Cutting Edge: Core Binding Factor β Is Required for Group 2 Innate Lymphoid Cell Activation

Group 2 innate lymphoid cells (ILC2) are long-lived innate effector cells residing at mucosal barriers such as lung and gut. Mature ILC2 functionally and transcriptionally mirror Th2 cells, but ILC2 lack clonal Ag receptors and can rapidly respond to the epithelial alarmin IL-33 in the absence of Ag stimulus (1). Upon activation, mature ILC2 quickly secrete large amounts of type-2 cytokines IL-5 and IL-13 as well as other effector molecules such as VEGFA (1, 2). Activated ILC2 are potent inducers of airway hyperresponsiveness (AHR) and are implicated in allergic asthma and asthma exacerbation. The molecular pathways that drive ILC2 activation, however, are not well understood.

Core binding factor β (CBFβ) is a non-DNA–binding subunit that binds with the RUNX proteins to form different CBF transcriptional complexes. CBFs are critical for early lymphocyte development. CBFβ is required for the generation of thymic T cell progenitors, bone marrow B cell precursors, and early innate lymphoid progenitors (3–7). CBFs are dispensable for the maintenance of mature T cells, but they control several aspects of T cell differentiation. CBFs are known to repress CD4⁺ Th2 differentiation (8–10). Deletion of CBFβ or RUNX proteins in T cells derepressed GATA3 and IL-4 expression in CD4⁺ T cells, resulting in elevated IgE and airway eosinophil inflammation in mice (8–10). Whether CBFs play a similar role in repressing the function of innate type-2 cytokine–producing cells, such as ILC2, remains unknown.

The current study was undertaken to examine the role of CBFβ in the activation and function of mature ILC2. To our surprise, contrary to its suppressive role in Th2 cell differentiation, our data indicate that CBFβ promotes ILC2 activation. Deletion or inhibition of CBFβ abrogated ILC2 activation during allergic airway inflammation and prevented ILC2-mediated AHR. The requirement of CBFβ in ILC2 function was cell intrinsic and involved both transcriptional and posttranscriptional gene regulatory mechanisms. These data establish a critical role for CBFβ in ILC2 activation and function. The positive regulation of ILC2 function by CBFβ is in contrast with the suppressive role of CBFβ in CD4⁺ Th2 cell differentiation (8, 9). Our data thus indicate that adaptive and innate type-2 immune responses are controlled by divergent molecular mechanisms. Such divergence in molecular control might help explain the extreme heterogeneity and complexity of human diseases and necessitates consideration of personalized medicine based on the specific immune responses involved.

CBFβ^f/f (Cbfb^tm2.1Ddg/J, line 028550), Id2^CreERt2(B6.129S(Cg)-Id2^{tm1.1(cre/ERT2)Blh/ZhuJ} line 016222), and Rosa26R^yfp (B6.129 × 1-Gt(ROSA)26Sor^tm1(EYFP)Cos/J, line 006148) mice were obtained from The Jackson Laboratory (11). CBFβ^f/f Id2^CreERt2 Rosa26R^yfp and control Id2^CreERt2 Rosa26R^yfp mice were bred at the Albany Medical College Animal Research Facility. BALB/c mice were purchased from Taconic Biosciences.

For deletion of CBFβ in ILC2, five doses of 5 mg of tamoxifen were administrated i.p. every other day. Twenty-eight days after the first tamoxifen treatment, mice were administrated with a single dose of IL-33 (400 ng; BioLegend) i.p. BALB/c mice received a single dose of intranasal administration of Alternaria extracts (100 μg; Greer) as described (2). CBFβ inhibitor Ro5-3335 (50 mg/kg; EMD) was administered i.p. 24 h before Alternaria challenge. AHR was measured 12 h after Alternaria challenge by a flexiVent system (SCIREQ). All mice experiments were approved by the Albany Medical College Institutional Animal Care and Use Committee.

