Diminished Expression of Transcription Factors Nuclear Factor κB and CCAAT/Enhancer Binding Protein Underlies a Novel Tumor Evasion Mechanism Affecting Macrophages of Mammary Tumor–Bearing Mice

Animals and tumors

Male and female BALB/c mice of 10 to 14 weeks of age, maintained by brother-sister mating in our facilities, were used. In all experiments, the mice were age and sex matched. The D1-DMBA-3 mammary adenocarcinoma is a transplantable tumor derived from a non-viral, non-carcinogen-induced preneoplastic alveolar nodule in a BALB/c mouse treated with 7,12-dimethylbenzanthracene (17). The immunogenic tumor is routinely transplanted in BALB/c mice by s.c. injection of 1 × 10⁶ tumor cells. Our institutional animal care and use committee approved the animal experiments.

Macrophage collection and culture

Normal mice and 4-week tumor-bearing mice were injected i.p. with 1.5 mL of 3% thioglycollate. On day 4, the peritoneal exudate cells were obtained by peritoneal lavage with 20 mL of ice-cold RPMI 1640; PEM from normal animals (N-PEM) and tumor bearers (T-PEM) were later purified and cultured as previously described (18).

Reagents

RPMI 1640 containing 10% FCS, 2 mmol/L l-gluthamine, and 100 units of penicillin and 100 µg/mL of streptomycin (all from Hyclone, Logan, UT) was used as culture medium in all experiments. LPS from Escherichia coli 055:B5 was purchased from Sigma-Aldrich (St. Louis, MO).

Cytokine ELISA

PharMingen (San Diego, CA) OptEIA sets for mouse IL-12p70 heterodimer, mouse IL-12p40 monomer, and mouse IL-10 were used to measure the amounts of IL-12p70, IL-12p40, and IL-10 proteins in the supernatants of PEM according to the instructions of the manufacturer.

Northern blot

Northern hybridizations were done as previously reported (11) with the following modifications: Total RNA was isolated using RNAwiz reagent (Ambion, Austin, TX) according to the instructions of the manufacturer and RNA samples (15 µg) were electrophoresed, transferred, and hybridized using the NorthernMax-Gly set, a glyoxal-based system for Northern blots from Ambion, following the instructions of the manufacturer. The cDNA probes for IL-12p40 and p35 genes, a generous gift from Dr. Giorgio Trinchieri (Laboratory of Parasitic Diseases, National Institute of Allergy and Infectious Diseases, Bethesda, MD), were radiolabeled with [α-³²P]dATP (3,000 Ci/mmol; New England Nuclear/Perkin Elmer, Boston, MA) by random primer labeling (Prime-a-Gene, Promega, Madison, WI). β-Actin levels were similarly examined using β-actin cDNA (Clontech Laboratories, Inc., Palo Alto, CA) and served as controls to normalize RNA quantity.

Electrophoretic mobility shift assay

Isolation of nuclear extracts, sequences, preparation of oligonucleotide probes, electrophoretic mobility shift assay (EMSA), and supershift assays were conducted as previously described (15, 16); all antibodies used were from Santa Cruz Biotechnology (Santa Cruz, CA).

Western blot

Whole-cell and nuclear extracts from N-PEM and T-PEM were used. To isolate whole-cell extracts, PEM were lysed using cold radioimmunoprecipitation assay buffer supplemented with protease inhibitor cocktail (Roche, Indianapolis, IN) and sodium vanadate (Roche, 1 mmol/L final concentration) as described in ref. 11. Nuclear extracts were isolated as described in ref. 15. Protein concentration was determined using the BCA Protein Assay Kit (Pierce, Rockford, IL) and Western blots were done as previously described (11). The primary antibodies used were rabbit anti-mouse polyclonals from Santa Cruz Biotechnology, except for IκB kinase (IKK)-β (mouse monoclonal from Imgenex, San Diego, CA) and for the phosphorylated form of IκBα (mouse monoclonal antibody from Santa Cruz Biotechnology); the presence of actin was detected by the use of a rabbit anti-mouse polyclonal antibody (Sigma-Aldrich). For the mouse monoclonal antibodies used, secondary antibodies were goat anti-mouse isotype matched to the corresponding primary antibodies.

