Abstract
Generation of allelic gene reporter mice has provided a powerful tool to study gene function in vivo. In conjunction with imaging technologies, reporter mouse models facilitate studies of cell lineage tracing, live cell imaging, and gene expression in the context of diseases. Although there are several advantages to using reporter mice, caution is important to ensure the fidelity of the reporter protein representing the gene of interest. In this study, we compared the efficiency of two Il9 reporter strains Il9citrine and Il9GFP in representing IL-9-producing CD4+ TH9 cells. Although both alleles show high specificity in IL-9–expressing populations, we observed that the Il9GFP allele visualized a much larger proportion of the IL-9–producing cells in culture than the Il9citrine reporter allele. In defining the mechanistic basis for these differences, chromatin immunoprecipitation and chromatin accessibility assay showed that the Il9citrine allele was transcriptionally less active in TH9 cells compared with the wild-type allele. The Il9citrine allele also only captured a fraction of IL-9–expressing bone marrow–derived mast cells. In contrast, the Il9citrine reporter detected Il9 expression in type 2 innate lymphoid cells at a greater percentage than could be identified by IL-9 intracellular cytokine staining. Taken together, our findings demonstrate that the accuracy of IL-9 reporter mouse models may vary with the cell type being examined. These studies demonstrate the importance of choosing appropriate reporter mouse models that are optimal for detecting the cell type of interest as well as the accuracy of conclusions.
Introduction
Cytokines are soluble messenger proteins that allow cell to cell communication among immune system cells and other cells in the body during homeostatic maintenance and inflammatory immune responses. There are various techniques such as ELISA, flow cytometry, and quantitative real-time PCR (qRT-PCR) that facilitate detection of cytokine mRNA and protein in vitro and ex vivo. However, these approaches have limitations for tracking cells in vivo and isolating cytokine-positive cells for functional analyses such as adoptive transfer. To overcome these challenges, reporter mouse models have been developed and widely used for easy detection and the lack of toxicity for the expressing cells.
Transgenic reporter mice represent a powerful approach to understand tissue development in vivo, signals regulating cell-fate decisions, cell lineage tracing, and gene function in diseases. In the study of cytokine biology, commonly used approaches in the generation of reporter mouse models for lineage tracing include introduction of internal ribosome entry site (IRES)–reporter under the control of the promoter of target gene or reporter gene knock-in into the first exon of the target gene (1). In the first approach, the target gene remains intact, whereas in the second approach, the target gene is disrupted. Both of these approaches allow for the detection of cells expressing the gene of interest (2). For example, IL-4 is a signature cytokine produced by TH2 cells that play a fundamental role in pathogenesis of humoral immunity, parasitic infections, and allergic inflammatory diseases (3–9). To study the role of TH2 cells and other IL-4–producing cells, several IL-4 reporter mouse models have been generated and each has their advantages and disadvantages. IL-4/GFP–enhanced transcript (4get) mice were one of the first cytokine reporter mice generated to detect IL-4 expression in situ (10). The 4get mice were made by inserting an IRES-GFP cassette into the 3′ untranslated region of the Il4 locus. The IL-4 reporter 4get mice have played an important role in identification of other IL-4–secreting cells populations, including NKT cells, basophils, eosinophils, and mast cells (11–13). Although the 4get mice identified IL-4 protein–expressing cells, the allele also expressed GFP when the allele was in an open and accessible conformation, but no Il4 mRNA was translated. To identify IL-4–secreting cells, a dual reporter system was generated by using 4get mice and human CD2/IL-4 reporter mice also known as knock-in hCD2 (KN2) mice (14). The KN2 mice were generated by replacing the first exon of mouse Il4 gene with the huCD2 cassette. The use of dual reporter system allowed identification of IL-4–competent cells that were GFP+ and IL-4–producing cells that were GFP+hCD2+. In a parallel approach to understand the function of IL-4 in vivo, GFP/IL-4 mice were also generated (15, 16). In this study, the first exon and first intron of the Il4 gene were replaced by the gene encoding–enhanced GFP–creating mice that are deficient in IL-4 yet allow detection of IL-4–producing cells. Each of the knock-in IL-4 reporter mouse models have proven to be advantageous in studies defining the function of IL-4 in inflammatory diseases (17–19).
