NFIL3/E4BP4 is a key transcription factor for CD8α⁺ dendritic cell development

Dendritic cells (DCs) are effective antigen-presenting cells that both initiate antigen-specific immune responses and generate and maintain self-tolerance.¹  DCs can develop from bone marrow (BM) progenitor cells and reside in the peripheral lymphoid tissues. Several DC subsets were identified by their cell-surface phenotypes, residential location, and functional differences.²  Splenic DCs are divided into 2 major populations: plasmacytoid DCs (pDCs) and conventional DCs (cDCs). The cDCs are further divided into CD8α⁺ and CD8α⁻ DCs. CD8α⁺ DCs predominantly produce IL-12 in response to microbial antigens, tumor cells, and virus-infected cells.^1-3  Importantly, CD8α⁺ DCs are specialized cells for the cross-presentation of antigens by MHC class I molecules.^4,5  Analyses of mice lacking transcription factor genes such as Irf4, Irf8, and Batf3 revealed their essential roles in the development of distinct subsets of DCs.^6-9 

NFIL3 (nuclear factor, IL-3 regulated; also called E4BP4) was originally identified as an adenovirus E4 promoter– and human IL3 gene promoter–binding protein, and its role in circadian regulation has been well studied.^10-12  However, the role of NFIL3 in the immune system was not largely elucidated until Nfil3^−/− mice were established recently. We have demonstrated that Nfil3^−/− mice have defects in IgE class switching by regulating Ig-ϵ germline transcription in B cells.¹³ Nfil3^−/− mice also lack natural killer (NK) cells, which suggests a role for NFIL3 in NK cell development.^13-15  Here, we demonstrate impaired CD8α⁺ DC development and impaired cross-priming activity of CD8⁺ T cells in response to cell-mediated antigen in Nfil3^−/− mice. Our results provide compelling evidence that NFIL3 is a key regulator of CD8α⁺ DC development.

Nfil3^−/− mice¹³  (C57BL/6 background) were used for all experiments. All experimental mouse protocols were approved by the Institutional Animal Care and Use Committee of University of Iowa.

cDNA was prepared from RNA isolated from cells by the SuperScript First-Strand Synthesis System (Invitrogen) according to the manufacturer's instructions and was subjected to real-time PCR with SYBR Green PCR Master Mix (Applied Biosystems) as described previously.¹³  Primers used are listed in the supplemental Methods (available on the Blood Web site; see the Supplemental Materials link at the top of the online article).

Cell suspensions from spleen, thymus, blood, and BM culture were stained with the antibodies listed in supplemental Methods. Cells were analyzed with LSR II (BD Biosciences), and data were interpreted with CellQuest (BD Biosciences) or FlowJo (TreeStar).

Spleen cDCs were isolated as described previously.⁹  For BM-derived DCs, BM cells were cultured with mouse Flt3 ligand (FL; 50 ng/mL) for 9 days.

FL-activated BM cells were spin-infected with retrovirus as described previously¹³  and were grown for another 7 days before analysis.

OVA_257-264-specific CD8⁺ T-cell responses in mice injected with irradiated splenocytes from act-mOVA/K^b−/− mice were determined by MHC class I–peptide tetramer staining.¹⁶  Detailed materials and methods can be found in the online supplemental Methods.

Nfil3 expression is high in NK cells and IL-4–stimulated B and T cells (Figure 1A).^14,15 Nfil3 expression in sorted splenic pDC and cDC subsets was similar to NK cells and IL-4–stimulated B and T cells, which suggests a potential role of NFIL3 in DC populations. Therefore, we examined DC populations in Nfil3^−/− mice. The percentages of both the cDC (CD11c^hiPDCA-1⁻) and pDC (CD11c^intPDCA-1⁺) populations of Nfil3^−/− mice were slightly increased compared with that of Nfil3^+/+ mice (Figure 1B). Strikingly, the CD8α⁺ cDC population of Nfil3^−/− mice was significantly reduced compared with Nfil3^+/+ mice (Figure 1B,D). CD8α⁺ cDCs also express DEC-205, CD24, and SIRP-α,¹⁷  and we confirmed the lack of CD8α⁺ cDCs in Nfil3^−/− mice by costaining for expression of these markers (Figure 1B,D). Furthermore, the frequency of the precursor cells (CD8α⁻CD24^hiSIRP-α^lo) of CD8α⁺ cDCs in the spleen¹⁸  is also considerably reduced in Nfil3^−/− mice. CD8α⁺ cDCs can develop from T-cell precursors in the thymus; therefore, we determined whether a lack of CD8α⁺ cDCs was observed in the thymus.^2,19  The frequency of pDCs was increased in the thymus of Nfil3^−/− mice but that of cDCs was decreased considerably compared with Nfil3^+/+ mice (Figure 1C). Moreover, Nfil3^−/− mice lacked CD8α⁺DEC-205⁺CD24^hiSIRP-α^lo cDC cells in their thymus (Figure 1C-D). Because CD8α⁺ cDCs constitute approximately 70% of the cDC population in the thymus,²⁰  the lack of CD8α⁺ cDCs likely reflects the reduction of the cDC population in Nfil3^−/− mice. These results suggest that NFIL3 regulates the development of both splenic and thymic CD8α⁺ cDCs in vivo.

