown in in contrast to DOP-treated mice, the DP thymocyte population was Fig. 2 suggest that the observed decline in DP thymocyte num- virtually missing in SGPLϪ/Ϫ mice both in terms of frequency and bers is largely due to a defect in early thymopoiesis. absolute cell numbers (Fig. 2A). Egress of premature T cells from T cell-independent contributions of the thymic microenvironment the thymus did not occur since no DP thymocytes were found in Ϫ/Ϫ the periphery. Other possible reasons for the observed deficiency Discontinuation of thymocyte development in SGPL1 mice in DP thymocytes include a developmental block at earlier devel- could either be an inherent problem of T cell progenitors or could opmental stages, increased susceptibility to apoptosis, or defects in positive or negative selection, or a combination thereof. Thus, we 6 The online version of this article contains supplemental material. 4296 THYMOCYTE DEVELOPMENT IN S1P-LYASE-DEFICIENT MICE Downloaded from

FIGURE 2. Disrupted T cell development in 4-wk-old SGPL1Ϫ/Ϫ mice. A, FACS analysis of thymocytes from 4-wk-old wt (ϩ/ϩ), heterozygous (ϩ/Ϫ), and SGPL1Ϫ/Ϫ (Ϫ/Ϫ) mice stained with Abs against http://www.jimmunol.org/ CD4 and CD8. B, Lineage-negative thymocytes were stained with Abs against CD44 and CD25 and costained with ic TCR␤ (C)or CD117 (D). by guest on October 1, 2021

be due to changes in the thymic microenvironment preventing thy- blood cells, or thymus were not altered in mice transplanted with mopoiesis or even the immigration of thymus-seeding progenitors. BM-derived cells from SGPL1Ϫ/Ϫ or control mice (supplemental To test these possibilities, BM-derived precursors from SGPL1Ϫ/Ϫ Fig. S3A). Thus, the observed block in thymopoiesis is likely due and wt donor mice were adoptively transferred into lethally irra- to changes in the thymic microenvironment. diated wt recipient mice. Analysis of thymi from recipients 8 wk Thymic T cell differentiation is dependent on Notch 1 receptor- after the adoptive transfer showed normal thymocyte development mediated signaling, and expression of its ligand DL4 by thymic with no signiﬁcant differences between T cell populations of epithelial cells is crucial for normal T cell development (30). Ex- SGPL1Ϫ/Ϫ and control mice, indicating that hematopoietic pro- pression of DL4 mRNA in capsules of thymocyte-depleted genitors from SGPL1Ϫ/Ϫ mice are largely normal (Fig. 4A). These SGPL1Ϫ/Ϫ thymi was 12 times lower than in wt controls, which results are in line with our observation of near-normal progenitor provides a potential explanation for loss of early thymocytes (sup- frequencies in both BM and circulation (supplemental Fig. S1, A plemental Fig. S3B). However, no accumulation of immature B and B). In addition, sphingolipid levels, including S1P, in plasma, cells was observed in SGPL1Ϫ/Ϫ thymi, as would be predicted The Journal of Immunology 4297 Downloaded from http://www.jimmunol.org/ by guest on October 1, 2021

FIGURE 4. Organ-dependent block in thymocyte development with altered levels of sphingolipids, increased DP thymocyte apoptosis, and macrophage counts in thymi of SGPL1Ϫ/Ϫ mice. A, Total number of DN, DP, and SP thymocytes 8 wk after adoptive transfer of BM-derived precursors Ϫ Ϫ FIGURE 3. Progressive loss of thymic T cell development in 1- to from SGPL1 / and wt donor mice in lethally irradiated wt littermates. 4-wk-old SGPL1Ϫ/Ϫ mice. A, Size (bar, 1 cm) and total number of thy- The presence of genomic DNA for ␤-galactosidase (Gal), which is specific Ϫ Ϫ mocytes from 1- to 4-wk-old wt (ϩ/ϩ), heterozygous (ϩ/Ϫ), and for SGPL1 / cells, and for SGPL1 was analyzed by PCR in thymi from SGPL1Ϫ/Ϫ (Ϫ/Ϫ) mice. B, Total number of thymocyte subsets. Shown are mice reconstituted with BM-derived precursors from wt donors (lanes 1 Ϫ Ϫ ETP (CD44ϩCD25ϪCD117ϩ), DN2 (CD44ϩCD25Ϫ), DN3 (CD44Ϫ and 2) and SGPL1 / donors (lanes 3 and 4). PCR products for SGPL1 are CD25ϩ), and DN4 (CD44ϪCD25ϪicTCR␤ϩ) lineage-negative thymocytes shown after 24 (left), 27 (middle), and 