CD226 is known to be expressed on many types of peripheral lymphoid cells and involved in T cell differentiation, activation, and cytotoxicity. In this study, we report that CD226 is also expressed on mouse thymocytes at varying developmental stages, and its expression is associated with resistance of thymocytes to apoptosis. The levels of CD226 expression appeared to be closely coupled with thymocyte development, in that it was preferentially expressed on CD4+CD8− and CD4−CD8+ thymocytes at all stages during mouse development, and was markedly increased on the cells in neonatal mice. Of the CD4+CD8+ population, CD226 was predominantly expressed by the cells also positive for CD69, suggesting that CD226 expression may be induced in thymocyte-positive selection. Inhibition of CD226 by short hairpin RNA in a fetal thymus organ culture model led to reduced thymus cellularity, which was associated with enhanced apoptotic cell death. In contrast, CD226-transgenic mice displayed enlarged thymus lobes resulting from increased thymus cellularity. CD226 on thymocytes seemed to play a role in regulating the expression of survivin, as inhibition of CD226 down-regulated survivin, but overexpression of CD226 rescued thymocytes from apoptosis through up-regulation of survivin. In addition, overexpression of CD226 reduced sensitivity of EL-4 thymoma cells to apoptosis by up-regulating the expression of survivin. Taken together, these results indicate that CD226 is an antiapoptotic molecule and may play an important role in murine thymocyte development.
The CD226, also named TliSA1, PTA1, or DNAM-1, is a member of the Ig superfamily containing two Ig V-like domains in its extracellular region and three tyrosine residues in its cytoplasmic portion (1, 2). In humans, CD226 is constitutively expressed in hematopoietic cells such as T cells, NK cells, NKT cells, a subset of B cells, monocytes/macrophages, and dendritic cells and is involved in a variety of immunological functions, including T cell differentiation and cytotoxicity, NK cell cytotoxicity, NKT cell apoptosis, monocyte extravasation, and dendritic cell maturation (1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12).
In the murine system, engagement of CD226 with its ligands (mCD112 and Tage4) is believed to mediate a costimulatory signal in Ag-specific T cells (13). It has been reported that Th1 cells differentiation are involved in increased levels of CD226 expression, whereas the levels of CD226 were down-regulated upon Th2 cell polarization in BALB/c mice. In addition, treatment with a specific mAb against CD226(10E5) was shown to inhibit T cell proliferation in spleens and lymph nodes in SJL/J mice (14).
Thymus is an organ for T lymphocyte development from early thymocyte progenitors to functional competent T cells ready for emigration to the periphery. To date, there has been no report on the expression status and the potential functional significance of CD226 in thymocytes. In view of this, we have examined the expression levels of CD226 on thymocytes during mouse development. We show in this report that CD226 is preferentially expressed on CD4+CD8− and CD4−CD8+ thymocytes at all stages during mouse development and is markedly increased in the cells in neonatal mice. Of the CD4+CD8+ population, CD226 is predominantly expressed by the cells also positive for CD69. Moreover, we demonstrate that CD226 is an antiapoptotic molecule that may play an important role in regulating thymus cellularity during murine development.
Materials and Methods
Cell culture and Ab
The EL-4 thymoma cell line and the HEK293 T cell line were maintained in DMEM (Invitrogen) supplemented with 2 mM l-glutamine, 50 μg/ml penicillin, and 50 μg/ml streptomycin and 10% heat-inactivated bovine serum (Invitrogen). All cultures were grown at 37°C in a humidified 5% CO2 air atmosphere.
A polyclonal Ab to mCD226 was generated as previously described (15).
Transgenic (TG) mice (tet-o-mCD226) were generated using conventional techniques as previously described (16, 17). Murine CD226 cDNA was cloned into pBI-5 vector, which contains the tetracycline response element generating tet-o-mCD226. TG founders were identified by PCR and Southern blot analysis of tail DNA and TG lines were established by backcrossing to C57BL/6 mice. The other TG mice (EμSR-tTA) which contain tTA were provided by D. W. Felsher (Division of Hematology-Oncology, Department of Medicine, University of California, San Francisco, CA) (18).
C57BL/6 mice were purchased from Charles River Laboratories and all animals were treated according to the guidelines of the National Institutes of Health for the care and use of laboratory animals (National Institutes of Health Publication 80-23).
