Ikaros promotes rearrangement of TCR alpha genes in an Ikaros null thymoma cell line

Summary

Ikaros is important in the development and maintenance of the lymphoid system, functioning in part by associating with chromatin-remodeling complexes. We have studied the functions of Ikaros in the transition from pre-T cell to the CD4⁺CD8⁺ thymocyte using an Ikaros null CD4⁻CD8⁻ mouse thymoma cell line (JE131). We demonstrate that this cell line carries a single functional TCR β gene rearrangement and expresses a surface pre-TCR. JE131 cells also carry non-functional rearrangements on both alleles of their TCR α loci. Retroviral re-introduction of Ikaros dramatically increased the rate of transcription in the α locus and TCR Vα/Jα recombination resulting in the appearance of many new αβTCR⁺ cells. The process is RAG dependent, requires SWI/SNF chromatin-remodeling complexes and is coincident with the binding of Ikaros to the TCR α enhancer. Furthermore, knockdown of Mi2/NuRD complexes increased the frequency of TCR α rearrangement. Our data suggest that Ikaros controls Vα/Jα recombination in T cells by controlling access of the transcription and recombination machinery to the TCR α loci. The JE131 cell line should prove to be a very useful tool for studying the molecular details of this and other processes involved in the pre-T cell to αβTCR⁺ CD4⁺CD8⁺ thymocyte transition.

Introduction

Key steps in T cell development are the rearrangements of the T-cell receptor (Tcr) loci through V(D)J recombination (1,2). In the early CD4⁻CD8⁻ double-negative (DN) stages of αβ T-cell development, the β chain locus is rearranged. Cell surface expression of an in-frame β chain together with the pre-Tα chain triggers a burst of proliferation and maturation to the CD4⁺CD8⁺ double-positive (DP) stage. This is followed by cell cycle exit and initiation of rearrangement of the TCR α loci, eventually producing a population of αβTCR⁺ DP thymocytes. The basic mechanisms of TCR gene rearrangement in the α and β loci have been well-studied, although many details remain obscure. The process involves the recruitment of numerous proteins including chromatin remodeling complexes, such as SWI/SNF (Switch/Sucrose Nonfermentable) (3), histone modifiers (4, 5), various transcription factors (6–10), and the enzymes RAG1 and RAG2 (11), which carry out V(D)J recombination. Also preceding the recombination events are sub-nuclear re-localization of the loci to areas of active chromatin (12), germ line transcription (13), and locus contraction, which brings V, D and J elements into close proximity (14, 15).

The Ikaros family of zinc finger proteins plays an integral role in the development of T and B cells (16). The expression of Ikaros itself is highly regulated during development and in mice lacking Ikaros, thymic development is very inefficient. Ikaros has been shown to be involved in regulating the expression of numerous genes such as Cd4 (17), Cd8 (18), TdT (19) and Ptcra (pre-TCRα) (20). While Ikaros was initially referred to as a transcription factor, one of its main functions appears to be recruitment of chromatin remodeling complexes, such as SWI/SNF, to specific loci (21). Ikaros has also been implicated in regulating both B-cell (22) and T-cell receptor gene recombination (10).

Mice lacking Ikaros develop thymomas, one of which, JE131, has a DN phenotype similar to pre-T cells. Re-introduction of Ikaros into JE131 sets in motion its rapid differentiation into a DP-like thymocyte and the appearance of many αβTCR⁺ cells (23). Here we show that the JE131 cell line expresses a surface pre-TCR with a single functional β chain and has many rearranged TCR α loci, nearly all of which encode out-of-frame sequences. Re-introduction of Ikaros results in the rapid increase in transcription from the locus and the appearance of new RAG-dependent in-frame rearrangements. The process requires the SWI/SNF remodeling complex and that is antagonized by the Mi2/NuRD (Nucleosome Remodeling and Deacetylase) complex. Our results suggest that Ikaros functions to open the TCR α loci setting in motion the processes that allow efficient recombination.

Results

Ikaros expression promotes TCR expression on JE131 cells

After confirming the CD4⁻CD8⁻C25⁺CD44⁻ surface phenotype of JE131 cells (23), we transduced the cells with retroviruses expressing Ikaros and GFP as separate proteins. We used expression of GFP to track transduced cells and to reflect approximate levels of Ikaros expression. An ‘empty’ retrovirus expressing GFP alone was used as a negative control. As shown previously (23), a small percentage of JE131 cells were TCR⁺ before transduction (Fig. 1A, top). However, enforced expression of Ikaros induced TCR surface expression robustly, as detected using a Cβ-specific antibody (Fig. 1A, top). The change in TCR expression was directly related to the level of GFP expression, indicating a dose-dependent effect of Ikaros. In cells with the highest dose of Ikaros, up to 44% of the cells became TCR⁺ (Fig. 1A, bottom). To address whether the increase in the frequency of TCR⁺ cells was due to preferential expansion of the pre-existing TCR⁺ cells, we depleted TCR⁺ cells by magnetic-activated cell sorting (MACS) prior to transduction with Ikaros retrovirus. Ikaros induced TCR expression in the TCR⁻ cells (Fig. 1B), suggesting that it did in fact cause de novo expression of TCR.

Figure 1. Ikaros promotes the generation of TCR⁺ JE131 cells.

