Peripheral blood monocytes migrate to and accumulate in hypoxic areas of inflammatory and tumor lesions. To characterize the molecular bases underlying monocyte functions within a hypoxic microenvironment, we investigated the transcriptional profile induced by hypoxia in primary human monocytes using high-density oligonucleotide microarrays. Profound changes in the gene expression pattern were detected following 16 h exposure to 1% O2, with 536 and 677 sequences showing at least a 1.5-fold increase and decrease, respectively. Validation of this analysis was provided by quantitative RT-PCR confirmation of expression differences of selected genes. Among modulated genes, 74 were known hypoxia-responsive genes, whereas the majority were new genes whose responsiveness to hypoxia had not been previously described. The hypoxic transcriptome was characterized by the modulation of a significant cluster of genes with immunological relevance. These included scavenger receptors (CD163, STAB1, C1qR1, MSR1, MARCO, TLR7), immunoregulatory, costimulatory, and adhesion molecules (CD32, CD64, CD69, CD89, CMRF-35H, ITGB5, LAIR1, LIR9), chemokines/cytokines and receptors (CCL23, CCL15, CCL8, CCR1, CCR2, RDC1, IL-23A, IL-6ST). Furthermore, we provided conclusive evidence of hypoxic induction of CCL20, a chemoattractant for immature dendritic cells, activated/memory T lymphocytes, and naive B cells. CCL20 mRNA up-regulation was paralleled by increased protein expression and secretion. This study represents the first transcriptome analysis of hypoxic primary human monocytes, which provides novel insights into monocyte functional behavior within ischemic/hypoxic tissues. CCL20 up-regulation by hypoxia may constitute an important mechanism to promote recruitment of specific leukocyte subsets at pathological sites and may have implications for the pathogenesis of chronic inflammatory diseases.
Peripheral blood monocytes (Mn)3 represent the early mononuclear phagocyte component of the leukocyte infiltrate at sites of inflammation, infection, and tumor growth, where they differentiate into inflammatory and tumor-associated macrophages (Mf) (1). Mn/Mf are potent regulators of immune and inflammatory reactions. They orchestrate the coordinated recruitment and activation of specific leukocyte subsets at pathological sites through the local secretion of low m.w. structurally related proteins, termed chemokines (1). Chemokines are classified into CXC, CC, C, and CX3C families, which bind to and activate members of a superfamily of 7-transmembrane domain, G protein-coupled receptors differentially expressed and regulated in leukocytes (2). CCL20 (also known as MIP-3α, liver and activation regulated chemokine, and Exodus) is a recently described Mn-derived CC-chemokine which selectively attracts immature dendritic cells (iDC), effector/memory T lymphocytes, and naive B cells through its specific receptor, CCR6, expressed on these cells (for a review, see Ref. 3).
Mononuclear phagocyte reactivity in pathological tissues is finely tuned by a complex interplay between stimulatory and inhibitory signals of various nature that include immune-derived stimuli (4, 5), viral/bacterial products (5, 6), cell metabolites (4, 7), and tissue-specific signals (8). A common denominator of many pathological processes and an important regulator of gene expression is represented by low partial oxygen pressure (pO2) (reviewed in Ref. 9). Hypoxia occurs in cardiovascular, hematological, and pulmonary disorders, ischemic wounds, arthritic joints, atherosclerotic plaques, and microbial infections, and experimental and clinical studies point toward its fundamental role in the pathogenesis of these diseases (8, 9, 10, 11). Areas of low pO2 are also present in solid tumors, where they have been associated with malignant progression, metastasis formation, resistance to therapy, and poor clinical outcome (8, 12, 13, 14). Transcriptional response to hypoxia is mediated primarily by the hypoxia-inducible factor-1 (HIF-1), a heterodimeric basic helix-loop-helix (bHLH) transcription factor composed of HIF-1β (also known as the aryl hydrocarbon receptor nuclear translocator (ARNT)), the constitutive subunit, and HIF-1α, 2α, or 3α, the oxygen-sensitive subunits (9, 15). The α subunits are posttranslationally stabilized under hypoxia and translocate to the nucleus where they dimerize with HIF-1β, transactivating the hypoxia responsive element present in the promoter of many O2-sensitive genes (9, 15). Regulation of HIF-1 expression and activity by hypoxia is a tightly regulated process which results from the activity of several oxygen-dependent enzymes and requires interaction and cooperation with various transcriptional cofactors and other transcription factors (15).