IL-9 is another pleotropic cytokine associated with immunity to helminthic parasites and tumors as well as a myriad of inflammatory diseases, including allergic airway inflammation and autoimmune disorders (20–24). In allergic airway diseases, IL-9 has been shown to increase accumulation of mast cells and eosinophils in the lung, promote IgE and type-2 cytokine production, and hyperresponsiveness (25, 26). Although the CD4+ T helper subset called TH9 cells are predominant producers of IL-9, IL-9–producing cell types also include type 2 innate lymphoid cells (ILC2s), mast cells, and NKT cells (26–31). With increasing evidence of the involvement of IL-9 and IL-9–producing cells in immunopathological diseases, several groups have generated IL-9 reporter mouse models to trace IL-9–producing cells in vivo (32, 33). The first IL-9 reporter mice successfully generated were IL-9 fate reporter mice. IL-9 fate reporter mice were made by introducing a first exon knock-in of Cre into an Il9 bacterial artificial chromosome (34). Surprisingly, the IL-9 fate reporter, crossed to mice with the Gt(ROSA)26Sortm1(EYFP)Cos/J allele, only detected 10% of the IL-9–secreting T cells cultured in vitro. Still, the lineage tracer identified IL-9–secreting innate lymphoid cell (ILC) and TH9 populations in several airway disease models.
Similar to IL-9 fate reporter mice, Gerlach et al. (35) used Il9citrine mice generated by insertion of a citrine-coding gene into the first exon of the endogenous Il9 gene, which disrupted expression of endogenous Il9 but allowed detection of IL-9–producing cells in vivo in an inflammatory bowel disease model, but only a fraction of IL-9–secreting cells from TH9 cultures. Similarly, Il99er mice, with enhanced YFP-IRES-Cre knock-in to first Il9 exon, have been used to identify IL-9–producing T cells and ILC populations in acute allergic lung inflammation (36). Using a different approach, Licona-Limón et al. (37) generated the IL-9 fluorescence-enhanced reporter (INFER) mice, Il9GFP, by inserting an IRES-EGFP cassette at the 3′ end of the Il9 gene. In vitro T cell differentiation analysis showed that around 80% of the IL-9–secreting population was also positive for the GFP reporter. The Il9GFP mice allowed detection of IL-9–secreting ILC and TH9 populations as well as quantitation of endogenous IL-9 production in a parasite model. Each of the IL-9 reporter mice generated thus far show fidelity in detecting IL-9–producing cell populations; however, the efficiency in reporting for IL-9 expression varies in different models. Furthermore, the relative efficiency of these reporter mice in accurately detecting IL-9 production in different cell lineages remains unclear. In this study, we demonstrate that there are differing efficiencies of Il9 reporter alleles in detecting IL-9 among distinct cell types.
Materials and Methods
Mice
Exon-1–knock-in knockout Il9tm1Anjm (Il9citrine) mice and INFER Il9tm2.1Flv (Il9GFP) mice that were previously described were obtained from Dr. A.N.J. McKenzie’s and Dr. R.A. Flavell’s laboratory (35, 37). The Il9citrine mice are on BALB/cJ genetic background whereas Il9GFP are on C57BL/6J (B6/J), accordingly BALB/cJ, C57BL6J (B6/J); 129S (129S1/SvlmJ) mice were obtained from Jackson Laboratory as background control. All experiments were done with 8- to 12-wk-old mice.
