RFX3 Modulation of FOXJ1 regulation of cilia genes in the human airway epithelium

Subjects were evaluated at the Department of Genetic Medicine Clinical Research Facility under the auspices of Weill Cornell and Rockefeller University NIH Clinical Translational Science Centers, using Weill Cornell Institutional Review Board clinical protocols approved for this study. Written informed consent was obtained.

Bronchoscopy was used to collect large airway epithelial cells by brushing the epithelium. The complete differentiated airway epithelium from 5 healthy, nonsmoking individuals was evaluated, and for 14 other individuals, the epithelium was cultured under conditions to obtain pure populations of basal cells as previously described [32‐34] (see Additional file 1: Additional Methods for more detail).

The basal cell culture and immunohistochemistry characterization protocol have been previously described [32]. Cells were sub-cultured at day 7 to a density of 10⁴ cells/cm². Cells from passage 1 to 5 were characterized and used in this study (see Additional file 1: Additional Methods for more detail).

TaqMan real-time RT-PCR was performed as previously described [35]. Relative expression levels were determined with the average value of untransfected or EGFP-control plasmid transfected basal cells as the normalizer (see Additional file 1: Additional Methods for more detail).

Basal cell cultures and exogenous FOXJ1 and RFX3 expression were assessed by Western analysis as previously described [35]. Immobilized proteins were reacted with anti-KRT5; anti-KRT14; anti-MUC5AC; anti-DNAI1; anti-acetylated TUBA; anti-FOXJ1; anti-RFX3 and anti-GAPDH antibodies (see Additional file 1: Additional Methods for more detail).

To assess expression profiles during the differentiation of human basal cells to ciliated cells, ciliated cell differentiation was induced from basal cells (n = 3 subjects) in vitro using air-liquid interface (ALI) cultures as previously described [32] (see Additional file 1: Additional Methods for more detail).

Human FOXJ1 and RFX3 cDNA were subcloned into expression plasmids to generate PGK.FOXJ1.IRES.EGFP, PGK.RFX3.IRES.EGFP, CMV.FLAG-FOXJ1 and CMV.FLAG-RFX3 expression plasmids. PGK.EGFP and CMV.EGFP expression plasmids were used as a control expression plasmid, respectively. Firefly luciferase (Luc) reporter gene plasmids driven by the direct upstream promoters of DNALI1, SPAG6, KRT14 and FOXJ1 were generated using standard cloning methods. DNAI1-Luc, TEKT1-Luc, RFX3-Luc, RFX2-Luc and Random sequence-Luc reporter gene plasmids were commercially available. A Renilla luciferase control reporter plasmid was used for normalization of transfection efficiency. The plasmids were transfected into primary human airway epithelial basal cells using lipofectamine LTX and promoter firefly luciferase activity was read in a luminometer. The data are reported as fold-induction (FOXJ1 compared to EGFP or FOXJ1 + RFX3 compared to FOXJ1) of at least three independent experiments read in triplicate, normalized to Renilla luciferase activity (see Additional file 1: Additional Methods for more detail).

To assess the interaction of human FOXJ1 and RFX3, 293A cells were transfected with PGK.FOXJ1 and CMV.FLAG-RFX3 expression plasmids. Proteins were immunoprecipitated using EZview Red Anti-FLAG M2 affinity gel and eluted with FLAG peptide (see Additional file 1: Additional Methods for more detail).

All data in this study are presented as mean ± standard error. Statistical comparisons between continuous variables were calculated using an unpaired, two-tailed t test with unequal variance. A p-value <0.05 was considered to be significant (see Additional file 1: Additional Methods for more detail).

To test the hypothesis that FOXJ1 is a key regulator of the motile multiciliated cell differentiation program in the human airway epithelium, we first established primary human airway epithelial basal cell cultures [32]. After 7 days in culture, immunohistochemistry of cytospin preparations demonstrated the cells expressed the basal cell marker cytokeratin (KRT) 5, but did not express the ciliated cell marker β-tubulin IV, the secretory cell marker mucin 5AC or the mesenchymal cell marker N-cadherin (Figure 1A). TaqMan quantitative real-time RT-PCR assessment of the basal cell markers KRT5, KRT14 and tumor protein p63 (TP63) confirmed significantly elevated mRNA levels of these markers in basal cell cultures compared to large airway epithelium recovered by airway brushing (Figure 1B). Expression of the basal cell markers KRT5 and KRT14 and a lack of the cilia markers dynein, axonemal, intermediate chain 1 (DNAI1) and acetylated α-tubulin and the secretory cell marker secretoglobin family 1A member 1 (SCGB1A1) expression were confirmed by Western analysis of basal cell cultures compared to the complete large airway epithelial cell population (Figure 1C).