(A) JE131 cells were transduced with empty MiG vector or MiG-Ikaros retrovirus. 3 days post-transduction, cells were stained for TCR-Cβ and analyzed by flow cytometry. Plots (gated on live cells) show GFP vs. TCR-Cβ expression. Data are representative of four independent experiments with similar results. The percentage of cells that are TCR-Cβ⁺ at increasing levels of GFP (Ikaros) expression is shown as mean ± SEM of n = 4. Gates are indicated in top panel. (■, Vector; Δ, Ikaros) (B) Non-transduced JE131 cells stained with an allophycocyanin-conjugated anti-TCR-Cβ antibody were depleted of TCR⁺ cells using MACS separation, using anti-allophycocyanin beads. Sorted TCR⁻ cells or unsorted cells were immediately transduced with MiG or MiG-Ikaros retrovirus, or MiG-Ikaros retrovirus. 3 days post-transduction, cells were stained for TCR-Cβ and analyzed by flow cytometry. Histograms from analysis of cells on Day 0 (gated on live cells) show TCR-Cβ expression on cells prior to or immediately after MACS sorting. Histograms from analysis of cells on Day 3 (gated on live and GFP^high cells) shows TCR-Cβ expression on unsorted or sorted cells 3 days after transduction with MiG or MiG-Ikaros. (High GFP expression is defined as fluorescence in the 4^th decade). (N.D., not done.) Data are representative of two experiments.

JE131 cells have a DN3 phenotype with a rearranged TCR β locus

The surface phenotype of JE131 cells (Fig. 2A, left) (23) suggested that they are similar to pre-T cells at the DN3 stage. To confirm this, we analyzed the cells by flow cytometry for co-expression of TCR β and pre-Tα chains on the cell surface. Just as DN3 thymocytes do in vivo, most (98%) JE131 cells expressed a pre-TCR, with a direct correlation (diagonal staining) between the levels of the pre-Tα and TCR β chains (Fig. 2A, right). Expression of Ptcra mRNA was confirmed by PCR analysis of cDNA generated from JE131 cells (Fig. 2B). However, as described above, a small percentage of JE131 cells appeared to have a higher level of TCR β expression prior to transduction with Ikaros (Fig. 1A and 2A). The lack of pre-Tα surface expression on these cells suggested that they displayed a mature αβTCR after functional α chains have replaced the pre-Tα chain. The frequency of this population of αβTCR⁺ cells increased dramatically upon re-expression of Ikaros with a corresponding disappearance of pre-TCR⁺ cells (Fig. 2C). These changes mirror events occurring in thymocytes in vivo as they progress from the DN stage to the DP stage, when TCR α gene rearrangement occurs.

Figure 2. JE131 cells have a DN3 pre-T cell phenotype.

(A) Non-transduced JE131 cells or B6 splenocytes were stained for CD44 or CD25 and analyzed by flow cytometry (left). Non-transduced JE131 cells were co-stained for TCR-Cβ and pre-Tα and analyzed by flow cytometry (right, top). Pre-Tα expression level of each population (gated on live and pre-TCR⁺ or αβTCR⁺ cells) is also shown (right, bottom). Data are representative of two independent experiments. (B) PCRs to amplify Ptcra were performed on cDNA prepared from JE131 cells. As positive or negative controls, PCRs were performed on cDNA prepared from TCRα^−/− mouse thymocytes or the plasmacytoma cell line uM2. Data are the results of one experiment. (C) JE131 cells were transduced with empty MiT vector or MiT-Ikaros retrovirus. 3 days post-transduction, cells were co-stained for TCR-Cβ and pre-Tα and analyzed by flow cytometry. Plots (gated on live cells and different levels of Thy1.1 expression) show TCR-Cβ vs. pre-Tα expression. Data are representative of two independent experiments.

To confirm that the low level of co-staining with pre-Tα- and Cβ-specific antibodies indicated the presence of functional TCR β chains in most, if not all, JE131, we transduced the cell line with a retrovirus expressing the fully functional TCR α chain of the DO-11.10 T-cell clone (DOα) (24). Surface TCR expression increased dramatically on all cells that expressed DOα (Fig. 3A). This confirmed that all JE131 cells express TCR β and that they express all components (e.g. the CD3 chains) necessary for surface expression of mature αβTCRs, except for functional TCR α chains.

Figure 3. JE131 cells express a single TCR β chain and are capable of expressing a mature αβTCR.

(A) JE131 cells were transduced with empty MiG vector or MiG-DOα retrovirus, encoding the DO-11.10 TCR α chain. 2 days later, cells were stained for TCR-Cβ and analyzed by flow cytometry. Histograms (gated on live and GFP⁺ cells) show TCR-Cβ expression. Data are representative of two independent experiments. (B) Non-transduced JE131 cells were co-stained for TCR-Cβ and Vβ8 and analyzed by flow cytometry. Plot (gated on live cells) shows TCR-Cβ vs. Vβ8 expression. (C) Non-transduced JE131 cells were cloned by limiting dilution and cultures of several expanded sub-clones were used to prepare cDNA. PCRs were performed on each sample using primers specific for Vβ8 and Cβ. Sequences of PCR products from all clones were identical. A portion of the sequence is shown. IMGT designations for each gene segment are in parentheses. Red lettering indicates nucleotides encoding the CDR3 region.

To determine if JE131 cells possess the same functional β chain rearrangement, we analyzed the cell line for expression of various Vβ’s and Cβ by flow cytometry. There was a direct correlation between the level of Cβ and Vβ8 expression (Fig. 3B). All Cβ⁺ cells were Vβ⁺, suggesting that these cells express the same Vβ8-containing TCR β chain. Confirming this result, the parent cell line and a number of independent JE131 clones all expressed the same mRNA encoding a functional TCR Vβ8 sequence comprised of Vβ8.1, Jβ2.5 and a single junctional CDR3 sequence (Fig. 3C).