Construction of lentivirus vector and virus transfection
Putative short interfering RNA (siRNA)4 sequences were designed against mCD226 using web software provided by Dharmacon RNA Technologies. Sequences were determined to be specific for the mCD226 gene by BLAST searches of GenBank database. Oligonucleotide sequences were designed for the sense and antisense sequences of the siRNA target sites of interest separated by a hairpin loop sequence. The oligonucleotide pairs were then annealed and cloned into the pLKO.1 vector at AgeI and EcoRI site (19). The oligonucleotide pair used to create the construct is: sense, 5′-CCGGGCACTGTGTGAAGAGACATTGCTCGAGCAATGTCTCTTCACACAGTGCTTTTTG-3′; antisense, 5′-AATTCAAAAAGCACTGTGTGAAGAGACATTGCTCGAGCAATGTCTCTTCACACAGTGC-3′. Lentiviral short hairpin (shRNA) vector and each packaging vector (Δ8.9, VSV-G) were cotransfected into HEK293 T cells. Supernatants containing transduction units were collected 48 h after transfection and filtered through a 0.45-μm filter (20).
The EL-4 cells and fetal thymus organs were cultured in medium with lentiviral vectors at a multiplicity of infection (MOI) of 30 or 50 at 37°C.
Fetal thymus organ culture (FTOC)
Thymic lobes dissected from 15-day fetal mice (embryonic day 15 (E15)) were placed on the surface of membrane disc filters (13 mm, 0.45-μm pore size; Whatman), floating on the surface of medium in 10 × 35-mm petri dishes. Culture medium consisted of IMDM (Invitrogen) supplemented with 10% FBS (Invitrogen), streptomycin (100 mg/ml), penicillin (250 mg/ml), nonessential amino acids (0.1 mM; Invitrogen), and lenti-mCD226 or mock at MOI of 50 as indicated. Cultures were grown in a humidified incubator in 5% CO2 at 37°C. Cells were harvested after 8 days of culture and thymus lobes were washed with medium and gently teased apart with the tips of sterile needles to release cells.
Flow cytometric analysis (FACS)
Single-cell suspensions were prepared from thymus and were stained using predetermined optimal concentrations of different fluorochrome-labeled Abs for 30 min at 4°C. Cell phenotypes were determined using the following Abs: anti-mouse CD4-PE (L3T4), anti-mouse CD8α-allophycocyanin (Ly-2), anti-mouse CD69-PerCP-Cy5.5, and anti-mouse CD62L-allophycocyanin-Cy7. All Abs were purchased from eBioscience.
To examine the expression of mCD226, cells were stained with rabbit anti-mCD226 polyclonal Ab, followed by FITC-conjugated goat anti-rabbit Ab. Cells were then stained with anti-mouse CD4-PE (L3T4), anti-mouse CD8α-allophycocyanin (Ly-2), anti-mouse CD69-PerCP-Cy5.5, and anti-mouse CD62L-allophycocyanin-Cy7. After washing, cells were analyzed with a FACSAria flow cytometer (BD Biosciences), and data were analyzed with the WinMDI 2.9 software (J Trotter, The Scripps Research Institute, La Jolla, CA).
Assessment of apoptotic thymocytes
Apoptotic cells were measured by staining with propidium iodide and annexin V with an Apoptosis Detection Kit (BD Biosciences) according to the manufacturer’s instruction.
The cells were lysed in lysis buffer (0.4% Nonidet P-40, 150 mM NaCl, 50 mM Tris-HCl, 1 mM sodium vanadate, 0.1 mM PMSF, and 2 μg/ml leupeptin and apoprotein, pH 8.0). Lysates were centrifuged at 10,000 rpm for 5 min at 4°C. Protein concentrations were measured using a Bio-Rad protein assay. Protein (30 μg) was electrophoresed onto 15% SDS-PAGE and transferred onto nitrocellulose membranes. Membranes were incubated with anti-caspase 3 (Stressgen), anti-mouse survivin (eBioscience), and anti-β-actin (Sigma-Aldrich). Analysis was conducted using ECL detection (Amersham Biosciences).
