| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
and NKT Cells: Subset-Specific Differences in IL-2 Regulation1
* Division of Biology 156-29, California Institute of Technology, Pasadena, CA 91125; and
The Scripps Research Institute, La Jolla, CA 92037
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
cells, as well as 
TCR-NKT cells, exhibit higher in vivo transgene expression levels than TCR cells. This deviates from patterns of normal IL-2 expression and from expression of an IL-2-GFP knock-in. Peripheral TCR cells accumulate GFP RNA faster than endogenous IL-2 RNA upon stimulation, whereas TCR cells express more IL-2 than GFP RNA. In TCR cells, IL-2-producing cells are a subset of the GFP-expressing cells, whereas in TCR cells, endogenous IL-2 is more likely to be expressed without GFP. These results are seen in multiple independent transgenic lines and thus reflect functional properties of the transgene sequences, rather than copy number or integration site effects. The high ratio of GFP: endogenous IL-2 gene expression in transgenic TCR cells may be explained by subset-specific IL-2 gene regulatory elements mapping outside of the 8.4-kb transgene regulatory sequence, as well as accelerated kinetics of endogenous IL-2 RNA degradation in TCR cells. The high levels and percentages of transgene expression in thymic and splenic TCR and NKT cells, as well as skin TCR-dendritic epidermal T cells, indicate that the IL-2-GFP-transgenic mice may provide valuable tracers for detecting developmental and activation events in these lineages. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
cells. All of the general principles of IL-2 induction were defined, in the 1990s, by exploitation of TCR thymoma lines and T cell clones, including the roles of NF-AT, NF-
B/Rel, AP-1, and Oct-1 as activators of the minimal IL-2 enhancer (1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11). Recent studies have provided evidence that chromatin remodeling events in and adjacent to the proximal promoter may be important for IL-2 transcription as well (12, 13, 14). In addition, several DNase I-hypersensitive regions have been identified which lie upstream of the proximal promoter region, and inclusion of 8.4-kb of sequence, including most of these sites, dramatically improves inducible reporter transgene expression over use of the 2-kb proximal promoter region alone (15). This result suggests the existence of additional regulatory elements, which, at minimum, moderate chromatin accessibility in vivo.
IL-2 is essentially T cell specific and, unlike most other cytokines, can be expressed by TCR as well as CD4+ and CD8+ T cell subsets; it is inducible even in immature thymocytes and is the major cytokine induced in the primary response of naive T cells (16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29). Although evidence suggests extensive sharing of key transcription factors (8, 17, 18), it is not yet clear whether IL-2 regulation in all cases is based on common molecular mechanisms that are accessible to all T cells, including important minority populations, such as TCR and NKT cells.
The set of transgenic (Tg)3 mouse lines, containing a green fluorescent protein (GFP) reporter gene controlled by 8.4 kb of the 5' flanking sequence from the murine IL-2 gene (IL2p8-GFP), which we derived to study IL-2 gene regulation, provides a new approach to compare different T cell subsets (15). As transgenics rather than GFP insertions into the intact IL-2 locus, these animals enhance detectability of IL-2 expression by three "artifacts" that prove useful for certain kinds of studies. First, T cell activation induces very high levels of GFP fluorescence over a wide dynamic range. Second, because the GFP mRNA and protein are more stable than native IL-2 mRNA and protein, GFP fluorescence can be used to track even rare cells that have recently undergone activation as well as cells caught in the middle of a transient activation response. These are both advantages for investigations of developmental contexts in which IL-2 is induced in response to poorly characterized stimuli, which cannot be readily synchronized or duplicated in vitro. Third, the transgene includes only a defined subset of the potential regulatory sequences that could control IL-2 expression in its native chromatin context. As we show in this article, these elements of artificiality in the Tg construct appear to render it a highly sensitive indicator of previously undisclosed differences among the mechanisms of IL-2 regulation in different T cell subsets, as well as providing a potentially valuable tracer for TCR and NKT cells.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
C57BL/6 (B6) mice were originally obtained from The Jackson Laboratory (Bar Harbor, ME) and bred and housed in specific pathogen-free conditions our mouse facility at Caltech. IL2p8-GFP Tg lines were previously generated, PCR typed, and characterized as described previously (15). The Tg lines used in this study are as follows (copy numbers given in parentheses): Tg4 (3 copies), Tg8 (4 copies), Tg186 (12 copies), Tg12 (16 copies), Tg170 (18 copies), Tg175 (23 copies), Tg177 (30 copies), and Tg17 (33 copies). Tg mice were bred and housed at the Caltech Transgenic Animal Facility and maintained as hemizygotes by backcrossing to B6 mice. Mice with the Gfp gene knocked into the Il2 locus (B6.IL2-GFPki-Rag2−/−) were generously provided by Dr. H. Gu (National Institute of Allergy and Infectious Diseases, Bethesda, MD). These mice were crossed to B6 mice to generate mice with normal T cell subsets as well as one GFP knock-in/IL2 knock-out allele and one wild-type IL-2 allele (IL-2-GFPki/wt).