30 cycles (right). GAPDH is shown and DP (CD4ϩCD8ϩ) total thymocytes. C, Relative amount of SP and DP as control. The relative amount of DNA amplified after 27 PCR cycles thymocytes. (SGPL1) and 35 PCR cycles (␤-galactosidase was quantified relative to GAPDH. B, Levels of S1P, Sph, and C16-Cer in thymi of 1- to 4-wk-old ,ء :wt (ϩ/ϩ), heterozygous (ϩ/Ϫ), and SGPL1 Ϫ/Ϫ (Ϫ/Ϫ) mice (p values Ϫ Ϫ Ϫ Ϫ from other studies (30). Thus, these data suggest that SGPL1 de- SGPL / vs wt and ϩ, SGPL / vs heterozygous). C, Relative amount of ficiency results in loss of ETP as the principal source for devel- AxV-positive apoptotic DP thymocytes. D, Amount of thymic macro- ␾ oping lymphocytes in thymus as a consequence of a defect in re- phages (M )per1goftissue weight and in percentage of total cells. plenishing thymocytopoiesis by extrathymic precursors. Lack of Notch signaling after lineage commitment could result in the loss of later DN populations. However, it cannot be excluded that the Potential role of impaired immigration of extrathymic lack of DL4 expression is not a direct effect of SGPL1 deficiency, precursors but rather due to the absence of cross-talk between DP thymocytes To evaluate the possible impact of deranged seeding of thymus- and thymic epithelial cells (31). seeding precursors on T cell development, thymi from SGPL1Ϫ/Ϫ 4298 THYMOCYTE DEVELOPMENT IN S1P-LYASE-DEFICIENT MICE and IL-7R ␣-deficient mice (IL-7RϪ/Ϫ) were transplanted under the kidney capsule of recipient nude mice. IL-7RϪ/Ϫ thymi were chosen as controls because of their similarly small size compared with SGPL1Ϫ/Ϫ thymi (32). All (six of six) transplanted IL-7RϪ/Ϫ thymi were engrafted and repopulated by T cells. In contrast, only 33% of SGPL1Ϫ/Ϫ thymi were engrafted (4 of 12). Those thymi developed lower percentages of DP thymocytes by trend compared with IL-7RϪ/Ϫ thymi and showed only minor differences in their lipid profile compared with controls (supplemental Fig. S4, A and B). Failed engraftment of SGPL1Ϫ/Ϫ thymi could have been the result of inefficient immigration by extrathymic progenitors, with a crippled thymus transplant left behind. However, we cannot completely exclude defective vascularization of transplants, since SGPL1 deficiency also translates into vascular defects (33). En- grafted SGPL1Ϫ/Ϫ thymi were rescued from severe S1P, Sph, and Cer accumulation, demonstrating that the phenotype was not maintained in the successfully transplanted tissues (supplemental Fig. S4B). Since SGPL1Ϫ/Ϫ mice died between 4 and 5 wk of age (34), they could not be successfully grafted with wt thymus. As an al- ternative approach, BM-derived precursors from wt donor mice Downloaded from were adoptively transferred into the liver of sublethally irradiated newborn wt, SGPL1ϩ/Ϫ, and SGPL1Ϫ/Ϫ recipient mice. All wt (33 of 33) and SGPL1ϩ/Ϫ (33 of 33) recipient mice survived and developed normally without any phenotypical alterations. SGPL1Ϫ/Ϫ recipient mice (10 of 10) died within 1 wk after sublethal irradi- ation and BM cell transfer, precluding analysis of wt hematopoi- http://www.jimmunol.org/ esis in a SGPL1Ϫ/Ϫ setting. Increased thymocyte apoptosis through Cer accumulation Loss of thymocytes occurred virtually simultaneously in all developmental stages (Fig. 3B), indicating that additional mechanisms might contribute to the depletion of DP thymocytes in SGPL1Ϫ/Ϫ, such as increased susceptibility to apoptosis. The analysis of sphingolipids in thymi from 4-wk-old SGPL1Ϫ/Ϫ mice manifested a by guest on October 1, 2021 Ͼ1000-fold increase in S1P, Ͼ100-fold in Sph, and Ͼ10-fold in C16-Cer (Fig. 4B). Cer are known to induce apoptosis (18, 19) and mice deficient for the Cer-producing enzyme acid sphingomyelinase are rescued from radiation-induced apoptosis (35). In fact, when analyzing the frequency of apoptotic DP thymocytes using annexin V (AxV) labeling, we observed increased frequencies of FIGURE 5. Enhanced apoptosis, caspase activity, and C16-Cer levels of AxV- positive cells (Fig. 