Macrophages are sentinel immune cells that survey the tissue microenvironment, releasing cytokines in response to both exogenous insults and endogenous events such as tumorigenesis. Macrophages mediate tumor surveillance and therapy-induced tumor regression; however, tumor-associated macrophages (TAM) and their products may also promote tumor progression. Whereas NF-κB is prominent in macrophage-initiated inflammatory responses, little is known about the role of p53 in macrophage responses to environmental challenge, including chemotherapy or in TAMs. Here, we report that NF-κB and p53, which generally have opposing effects in cancer cells, coregulate induction of proinflammatory genes in primary human monocytes and macrophages. Using Nutlin-3 as a tool, we demonstrate that p53 and NF-κB rapidly and highly induce interleukin (IL)-6 by binding to its promoter. Transcriptome analysis revealed global p53/NF-κB co-regulation of immune response genes, including several chemokines, which effectively induced human neutrophil migration. In addition, we show that p53, activated by tumor cell paracrine factors, induces high basal levels of macrophage IL-6 in a TAM model system [tumor-conditioned macrophages (TCM)]. Compared with normal macrophages, TCMs exhibited higher p53 levels, enhanced p53 binding to the IL-6 promoter, and reduced IL-6 levels upon p53 inhibition. Taken together, we describe a mechanism by which human macrophages integrate signals through p53 and NF-κB to drive proinflammatory cytokine induction. Our results implicate a novel role for macrophage p53 in conditioning the tumor microenvironment and suggest a potential mechanism by which p53-activating chemotherapeutics, acting upon p53-sufficient macrophages and precursor monocytes, may indirectly impact tumors lacking functional p53. Cancer Res; 74(8); 2182–92. ©2014 AACR.
The immune system is intricately connected to various stages of tumorigenesis. On one hand, it is critical in the clearance of senescent and damaged tumor cells after therapy and in tumor prevention through detection and elimination of early-stage cancer cells (1, 2). Work from Xue and colleagues (1) is exemplary of this phenomenon, because they show that p53 stabilization triggers senescence in tumor cells in mice and subsequent activation of innate immune cells that ultimately clear the senescent tumor cells, leading to tumor regression. On the other hand, immune cells associated with the tumor microenvironment of advanced cancers have tumor-promoting functions through secretion of soluble factors that promote survival, proliferation, angiogenesis, and metastasis (3). Underlying central components to this dual function of the immune system in tumorigenesis are macrophages.
Macrophages are key players in innate immunity, and their functions depend on the environment in which they reside. In a noncancerous microenvironment, macrophages are central detectors of infectious and noninfectious exogenous stress, including DNA-damaging agents such as chemotherapeutics. Upon stimulation, macrophages trigger cascades of cell–cell signaling that result in synthesis and secretion of proinflammatory cytokines and chemokines, and, consequently, recruitment of other effector immune cells. In the context of a tumor, tumor cells secrete soluble factors that recruit and program tumor-associated macrophages (TAM) to support tumor growth (3). TAMs are phenotypically distinct from classical macrophages in that they exhibit different morphology and expression markers. For example, TAMs have high expression of interleukin (IL)-6, CXCL1, IL-8, and CCL2 (3–5). The secretion of cytokines and chemokines constitute a major mechanistic feature of macrophage function; therefore, understanding the precise mechanisms that drive the induction of proinflammatory genes is crucial.
NF-κB plays an essential role in inflammation, innate immunity, and cancer (6, 7). Activated by inflammatory stimuli such as pathogen-associated molecular patterns (PAMP) and various cytokines, including TNF-α, NF-κB enhances transcription of several proinflammatory cytokines such as IL-6 and IL-8, which are secreted from the cell and propagate the immune response by acting on neighboring immune cells (6). In addition, NF-κB is found constitutively activated in several types of human cancers and has been shown to promote cancer cell growth and survival, for example, by regulating the transcription of antiapoptotic genes (7).
Another master regulator of stress response, the tumor suppressor p53, also has roles in inflammation and immunity (8, 9). Recently, we reported that p53 can upregulate most members of the Toll-like receptor (TLR) family and consequently enhance TLR-dependent production of proinflammatory cytokines (10, 11). Surprisingly, p53 regulation of the TLRs is restricted to human cells because the p53 response elements (p53REs) in the TLR promoter regions are not conserved in mice (10), suggesting that some p53-related immune responses can only be addressed in human material. These results highlight the fact that p53 has an important physiologic role in the immune system in addition to its well-characterized role as a tumor suppressor, providing a new dimension to the broad role that p53 plays in human biology.