Mononuclear phagocytes accumulate preferentially in hypoxic/ischemic areas of diseased tissues (1), and hypoxic conditions have been shown to profoundly affect their proinflammatory and immunoregulatory responses by modulating the expression of genes coding for angiogenic factors, inflammatory cytokines, and extracellular matrix (ECM) components/regulators (reviewed in Ref. 1). Recent evidence indicates that hypoxia can also strictly control the chemokine network in cells of the monocytic lineage not only by regulating the production of specific chemokines but also controlling their action through the modulation of their receptors. Up-regulation of CCL3 (16), CXCL1 (1), CXCL8 (1), and CXCR4 (17), and inhibition of CCL2 (16) and CCR5 (18) under hypoxia were reported.
In the last few years, microarray technology has become an important tool for the characterization of the molecular bases underlying cell response to stimulation (19). Recent investigations have defined the transcriptional profile induced by hypoxia in in vitro-derived human Mf (hMDM) (20, 21). Given their critical role in the regulation of the initial phases of inflammation (1), it was important to study primary Mn as a model of the early response to the hypoxic environment. In this study, we report the first transcriptome analysis of primary human Mn following hypoxia exposure. Our results reveal the regulation by hypoxia of a cluster of novel genes relevant to inflammation and immunity coding for surface molecules/markers, inflammatory cytokines/chemokines, and their receptors, and identify CCL20 as a new hypoxia-inducible gene.
Materials and Methods
Cells and culture conditions
PBMC were isolated from platelet apheresis of healthy donors, obtained by the Blood Transfusion Center of the Gaslini Institute (Genova, Italy), by density gradient centrifugation over a Ficoll cushion (Ficoll-Paque PLUS; Amersham Biosciences). Mn were then purified by countercurrent centrifugal elutriation using a Beckman JE-6 elutriation chamber and Avanti J-20XP rotor system (Beckman Coulter), as described (22), followed by MACS magnetic bead separation (Human Monocyte Isolation kit-II; Miltenyi Biotec). The purity of Mn preparations was ≥95%, as assessed by morphology on Giemsa-stained cytocentrifuge slides and flow cytometry with anti-CD14 mAb. Viability, determined by flow cytometry after DNA staining with propidium iodide (5 μg/ml in PBS), was >98%.
Mn were plated in Costar plates (Celbio) in RPMI 1640 (Euroclone; Celbio) supplemented with 10% heat-inactivated FCS (HyClone; Celbio), 2 mM l-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin (Celbio) and maintained at 37°C in a humidified incubator containing 21% O2, 5% CO2, and 75% N2, referred to as normoxic conditions. Hypoxic conditions (i.e., 1% O2) were achieved by incubating and handling the cells at 37°C in a humidified anaerobic work station incubator (Bug Box; ALC International) flushed with a mixture of 94% N2, 5% CO2, and 1% O2. Culture medium was allowed to equilibrate for 3 h in a loosely capped flask in the hypoxic incubator before cell exposure, and pO2 was monitored using a portable trace oxygen analyzer (Oxi 315i/set; WTW). The pO2 in normoxic medium ranged between 149 and 150 mm Hg, values that correspond to a 21% O2 concentration in an aqueous solution at 37°C and at a barometric pressure of 760 mm Hg, whereas the pO2 attained in the medium under hypoxic conditions was ∼7.1 mm Hg, which is equivalent to an O2 concentration of 1% and is in the range of the hypoxic levels found in inflammatory tissues (8, 9, 10, 11, 12). The endotoxin content, determined by the Limulus amebocyte lysate test (QCL-1000; Bio-Whittaker), was <0.125 endotoxin units/ml in all reagents.
