Twist1 and Twist2 Contribute to Cytokine Downregulation following Chronic NOD2 Stimulation of Human Macrophages through the Coordinated Regulation of Transcriptional Repressors and Activators

This article is featured in In This Issue, p.1

Human nucleotide-binding oligomerization domain 2 (NOD2), an intracellular sensor of bacteria-derived muramyl dipeptide (MDP; a component of peptidoglycan), confers the greatest genetic risk of developing Crohn’s disease (CD), a disease of chronic intestinal inflammation (1). When peripheral monocytes enter mucosal sites, such as the intestinal lamina propria, they are continuously exposed to bacterial products, including the NOD2 ligand, peptidoglycan/MDP (2, 3). Initial NOD2 stimulation results in cytokine secretion (3–5). However, ongoing NOD2 stimulation significantly downregulates cytokine secretion upon restimulation through pattern recognition receptors (PRRs) (3–7). This downregulation is impaired in individuals with CD-associated NOD2 polymorphisms (4, 5). Cytokine secretion in intestinal macrophages is similarly attenuated upon PRR stimulation (8), which is important for intestinal immune homeostasis. Moreover, chronic MDP administration to mice in vivo attenuates subsequent experimental colitis (6), thereby demonstrating the beneficial effects of chronic NOD2 stimulation in intestinal immune regulation.

Mechanisms contributing to cytokine downregulation after chronic NOD2 stimulation in human myeloid–derived cells include the upregulation of the intracellular inhibitors IRAK-M (4), Tollip (3), and IRF4 (6, 9), and the Tyro3, Axl, and Mer tyrosine kinase receptors (10); the NF-κB1–dependent upregulation of the transcriptional repressor activating transcription factor (ATF)3 (3); and the secretion of the inhibitory mediators IL-10 and TGF-β (5). Each of these mechanisms contributes only partially to cytokine downregulation and is operational to varying degrees in different individuals (e.g., IRAK-M) (4, 11). Given the dramatic alterations in macrophage functions and importance of downregulating PRR-initiated pathways upon chronic microbial stimulation, we hypothesized that additional critical mechanisms mediating these changes have yet to be identified, including those involving transcriptional regulators.

Twist1 and Twist2 are basic helix-loop-helix transcriptional repressors that bind to E boxes in gene promoters, acting as master regulators in a variety of biological processes, including organogenesis, osteogenesis, cancer progression, and hematopoietic cell development (12–14). As such, Twist1^−/− mice are embryonic lethal because of a variety of developmental defects (15). Mutations in Twist1 are associated with Saethre–Chotzen syndrome (16, 17), an autosomal dominant disorder characterized by craniofacial and limb anomalies. Although the major focus on Twist function has been in the context of the critical cellular processes described above, there is evidence for a role for Twists in regulating inflammation (14, 18–22). Twist2^−/− mice demonstrate increased myeloid lineage development (14). Moreover, Twist2^−/− mice or Twist1^+/−/Twist2^+/− mice demonstrate increased proinflammatory cytokines associated with increased NF-κB pathway signaling and perinatal death and defects in the type I IFN–mediated suppression of proinflammatory cytokines in macrophages (18, 22). In T cells, Twist1 limits Th1, Th17, and T follicular helper cell development (20, 21). However, the role and mechanisms through which Twist1 and Twist2 regulate PRR-induced cytokine outcomes is poorly understood; Twist regulation of NOD2-induced outcomes has not been reported. Given the repressive functions of the Twist proteins and evidence for their ability to limit inflammation in select situations, we hypothesized that Twist1 and/or Twist2 would contribute to downregulating cytokines during both acute and chronic NOD2 stimulation in primary human macrophages.