In vitro T cell differentiation
Naive CD4 T cells were isolated from mouse spleens using CD4+CD62L+ T Cell Isolation Kit provided by the supplier (Miltenyi Biotec). Cells were cultured in complete RPMI 1640 media on anti-CD3 (2 U/ml 145-2C11; BioXCell)–coated-plates and soluble anti-CD28 (2.5 μg/ml; BD Pharmingen) under TH9-polarizing conditions including the following: hTGF-β1 (2 ng/ml), IL-4 (20 ng/ml), hIL-2 (50 U/ml), anti–IL-10R (10 μg/ml), and anti–IFN-γ (10 μg/ml) (all cytokines were obtained from PeproTech, and Abs were obtained from BioXCell). On day 3, cells were expanded into fresh media containing the original concentrations of cytokines in the absence of costimulatory signals for an additional 2 d. On day 5, mature TH9 cells were harvested for further analysis.
In vitro bone marrow–derived mast cell generation
To generate bone marrow–derived mast cells (BMMCs), bone marrow cells from wild-type (WT) or Il9citrine/citrine mice were isolated and RBCs were lysed using ammonium-chloride-potassium lysis buffer (LONZA). Cells were cultured in complete RPMI 1640 with IL-3 (10 ng/ml) and stem cell factor (30 ng/ml) and allowed to mature for 21 d.
ILC isolation
For ILC2 isolation, WT and Il9citrine/citrine mice were treated intranasally (i.n.) with 0.5 μg of IL-33 for 3 d. On day 4, lungs were harvested, minced, and incubated with 0.5 mg/ml collagenase A (Roche Diagnostics) in DMEM at 37°C for 45 min. To generate single-cell suspension, digested lung tissue was passed through stainless steel–meshed strainers, and RBCs were removed by using ammonium-chloride-potassium lysis buffer. Cells were further washed with PBS containing 0.5% BSA and filtered through 70-μm nylon mesh to remove debris. For intracellular staining of IL-9 in ILC2 subset, lung cells were restimulated with 50 ng/ml IL-33 for 5 h and treated with monensin for 2 h. ILC2 were identified as Lineage− (Lin−), CD45.2+, CD90.2 (Thy1.2)+, Sca-1+, Klrg-1+, and ST2+ and were stained for IL-9. For isolation of ILC2 subset, lung-cell suspension was lineage depleted using biotinylated Abs, including Mouse Lineage Depletion Kit and CD11c, CD19, DX5, and NK1.1 from Miltenyi Biotec. Lineage-depleted lung cells were further stained for ILC2 markers, including CD45.2, 104 (BD Horizon); Klrg1, 2F1 (BD Horizon); Sca-1, D7 (BD Horizon); and ST2, U29-93 (BD Horizon), and sorted on a BD FACS Aria II at the Indiana University School of Medicine flow cytometry core.
qRT-PCR
Total RNA was isolated from cells using TRIzol (Life Technologies). RNA was reverse transcribed according to manufactures directions (Quantabio, Beverly, MA). qRT-PCR was performed with commercially available primers (Life Technologies) with a 7500 Fast-PCR Machine (Life Technologies). Gene expression was normalized to housekeeping gene expression (β2-microglobulin). In the case of quantitative PCR (qPCR) for chromatin immunoprecipitation (ChIP) assay, SYBR Green master mix (Applied Biosystems) was used for analysis (38).
Flow-cytometric analysis
For cytokine staining, CD4+ T cells were stimulated with PMA (50 ng/ml; Sigma-Aldrich) and ionomycin (1 μg/ml; Sigma-Aldrich) for 3 h followed by monensin (2 μM; BioLegend) for a total of 6 h at 37°C. BMMCs were treated with monensin and restimulated with IL-33 (50 ng/ml) for 4 h. After restimulation, cells were washed with FACS buffer (PBS with 0.5% BSA). CD4+ T, BMMCs, or total lung cells were then stained with a fixable viability dye (eFluor780; eBioscience) and surface markers (CD4, GK1.5 [BioLegend]; CD90.2, 53-2.1 [BioLegend]; CD117 [c-kit], 104D2 [BioLegend]; FcεR1a, MAR-1 [BioLegend]; CD45.2, 104 [BD Horizon]; Klrg1, 2F1 [BD Horizon]; Sca-1, D7 [BD Horizon]; ST2, U29-93 [BD Horizon]; and mouse lineage Ab mixture [BD Pharmigen]) for 30 min at 4°C followed by washing and fixation with 4% formaldehyde for 10 min at room temperature (RT). For cytokine staining, cells were then permeabilized with permeabilization buffer (eBioscience) for 30 min at 4°C and stained for cytokines (IL-9, RM9A4; BioLegend) for 30 min at 4°C.