Cell culture and stimulations
Splenocytes and purified T cells were isolated from individual animals and cultured as described previously (15). Cells were cultured in DMEM supplemented with 10% FBS, 2 mM L-glutamine, 5 x 10−5 M 2-ME, 100 U/ml penicillin, and 100 µg/ml streptomycin (Invitrogen, Gaithersburg, MD). For stimulations, cells were either added to 96-well flat-bottom plates (Corning, Corning, NY) coated with 10 µg/ml purified Abs to CD3 and CD28 (BD PharMingen, San Diego, CA) or incubated with 175 nM calcium ionophore, A23187, and 10 ng/ml PMA (Sigma-Aldrich, St. Louis, MO). For mRNA stability studies, 250 ng/ml cyclosporin A (Novartis Pharmaceuticals, East Hanover, NJ) was added to cells 3 h after stimulation.
Flow cytometric staining and cell sorting
Single-cell suspensions of fresh or cultured splenocytes or thymocytes from individual mice were stained as described previously (15). The following surface-staining reagents were used: TCR-allophycocyanin, NK1.1-PE (BD PharMingen), and TCR-PE-Cy5 (Accurate Chemicals, Westbury, NY). Anti-IL2-allophycocyanin and the BD Cytofix/Cytoperm kit with GolgiPlug (brefeldin A; BD Biosciences, Mountain View, CA) were used for intracellular IL-2 staining. FACS analyses were performed using a BD FACSCalibur. Cells were sorted using a BD FACSVantage Cell Sorter.
RNA purification and real-time quantitative RT-PCR (Q-PCR)
Total RNA was extracted from cells after various treatments using RNAzol (Leedo Medical, Houston, TX) following the manufacturer’s instructions. RNA was treated with RNase-free DNase to remove residual genomic DNA. First-strand cDNA synthesis reactions were then conducted using Superscript Reverse Transcriptase II (Invitrogen) and 250 ng random hexamers (Pharmacia, Uppsala, Sweden) following standard protocols.
Real-time fluorescent PCR (Q-PCR) analysis was conducted using the ABI Prism 7700 Sequence Detection System (PE Applied Biosystems, Foster City, CA) as previously described (15). A GFP-specific oligonucleotide probe, 6-FAM-GCTTTACTTGTACAGCTCGTCCATGCCGA-TAMRA, and transgene-specific primers (5'-CACATGGTCCTGCTGGAGTTC-3') and (5'-CAGCACACAGACCAGCACGTT-3') were used as well as TaqMan probes and primers for murine GAPDH (VIC labeled) and IL-2 (FAM labeled; PE Applied Biosystems). Thirty-microliter reactions were conducted in TaqMan Universal PCR buffer using the following thermocycling conditions: 50°C for 2 min, 95°C for 10 min, then 40 cycles of 95°C for 15 s and 60°C for 1 min. Control GAPDH cycle threshold (CT) values were subtracted from GFP and IL-2 CT values for normalization of each sample. Relative mRNA levels were calculated as 2−
CT and values for TCR and NKT cells were adjusted relative to the TCR sample from the same animal. Additionally, ratios of GFP:IL-2 RNA, in the same samples, were also calculated and adjusted relative to the TCR sample from the same animal, controlling for normalization artifacts.
Ear dendritic epidermal T cell and epidermal sheet isolation and immunofluorescence
DETC and epidermal sheets were isolated from the skin of 10-wk-old Tg 17 mice as previously described (30, 31). Epidermal cells were stained with Abs to CD3, TCR, and I-Ab (BD PharMingen). FACS analysis was performed using a BD FACSCalibur and analyzed using BD Cell Quest software. Unfixed epidermal sheets were stained with PE conjugates of Abs to TCR or I-Ab (BD PharMingen). Samples were mounted with PBS and imaged within 2 h of staining using a Zeiss Axiovert 100TV microscope (Zeiss, Oberkochen, Germany). Digital images were collected using a SPOT (Diagnostics Instruments, Sterling Heights, MI) camera and software.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
vs TCR cellsWe previously reported that an extended 5' flanking region of the IL-2 gene, up to 8.4 kb upstream of the transcriptional initiation site, could drive consistent expression of a GFP reporter transgene (IL2p8-GFP) in activated T cells in nearly all Tg mouse founder lines (15). This represented a substantial increase in efficiency of expression over Tg constructs with only 2 kb of 5' flanking region, even though there was no obvious increase in expression level in transient transfection assays. The results thus suggested that some elements of locus control-like function could map between 2 and 8.4 kb upstream of the IL-2 promoter. To determine whether this region actually includes all elements needed for IL-2 regulation, we investigated the expression of the transgene in diverse subsets of T cells in more detail.