JE131 cells have undergone 1° TCR α rearrangements

Given that a small percentage of JE131 cells appear to express mature αβTCRs, we determined whether JE131 cells have undergone rearrangement at the TCR α loci in the absence of Ikaros. To test this possibility in a simple qualitative way, we asked whether the 5’ Jα segments were present in genomic DNA purified from JE131 cells. We performed PCRs on purified genomic DNA using a series of primers that matched various regions of the Jα locus (Fig. 4A, top). Surprisingly, the 5’ Jα segments upstream of Jα16 were missing (Fig. 4B), suggesting that primary VαJα rearrangements in JE131 cells have deleted these upstream Jα segments (Fig. 4A, bottom). Furthermore, since only a small percentage of cells expressed a mature αβTCR, most of the pre-T cell-like (αβTCR⁻) cells likely carry non-productive α chain rearrangements.

Figure 4. Primary rearrangements have deleted the 5’ Jα gene segments.

(A) Schematic representation of the TCR α locus. Horizontal black bars show approximate locations of sequences corresponding to various primer pairs used in PCRs (below). (Eα, TCR α enhancer) (B) Genomic DNA was prepared from non-transduced JE131 cells or, as a positive control, a γδ T cell hybridoma. Purified gDNA was used a template in separate PCR reactions, using primers in or near various Jα segment exons. A PCR for the Cα region was used as a positive control. Data shown are representative of two independent experiments. (C) 14 JE131 sub-clones were stained for TCR-Cβ and analyzed by flow cytometry 18 days after cloning. Histograms show TCR-Cβ expression on four representative clones. (D) Five sub-clones were transduced with a retroviral vector encoding a RSS-GFP-RSS recombination substrate (schematic diagram of the substrate shown at right). The GFP gene is in an inverted orientation and must be inverted by RAG-mediated recombination to be expressed. Two days post-transduction, cells were stained for hCD4 (transduction marker) and analyzed by flow cytometry for GFP expression. The results for one representative clone are shown. In a separate experiment, as a negative control, M12C3 B-cell lymphoma cells were transduced with the same virus and stained and analyzed as before. Histograms (gated on live and hCD4⁺ cells) show GFP expression.

JE131 cells undergo constitutive but inefficient TCR α rearrangement

Given that primary rearrangements have occurred in JE131 cells, we tested whether they are competent to generate spontaneous TCR α rearrangements. We analyzed 10 expanded sub-clones and all had a small percentage of cells that expressed a mature αβTCR (Fig 4C). Given that most of the clones we isolated were likely initially TCR⁻, the presence of TCR⁺ cells in all sub-clones suggests that subsequent TCR α gene rearrangements occurred. To confirm that the cells were competent to rearrange genes, we asked whether they could rearrange an exogenous recombination substrate, the product of which is detectable by flow cytometry. We transduced several sub-clones with a retrovirus encoding an inverted GFP gene flanked by RSSs (Fig 4D, right) that is activated following inversion by RAG1/2 (25). Indeed, transduced cells became GFP⁺ at a high frequency (Fig 4D, left).

Ikaros induces TCR by inducing TCR α rearrangement

Because JE131 cells are not defective for surface expression of exogenous TCR α chains (Fig. 3A), we hypothesized that the Ikaros-mediated induction of TCR is due to de novo TCR α rearrangement. To address this, we compared the TCR α repertoire in mock-transduced cells to that of cells transduced with Ikaros. We first depleted the αβTCR⁺ cells from the culture in order to remove cells that harbored an in-frame TCR α sequence. We then either mock-infected the TCR⁻ cells or infected them with retrovirus encoding Ikaros. After 3 days, cells were harvested and sorted. Three populations were sorted for comparison: mock-infected TCR⁻ cells, Ikaros-infected GFP⁺TCR⁻ cells, and Ikaros-infected GFP⁺TCR⁺ cells. We generated cDNA from each sorted population, amplified TCR α transcripts by PCR with a Cα primer and a set of Vα primers, then cloned and sequenced the resulting PCR products. If Ikaros were inducing new TCR α gene rearrangements, we would expect to detect diverse in-frame TCR α sequences in TCR⁺ cells (and new out-of-frame sequences too), while TCR⁻ cells would contain only out-of-frame sequences. In the Ikaros-transduced TCR⁺ cells, we identified many unique in-frame TCR α sequences (Table 1), as well as several out-of-frame sequences. As expected, we found several out-of-frame sequences in both the mock-transduced TCR⁻ cells (Supporting Information Table 1) and the Ikaros-transduced TCR⁻ cells (Supporting Information Table 2). As predicted by our analysis of the Jα locus (Fig. 4B), all of the TCR α rearrangements included Jα segments from the 3’ end of the Jα cluster. Together, these data suggest that Ikaros promotes secondary TCR α rearrangements in JE131 cells. We also confirmed that Ikaros did not induce rearrangement by inducing quiescence in JE131 cells. Cells transduced with retroviruses encoding p27 or Ikaros exhibited slowed growth (Supporting Information Fig. 1, bottom); however, only the expression of Ikaros induced TCR (Supporting Information Fig. 1, top).

Table 1.