Mechanistically, activation of p53 and NF-κB is similar and involves stress-induced degradation of inhibitors. Specifically, Mdm2 (murine double-minute 2 or human hMdm2) binds to p53 and targets p53 for proteasomal degradation. During cellular stress, Mdm2 and p53 are modified and can no longer bind to each other, leading to p53 stabilization (12). For NF-κB, the subunits are held in the cytoplasm through binding to IκB (inhibitor of NF-κB) proteins. Upon stimulation, IκB is phosphorylated by the upstream IKK (inhibitor of nuclear factor-κB kinase) complex, leading to ubiquitination and degradation of IκB, which allows the NF-κB subunits to enter the nucleus and modulate transcription (reviewed in ref. 6). Although both transcription factors are activated by similar stimuli, the resulting biologic consequences are generally considered to be opposing in that NF-κB signaling is mostly associated with prosurvival cues and p53 induction is proapoptotic (13). In addition, signaling cross-talk between p53 and NF-κB is antagonistic (13). However, there are a few examples of p53 and NF-κB working together (14, 15).
Because macrophages are key players in the immune response to tumor cells, we have addressed the potential role for p53 regulation of immune genes and possible interactions with the NF-κB network within human macrophages from healthy human subjects and in a TAM in vitro model system, tumor conditioned macrophages (TCM). We show that p53 and NF-κB coregulate the expression of several proinflammatory genes, including cytokines (such as IL-6) and chemokines (such as CXCL1) in TCMs and in normal macrophages in response to p53 and NF-κB–activating agents. Taken together, these results identify a unique p53/NF-κB interaction within macrophages that we propose may be important in amplifying the immune responses associated with various stages of tumorigenesis and after chemotherapy and radiotherapy.
Materials and Methods
Alveolar macrophages were isolated as previously described (16). Monocyte-derived macrophages (17) and neutrophils were harvested with Histopaque-1077 (Sigma-Aldrich) from the blood of healthy volunteers as previously described (17). For TCMs, monocytes were cultured with 50% conditioned media from MDA-MB-231 cells for the first 4 days as previously described (IRB #: 10-E-0063; ref. 5).
The following reagents were used: 10 μmol/L Nutlin-3, 0.3 μmol/L doxorubicin, 300 μmol/L 5-fluorouracil, 1 μ/mL Actinomycin D (Sigma Aldrich), 50 μmol/L pifithrin-α, 50 μmol/L Bay-11-7082 (Sigma-Aldrich); 100 ng/mL LPS (InvivoGen); p65 S276 blocking peptide or control peptide (Imgenex, per manufacturer's protocol).
Chromatin immunoprecipitation assay
Chromatin immunoprecipitation (ChIP) assays were performed as described in our previous work (18). SYBR Green assays were performed using IL-6–specific primers (Cat. # NM_000600.2 (-)01Kb: GPH1011803(-)01A; Qiagen Inc.).
TaqMan gene expression assays or SYBR green assays were run with Applied Biosystems 7900HT Fast Real-Time PCR System and results were analyzed with SDS2.4 software. Primers are described in the Supplementary Data.
Microarrays were performed with the Affymetrix GeneChIP Instrument system using procedures previously described (18) and analyzed using a mixed linear ANOVA model: , where Yijkl represents the log base 2 intensity value for the lth observation on the ith treatment for the jth subject and kth biologic replicate. μ is the grand mean expression and εijkl represents the random error. The errors are assumed to be normally and independently distributed with mean 0 and SD δ for all measurements. Thresholds used for detecting differentially expressed genes: FDR < X, FC ± Y. Treatment is a fixed effect, whereas subject and replicate are random effects. The following contrasts were performed to compare mean intensity values: LPS vs. non-treated; Nutlin vs. DMSO. GEO accession: GSE43596.
Secreted IL-6 and CXCL1 protein from treated macrophages was quantified using the IL-6 Human ELISA Kit from Life Technologies and the Human CXCL1/GROa Immunoassay from R&D Systems following the manufacturers' protocol.
Macrophages were lysed in radioimmunoprecipitation assay buffer for whole-cell extracts or with NE-PER kit (Thermo Scientific) for nuclear and cytoplasmic extracts, and immunoblots were performed as described in ref. 10. Specific antibodies are listed in the Supplementary Data.