We found that Twist1 and Twist2 were required for the downregulation of cytokines during both acute and, in particular, chronic NOD2 stimulation of human monocyte–derived macrophages (MDMs). Consistently, Twist1 and Twist2 were each significantly upregulated after prolonged NOD2 stimulation; this upregulation required autocrine IL-10 and TGF-β and was defective in MDMs from CD-associated NOD2 risk carriers. Binding of the upregulated Twist1 and Twist2 transcriptional repressors to cytokine promoters was increased in a cooperative manner after chronic NOD2 stimulation. Moreover, Twist1 and Twist2 upregulated the expression and promoter binding of additional inhibitory transcriptional regulators, including c-Maf and Bmi1, while decreasing the expression and promoter binding of essential activating transcription factors, including ATF4, C/EBPα, Runx1, and Runx2. Complementation of the reduced c-Maf and Bmi1 expression, or reduction in the upregulated ATF4, C/EBPα, Runx1, and Runx2 under Twist-deficient conditions were each sufficient to restore cytokine downregulation under chronic NOD2 stimulation conditions. Therefore, Twist1 and Twist2 contribute to cytokine downregulation during chronic NOD2 stimulation through the coordination of the reciprocal expression and cytokine promoter binding of additional transcriptional repressors and activators.

Informed consent was obtained as approved by the Yale University Institutional Review Board. Monocytes were purified from healthy individuals or CD patients (Supplemental Fig. 1E) and differentiated to MDMs for 7 d with 10 ng/ml M-CSF (Shenandoah Biotechnology, Warwick, PA) as in Ref. 4. We performed NOD2 genotyping by TaqMan (Applied Biosystems, Foster City, CA) or Sequenom platform (Sequenom, San Diego, CA).

Total RNA was isolated and reverse transcribed, and quantitative PCR was performed as previously described (4) with normalization to GAPDH. Primers are available upon request.

For tolerance induction in vitro, human MDMs (0.5 × 10⁶) were pretreated with 100 μg/ml MDP (Bachem, King of Prussia, PA) for 48 h prior to extensive wash and retreated for 24 h with 100 μg/ml MDP. In some cases, anti–IL-10 or anti–TGF-β–neutralizing Abs (R&D Systems, Minneapolis, MN) or lipid A (Peptide International, Louisville, KY) was used. Supernatants were assayed for cytokine secretion per manufacturer instructions using the following Abs: TNF, IL-6, IL-8, IL-10 (BD Biosciences, San Jose, CA), or IL-12 (eBioscience, San Diego, CA).

Western blot analysis was performed as in Ref. 4 using Abs to Twist1/Twist2 (ab50887; Abcam, Cambridge, MA), c-Maf (sc7866; Santa Cruz Biotechnology, Santa Cruz, CA), Bmi1 (05-1322; EMD Millipore, Billerica, MA), or anti-GAPDH (EMD Millipore).

Primary human MDMs were transfected with 100 nM scrambled or ON-TARGETplus SMARTpool small interfering RNA (siRNA) against Twist1, Twist2, c-Maf, Bmi1, ATF4, C/EBPα, Runx1, or Runx2 (Dharmacon, Lafayette, CO) (four pooled siRNAs for each gene) or with 2 μg pT3-EF1a-Bmi1 [Addgene plasmid 31783 kindly deposited by X. Chen (23)], c-Maf (GeneCopoeia, Rockville, MD), or empty vector using Amaxa nucleofector technology (Amaxa, San Diego, CA).

Chromatin immunoprecipitation (ChIP) analysis of MDMs was performed according to a modified protocol (3). Primers were designed to amplify genomic sequences at the cytokine gene promoter region (available upon request). Abs used included the following: Twist1 (ab50887), Twist2 (ab66031), Runx1 (ab23980), ATF4 (ab85049), Bmi1 (ab14389) (Abcam), Runx2 (sc10785x), C/EBPα (sc-61x), or c-Maf (sc-7866x) (Santa Cruz Biotechnology).