TCR α chains expressed in Ikaros-transduced TCR⁺ cells

Valpha	CDR3	Jalpha	Times Observed
In-frame
TRAV-6	CVLGMLSGSFNKLTF	TRAJ-4	1
TRAV-7	CAASDLSGGNYKPTF	TRAJ-6	1
TRAV-10	CAANPGGNYKPTF	TRAJ-6	1
TRAV-6	CALGRTSGGNYKPTF	TRAJ-6	1
TRAV-10	CAATSGGNYKPTF	TRAJ-6	1
TRAV-7	CAASRPSGGNYKPTF	TRAJ-6	1
TRAV-6	CVLGETSGGNYKPTF	TRAJ-6	1
TRAV-10	CALRTSGGNYKPTF	TRAJ-6	2
TRAV-7	CAVSGMGYKLT	TRAJ-9	1
TRAV-6	CALEDMGYKLTF	TRAJ-9	1
TRAV-12	CALSRNMGYKLT	TRAJ-9	1
TRAV-10	CAASRTRNMGYKLTF	TRAJ-9	1
TRAV-6	CVLGDQDMGYKLT	TRAJ-9	1
TRAV-7	CAASEGQGGRALIF	TRAJ-15	1
TRAV-7	CAASEGGRALIF	TRAJ-15	1
TRAV-7	CAASAYQGGRALIF	TRAJ-15	1
TRAV-7	CAASSYQGGRALIF	TRAJ-15	1
TRAV-7	CAASSYQGGRALIF	TRAJ-15	1
TRAV-7	CAASQGGRALIF	TRAJ-15	1
TRAV-7	CAASEHDQGGRALIF	TRAJ-15	1
TRAV-7	CAASERQGGRALIF	TRAJ-15	1
TRAV-7	CAASELGGRALIF	TRAJ-15	1
TRAV-7	CAASEPYQGGRALIF	TRAJ-15	1
TRAV-7	CAASEVYQGGRALIF	TRAJ-15	1
TRAV-7	CAASELGQGGRALIF	TRAJ-15	1
TRAV-7	CAASEAYQGGRALIF	TRAJ-15	1
TRAV-7	CAASEQGGRALIF	TRAJ-15	2
TRAV-7	CAASAQGGRALIF	TRAJ-15	2
TRAV-7	CAASEHQGGRALIF	TRAJ-15	2
Out-of-frame
TRAV-10	CAASP…TSGGNYKPTF	TRAJ-6	2
TRAV-7	CAASE…GLQQQQTYF	TRAJ-7	1
TRAV-6	CALSE…GLQQQQTYF	TRAJ-7	1
TRAV-14	CAAS…GLQQQQTYF	TRAJ-7	1
TRAV-7	n.a.	TRAJ-9	3
TRAV-14	CAAS…MGYKLTF	TRAJ-9	1
TRAV-12	CALS…MGYKLTF	TRAJ-9	15
TRAV-7	CAARD…GTGGYKVVF	TRAJ-12	1
TRAV-7	CE…GGYKVVF	TRAJ-12	1

Interestingly, in both the mock-transduced TCR⁻ cells and the Ikaros-transduced TCR⁻ cells, we detected at high frequency an in-frame TRAV1 (Vα19)-containing transcript. Since TRAV1 is the most 5’-Vα gene segment, rearrangements involving TRAV1 are terminal (no further secondary rearrangements are possible on that chromosome). As such, a TRAV1 rearrangement could be expected to accumulate in JE131 cells, since it could not be replaced by another rearrangement. This TRAV1-containing α chain cannot pair with the JE131’s Vβ8.1-containing β chain to assemble surface αβTCRs (data not shown). Therefore, rearrangements of TRAV1 are enriched in the TCR⁻ populations. In vivo, TRAV1 is almost exclusively expressed in a canonical rearrangement to TRAJ33 (rather than to the TRAJ9 seen in JE131 cells), where it is part of the TCR of the MAIT population of intestinal T cells (26). Rearrangement to other TRAJ elements, even if in frame, may often be unable to pair effectively with most TCR β chains (27).

To confirm that Ikaros is in fact inducing de novo TCR α rearrangement, and not simply expression, we examined TCR induction by Ikaros in JE131 cells after diminishing the level of RAG1, one of the two components of the V(D)J recombinase complex required for recombination, via short-hairpin RNA(shRNA)-mediated knock-down. If the Ikaros-induced appearance of TCR⁺ cells requires de novo rearrangement, this process should be dependent on RAG1. As expected, knock-down of Rag1 reduced Rag1 expression, as these cells exhibited a dramatic reduction in their ability to activate the RSS-GFP-RSS recombination substrate (Fig. 5B). Knock-down of Rag1 also greatly reduced Ikaros-mediated increase of TCR⁺ cells (Fig. 5A, bottom), indicating that new TCR⁺ cells arose from new RAG-1-dependent α rearrangements. Knock-down of Rag1, in the absence of Ikaros, had no effect on the pre-existing TCR⁺ cells (Fig. 5A, top). These data show that the induction of TCR by Ikaros is due to the activation of TCR α gene rearrangement.

Figure 5. Ikaros-driven appearance of TCR⁺ cells is RAG1/2-dependent.

(A) JE131 cells were depleted of TCR⁺ cells by MACS separation and transduced with empty pQXCIP retrovirus (blue) or one encoding the RAG-1 shRNA (red). 2 days later, each culture was transduced with empty MIT or MiT-Ikaros retroviruses. Three days later CFP⁺/Thy1.1⁺ cells were analyzed for TCR Cβ expression. Results shown are representative of two independent experiments. (B) JE131 cells were transduced with empty pQXCIP retrovirus (blue) or one encoding shRNA against RAG-1 (red). 2 days later the cells were transduced with RSS-GFP-RSS IRES hCD4 retroviruses. Three days later CFP⁺/hCD4⁺ were analyzed for GFP expression. Results shown are representative of two independent experiments.

Ikaros acts directly on the TCR α locus, not by induction of Rag1/2 expression

Although the Rag1/2 genes are constitutively expressed in JE131 cells, re-introduction of Ikaros might further increase Rag1/2 gene expression or activity enhancing TCR α rearrangements indirectly. However, Ikaros did not increase Rag1/2 transcript levels when compared to cells transduced with empty vector control (Fig. 6A). Similarly, Ikaros did not increase RAG1/2 activity. In cells transduced with retroviruses encoding RSS-GFP-RSS (25), Ikaros did not alter the percentage of cells that rearranged the substrate (the percentage of GFP⁺ cells) (Fig. 6B). These data are consistent with the idea that Ikaros acts directly on the TCR α locus.