Neutrophil migration assay
The QCM Chemotaxis Cell Migration Assay (EMD Millipore) was used following the manufacturer's protocol. Neutrophils were seeded at 20 × 104 cells per well and allowed to migrate for 2 hours. Fluorescence was measured as relative fluorescence units (RFU).
For analysis of expression changes (treated compared with untreated or vehicle control), ChIP binding, and neutrophil migration, the data were log transformed and then a one sample t test was performed for each pair to test if the mean fold change is ≤ to 0, or alternatively, the mean is >0. Assuming a normal distribution with mean = μ and variance σ2, test statistic is defined as: where is the mean, n is the sample size, and S is the SD of the values. We set μ0 equal to 0 and perform the test at α = 0.05. Significance (indicated with an asterisk) is determined if tc > the critical value of a t-distribution with n − 1 degrees of freedom and α = 0.05.
Activation of p53 induces IL-6, TNFα, and IL-8 expression
In our initial experiments, we assessed the effect of p53 activation on gene expression in primary human alveolar macrophages using the chemotherapeutic drugs doxorubicin and 5-fluorouracil and the small molecule Nutlin-3, which stabilizes p53 (Supplementary Fig. S1A; ref. 19). Treatment with doxorubicin highly induced mRNA levels of the cytokine gene IL-6, a proinflammatory gene important for macrophage function (Fig. 1A). As shown in Fig. 1B, Nutlin-3 treatment for 6 hours also enhanced mRNA levels of the cytokines IL-6 and TNFα and the chemokine IL-8 in macrophages from every donor tested; however, there were large differences among the genes and donors. The expression of the p21 gene, a well-known p53 target that identifies p53 activation, was also increased by Nutlin-3 treatment (Fig. 1B) but with less variation. Notably, IL-6 mRNA was induced in some donors to much higher levels than TNFα or IL-8 (Fig. 1B). Because IL-6, TNFα, and IL-8 protein were all induced (Fig. 1C and S1B), p53 activation is likely to have a strong functional impact.
Rapid induction of IL-6 by Nutlin-3
Because of low cell recovery and availability, experiments with human alveolar macrophages are limiting. Therefore, we utilized macrophages differentiated ex vivo from primary human blood monocytes (monocyte-derived macrophages). Similar to results with alveolar macrophages, induction of IL-6 after treatment with doxorubicin, 5-fluorouracil, and Nutlin-3 was also observed in monocyte-derived macrophages (Fig. 2A). All three treatments also induced p21, demonstrating p53 responsiveness (Fig. 2A). Because IL-6 induction was the highest and most consistent cytokine product among donors after Nutlin-3 treatment, our subsequent experiments focused on Nutlin-3 activation of p53.
Because p53 nuclear accumulation is observed within 1 hour of Nutlin-3 treatment (Supplementary Fig. S1A), we assayed the kinetics of IL-6 induction by Nutlin-3 in monocyte-derived macrophages. IL-6 mRNA induction levels were measured after 2, 4, and 6-hour Nutlin-3 treatments of macrophages from each donor (Fig. 2B). Within 2 hours, the level of IL-6 mRNA was dramatically increased, up to 10,000-fold, in cells of all but one donor (which showed an ∼8-fold induction). The rapid and large increases are similar to cytokine production triggered through TLR signaling by PAMPs such as LPS (20). In addition, IL-6 induction by Nutlin-3 treatment was dependent on transcription since pretreatment with actinomycin D, a transcription inhibitor, abolished IL-6 mRNA induction (Supplementary Fig. S1C).
The kinetics and level of IL-6 induction by Nutlin-3 varied considerably among donors, suggesting large differences in individual responses to stresses that activate p53. For example, macrophages from donors 117 and 120 were high responders with IL-6 induction levels in the 10,000–25,000-fold range, whereas donors 123 and 125 were relatively lower responders with induction levels reaching 10–100-fold (Fig. 2A and B). There was much less variability in the kinetics of p21 induction compared with IL-6 induction by Nutlin-3 (Fig. 2A and B), suggesting that the differences are not due to p53 level.
