Differentiation of Th cells from naive precursors is a dynamic process that involves multiple transcription factors acting at specific time points to regulate gene expression. In this study we show that the homeobox transcription factor Hlx is up-regulated early in Th1 cell differentiation. Mice constitutively expressing an Hlx transgene driven by a CD4 promoter showed marked reduction in the CD4+CD8+ thymocyte population. The Hlx transgenic mice generated increased numbers of Th1 cells in response to keyhole limpet hemocyanin immunization. After differentiation under Th2-polarizing conditions in vitro, the transgenic CD4 T cells expressed high levels of IFN-γ. Intracellular cytokine staining revealed that in addition to Th2 cells, large numbers of Th0 and Th1 cells were generated from such in vitro differentiated transgenic CD4 T cells. Retrovirally overexpressed Hlx also induced the aberrant expression of IFN-γ in normal CD4 T cells differentiated under Th2-polarizing conditions. This effect was apparent only when Hlx was introduced into the cells by retroviral infection at an early time point that led to the expression of the retrovirally transferred Hlx gene at a time comparable to that of the up-regulation of the endogenous Hlx during Th1 cell differentiation. Later infection with Hlx-expressing retrovirus showed no effect. Thus, the induction of IFN-γ expression by Hlx depends on a permissive epigenetic state of the IFN-γ gene locus and/or the molecular context of the immature Th cells.
Generation of appropriate Th cell responses is critical for the success of the immune system to defend its host against various infections, whereas inappropriate Th cell responses often lead to infectious and autoimmune diseases. Th1 cells produce predominantly IFN-γ and lymphotoxin-α. They are important for the clearance of intracellular infections and for organ-specific autoimmune diseases. In contrast, Th2 cells predominantly produce IL-4, IL-5, and IL-13. They are responsible for the clearance of helminth infections and allergic reactions (1, 2, 3).
The mutually exclusive pattern of cytokine expression of Th1 and Th2 cells is acquired through a differentiation process from naive CD4 T cells that leads to subset-specific expression of transcription factors. Naive CD4 T cells differentiate into Th2 cells after TCR stimulation in the presence of IL-4 (4, 5, 6). Via Stat6, IL-4 up-regulates the expression of GATA-3 (7, 8), which triggers an autoregulatory loop to maintain its own expression (9, 10). GATA-3 drives Th2 and blocks Th1 cell differentiation (8, 10, 11, 12, 13). In differentiated Th2 cells, GATA-3 is necessary for the expression of most, if not all, Th2 cytokines (8). In addition, Th2 cells specifically express the transcription factor c-Maf (14). C-Maf is required for the optimal expression of IL-4, but not other Th2 cytokines (14, 15, 16).
T-bet is the key transcription factor that drives Th1 cell differentiation (17). The differentiation of Th1 cells from naive CD4 T cells requires polarizing signals of IFN-γ and IL-12 (18, 19, 20, 21, 22). IFN-γ and TCR stimulation synergistically up-regulate T-bet expression (23, 24, 25). In addition, IFN-γ induces the expression of functional IL-12R (26, 27). The induction of functional IL-12R allows the naive CD4 T cells to respond to IL-12 to activate Stat4 (28, 29, 30). However, how Stat4 regulates Th1 cell differentiation is not very clear. Stat4 is not essential for T-bet up-regulation and Th1 cell differentiation (24, 25), but it is required for generating optimal Th1 responses (31, 32, 33).
Although T-bet is critical for Th1 cell differentiation and IFN-γ expression in CD4 T cells (34), how it activates IFN-γ gene expression is not fully understood. In the original study, it was found that T-bet activated an IFN-γ promoter (17), so it was believed that T-bet directly activated the endogenous IFN-γ gene. However, later studies showed that T-bet transcriptional activity is not required for IFN-γ expression in differentiated Th1 cells (35). In addition, although retroviral expression of T-bet in CD4 T cells undergoing Th2 cell differentiation can induce IFN-γ expression (17), the expression is not as strong as in bona fide Th1 cells (24, 25). These studies stressed the importance of additional transcriptional regulators in the differentiation of Th1 cells as well as in the optimal expression of IFN-γ in differentiated Th1 cells. To date, two additional Th1 cell-specific transcription factors have been reported. The Ets transcription factor ERM is specifically induced by IL-12 during Th1 cell differentiation (36). It appears to enhance IFN-γ expression in Stat4 haplo-insufficient Th1 cells. More recently, Mullen et al. (35) reported that the homeobox protein Hlx was specifically expressed in Th1 cells. When retrovirally coexpressed with T-bet in developing Th2 cell, Hlx enhanced T-bet-induced IFN-γ expression.