Significance was assessed using a two-tailed Student t test. A p value < 0.05 was considered significant. A Bonferroni correction was applied for multiple comparisons. To maintain a consistent picogram per milliliter scale for cytokine concentration measures, the higher levels observed for IL-8 and IL-6 (nanogram per milliliter levels) are shown with a multiplication factor.

To elucidate the role of Twist1 and Twist2 in cytokine downregulation with acute NOD2 stimulation in human MDMs, we used siRNA to reduce the expression of Twist1 and Twist2, alone and in combination. We assessed expression of each Twist with each knockdown condition and found selective downregulation of the targeted Twist (Supplemental Fig. 1A), such that Twist1 and Twist2 did not coregulate the expression of the other. We further confirmed knockdown of Twist1 and Twist2 proteins (Supplemental Fig. 1B). Knockdown of Twist1 and Twist2, alone and in combination, led to increased secretion of the proinflammatory cytokines TNF, IL-6, IL-8, IL-12p40 (Fig. 1), and IL-1β (Supplemental Fig. 1C), as well as the anti-inflammatory mediators IL-10 (Fig. 1) and IL-1Ra (Supplemental Fig. 1C), upon acute NOD2 stimulation of human MDMs. We next evaluated the role of Twist1 and Twist2 in the cytokine downregulation observed after chronic NOD2 stimulation. We previously found that stimulation of NOD2 for 48 h is optimal for downregulating cytokine secretion upon restimulation through NOD2 or other PRRs (4). We used 100 μg/ml MDP as this concentration approximates muramic acid stool levels (24), has been used by us and others, and results in optimal cytokine downregulation after chronic NOD2 stimulation (3–5). Upon knockdown of Twist1 or Twist2 after initiation of chronic NOD2 stimulation and then restimulating through NOD2 (Fig. 1A), there was significant reversal in the downregulation of both proinflammatory cytokines and anti-inflammatory cytokines (Fig. 1B). Reversal was further enhanced when Twist1 and Twist2 were knocked down in combination (Fig. 1B). Taken together, Twist1 and Twist2 cooperate to downregulate cytokines upon both acute and chronic NOD2 stimulation of human MDMs.

Given the role for Twist1 and Twist2 in downregulating cytokines upon chronic NOD2 stimulation, we questioned whether Twist expression is upregulated with chronic NOD2 stimulation of human MDMs. Twist1 and Twist2 mRNA expression was enhanced in 48-h MDP-pretreated MDMs and further increased 4 h after restimulation (Fig. 2A). Twist1/2 protein expression was similarly increased after restimulation of 48-h MDP-pretreated MDMs (Fig. 2B). We next sought to define the mechanisms upregulating Twist1 and Twist2 expression upon chronic NOD2 stimulation in human MDMs. We have previously shown that the early secretion of IL-10 and TGF-β is required for subsequent pro- and anti-inflammatory cytokine downregulation during chronic NOD2 stimulation (5). Furthermore, recent studies have demonstrated a necessity for IL-10R on intestinal macrophages for the generation of anti-inflammatory macrophages (25, 26), thereby highlighting the importance of IL-10 in conditioning the inhibitory mechanisms generated in intestinal macrophages. We found that autocrine IL-10 and TGF-β were each required, and moreover cooperated with each other, for optimal upregulation in Twist1 and Twist2 expression upon chronic NOD2 stimulation in human MDMs (Fig. 2C). Finally, we questioned whether MDMs from CD patients similarly upregulate Twist1 and Twist2 upon chronic NOD2 stimulation (see Supplemental Fig. 1E for patient characteristics). Twist1 and Twist2 expression in MDMs from wild-type NOD2 CD patients was induced to a similar degree as in MDMs from healthy controls (Fig. 2D). Furthermore, the range of Twist1 and Twist2 expression upon chronic NOD2 stimulation in these individuals was not significantly different (Supplemental Fig. 1D). In contrast, Twist1 and Twist2 expression failed to increase after chronic NOD2 stimulation of MDMs from CD loss-of-function NOD2 risk carriers (Fig. 2D). MDMs from these NOD2 risk carriers were responsive to stimulation through alternative PRRs, such that Twist1 and Twist2 expression increased in these MDMs upon chronic lipid A treatment (Fig. 2D). Therefore, Twist1 and Twist2 increase in an IL-10– and TGF-β–dependent manner upon chronic NOD2 stimulation.