4C). Similarly, DN and ETP populations thymocytes incubated with C8-Cer. Thymocytes (A) and splenocytes (B) Ϫ/Ϫ from SGPL mice contained higher amounts of AxV-positive from wt mice were incubated without (control (Cont)) and with 10 ␮M Sph cells compared with wt and heterozygous littermates (supplemen- and 60 ␮M C8-Cer. Since C8-Cer shows only moderate cell permeability, tal Fig. S5A). Macrophage numbers were not dramatically altered a higher concentration of C8-Cer than Sph was used. Apoptosis of CD4 and based on thymus weight, but due to thymocyte atrophy the ratio of CD8 DP and SP thymocytes and splenocytes was detected by staining with C8-Cer vs control and ,ء :macrophages:T cells was 21 times increased in SGPL1Ϫ/Ϫ thymi anti-CD4 and anti-CD8 Abs and AxV (p values ϩ (Fig. 4D). In vitro incubation of thymocytes with C8-Cer consid- , C8-Cer vs Sph). C, Comparison of C16-Cer levels in thymus, spleen, ϩ ϩ erably enhanced apoptosis, which underlies the proapoptotic ac- and peripheral lymph nodes (pLN) from 4-wk-old wt ( / ), heterozygous (ϩ/Ϫ), and SGPL1 Ϫ/Ϫ (Ϫ/Ϫ) mice. D, Cellular concentrations of Sph tivity in T cells (Fig. 5A). In this assay, DN populations including (left), S1P (middle and left), and C16-Cer (middle and right) of thymocytes ETP as well as DP thymocytes proved to be very susceptible to and C16-Cer of splenocytes (right) incubated without (control) and with 10 C8-Cer vs control and ϩ, C8-Cer ,ء :Cer-induced apoptosis, whereas CD4 and CD8 SP thymocytes ␮M Sph and 60 ␮M C8-Cer (p values were little or not affected (Fig. 5A and supplemental Fig. S5B). vs Sph). E, Activity of caspases 3, 8, and 9 in thymocytes incubated for Mature CD4 and CD8 T cells isolated from spleen did not show 16 h without (control) and with 10 ␮M Sph and 60 ␮M C8-Cer. increased apoptosis after similar treatment with C8-Cer (Fig. 5B). Furthermore, Cer concentrations in SGPL1Ϫ/Ϫ mice were maxi- mally enhanced in thymus and less pronounced in spleen and pe- and 9 (Fig. 5E). Thus, increased Cer-induced apoptosis was likely ripheral lymph nodes (Fig. 5C). Notably, similar incubations with to substantially contribute to the simultaneous loss of developing Sph did not result in extensive apoptosis and emphasize the spe- thymocytes. cific role of Cer in inducing T cell apoptosis (Fig. 5, A and B). Incubation of thymocytes with C8-Cer led to a significant increase Dispensable role of S1P receptors for thymus atrophy in C16-Cer, which most likely derives from endogenous cellular S1P is the endogenous ligand of five S1P receptors that are po- production in apoptotic cells (Fig. 5D). Again, the increase in C16- tential candidates to mediate the observed loss of DN and DP Cer was lower in similarly treated splenocytes. Cer-induced apo- thymocytes of SGPL1Ϫ/Ϫ mice. Although knockout mice of all ptosis in thymocytes was mediated by activation of caspases 3, 8, single S1P receptors were immunologically characterized, none of The Journal of Immunology 4299

and triggers the closure of postulated endothelial cell portals for exiting lymphocytes (20–22). Both events contribute to the immunosuppressive effect of SGPL1 inhibitors like LX2931 (Lexi- con Pharmaceuticals), which is tested for treatment of the autoimmune disorder rheumatoid arthritis. Although SGPL1Ϫ/Ϫ mice principally mirror the lymphopenic phenotype of DOP treatment, thymus atrophy has been addition- ally observed in these mice (34), which was not induced by DOP treatment (Fig. 6A). DOP is distributed to tissue cells via the bloodstream and needs to be phosphorylated to be an active vita-

min B6 antagonist (41). Sphingolipid measurements of thymi from DOP-treated mice revealed 10 times lower increases in S1P and Sph levels compared with SGPL1Ϫ/Ϫ mice (Fig. 6B). SGPL1 inhibition was therefore more complete in SGPL1Ϫ/Ϫ mice, which most likely also accounts for the additional thymus atrophy in these mice (Fig. 3A). The even higher S1P and Sph levels in SGPL1Ϫ/Ϫ mice expedited the generation of Cer as the closest metabolites of Sph (Fig. 4B). FIGURE 6. Effect of FTY720 and DOP on thymocyte development. A,