RNA isolation and cRNA synthesis
24 primer (high purity salt-free purified) was used in place of the oligo provided with the kit. cDNA was purified and used for in vitro transcription with the BioArray High Efficiency RNA Transcript Labeling kit (Enzo Life Sciences) in the presence of biotin-11-CTP and biotin-16-UTP. Labeled cRNA was cleaned up using the Qiagen RNeasy Mini kit, checked for quality, and fragmented by incubation in mild alkaline buffer.
GeneChip hybridization and data analysis
23). Comparative analysis of hypoxic relative to normoxic expression profiles was conducted on GeneSpring 7.2 software (Silicon Genetics). Gene expression data for each replicate experiment were normalized using the “per chip normalization” and “per gene normalization” algorithms implemented in the GeneSpring program. First, each signal was normalized based upon the median signal in that chip (“per chip normalization”). Each corrected value was then normalized based upon the median of the measurements for that gene in all samples (“per gene normalization”). This normalization method, which removes the differing intensity scales and binding rates from multiple experimental readings, allows the comparison of multiple GeneChip hybridizations (24). Finally, gene expression levels of replicate experiments were averaged, and only genes that were modulated by at least 1.5-fold in hypoxic relative to normoxic samples (means of three experiments) were considered differentially expressed. The significance of gene expression differences between the two experimental conditions was calculated using a one-way ANOVA. Only genes that passed a Student’s two-sample t test at a confidence level of 95% (p value <0.05) were considered significant. The complete data set for each microarray experiment was lodged in the ArrayExpress website 〈www.ebi.ac.uk/arrayexpress〉. Gene Ontology (GO) data mining (25) for biological process at level 3 and Expression Analysis Systematic Explorer (EASE) biological theme analysis (26) were conducted online at 〈http://david.niaid.nih.gov〉 using the Database for Annotation, Visualization, and Integrated Discovery (DAVID) 2.0 program (27).
Real time quantitative PCR (qRT-PCR) of reverse-transcribed cDNA was performed on an I-Cycler (Bio-Rad), using iQ Supermix supplemented with 10 nM fluorescein (Bio-Rad), 0.1× Sybr-Green I (Sigma-Aldrich), and 300 nM sense and antisense oligonucleotide primers (TIBMolbiol). All primer pairs (listed in Table I⇓) were designed using Primer-3 software (28) from sequences in GenBank with a Tm optimum of 60°C and a product length of 80–150 nt and tested before use to confirm appropriate product size and optimal concentrations. qRT-PCR was conducted in triplicate for each target transcript under the following cycling conditions: initial denaturation of 3 min during which the well factor was measured, 50 cycles of 15 s at 95°C followed by 30 s at 60°C. Fluorescence was measured during the annealing step in each cycle. After amplification, melting curves with 80 steps of 15 s and 0.5°C increase were performed to monitor amplicon identity. Expression data were normalized on the values obtained in parallel for three reference genes (actin-related protein 2/3 complex 1B (ARPC1B); lysosomal-associated multispanning membrane protein-5 (LAPTM5); and thrombospondin 1 (THBS129). Relative expression values were calculated using Q-gene software (30).
A total of 1 × 105 Mn were applied to polysine glass slides by cytocentrifuging at 900 rpm for 5 min. Cytospin preparations were fixed in 4% paraformaldehyde in PBS for 30 min at room temperature, and permeabilized with 0.1% Triton X-100 in PBS for 5 min. Endogenous peroxidases were blocked with 0.3% hydrogen peroxide for 15 min. After rinsing in PBS, the slides were preincubated for 30 min in blocking buffer (PBS supplemented with 2% human AB serum) and then incubated for 1 h with anti-human CCL20 mAb or an isotype-matched mAb (IgG1; R&D System) in blocking buffer. mAbs were detected with DakoCytomation Envision+ System Labeled Polymer-HRP anti-mouse. Peroxidase staining was revealed by 3-amino-9-ethylcarbazole (DakoCytomation), as a substrate. Slides were counterstained with hematoxylin, coverslipped with 80% glycerol in PBS, and examined with a phase contrast microscope (Olympus Italia). Photomicrographs were taken with a Zeiss camera.