ChIP assay
In vitro–differentiated Th cells were activated with anti-CD3 for 3 h and were crosslinked for 15 min with 1% formaldehyde at RT with rotation. The reaction was quenched by adding 0.125 M glycine and incubated at RT for 5 min. Fixed cells were lysed with cell lysis buffer followed by nuclear lysis buffer. Nuclei were degraded, and chromosomal DNA were fragmented to a size range of 200–500 bp through an ultrasonic processor (Vibra-Cell). After sonication, the supernatant was diluted 10-fold with ChIP dilution buffer. After preclearing, the supernatant was incubated with the ChIP Abs at 4°C overnight with rotation. The following day, immunocomplexes were precipitated with protein agarose A or G beads at 4°C for 2–4 h. Immunocomplexes were washed with low salt, high salt, LiCl, and two times with TE buffer. After elution followed by reverse crosslinks, DNA was purified and analyzed by qPCR. After normalization to the input DNA, the amount of output DNA of each target protein was calculated by subtracting that of the IgG control. Quantification of ChIP assay by qPCR was performed using primers as described (38).
Chromatin accessibility assay
Chromatin was isolated from in vitro-cultured T cells and restimulated for 3 h with anti-CD3. Chromatin was digested with nuclease (Nse) mix using the EpiQuik Chromatin Accessibility Assay Kit (EpiGentek). Isolated chromatin was divided into two, one for Nse treatment and another for nontreatment. After incubation at 37°C for 4 min, reaction was quenched by adding reaction stop solution. Samples were incubated with proteinase K at 65°C for 15 min to degrade any contaminating proteins. DNA was purified followed by qPCR to amplify DNA fragment with primers for Il9 conserved noncoding sequence regions (same as ChIP primers) or Hbb-bs for negative control (forward: 5′-GAGTGGCACAGCATCCAGGGAGAAA-3′; reverse: 5′-CCACAGGCCAGAGACAGCAGCCTTC-3′) The fold enrichment (FE) was calculated by the following formula: FE = 2(Nse CT − no Nse CT) × 100%, where CT stands for cycle threshold.
Statistical analysis
All the data were analyzed using two-tailed Student t test to generate p values. Post hoc Tukey test was used for multiple comparisons. The p value ≤0.05 was considered statistically significant.
Results
Il9+/citrine reporter mice inadequately report for IL-9 expression in TH9 cells differentiated in vitro
To characterize TH9 cells in more detail, we sought to profile IL-9 reporter expression in two distinct IL-9 reporter mice, Il9+/citrine or Il9+/GFP, that have been previously used to identify IL-9–producing cells in disease models. Naive CD4 T cells from Il9+/citrine or Il9+/GFP were cultured under TH9-polarizing conditions for 5 d in vitro and IL-9 or reporter expression was monitored through each day of TH9 differentiation. Flow-cytometric analysis showed that in the Il9+/citrine cultures, there were small populations of IL-9+ and citrine+ cells, which were largely mutually exclusive. Because each allele produces only one of the detected proteins, it suggests there is monoallelic expression of Il9 alleles within cells (Fig. 1A). There was a large increase in IL-9 staining on day 4, but far less of an increase in citrine expression, with citrine detecting only about a third of the IL-9–producing cells (Fig. 1A). On day 2 and 3 of Il9+/GFP cell culture, there were significant amounts of GFP detected in the absence of IL-9 production, suggesting that the reporter is indicating gene competence. By day 4, there is a predominant population of IL-9+ GFP+ cells in the Il9+/GFP cell culture (Fig. 1A). To further quantify expression from these alleles, we performed qRT-PCR. The relative amount of GFP transcript expressed by the Il9+/GFP allele was about half that of Il9 (Fig. 1B). However, citrine expression was <10% of Il9 transcript levels (Fig. 1B). In addition to examining reporter expression in TH9 cells in heterozygous IL-9 reporter mice, we also assessed the reporter expression in reporter allele homozygous mice. Similar to Il9+/citrine TH9 cultures, the Il9citrine/citrine cells reported only 7% of the TH9 culture (Fig. 1C), whereas, Il9GFP/GFP cells reported for about half of the TH9 culture (Fig. 1C). The distinct function of the reporters was not due to differences in the background of the mice, as TH9 cultures from BALB/c mice, C57BL6 mice, and 129S mice [which were used to generate the citrine allele (35)] had similar percentages of IL-9+ TH9 cells (Fig. 1D).