We also identified Hlx as a Th1 cell-specific transcription factor in a study of cDNA subtraction between Th1 and Th2 cells. Kinetic studies showed that Hlx is expressed at a steady level in naive CD4 T cells, initially down-regulated during both Th1 and Th2 differentiation, and up-regulated upon commitment to Th1 fate. Transgenic mice constitutively expressing Hlx in CD4 T cells generated increased numbers of Th1 cells in response to keyhole limpet hemocyanin (KLH)5 immunization. Large numbers of cells in Hlx transgenic CD4 T cells differentiated in vitro under Th2-polarizing conditions produced both IFN-γ and IL-4, and Th1 cell differentiation was also enhanced under the same conditions. Similar results were obtained by retroviral overexpression of Hlx alone in normal CD4 T cells at an early, but not a later, time during Th2 cell differentiation.
Materials and Methods
Preparation of naive CD4 T cells and APC
CD4 T cells and APC were prepared as previously described (8). CD62Lhigh naive CD4 T cells were isolated either by MACS (Miltenyi Biotec, Auburn, CA) or FACS sorting of the double-stained CD4+CD62Lhigh population.
In vitro Th cell differentiation
Naive CD4 T cells (5 × 105/ml) and APC were mixed at 1/1 ratio and stimulated with Con A (2.5 μg/ml; Roche, Mannheim, Germany) plus human IL-2 (50 U/ml; Chiron, Emeryville, CA) in IMEM containing 10% FBS. Alternatively, naive CD4 T cells were stimulated with plate-bound anti-TCRβ Ab (H57; gift from Dr. D. Fowell, University of Rochester, Rochester, NY) and anti-CD28 Ab (BD PharMingen, San Diego, CA). For Th1 cell differentiation, cultures were supplemented with 5 ng/ml IL-12 (gift from Genetics Institute, Cambridge, MA) plus anti-IL-4 Ab (11B11). For Th2-polarizing conditions, cultures were supplemented with 30 ng/ml IL-4 (BD PharMingen) plus Abs against IFN-γ (XMG1.2) and IL-12 (BD PharMingen). The cells were cultured for 4 days, then rested in fresh medium containing 50 U of IL-2/ml for 1.5 days. APC from B6C3F1 were used for transgenic mice and their littermates. The differentiated cells were restimulated with Con A (2.5 μg/ml) for only 20 h, and supernatants were harvested to measure cytokine secretion by ELISA. Alternatively, the cells were stimulated with PMA and ionomycin (Calbiochem, La Jolla, CA) for intracellular cytokine staining.
Representational difference analysis (RDA)
RDA analysis was performed as previously described (37). Briefly, cDNA from the in vitro differentiated Th1 and Th2 cells was digested with DpnII and ligated to the R-Bgl-12/24 adapters. Representations were prepared by PCR amplification of the cDNA fragments using R-Bgl-24 as primer. Subtraction was performed using Th1 cell representation as tester and Th2 representation as driver. The tester/driver ratios are 1:100, 1:800, and 1:400,000 for the first, second, and third subtractions, respectively. Southern blot analysis of the RDA products was conducted according to standard procedures. The third difference products were cloned, and individual clones were sequenced to identify the gene fragments in the RDA final difference analysis.
Complementary DNA library construction and screening
MACS-isolated naive CD4 T cells were stimulated with Con A and differentiated in vitro under Th1-polarizing conditions for 4 days. Total RNA was isolated by CsCl2 gradients. The cDNA synthesis and library construction were conducted using a ZAP Express/Gigapack III Gold Cloning Kit (Stratagene, La Jolla, CA). Recombinant phage colonies (∼1 × 106) were screened with Th1.Dp3 as probes using the two-step strategy described previously (8).