The very rapid and dramatic IL-6 induction by Nutlin-3 was surprising because this deviates from the typical kinetics of p53-dependent gene regulation described in the literature. Whereas gene regulation by other transcription factors such as NF-κB can be rapid (within minutes; ref. 21), p53 regulation of transcription generally occurs more slowly (over several hours; ref. 22). In addition, the increases in protein levels for IL-6, IL-8, and TNFα after Nutlin-3 treatment were comparable with the levels induced by LPS treatment, a well-known stimulator of the inflammatory response and activator of NF-κB (Supplementary Fig. S1B). To confirm that the induction of IL-6 by Nutlin-3 was p53-dependent, we examined the effects of the p53 inhibitor pifithrin-α (PFT-α; ref. 23). Several attempts to significantly reduce p53 using siRNA approaches were unsuccessful due to low transfection efficiency of primary human macrophages. In addition, as reported by Liu and colleagues (24) and from our preliminary experiments, Nutlin-3 treatment alone did not induce IL-6 expression in bone marrow–derived macrophages from wild-type mice, preventing our use of p53−/− murine macrophages in this study. As expected, PFT-α reduced Nutlin-3–induced p21 levels in all donors (Supplementary Fig. S1D). Because, as shown in Fig. 2C, the PFT-α pretreatment dramatically lowered Nutlin-3–inducedIL-6 mRNA levels, we conclude that p53 is required for IL-6 induction by Nutlin-3 in human macrophages.
Induction of IL-6 by p53-activating agents is limited to the monocyte/macrophage cell lineage
Several other types of primary cells and cancer cell lines were tested for the effect of p53-activating agents on IL-6 induction. As shown in Supplementary Table S1, short (e.g., 2 hours) and long (e.g., 24 hours) incubations with Nutlin-3, doxorubicin, and 5-fluorouracil did not induce IL-6 expression in cancer cell lines of various origins and p53 status, such as breast (MCF-7, p53+), bone (Saos2, p53− and U2OS, p53+), lung (H1299, p53−), and colon (HCT-116, p53+). As expected, p21 was upregulated in all cell lines carrying a functional TP53 gene. Other types of human primary cells were also treated with Nutlin-3, doxorubicin, and 5-fluorouracil, including human diploid fibroblasts, CD3+ T-lymphocytes, etc. (Supplementary Table S1). Whereas all of these cells showed p53 responsiveness (i.e., p21 induction; ref. 10), the only human primary cells that showed induction of either IL-6, TNFα, or IL-8 after treatment with p53-activating agents were CD14+ monocytes, which are macrophage precursors. Altogether, p53 regulation of IL-6 appears restricted to human monocytes and macrophages.
Activation of NF-κB is required for Nutlin-3 induction of IL-6 in primary macrophages
Having established that p53 is required for Nutlin-3 induction of IL-6, we next investigated whether NF-κB is also required. First, NF-κB is a well-known regulator of the transcription of the proinflammatory genes IL-6, TNFα, and IL-8 (25, 26). Second, the kinetics of IL-6 induction by Nutlin-3 are comparable to those for cytokine induction by PAMPs, which are largely driven by NF-κB (20). Third, there are reports describing cooperation between p53 and NF-κB in the transcription of certain genes (15, 27). Finally, NF-κB and p53 are known to be coordinately activated by certain stimuli, such as DNA damage, oxidative stress, and immune stimuli (28).
We first evaluated whether NF-κB is indeed activated by Nutlin-3 in human macrophages. In the canonical NF-κB pathway, the inhibitor IκBα is phosphorylated and degraded, which frees the NF-κB subunits to translocate to the nucleus to promote transcription of target genes, including the IκBα gene (6). We found that IκBα protein decreases within 30 minutes of Nutlin-3 addition to monocyte-derived macrophages, followed by an increase to almost basal levels at 6 hours (Fig. 3A). Similar results were seen with alveolar macrophages (Supplementary Fig. S2A). In addition, IκBα phosphorylation on Ser32 is seen as early as 30 minutes after Nutlin-3 treatment (“pIκBα,” Fig. 3A).
In addition to regulation by the inhibitor IκB proteins, NF-κB subunits are further regulated by posttranslational modifications such as phosphorylation. Serine 276, an important activating phosphorylation site of the NF-κB subunit p65, (29) is phosphorylated after a 2-hour treatment of Nutlin-3 (Fig. 3A, “pp65”). Cytokine induction by nutlin requires phosphorylation of both IκBα and p65, as IL-6 induction is markedly reduced by pretreatment either with Bay-11-7082 (Fig. 3B), a small molecule that inhibits IκBα phosphorylation and p65 nuclear translocation (30), or with a p65 decoy peptide that specifically competes with and blocks phosphorylation of S276 on the endogenous p65 protein (Supplementary Fig. S2B). Thus, Nutlin-3 represents a useful tool to probe the effect of dual p53 and NF-κB activation on gene regulation in macrophages.