CD8α⁺ DCs produce IL-12 in response to stimulation of TLR3, TLR4, and TLR9 by their specific ligands.^17,21  Although TLR4 and TLR9 stimulation induced IL-12 in Nfil3^−/− splenic cDCs, TLR3 stimulation did not induce IL-12 in Nfil3^−/− cDCs (Figure 1E). These results suggest the lack of functional TLR3-expressing CD8α⁺ DCs in Nfil3^−/− mice.

In vitro BM culture with FL generates cDCs that are phenotypically and functionally equivalent to splenic CD8α⁺ and CD8α⁻ cDCs.¹⁷  We next determined whether the Nfil3 deficiency in BM cells affected FL-dependent DC development in vitro. Although both cDCs (CD11c⁺CD45RA⁻) and pDCs (CD11c⁺CD45RA⁺) were generated from Nfil3^+/+ and Nfil3^−/− BM cells, NFIL3 deficiency resulted in significantly fewer CD8α⁺-equivalent FL-cDCs (CD24^hiSIRP-α^lo) compared with Nfil3^+/+ BM cells (Figure 1F). FL-cDCs generated from Nfil3^−/− BM cells also failed to produce IL-12 by stimulation with TLR3, which suggests the impaired development of CD8α⁺-equivalent FL-cDCs in the absence of NFIL3 (Figure 1G).

Administration of recombinant FL or a B16 melanoma cell line expressing FL (B16-FL) to mice resulted in the strong induction of pDCs and CD8α⁺ cDCs in lymphoid tissues.^22,23  Despite the strong induction of splenic CD11c⁺ cells, B16-FL–injected Nfil3^−/− mice failed to produce CD8α⁺DEC-205⁺ cDCs (Figure 1H). Taken together, these results suggest that NFIL3 is required for FL-dependent CD8α⁺ cDC development in vivo and in vitro.

CD8α⁺ DCs can arise from different types of progenitor cells in BM and T-cell precursors in the thymus.^2,24,25  Interestingly, Nfil3^−/− mice have normal numbers of both cDC progenitors in BM (Figure 1I) and T-cell precursors in the thymus,¹³  which suggests that lack of CD8α⁺ cDCs is not due to a defect in progenitor cells. In contrast, Nfil3^−/− mice are deficient in the immediate splenic precursors of CD8α⁺ cDCs (Figure 1B). Therefore, NFIL3 may function during the development or expansion of pre-cDCs to the CD8α⁺-committed precursor stage.

To elucidate the mechanisms by which NFIL3 regulates CD8α⁺ DC development, pre-cDCs from Nfil3^+/+ and Nfil3^−/− mice were purified, and gene expression of transcription factors essential for CD8α⁺ DC development was examined. Expression of Batf3 but not Irf8 and Id2 was significantly reduced in Nfil3^−/− pre-cDCs, which suggests that BATF3 could be downstream of NFIL3 function (Figure 1J). Consistent with this interpretation, BATF3 transduction into Nfil3^−/− BM progenitors was able to substantially restore CD8α⁺ DC development (Figure 1K). These results suggest that NFIL3 regulates CD8α⁺ DC development in part through BATF3 expression, although other factors appear to be required to fully rescue CD8α⁺ DC deficiency in the absence of NFIL3.

An important function of CD8α⁺ DCs is the cross-presentation of cell-associated antigens to CD8⁺ T cells.²⁶  Therefore, we examined whether Nfil3^−/− mice exhibit an impaired ability to cross-prime antigen-specific CD8⁺ T cells. Injection of irradiated act-mOVA/K^b−/− splenocytes, which cannot directly present the OVA_257-264 epitope, into Nfil3^+/+ mice but not Nfil3^−/− mice resulted in induction of OVA_257-264-specific CD8⁺ T cells in both the blood (Figure 2A-B) and spleen (Figure 2C-E). These results suggest that Nfil3^−/− mice, which lack a CD8α⁺ DC population (DEC-205⁺SIRP-α^lo) in the spleen (Figure 2F), are defective in cross-priming antigen-specific CD8⁺ T cells against cell-associated antigens. Taken together, these results indicate that Nfil3^−/− mice lack functional CD8α⁺ DCs in the spleen.

Here, we demonstrated that Nfil3 expression is essential for FL-dependent CD8α⁺ cDC development in vivo and in vitro and that Nfil3^−/− mice exhibited an impaired capacity to cross-prime CD8⁺ T cells against cell-associated antigens. Thus, we conclude that NFIL3 plays a critical role in the development of CD8α⁺ cDCs, in addition to NK cells.^13-15  Both CD8α⁺ cDCs and NK cells can develop from T-cell precursors in the thymus, which are different lineages from BM-derived CD8α⁺ cDCs and NK cells.^19,27  Therefore, the lack of CD8α⁺ cDCs and NK cells in Nfil3^−/− mice suggests that NFIL3 could be important for the development of common progenitors of these cells. Control of the expression of the Nfil3 gene or its target genes in progenitor cells for CD8α⁺ cDCs and NK cells may provide new avenues to dissect the role of DC in regulating immune responses.

This work was supported by grants R01 AI54821 (P.B.R) and AI46653 and AI50073 (J.T.H) from the National Institutes of Health.

National Institutes of Health

Contribution: M.K., N.-L.L.P., and L.L.P. performed the experiments and analyzed data; M.K., J.T.H., and P.B.R. designed the experiments; and M.K. and P.B.R. wrote the paper.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Paul B. Rothman, Roy J. and Lucille A. Carver College of Medicine, University of Iowa, 212 CMAB, 451 Newton Rd, Iowa City, IA 52242; e-mail: paul-rothman@uiowa.edu.