Both protein and mRNA levels of proinflammatory molecules are decreased in peritoneal macrophages from tumor bearers

LPS-activated T-PEM from 4-week tumor bearers exhibit an impaired capacity to produce the IL-12p70 heterodimer as measured by ELISA compared with the IL-12 expression in N-PEM (Fig. 1A). Because IL-12 is composed of two chains, p40 and p35, we sought to determine whether the IL-12p40 protein synthesis/secretion was also affected in these cells (there is no available antibody to study p35 protein expression). IL-12p40–specific ELISAs were done using 24-hour supernatants of macrophages from normal or 4-week tumor-bearing mice cultured alone or with LPS. As seen in Fig. 1B, LPS-activated T-PEM exhibited a down-regulated IL-12p40 monomer production compared with similarly activated cells from normal animals, accounting, at least in part, for the decreased production of IL-12p70 in these cells. It was observed that the levels of IL-12p40 were higher than the levels of IL-12p70 produced by both cell types, as expected (19). Our previous studies have shown that the protein levels of iNOS in T-PEM stimulated with LPS are also down-regulated (11).

To analyze whether IL-12p40 and p35 mRNA expressions were affected in T-PEM, Northern blot experiments were done. The results of these Northern hybridizations (Fig. 1C) showed a substantial decrease in the expression of the p40 gene, together with a slightly diminished p35 gene expression in LPS-activated T-PEM, as compared with similarly activated N-PEM. Furthermore, no alternatively spliced or aberrant mRNAs were observed in T-PEM for either gene (i.e., only the expected transcripts of 2.3 kb for p40 and 1.3 kb for p35 were detected). As was the case with the protein, we have also previously shown that the iNOS mRNA levels are decreased in T-PEM as compared with the levels in N-PEM (11).

Peritoneal macrophages from tumor bearers exhibit decreased nuclear factor κB and CCAAT/enhancer binding protein bindings to gene promoters of proinflammatory genes

To further analyze the mechanism of the decreased mRNA expression of the IL-12 genes in T-PEM, we studied the functionality of the most relevant transcription factors involved in their expression. Because the IL-12p40 gene was shown to be the most affected component involved in the IL-12 deficiency in T-PEM, we focused our analysis in this particular gene. The mouse IL-12p40 gene promoter has been well characterized (20–22) and the transcription factors NFκB and C/EBP are among the major ones involved in its regulation. A study of the binding activities of these two transcription factors to the IL-12p40 promoter was done by EMSA with nuclear protein extracts isolated from N-PEM and T-PEM, with and without LPS activation, using synthetic double-stranded oligonucleotides matching known sequences of the mentioned transcription factors sites on the mouse IL-12p40 gene promoter. Our results, using a ³²P probe for the NFκB “half site” of the IL-12p40 murine promoter (20), revealed a profound decrease in the binding activity of NFκB in LPS-activated T-PEM (Fig. 2A) compared with N-PEM. Competition experiments with cold NFκB and irrelevant Oct-1 binding sites showed the specificity of the binding to the NFκB site. Supershift experiments were done with nuclear extracts of LPS-activated N-PEM and T-PEM to characterize the contribution of the different members of the NFκB transcription factor family in these cells using antibodies against the various NFκB members (p50, p52, p65, c-rel, and relB; Fig. 2B). These experiments revealed that p50, p65, and c-rel compose the two major NFκB complexes (designated a and b, respectively) in PEM as addition of antibodies against these three components caused alterations in the mobility of these complexes. Band b, with the increased mobility, represents p50/p50 homodimers whereas the slower band a reflects the presence of p50/p65 and p50/c-rel heterodimers. T-PEM exhibit a dramatic decrease in the intensity of band b, illustrating a significant reduction of p50/p50 in tumor bearers (Fig. 2B). Furthermore, minor participation of relB and p52 can also be observed as expressed by the lower density of the corresponding band patterns after the addition of the respective antibodies. Regarding iNOS, our data, using a ³²P probe for the NFκBd unique site from the NFκB consensus sequence (23), revealed a profound decrease in the binding activity of NFκB to the iNOS promoter in LPS-activated T-PEM compared with N-PEM (Fig. 3A). Supershift experiments showed, in both N-PEM and T-PEM activated with LPS, that nearly the entire NFκBd complex was shifted in response to anti-p50. Addition of antibodies against p65 and c-rel, respectively, induced a decrease in the intensity of the upper band of the duplex in each case. Antibodies against Rel B or p52 did not affect the mobility or the intensity of the complexes (Fig. 3B).