Transient cDNA transfection and small interfering RNA
EL-4 thymoma cells were transfected with pcDNA3/CD226 and/or the siRNA construct using Amaxa nucleofection technology. Briefly, cells were resuspended in solution following the Amaxa guidelines for cell transfection and transfected with pcDNA3/CD226 and/or 50–100 nmol/L siRNA. The siRNA sequences were designed against survivin using web software provided by Dharmacon RNA Technologies. The sequences contained: 1) sense, CGAUAGAGGAGCAUAGAAATT and antisense, UUUCUAUGCUCCUCUAUCGGG; 2) sense, CCGUCAGUGAAUUCUUGAATT and antisense, UUCAAGAAUUCACUGACGGTT; and 3) sense, AGGGCCUCCUAGCAGGAUCUUATT and antisense, UAAGAUCCUGCUAGGAGGCCCUTT. The nontargeting siRNA control was also obtained from Dharmacon. Expression of CD226 and efficiency of siRNA were measured by flow cytometry and Western blot analysis 60 h after transfection.
CD226 is differentially expressed on thymocytes of different subsets in adult mice
We studied whether murine thymocytes from adult C57BL/6 mice (6–8 wk) express CD226 on their surface with a polyclonal Ab against mouse CD226 by flow cytometry. Abs against mouse CD4 and CD8 were also used to identify CD4+ and CD8+ thymocytes, respectively. Fig. 1,A shows that a small proportion (∼12%) of adult murine thymocytes were positive for CD226. Analysis of distributions of CD226+ cells among different subpopulations of thymocytes showed that ∼60% of CD4+CD8− and 75% of CD4−CD8+ thymocytes expressed CD226 on their surface, whereas only ∼5% of CD4−CD8− and CD4+CD8+ thymocytes had CD226 expression, respectively (Figs. 1 A). These results indicated that CD226 may be preferentially expressed by positively selected thymocytes.
To study whether CD226 plays a role in positive selection of thymocytes, we tested the expression levels of CD226 on CD4+CD8+CD69−, CD4+CD8+CD69+, single-positive (SP) CD69+. and SP CD69−CD62L+ thymocytes, as CD69+ thymocytes are believed to be undergoing positive selection. The results showed that CD226 was mainly expressed on CD69+ double-positive (DP) thymocytes (∼30%), whereas it was hardly detected on CD69− DP thymocytes. There was no difference in the levels of CD226 expression between the CD4+CD8−CD69+ and CD4+CD8−CD69−CD62L+ populations (54.35 and 51.78%, respectively). CD226 was also detected in both the CD4−CD8+CD69+ (85.60%) and CD4−CD8+CD69−CD62L+ (70.22%) populations. However, as can be seen, ∼30% of the latter remained negative for CD226. Collectively, these results suggest that CD226 could be up-regulated during positive selection of DP thymocytes (Fig. 1 B).
CD226 expression on positively selected thymocytes is increased in neonatal mice
To further study the potential relationship between CD226 expression and T cell development, we examined the levels of CD226 expression on different subsets of thymocytes from C57BL/6 mice at early development stages and compared the results with those from adult mice. As shown in Fig. 1,A, CD226 was preferentially expressed on CD4+CD8− and CD4−CD8+ thymocytes at all stages during mouse development, even though there were wide variations in the levels of expression among different stages. At embryonic day 18 (E18), only 15% of CD4+CD8− thymocytes expressed CD226. However, the proportion of CD4+CD8− thymocytes expressing CD226 increased markedly after birth with 43.25% of the cells being positive for CD226 at age day 1 and ∼50.67% at age day 6. By the second week after birth (day 14), ∼58.40% of CD4+CD8− thymocytes expressed CD226, similar to the levels seen in adult mice (Fig. 1,A). Similarly, the percentage of CD4−CD8+ thymocytes expressing CD226 also increased substantially after birth, albeit with different kinetics. The proportions of CD4−CD8+ thymocytes positive for CD226 remained comparable between E18 and day 1 neonates. However, the levels increased markedly at day 6 after birth. By week 2, 62.53% of the CD4−CD8+ thymocytes expressed CD226, similar to the level in adult mice (Fig. 1 A). These results suggested that CD226 expression is closely coupled with thymocyte development.