Gene expression profile of hypoxic Mn
Mn purified from 15 independent donors were cultured for 16 h in 1% O2, a condition previously shown to effectively modulate gene expression in these cells (17, 31), and the mRNA for the angiogenic factor, vascular endothelial growth factor (VEGF), was assessed by RT-PCR as an index of the response to hypoxia. Fig. 1⇓ shows VEGF mRNA levels in a representative subset of samples. Mn exposed to normoxic conditions expressed basal levels of VEGF mRNA, though showing some degree of donor-to-donor variation. Incubation under hypoxia caused a strong and consistent VEGF up-regulation in all the samples, in agreement with previous observations (31).
Functional classification of HMGs
Genes displaying at least 1.5-fold differential expression levels were classified into various categories based on the biological function(s) of the encoded protein to determine the global direction of the molecular response to hypoxia. According to GO data mining for biological processes (Fig. 2⇓), the transcriptional profile induced by hypoxia in Mn was mainly related to cell growth and/or maintenance, signal transduction, nucleic acid, and protein metabolism, these functional categories being the most enriched in both up- and down-regulated genes. Moreover, hypoxic Mn showed a prominent expression of genes involved in organogenesis, response to stress, biosynthesis, and phosphorus metabolism. Interestingly, a significant number of HMGs coded for proteins implicated in cell response to external stimuli, immune response, cell adhesion and motility, and cell-cell signaling, revealing a trend toward inflammation and immunity (Fig. 2⇓).
Comparison of microarray data with known hypoxia-induced changes in gene expression
In an initial verification of microarray data, we have cross-referenced our results with those of other studies investigating HMGs. As summarized in Table II⇑, a large cluster of known HMGs was affected by hypoxia. In addition to the reference gene VEGF, we demonstrated up-regulation of 20 genes related to angiogenesis, cell adhesion, transcription, and inflammatory responses, which have been previously identified as hypoxia inducible in cells of the Mn lineage (Table II⇑). The majority of them, including adrenomedullin (ADM), arginase-1 (ARG1), coagulation factor III (F3), fibronectin-1 (FN1), IL-1α (IL-1A), IL-6, TNF-α, macrophage migration inhibitory factor (MIF), matrix metalloproteinase-1 (MMP-1), osteopontin (SPP1), and VEGF receptor-1 (FLT1), were reported to be increased by hypoxia in hMDM and/or mouse Mf (20, 32), but not in primary Mn. The expression of another 46 known HMGs characterized in cells types other than mononuclear phagocytes was also triggered in hypoxic Mn (Table II⇑), including classical genes involved in glycolytic metabolism and glucose transport (e.g., glucose transporter 1 and 3 (GLUT1, 3) (9, 13, 15, 33), or associated with nonglycolytic metabolism and ion transport (e.g., carbonic anhydrase XII (CA12) (13, 33). Genes coding for two novel HMGs, hypoxia-inducible protein 2 (HIG2) (34) and HIF-1-responsive RTP801 (DDIT4) (15), and other HIF-1 target genes implicated in HIF-1α hydroxylation, such as EGL-nine homolog-1 (EGLN1) (35) and proline 4-hydroxylases-AI,II (P4HA1, A2) (33, 36), were increased in hypoxic Mn (Table II⇑). Finally, several genes related to apoptosis, cell cycle, transcription, and immune responses, whose responsiveness to hypoxia was previously demonstrated in normal and malignant cells, were also up-regulated in hypoxic Mn. BACH1 transcription repressor (37), bHLH domain-containing B2 transcription factor (BHLHB2/DEC1) (15), BCL2/adenovirus E1B-interacting protein 3 (BNIP-3) (15), hepatocyte growth factor receptor (MET) (15), IL-4 (38), IL-1 receptor antagonist (IL-1RN) (39), MAX-interacting protein-1 (MXI1) (33), and N-myc downstream-regulated gene-1 (NDRG1) (33) represent a few examples (Table II⇑). Interestingly, we also demonstrated inhibition by hypoxia of a cluster of inflammatory genes previously found down-regulated in mMf and hMDM, such as CCL2 (16), CCR5 (18), cathepsin C (CTSC) (21), 2,5-oligoadenylate synthetase (OASE) (40), and the ras family member, RAB7 (21). Overall, these data demonstrate that primary Mn share with Mf and cells of other lineages a cluster of hypoxia-responsive genes.
