Macrophages are innate immune cells that exhibit remarkable phenotypic plasticity, adopting contrasting polarization states over the course of an infection to best contribute to tissue defense and repair (1). During an infection’s opening stages, exposure to interferon-γ (IFNγ) and/or pathogen-associated molecular patterns (PAMPs), such as lipopolysaccharide (LPS), promotes M1 polarization of naïve (M0) macrophages (2). In this state, macrophages secrete pro-inflammatory cytokines and exhibit heightened antimicrobial activity, partially driven by inducible nitric oxide synthase (iNOS) expression, which produces nitric oxide (NO) as a precursor for microbicidal reactive nitrogen species (RNS) (3–5). As an infection resolves, rising levels of interleukin (IL)-4 and IL-13 repolarize macrophages to the M2 state, which is associated with tissue repair and anti-inflammatory cytokine secretion. Additionally, the production of cytotoxic RNS is attenuated through arginase-1 (Arg1) expression, which competes with iNOS for arginine, the common substrate of both enzymes (6).

The transition between polarization states involves broad transcriptional reprogramming affecting >1,000 genes (7, 8). The molecular mechanisms governing these changes are of particular importance to human health as the failure to downregulate pro-inflammatory gene expression programs or activate them results in adverse outcomes. Specifically, inappropriate M1 activation contributes to inflammatory disorders, such as atherosclerosis, asthma, diet-induced obesity, and insulin resistance (9–13). Conversely, intracellular pathogens that colonize macrophages, including Toxoplasma gondii and Salmonella enterica serovar Typhimurium, utilize effectors that target these same mechanisms to block M1 or promote M2 polarization of host cells, creating a more permissive environment for growth and persistence (14–19).

While multiple transcription factors, including Krüppel-like factors (KLFs), hypoxia-inducible factor 1 (HIF1), and p53 (20–27) contribute to gene expression changes during polarization, the nuclear factor kappaB (NF-κB) and Janus kinase (JAK)/signal transducer and activation of transcription (STAT):interferon regulated factor (IRF) pathways are considered the primary controllers (28–30). Each is regulated by their own pathway-specific feedbacks that amplify or attenuate the response to polarizing stimuli. These include reciprocal positive feedback loops between Stat1 and Irf1 (31), and the various negative feedbacks (i.e., inhibitor-kappaB (IκB)α, ϵ, and A20) that prevent premature or prolonged NF-κB activity (32–39). However, these are supplemented by feedbacks that impact the activity of multiple polarization regulatory pathways, providing a higher level of control and coordination (40). For example, suppressor of cytokine signaling 1 (SOCS1) attenuates STAT1 activity by inhibiting JAKs and blocking NF-κB activity through recruitment to specific NF-κB-regulated gene promoters, stimulating nuclear p65 (RelA) NF-κB protein degradation (41–43). In this review, we will bring together recent findings showing how the CBP/p300-interacting transactivator with glutamic acid/aspartic acid-rich carboxy-terminal domain (CITED) family of transcriptional co-regulators contribute to this broader level of control. Specifically, we will describe how CITED1 and 2 proteins operate as common regulators of NF-κB, HIF1, IRF1, and the STAT family of transcription factors and explore how the antagonistic activities of these proteins shape the transcriptional response of macrophages to polarizing signals.

Originally named melanocyte-specific gene 1 (msg1) following its discovery in highly pigmented B16-F1 murine melanoma cells, Cited1 was the first member of the Cited gene family to be described (44), with the same study also identifying Cited2, initially dubbed msg-related gene 1 (mrg1) based on sequence homology. This was quickly followed by Cited3, present in avian, amphibian, and fish species (45, 46) but absent in mice and humans, and Cited4, the only other family member expressed in mammalian systems (47, 48). These genes participate in different and diverse biological processes, such as melanocyte pigmentation (44), placentation (49), embryonic development (50), and differentiation (48). They are even implicated in different sets of pathological conditions, including congenital heart disease and cancer (51–64). In all of these biological processes, they function as transcriptional co-regulators (48, 65, 66).

As transcriptional co-regulators, CITED proteins lack DNA binding domains and are not known to form dimers or multimers. Instead, these proteins contain sets of conserved region (CR) motifs including the centrally located CR1 motif (~14 aa) present in CITED1, 2, and 3, a shorter CR3 motif (~6 aa) at the N-terminus of CITED2, 3, and 4, and the longer C-terminal CR2 domain (48-55 aa) common to all CITED proteins ( Figure 1A ) (47, 65). Outside these regions, CITED proteins exhibit minimal sequence similarity and are largely unstructured ( Figure 1B ), but can contain protein interaction motifs, such as the Smad-interacting domain (SID) identified in CITED1 (65).

While no clear function has been ascribed to CR1 and 3, the CR2 domain is essential for transcriptional co-regulatory activity (65, 69). Analogous to the transactivation domains (TADs) of STAT, NF-κB, and HIF1 proteins, CR2 enables CITED proteins to associate with one or more of the three cysteine-histidine-rich (CH) domains within the histone acetyltransferase (HAT), CREB binding protein (CBP) and its paralogue, p300 (65, 70–72) ( Figure 2A ). In this way, CITED proteins can inhibit or facilitate the formation of transcription factor:CBP/p300 complexes at gene cis-regulatory sites by blocking access to specific CH domains or functioning as a co-factor for transcription factors that weakly associate with CBP/p300 (66, 73–76) ( Figures 2B, C ).