Nutlin-3 treatment induces p53 and p65 binding to the promoter region of IL-6
In silico analysis of the IL-6 gene promoter region revealed a potential p53 response element (p53RE) consisting of a decamer half-site located 362 base pairs (−362 bp) upstream from the TSS in addition to a previously reported κB site (−138 bp; Fig. 3C; ref. 31). Because we previously established p53 responsiveness at half-sites (32), we performed ChIP assays to determine whether p53 and NF-κB could cooperate in cis to mediate IL-6 expression. As depicted in Fig. 3D and E, both p53 and p65 bound to the promoter of the IL-6 gene after 1 hour of Nutlin-3 treatment of monocyte-derived macrophages. However, when compared with p65, occupancy of p53 was lower; the levels of DNA pulled down by the p65 ChIP were 0.7%–1.25% of the input as compared with 0.3% to 0.5% of the input for p53 ChIP. One possible explanation for weaker p53 binding is that the p53RE is a half-site, which generally shows less binding than full sites (32). As expected, p53 and p65 also bind to the p21 promoter (Supplementary Fig. S2C), as previously shown in other cell systems (33, 34). Taken together, these data show that both p65 and p53 bind to regions near the IL-6 gene, suggesting that they work together in cis to regulate IL-6 gene transcription and expression.
Chemokines are highly induced by both Nutlin-3 and LPS
We hypothesized that if p53 and NF-κB coregulate IL-6 gene transcription in human macrophages, there is likely a larger set of genes that are also subject to the coregulation. To investigate this possibility, we performed microarray analyses with monocyte-derived macrophages from two different donors treated with either LPS or Nutlin-3 for 2 hours. We reasoned that because LPS is a well-known activator of NF-κB, genes regulated by both LPS and Nutlin-3 might have responses similar to that of IL-6. We chose to use macrophages from donors with similar demographics (Supplementary Fig. S3A).
Principal component analysis showed that the microarrays were of good quality and that even with similar demographics the donors grouped separately for all treatments (Supplementary Fig. S3B). Nonetheless, gene regulation in response to 2-hour treatments with either Nutlin-3 or LPS was remarkably similar in both donors as shown in the heat map in Supplementary Fig. S3C. Although most studies have addressed p53 networks with Nutlin-3 at later times (35), we also find elements of p53 signature (such as induction of p21 and Mdm2) at the much shorter time of 2 hours in macrophages. Because our goal was to identify genes regulated like IL-6, we chose to focus only on the upregulated genes with a threshold 1.7-fold change (Nutlin-3/DMSO or LPS/Untreated). Strikingly, most of the genes upregulated by Nutlin-3 (155/175) were also upregulated by LPS (Fig. 4A). Thus, by using Nutlin-3 as an agent to activate p53 and LPS to identify NF-κB–driven changes in gene expression, we identify a subset of 155 genes that are candidates for cooperative regulation by p53 and NF-κB.
Gene ontology analysis of the gene subset upregulated by both Nutlin-3 and LPS revealed that approximately 40% of these upregulated genes could be classified as functioning in the immune response where immune and inflammatory response functions were the most significant (with the lowest P values) as determined by the g:profiler program (Supplementary Table S2; ref. 36). Importantly, if genes within this subset are indeed regulated in a similar manner as IL-6, then there should be response elements for NF-κB and p53 near these genes. In a search for conserved transcription factor binding sites 1 kB upstream of the TSS for each gene we found several that would be capable of binding NF-κB subunits (Supplementary Table S3). The motif with the most significant P value was the p65 motif (P value of 1.58e−7; Supplementary Table S3). Because conventional prediction programs are poor identifiers of p53REs, we used the p53 Scan software to search for predicted p53REs and found that a large subset of the 155 genes had associated p53REs (40/155 genes), and 25/155 genes had both associated p53RE and κB sites (Supplementary Table S4).