Figure 6. Ikaros acts directly on the TCR α locus.

(A) Analysis of Rag1/2 expression by real-time PCR. JE131 cells were transduced with MiT or MiT-Ikaros retrovirus and MACS-sorted on day 3 based on Thy1.1 expression. For the sort, MiT- or MiT-Ikaros-transduced cells were stained with a PE-conjugated anti-Thy1.1 antibody, and cells were sorted using anti-PE beads. The elution (Thy1.1⁺ cells) (MiT-Ikaros sample) or flow-through (Thy1.1⁻ cells) (MiT sample) from each column was collected and used to prepare cDNA. Data are the results of one experiment using three technical replicates. (B) JE131 clone 6 cells were mock-transduced or transduced with MiT-Ikaros retrovirus. 24 hours later, samples were transduced with the RSS-GFP-RSS recombination substrate and monitored for GFP expression by flow cytometry. Histograms (gated on live and hCD4⁺ cells) show GFP expression 4 days after Ikaros transduction (3 days after RSS-GFP-RSS transduction). The percentage of GFP cells was also similar between the two samples on the previous 2 days. Data are representative of two independent experiments. (C) JE131 cells were transduced with MiT or MiT-Ikaros. 3 days later cells were stained with a PE-conjugated anti-Thy1.1 antibody. Thy1.1⁺ cells were obtained by MACS separation with anti-PE beads. Sorted cells were cross-linked, lysed, and Flag-Ikaros:DNA complexes were precipitated with anti-Flag antibody. Purified DNA was analyzed by real-time PCR. Graph shows relative enrichment of Eα sequences in empty vector or Ikaros-transduced samples. Results shown are representative of two independent experiments; real-time PCR performed using three technical replicates per sample.

Ikaros binding sites can be found in many DNA regulatory regions within the TCR α locus. Ikaros binds to the enhancer (Eα) in vivo (10), a region that is essential for recombination (28). Thus, to determine whether Ikaros promotes rearrangement in JE131 cells through interactions with the enhancer, we performed ChIP using FLAG-tagged Ikaros-transduced cells and FLAG-specific antibodies. Indeed, DNA sequence from the Eα was highly enriched in DNA associated with Ikaros in the ChIP (Fig. 6C), suggesting that Ikaros promotes rearrangement through direct effects on the TCR α locus and that Ikaros’ interaction with Eα may be important for the induction of rearrangement.

SWI/SNF chromatin-remodeling complexes are required for induction of TCR α rearrangement

If Ikaros acts directly on the TCR α locus to promote new rearrangements, it is plausible that it does so by enhancing accessibility via chromatin remodeling. Ikaros associates with SWI/SNF complexes in T cells (21), and SWI/SNF complexes play a critical activating role in TCR β gene rearrangement in vivo (3, 4). Thus, we asked whether SWI/SNF is important for Ikaros-mediated induction of TCR α rearrangement in JE131 cells. To answer this question, we examined TCR induction by Ikaros after diminishing the level of SWI/SNF via shRNA-mediated knock-down of Brg1 and Brm, the ATP-dependent catalytic subunits of SWI/SNF complexes. Knock-down of Brg1 and Brm (Fig. 7B) greatly reduced Ikaros-mediated induction of TCR α rearrangements significantly. Knock-down of Brg1 and Brm, in the absence of Ikaros, also had a slight effect on TCR expression (Fig. 7A, middle). Notably, knock-down of Brg1 and Brm did not reduce RAG1/2 expression or activity, as these cells did not exhibit a reduction in their ability to activate the RSS-GFP-RSS recombination substrate (Fig. 7C). These data show that the activation of TCR α gene rearrangement by Ikaros requires SWI/SNF.

Figure 7. Ikaros-mediated TCR induction requires SWI/SNF and is antagonized by Mi2β/NuRD.

(A) JE131 cells were transduced with empty pQXCIP vector or retrovirus encoding GFP (as a marker for transduction) and a shRNA directed against either Brg/Brm or Mi2β. 24 hours later, each culture was mock-transduced or transduced with MiT-Ikaros retrovirus. 3 days later (4 days after transduction with the shRNAs and 3 days after transduction with MiT-Ikaros) the cells were stained for Thy1.1 and TCR-Cβ and analyzed by flow cytometry. Histograms (gated on live, Thy1.1⁺, and GFP^high cells) show TCR-Cβ expression on mock- or Ikaros-transduced cells that are also expressing either shRNA (or the empty vector control). Results shown are representative of two independent experiments. (B) Western blot demonstrating efficacy of shRNA constructs. JE131 cells were transduced with retroviruses and stained as above. For doubly-transduced cells, live, Thy1.1⁺, GFP⁺ cells were MoFlo-sorted. For singly-transduced cells (shRNA only), live, GFP⁺ cells were MoFlo-sorted. Whole-cell lysates of sorted cells were run on a SDS-PAGE gradient gel and transferred to PVDF membrane. Membranes were probed with antibodies against Brg, Brm, or Mi2β. HDAC2 was used as a loading control. (C) JE131 cells were either mock transduced or transduced with retrovirus shRNAs directed against either Luciferase(negative control), Brg/Brm, or Mi2β. These viruses also expressed CFP (cyan fluorescent protein) as a marker of transduction. 24 hours later, each culture was transduced with virus encoding the recombination substrate RSS-GFP-RSS (Fig. 4D). This virus also encoded hCD4 as a marker of transduction. 3 days later (4 days after transduction with the shRNAs and 3 days after the RSS-GFP-RSS transduction) cells were analyzed by flow cytometry for recombination of the substrate . Mock transduced cells were gated on live, hCD4⁺ cells, while the shRNA transduced cells were gated on live, hCD4⁺, and CFP⁺ cells. The histograms show GFP expression from the recombined substrate in the gated cells. Results shown are representative of two independent experiments with similar results. (D) cDNA from the same experiment shown in Supporting Information Tables 1 and 2 was used to determine the relative expression of the TRAV1:TRAJ9 mRNA using real-time PCR. The relative expression of this rearrangement is shown for the sorted TCR⁻ cells that were either mock- and Ikaros- transduced. Data are representative of three independent real-time PCR experiments with similar results, each using three technical replicates.