The first indication that CITED proteins function as transcriptional co-repressors was provided by Bhattacharya et al. (77). In this study, p35srj, a CITED2 isoform containing a 39 aa serine-glycine-rich junction (SRJ) between CR1 and 2 ( Figure 1A ), was identified in a screen for proteins interacting with p300-CH1 (TAZ1). Endogenous p35srj was shown to be constitutively nuclear in U2OS human osteosarcoma cells with almost the entire cellular pool found in CBP/p300 complexes, and the minimal binding region necessary for p300-CH1 interaction completely contained within the CR2 domain (224-255 aa). This study also showed that p35srj could inhibit HIF1-dependent gene transactivation by competing with the C-terminal TAD (CAD) of HIF1α for CH1 binding ( Figures 2A, B ). This is consistent with subsequent work showing synthetic membrane-permeable peptides, based on the CITED2 CR2 domain, inhibit HIF1α:p300 complex formation and HIF1-dependent gene expression (73).

One of the more important outcomes of the Bhattacharya study was the demonstration that CITED2 preferentially binds the CH1 domain of CBP/p300, distinguishing it from CITED1, which primarily associates with CH2 (65) ( Figure 2A ). Specifically, Bhattacharya et al. showed that p35srj did not reduce the activity of transcriptional regulators binding p300 outside the CH1 domain, including SREBP2 and src-1, which associate with the CREB-binding domain (KIX) and glutamine-rich domains of CBP/p300, respectively (78, 79). This result was supported by subsequent studies showing that CITED2 blocks the binding of the HIF1α N-terminal TAD (NAD) to p300-CH1 but not CH3 (74). More recently, NMR spectroscopy was used to identify the specific CH1 binding surface utilized by CITED2 and it was found that while the CITED2 CR2 domain and the HIF1α CAD bind to CH1 with similar affinity, they occupy distinct but overlapping binding positions (70). Collectively, these data indicate that CITED2 proteins may function as general inhibitors of transcription factors that use the CH1 domain for docking with CBP/p300.

While Cited2 expression is ubiquitous, Cited1 is largely restricted to a small number of tissues (113), and was thought to be absent in macrophages (85). This perception was changed in 2020 when Cited1 transcripts were detected in M1-polarized RAW264.7 murine macrophage-like cells infected with the fungal pathogen, Cryptococcus neoformans (114). It was subsequently determined that exposure of these cells to IFNγ for periods greater than 24 hours stimulated expression of CITED1 proteins, regulated by both STAT1 and IRF1 at the transcriptional level. Cited1 expression was unaffected by other M1 polarizing stimuli (i.e., LPS), indicating that it operates as an interferon-stimulated gene (ISG) (113). This pattern of regulation is in stark contrast to Cited2, which is transiently downregulated following short periods of exposure (6 hours) to M1 polarizing stimuli (LPS or IFNγ) and upregulated by M2 stimuli (IL-4 or IL-13) (85), indicating that the circumstances under which CITED1 and 2 operate may be minimal or non-overlapping.

A further regulatory difference between CITED1 and 2 is their subcellular localization under varying polarizing conditions. While CITED2 has been shown to be constitutively nuclear (97), CITED1 is predominantly cytoplasmic in unstimulated cells but becomes enriched in the nucleus following IFNγ stimulation (113). Quite how this shift in subcellular distribution is controlled is presently unclear, but it may be connected to the increased phosphorylation of the protein observed in IFNγ-stimulated macrophages as CITED1 phosphorylation at Ser79 following parathyroid-hormone-stimulation in murine osteoblastic cells is known to be associated with its nuclear translocation (115, 116).

The differences between CITED1 and 2 go beyond the regulation of their expression and localization and extends to their function with the two transcriptional regulators exerting opposing effects on macrophage pro-inflammatory gene expression. This was shown by Subramani et al. using a complementary loss- and gain-of-function approach, where it was found that CITED1 enhances STAT1- and IRF1-regulated gene expression in IFNγ-stimulated macrophages, increasing transcript levels of Ccl, Isg, and Ifit family members amongst many others (113). These data suggest that CITED1 is functioning as a transcriptional co-enhancer ( Figure 2C ), versus the largely co-inhibitory effects of CITED2 in the same context. This is perhaps unsurprising as (i) neither STAT1 or IRF1 bind CBP/p300 via CH2, which is thought to be the preferred CITED1 interaction site (72, 117), and (ii) CITED1 has been shown to function as a co-enhancer in other contexts (65, 69, 118). For example, CITED1 enhances Smad4-dependent gene expression by stabilizing CBP/p300:Smad4 complexes, interacting with both proteins (65). Similarly, CITED1 acts as a co-factor for the assembly of STAT3:estrogen receptor-α (ERα) transcriptional complexes in neurons as part of leptin signaling, contributing to whole-body energy homeostasis (118). This is of relevance as STAT3 contributes to IFNγ-stimulated gene expression in macrophages (119, 120). It also illustrates the close relationship that both CITED1 and 2 share with the STAT family of transcriptional regulators.