Profile of fluorescent protein expression in Il9 reporter mice.
(A) Reporter expression measured over a period of 5 d during TH9 differentiation in vitro from Il9+/citrine or Il9+/GFP mice. (B) mRNA expression of Il9 and respective reporter (citrine or GFP) measured on day 4 of TH9 differentiation. (C) IL-9 reporter expression by day 5 TH9 cells cultured in vitro from in Il9citrine/citrine and Il9GFP/GFP mice. (D) IL-9 production by day 5 TH9 cells cultured in vitro from BALB/c, C57BL6, and 129S mice. Data are mean ± SEM of three mice per experiment and representative of at least two independent experiments. *p < 0.05.
Disruption of exon-1 region in Il9 gene in Il9citrine/citrine reporter mice reduced transcriptional activity at the Il9 locus in TH9 cells
The decreased ability of the citrine allele to report for IL-9 production suggested that targeting the exon-1 region at Il9 locus may have interfered with regulation of the reporter allele. Recent studies have identified and reported conserved noncoding sequence regions across the Il9 locus that play an important role in Il9 gene transcription and regulation (38, 39) (Fig. 2A). To define how targeting the first exon of Il9 affected gene activity, we measured chromatin accessibility across the conserved noncoding sequence regions at the Il9 locus in both BALB/c and Il9citrine/citrine mice (38). The Il9 locus in TH9 cells cultured from Il9citrine/citrine mice was much less accessible than the WT allele (Fig. 2B). To determine whether the disruption of the exon-1 region of the Il9 locus in Il9citrine mice affected transcriptional activity, we performed ChIP assays to assess histone modifications at the intronic, exonic, and conserved noncoding sequence regions of the Il9 locus in BALB/c, C57BL/6, 129S, Il9citrine, and Il9GFP TH9 cultures. H3K27 acetylation, a mark of open chromatin was similar among TH9 cells from BALB/c, C57BL/6, 129S, and Il9GFP although it was considerably lower in the Il9citrine cultures (Fig. 2C, 2D). H3K4 trimethylation, a mark of active transcription, demonstrated a similar pattern in the WT strains but was lower in the Il9GFP TH9 cells and almost undetectable in the Il9citrine cultures (Fig. 2C, 2D). The pattern of RNA polymerase II (pol II) binding across the locus varied greatly among the WT strains, suggesting strain-specific patterns of pol II pausing on the allele (Fig. 2C, 2D). The Il9GFP TH9 cells showed a unique peak in the 3′ end, suggesting a link to the IRES element (Fig. 2D). In contrast, the Il9citrine TH9 cells had barely detectable RNA pol II binding (Fig. 2C). Together these findings demonstrated that disruption of the exon-1 region of Il9 gene impaired transcriptional activity of the Il9citrine reporter allele in TH9 cells.
Disruption of Il9 exon 1 impaired transcriptional activity.