The plasmid pEF-Hlx-MC1neoPA containing the Hlx cDNA was a gift from Dr. J. Adams (Walter and Elixa Hall Institute of Medical Research, Melbourne, Australia) (38). The full-length Hlx cDNA was excised from pEF-Hlx-MC1neoPA and cloned to the SalI site of the p37.1 plasmid (a gift from Dr. D. Littman, New York University, New York, NY) downstream of the CD4 promoter to generate plasmid pCD4-Hlx+. The NotI fragment containing the CD4-Hlx sequence was isolated from pCD4-Hlx+ and used to generate Hlx transgenic mice. Transgene-positive mice were identified by detecting the 1.9-kb BglII fragment of the transgene by Southern hybridization using the Hlx cDNA probe. The mice were initially created on the B6C3 background, then extensively backcrossed to C57BL/6. Transgenic lines were named after their original founders.
MHC class II Ag− CD4 T cells were isolated by negative selection with MACS (Qiagen, Valencia, CA). CD4 T cells (1 × 107) were mixed with 2 × 107 T cell-depleted spleen and lymph node cells. The cells were suspended in 0.5 ml of plain HBSS and i.v. injected into a B6.129S7.Rag1−/− mouse (National Institutes of Health, Bethesda, MD). Twenty-four hours later, the recipient mice were immunized with KLH.
Mice were immunized at the footpads of hind legs and tail base with 50 μg of KLH (Sigma-Aldrich, St. Louis, MO) in CFA (Difco, Detroit, MI) per site. On day 10 after immunization, draining lymph nodes (DLN) were collected. DLN cells (5 × 106/ml) were stimulated in vitro with KLH at 100 μg/ml for 45 h. Supernatants from the in vitro stimulated DLN cells were harvested for analysis of cytokine production by ELISA. For intracellular cytokine staining, mice were immunized in the same way. Ex vivo DLN cells were stimulated with PMA plus ionomycin, then stained for CD4 and IFN-γ or IL-4.
Analysis of cytokine expression
Cytokine levels in cell culture supernatants were analyzed by ELISA based on standard procedure. Briefly, supernatants were added to Nunc-Immuno plates (Nalge Nunc International, Copenhagen, Denmark) coated with either anti-IL-4 (BVD4-1D11) or anti-IFN-γ (R4-6A2) Abs (BD PharMingen). The captured cytokines were then incubated with biotinylated anti-IL-4 (11B11) or anti-IFN-γ (XMG1.2) Abs (BD PharMingen). The cytokines were detected with avidin-conjugated HRP (Vector Laboratories, Burlingame, CA) using 3,3′,5,5′-tetramethylbenzidine substrate and stop solution (Kirkegaard & Perry Laboratories, Gaithersburg, MD). The ELISA plates were read on a Vmax Microplate Reader, and data were analyzed with SOFTmax Pro software (Molecular Devices, Sunnyvale, CA). For intracellular cytokine staining, resting CD4 T cells were stimulated with PMA (50 ng/ml) and ionomycin (1 μM). After 2 h of stimulation, brefeldin A (10 μg/ml; Sigma-Aldrich) was added to the cultures, and cells were stimulated for an additional 4 h. Cells were then fixed and permeabilized using CytoFix/CytoPerm solutions (BD PharMingen), and were stained with allophycocyanin-conjugated anti-CD4 Abs, anti-IL-4, and/or anti-IFN-γ Abs conjugated with FITC or PE (BD PharMingen). Cells were analyzed on a FACSCalibur.