In addition to the spontaneous expression of the 8.4-kb Tg construct in splenic TCR cells reported previously (15), surprisingly high levels of GFP were observed in splenic TCR and NKT cells from the same mice. Fig. 1A shows typical patterns of GFP expression in freshly isolated unstimulated splenic T cell subsets from two Tg lines, low copy Tg4 (3 copies) and high copy Tg17 (33 copies), distinguished by their expression of TCR, TCR, and NK1.1. Whereas <10% of the conventional TCR cells (TCR+NK1.1−) spontaneously express detectable GFP in most of the Tg lines, 40–60% of TCR+ cells do so (Fig. 1A). GFP expression is low compared with the level of expression that can be reached in acutely stimulated cells (MFI 
50–200 vs MFI of 
1000), but it is clearly distinguishable over background. There is also significant expression of GFP in an elevated fraction of NKT cells (TCR+NK1.1+) in the same animals. The TCR+ population also typically includes 20% NK1.1+ cells, and these are similar to other TCR cells in GFP expression percentages and levels (data not shown). Spontaneous transgene expression in TCR and NKT cells in vivo has been seen in every 8.4-kb Tg animal analyzed in at least four additional, independent comparisons including cells from >20 individual mice representing eight different IL2p8-GFP Tg founder lines.
|
as compared with TCR cells, splenic GFP-negative TCR and TCR cells were purified by flow cytometry and stimulated in vitro. Fig. 1B shows representative results with two Tg lines, one low copy number line, Tg4 (3 copies), and one high copy line, Tg175 (23 copies). A substantial fraction of the TCR cells strongly activated expression of the transgene in response to stimulation overnight either by CD3/CD28 cross-linking or by addition of PMA/A23187 (Ca2+ ionophore). However, in each case the fraction of TCR cells responding was higher and the intensity of expression, measured by mean fluorescent intensity of the GFP signal, was also generally higher. TCR cells gave stronger transgene expression than TCR cells from the same mice in repeated experiments analyzing Tg mice from another four different founder lines. These results show that not only are the TCR cells from the Tg mice more likely to express GFP than TCR cells in vivo, but this difference is a reflection of greater transgene inducibility in TCR cells in response to stimulation. Subset-specific derepression of the IL-2-GFP transgenes
Transgene expression could reflect a previously unappreciated level of IL-2 gene expression in NKT and TCR cells, higher than in conventional TCR cells, or it could reflect a deviation of transgene expression from the endogenous IL-2 expression levels in these cell types. To distinguish between these possibilities, we isolated conventional TCR, NKT, and TCR cells from four different Tg lines and non-Tg controls and quantified the levels of IL-2 and GFP RNA expression induced in these populations in response to short-term stimulation. RNA analysis was used to ensure that the measures of transgene expression were strictly comparable with measures of endogenous gene expression, and an early time point of 5 h was used to minimize effects of any differences in RNA half-lives between the GFP Tg RNA and the IL-2 endogenous RNA (see below).
The results, normalized to the level of RNA induced in the TCR conventional population are shown in Fig. 2. Stimulated TCR cells expressed approximately two to five times less IL-2 RNA than the stimulated TCR cells from the same animals, with or without the transgene (Fig. 2A). As a rule, TCR cells appear to limit their endogenous IL-2 expression more severely than TCR cells. However, in the same samples, the levels of GFP RNA were higher in the TCR cells than in the TCR cells in all six Tg animals analyzed (Fig. 2B). When relative levels of GFP to IL-2 in each sample were calculated, minimizing any effects due to normalization, relative differences in expression of the two genes in the two cell populations were apparent (Fig. 2C; note the log scale for this panel). Thus, under conditions where IL-2 is expressed better than GFP in TCR cells, GFP is expressed better than IL-2 in TCR cells from the same Tg animal.
|
T cells. Again, the levels of endogenous IL-2 RNA expression in stimulated NKT cells are usually low compared with those in similarly stimulated conventional TCR cells, although there is more mouse-to-mouse variability than in the case of the TCR cells (Fig. 2A). GFP RNA levels in the NKT cells, however, are similar to or greater than those in the TCR cells from the same mice (Fig. 2B). The ratios of GFP:IL-2 RNA, shown in Fig. 2C, also show that NKT cells have similar or higher ratios than TCR cells. There may also be an effect of transgene copy number, with the low copy number animals showing higher ratios in NKT cells relative to TCR cells and the higher copy animals showing similar ratios (Tg4, 3 copies; Tg186, 12 copies vs Tg175, 23 copies; Tg17, 33 copies). The preferential expression of transgenes relative to the endogenous genes may thus separate NKT, as well as TCR cells, from conventional TCR cells.