Further classification of the 155 gene subset based on the level of fold change revealed that several of the genes were highly upregulated (more than 10-fold) similar to IL-6 (Fig. 4B). As expected, IL-6, IL-8, and TNFα were highly induced by both Nutlin-3 and LPS (Supplementary Table S5). In addition, chemokines including CXCL1, CCL20, IL-8, CCL4, and CXCL3 were among the highest induced genes after Nutlin-3 treatment (Supplementary Table S5, the * refers to chemokines). Induction of these chemokines was validated with real-time (RT)-PCR (Fig. 4C and D), and several of these chemokines have potential p53REs, some of which are listed in Supplementary Table S6. Some genes, like CXCL3, have multiple p53REs. In addition, CCL5 and CCL3 had perfect (i.e., no mismatches) half-site p53REs (bold letters; Supplementary Table S5). p53 and NF-κB are both required for Nutlin-3 induction of these chemokines, because inhibition of p53 or NF-κB by pretreatment with PFT-α or Bay-11-7082, respectively, significantly reduced Nutlin-3–induced CXCL3, CCL3, and CCL4 levels (Supplementary Fig. S3D). Together, our data suggest that these chemokines are regulated similarly to IL-6 in primary human macrophages.
Nutlin-3 treatment of macrophages enhances chemokine secretion and induces neutrophil migration
Having established that activation of p65 and p53 in primary human macrophages enhances the expression of specific cytokines and chemokines, we investigated the biologic consequences. Because chemokines such as CXCL1, IL-8, and CXCL3 secreted from macrophages can bind to receptors on the surface of neutrophils to promote migration (37), we tested whether Nutlin-3 treatment increased chemokines in the culture supernatant and enhanced neutrophil migration. As shown in Fig. 5A and Supplementary Fig. S1B, treatment of macrophages with Nutlin-3 and LPS (used as a positive control) for 24 hours induced robust CXCL1 and IL-8 secretion.
To determine whether these secreted chemokines are functional, a migration assay was performed to test whether the chemokine-rich culture supernatant resulting from Nutlin-3 treatment was capable of promoting neutrophil migration. Migration of primary human neutrophils toward supernatants from macrophages treated for 24 hours with Nutlin-3 or LPS was significantly higher than the untreated and DMSO control supernatants (Fig. 5B). Taken together, these data show coregulation of cytokine and chemokine expression by p53 and NF-κB in primary human macrophages. While the role of NF-κB in this process is well described, our data highlight a novel role of p53 in macrophage function.
p53 is activated and drives IL-6 expression in tumor-conditioned macrophages
Having identified a subset of chemokines and cytokines regulated by p53 in normal macrophages we next assessed whether p53 plays a similar role in gene regulation in TAMs. To address this issue, we used an established in vitro model system of TAMs in which human peripheral blood monocytes (PBMC) or isolated monocytes are differentiated into macrophages by culturing them with conditioned media from cultured cancer cells lines or primary tumor cells, resulting in TCMs (4, 5). Confirming successful induction of TCMs, macrophages cultured under these conditions exhibited a distinct, elongated morphology compared with normal macrophages cultured under standard conditions (MAC) and showed upregulation of several TAM/TCM markers such as IL-6, IL-8, CXCL-1, and CCL2 (Fig. 6A and B; ref. 38). Interestingly, most of these TAM/TCM markers, including IL-6, are genes we had identified as coregulated by p53 and NF-κB and have also been shown to contribute to tumorigenesis (4, 5, 39). However, while NF-κB activity has been shown to regulate the expression of these genes in TAMs/TCMs (3), the role of p53 in this process is unknown. We first tested whether p53 was activated in TCMs and whether it contributes to their high basal expression levels of IL-6. As shown in Fig. 6C, p53 protein levels are indeed higher in TCMs compared with normal MACs. Furthermore, four classical p53 targets, p21, Mdm2, Puma, and Bax, were consistently increased between 2- and 10-fold in TCMs compared with MACs depending on the gene and donor, suggesting that the stabilized p53 is functional. Interestingly, whereas p53 activation generally occurs through stabilization of p53 protein, the TCMs showed increased p53 mRNA compared with MACs, which could contribute to the larger level of p53 proteins (Supplementary Fig. S4A). Furthermore, p53 binding to the IL-6 promoter region was increased in TCMs, and treatment with PFT-α for 4 hours reduced IL-6 levels in TCMs (Fig. 6E and F). Interestingly, additional p53 stabilization by treatment with Nutlin-3 did not induce IL-6 in TCMs, which might be due to high IL-6 basal levels (Supplementary Fig. S4B). Taken together, these data indicate that p53, activated by paracrine factors released by tumor cells, promotes IL-6 expression in TCMs.