Sorted lung ILC2 or ILC2 line were cultured with 10 ng/ml IL-2, IL-7, and IL-33 for 5 d. One hundred nanomolar 4-hydroxytamoxifen (4-OHT) (Sigma-Aldrich) was added to primary lung ILC2 culture at day 6. Cells were cultured for nine more days in the presence of 4-OHT. Cytokine production of YFP⁺ cells was examined by intracellular staining and flow cytometry analysis. In some experiments, YFP⁺ cells were sorted after 5 d of culture with 4-OHT and cultured for another 4 d in the presence of 4-OHT. Growth rate of YFP⁺ cells were measured as the inverse of doubling time. In vitro Th2 differentiation was performed as described (12).

Abs were purchased from eBioscience, BioLegend, or MD Biosciences. Flow cytometric analysis was performed on FACSCanto (BD Biosciences). Flow cytometry cell sorting was performed on a FACSAria II (BD Biosciences).

mRNA was extracted by Qiagen RNeasy kits. Gene expression was examined by quantitative PCR (qPCR). Microarray analysis was performed at Boston University Microarray and Sequencing Resource (accession number GSE116062, https://www.ncbi.nlm.nih.gov/geo). Gene enrichment analyses were provided by Boston University Clinical and Translational Science Institute translational bioinformatics consultation service (CTSA Grant U54-TR001012).

Sucrose gradients were performed to examine gene translation. Specifically, cells were treated with 100 μg/ml cycloheximide (Sigma-Aldrich) for 10 min and lysed on ice. Cell lysates were layered onto 10–50% sucrose gradients and centrifuged in an SW 41 Ti rotor at 35,000 rpm for 2 h. Sixty fractions were carefully collected. OD260 absorption for each fraction was recorded. Unassociated RNA (pooled from fractions 1–9) and polysome-associated RNA (pooled from fractions 41–59) were extracted and examined by qPCR. Translation was determined by the ratio of polysome-associated RNA versus -unassociated RNA and normalized to Gapdh.

ILC2 cell line was described previously (13). Specifically, an immortalized cell line was established with small intestinal lamina propria ILC2 by selection of spontaneous mutants. The ILC2 line cells have been maintained with IL-2 and IL-33, and therefore they phenotypically and functionally resemble activated primary ILC2. Chromatin immunoprecipitation (CHIP) was performed as we previously described (13). LentiCRISPRv2GFP was a gift from David Feldser (plasmid no. 82416; Addgene). Guide RNA (gRNA)-encoding sequences were cloned into LentiCRISPRv2GFP vectors to achieve gene knockout effects as described (14). Two gRNA sequences targeting Cbfb achieved effective knockout effects. The gRNA (5′-CGATCTCCGAGCGACCGTCG-3′) was used throughout this study, but major findings were verified using the other gRNA (5′-ACCGCCTCACCTCGCACTCG-3′) with similar results. Nontarget control (5′-TGCGAATACGCCCACGCGATGGG-3′) was used to generate control CRISPR constructs. Lentiviral transduction of ILC2 was performed as described (15). Gene knockout effects were verified by Western blotting.

Data are shown as mean ± SD. Unpaired t test was used to compare the difference of two groups.

To examine the role of CBFβ in mature ILC2 maintenance and function, we generated CBFβ^f/f Id2^CreERt2 Rosa26R^yfp (CBFβ^f/f) and control Id2^CreERt2 Rosa26R^yfp (CBFβ^+/+) mice. Because mature ILC2 express high amounts of Id2, tamoxifen administration results in deletion of Cbfb in ILC2 of CBFβ^f/f mice. The deletion efficiency is indicated by expression of YFP. Comparable numbers of YFP⁺ ILC2 appeared in CBFβ^f/f and control CBFβ^+/+ mice 4 wk after tamoxifen treatment, indicating that the loss of CBFβ did not affect the survival and maintenance of ILC2 at homeostasis (Fig. 1A–C). qPCR verified effective deletion of Cbfb in YFP⁺ ILC2 of CBFβ^f/f mice (Fig. 1D).