Coexpression of Tg and endogenous loci at higher frequencies in TCR than in TCR cells
Under limiting conditions of stimulation, IL-2 may be expressed from only one allele at a time (32, 33), but the mechanism governing the separate activation thresholds for the two alleles is not well defined. Transgenes are expected to be integrated at one site in the DNA as a tandem concatemer, and this organization may be helpful to distinguish whether the crucial threshold-setting mechanisms act at the level of individual promoter/enhancers or at the level of a more extended chromosomal region. In initial staining experiments to determine whether endogenous IL-2 is coexpressed with the transgene, we noted that the majority of individual cells in unfractionated, Tg T cell populations expressed endogenous IL-2 protein or the GFP transgene, but not both (Fig. 3 and data not shown). This observation, plus the fact that the total percentage of GFP+ cells and the percentage of those GFP+ cells that are also expressing IL-2 do not differ significantly between the low copy line (Tg4, 3 copies) and the high copy line (Tg17, 33 copies), could be interpreted as evidence that all of the transgenes are handled by most TCR cells as a single "locus," comparable to a single endogenous allele. As shown below, the threshold-setting mechanism appears to be applied differently in TCR cells than in TCR cells.
|
T cells from controls and Tgs alike could be shown to contain IL-2 protein, and 13–16% were positive for GFP. Much of the IL-2 detected at this time point is induced during the restimulation, since the percentage positive is >5-fold higher than could be detected without acute restimulation. In contrast, the percentage of cells expressing GFP is the same with or without restimulation (Fig. 3).
As shown in Fig. 3, some activated TCR cells express IL-2 without detectable GFP, some express GFP without detectable IL-2, some express both, and some express neither. The percentage of double-expressing cells scored depends on the efficiency with which each gene is detected, but under the double stimulation condition, it is more than half of the GFP+ cells. In most TCR samples, however, with or without restimulation, IL-2 is expressed by approximately three to four times as many GFP− cells as GFP+ cells. This suggests that many cells can activate an endogenous IL-2 gene without having activated expression of the transgenes.
The pattern of IL-2 and GFP expression is markedly different in the TCR-enriched populations than in TCR populations from the same mice under the same conditions of stimulation. Fig. 4 presents a summary of FACS analyses of IL-2 and GFP expression in TCR- and TCR-enriched cell populations from seven different IL2p8-GFP Tg lines and a non-Tg animal. Under the conditions of the assay shown, the most common responder type in the TCR populations of all Tg lines was a cell expressing IL-2 alone without detectable GFP (Fig. 4A). The percentage of GFP- and IL-2-expressing cells is comparable to the number expressing GFP alone, both groups being much lower in all but one case than IL-2 alone. Restimulation conditions alone, 2-h stimulation with PMA/A23187, were not adequate to induce many TCR cells to express IL-2 over the background seen in unstimulated cells (Fig. 4, C vs E). Detection of IL-2 is thus dependent upon both primary stimulation (16 h of anti-CD3/CD28) plus restimulation (2 h of PMA/A23187) during the time of brefeldin A treatment and IL-2 accumulation.
|
cells, TCR-enriched cells from six of seven of the IL2p8-GFP Tg lines assayed were more likely to express GFP alone than to express IL-2 alone, and in five of seven, they were more likely to express IL-2 along with GFP than to express IL-2 alone (Fig. 4B). Some of the GFP expression in TCR-enriched cells is due to in vivo background expression (Fig. 4D and below), but this does not account quantitatively for the results seen. These results imply that the TCR-enriched cells that are competent to express IL-2 on stimulation are a subset of those competent to express GFP under the same conditions. Moreover, in contrast to the TCR cells, restimulation conditions alone were sufficient to result in percentages of GFP+TCR cells comparable in many cases to those induced by primary anti-CD3/28 stimulation plus restimulation (Fig. 4, B vs F).