(A) Schematic of conserved noncoding sequences (CNS) at the Il9 locus. (B–D) Naive CD4+ T cells from WT or Il9citrine mice were cultured in TH9 conditions for 5 d and restimulated with PMA and ionomycin for 3 h upon harvest on day 5. (B) Chromatin accessibility assessed at Il9 locus in BALB/c and Il9citrine/citrine allele. (C) Relative amounts of H3K27ac, H3K4me3, and RNA pol II at Il9 locus in TH9 cells examined ChIP-qPCR in BALB/c and Il9citrine/citrine mice. (D) Relative amounts of H3K27ac, H3K4me3, and RNA pol II at Il9 locus in TH9 cells examined ChIP-qPCR in C57BL6, 129S, and Il9GFP/GFP mice. Data are mean ± SEM of three mice per experiment and representative of two independent experiments.
In contrast to TH9 and mast cells, Il9citrine/citrine reporter mice demonstrated higher IL-9 reporter expression in ILC2 population compared with WT allele
In addition to TH9 cells, mast cells, and ILC2s have also been described as potent IL-9 producers (28, 33, 40). To investigate whether Il9citrine/citrine mice differentially reported for other IL-9–producing cells we looked at IL-9 and reporter expression in BMMCs restimulated with IL-33. As in TH9 cells, Il9citrine/citrine reported for less than half of the IL-9 production in c-kit+ FcεR1α+ BMMCs cultured in vitro (Fig. 3). We were unable to detect consistent IL-9 production in mast cell cultures from WT and Il9GFP/GFP mice on the C57BL/6 background in these studies.
Il9 reporter expression in BMMCs compared.
IL-9 and citrine expression in BMMCs derived from WT or Il9citrine/citrine mice. Data are mean ± SEM of three mice per experiment and representative of two independent experiments. *p < 0.05.
To examine whether the Il9citrine/citrine and Il9GFP/GFP allele exhibited a similar pattern of reporter expression in ILC2 populations, Il9citrine/citrine, and Il9GFP/GFP mice along with their respective WT background (BALB/c and C57BL/6) mice were i.n. challenged with rIL-33 for 3 d (Fig. 4A). On day 4, lungs were harvested, and samples were analyzed for IL-9 or reporter expression in the ILC2 population defined as Lin− CD45.2+ Sca1+ Klrg1+ ST2+ Thy1.2+ cells. Using flow cytometry, we observed that in contrast to TH9 and mast cells, both Il9citrine/citrine and Il9GFP/GFP mice reported higher percentages of IL-9–producing ILC2s than WT mice, although the increase in Il9GFP/GFP mice was NS (Fig. 4B–D). We then sorted ILC2 populations from the lungs of Il9+/citrine and Il9GFP mice i.n. challenged with rIL-33 for 3 d and measured transcript level expression of IL-9 and the respective reporter and compared expression to in vitro–derived TH9 cells examined in Figs. 1, 2. Comparing the ratio of reporter/Il9 gene expression in TH9 and ILC2 populations in the respective reporter mice, we observed that the Il9citrine allele preferentially reported for IL-9 in ILC2s when compared with TH9 cells, and the Il9GFP allele trended in that direction (Fig. 4E). Importantly, the Il9citrine allele in ILC2s reported ∼100 times more expression than in the TH9 population, whereas the Il9GFP allele showed differences that were more modest and not statistically significant. These findings indicate that there are important differences in how a reporter allele identifies cytokine expression in distinct cells types and that knock-in to a coding sequence or modifying the target gene locus has the potential to interfere with locus expression. This is an important caveat in interpreting experiments using various reporter mice.
Il9citrine/citrine and Il9GFP/GFP mice selectively reported IL-9 production in ILC2s.
(A) Schematic of rIL-33 i.n. challenge in Il9 reporter mice for generation and detection of lung ILC2 population. (B) Gating strategy for identification of ILC2 population in the lung defined as Lineage−CD45+Klrg1+Sca1+Thy1.2+ST2+. (C) Detection of IL-9 and citrine reporter production in lung ILC2s. (D) Detection of IL-9 and GFP reporter production in lung ILC2s. (E) Ratio of reporter/Il9 allelic expression in TH9 cells (as generated in Fig. 1) and ILC2 subsets. Data are mean ± SEM of three mice per experiment and representative of two independent experiments. *p < 0.05.