Regarding the C/EBP transcription factor, three major DNA-protein complexes were revealed using a synthetic probe for its binding site on the murine IL-12p40 promoter of N-PEM (21), designated a, b, and c in Fig. 2C. In T-PEM, these binding activities were strongly diminished in both resting and LPS-activated cells. Competition experiments with cold C/EBP and irrelevant Oct-1 binding sites showed a specific binding for the C/EBP species. Supershift analyses were done with the nuclear extracts of LPS-activated N-PEM and T-PEM. Figure 2D shows a representative experiment done with antibodies against the different members of this transcription factor family, C/EBPα, β, and δ. The three main DNA-protein complexes formed with the C/EBP synthetic probe correspond to homodimers and heterodimers between C/EBPα, β, and δ (21). Our data revealed a major involvement of α and β and, to a lesser extent, of δ in both cell types. No cooperation between C/EBP with NFκB or activator protein transcription factors was observed in either cell type. Concerning iNOS, we analyzed the binding of nuclear proteins to the NF-IL6 (−153/−142) site of the iNOS promoter (24). Nuclear extracts from T-PEM, as compared with N-PEM, show a significantly diminished binding to the C/EBP sequence in the constitutive as well as the inducible binding after stimulation with LPS (Fig. 3C). Supershift experiments revealed, in both N-PEM and T-PEM activated with LPS, that antibodies against C/EBPα and β clearly affected the mobility or the intensity, respectively, of the band complex whereas antibody against C/EBPδ did not induce major changes (Fig. 3D)

Diminished binding activities of transcription factors nuclear factor κB and CCAAT/enhancer binding protein are associated with their decreased protein expressions in peritoneal macrophages from tumor bearers

Several reasons might account for the fact that NFκB and C/EBP isolated from T-PEM bind deficiently to synthetic oligonucleotides representing their actual sites in the IL-12p40 and iNOS promoters. One of them is the existence of diminished amounts of these transcription factors in T-PEM. Using whole-cell extracts isolated from resting N-PEM and T-PEM, we determined by Western blot analysis whether there was a down-regulated protein production of NFκB and C/EBP in T-PEM. Figure 4A shows the results of a representative experiment. NFκB c-rel protein is profoundly down-regulated in T-PEM, followed by NFκB p50 and p65 to lesser extents. In addition, C/EBPα and β exhibit decreased expressions in these cells as compared with nonactivated N-PEM.

Peritoneal macrophages from tumor bearers do not exhibit altered amounts of IκBα or its phosphorylated form

Another reason that could account for the decreased binding activity of NFκB in T-PEM is the existence of altered amounts of Iκβα protein or its phosphorylated form in these cells. This possibility was examined by Western blot using whole-cell extracts of resting as well as LPS-activated N-PEM and T-PEM. Iκβα was expressed constitutively in both N-PEM and T-PEM at similar levels (Fig. 4B); yet, on LPS activation, its amount decreased significantly to similar levels in both cell types. When the phosphorylated form of IκBα (p-IκBα) was examined, it was observed that both cell types exhibited increased levels of phosphorylated IκBα on LPS activation; however, there were no prominent differences in its expression between N-PEM and T-PEM. Thus, both IκBα and its phosphorylation remain intact in T-PEM and do not explain the down-regulated binding activity of NFκB observed in T-PEM. The presence of IKKα and IKKβ proteins, the two kinase domains of the IKK complex responsible for the IκB phosphorylation (25), was also assessed in both groups of macrophages. As shown in Fig. 4B, IKKα was strongly expressed in both N-PEM and T-PEM but was not modulated by LPS in either cell type, and there were no differences in its expression levels between N-PEM and T-PEM. On the other hand, IKKβ protein could be detected neither in N-PEM nor in T-PEM.

Nuclear factor κB and CCAAT/enhancer binding protein nuclear translocations are reduced in peritoneal macrophages from tumor bearers

The diminished binding activities of NFκB and C/EBP to the IL-12p40 and iNOS promoter regions in T-PEM may be also due to their decreased nuclear translocations in these cells. To examine this possibility, we analyzed nuclear extracts from N-PEM and T-PEM, with and without activation with LPS, by Western blotting. Figure 4C shows that unstimulated N-PEM show very little NFκB nuclear protein whereas after LPS activation for 1 hour, NFκB p50, p65, and c-rel translocate to the nucleus. However, a dramatic decrease in the amounts of p65, p50, and c-rel proteins can be observed in the nucleus of T-PEM after 1 hour of LPS stimulation, especially in the case of p65. Furthermore, the nuclear translocation of both C/EBPα and β was substantially reduced in T-PEM as compared with N-PEM, especially C/EBPα, although LPS did not modulate C/EBP in either cell type, as was also shown by EMSA.

Although peritoneal macrophages from tumor bearers exhibit a decreased production of interleukin-12, they do not seem to be M2 macrophages

Low IL-12 production accompanied by increased IL-10 expression is associated with macrophages of the M2 phenotype (26). To analyze whether T-PEM isolated from mice bearing advanced (4-week) tumors behave as M2 macrophages, their IL-10 expression levels were studied. Figure 5 shows that the production of IL-10, although somewhat variable, revealed no significant differences between the amounts produced by N-PEM and the levels expressed by T-PEM at various times after tumor implantation.