Regular and real-time RT-PCR
Total RNA extracted from cells was reverse transcribed using Superscript II reverse transcriptase (Invitrogen, Carlsbad, CA) according to the manufacturer’s instructions. Sequences for all PCR primers for regular and semiquantitative RT-PCR are as follows: β-actin 5′, gtg ggc cgc tct agg cac ca; β-actin 3′, cgg ttg gcc tta ggg ttc agg ggg g; CD4e1, gtg aag gaa gga ctg gcc ag; Hlxtg+, ccc acc agt ccc tct gca gc; hypoxanthine phosphoribosyltransferase (HPRT) 5′, gca tca att cca aga cac att tcc; HPRT 3′, tcg gat atc cgg tcg gat ggg ag; Hlxi 5′, gca tca att cca aga cac att tcc; and Hlxi 3′, tcg gat atc cgg tcg gat ggg ag. Primers for real-time PCR were: Tubb5.211F, gac cga atc tct gtg tac tat aat g; Tubb5.365R, cca gac tga ccg aaa acg aag t; Hlx.1360F, gtg acc aag cca gac cga aag; and Hlx.1497R, ctt cct tgt cct tgt cct tct g. For semiquantitative PCR amplification of Hlx, the amounts of input cDNA samples were normalized using HPRT as a reference. Real-time PCR was performed using the SYBR Green PCR Master Mix (PR Applied Biosystems, Foster City, CA) according to the manufacturer’s instructions. The reactions were run on an ABI PRISM 7900HT Sequence Detection System (PE Applied Biosystems). Each sample was amplified in triplicate. Data were analyzed with SDS software. Dissociation curves were generated to ensure a single peak of PCR products in each reaction. The transcripts of each gene relative to the β-tubulin V gene were determined using the 2−ΔCt method as previously described (39).
Retroviral infection of T cells
The retroviral vector MigR1, and Hlx-MigR1 for bicistronic expression of Hlx and green fluorescence protein (GFP) were gifts from Dr. S. Reiner (University of Pennsylvania, Philadelphia, PA). The packaging cell line Phoenix Eco was originally developed in Dr. G. Nolan’s laboratory (Stanford University, Stanford, CA), and was purchased from American Type Culture Collection (Manassas, VA). The packaging cells were transfected with retroviral vectors using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s instruction. Forty-eight hours post-transfection, culture supernatants were harvested and centrifuged at 6000 × g to concentrate the viruses. For T cell infection, naive CD4 T cells were stimulated under specific differentiation conditions for 12 or 36 h. Culture supernatants were removed and kept at 4°C. Cells were mixed with fresh prepared viruses plus 8 μg/ml polybene (American Bioanalytical, Natick, MA) and 50 U/ml IL-2, then centrifuged at 1800 rpm for 45 min at room temperature. After 12 h of incubation at 37°C, viruses were removed, and the saved culture supernatants were added back to the cells. Three days later, the cultures were split 1/3, and cells were expanded for 2 more days. Then, the cells were washed and rested in fresh medium plus 50 U/ml IL-2 for 36 h. Resting cells were restimulated for intracellular cytokine staining or ELISA detection of cytokine production.
Expression of Hlx in naive CD4 T cells and developing Th cells
To understand the molecular mechanisms underlying Th1 cell differentiation, we isolated genes that are specifically expressed in Th1 cells by cDNA RDA (37). Naive CD4 T cells were isolated by FACS sorting and induced to differentiate into Th1 or Th2 cells in vitro. After 4 days of differentiation, mRNA was extracted from the cells to prepare cDNA representation of Th1 and/or Th2 cells. Subtractive hybridization was performed using the Th1 representation as tester and the Th2 representation as driver. After three rounds of subtraction, we confirmed that genes common to both cell types had been significantly depleted by showing that the γ-actin signals in the Southern blot of the RDA difference products had been greatly reduced or eliminated (Fig. 1,A, lower panel). Sequencing of the cloned final difference products revealed that one of the most abundant Th1 cell-specific genes was the homeobox protein gene Hlx. We later used the total final Th1 difference products (Dp3) as probe to screen ∼1 × 106 recombinant clones of a Th1 cDNA library. Sequencing of positive clones identified a number of Hlx clones, many of which contained full-length cDNA of Hlx. Their sequences were all identical to the published Hlx cDNA sequence. Southern blot of the RDA difference products hybridized to Hlx cDNA probe showed a sequential increase in Hlx signals after each round of subtraction (Fig. 1 A, middle panel), demonstrating enrichment of the Hlx gene by RDA.