Potential contributory factors for subset regulatory differences: IL-2 mRNA instability in TCR cells
Several mechanisms could contribute to the subset-specific differences between TCR and TCR cells in their regulation of the IL-2-GFP transgenes as compared with endogenous loci. One could be RNA stability. The GFP transcript from the transgene is designed to lack the AU-rich sequences that target endogenous IL-2 RNA for rapid degradation by using the -globin 3' splice and poly(A) addition sites. This makes the GFP mRNA more stable than IL-2 mRNA, as shown by real-time RT-PCR quantitation of the RNA levels after a 3-hour initial stimulation followed by a cyclosporin A chase (Fig. 5A). As shown in Fig. 5A, both GFP and IL-2 RNA are sharply induced in the initial 3-h stimulation period, but while the IL-2 RNA decreases rapidly over the subsequent 3 h of chase, GFP RNA decays more slowly. Thus, any subset-specific mechanism that exaggerates the difference in mRNA stability between IL-2 and GFP could appear to enhance transgene expression relative to endogenous IL-2. In fact, the half-life of IL-2 mRNA in splenic TCR cells in normal mice is even shorter than in TCR cells, as shown in Fig. 5B. Thus, the ongoing rate of IL-2 transcription in stimulated TCR cells is likely to be underestimated by their steady-state IL-2 RNA accumulation, relative to that in TCR cells.
|
cells, as measured after prolonged stimulation. However, other mechanisms are likely to be responsible for the difference seen at early time points of stimulation, e.g., as shown in Figs. 2 and 4, E and F. Furthermore, as described below, the Tg GFP reporter is expressed differently from a similar GFP reporter that has been knocked into the endogenous IL-2 locus. Thus, the turnover kinetics of the reporter transcript may enhance, but cannot fully explain, its differential expression in TCR vs TCR cells.
Preferential activation of IL-2-GFP transgene expression in TCR and NKT cells during thymic ontogeny
The trigger for transgene expression in peripheral TCR and NKT cells has not been determined rigorously. Because these Tg mice are not maintained under strictly aseptic conditions, a fraction of the observed response could have been triggered by stochastic encounters with low-level pathogens. However, both TCR and NKT cells also exhibit strong transgene expression in the thymus (Fig. 6A), where expression is more likely due to developmental programming or normal microenvironmental signals.
|
+ thymocytes with phenotypes that indicate they are undergoing positive selection to the CD4 single-positive (SP) and CD8 SP cell types (15). However, the percentage of TCR and NKT thymocytes expressing the transgene in the same mice is as much as 20–40 times higher (Fig. 6A). Analysis of transgene expression was done using TCR, TCR, and NK1.1 to subdivide TCR+ thymocytes according to presumptive lineage. In TCRhigh cells that do not express NK1.1 (primarily conventional CD4 and CD8 SP cells), <1% of the cells express the transgene (Fig. 6A, left panels). This agrees well with our previous findings. However, the percentage of TCR+ thymocytes that express GFP is >40%, and the percentage of TCR+ NK1.1+ cells expressing GFP is comparable (Fig. 6A, middle and right panels). The mean fluorescence intensities of the three groups of spontaneously expressing cells are similar, with the level depending on the Tg line. This elevated expression of GFP in thymic TCR and NKT cells has been seen in all eight founder Tg lines analyzed.
To assess the developmental stages of the cells expressing GFP in the thymus, surface heat-stable Ag (HSA; CD24) was used as a developmental marker for TCR and TCR cells. Shown in the left panels of Fig. 6B is the relationship between TCR and HSA levels in thymocytes. High levels of HSA are expressed both on immature double-positive (DP) thymocytes, which have intermediate amounts of surface TCR (Fig. 6B left, lower right quadrant), as well as on T cells that have recently undergone positive selection and have high TCR levels (Fig. 6B left, upper right quadrant). HSA levels then decline after positive selection and with continuing thymic maturation (Fig. 6B left, upper left quadrant). Stained thymocytes, from several different Tg lines, show the presence of both HSA+ and HSA− cells among GFP+-gated cells (Fig. 6B, middle panels). The HSA+TCR+ cells expressing GFP (Fig. 6B middle, upper right quadrants) have relatively high levels of TCR and appear to be DP or early SP cells, which may have up-regulated GFP in response to selective signals, in agreement with our previous report (15). The GFP+HSA− cells, on the other hand, are intermediate for levels of surface TCR (Fig. 6B middle, upper left quadrants) and represent the NKT cells, as shown by separate staining experiments (data not shown). The GFP+TCR− cells (Fig. 6B middle, lower right quadrants) are predominantly TCR cells. Although few maturation markers are available for these cells, HSA is also a marker for immature TCR thymocytes (34). Fig. 6B (right panels) shows the HSA and GFP levels of gated TCR thymocytes from the same Tg and non-Tg animals. Not only do the majority of thymic TCR cells still express HSA, but a disproportionate percentage of the GFP+ cells are HSA+. This result suggests that a high percentage of thymic TCR cells are induced to express the GFP transgene while they are still immature, as an outcome of normal developmental events in the thymus rather than due to exogenous Ag exposure.