The apoptosis-inducing function of Cer is generally accepted, Downloaded from Total number of thymocyte subsets from 4-wk FTY720-treated (FTY) and although its contribution to T cell apoptosis is controversially dis- untreated mice (control (Cont)) and from 5-wk DOP-treated and untreated ϩ Ϫ ϩ cussed. Early reports demonstrating a role for Cer in Fas-induced mice (control). Shown are lineage-negative ETP (CD44 CD25 CD117 ) and T cells in DN4 stage (CD44ϪCD25ϪicTCRϩ and SP CD4 (CD4ϩ apoptosis were questioned by later studies showing the indepen- CD8Ϫ), CD8 (CD4ϪCD8ϩ), and DP (CD4ϩCD8ϩ) total thymocytes. B, dence of Fas-induced T cell apoptosis from Cer generation (42, Concentrations of S1P, Sph, and C16-Cer in thymi from 4-wk FTY720- 43). In fact, T cells seem to have different dispositions for Cer- treated (FTY) and untreated mice (control) and from 5-wk DOP-treated

induced apoptosis with developing DP thymocytes being more http://www.jimmunol.org/ and untreated mice (control). susceptible than mature CD4 and CD8 SP T cells (Fig. 5, A and B). Similar differences in the responsiveness of developing and mature thymocytes were also reported for corticosterone (29) and may them exhibit the described phenotype except for S1P , the deletion 1 result from a general proapoptotic state of developing thymocytes of which is embryonically lethal (36–38). Fetal liver chimeras that before positive selection. Whereas DP thymocytes are prone to die lack expression of S1P in the hematopoietic system however were 1 and require active survival signaling, CD4 and CD8 SP thymo- able to repopulate the thymus, which thereby excludes a major role cytes are thought to be prone to survival and require active signals for S1P1 expression in thymus entry and settling of extrathymic progenitors, as well as intrathymic development (11). To test to be negatively selected, i.e., to be deleted by apoptosis, or they are already in a postselection state (44). Notably, Cer has been by guest on October 1, 2021 whether or not S1P1 expression on nonhematopoietic cells is in- volved in loss of ETP and later thymocyte subsets, wt mice were suggested to play a role in glucocorticoid-mediated cell death of treated for 4 wk with 20 mg/L of the immune modulator FTY720 DP thymocytes (45, 46). In combination with the highest increase Ϫ/Ϫ in drinking water. FTY720 is phosphorylated in vivo to FTY-P, of Cer levels in thymi of SGPL1 mice (Fig. 5C), thymocyte development is severely impaired in SGPL1Ϫ/Ϫ mice due to the which is a potent agonist for S1P1 among others and accumulates in lymphoid tissues including thymus (39, 40). Flow cytometric consequent loss of immature T cell precursors in the thymus by analysis of thymocytes revealed no major differences of develop- Cer-induced apoptosis (Fig. 3). ing thymocyte populations in treated mice except for the accumu- Another possible explanation for the observed thymus atrophy Ϫ/Ϫ lation of SP CD8 mature thymocytes and slightly reduced ETP in SGPL1 mice would be stress-induced increase of cortico- (Fig. 6). In addition, treatment of wt mice with DOP for 5 wk also steroids, which can also lead to thymic involution (29). However, Ϫ/Ϫ did not result in loss of developing thymocytes, but led to an ac- corticosterone levels in plasma of SGPL1 mice were not ele- cumulation of thymic SP T cells. Compared with untreated mice, vated (supplemental Fig. S1C). Furthermore, no analogous invo- Ϫ/Ϫ thymic levels of S1P increased 100-fold, Sph 10-fold, and C16-Cer lution of spleens from SGPL1 mice was observed (data not 2-fold (Fig. 6). FTY-P and S1P levels were therefore high enough shown), as reported for mice with increased corticosterone plasma levels (29). Stress can therefore be excluded as the primary reason to activate and down-regulate S1P1 receptors in SP thymocytes with their consequent accumulation due to the block in egress (14). for thymus atrophy. Ϫ Ϫ In contrast, the