Assessment of expression of the putative cilia-associated transcriptional regulators FOXJ1, NOTO and RFX3 during human airway basal cell to ciliated cell differentiation on air liquid interface cultures (ALI) in vitro demonstrated that the expression pattern over time for both FOXJ1 and RFX3 was characteristic of genes participating in the ciliated cell differentiation program [36]. In contrast, NOTO, a suggested FOXJ1-upstream master regulator in mouse notochord ciliogenesis [16], was not expressed at any time-point (Figure 1D). Interestingly, NOTO was also not expressed in the complete airway epithelium, suggesting that the signaling network regulating airway epithelial motile ciliogenesis in the human airway epithelium is somewhat different than notochord ciliogenesis in mice and zebrafish (not shown).

To address the role of FOXJ1 in human airway epithelial cell differentiation, we first assessed the efficiency by which FOXJ1 expression could be induced by plasmid-mediated gene transfer of a FOXJ1-expression plasmid to primary human airway basal cells. TaqMan quantitative real-time RT-PCR demonstrated a dose-dependent relationship between the amount of transfected FOXJ1 plasmid and the mRNA levels of FOXJ1 with 10⁴-times higher FOXJ1 mRNA levels in FOXJ1 (1000 ng) expression plasmid transfected basal cells compared to untransfected (mock) or EGFP (1000 ng) transfected basal cells (Figure 2A). The transfection protocol typically yielded 55-65% transfection efficiency. Western analysis of FOXJ1 and GAPDH on untransfected basal cells and basal cells transfected with either the EGFP control or FOXJ1 expression plasmids confirmed a robust FOXJ1 expression in FOXJ1 transfected basal cells (Figure 2B).

To investigate whether gene transfer of FOXJ1 to basal cells under conditions that normally prevent cell differentiation was sufficient to induce differentiation towards the multiciliated cell lineage, TaqMan quantitative real-time RT-PCR was used to assess EGFP or FOXJ1 expression plasmid transfection of basal cells maintained under standard basal cell culture conditions. FOXJ1 transfection had no effect on the mRNA levels of the basal cell markers KRT5, KRT14 and TP63 or expression of the secretory cell markers MUC5AC, MUC5B and SCGB1A1, not normally expressed in basal cell cultures (Figure 3A, B). In contrast, FOXJ1 expression in basal cells induced the expression of a panel of cilia-associated genes, including centrin 2 (CETN2); dynein, axonemal, heavy chain 11 (DNAH11); dynein, axonemal, intermediate chain 1 (DNAI1); dynein, axonemal, light intermediate chain 1 (DNALI1); EF-hand domain, C-terminal, containing 1 (EFHC1); sperm associated antigen 6 (SPAG6); tektin 1 (TEKT1), TEKT2 and tubulin, alpha 1a (TUBA1A; Figure 3C and Additional file 2: Table S1). Consistent with this data, co-transfection of the control EGFP or the FOXJ1 expression plasmids with plasmids carrying the firefly luciferase reporter gene driven by promoters of the ciliated cell-associated genes DNAI1, DNALI1, SPAG6 and TEKT1 demonstrated that FOXJ1 significantly induced promoter activity from the ciliated cell-associated genes while promoter activity from the basal cell specific gene KRT14 or a random sequence remained unaffected (Figure 3D). We also attempted to assess gene expression in the human samples with the available antibodies, but the antibodies did not work well, with weak staining and lack of specificity.

To investigate the role of cilia-associated transcriptional regulator RFX3 in human airway epithelial motile ciliogenesis, TaqMan quantitative real-time RT-PCR was used to assess RFX3 expression in basal cells transfected with FOXJ1 expression plasmid and RFX3 promoter activity was assessed by co-transfection with a plasmid carrying the firefly luciferase reporter gene driven by the RFX3 promoter (Figure 4A, B). The data demonstrated increased RFX3 promoter activity and mRNA expression in cells transfected with the FOXJ1 expression plasmid compared to cells transfected with the control EGFP plasmid, suggesting a role for RFX3 downstream of FOXJ1 in human airway epithelial ciliogenesis. To investigate whether RFX3 could, like FOXJ1, induce differentiation from the basal cell lineage towards the multiciliated cell lineage, a RFX3 expressing plasmid was transfected into the basal cells. While RFX3 expression was demonstrated by TaqMan quantitative real-time RT-PCR and Western analysis in basal cells transfected with a RFX3 expression plasmid (Figure 4C, D), none of the wide panel of FOXJ1-induced cilia associated genes or the basal and secretory cell markers were induced in RFX3 transfected basal cells (Additional file 3: Figure S1A-C). These data suggest that although FOXJ1 induces RFX3 expression in human airway epithelial cells, RFX3 expression alone is not sufficient to drive the human basal cell to multiciliated cell differentiation process.