siRNA knockdown of CD226 induces apoptosis of thymocytes of cultured fetal thymus
We studied whether expression of CD226 confers a survival advantage of thymocytes by knockdown of CD226 with shRNA. To this end, we first examined the efficiency of lentiviral transduction of shRNA for CD226 in the mouse thymoma cell line EL-4, which was derived from a 9,10-dimethyl-1,2-benzanthracine-induced thymoma in a C57BL/6 N mouse. Fig. 2 A showed that the CD226 shRNA markedly reduced the level of the CD226 transcript in EL-4 cells. Similarly, the proportion of EL-4 cells positive for CD226 was also decreased by the shRNA. Although 27% of EL-4 cells with mock transduction expressed CD226, only 7% of the cells transduced with the CD226 shRNA had CD226 expression on their surface.
We next studied the effect of inhibition of CD226 on thymocyte development in a FTOC model, which is a commonly used system for examining developmental biology of thymocytes. Fetal thymus from E15 was cultured in the medium containing lentiviral particles carrying the CD226 shRNA or the control shRNA. As shown in Fig. 2,B, the addition of CD226 shRNA resulted in substantially smaller thymi in comparison to those cultured in the presence of the control shRNA. Quantitation of thymocyte numbers showed that they were reduced by 42% in total and by 15, 45, 71, and 29% in CD4−CD8−, CD4+CD8+, CD4+CD8−, and CD4−CD8+ cells, respectively, in thymi cultured with the CD226 shRNA in comparison to those cultured with the control shRNA (Fig. 2,C). Analysis of CD226 expression revealed that the CD226 shRNA reduced the level of CD226 expression on CD4−CD8−, CD4+CD8+, CD4+CD8−, and CD4−CD8+ thymocytes by ∼49, 68, 46, and 66%, respectively (Fig. 2 D). The CD226 expression level by percentage positive was compared with that by mean fluorescence intensity (MFI), and it has been found that the similar results were gotten by these two ways (supplemental Fig. 2S5).
To study whether reduction in cell numbers in thymus cultured with CD226 shRNA was due to increased induction of apoptosis, we measured externalization of phosphatidylserine with FITC-conjugated annexin V. As shown in Fig. 2,E, shRNA inhibition of CD226 resulted in significant increases in the proportion of thymocytes undergoing apoptosis. This was confirmed by increased activation of caspase 3 as shown by Western blot analysis (Fig. 2,F). Enhancement of apoptotic cell death by inhibition of CD226 was more pronounced in CD4+CD8+ and, in particular, CD4+CD8− thymocyte populations, with the percentage of apoptotic cells being increased by ∼50 and 300%, respectively (Fig. 3 G). Taken together, these results suggested that CD226 may play a role in regulating thymocyte development by antagonizing apoptotic signaling.
Overexpression of CD226 enhances thymopoiesis by inhibition of apoptosis of thymocytes
To further confirm the role of CD226 in murine thymocyte development, we created CD226 TG mice by crossing individuals of the tet-o-mCD226 line with EμSR-tTA line. As shown in Fig. 3, A and B, CD226 expression was elevated on the cell surface of thymocytes from TG mice (6 wk), with approximate increases of 2.46, 2.44, 1.41, and 1.45 times being observed on CD4−CD8−, CD4+CD8+, CD4+CD8−, and CD4−CD8+ cells, respectively. Of significance, adult TG mice (6 wk) exhibited enlarged thymi in comparison to those from littermate control mice (Fig. 3,C). This appeared to be due to increased thymus cellularity as shown in Fig. 3,D. Quantitation of thymocytes showed that CD4−CD8−, CD4+CD8+, CD4+CD8−, and CD4−CD8+ cells were increased by 1.89, 2.14, 2.12, and 1.57-folds, respectively, with the total number of thymocytes being increased by ∼2.06 times (Fig. 3 D).
We tested whether increased cellularity of thymi from TG mice (6 wk) is due to inhibition of apoptosis of thymocytes. Fig. 3, E and F, show that the levels of apoptotic cell death, as measured by externalization of phosphatidylserine and activation of caspase 3 in thymi from CD226 TG mice, were significantly reduced in comparison to that from littermate control mice. Collectively, these results confirmed the role of CD226 in antagonizing apoptotic cell death during murine thymocyte development.