CBP/p300 serves as a convergence point for the principal signaling pathways that control macrophage polarization, with their associated transcription factors either competing for shared CBP/p300 binding surfaces or acting cooperatively to recruit it to enhancers. The CITED family of transcriptional regulators largely operate at this level and their differing binding preference for the three CH domains within CBP/p300 enable them to hinder or facilitate this process, shaping macrophage pro- and anti-inflammatory gene expression and phenotype. This process is assisted by the contrasting regulation of Cited1 and 2, which results in the highest expression of CITED2 in the absence of M1 or presence of M2-polarizing stimuli, dampening pro-inflammatory gene expression. Whereas CITED1 is only expressed at detectable levels in cells exposed to IFNγ for extended periods, enhancing ISG expression and maintaining the pro-inflammatory response.

The contributions of CITED2 to macrophage polarization and its role in a variety of inflammatory disorders, including atherosclerosis and diet-induced obesity, has been well explored using mouse models and primary cells (85, 96, 97, 99, 107). Additionally, the molecular interactions of this protein with CBP/p300 are well understood to the extent where the CITED2 CR2 domain has been used as the basis for cell-permeable peptide inhibitors of HIF1:CBP/p300 interaction with clinical potential (73). Far less is known about CITED1 in this context and our current knowledge is based exclusively on studies in murine macrophage cell lines (113, 114). Those in vivo studies that have been performed have instead focused on the impact of Cited1 gene deletion or the effect of CITED1 mutants on development and cancer rather than its role in myeloid cells (60, 121). Perhaps most importantly, the CITED1:CBP/p300 interaction has not been fully resolved with the proposed CH2-binding preference of CITED1 based only on qualitative in vitro pull-down mapping experiments using large GST-fused portions of p300 containing more than just the CH domains (65). While the CR2 domains of CITED1 and 2 are highly similar, (82% amino acid identity based on aa 161-209 of human CITED2 (44);), there are notable differences, including a 7 aa insert at the end of the domain that is absent from CITED1, indicating that a difference in CH binding preference is at least plausible. Future in-depth structural studies will provide clarity in this area. Should these differences in binding preference prove to be valid then their contributions to the seemingly opposite effects of CITED1 and 2 on IFNγ-stimulated gene expression could be explored using CITED mutants where the CR2 domains have been exchanged. Hypothetically, this would convert CITED1 from a transcriptional co-enhancer to a co-repressor.

A further unresolved question concerns the potentially differing relationships between the CITED proteins and the HATs, CBP and p300. It is currently unclear whether they exhibit a binding preference for one or the other, which is potentially significant as while the activities of CBP and p300 are at least partially overlapping, they are known to have separate functions and regulate distinct collections of genes (102). This is relevant within the context of macrophage polarization as CBP and p300 interact differently with STAT1. As an example of this, CBP, but not p300, has been shown to directly acetylate the STAT1 DNA binding domain at Lys410 and 413 (122). This has two important consequences; i) acetylated STAT1 binds nuclear NF-κB dimers, reducing their DNA-binding and transcriptional regulatory activity, and ii) signals for the nuclear export and dephosphorylation of STAT1 as part of a phosphorylation-acetylation switch that deactivates the protein (123). This would suggest that by interfering with the STAT1:CBP interaction, CITED proteins could, perhaps counterintuitively, extend STAT1 nuclear occupancy and enhance NF-κB activity.

As a similar concern, our understanding of how post-translational modifications (PTMs) of the CITED proteins themselves impact their function and activity remains incomplete. Most of those reported to date are not known to occur within myeloid cells and include the MAPK1-dependent phosphorylation of CITED2 at T166 at the N-terminal end of the SRJ domain (124), and the cell cycle-regulated phosphorylation of CITED1 mapped to five serine residues throughout the protein (125), with these modifications altering the coactivator functions of the proteins by influencing CBP/p300 binding. So far only CITED1 has been shown to be phosphorylated in macrophages in response to polarizing signals (in this case, IFNγ), but the modification sites, kinases responsible, and the functional consequences have yet to be elucidated. Additionally, just as many of the transcription factors that interact with CBP/p300 are acetylated by these HATs as an important part of their regulation [e.g., STAT1 and p65 NF-κB (98, 122, 123)], it seems plausible that the CITED proteins will also be acetylation targets or subject to other PTMs.

In closing, much remains to be discovered about how CITED proteins contribute to the control of macrophage polarization. Continued efforts in this area will help us better understand the various CBP/p300-dependent and independent mechanisms used by these proteins to tune pro- and anti-inflammatory gene expression and how the CITED proteins themselves are regulated.

DW: Writing – original draft, Writing – review & editing. JM: Writing – original draft, Writing – review & editing. DN: Writing – original draft, Writing – review & editing, Conceptualization, Funding acquisition, Supervision.