To confirm Th1 cell-specific expression of Hlx, we first performed semiquantitative RT-PCR analysis of Hlx expression in naive and newly committed Th1 and Th2 cells. As shown in Fig. 1,B, naive CD4 T cells expressed significant levels of Hlx. After 2 days of Th1 cell differentiation, Hlx was up-regulated, whereas in cells differentiated under Th2 conditions, it was down-regulated to an undetectable level. To further study the kinetics of Hlx expression in the course of Th1 and Th2 cell differentiation in a more quantitative way, we performed real-time PCR analysis of Hlx expression at different time points of differentiation (Fig. 1,C). For internal reference, we used the β-tubulin V gene, which is constitutively expressed in T cells (40). Consistent with the semiquantitative RT-PCR results, Hlx expression was detected at a significant level in naive CD4 T cells. After 1 day of differentiation, Hlx expression was down-regulated in cells of both Th1 and Th2 cultures (Fig. 1,C, inset). Unexpectedly, this initial down-regulation of Hlx was greater in Th1 culture than in Th2 culture. However, after 2 days of Th1 differentiation, Hlx was up-regulated to a level >10-fold that in naive CD4 T cells. In contrast, Hlx expression remained at low levels during Th2 cell differentiation. Thus, the kinetics showed three phases of Hlx expression in CD4 T cells: steady state expression in naive CD4 T cells, initial down-regulation during Th1 and Th2 cell differentiation, and up-regulation upon commitment to the Th1 fate. Finally, we compared Hlx expression in long term Th1 and Th2 clones, AE7 and D10. Albeit at a significantly lower level than in newly differentiated Th1 cells, the Hlx expression level was higher in AE7 cells than in D10 cells (Fig. 1 D).
Hlx transgenic mice and their T cell development
To study the functions of Hlx in Th cell differentiation, we created transgenic mice to constitutively express Hlx under the control of the CD4 promoter (Fig. 2,A). Initially we established four independent transgenic lines, two of which (lines 82 and 122) were used in the current study. The expression of the transgene in transgenic mice of these two lines in both the lymph nodes and thymi was verified by RT-PCR using primers derived from the first exon of the CD4 gene and the 5′ region of Hlx cDNA in the transgene construct (Fig. 2 B).
We first noticed that the Hlx transgenic mice of all four initial lines showed various degrees of reduction of the CD4+CD8+ population in the thymus. Of the two lines used in subsequent studies, transgene-positive mice of line 82 showed a mild decrease in the CD4+CD8+ population, whereas line 122 showed a severe diminution of this population (Fig. 2,C). Nonetheless, both lines produced CD4 and CD8 single-positive T cells. Total thymocytes were also decreased in the transgenic mice. Typically, the total numbers of thymocytes of transgenic mice of line 82 were ∼70% those of the wild type, whereas those of line 122 were ∼20%. In the periphery, a similar reduction of cellularity was found in the lymph nodes of transgenic mice (data not shown). In line 82, the distribution of CD4 and CD8 T cells in the lymph nodes of transgenic mice was not different from that in wild-type mice, whereas in line 122, the percentages of CD4 and CD8 T cells were significantly reduced in transgenic mice (Fig. 2,D). In both lines, the percentages of peripheral CD4 T cells of the CD62Lhigh naive phenotype were markedly lower than those in their wild-type littermates (Fig. 2 E).
Enhanced Th1 response to KLH immunization in Hlx transgenic mice
To investigate the roles of Hlx in Th cell differentiation, we first sought to examine in vivo the Th cell response to immunization with the protein Ag KLH in CFA. Ten days after the immunization, DLN cells were harvested and restimulated with KLH in vitro to measure cytokine production. As shown in Fig. 3 A, we detected strikingly greater amounts of IFN-γ produced by DLN cells of transgene-positive mice than by DLN cells of their wild-type littermates. In contrast to the robust production of IFN-γ, IL-4 production was minimal in both transgenic and wild-type mice, perhaps due to the C57BL/6 genetic background of the mice.
As DLN cells consisted of mixed populations of cells, to further analyze IFN-γ expression exclusively in CD4 T cells, we performed intracellular cytokine and CD4 staining of ex vivo DLN cells (Fig. 3 B). As expected, higher percentages of IFN-γ-producing CD4 T cells were found in the transgenic DLN cells than in the wild-type DLN cells in both lines. Consistent with only minimal IL-4 levels detected by ELISA, essentially no IL-4-producing CD4 T cells were detected by intracellular cytokine staining in both wild-type and transgenic DLN cells (data not shown). These results demonstrated that Hlx transgenic mice generated heightened Th1 responses to KLH immunization.