We have identified a novel role for p53 that is specific to the regulation of several proinflammatory genes in human macrophages, including IL-6, IL-8, and CXCL1. Importantly, NF-κB coactivation is essential for this regulation. While previous investigations have suggested the potential for p53/NF-κB interactions (15), our study is the first to investigate such interactions in human primary immune cells. Importantly, the only cells in which we were able to detect p53/NF-κB co-regulation of cytokines were macrophages and precursor monocytes. Presented in Fig. 7 is a model summarizing our finding on potential p53 and NF-κB responses in normal macrophages and TAMs. In macrophages associated with normal tissue, enhanced cytokine/chemokine induction due to activated p53 and NF-κB may lead to neutrophil recruitment and inflammation. Given that p53 and NF-κB are activated by many exogenous stimuli, this coregulatory transcriptional mechanism might serve as an important amplifying element in the macrophage's response to a wide array of tissue-damaging insults. In the tumor microenvironment, macrophage p53 appears to be activated in response to paracrine tumor signals and, in part, facilitates IL-6 expression. In this context, the tumor-derived growth factors may induce a p53 response in the macrophages, which in turn alters cellular behavior of the tumor cells. This p53 response could possibly have some beneficial consequences in tumors, which are often p53 defective. On one hand, p53 may function to promote senescence, especially because p53 and NF-κB have been shown to cooperate in senescence and NF-κB is important in cytokine/chemokine gene regulation in TAMs (3, 40). On the other hand, the protumor effects of IL-6 and chemokines, including CXCL1, on tumor cell survival, angiogenesis, and recruitment of protumor tumor-associated neutrophils (TAN) have been well described (3). Thus, p53 within TAMs might contribute to tumor promotion through supporting a paracrine axis of secreted cytokines and chemokines (including IL-6 and CXCL1) between TAMs and tumor cells. The impact of activated p53 on cytokine/chemokines in TAMs along with the tumorigenic consequences as described in our model will be of interest in future studies.
Given the rapid kinetics of proinflammatory gene induction and the high levels of IL-6 induction, our findings may also be more broadly relevant to tissue responses induced during p53-activating cancer therapies. For example, circulating IL-6 is detected early after radiotherapy and can be used as a predictive diagnostic test for radiation pneumonitis (41, 42). Here, we found that the level of IL-6 induction by p53 and NF-κB activation was among the highest of all cytokines and occurred within 2 hours in macrophages from all donors. Our finding that the magnitude of IL-6 induction is donor-dependent implies that there may be underlying genetic determinants of IL-6 activation by p53 and NF-κB. For example, there may be polymorphisms in one or both of these transcription factors or their response elements. Our finding that the induction of IL-6 was dramatically more variable than that of the classical p53 target gene p21 may suggest that polymorphic p53 response elements in the IL-6 promoter are likely. We propose that such genetic variants, if identified, could perhaps be used as predictors of the inflammatory response to chemotherapy or radiotherapy.
This is the first report, to our knowledge, showing that Nutlin-3 activates NF-κB. While the mechanism of Nutlin-3 activation needs to be further defined, it may be through the DNA damage response (DDR). Dey and colleagues (43) have reported that Nutlin-3 can inhibit rather than enhance NF-κB function via a p53-dependent mechanism; however, that was found in lung cancer cells. Nutlin-3 activation of NF-κB is likely cell type-dependent and may depend on whether Nutlin-3 initiates the DDR, which also appears to be determined by cell type and Nutlin-3 dose (44–47). Furthermore, Nutlin-3 and a similar small molecule, RITA, which has also been shown to induce the DDR (48), are currently being evaluated for cancer therapy and may promote macrophage-mediated inflammatory responses that could serve as adjuvants in anticancer responses.