Mi2/NuRD chromatin-remodeling complexes antagonize induction of TCR α rearrangement

Mi2/NuRD complexes have been associated with the inactivation of transcription at various loci (29). Ikaros associates with Mi2/NuRD complexes in T cells (21, 30), but it is unknown whether these chromatin remodeling complexes play a role in V(D)J recombination. Therefore, we tested whether Mi2/NuRD plays a role in Ikaros-mediated induction of TCR α rearrangement in JE131 cells. We examined TCR induction by Ikaros after diminishing the level of Mi2/NuRD complexes via shRNA-mediated knock-down of Mi2β, the ATP-dependent catalytic subunit of the NuRD complexes. Knock-down of Mi2β (Fig. 7B) greatly enhanced Ikaros-mediated induction of TCR α rearrangements. Knock-down of Mi2β in the absence of Ikaros had little effect (Fig. 7A, bottom). These data show that Mi2/NuRD complexes have an antagonistic effect on Ikaros-mediated TCR α gene rearrangement. Furthermore, these data confirm that Mi2/NuRD acts in opposition to SWI/SNF complexes in V(D)J recombination, similar to their antagonism in the regulation of lymphocyte-specific transcription (31).

Ikaros enhances transcription of a rearranged TCR α gene in JE131 cells

Another way that Ikaros may promote TCR α rearrangement is by enhancing transcription. As a first step to answering this question, we asked whether Ikaros had any effect on transcription of rearranged TCR α genes. To do so, we compared mRNA levels of the terminal, irreplaceable in-frame JE131 TRAV1:TRAJ9 rearrangement described above (Supporting Information Tables 1 and 2) in Mock- and Ikaros-transduced cells (using the same cDNA as before) by real-time PCR. Spontaneous rearrangement in JE131 cells occurs very slowly, and because this rearrangement was also strongly represented in the Mock-transduced population, it is likely that the rearrangement existed in most cells prior to transduction. Interestingly, mRNA levels of the TRAV1:TRAJ9 rearrangement were dramatically increased in Ikaros-transduced cells (Fig. 7D) consistent with the well establish linkage between transcription and recombination in the loci of immune receptors.

Discussion

The transition of pre-T cells to DP TCR⁺ thymocytes involves an enormous re-programming of expression of multiple genes throughout the genome. However, because this transition takes place asynchronously in a small subpopulation of thymocytes, studying the coordination of gene expression or the biochemistry of the process has been very difficult. The JE131 cell line, derived from an Ikaros^−/− thymoma, offers a unique tool that partially overcomes this problem. JE131 cells appear bottlenecked at this transitional stage and have a pre-T cell-like phenotype. However, re-introduction of Ikaros into these cells leads to their differentiation to a DP TCR⁺ phenotype accompanied by many of the changes in gene expression seen in normal thymocytes. The process is very rapid, directly related to the level of Ikaros expression and amenable to studying the effects of specific inhibitors or introduced genes on the transition.

However, like thymocytes in Ikaros^−/− mice, the bottleneck is not absolute. First of all, in what appears to be a stochastic, uncoordinated manner, a small portion of individual JE131 cells spontaneously up- or down-regulate particular genes associated with the transition (e.g. Cd4, Cd8, Cd25, etc.) (23). Secondly, unlike normal pre-T cells, JE131 cells constitutively express Rag1/2 at functional levels, which has led to the eventual accumulation of rearrangements of the TCR α chain locus on both chromosomes, nearly all of which involve out-of-frame VαJα junctions. Secondary α locus rearrangements continue to occur in culture at a low rate in a manner uncoordinated with occasional up- or down-expression of the other transitional genes. These results suggest that Ikaros enhances, but is not absolutely required for the expression of many of the genes it regulates.

Despite its differences from normal pre-T cells, we show here that JE131 can be used to study how Ikaros activates new rearrangements within the TCR α locus, while promoting the transition to a DP phenotype. We demonstrate that these cells produce a single functional Vβ8.1⁺ β chain, which in most cells is expressed on the cell surface with the pre-Tα chain. Retroviral re-introduction of Ikaros into JE131 not only sets in motion the coordinated changes in transitional genes such as Cd4 and Cd8, but also leads to rapid secondary rearrangements in the α locus, producing many new functional, in frame α chains. This is accompanied by a dramatic increase in transcription from the α locus and in αβTCR⁺ cells with the loss of pre-Tα surface expression.

Ikaros has many potential binding sites throughout the TCR α locus and it is likely that Ikaros acts directly. While our studies do not formally demonstrate that Ikaros induces VJ recombination via its direct interaction with the TCR α locus, others have reported the interaction of Ikaros with the TCR α and β loci in normal T cells (10) and we show here that Ikaros binds to a critical enhancer within the α locus in JE131 cells. In B cells, Ikaros promotes V(D)J recombination of the Igh locus during development in a seemingly direct manner (22) and Ikaros directly regulates class-switch recombination (CSR) in mature cells (33). Thus, by analogy, it is reasonable to postulate that Ikaros also directly regulates the recombination of Tcr genes.