Twist1 and Twist2 transcriptional repressors were upregulated with chronic NOD2 stimulation (Fig. 2) and were required for optimal cytokine downregulation after chronic NOD2 stimulation (Fig. 1). We therefore questioned whether the binding of these transcriptional repressors to cytokine promoters was enhanced after chronic NOD2 stimulation, and if so, whether Twist1 and Twist2 cooperated with each other for this enhanced binding. Twist1 and Twist2 binding to both proinflammatory and anti-inflammatory cytokine promoters increased particularly 48 h after NOD2 stimulation and even more so upon restimulation of MDP-pretreated MDMs (Fig. 3A). Upon examining restimulated MDP-pretreated MDMs where Twist1 and Twist2 binding to cytokine promoters was most enriched, we observed that knockdown of each Twist affected binding not only of itself but also of the other Twist protein to cytokine promoters (Fig. 3B). Taken together, chronic NOD2 stimulation results in significantly enriched and cooperative binding of the transcriptional repressors Twist1 and Twist2 to proinflammatory and anti-inflammatory cytokine promoters.

In addition to their direct suppressor activity, transcriptional repressors, including the Twist proteins, can both increase the expression of additional inhibitory molecules and suppress the expression of essential activating molecules; in combination, these enable optimal suppressive outcomes (12). We therefore first investigated whether Twist1 and Twist2 might be contributing to cytokine downregulation after chronic NOD2 stimulation through upregulating inhibitory molecules previously identified to play a role in NOD2-induced cytokine downregulation. However, Twist1 and Twist2 knockdown, either alone or in combination, did not impair the upregulation of NFκB1, ATF3, IRAK-M, or Tollip after chronic NOD2 stimulation (Supplemental Fig. 2A). We further examined proximal signaling pathways that are induced upon acute NOD2 stimulation and found that combined Twist1 and Twist2 knockdown did not regulate acute NOD2–induced activation of the ERK, p38, or JNK pathways (Supplemental Fig. 2B) or the early secretion of IL-1β (Supplemental Fig. 2C), which we have found to be a measure of NOD2-induced caspase-1 activation (27, 28).