Subset-biased developmental expression is unique to IL-2-GFP transgenes and not detected using GFP in the endogenous IL-2 locus
Previous data have indicated some IL-2 expression in thymic TCR cells (20), but have not determined its frequency nor its occurrence in the NKT lineage. To resolve whether the transgene is reflecting the expression status of the endogenous locus in these contexts, mice with a single allele of GFP knocked into the IL-2 locus (IL-2-GFPki/wt) (35) were compared with the IL2p8-GFP Tg animals. In contrast with Tg lines, thymocytes from IL-2-GFPki/wt mice do not express detectable GFP spontaneously in TCR and NKT cells, while GFP was clearly detectable in TCR thymic cells, although the percentages and fluorescence intensities were approximately 10 times lower than those seen in the Tg cells (Fig. 6A, fourth row). Furthermore, TCR and NKT cells from IL-2-GFPki/wt spleens express little if any GFP, while GFP is spontaneously detected in TCR spleen cells from the same animals (data not shown). Thus, the knock-in GFP gene, having all of the regulatory elements of the endogenous IL-2 gene, does not exhibit the spontaneous preferential expression in TCR or NKT cells found in the IL2p8-GFP Tg lines.
Fetal thymus-derived skin TCR cells express the IL2p8 transgene
Fetal thymic-derived DETC express a canonical V3V1 TCR and are developmentally distinct from adult thymic TCR cells (36, 37). Additionally, these cells have to been shown to be able to express abundant IL-2 protein and mRNA, especially in the first month of postnatal life (20). To determine whether GFP is expressed in skin, epidermal sheets were prepared, stained with Abs to TCR to detect DETC or I-Ab to detect dendritic Langerhans cells, and visualized using fluorescent microscopy. TCR+ DETC were found to coexpress GFP, while I-Ab+ Langerhans cells did not (Fig. 7A). GFP was primarily expressed in CD3+ cells, based on FACS analysis of isolated skin cells (Fig. 7B). Furthermore, these isolated GFP+ cells are positive for surface TCR but negative for I-Ab (data not shown). Two GFP+ populations were observed (Fig. 7B). The cells expressing intermediate levels of GFP are also intermediate for TCR levels and high for CD69 (data not shown), indicating that they may be recently activated cells. The GFPhigh cells have high levels of TCR and intermediate levels of CD69, suggesting that they may be previously activated cells. Some skin-infiltrating TCR+GFP+ cells were found but represent a very minor population (<0.4%) of total skin cells (data not shown). Thus, the GFP transgene is also readily induced in a developmentally distinct and fetal thymus-derived T cell population, suggesting a common regulatory mechanism in TCR cells as a whole.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
and TCR classes of T cells and shared between CD4 and CD8 SP T cells. All T cells require stimulatory signals to turn on IL-2 and these are mediated, largely or entirely, by a shared complement of signal-dependent transcription factors (2, 3, 4, 8, 17, 18). Superimposed on this activation dependence is likely to be a developmentally controlled mechanism of locus accessibility, and it would be reasonable to expect this to be a single mechanism inherited by all T cell subsets from a common T lineage precursor. In this study, however, we have presented data indicating that different T cell populations utilize different mechanisms in their control of the magnitude, and even the probability, of IL-2 expression. The results indicate that IL-2 gene regulation is conducted by mechanisms that are more cell-type specific than had been suspected previously.
The evidence presented shows that the same transgenes, in the same integration sites, are handled differently by TCR and NKT cells than by conventional TCR cells in the same mouse. The differences do not depend on an assumption that the stimuli used are identically efficient for the different classes of T cells, because they are seen even when the responses of the subsets are normalized to corresponding levels of endogenous IL-2 induction. Furthermore, high percentages of TCR and NKT cells were found to express the transgene in vivo. The question is what kinds of mechanisms could cause the differences in IL2p8-GFP transgene usage between subsets. The differences that we see between these subsets in their expression of the IL2p8-GFP transgenes could be influenced by signal-dependent mechanisms acting at transcriptional or posttranscriptional levels, as well as longer term mechanisms involving transcription factor pools and chromatin structure established during divergence of TCR, TCR, and NKT cell precursors.