Ikaros could influence α locus recombination in several ways. Ikaros-mediated recruitment of SWI/SNF complexes and other subsequently attracted chromatin remodeling proteins (i.e., histone acetyl- or methyl-transferases) may make the Vα and Jα recombination signal sequences (RSSs) more accessible (Supporting Information Fig. 2A). Our data suggest that the ability of Ikaros to promote secondary TCR α rearrangements requires the SWI/SNF chromatin-remodeling complex (and that this process is inhibited by the Mi2/NuRD complex). Ikaros is known to recruit SWI/SNF to other target genes (21), and chromatin-remodeling is known to be important for TCR gene rearrangement (3, 4, 32). Our data also show that Ikaros dramatically promotes transcription of a rearranged TCR α gene in JE131 cells again consistent with increased accessibility of the transcription machinery to the locus following Ikaros mediated chromatin remodeling. Given the long standing linkage between recombination and transcription (22, 33), it is possible that germline transcripts from the locus play a direct role in the initiation of recombination. Ikaros may also facilitate recombination by enhancing the locus contraction required to juxtapose Vα and Jα RSSs for RAG-mediated recombination (Supporting Information Fig. 2B). Ikaros appears to act via the same mechanism to promote Igh locus rearrangement during B cell development (22). A third way in which Ikaros may promote rearrangement is by controlling the sub-nuclear localization of the TCR α locus, perhaps by recruiting the TCR α loci to chromatin compartments that are permissive to recombination (Supporting Information Fig. 2C).

Our results do not formally speak to the question of the role Ikaros in initial α rearrangements during the transition from pre-T cell to DP αβTCR+ thymocyte. Since, these initial rearrangements occurred at some earlier point during the development of the JE131 cell line and also can occur in Ikaros null^−/− thymocytes, Ikaros is not absolutely required to begin the process. Perhaps in T cells other members of the Ikaros family, such as Helios or Aiolos, can partially provide this function.

In summary, we present a detailed characterization of the JE131 cell line and data which identifies Ikaros as a regulator of TCR α gene rearrangement. We also present preliminary evidence suggesting that the mechanism responsible for Ikaros-dependent activation of TCR α gene rearrangement involves chromatin-remodeling. The mechanistic details have yet to be elucidated and confirmed in vivo; however we have demonstrated that the JE131 cell line is a powerful system that can be used to study the events taking place during the transition of thymocytes from the late DN to DP stage. The fact that this cell line can be very efficiently transduced with retroviruses creates a unique tool for rapid evaluation of specific genes in this process. This cell line may also prove useful for studying other functions of Ikaros in T-cell development such as pre-TCR signaling and the regulation of pre-Tα expression.

Materials & Methods

Reagents and Cell Lines

JE131 cells were obtained from S. Winandy and maintained as described (23). Phoenix Eco packaging cells were used for production of retroviruses. M12C3 cells are a B cell lymphoma line kindly provided by L. Glimcher (34). µM2 cells, a plasmacytoma line, have been described previously (35). TCRα^−/− thymocyte RNA was provided by T. Potter. The γδT cell hybridoma was provided by W. Born and R. O’Brien. Oligonucleotides were synthesized in the NJH Molecular Resource Center (sequences are listed by number in Supporting Information Table 3).

Retroviral Constructs

MSCV-IRES-GFP and MSCV-Ikaros-IRES-GFP were provided by S. Winandy (23). MSCV-Ikaros-IRES-Thy1.1 was generated by sub-cloning the Ikaros insert from MSCV-Ikaros-IRES-GFP into MSCV-IRES-Thy1.1. MSCV-DOα-IRES-GFP encodes the full length TCR α chain of DO-11.10 (36). MSCV-RSS-GFP-RSS-IRES-hCD4 was provided by B. Sleckman (25). pQXCIP-IRES-GFP and pQXCIP-shRNA-IRES-GFP (encoding shRNAs against Brg/Brm or Mi2β) have been previously described (31). For inhibition of Rag1 with shRNA, we annealed oligonucleotides 1 and 2 and cloned the fragment into the HpaI/XhoI sites pQXCIP-shRNA-IRES-CFP. MSCV-p27-IRES-GFP, encoding mouse p27, was purchased from AddGene.

Retroviral Packaging

10 µg of retroviral plasmids were co-transfected into Phoenix Eco cells with 2 µg of the pCLEco accessory plasmid using Lipofectamine 2000 (Invitrogen) in 10 cm petri dishes. Retrovirus-containing supernatants were harvested 48 and 72 hr post-transfection, 0.45 µm filtered, and stored at −80°C.

Retroviral Transductions

Viral supernatants were used to infect JE131 cells using 1 mL of viral supernatant (plus 8 µg polybrene) per 2.5×10⁶ cells. Cells were spinfected in 24-well plates at 650 × g for 2 hours at 37°C. Supernatants were removed and cells were maintained in RPMI complete medium.

PCR

cDNA was generated from total JE131 RNA with a Superscript kit (Invitrogen). Genomic DNA was isolated by lysing cells in 100mM Tris pH 8.0, 200mM NaCl, 0.2% SDS, 5 mM EDTA, boiling at 95°C for 5 min., incubating with Proteinase K for 1 hour at 55°C, and isopropanol-precipitating.

For standard PCR, the following primer pairs were used: Ptcra-3/4; Sequences of primers used to amplify genomic segments: Cα-11/12, Jα49-13/14, Jα30-15/16, Jα20-17/18, Jα18-19/20, Jα18-21/22, Jα9-23–24.