concentration of apoptosis-inducing Cer was not Cer generation in SGPL1 / mice was not expected because sufﬁcient to abrogate thymocyte development. S1P and Sph are not part of the de novo synthesis of Cer starting from serine and palmitoyl-CoA, nor do they play a role in the Discussion sphingomyelinase pathway. On the other hand Sph serves as sub- strate for Cer synthases and has been implicated in Cer formation Treatment of mice with the vitamin B6 antagonist DOP is immunosuppressive by inducing lymphopenia in blood (14). DOP in- from the breakdown of complex sphingolipids in lysosomes and hibits SGPL1, leading to substantial increases of S1P in lymphoid late endosomes via the salvage pathway and from recycling of tissues (14). S1P in blood and lymph mediates lymphocyte egress exogenous Cer via ceramidase activity (reviewed in Ref. 47). It is from lymphoid organs by stimulating the S1P1 receptor on the therefore likely that direct formation of Cer from Sph via Cer surface of T and B cells, thus providing the required exit signal synthases accounts for the observed Cer accumulation in (15). Increased S1P levels in lymphoid tissues prematurely down- SGPL1Ϫ/Ϫ mice. This pathway typically does not contribute much regulate S1P1 receptor surface expression on lymphocytes, pre- to the endogenous Cer pool under normal conditions. venting them from recognizing the exit signal (13). Furthermore, it S1P accumulation leads to the block of mature SP thymocyte activates S1P1 on sinus-lining endothelial cells in lymphoid organs egress and their correspondent enrichment in the thymic medulla 4300 THYMOCYTE DEVELOPMENT IN S1P-LYASE-DEFICIENT MICE

(Figs. 1C,2A, and 3B). Whereas this phenotype can be fully ex- 2. Igarashi, H., S. C. Gregory, T. Yokota, N. Sakaguchi, and P. W. Kincade. 2002. plained by its ligand properties for S1P , acting either as a func- Transcription from the RAG1 locus marks the earliest lymphocyte progenitors in 1 bone marrow. Immunity 17: 117–130. tional antagonist that down-regulates S1P1 surface expression on 3. Adolfsson, J., O. J. Borge, D. Bryder, K. Theilgaard-Monch, I. Astrand-Grundstrom, lymphocytes (13–15) or as a classical agonist that stimulates the E. Sitnicka, Y. Sasaki, and S. E. Jacobsen. 2001. Upregulation of Flt3 expression within the bone marrow LinϪSca1ϩc-kitϩ stem cell compartment is accompanied by closure of postulated endothelial cell barriers (20–22), both inhib- loss of self-renewal capacity. Immunity 15: 659–669. iting thymocyte egress, the loss of the ETP population cannot be 4. Karsunky, H., M. A. Inlay, T. Serwold, D. Bhattacharya, and I. L. Weissman. 2008. Flk2ϩ common lymphoid progenitors possess equivalent differentiation explained by modulation of S1P1 function. This was also sup- potential for the B and T lineages. Blood 111: 5562–5570. ported by the lack of thymus atrophy after treatment with the im- 5. Krueger, A., and H. von Boehmer. 2007. Identification of a T lineage-committed munomodulator FTY720 (Fig. 6A). Treatment of wt thymocytes progenitor in adult blood. Immunity 26: 105–116. with C8-Cer mimicked the observed higher apoptotic rate and de- 6. Wu, L., R. Scollay, M. Egerton, M. Pearse, G. J. Spangrude, and K. Shortman. Ϫ/Ϫ 1991. CD4 expressed on earliest T-lineage precursor cells in the adult murine creased cell number of thymocytes in SGPL mice, indicating thymus. Nature 349: 71–74. that Cer-induced apoptosis may also account for the reduction of 7. Moore, T. A., and A. Zlotnik. 1995. T-cell lineage commitment and cytokine thymic ETP (Figs. 4C and 5, A–C). The lack of B cell development responses of thymic progenitors. Blood 86: 1850–1860. Ϫ/Ϫ 8. Allman, D., A. Sambandam, S. Kim, J. P. Miller, A. Pagan, D. Well, A. Meraz, in SGPL thymi despite the reduced amount of DL4 and sub- and A. Bhandoola. 2003. Thymopoiesis independent of common lymphoid pro- sequent impairment of Notch 1 receptor-mediated signaling points genitors. Nat. Immunol. 