Twist1 and Twist2 have not been well studied in PRR-induced cytokine regulation. However, they have been well investigated in carcinogenesis, osteogenesis, and hematopoietic cell development. We therefore questioned whether inhibitory molecules upregulated by Twist1 and/or Twist2 identified through these other processes might be similarly Twist dependent upon chronic NOD2 stimulation in human MDMs and thereby contribute to the cytokine downregulation observed. During myeloid lineage development, Twist2 can upregulate the proinflammatory transcriptional inhibitor c-Maf (14); Twist2 can also bind to the c-Maf promoter with LPS treatment in mouse macrophages (29). However, c-Maf regulation of PRR-initiated pathways has not been well defined. Twist1 regulates Bmi1 expression and cooperates with Bmi1 in repressor functions during epithelial–mesenchymal transition relevant to cancer metastasis (30); a role for Bmi1 in PRR signaling and outcomes is poorly understood. Furthermore, Twist regulates PI3K pathways (31), and Akt1 can contribute to an anti-inflammatory macrophage phenotype (32). We found that c-Maf, Bmi1, and Akt1 were upregulated with chronic NOD2 stimulation (Fig. 4A). However, only c-Maf and Bmi1 transcript upregulation was Twist1 and Twist2 dependent (Fig. 4B); c-Maf and Bmi1 protein upregulation was similarly Twist dependent (Fig. 4C). Furthermore, the binding of c-Maf and Bmi1 to both proinflammatory and anti-inflammatory cytokine promoters was enriched with chronic NOD2 stimulation (Fig. 4D); this enhanced promoter binding was Twist1 and Twist2 dependent (Fig. 4D). Importantly, through c-Maf and Bmi1 knockdown (Fig. 5A), we established that c-Maf and Bmi1 each contributed to the cytokine downregulation observed after chronic NOD2 stimulation (Fig. 5B). Moreover, they cooperated with each other for optimal cytokine downregulation (Fig. 5B). Finally, to clearly establish the role of c-Maf and Bmi1 in Twist-dependent contributions to chronic NOD2-induced cytokine downregulation in MDMs, we restored both c-Maf and Bmi1 expression in Twist-deficient MDMs to levels similar to those observed after chronic NOD2 stimulation (Fig. 4C). Despite the absence of Twist1 and Twist2, reconstitution of each c-Maf and Bmi1 could partially restore the cytokine downregulation observed after chronic NOD2 stimulation (Fig. 5C). Moreover, the restoration was further improved with combined expression of both c-Maf and Bmi1 (Fig 5C). Therefore, one mechanism through which Twist1 and Twist2 downregulate cytokines after chronic NOD2 stimulation is through the upregulation in expression and cytokine promoter binding of the transcriptional repressors c-Maf and Bmi1.

Twist1 and Twist2 can both decrease the expression and compete with promoter binding of activating transcription factors in osteogenesis, carcinogenesis, and myeloid differentiation (12). We therefore hypothesized that Twist1 and Twist2 might be similarly downregulating activating transcription factors upon chronic NOD2 stimulation where cytokines are dramatically reduced. How activating transcription factors are regulated with chronic NOD2 stimulation has not been well defined. We therefore considered the Twist-regulated activating transcription factors defined in other cellular processes. Twist1 can attenuate ATF4 binding to the osteocalcin promoter in osteoclasts (33). Furthermore, Twist1 and Twist2 can inhibit Runx2 transactivation functions during osteoblast differentiation (34). Moreover, Twist2 can negatively regulate myeloid lineage differentiation by interacting with and inhibiting the transcription factors Runx1 and C/EBPα (14). Twist proteins can also regulate Akt2 expression in breast cancer cells (35). To our knowledge, roles for ATF4, C/EBPα, Runx1, and Runx2 in NOD2-induced outcomes have not been reported, and we therefore first examined the regulation of each of these transcription factors with acute and chronic NOD2 stimulation of MDMs.

With NOD2 stimulation of MDMs, ATF4, C/EBPα, Runx1, Runx2, and Akt2 were each upregulated acutely but downregulated by 48 h, and they failed to increase on restimulation of NOD2 (Fig. 6A), consistent with the attenuated cytokines at this time. We further assessed the Twist dependency for these downregulated transcription factors and found that ATF4, C/EBPα, Runx1, and Runx2 downregulation was dependent on the combined expression of Twist1 and Twist2 (Fig. 6B). Akt2 downregulation after chronic NOD2 stimulation, in contrast, was Twist independent (Fig. 6B). We therefore pursued the role and regulation of the Twist-dependent transcription factors. ATF4, C/EBPα, Runx1, and Runx2 bound to cytokine promoters with acute NOD2 stimulation, and this binding was dramatically reduced after chronic NOD2 stimulation, consistent with the reduced cytokines observed under these conditions (Fig. 6C). Moreover, upon Twist1 and Twist2 knockdown during chronic NOD2 stimulation, the binding of these transcription factors to cytokine promoters failed to undergo reduction relative to scrambled knockdown cells, consistent with the impaired cytokine downregulation (Fig. 6C). Finally, to ensure that the relative increase in expression and binding of ATF4, C/EBPα, Runx1, and Runx2 to cytokine promoters in Twist-deficient MDMs was contributing to the failure to downregulate cytokines upon chronic NOD2 stimulation, we used siRNA to each of the transcription factors (Fig. 7A), alone and in combination. Reduced expression of each ATF4, C/EBPα, Runx1, and Runx2 partially restored cytokine downregulation, and combined knockdown of these Twist-regulated transcription factors fully restored cytokine downregulation in Twist-deficient macrophages after chronic NOD2 stimulation (Fig. 7B). Cells were viable under these conditions (data not shown). Therefore, Twist1 and Twist2 contribute to the downregulation in expression and promoter binding of various transcription factors, in particular ATF4, C/EBPα, Runx1, and Runx2, which in turn is necessary for optimal cytokine downregulation after chronic NOD2 stimulation in MDMs.