Another complicating in vivo factor were APC that might have functioned differently and affected Th cell differentiation in the DLN of the transgenic mice. To investigate whether Hlx transgene directly affects CD4 T cells, we performed adoptive transfer experiments. MHC class II Ag-negative CD4 T cells were isolated from transgene-negative and -positive mice and transferred to B6.129S7.Rag1−/− recipients. The recipient mice were then immunized with KLH, and DLN cells were analyzed for cytokine expression by intracellular cytokine staining (Fig. 3,C). As shown in Fig. 3 B, a higher percentage of IFN-γ-expressing CD4 T cells was detected in mice adoptively transferred with Hlx transgenic CD4 T cells than in mice transferred with transgene-negative CD4 T cells. Again, no significant number of IL-4-producing CD4 T cells was detected in either group. Thus, the Hlx transgene exerted a direct effect on CD4 T cells to induce IFN-γ expression.
Strong IFN-γ expression by Hlx transgenic CD4 T cells differentiated under Th2-polarizing conditions in vitro
The minimal amounts of IL-4 produced in the KLH immunization studies made it difficult to fully assess the impact of the Hlx transgene on Th2 cell differentiation (as similar results were obtained from lines 82 and 122, we used only line 82 in subsequent studies). To better address this issue, we compared the outcomes of in vitro Th2 cell differentiation from naive CD4 T cells isolated from Hlx transgenic mice and wild-type mice. Cytokine production was first analyzed by ELISA (Fig. 4 A). As expected, wild-type naive CD4 T cells differentiated into Th2 cells that produced large amounts of IL-4 and minimal amounts of IFN-γ as detected by ELISA. Cells differentiated from Hlx transgenic naive CD4 T cells produced slightly less IL-4 than their wild-type counterparts. However, these cells also produced large amounts of IFN-γ.
The coproduction of IL-4 and IFN-γ in the transgenic cells could be the result of either a mixture of Th1 and Th2 populations or the generation of IL-4, IFN-γ double-producing Th0 cells. To discern the cytokine profiles of the in vitro differentiated CD4 T cells, the same cells in Fig. 4,A were analyzed by intracellular cytokine staining (Fig. 4 B, upper panel). Similar percentages of IL-4 single-producing cells were detected in transgene-negative and -positive cells, whereas a higher percentage of Th1 cells was detected in Hlx transgenic CD4 T cells. In addition, a large percentage of transgenic CD4 T cells expressed both IFN-γ and IL-4. Therefore, Hlx transgene induced ectopic expression of IFN-γ, but did not block IL-4 expression in CD4 T cells differentiated under extreme Th2-polarizing conditions. We noted that the total percentage of IL-4-producing cells was higher in transgenic cells than in wild-type cells in the cytokine staining experiments, whereas the net amount of IL-4 in the supernatant of transgenic T cells as detected by ELISA was slightly lower than that in the supernatant of the wild-type cells. This discrepancy was perhaps due to the combination of autocrine inhibition of IL-4 expression by IFN-γ in the transgenic cells (41) and assumption of IL-4.
Differentiation of Hlx transgenic CD4 T cells under Th1-polarizing or Th0 conditions
In the KLH immunization experiments, we observed stronger Th1 responses in Hlx transgenic mice. To test whether such an enhanced Th1 response can be recapitulated in vitro, we performed in vitro Th1 cell differentiation studies using purified naive CD4 T cells. Intracellular cytokine staining of the differentiated cells showed that only IFN-γ-producing, but not IL-4-producing, cells were generated from both transgene-positive and -negative CD4 T cells, and no significant difference was observed between these two types of cells (Fig. 5,A). We also performed studies of differentiation of transgene-positive and -negative CD4 T cells under neutralizing or Th0 conditions (Fig. 5 B). Under such conditions, essentially only Th1 cells were generated in both types of cells, and their frequencies were similar, an outcome that is perhaps dictated by the genetic background of the cells (42). Importantly, the Th0-like cells observed in the transgenic CD4 T cells under Th2-polarizing conditions were not observed under Th0 conditions. Therefore, it appears that generation of Th0-like cells requires the presence of IL-4 in the initial culture.