4: 168–174. to impaired thymus seeding by extrathymic precursors (supple- 9. Rossi, F. M., S. Y. Corbel, J. S. Merzaban, D. A. Carlow, K. Gossens, J. Duenas, L. So, L. Yi, and H. J. Ziltener. 2005. Recruitment of adult thymic progenitors is mental Fig. S3B) (30). The progressive and complete loss of thy- regulated by P-selectin and its ligand PSGL-1. Nat. Immunol. 6: 626–634. mic ETP with near-normal frequencies of T cell progenitors in 10. Uehara, S., A. Grinberg, J. M. Farber, and P. E. Love. 2002. A role for CCR9 in blood and BM in SGPL1Ϫ/Ϫ mice reveals a nonredundant molec- T lymphocyte development and migration. J. Immunol. 168: 2811–2819. 11. Matloubian, M., C. G. Lo, G. Cinamon, M. J. Lesneski, Y. Xu, V. Brinkmann, Downloaded from ular mechanism essential for the earliest step in thymic T cell M. L. Allende, R. L. Proia, and J. G. Cyster. 2004. Lymphocyte egress from development. thymus and peripheral lymphoid organs is dependent on S1P receptor 1. Nature A recent study suggested S1P in plasma as one potential feed- 427: 355–360. 12. Allende, M. L., J. L. Dreier, S. Mandala, and R. L. Proia. 2004. Expression of the back signal to sense lymphopenia and induce thymic receptivity sphingosine 1-phosphate receptor, S1P1, on T-cells controls thymic emigration. for T cell precursors via S1P-controlled expression of thymic P- J. Biol. Chem. 279: 15396–15401. selectin/CCL25 (48). Indeed, disruption of the S1P gradient be- 13. Lo, C. G., Y. Xu, R. L. Proia, and J. G. Cyster. 2005. Cyclical modulation of sphingosine-1-phosphate receptor 1 surface expression during lymphocyte recir- http://www.jimmunol.org/ tween plasma and thymus by FTY720 and DOP treatment resulted culation and relationship to lymphoid organ transit. J. Exp. Med. 201: 291–301. in mildly reduced P-selectin mRNA levels and concomitant re- 14. Schwab, S. R., J. P. Pereira, M. Matloubian, Y. Xu, Y. Huang, and J. G. Cyster. duced thymic receptivity. This is in line with our finding that the 2005. Lymphocyte sequestration through S1P lyase inhibition and disruption of S1P gradients. Science 309: 1735–1739. same treatment leads to a mild reduction of ETP numbers (Fig. 15. Pappu, R., S. R. Schwab, I. Cornelissen, J. P. Pereira, J. B. Regard, Y. Xu, 6A). Notably, thymocyte development in SGPL1Ϫ/Ϫ mice is much E. Camerer, Y. W. Zheng, Y. Huang, J. G. Cyster, and S. R. Coughlin. 2007. Promotion of lymphocyte egress into blood and lymph by distinct sources of more impaired when compared with treatment with these com- sphingosine-1-phosphate. Science 316: 295–298. pounds, which may be due to elevated Cer levels in thymi of these 16. Ha¨nel, P., P. Andreáni, and M. H. Gra¨ler. 2007. Erythrocytes store and release mice (Fig. 4B). Alternatively, progenitor entry could also be ham- sphingosine 1-phosphate in blood. FASEB J. 21: 1202–1209. pered by other mechanisms, such as defective vascularization (33). 17. Cuvillier, O., G. Pirianov, B. Kleuser, P. G. Vanek, O. A. Coso, S. Gutkind, and S. Spiegel. 1996. Suppression of ceramide-mediated programmed cell death by by guest on October 1, 2021 Furthermore, the described S1P-controlled expression of thymic sphingosine-1-phosphate. Nature 381: 800–803. P-selectin/CCL25 is not sufficient to explain the virtually simulta- 18. Neumeyer, J., C. Hallas, O. Merkel, S. Winoto-Morbach, M. Jakob, L. Thon, Ϫ/Ϫ D. Adam, W. Schneider-Brachert, and S. Schutze. 2006. TNF-receptor I defective neous loss of all DN populations in SGPL1 mice (Fig. 3B). in internalization allows for cell death through activation of neutral sphingomy- Earlier studies postulated that Cer and S1P establish a rheostat elinase. Exp. Cell Res. 312: 2142–2153. that ultimately determines the fate of a given cell (17). Accumu- 19. Thon, L., H. Mohlig, S. Mathieu, A. Lange, E. Bulanova, S. Winoto-Morbach, S. Schutze, S. Bulfone-Paus, and D. Adam. 2005. Ceramide mediates caspase- lation of S1P would favor cell survival, whereas an increase of Cer independent programmed cell death. FASEB J. 19: 1945–1956. would induce cell death. Our data suggest that Cer accumulation 20. Sanna, M. G., J. Liao, E. Jo, C. Alfonso, M. Y. Ahn, M. S. Peterson, B. Webb, alone is sufficient to induce apoptosis in developing T cells re- S. Lefebvre, J. Chun, N. Gray, and H. Rosen. 