Because intestinal macrophages are exposed to multiple PRRs, we next assessed whether chronic stimulation of other PRRs similarly upregulates Twist1 and Twist2 expression in human MDMs, and whether these other PRRs cooperate with NOD2 to regulate Twist1- and Twist2-mediated outcomes. Similar to chronic NOD2 stimulation, cytokines are downregulated upon restimulation of MDM after chronic stimulation of TLR2, TLR4, or TLR9 (Supplemental Fig. 3A). We found that upon pretreatment through TLR2, TLR4, or TLR9, and then restimulation of each respective TLR, Twist1 and Twist2 expression was increased (Supplemental Fig. 3B). The Twist-dependent pathways we had identified were similarly regulated under these conditions, with upregulation of the transcriptional repressors c-Maf and Bmi1 (Supplemental Fig. 3C) and downregulation of the transcriptional activators ATF4, C/EBPα, Runx1, and Runx2 (Supplemental Fig. 3D). Although low-dose MDP treatment synergized with each of the TLR ligands to increase expression of cytokines during acute treatment, cytokines were downregulated upon chronic stimulation of NOD2 with each of the TLRs (Supplemental Fig. 3A). Moreover, chronic MDP treatment synergized with each TLR to upregulate Twist1 and Twist2 and the transcriptional repressors c-Maf and Bmi1 (Supplemental Fig. 3B, 3C). Taken together, under the chronic microbial ligand exposure that occurs in the intestinal environment, NOD2 can synergize with other PRRs to upregulate Twist1 and Twist2 expression and Twist-dependent mechanisms that lead to cytokine downregulation.

In this study, we demonstrate that Twist1 and Twist2 contribute to the downregulation of proinflammatory cytokines upon acute and particularly chronic NOD2 stimulation of human MDMs. Twist1 and Twist2 expression increased with chronic NOD2 stimulation, consistent with the important role for these transcriptional repressors in the downregulation of cytokines observed after prolonged NOD2 stimulation. Twist upregulation is similarly observed in MDMs from CD patients but is impaired in MDMs from CD NOD2 risk carriers. Early secreted IL-10 and TGF-β were critical for Twist1 and Twist2 expression upregulation with chronic NOD2 stimulation. Binding of Twist1 and Twist2 to cytokine promoters was particularly enhanced after chronic NOD2 stimulation, and the two proteins cooperated with each other for optimal promoter binding. In addition to the direct binding of the Twist1 and Twist2 transcriptional repressors to cytokine promoters, they also orchestrated a coordinated program in which the expression and cytokine promoter binding of additional transcriptional repressors including c-Maf and Bmi1 were increased, whereas the expression and cytokine promoter binding of various activating transcription factors including ATF4, C/EBPα, Runx1, and Runx2 were reduced. These combined Twist-mediated mechanisms contributed to the ability of Twist1 and Twist2 to downregulate NOD2-induced cytokines (Supplemental Fig. 4). Furthermore, NOD2 could synergize with various TLRs to mediate these outcomes, consistent with the cooperative role for NOD2 with other PRRs and cytokines in enhancing downstream immune outcomes (36–42). Therefore, we now elucidate a clear role for Twist1- and Twist2-mediated mechanisms in downregulating cytokines upon acute and chronic NOD2 stimulation of primary human MDMs.