For TCR α sequencing, various Vα forward primers, 25–37, were used in separate PCR reactions with a common Cα reverse primer, 38, to amplify TCR α sequences from cDNA. Forward Vα primers were designed to amplify nearly all Vα sequences. To increase yield and purity, PCR products were gel-purified and gel slices were used as template in a second nested PCR reaction, using the same forward primers and an internal Cα reverse primer, 39. PCR products were gel-purified, pooled, and cloned into pCR-2.1-TOPO vector [TOPO-TA Cloning Kit (Invitrogen)]. TOPO cloning reactions were transformed into XL-1 Blue (Stratagene) or Top10 cells, plated on LB/Carbenicillin plates, and grown overnight at 37°C. Colonies were screened for the presence of an insert by PCR using M13 primers, 40/41. PCR products from positive clones were purified using S-300 columns (GE Biosciences) and sequenced using either of the M13 primers above.

To sequence the TCR β chain, cDNA was prepared as above from the parent JE131 cell line or sub-clones. cDNA was used as a template in standard PCRs with primers 42/43. PCR fragments were sequenced with primer 42.

For quantitative real-time PCR, the following primers were used: Rag1-44/45; Rag2-46/47, Hprt-5/6, Trav1:Traj9-37/50. PCR was performed using SYBR Green 2x master mix (Fermentas).

Flow Cytometry

For analysis and MoFlo sorting, cells were incubated with FcR-blocking antibody (24G2) and each fluorochrome-conjugated antibody for 20 minutes on ice, washed once with BSS, and immediately analyzed/sorted. Antibodies used for analysis/MoFlo sorting were: anti-Cβ-allophycocyanin (H57–597)(BD Biosciences), anti-Thy1.1-PacBlue (OX-7)(BioLegend), anti-Thy1.1-PerCP (OX-7)(BD Biosciences), anti-Thy1.1-PE-Cy7 (HIS51)(eBioscience), anti-CD8α-A647 (53-6.7)(made in house), anti-CD4-allophycocyanin (eBioscience)(GK1.5) anti-CD25-allophycocyanin (PC61)(BD Biosciences), anti-Vβ8-PE (F23.1)(BD Biosciences), anti-hCD4-PacBlue (OKT4)(eBioscience), anti-hCD4-allophycocyanin (RPA-T4)(BD Biosciences), anti-pTα (2F5)(BD Biosciences) or isotype control (MOPC-31C)(BD Biosciences) with biotin-anti-IgG1 secondary (A85-1)(BD Biosciences) and streptavidin-PerCP tertiary (BD Biosciences). FACS data were analyzed and graphed using FlowJo software (Tree Star). General gating strategies are described in the Supplemental Methods. More involved gating strategies are described in individual figure legends.

Staining for MACS sorting was performed according to the manufacturer’s instructions (Miltenyi Biotec) using anti-Cβ-allophycocyanin (same as above) and anti-Thy1.1-PE (OX-7)(BD Biosciences).

Western Blot Analysis

A detailed description of Western blot analysis can be found in the Supporting Information. Transduced, sorted cells were washed in PBS and stored at −80°C. Cells were thawed and lysed for 10 minutes on ice in lysis buffer with protease inhibitors. Cell debris was removed by centrifugation. Supernatants were mixed with 4x SDS-PAGE loading dyes then boiled. Equal numbers of cell equivalents of each sample were loaded onto a gradient gel and proteins were separated by gel electrophoresis. Proteins were transferred to PVDF membranes (Millipore), blocked in TBST+5% milk, then probed with 1° antibodies to either Brg1 (Upstate USA, Inc), Brm (BD Biosciences), Mi2β (Santa Cruz Biotechnology), or HDAC2 (Invitrogen). Membranes were washed and incubated in HRP-conjugated 2° antibody (donkey-anti-rabbit or goat-anti-mouse) (Jackson ImmunoResearch Laboratories) in TBST+5% milk, washed, incubated in ECL+ chemiluminescent reagent (GE Biosciences), and exposed to x-ray film (Thermo Scientific).

Chromatin Immunoprecipitation

A detailed description of the ChIP experiment can be found in the Supporting Information. In brief, MiT- or MiT-Ikaros-transduced JE131 cells were stained and MACS sorted to isolate Thy1.1⁺ (transduced) cells. Cells were resuspended in RPMI complete medium and protein:DNA complexes were cross-linked with formaldehyde. Cross-linking was terminated with glycine. Cells were harvested by centrifugation, washed twice in cold PBS, and lysed in lysis buffer with protease inhibitors. Chromatin was sheared by sonication. Insoluble material was removed by centrifugation. Absorbance at 260nm was used to quantify the chromatin in each sample. Samples were diluted to normalize chromatin content. 10% of the total volume of each IP was removed and stored at −80°C. Samples were then incubated with the appropriate antibody. Immunocomplexes were collected with protein A/G beads (Diagenode). Beads were pelleted and washed. Immunocomplexes were eluted twice in elution buffer. Cross-linking was reversed and DNA was recovered by phenol:chloroform extraction and precipitation. DNA pellets were resuspended in water, passed through a MiniElute column (Qiagen), and stored at −20°C. Eα primers used for the detection of immunoprecipitated DNA (by real-time PCR) were 48/49.

Abbreviations

SWI/SNF

Switch/Sucrose Nonfermentable

NuRD

Nucleosome Remodeling and Deacetylase

shRNA

short hairpin RNA

Ptcra

pre-TCRα

GFP

green fluorescent protein

CFP

cyan fluorescent protein