2004. Sphingosine 1-phosphate (S1P) receptor subtypes S1P1 and S1P3, respectively, regulate lymphocyte re- gardless of how much S1P is present (Figs. 3B,4,B and C, and 5, circulation and heart rate. J. Biol. Chem. 279: 13839–13848. A–C). Accumulation of S1P by SGPL1 inhibition can reach a level 21. Wei, S. H., H. Rosen, M. P. Matheu, M. G. Sanna, S. K. Wang, E. Jo, where S1P- and S1P receptor-mediated immunosuppression is ac- C. H. Wong, I. Parker, and M. D. Cahalan. 2005. Sphingosine 1-phosphate type 1 1 receptor agonism inhibits transendothelial migration of medullary T cells to companied with ongoing Cer-induced apoptosis predominantly of, lymphatic sinuses. Nat. Immunol. 6: 1228–1235. but likely not limited to, developing T cells. Pharmaceutical ap- 22. Sanna, M. G., S. K. Wang, P. J. Gonzalez-Cabrera, A. Don, D. Marsolais, plication of SGPL1 inhibitors for treatment of autoimmune disor- M. P. Matheu, S. H. Wei, I. Parker, E. Jo, W. C. Cheng, et al. 2006. Enhancement of capillary leakage and restoration of lymphocyte egress by a chiral S1P1 an- ders like rheumatoid arthritis should therefore be carefully evalu- tagonist in vivo. Nat. Chem. Biol. 2: 434–441. ated, considering thresholds of sphingolipid concentrations for 23. Brown, P., K. Frazier, D. Augeri, D. W. Walke, C. Pappas, B. Brooks, M. Donoviel, and T. Oravecz. 2008. LX2931: a potential small molecule treat- different phenotypic implications including thymocyte abundance, ment for autoimmune disorders. In 2008 Annual Meeting of the American College development, apoptosis, and trapping. of Rheumatology. Arthritis Rheum. 58(9 Suppl.): 351. 24. Van Veldhoven, P. P. 2005. Abstracts from the 46th International Conference on the Bioscience of Lipids, Ajaccio, Corsica, September 20–24, 2005. Chem. Phys. Acknowledgments Lipids 136: 164–165. We thank Michaela Friedrichsen and Stefanie Willenzon for technical help 25. Zambrowicz, B. P., G. A. Friedrich, E. C. Buxton, S. L. Lilleberg, C. Person, and and Ulrich H. von Andrian, Reinhold Fo¨rster, Immo Prinz, and Vijay A. T. Sands. 1998. Disruption and sequence identification of 2,000 genes in Kumar Chennupati for helpful discussions. mouse embryonic stem cells. Nature 392: 608–611. 26. Andreáni, P., and M. H. Gra¨ler. 2006. Comparative quantification of sphingolipids and analogs in biological samples by high-performance liquid chromatogra- Disclosures phy after chloroform extraction. Anal. Biochem. 358: 239–246. The authors have no financial conflict of interest. 27. Benz, C., V. C. Martins, F. Radtke, and C. C. Bleul. 2008. The stream of precursors that colonizes the thymus proceeds selectively through the early T lineage precursor stage of T cell development. J. Exp. Med. 205: 1187–1199. References 28. Schwarz, B. A., and A. Bhandoola. 2004. Circulating hematopoietic progenitors 1. Perry, S. S., H. Wang, L. J. Pierce, A. M. Yang, S. Tsai, and G. J. Spangrude. with T lineage potential. Nat. Immunol. 5: 953–960. 2004. L-selectin defines a bone marrow analog to the thymic early T-lineage 29. Chen, Y., B. Dabovic, C. Colarossi, F. R. Santori, M. Lilic, S. Vukmanovic, and progenitor. Blood 103: 2990–2996. D. B. Rifkin. 2003. Growth retardation as well as spleen and thymus involution The Journal of Immunology 4301