Discussion
Reporter mice serve as a remarkable tool to identify gene function and detect cells expressing the gene of interest in vivo with minimal toxic effects. However, the fidelity of the reporter to identify populations of cells expressing the gene of interest and the efficiency in reporting for gene expression needs to be defined for each model as well as cell type. In this study, we examined the fidelity and efficiency of two IL-9 reporter mice in identifying the IL-9–secreting T cells cultured in vitro. Our results showed that although both Il9GFP and Il9citrine mice identified IL-9–secreting cells, the Il9GFP mice detected a majority of the IL-9–producing T cells, whereas the Il9citrine mice identified ∼20% of the IL-9–secreting T cells in vitro. Thus, different approaches for targeting a gene of interest can result in differing efficiencies of reporting.
The mechanisms for the functional differences between reporter alleles are still not entirely clear. Although the Il9citrine allele maintains specificity for expression in TH9 cells (35, 41), it is likely that the Il9citrine allele disrupts a regulatory element that contributes to Il9 expression in T cells and mast cells, consistent with the altered chromatin modifications observed at the Il9citrine allele. Deletion of the Il9 enhancer region was similarly crucial for Il9 gene regulation in T cells and mast cells but not in the ILC2 population (38, 39). Thus, these findings indicate that the first exon of the Il9 locus may be important in transcriptional regulation of Il9 in T cells. In contrast to the lack of expression of the Il9citrine allele in T cells, ILC2s show enhanced reporter protein and mRNA compared with IL-9 protein and Il9 mRNA for both reporter strains. The differences in protein could result from differences in protein half-lives of reporter versus cytokine and of fluorescent proteins among cell types. The differences in mRNA might result from altered regulation of the gene, but it is also possible that the gene encoding the fluorescent protein confers greater stability to the mRNA in ILC2s. These distinctions would require more detailed analyses.
The expression pattern observed in the Il9citrine mice is similar to the previously generated Il9Cre fate reporter mice in which only 10% of IL-9–secreting T cells were identified by the IL-9 fate reporter mice (34). Both Il9citrine and IL-9 fate reporter mice were generated by knocking in the reporter into the first exon of Il9 gene. Consistent with these observations, Schwartz et al. (42) noted substantial differences between gene expression patterns observed in IL-9 fate reporter mice compared with Il9GFP mice in a papain-induced allergic airway disease model. These identified differences in expression efficiency among Il9 reporter allele across cell lineages requires a reexamination of conclusions regarding the relative contributions of cells producing IL-9. In models of airway inflammation and rheumatoid arthritis, it was concluded that ILC2s are the primary producer of IL-9 (34, 43, 44). The IL-9 fate reporter and Il9citrine mice used in these studies more efficiently identifies IL-9 expression in ILC2 than TH9 cells, potentially underrepresenting the role of TH9 cells. Until IL-9 conditional mutant mice are used to selectively remove IL-9 production in specific populations, the definitive answer to the major or required source of IL-9 will be unclear. More than likely, the cellular source of IL-9 will vary depending on the model and type of inflammation.
Another potential concern in reporter alleles is elimination of the endogenous gene, as occurs in the Il9citrine mice, if there is positive feedback on gene expression. Indeed, in mast cells, IL-9 can amplify its own production as well as the production of other type 2 cytokines (45). However, in multiple experiments with TH9 cultures in which IL-9 is blocked during culture or supplemental IL-9 is added during culture, we have not observed any changes in IL-9 production at the end of the culture (data not shown). Thus, the reduced IL-9 production in the Il9citrine TH9 cultures is unlikely to be from a lack of a positive feedback from IL-9 itself.