A role for Twist repressors in PRR-induced outcomes has been relatively unexplored. However, Twist1- and Twist2-mediated transcriptional mechanisms identified to regulate suppression in biological processes such as osteogenesis, hematopoiesis, and carcinogenesis provided the opportunity to now dissect if these pathways are Twist dependent and similarly or differentially regulate outcomes upon PRR stimulation of human macrophages. A number of the downstream molecules contributing to Twist-dependent regulation in these other cell processes have not been previously implicated in NOD2-induced outcomes. One such example is the role we now identify for Twist-dependent c-Maf regulation in downregulating cytokines after chronic NOD2 stimulation. c-Maf has been predominantly studied in the context of T cell differentiation, in particular Tr1 cells (43). However, a few studies have also identified a role for c-Maf as an anti-inflammatory transcription factor in myeloid cells; c-Maf downregulates transcriptional activation of proinflammatory cytokines while upregulating transcriptional activation of IL-10 in myeloid-derived cells (44, 45). In contrast, we found that in primary human macrophages, c-Maf downregulated both pro- and anti-inflammatory cytokines upon NOD2 stimulation (Fig. 5). Studies on Bmi1 have focused largely on its contributions to carcinogenesis. The role of Bmi1 in regulating PRR-induced outcomes has been poorly defined. To our knowledge, there is only one such study; this study identified a role for Bmi1 in suppressing IL-10 in mouse macrophages (46). We now define a broader role for Bmi1 in suppressing a range of both pro- and anti-inflammatory cytokines upon PRR stimulation, with a Twist-dependent role for Bmi1 in the robust downregulation of cytokines after chronic PRR stimulation (Fig. 5). Although neither Akt1 nor Akt2 was regulated by Twist proteins, we observed that Akt1 was upregulated (Fig. 4A), whereas Akt2 was downregulated (Fig. 6A) with chronic NOD2 stimulation. This pattern of expression regulation is consistent with the role of Akt1 in polarization to anti-inflammatory macrophages and of Akt2 in polarization to proinflammatory macrophages (32).

We identified a number of transcription factors, including ATF4, C/EBPα, Runx1, and Runx2 whose expression and binding to cytokine promoters was upregulated with acute NOD2 stimulation but downregulated in a Twist-dependent manner with chronic NOD2 stimulation. Restoring the downregulation of these transcription factors in Twist-deficient MDMs was able to restore the cytokine downregulation observed with chronic NOD2 stimulation. Runx1 and Runx2 have been well studied in cell proliferation and differentiation (47), and C/EBPα has been identified as a lineage-specific transcription factor in the hematopoietic system (48). However, the role of these transcription factors in PRR-induced outcomes has been poorly defined. We now identify each of these transcription factors as being essential in the induction of cytokines after acute NOD2 stimulation; their downregulation after chronic NOD2 stimulation contributes to the overall cytokine downregulation under these conditions. Limited studies have examined the role of ATF4 downstream of PRRs. ATF4 was found to contribute to cytokine secretion upon acute TLR4 stimulation (49). In contrast, upon prestimulation of macrophages through TLR4, TRIF-dependent signaling led to the downregulation of the ATF4-C/EBP homology protein branch of the unfolded protein response (50). We similarly observed differential regulation of ATF4 with acute and chronic NOD2 stimulation. Taken together, these studies elucidate an important role for Twist1 and Twist2 in regulating NOD2- and PRR-induced cytokine secretion under both the initial acute stimulation as well as under the chronic stimulatory conditions that are present in the intestinal environment, thereby providing insight into the inhibitory mechanisms critical for regulating the CD-associated NOD2.

The authors have no financial conflicts of interest.