in latent TGF-␤ binding protein (Ltbp)-3 null mice. J. Cell Physiol. 196: (S1P) receptors, S1P(2)/LP(B2)/EDG-5 and S1P(3)/LP(B3)/EDG-3. J. Biol. 319–325. Chem. 277: 25152–25159. 30. Koch, U., E. Fiorini, R. Benedito, V. Besseyrias, K. Schuster-Gossler, M. Pierres, 39. Sensken, S. C., C. Bode, and M. H. Gra¨ler. 2008. Accumulation of FTY720 in N. R. Manley, A. Duarte, H. R. Macdonald, and F. Radtke. 2008. Delta-like 4 is lymphoid tissues contributes to prolonged efficacy. J. Pharmacol. Exp. Ther. 328: the essential, nonredundant ligand for Notch1 during thymic T cell lineage com- 963–969. mitment. J. Exp. Med. 205: 2515–2523. 40. Mandala, S., R. Hajdu, J. Bergstrom, E. Quackenbush, J. Xie, J. Milligan, 31. Fiorini, E., I. Ferrero, E. Merck, S. Favre, M. Pierres, S. A. Luther, and R. Thornton, G. J. Shei, D. Card, C. Keohane, et al. 2002. Alteration of lym- H. R. MacDonald. 2008. Cutting edge: thymic crosstalk regulates delta-like 4 phocyte trafficking by sphingosine-1-phosphate receptor agonists. Science 296: expression on cortical epithelial cells. J. Immunol. 181: 8199–8203. 346–349. 32. Peschon, J. J., P. J. Morrissey, K. H. Grabstein, F. J. Ramsdell, E. Maraskovsky, 41. Coburn, S. P., and J. D. Mahuren. 1976. In vivo metabolism of 4Ј-deoxypyri- B. C. Gliniak, L. S. Park, S. F. Ziegler, D. E. Williams, C. B. Ware, et al. 1994. doxine in rat and man. J. Biol. Chem. 251: 1646–1652. Early lymphocyte expansion is severely impaired in interleukin 7 receptor-defi- 42. Watts, J. D., M. Gu, A. J. Polverino, S. D. Patterson, and R. Aebersold. 1997. cient mice. J. Exp. Med. 180: 1955–1960. Fas-induced apoptosis of T cells occurs independently of ceramide generation. 33. Schmahl, J., C. S. Raymond, and P. Soriano. 2007. PDGF signaling specificity is Proc. Natl. Acad. Sci. USA 94: 7292–7296. mediated through multiple immediate early genes. Nat. Genet. 39: 52–60. 43. Gulbins, E., R. Bissonnette, A. Mahboubi, S. Martin, W. Nishioka, T. Brunner, 34. Vogel, P., M. S. Donoviel, R. Read, G. M. Hansen, J. Hazlewood, S. J. Anderson, G. Baier, G. Baier-Bitterlich, C. Byrd, F. Lang, et al. 1995. FAS-induced apo- W. Sun, J. Swaffield, and T. Oravecz. 2009. Incomplete inhibition of sphingosine ptosis is mediated via a ceramide-initiated RAS signaling pathway. Immunity 2: 1-phosphate lyase modulates immune system function yet prevents early lethality 341–351. and non-lymphoid lesions. PLoS ONE 4: e4112. 44. Sebzda, E., S. Mariathasan, T. Ohteki, R. Jones, M. F. Bachmann, and 35. Santana, P., L. A. Pena, A. Haimovitz-Friedman, S. Martin, D. Green, P. S. Ohashi. 1999. Selection of the T cell repertoire. Annu. Rev. Immunol. 17: M. McLoughlin, C. Cordon-Cardo, E. H. Schuchman, Z. Fuks, and R. Kolesnick. 829–874. 1996. Acid sphingomyelinase-deficient human lymphoblasts and mice are defec- 45. Cifone, M. G., G. Migliorati, R. Parroni, C. Marchetti, D. Millimaggi, A. Santoni, tive in radiation-induced apoptosis. Cell 86: 189–199. and C. Riccardi. 1999. Dexamethasone-induced thymocyte apoptosis: apoptotic 36. Girkontaite, I., V. Sakk, M. Wagner, T. Borggrefe, K. Tedford, J. Chun, and signal involves the sequential activation of phosphoinositide-specific phospho- K. D. Fischer. 2004. The sphingosine-1-phosphate (S1P) lysophospholipid re- lipase C, acidic sphingomyelinase, and caspases. Blood 93: 2282–2296. ceptor S1P3 regulates MAdCAM-1ϩ endothelial cells in splenic marginal sinus 46. Ashwell, J. D., F. W. Lu, and M. S. Vacchio. 2000. Glucocorticoids in T cell organization. J. Exp. Med. 200: 1491–1501. development and function. Annu. Rev. Immunol. 18: 309–345. Downloaded from 37. Liu, Y., R. Wada, T. Yamashita, Y. Mi, C. X. Deng, J. P. Hobson, 47. Kitatani, K., J. Idkowiak-Baldys, and Y. A. Hannun. 2008. The sphingolipid H. M. Rosenfeldt, V. E. Nava, S. S. Chae, M. J. Lee, et al. 2000. Edg-1, the G salvage pathway in ceramide metabolism and signaling. Cell Signal. 20: protein-coupled receptor for sphingosine-1-phosphate, is essential for vascular 1010–1018. maturation. J. Clin. Invest. 106: 951–961. 48. Gossens, K., S. Naus, S. Y. Corbel, S. Lin, F. M. Rossi, J. Kast, and H. J. Ziltener. 38. Ishii, I., X. Ye, B. Friedman, S. Kawamura, J. J. Contos, M. A. Kingsbury, 2009. Thymic progenitor homing and lymphocyte homeostasis are linked via A. H. Yang, G. Zhang, J. H. Brown, and J. Chun. 2002. Marked perinatal lethality S1P-controlled expression of thymic P-selectin/CCL25. J. Exp. Med. 206: and cellular signaling deficits in mice null for the two sphingosine 1-phosphate 761–778. http://www.jimmunol.org/ by guest on October 1, 2021