To examine whether CBFβ regulates ILC2 function, we activated ILC2 in vivo by treating mice with IL-33. A single dose of IL-33 induced production of IL-5 and IL-13 in mature lung ILC2 within 12 h (Fig. 1E–G). CBFβ-deficient ILC2, however, failed to efficiently produce IL-5 or IL-13 in vivo (Fig. 1E–G). Thus, CBFβ is required for ILC2 activation in vivo.

To confirm the relevance of CBFβ-mediated ILC2 activation, we used a model of acute Alternaria allergen inhalation. Alternaria alternata is a fungal allergen that triggers acute and severe asthma attacks. As described (2, 16), a single dose of Alternaria inhalation rapidly upregulated IL-5 and IL-13 expression in lung ILC2 within 12 h (Supplemental Fig. 1). CD4⁺ Th2 cells were not activated at this time point (2). Treatment with CBFβ inhibitor (Ro5-3335) reduced production of IL-5 and IL-13 from ILC2 (Supplemental Fig. 1). Alternaria inhalation rapidly induced airway eosinophil and neutrophil infiltration that was prevented by Ro5-3335 (Fig. 1H). Acute Alternaria inhalation also rapidly elicited strong AHR that was suppressed by Ro5-3335 (Fig. 1I). Together, CBFβ inhibition repressed ILC2 activation in response to acute Alternaria allergen inhalation and prevented ILC2-mediated AHR.

We sought to understand whether CBFβ controls ILC2 activation through cell-intrinsic mechanisms. We cultured highly purified lung ILC2 from CBFβ^f/f or control CBFβ^+/+ mice in the presence of IL-7, IL-2, IL-33, and 4-OHT. Induction of CRE expression by 4-OHT resulted in the appearance of YFP⁺ cells in both CBFβ^f/f and control CBFβ^+/+ ILC2 cultures. However, CBFβ-deficient ILC2 (YFP⁺ CBFβ^f/f) proliferated at a much lower rate than control ILC2 (YFP⁺ CBFβ^+/+) (Fig. 2A). CBFβ-deficient ILC2 also failed to efficiently produce IL-13 or IL-5 in response to IL-33 (Fig. 2B, 2C). Thus, cell-intrinsic CBFβ is required for ILC2 activation. Similar results have been obtained with CRISPR technique–mediated knockout of CBFβ in an ILC2 cell line, verifying an essential cell-intrinsic role for CBFβ in ILC2 activation (Supplemental Fig. 2A–D).

To determine the mechanisms through which CBFβ controls ILC2 activation, we examined gene expression of purified YFP⁺ ILC2 from the lungs of CBFβ^f/f and CBFβ^+/+ mice that were treated with tamoxifen and challenged with IL-33. GATA3 is a critical transcriptional controller in ILC2. However, deletion of Cbfb did not affect Gata3 mRNA levels in ILC2 in vivo, indicating that CBFs might control ILC2 activation through GATA3-independent mechanisms (Fig. 3A). Interestingly, mRNA expression for Il13 and Vegfa, but not Il5, was reduced in CBFβ-deficient ILC2 (Fig. 3A). Conserved CBF binding sites were identified at the −90-bp Il13 promoter and the −7.4-Kb Vegfa enhancer regions. CHIP indicated that RUNX proteins directly bound to both conserved regulatory regions in ILC2 line cells (Fig. 3B). Thus, CBF transcriptional complex may directly control the transcription of some ILC2 genes.