The Il9citrine allele is clearly less active in T cells. Compared with either BALB/c or 129S control cells, the allele was less accessible, had less H3K27 acetylation, and less H3K4 trimethylation. These markers of active genes also correlated with less RNA pol II binding. In contrast, the Il9GFP allele had normal H3K27 acetylation. The decreased H3K4 trimethylation at the Il9GFP allele was surprising given the expression of the allele. However, this might be a result of IRES effects on the rest of the gene. In fact, we observed unique peaks at exon 5 for RNA pol II binding in the Il9GFP allele that support a role for the IRES in altering transcription patterns at the Il9 locus. Together, these results suggest that reporter alleles, regardless of targeting strategy, likely alter the chromatin structure of the allele, and that inadvertently targeting regulatory elements or inserting IRES sequences will both have effects.
Reporter alleles can also help define the monoallelic expression observed in some cytokine genes. For example, Holländer et al. (46) first reported monoallelic expression of IL-2 in mature thymocytes and T cell populations identified using polymorphisms in alleles (47). Interestingly Naramura et al. (48) showed that CD4 T cells cultured from IL-2–GFPki mice (GFP knocked into the Il2 locus) failed to exhibit monoallelic expression or allelic exclusion. These observations suggest that gene-regulatory elements including exons, introns, promoter, and enhancer regions can influence allelic expression in specific cells types. The Il9citrine TH9 cultures early in differentiation do show early exclusion, showing preferential expression of either IL-9 or citrine. However, at the later stages of differentiation, there are double-positive cells, suggesting that exclusion is not complete. It is possible that while gaining competence for expression, there is monoallelic expression, but as the cell becomes more differentiated, both alleles are expressed. Importantly, the early exclusion is not observed in the Il9GFP allele because both IL-9 and GFP can be generated from the same allele.
There is also the question of whether alleles report strictly for expression or also for competence of the allele, as observed in 4get mice (10). In the Il9GFP allele, generated with an IRES insertion like the 4get allele, there are large populations of GFP+ cells during early TH9 differentiation when IL-9 protein secretion is absent. That observation parallels data from the 4get mice that showed GFP expression where there was no Il4 mRNA. This parallel suggests that competency of the locus is being reported and that the allele might be useful in analysis of changes in chromatin structure at the Il9 locus.
Our results have demonstrated that different approaches to targeting the Il9 locus in the generation of a reporter allele result in reporters that have efficiencies that vary with cell type. This is an important consideration in selecting alleles for the studies of specific cell types and presents caveats for studies in which positive cells could not be identified. In these situations, it is possible that technical issues with a reporter allele could be responsible for negative results. Whether this paradigm is restricted to the Il9 locus or whether it is broadly applicable to other reporter alleles will be defined as studies with these reagents continues.
Disclosures
The authors have no financial conflicts of interest.
Acknowledgments
We thank Andrew McKenzie for providing the Il9citrine mice and for comments on the manuscript.
Footnotes
This work was supported by National Institutes of Health Public Health Service Grants (NIH PHS) R01 AI057459, R01 AI129241, and R03 AI135356 (to M.H.K.). B.J.U. was supported by NIH PHS Grants T32 AI060519 and F30 HL147515. A.A.Q. was supported by NIH PHS Grant T32 DK007519. Core facility usage was also supported by Indiana University Melvin and Bren Simon Cancer Center Support Grant P30 CA082709 and NIH PHS Grant U54 DK106846. Support provided by the Herman B Wells Center was in part from the Riley Children’s Foundation.
Abbreviations used in this article:
- BMMC
- bone marrow–derived mast cell
- ChIP
- chromatin immunoprecipitation
- 4get
- IL-4/GFP–enhanced transcript
- ILC
- innate lymphoid cell
- ILC2
- type 2 innate lymphoid cell
- i.n.
- intranasal(ly)
- IRES
- internal ribosome entry site
- Nse
- nuclease
- pol II
- RNA polymerase II
- qPCR
- quantitative PCR
- qRT-PCR
- quantitative real-time PCR
- RT
- room temperature
- WT
- wild-type.
- Received September 30, 2019.
- Accepted May 7, 2020.
- Copyright © 2020 The Authors
This article is distributed under the terms of the CC BY 4.0 Unported license.
References