An intriguing finding in our study is that p53 and NF-κB work together to regulate gene expression, contrary to most previous studies. There have been a few reports of small degrees of p53 and NF-κB cooperativity (14, 15) unlike that seen here, and the nature of the p53/NF-κB is likely context-dependent (27). Most of the research describing the opposing interaction between p53 and NF-κB has been conducted in either human cancer cells or mouse primary cells. Few reports have addressed p53/NF-κB interactions in primary human cells, especially in relation to macrophage function during chemotherapy. Our finding that proinflammatory gene induction by Nutlin-3 is limited to primary human macrophages and monocytes highlights the fact that interactions between p53 and NF-κB are different depending on context, for example, cancer cells versus noncancerous “normal” cells and lymphocytes versus macrophages. The dramatic and specific responsiveness of macrophages identifies programmatic differences from other cells, possibly to assure ROS-enhanced immune responses. Because tissue macrophages are also relatively resistant to genotoxic injury (49), our results suggest that tissue macrophages may serve a dual function as sensors and as effectors of an amplified response to genotoxic injury. Expression of p53 in response to DNA damage could potentiate NF-κB–induced inflammatory response and immune cell recruitment. The underlying mechanisms for assuring these programmatic differences between macrophages and other cell types as well as the absence in rodents are intriguing issues that have larger regulatory as well as evolutionary implications.
Another unexpected finding is that p53 is stabilized in TAM-like macrophages and enhances IL-6 expression. While the roles of wild-type and mutant p53 are well described in tumorigenesis, there have been no studies to our knowledge addressing p53 functions in tumor-associated cells like TAMs. p53 is commonly mutated in tumor cells; however, TAMs presumably have wild-type p53 because they originate from circulating monocytes that are recruited into the tumor microenvironment by soluble factors (3). This suggests that the role of p53 in TAMs is likely different from its role within tumor cells. Here, we show that p53 is activated in the absence of exogenous stress in TAM-like cells, but the mechanism of this activation is currently unknown. Because NF-κB is activated in TAMs and can bind to κB sites in the TP53 gene to regulate p53 transcription (3, 50), one possibility is that NF-κB enhances p53 transcription in TAMs. It is also possible that signaling within the macrophage by tumor cell-secreted soluble factors like M-CSF and CCL2 (4) activates p53 independently of NF-κB. Furthermore, the role of p53-driven proinflammatory gene expression in TAMs is unclear. As mentioned above, p53 within TAMs could conceivably have protumor functions. However, it is also possible that macrophage p53 could promote senescence in TAMs, as secretion of cytokines and chemokines is associated with senescence (senescence associated secretory phenotype) and there is continuous turnover of TAMs in the tumor microenvironment (3). Thus, the mechanism of p53 activation in TAMs, the functional role of p53 in TAMs, as well the role of TAM p53 in tumorigenesis warrant further investigation.
In summary, our results show that p53 and NF-κB together drive expression of cytokines and chemokines, including IL-6 and CXCL1, in macrophages. The secretion of these factors may have distinct biologic effects depending on the microenvironment, promoting inflammation in normal tissues upon injury and modifying cancer cell responses in the tumor microenvironment. Importantly, we found that Nutlin-3 had no apparent effect on the expression of IL-6 in TAM-like cells, possibly because p53 is already present. This may suggest that systemic treatment with p53-activating chemotherapeutics might enhance antitumor responses of normal macrophages without augmenting protumor functions of TAMs. The net effect of TAM p53 upon tumorigenesis remains to be defined. Nonetheless, our studies suggest that therapies based on p53 activation may impact tumor cell biology through p53-positive macrophages even in patients with tumors lacking functional p53.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Conception and design: J.M. Lowe, D. Menendez, S. Garantziotis, M.B. Fessler, M.A. Resnick
Development of methodology: J.M. Lowe, M.A. Troester
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): J.M. Lowe, M. Shatz, E.L. Kirk, M.A. Troester, S. Garantziotis, M.B. Fessler
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J.M. Lowe, D. Menendez, P.R. Bushel, M. Shatz, E.L. Kirk, M.A. Troester, M.B. Fessler, M.A. Resnick
Writing, review, and/or revision of the manuscript: J.M. Lowe, D. Menendez, P.R. Bushel, M. Shatz, M.A. Troester, S. Garantziotis, M.B. Fessler, M.A. Resnick
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): J.M. Lowe
Study supervision: D. Menendez, S. Garantziotis, M.B. Fessler, M.A. Resnick
This study was supported by funding from NIH, NIEHS, Breast Cancer and the Environment Research Program grant, U01-ES109472 (M.A. Troester and E.L. Kirk) and the NIEHS Intramural Research Program.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The authors thank Andrew Ghio, Annette Rice, and Jamie Marshburn for performing macrophage/monocyte isolations and Dr. Shyamal Peddada for statistical analysis assistance.
Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).
- Received April 12, 2013.
- Revision received February 5, 2014.
- Accepted February 9, 2014.
- ©2014 American Association for Cancer Research.