Differential regulation of basal expression of inflammatory genes by NF-κB family subunits

The nuclear factor kappaB (NF-κB) family of transcription factors regulates immune and inflammatory responses and controls cell survival, proliferation, differentiation, and apoptosis.¹ The five subunits of the NF-κB family, (RelA, RelB, c-Rel, p105/p50, and p100/p52), are categorized as activating (RelA, RelB, and c-Rel), or inhibitory (p105/p50 and p100/p52), based on the presence or absence of the transactivation domain and form homo- and heterodimers. Stimulus-dependent activation of NF-κB, either by the classical pathway involving the degradation of the inhibitor of NF-κB (IκB) or by the alternative pathway involving the partial degradation of p100 to p52 and formation of the p52/RelB dimer, has emerged as a central theme in the regulation of NF-κB-dependent genes.^2–4 However, in Sertoli cells and some neurons, lymphocytes, and malignant cells, NF-κB has been shown to be constitutively active in the nucleus^5,6 (www.bu.edu/nf-kb). Moreover, several recent studies have also identified disease-associated mutations in the NF-κB subunits that affect their expression level and transcriptional activity.^2,7 Our understanding of the ways in which the changes in the basal level of NF-κB affect the subsequent basal regulation of the expression of NF-κB-dependent genes remain poorly defined.

In this study, we conducted a preliminary exploration of the role that each NF-κB subunit plays in the constitutive or basal expression of selected NF-κB-dependent genes. We examined the expression levels of several genes associated with inflammation in wild-type mouse embryonic fibroblast cells (MEFs), which act as sentinel cells in the inflammatory response,⁸ and compared these levels to those in knockout cells for each NF-κB subunit. We assessed the relative mRNA abundance to represent the actual amount of each gene transcript present in unstimulated cells. We summarized our findings for use as a basis for further understanding NF-κB-mediated inflammatory pathologies⁹ and targeted future investigations of NF-κB-dependent transcriptional regulation.

Here, we present the gene expression data categorized based on the deficiency of the transcriptionally activating RelA, RelB, and c-Rel subunits (Fig. 1) and transcriptionally suppressive NF-κB p50 and p52 subunits (Fig. 2). First, we confirmed the deficiency of the Rel subunits in the respective cell lines by western blotting (Fig. 1a), and we then measured the relative increase or decrease in the mRNA expression of each gene in NF-κB subunit knockout cells compared to wild-type cells. As summarized in the Venn diagram (Fig. 1b), we found that a substantial number of genes showed enhanced basal expression in both RelB-deficient cells (20 of 33) and c-Rel-deficient cells (19 of 33), while fewer genes (10 of 33; A20, CJUN, CXCL10, IRF1, CCL5, MMP10, IFIT1, ZFP36 (Fig. 1c), IKBE, and TNF (Fig. 1g)) showed enhanced expression in RelA-deficient cells. Enhanced expression of these ten genes overlapped either RelB or c-Rel deficiency (Fig. 1b left), and no genes were found to exhibit enhanced expression in RelA-deficient cells that were solely regulated by RelA (Fig. 1b left). We also found several genes (7 of 33; SAA3, C3, BIRC3, CCL20, CCL7, CFOS, and BIMS) with basal expression levels predominantly dependent on RelA because these genes were not significantly downregulated in RelB or c-Rel knockout cells (Fig. 1d), while few other genes (3 of 33; CXCL1, CXCL2, and EDN1) showed partial dependence on RelA (Fig. 1e). Overall, only 15 of 33 genes were found to be downregulated in cells with knockout of any one of the transactivating subunits (Fig. 1b, right). Interestingly, the expression levels of five genes—VCAM1, IL-4, MMP3, Serp3G, and IKBA—were consistently found to decrease when RelA, RelB, or c-Rel were individually knocked out, suggesting that their basal expression depends on all three transactivating subunits (Fig. 1f). While no genes were found to be uniquely increased in RelA-deficient cells, genes predominantly regulated by either RelB or c-Rel were identified: the expression levels of IL-6, IKBE, TNF (Fig. 1g), and CXCL2 (Fig. 1e) were significantly enhanced in RelB knockout cells, and those of ZFP36 (Fig. 1c) XIAP, ICAM1 (Fig. 1h), and BIRC3 (Fig. 1d) were significantly enhanced in c-Rel knockout cells. Genes such as CCL2, MCSF, MMP13 (Fig. 1i), SAA3, C3, CCL20, CCL7, CFOS, and BIMS (Fig. 1d) were significantly upregulated both in RelB and c-Rel knockout cells, and no significant changes were observed in cFLIP and BCLXL expression (Fig. 1j).