RFX3 has an established role in the ciliogenetic process in animal models [16, 17, 20, 24, 30], and although RFX3 alone does not seem to induce human airway epithelial differentiation (Additional file 3: Figure S1A-C), we questioned whether RFX3 could play a role in human airway ciliogenesis by mediating FOXJ1-induced multiciliated cell differentiation. Interestingly, RFX3 significantly enhanced the FOXJ1-induced mRNA expression of the ciliated cell associated genes DNAI1, DNALI1, SPAG6, TEKT1 and TEKT2 (all p < 0.004) in basal cells transfected with FOXJ1 and RFX3 expression plasmids together compared to basal cells transfected with FOXJ1 and EGFP expression plasmids together (Figure 5A). In addition, the expression of these cilia associated genes were induced in a dose-dependent manner by FOXJ1 (all p < 0.04) in basal cells transfected with an increasing dose of FOXJ1 expression plasmid (Figure 5A), demonstrating a tight FOXJ1-RFX3 mediated regulation of these genes. Expression of the basal cell markers KRT5, KRT14 and TP63 remained unaffected (Figure 5B). Consistent with this data, promoter activity of the cilia-associated genes DNAI1, DNALI1, SPAG6 and TEKT1 was enhanced (p < 0.02) in basal cells transfected with FOXJ1 and RFX3 together compared to cells transfected with FOXJ1 and EGFP together when co-transfected with the respective firefly luciferase reporter plasmids (Figure 5C). Promoter activity of the basal cell marker KRT14 and a random sequence were unaffected by FOXJ1 and RFX3 co-expression (Figure 5C). These data suggest that RFX3 alone is not sufficient to induce human airway basal cells to differentiate to multiciliated cells, but rather acts as a transcriptional co-activator to FOXJ1 that magnifies the induction of cilia associated genes.

Western analysis was used to address whether RFX3 interacts with FOXJ1 via protein-to-protein interaction. Immunoprecipitation with anti-FLAG affinity beads of functionally intact (not shown) FLAG-tagged RFX3 co-precipitated FOXJ1 in 293A cells transfected with FLAG-RFX3 expression plasmids together with FOXJ1 expression plasmids, compared to mock, FOXJ1, or FLAG-RFX3 alone transfected cells (Figure 6). Analysis of input lysates before immunoprecipitation using antibodies against FOXJ1 and RFX3 revealed robust expression of each in the appropriate lysate. Detection of GAPDH confirmed an equal amount of total protein in each (Figure 6). These data demonstrate that RFX3 can interact and function as a co-activator to FOXJ1 via direct protein-to-protein interaction between the two factors or indirect interaction within an intracellular protein complex.

To address whether or not FOXJ1 and RFX3 expression is self-regulated in human airway epithelial cell differentiation, we assessed the RFX3 and FOXJ1 promoter activity in basal cells transfected with FOXJ1 and/or RFX3 expression plasmid and co-transfection with plasmids carrying the firefly luciferase reporter gene driven by either the FOXJ1 or the RFX3 promoter (Figure 7A). Corroborating the FOXJ1 induced RFX3 mRNA expression (Figure 4A), both FOXJ1 and RFX3 promoter activity were induced 6.0- (p < 0.01) and 7.9-fold (p < 0.02) in basal cells transfected with FOXJ1 expression plasmid compared to cells transfected with the control EGFP expression plasmid. RFX3 co-expression further enhanced the FOXJ1-induced activity of the FOXJ1 promoter 2.9-fold (p < 0.008) and the RFX3 promoters 2.7-fold (p < 0.008). In contrast to the cilia-associated genes, RFX3 expression was sufficient to induce both FOXJ1 2.2-fold (p < 0.008) and RFX3 promoter activity 2.4-fold (p < 0.05) when expressed alone, although at lower levels than FOXJ1. RFX3 did not, however, significantly induce endogenous FOXJ1 mRNA expression in basal cells transfected with RFX3 expression plasmid alone (Figure 7B). In summary, these data suggest that both FOXJ1 and RFX3 can regulate their own expression to different degrees and that FOXJ1 is the major inducer of a positive auto-regulatory feedback mechanism where FOXJ1 is upstream of RFX3.