Induction of IFN-γ by Hlx depends on its early expression during Th cell differentiation
So far our studies have demonstrated a positive effect of transgenic overexpression of Hlx on IFN-γ expression, most strikingly, ectopic expression of large amounts of IFN-γ under Th2-polarizing conditions. This result was seen in a transgenic line with only minimal abnormalities of thymic development. Therefore, it is unlikely that the effect on Th cell differentiation is consequently related to thymic abnormalities. Nonetheless, this possibility cannot be fully excluded using transgenic mice.
To better address this issue, we used retrovirus to overexpress Hlx in normal CD4 T cells undergoing Th2 cell differentiation (Fig. 6). We first followed the widely used protocol of retroviral infection of T cells. At 36 h after Th2 cell differentiation, CD4 T cells were infected with control retrovirus or retrovirus expressing Hlx. After further differentiation under Th2-polarizing conditions, the cells were rested, then restimulated for intracellular cytokine staining. As shown in Fig. 6,A, no significant changes in IFN-γ- or IL-4-producing cells were detected between cells infected with control and Hlx-expressing viruses. To confirm the expression of Hlx in the infected cells, GFP-positive cells were sorted. The level of Hlx expression in the sorted cells was analyzed by real-time RT-PCR and compared with the levels in normal Th1 cells and in Hlx transgenic CD4 T cells differentiated under Th2 conditions. Among these three types of cells, cells infected with Hlx-expressing viruses showed the highest level of Hlx (Fig. 6 B). Therefore, the lack of effect of Hlx in these cells was not due to insufficient expression of Hlx.
The lack of induction of IFN-γ by retroviral overexpression of Hlx was in sharp contrast to the strong expression of IFN-γ observed in Hlx transgenic CD4 T cells differentiated under Th2-polarizing conditions. We noticed that in the standard retroviral infection protocol, the expression of the virally transferred gene became significant around day 3 after initial stimulation of CD4 T cells, whereas significant up-regulation of Hlx during Th1 cell differentiation occurred between days 1 and 2 after initial T cell stimulation. Therefore, a delay of Hlx overexpression could be responsible for the lack of effect on IFN-γ expression by viral infection. To better match the natural kinetics of Hlx up-regulation, we modified the viral infection protocol to infect T cells with Hlx-expressing retrovirus at 12 h of Th2 differentiation. This allowed significant expression of viral Hlx on day 2 of in vitro differentiation, a time when initial up-regulation of Hlx was observed during Th1 cell differentiation. We then sorted the GFP-positive cells at this time and allowed them to further differentiate under Th2-polarizing conditions. Intracellular cytokine staining of the differentiated cells showed a significant increase in the Th1 population in cells infected with Hlx-expressing viruses. A similar increase in the Th0 population was observed in these cells (Fig. 6 C). In contrast, these two populations were minimal in cells infected with control viruses. Thus, the induction of IFN-γ expression by Hlx is dependent on its early expression during Th cell differentiation.
The differentiation of Th1 or Th2 cells is a dynamic process in which multiple genes are turned on or off at specific time points. We found that the homeobox protein Hlx is specifically expressed in Th1 cells at high levels. Its expression kinetics showed three phases: steady state expression in naive CD4 T cells, initial down-regulation during both Th1 and Th2 cell differentiation, and up-regulation upon commitment to the Th1 lineage. Recently, Mullen et al. (35) showed that Hlx enhanced T-bet-induced IFN-γ expression in vitro. In the present study we provided both in vitro and in vivo evidence for the positive regulation of IFN-γ expression by Hlx. Importantly, we showed that Hlx alone is sufficient to induce IFN-γ expression when over expressed at an early time point that matches the natural time course of Hlx up-regulation during Th1 cell differentiation. Unlike T-bet, Hlx did not appear to directly block IL-4 expression so that in addition to Th1 cells, Th0-like cells were induced by Hlx over expression in CD4 T cells differentiated under Th2-polarizing conditions.
Our studies demonstrated the importance of early expression of Hlx for inducing IFN-γ expression in CD4 T cells. In an earlier study, Mullen et al. (35) found T-bet in combination with Hlx induced more IFN-γ producing cells in differentiating Th2 cells than T-bet alone. However, Hlx alone did not induce IFN-γ expression. In our initial retroviral infection experiments we confirmed this finding. However, when the cells were infected earlier with Hlx-expressing retrovirus, and overexpression of Hlx matched the natural time course of Hlx up-regulation in Th1 cell differentiation, IFN-γ was induced in CD4 T cells undergoing Th2 cell differentiation. Thus, it appears that a window of time exists during early Th cell differentiation for Hlx to induce IFN-γ expression. Such a window of time is captured by early viral and transgenic overexpression of Hlx.