Nevertheless, the changes in Il5 and Il13 mRNA were rather moderate when compared with the remarkable decrease in protein expression in CBFβ-deficient ILC2 (Fig. 1E). Thus, additional posttranscriptional mechanisms are likely to be involved. We further examined the gene expression of purified CBFβ-deficient and control ILC2 from the lungs of tamoxifen-treated mice (Fig. 4A). Interestingly, CBFβ-deficient ILC2 had reduced expression of many ribosomal protein genes (Fig. 4A). They included the ribosome structural genes (rpl3, rpl5, rps2, rps15a), p component molecules (pop4, rpp30, and rpp38), and other key molecules involved in ribosome biogenesis and function (riok2, ddx21, rrp1b, rcl1, and nol11). Microarray analysis with CBFβ knockout and control ILC2 line cells verified that loss of CBFβ expression led to extensive reduction in mRNA expression for genes related to ribosome biogenesis, RNA processing, and translation (Fig. 4B). Thus, CBFβ promotes the expression of ribosomal protein genes in activated ILC2.

To directly examine whether CBFβ promotes gene translation in ILC2, we deleted Cbfb in cultured ILC2 by in vitro 4-OHT treatment and then performed sucrose gradient centrifugation to separate polysome-associated RNA from -unassociated RNA (Fig. 4C). We measured the efficiency of gene translation as the ratio of polysome-associated RNA versus -unassociated RNA. The translational levels of Il5, Il13, and Vegfa were all reduced in CBFβ-deficient ILC2, indicating that CBFβ is required for optimal gene translation in ILC2 (Fig. 4C). Similar results have been obtained with CBFβ knockout ILC2 line cells (Supplemental Fig. 2E, 2F). In contrast, CRISPR-mediated knockout of CBFβ did not affect the translation of Il13 in CD4⁺ T cells cultured under Th2 polarization condition, indicating that CBFβ was not required for Il13 translation in Th2 cells in vitro (Supplemental Fig. 2G). We next compared the ribosome profiles of CBFβ knockout and control ILC2 line cells. CBFβ knockout ILC2 had much smaller polysome-associated peaks, indicating defects in ribosome biogenesis and/or function (Fig. 4D). Together, these data indicate that CBFβ promotes ribosome biogenesis and/or function, thus enhancing gene translation during ILC2 activation.

Our study revealed an essential role for CBFβ in ILC2 activation and function. Of note, the positive regulation of ILC2 function by CBFβ is opposite to the repressive role of CBFβ in CD4⁺ Th2 differentiation (8, 9). Thus, despite striking functional similarity between ILC2 and CD4⁺ Th2 cells, there are important differences in molecular control between these two closely related lineages. The diametric regulation of innate and adaptive type-2 immunity by CBFβ might help balance the delicate immune responses at mucosal barrier sites. Such complexity in molecular control necessitates the search for new immune response endotypes in complicated human diseases such as asthma and allergy and dictates the importance of personalized medicine based on the specific immune responses involved.

CBFs appear to play diverse roles in different cell types and at distinct developmental stages. CBFs are required for the generation of early innate lymphoid progenitors, the maintenance of ILC1, and the specification of fetal lymphoid tissue inducer progenitors (5, 17, 18). However, they are dispensable for the maintenance of ILC2 and ILC3 at homeostasis. In addition, our data indicate that CBFβ promotes ribosome biogenesis and gene translation in mature ILC2. Ribosome biogenesis is also reduced in RUNX1-deficient hematopoietic stem cells (19). However, deficiency of RUNX proteins does not affect the ribosome biogenesis of CD34⁺flt3⁺ multipotent progenitors and even results in enhanced rRNA transcription and protein synthesis in mouse calvaria cells (19, 20). Thus, CBFs control the expression of ribosome biogenesis genes through cell type–specific mechanisms. Similarly, CBFs repress the expression of type-2 cytokines in CD4⁺ T cells but promote cytokine production in ILC2. The existence of lineage-specific coactivators and corepressors might help explain the differential and even opposing roles of CBFs in different cell types. Searching for ILC2-specific functional partners of CBFβ might reveal unique lineage-specific molecular pathways that control ILC2 activation and function.

The authors have no financial conflicts of interest.