CD226 expression is associated with elevated expression of survivin in murine thymocytes
To examine the mechanism(s) by which CD226 protects murine thymocytes from apoptosis, we examined the expression levels of some important substrates of the apoptosis pathways in thymocytes with or without CD226 being inhibited by shRNA. Western blot showed that the active form of caspase 3 (p20) was significantly increased after knockdown of CD226 in FTOC (Fig. 2,F). Of note, the levels of a member of the inhibitor of apoptotic protein family, survivin, were also markedly reduced in thymocytes with CD226 knocked down by shRNA (Fig. 4 A).
To determine whether survivin is involved in CD226-mediated protection of thymocytes against apoptosis, we generated CD226 TG mice. As shown in Fig. 4,B, thymocytes isolated from CD226 TG mice displayed significantly increased survivin expression. Consistently, the active form of caspase 3 was markedly reduced (Fig. 3 F). Taken together, these results suggest that regulation of survivin by CD226 is responsible, at least in part, for its role in antagonizing against apoptosis in murine thymocytes.
To confirm the role of survivin in the CD226-mediated antiapoptotic effect, we transfected pcDNA3/CD226 and/or survivin-specific siRNA into EL-4 thymoma cells (Fig. 5, A and B). As shown in Fig. 5,C, inhibition of survivin by siRNA enhanced apoptosis of EL-4 cells. In contrast, overexpression of CD226 inhibited apoptosis of EL-4 through up-regulating the expression of survivin (Fig. 5, D and E).
Only a small proportion of thymocytes survive T cell selection in the thymus and leave the thymus migrating to peripheral lymphoid tissues as mature T cells (21). In the murine thymus, it has been shown that most thymocytes that fail selection undergo apoptosis. A number of proapoptotic and antiapoptotic proteins have been suggested to be involved in this process (22, 23, 24).
CD226 is a molecule of the Ig superfamily that was initially cloned in humans in 1996 (1). mCD226 was subsequently cloned in 1999 (Zhang, X., D. Li, W. Ouyang, W. Jia, L. Chen, X. Xie, S. Ning, Y. Zhang, and B. Jin, GenBank accession no.AF416980, http://www.ncbi.nlm.nih.gov/nuccore/22671456; GenBank accession no. AF421198, http://www.ncbi.nlm.nih.gov/nuccore/22671459; GenBank accession no. AF421199, http://www. ncbi.nlm.nih.gov/nuccore/22671462; and GenBank accession no. AF421200 http://www.ncbi.nlm.nih.gov/nuccore/22671465). It has been shown that mCD226 interacted with its ligands mCD112 and Tage4 (PVR homologs) and mediated a costimulatory signal in Ag-specific T cells (13). Furthermore, mCD226 has been demonstrated to be predominantly expressed on T cells obtained from lymph nodes and spleens from BALB/c mice and specifically expressed on the surface of differentiated Th1 cells. Anti-mCD226 treatment delayed the onset and reduced the severity of a Th1-mediated autoimmune disease, experimental autoimmune encephalomyelitis (14). Because the thymus is of central importance in the initial establishment of diverse T cell repertoire, it is of importance to determine the CD226 expression status on thymocytes and to ascertain its biological significance during the development of the thymus.
The initial finding in this study that CD226 was mainly expressed on the CD4 SP or CD8 SP subpopulations in thymi of adult C57BL/6 mice promoted us to examine whether CD226 is involved in positive selection of thymocytes. We compared the expression levels of CD226 on CD4+CD8+CD69−, CD4+CD8+CD69+, SPCD69+, and SPCD69−CD62L+ thymocytes, as CD69+ thymocytes are thought to be undergoing positive selection. The results showed that mCD226 was predominantly expressed on CD69+DP thymocytes. There was no notable difference in the levels of CD226 expression between CD4+CD8−CD69+ and CD4+CD8−CD69−CD62L+ populations. In addition, CD226 was also detected in both CD4−CD8+CD69+ and CD4−CD8+CD69−CD62L+ thymocytes at comparable levels. Nevertheless, a subset of CD4−CD8+CD69−CD62L+ appeared not to express CD226. These data suggest that CD226 expression could be induced in DP thymocytes during positive selection, and they are partly in accordance with the data that CD226 is mainly expressed at very high percentage among positively selected SP thymocytes. Because CD8 SP thymocytes are composed of both immature and mature cells, the expression level of CD226 on CD4−CD8+TCRhigh and CD4−CD8+TCRlow thymocytes was compared. It was found that CD226 was mainly expressed on CD4−CD8+TCRhigh thymocytes, whereas there was little expression of CD226 on CD4−CD8+TCRlow cells, and this may be the reason that CD226 expression is unimodal on CD4 SP thymocytes but biphasic on CD8 SP thymocytes (supplemental Fig. 3S). Interestingly, CD226 expression on CD4 SP thymocytes was found to be dramatically increased at day 1 after birth, whereas its expression on CD8 SP thymocytes was increased gradually during the development. It is conceivable that the expression of CD226 on CD4 SP was constitutive, while the expression of CD226 on CD8 SP could be regulated during development. However, the mechanism(s) involved remains to be elucidated.