Fig. 1.

a Total cell lysates of wild-type and RelA, RelB and c-Rel knockout MEFs were analyzed by western blotting as indicated. Antibodies against RelA, RelB, and β-actin were purchased from Santa Cruz Biotechnology, and antibodies against c-Rel were obtained from Santa Cruz Biotechnology and Biolegend. b–j Real-time PCR analysis of inflammatory genes in unstimulated wild-type and RelA, RelB, and c-Rel knockout MEFs. b Venn diagram summarizing the gene expression data generated using Integrated Omics’ Venn Plotter software. c–j Relative mRNA abundance of the indicated genes. qPCR was performed as previously described.¹⁰ All expression values were quantified relative to the expression of the housekeeping gene ribosomal protein L32. The data are representative of at least three independent experiments, each performed in duplicate or triplicate, and are presented as the means ± standard errors of the mean (SEMs). p values were obtained by unpaired Student’s t test; ****p < 0.0001, ***p < 0.001, **p < 0.01, and *p < 0.05

Fig. 2.

a Total cell lysates of wild-type and p105/p50 and p100/p52 knockout cells were analyzed by western blotting as indicated. Antibodies against p100/p52 and p105/p50 were obtained from Cell Signaling Technology. b Real-time PCR analysis of inflammatory genes in unstimulated wild-type and p105/p50, and p100/p52, knockout fibroblasts. b Venn diagram summarizing the gene expression data. c–h Relative mRNA abundance of the indicated genes. i Expression levels of the indicated genes in three different wild-type MEFs

We also examined how the deficiency of NF-κB subunits lacking a transactivation domain, i.e., p50 and p52 (Fig. 2a), influenced the basal expression of the genes studied here. As summarized in the Venn diagram (Fig. 2b), the basal expression levels of only two genes (CCL20 and ICAM1) and only six genes (cFLIP, Serp3G, CFOS, XIAP, BCLXL, and BIRC3) were uniquely enhanced in p50 (Fig. 2c) and p52 (Fig. 2d) knockout cells, respectively. However, a substantially larger number of genes (13 of 33; CCL5, ZFP36, A20, SAA3, CXCL10, CXCL2, MMP10, IFIT1, IL6, CJUN, MMP13, BIMS, and MCSF) showed overlap in enhanced expression, indicating that p50:p52 heterodimers may act as the primary suppressive NF-κB dimers in fibroblasts (Fig. 2b, e). Interestingly, although p50 and p52 are categorized as repressive subunits, we found that their deficiency decreased the expression of ten selected genes, suggesting their role in the basal expression of these genes, most likely by dimerizing with an activating subunit (Fig. 2b). Serp3G showed a unique dependence on p50, as its expression was found to decrease in p50 knockout cells (Fig. 2b, d). The expression levels of IRF1, CCL2, and CCL7 were uniquely decreased in p52 knockout cells (Fig. 2b, f). Other genes, such as CXCL1, IKBE, C3, VCAM1, MMP3, and EDN1, showed decreased expression when either p50 or p52 was deleted (Fig. 2b, g). No significant changes in basal expression were observed for 3 of the 33 genes (IKBA, TNF, and IL4) in either p50 or p52 knockout cells (Fig. 2c, h).

Overall, we observed a greater role for the individual NF-κB subunits in suppressing the basal level of inflammatory gene expression, as the number of genes that showed enhanced expression in single-knockout cells was higher than the number of genes that showed decreased expression. Interestingly, CCL5 was found to be the only gene showing universal enhancement in the absence of each subunit, indicating that all five NF-κB subunits are somehow involved in its basal suppression. This pattern is unique among other cytokines in its class, such as CCL2, CCL7, CXCL1, and CXCL2. While this effect could be specific to fibroblasts, it warrants further investigation. However, the expression of VCAM1 and MMP3 was found to be significantly decreased when any one of the NF-κB subunits was deleted, suggesting that each subunit has a nonredundant role in regulating the basal expression of these genes. This five-way redundancy in the expression of these genes suggests that further detailed transcriptional studies may reveal NF-κB binding sites and patterns that would ensure targeting by all five subunits and, consequently, the fifteen different dimers of NF-κB (2).

To avoid biased conclusions based on comparisons of different knockouts to single wild-type cells and to ensure scientific rigor, we compared three different wild-type fibroblast cells and found no significant differences in basal gene expression in these cells (Fig. 2i). The current study is limited by uncontrollable genetic and developmental differences between the mouse lines from which the immortalized cell lines were derived. Our future studies will include studying this basal regulation by NF-κB proteins in hematopoietic and nonhematopoietic cell lines using CRISPR/Cas9-mediated knockout of individual subunits to ensure that each of the knockouts are compared with the same parental cell line. In addition, comprehensive biochemical experiments are necessary to demonstrate the type of NF-κB dimers actively involved in the basal transcription of inflammation-associated genes and how the IκB proteins regulate these dimers.

In conclusion, these results suggest that genetic polymorphisms or other factors that may affect the expression or function of the individual NF-κB family subunits may alter the basal expression of several proinflammatory genes. To maintain homeostasis and prevent a chronic inflammatory state or unnecessary inflammatory responses, the transcription of proinflammatory genes must be tightly regulated at a basal level. Changes in the basal transcription of inflammatory genes may lead to either an increased or decreased initial threshold for the inflammatory response, which may result in excessive or insufficient levels of inflammatory mediators, both of which are detrimental to the host.

Author contributions

T.J.D. planned and performed the experiments, analyzed and interpreted the data, and wrote the manuscript. J.T.C. performed the experiments and analyzed and interpreted the data. P.R. conceived the project, planned the experiments, analyzed the data, and edited the manuscript.

Competing interests

The authors declare no competing interests.