Several potential levels of transcriptional regulation are likely to impinge on the IL-2 gene and the IL-2 regulatory sequence transgenes. First is the mechanism that actually turns on IL-2 transcription, once the chromatin is open, a mechanism that is explained in terms of the activation and nuclear concentrations of NF-AT, NF-B/Rel, AP-1, and Oct family transcription factors acting on the proximal promoter. This mechanism is not only essential for expression but should also control the rate of transcription from the gene once activated. Another, less understood mechanism apparently sets an all-or-none threshold for letting the gene be transcribed from a particular allele (32, 33), possibly by demanding a certain quorum of positive transcription factors ("and" logic) and/or requiring the removal of inhibitory transcription factors or inhibitory chromatin modifications. There may be a third mechanism that makes IL-2 eligible, at all, for activation in a given cell type. This basally permissive state is only known to be established in T cells and several other hemopoietic cell types, B cells (8), mast cells (38, 39), and eosinophils (40, 41). In principle, all three of these mechanisms could act on the IL-2 sequence transgenes as well as the endogenous genes. Cell type-specific alterations in any of these mechanisms could contribute to the differences between transgene expression in the TCR and TCR cells.
Posttranscriptional differences provide another potential mechanism by which cell type-specific differences in gene regulation might occur. The control of IL-2 RNA degradation and translation in TCR cells is complex, depending on precise stimulation conditions and on sequences in the coding and 3' untranslated regions of the gene (42, 43, 44, 45, 46). We show initial evidence that one difference between TCR and TCR cells may affect the posttranscriptional stability of IL-2 transcripts expressed in these cells. Much remains to be learned about the stimulation conditions that are optimal for IL-2 expression in TCR cells, but at least under conventional conditions of CD3/CD28 stimulation, splenic TCR cells degrade their IL-2 mRNA faster than TCR cells. This implies that measured levels of IL-2 RNA accumulation are likely to underestimate the actual rate of transcription of endogenous IL-2 genes by the TCR population under these conditions. Thus, one factor contributing to make the ratio of transgene:endogenous gene expression higher in the TCR cells may be the greater proportional impact of expressing a stable GFP transcript instead of an unstable IL-2 mRNA. It will be of great interest to determine which of the components of these posttranslational mechanisms are differentially applied in TCR and TCR cells.
Three pieces of evidence suggest that the differences in transgene expression in TCR vs TCR cells cannot be explained solely in terms of posttranscriptional mechanisms, however. One is the fact that transgene expression in TCR cells initiates more readily than in TCR cells, even in the first few hours of stimulation when degradation kinetics have a negligible impact (Fig. 4). The other is the frequency of expression of GFP in individual TCR vs TCR cells. Even if levels of detectable IL-2 RNA and protein accumulation in TCR cells are reduced disproportionately by RNA degradation, this does not explain why the transgenes are expressed by so much larger a fraction of the population of TCR cells than of TCR cells (Fig. 4). The transgenes must also be more readily induced at a transcriptional level in TCR than in TCR cells, or the TCR cells must have a mechanism for repressing transgene expression that does not apply in TCR cells. Finally, mice with the GFP gene knocked into the IL-2 locus, using an SV40 intron and polyadenylation signal sequence but otherwise retaining all of the upstream, and possibly downstream, IL-2 regulatory elements, do not express GFP spontaneously in thymic or splenic TCR or NKT cells, in sharp contrast to the Tg mice. This suggests that differences in the IL-2 regulatory sequences outside both the 8.4-kb promoter sequences and the coding or 3' untranslated regions are critical for the differences in expression between the two types of GFP-expressing animals.
Our staining data on the coexpression of endogenous IL-2 with the IL-2-GFP transgenes link the difference between TCR and TCR cells with the second mechanism, the one that sets thresholds for opening one or another IL-2 locus in an otherwise competent cell. Recent data indicate that this threshold is partially controlled by the methylation of critical sites in the first <1 kb of IL-2 5' flanking sequence (14), but roles for more distant regulatory sites have not been ruled out. The most likely explanation for the lowered transgene expression threshold in TCR cells is that the transgene is missing some negative regulatory sequence, either upstream of −8.4 kb or downstream of the translational start site, which normally sets activation thresholds for IL-2 expression, specifically in splenic TCR cells and perhaps in NKT cells as well. This missing sequence may also act as an activator for TCR cells.
This interpretation is made tenable by three lines of evidence. First, our previous DNase I hypersensitivity data showed the presence of distinct sites at approximately −8 and −10 upstream of the IL-2 gene start site, one or both of which lies outside of the 8.4-kb IL-2 sequence used in the transgenes (15). Second, analysis of additional Tg lines supports the idea that such subset-specific negative regulatory sequences can exist. For example, deletion of a specific 800-bp sequence within the 8.4-kb upstream region seems to be capable of further derepressing transgene expression in thymic NKT cells, without enhancing expression in any other subsets tested (M.A.Y. and E.V.R., unpublished data). The expression in thymic NKT cells is particularly notable since this subset of thymocytes, which is known for its high-level IL-4 expression and general state of activation, has not previously been noted to express IL-2. Finally, further evidence comes from comparative sequence analysis (47), which has recently become possible with the release of the human, murine, and rat (draft) genomic sequences. This approach indicates several regions of unusually high conservation among the murine, human, and rat genomes, beyond the regulatory region included in the 8.4-kb transgenes. Work is currently in progress to map these regions and to characterize them functionally (S. Adachi, M. A. Yui, and E. V. Rothenberg., unpublished data).