The role of CD226 was further confirmed in studies with the shRNA technique in a murine FTOC model. The results showed that inhibition of CD226 by shRNA resulted in smaller thymus lobes, which was accounted for by reduced thymus cellularity. In particular, the numbers of DP and CD4 SP thymocytes appeared to be decreased more markedly. Consistently, CD226 TG mice displayed enlarged thymus lobes due to increased cellularity. These results suggested that CD226 may play a role in regulating the number of thymocytes. We have found in a separate study that in comparison to the cell number in the FTOC human Ig group, the cell number in the CD226/Ig group decreased significantly (data not shown).
It has been reported that treatment with a CD226 Ab could inhibit proliferation of T cells in mouse spleens and lymph nodes (14). Consistently, we found that knockdown of CD226 by shRNA inhibited EL-4 cell proliferation due to a cell cycle arrest at the G2-M phase (data not shown). Nevertheless, knockdown of CD226 enhanced apoptotic cell death, whereas increased CD226 expression resulted in reduced apoptosis, suggesting that the apoptosis machinery also plays a part in regulation of the number of thymocytes. In support of this, CD226 is known to play an antiapoptotic role during NKT cell development (9). Our results suggest that the effect of CD226 on apoptosis of thymocytes is not Ag specific, in that the levels of apoptosis were comparable in thymocytes with similar levels of CD226, even when they were cultured in medium deprived of serum, or in medium containing dexamethasone (supplementary Fig. 1S).
The antiapoptotic effect of mCD226 appeared to be related to survivin. This was first shown by an association between the levels of CD226 and the levels of survivin expression. Survivin, or TIAP, is a member of the inhibitor of apoptotic protein family that is a major class of modulators of apoptosis (25, 26, 27). In humans, CD226 deficiency has been reported to cause increased sensitivity of NKT cells to apoptosis through down-regulation of survivin (9). In this study, knockdown of CD226 in FTOC decreased the expression of survivin and enhanced apoptosis of thymocytes. In contrast, overexpression of CD226 rendered thymocytes more resistant to apoptosis through up-regulation of survivin in TG mice. The role of survivin in CD226-mediated protection of thymocytes was further confirmed in EL-4 thymoma cells overexpressing CD226. Inhibition of survivin in EL-4 cells markedly attenuated the antiapoptotic effect of CD226. In contrast, overexpression of CD226 rescued EL-4 cells from apoptosis via up-regulation of survivin. Collectively, these results indicate that CD226 antagonizes apoptotic signaling in murine thymocytes by, at least in part, up-regulation of survivin.
Beside binding its ligands (CD112 and Tage4), CD226 has been shown to associate with LFA-1 (28). These binding partners of CD226 are commonly expressed in thymocytes. Upon engagement, CD226 is recruited into lipid rafts in the cell membrane where most signaling machinery is usually located (29, 30). Further studies are needed to clarify whether regulation of survivin expression by CD226 is mediated by these signaling pathways.
We thank Dr. Xu Dong Zhang for critical reading of this manuscript.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by the National Natural Science Foundation of China (Grants 30030130, 30570268, and 30672370).
Abbreviations used in this paper: siRNA, short interfering RNA; shRNA, short hairpin RNA; DP, double positive; DN, double negative; FTOC, fetal thymus organ culture; MFI, mean fluorescence intensity; MOI, multiplicity of infection; SP, single positive; E15, embryonic day 15; TG, transgenic; E18, embryonic day 18.
The online version of this article contains supplemental material.