RFX-family members other than RFX3 have also been implicated in cilia homeostasis in various model organisms [18, 20, 26, 37]. Assessment of expression of the different RFX-transcription factor members during human airway basal cell to ciliated cell differentiation on air liquid interface cultures in vitro demonstrated that the expression patterns over time for RFX2 were also characteristic of genes participating in the ciliated cell differentiation program [36], in similarity with FOXJ1 and RFX3 (Figure 1D, Additional file 4: Figure S2A). We further assessed the ability of FOXJ1 and /or RFX3 to induce endogenous RFX2 mRNA expression in basal cells transfected with FOXJ1 and /or RFX3 expression plasmids, as well as induced RFX2 promoter activity in basal cells co-transfected with FOXJ1 and /or RFX3 expression plasmids and a plasmid carrying the firefly luciferase reporter gene driven by the RFX2 promoter (Additional file 4: Figure S2B,C). The data demonstrated increased RFX2 promoter activity (p < 0.0005) and mRNA expression (p < 0.04) in cells transfected with the FOXJ1 expression plasmid alone and no significant induction of either RFX2 promoter activity or mRNA expression in cells transfected with RFX3 alone, compared to cells transfected with the control EGFP plasmid. Cells transfected with RFX3 seemed to have elevated RFX2 mRNA expression, although to a variable degree and not significant. When RFX3 was co-transfected with FOXJ1, however, the FOXJ1 induced RFX2 promoter activity (p < 0.03) was further enhanced, suggesting a similar and potentially partially overlapping role for RFX2 and RFX3 downstream of FOXJ1 in human airway epithelial ciliogenesis.

Much of our knowledge of motile monocilium and multiciliated cell differentiation is based on studies of the mono motile ciliated cells of the mouse node and its homologues in zebrafish and xenopus [16‐19]. FOXJ1 is one of the most well characterized transcription factors involved in motile ciliated cell differentiation, where FOXJ1 is known to regulate programs promoting basal body docking and axoneme formation [19‐29]. Also the RFX-family of transcription factors has been shown to be a major regulator of ciliogenesis in multiple organisms, controlling the expression of the many essential genes required for making cilia [38‐42] and has recently been shown to co-regulate cilia gene expression together with FOXJ1 in Drosophila [31]. Beckers et al [16] demonstrated evidence that the murine nothochord motile monocilium differentiation program is regulated by the transcription factor NOTO, acting upstream of both FOXJ1 and RFX3. The novel nuclear protein multicilin was also recently shown to drive the expression of both FOXJ1 and multiciliated cell specific genes in frog and mouse multiciliated epithelia [43]. Several investigators, using murine models and transformed human cell lines, have demonstrated that FOXJ1 is a key regulator of airway epithelial ciliated cell differentiation [21, 22, 27, 29, 44‐50]. In the airway epithelium, Notch signaling had been demonstrated to be the most upstream cue for transcriptional events during the differentiation of multiciliated and other airway epithelial cell types [41].

Murine models of airway development and cellular differentiation have been critical to our basic understanding of the mucociliary clearance apparatus during development or after injury [21, 51], but the human airways differ from those of mice to some extent [13]. The networks regulating cell commitment in the airway epithelium in humans are therefore likely to differ somewhat from those of mice and other animal models.

The human FOXJ1 gene was first cloned over 10 yr ago [52, 53], yet the function of FOXJ1 in the human airway epithelium remains relatively unknown. Studies have addressed whether mutations in the FOXJ1 gene could be linked to primary ciliary dyskinesia [54, 55], although this hypothesis has not been supported. The activity of the human FOXJ1 promoter has been demonstrated to be specific to ciliated cells in the mouse airway epithelium in vivo[56]. The mouse FOXJ1 promoter has been shown to be active in primary human airway epithelial cells and the human airway epithelial cell line H441, but not in the human fibroblast cell line HT1080 [57]. The murine and human FOXJ1 gene have been transfected into the transformed human airway epithelial BEAS-2B cell line and in the human embryonic kidney (HEK-293) cell line [21, 45, 57, 58]. Expression of murine FOXJ1 in BEAS-2B cells failed to induce cilia development [21], but consistent with the findings in the present study, murine FOXJ1 induced SPAG6 promoter activity in these cells [58].

FOXJ1 and RFX3 are highly expressed in the small airway epithelium [32], consistent with the concept that FOXJ1 and RFX3 together may play a pivotal role in maintaining homeostasis of the multiciliated cell population in the adult human airway epithelium. In the current study, we found that NOTO is not expressed in the human airway epithelium nor is its expression induced by ALI-mediated airway epithelial differentiation in vitro. Hence, in contrast to murine notochord ciliogenesis, FOXJ1 and RFX3 are not regulated upstream by NOTO in the human airway epithelium. We can only speculate on the identity of these players, but possible candidates include Notch, multicilin, FOXA2 and SOX2 [41, 43, 59‐61].