The difference in transgene utilization by TCR and TCR cells could also be attributable to an alternative mechanism. If transcription of the transgene in the TCR cells is "normal" (but detected with heightened sensitivity due to the stability of GFP RNA), then the focus would shift to the lower ratio of transgene:endogenous IL-2 gene expression seen in conventional TCR cells. Multicopy transgenes are known to integrate as concatenates, and these could be subject to an additional transgene-specific effect: namely, the silencing of repeated gene arrays (48). In this case, the TCR cells may be more likely to silence the transgene arrays completely, either due to a more stringent repeat-induced silencing mechanism or to a requirement for additional positive regulatory sequences that are not included in the transgene. The higher ratios of GFP:IL-2 mRNA observed in the lower transgene copy lines make this hypothesis less likely but it cannot currently be ruled out.
In either case, the data we have presented here indicate that the new, subset-specific function affects the probability that the locus, transgene or endogenous, will be transcribed at all in a cell, during a given response to activation, rather than offering amplification over a continuous range of expression levels. Determining whether one or both of the two models are operating on the IL-2 gene and transgene will require further investigation. These results demonstrate that the mechanisms of IL-2 transcriptional regulation are more complex and developmentally nuanced than has been previously recognized.
In addition to revealing T cell subset differences in IL-2 transcriptional regulation, an important implication of this work is that the IL2p8-GFP Tgs may provide a novel tool for tracking T cell subsets, however rare, undergoing developmental and immunological responses. The 8.4-kb IL-2-GFP transgene cannot only be used to track activated TCR cells, but can also preferentially track the activation-induced responses of relatively rare peripheral TCR cells, skin TCR DETC, and NKT cells. These classes of T cells are the focus of increasing interest due to their roles in innate immunity, yet their rarity and idiosyncratic responses to stimulation in vitro make them hard to study. Because the target Ags are mostly unknown for all of these T cell subsets, it is difficult to determine to what extent particular responses are due to innate gene expression programming, as opposed to normal selection or activation events. The in vivo expression of the transgene in skin DETC further emphasizes the high level of expression and detectability of the transgenes in a developmentally distinct and diffusely distributed T cell subset. These cells express a canonical V3V1+ TCR and are only produced during the first wave of fetal thymic rearrangement, beginning at day 14 of embryogenesis (49). After exiting the thymus during fetal development, DETC reside in murine skin where they recognize an uncharacterized keratinocyte Ag and have roles in immune surveillance and homeostasis (50). These cells are known to be capable of expressing abundant IL-2 protein and mRNA during development (20), and the GFP+ cells appeared to be undergoing activation, as CD69 was coexpressed in many of the GFP+ cells. However, here as in other TCR cell lineages, the transgenes could be a more sensitive indicator of activation than endogenous IL-2 expression, since high levels of GFP expression were detected here in postweaning animals at an age when the endogenous expression would have been greatly diminished (20).
Finally, the high level of GFP expression during the development of immature and minority thymic T cells, as well as NKT cells, is intriguing and may prove to be a useful developmental marker. If GFP is only expressed, however aberrantly, upon activation (possibly via TCR), this result suggests that the transgene detects normal activating interactions in the thymus, possibly akin to positive selection. NKT cells appear to undergo thymic selection based upon TCR-CD1d interactions (reviewed in Refs. 51 and 52), but T cell developmental signaling events have proven to be more elusive. Thus, the inducible, profuse, and subset-specifically enhanced expression of the GFP transgene may prove to be a useful probe for the detection of activating events occurring in the thymus during normal development, as well as during peripheral immunological responses.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Ellen V. Rothenberg, Division of Biology, 156-29, California Institute of Technology, Pasadena, CA 91125. E-mail address: evroth{at}its.caltech.edu
3 Abbreviations used in this paper: Tg, transgenic; GFP, green fluorescent protein; Q-PCR, quantitative PCR; CT, cycle threshold; DETC, dendritic epidermal T cell; SP, single positive; HSA, heat-stable Ag; DP, double positive.
Received for publication July 23, 2003. Accepted for publication January 29, 2004.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
B and AP-1 trans-activation in response to minimal TCR engagement by a partial agonist in naive CD8 T cell agonist in naive CD8 T cells. J. Immunol. 163:5219.
), GFP (
), or both (
) and percentages for each group are given for each animal.