We believe that the early time window reflects the epigenetic state of the IFN-γ gene that is permissive to regulation by Hlx. As epigenetic modification is characteristic of Th cell differentiation, it is conceivable that repressive epigenetic modification at the IFN-γ gene may accumulate during Th2 cell differentiation to transform this locus to a transcriptionally hyper-repressed state that is nonpermissive to activation by Hlx. Such repressive modification could be reversed by T-bet, a possibility consistent with a role for T-bet in chromatin remodeling (24, 25). Thus, coexpression of T-bet in later developing Th2 cells allows Hlx to induce additional cells to express IFN-γ. Alternatively, a unique molecular context in early developing Th cells, which can also be induced by T-bet at later time points, is required for Hlx to induce IFN-γ expression. Future studies to characterize the differences in epigenetic modification of the IFN-γ locus and gene expression patterns between early and late developing Th2 cells will provide a better understanding of the biochemical mechanisms by which Hlx regulates IFN-γ expression.
In response to KLH immunization, Hlx transgenic mice generated greater numbers of Th1 cells than wild-type mice. Adoptive transfer experiments demonstrated that this was a result of the direct effect of the Hlx transgene on CD4 T cells. In contrast, no significant increase in IFN-γ-producing cells was observed in Th1 cell differentiation cultures of Hlx transgenic naive CD4 T cells in vitro. In the in vitro Th1 cell differentiation model, the culture conditions were extremely polarized. It was therefore likely that the frequencies of IFN-γ-producing cells were maximized, so that overexpression of Hlx could not further increase the frequencies. Unlike the in vitro differentiation cultures, conditions in the DLN in vivo were unlikely to be extremely polarized. In addition, the exact factors for triggering Th1 cell differentiation in the DLN are not clear. It is possible that a factor(s) other than IL-12, for example, IL-23, may contribute to the induction of Th1 cell differentiation. Under these circumstances, especially when such factors are also limited, Th1 cell differentiation may be suboptimal, so that overexpression of Hlx can induce more CD4 T cells to express IFN-γ and differentiate into Th1 cells.
Finally, in addition to the effect on Th cell differentiation, transgenic overexpression of Hlx caused abnormal thymic development, most notably, the reduction of the CD4, CD8 double-positive population. Although our retroviral studies using normal CD4 T cells clearly demonstrated a role for Hlx in inducing IFN-γ expression independent of its effects on thymic development, it cannot be fully excluded that in the Hlx transgenic mice the enhanced Th1 responses were the consequence of abnormal thymic development. We hope that our Hlx transgenic mice can serve as useful tools to dissect the elements of thymic development that can influence Th cell differentiation in the periphery. In addition, Hlx transgenic mice can be useful animal models for studies of immunological diseases in which heightened IFN-γ production by CD4 T cells is critical.
We thank Dr. Charles A. Janeway, Jr., and Grigoriy Losyev for the Abs used in the preparation of CD4 T cells and APC, Patricia Ranney for AND TCR transgenic mice, Drs. Kim Bottomly and Stephanie Constant for technical advice and reagents, Dr. Deb Fowell for anti-TCR Ab, Dr. Jerry Adams for Hlx cDNA, Dr. Dan Litterman for CD4 promoter, Dr. Steve Reiner for retroviral vectors, Dr. Garry Nolan for permission to use the Phoenix packaging cells, and Genetic Institute for IL-12.
This work was supported by a Cancer Research Institute postdoctoral fellowship and National Institutes of Health Grant AI47263 (to W.P.Z.). W.P.Z. and B.Y.L. were initially associates, and R.A.F. and D.G.S. are investigators with Howard Hughes Medical Institute.
Abbreviations used in this paper: KLH, keyhole limpet hemocyanin; DLN, draining lymph node; GFP, green fluorescence protein; HPRT, hypoxanthine phosphoribosyltransferase; RDA, representational difference analysis.