Identification of novel hypoxia-modulated genes in primary Mn
To identify genes not previously characterized in terms of responsiveness to hypoxia, the hypoxic transcriptome of primary Mn was further investigated after excluding known HMGs listed in Table II⇑. A list of selected novel hypoxia-modulated genes is presented in Table III⇑⇑⇑. Of interest, several were enzymes or other molecules implicated in lipid metabolism with a role in the regulation of fatty acid and/or cholesterol biosynthesis or transport (Table III⇑⇑⇑). The most relevant are: apolipoprotein B48 receptor (APOB48R, up-regulated), apolipoprotein E (APOE, down-regulated), fatty acid acyl-CoA synthase long-chain family 1 and 3 (ACSL1,3, down-regulated), fatty acid binding protein 4 (FABP4, up-regulated), low density lipoprotein receptor (LDLR, down-regulated), oxysterol binding protein-like 10 (OSBPL10, up-regulated), and steroid 27-hydroxylase (CYP27A1, down-regulated). Modulation of these genes under low pO2 may have implications for the pathogenesis of atherosclerosis and Alzheimer’s disease, where a role for hypoxia has been suggested (9, 10, 41).
Other genes differentially expressed in hypoxic Mn coded for cytoskeleton and ECM components/regulators (Table III⇑⇑⇑). Inducible genes include adhesive proteins of the desmosome type of cell-cell junction, actin-interacting/regulatory transmembrane molecules, and secreted proteins. Of note are galectin 8 (LGALS8), MMP-16 (MT3-MMP), MMP-19, regulator of G-protein signaling 1 (RGS1), tensin 1 (TNS1), and tissue factor pathway inhibitors 1 and 2 (TFPI1,2). Among the most significantly down-regulated genes, we identified advillin (AVIL), autotaxin (ENPP2), galectin 2 (LGALS2), myosin heavy polypeptide 11 (MYH11), and VAV3. Modulation of these genes is likely to regulate Mn adhesion, motility, and tissue remodeling.
The hypoxic profile also revealed hypoxia inducibility of a set of genes with transcription regulatory activity, including activating transcription factor (ATF)-2 and ATF-5, Fos homolog B (FOSB), Fos-like Ag 2 (FRA2), and runt-related transcription factors 1 and 2 (RUNX1,2) (Table III⇑⇑⇑). Moreover, several transcription factor- and cofactor-encoding genes were inhibited by hypoxia. Of relevance are the aryl-hydrocarbon receptor (AHR), ARNT2, CCAAT enhancer-binding protein α (C/EBPα), CREB-binding protein (CBP), HIF-1α, microphthalmia-associated transcription factor (MITF), MYC oncogene, nuclear receptor coactivator 1 (NCOA1/SRC-1), p53 tumor Ag (TP53), and zinc finger protein 197 (ZNF197/VHLaK) (Table III⇑⇑⇑). These data are indicative of major, coordinated changes in transcription and suggest the existence of both positive and negative O2-driven feedback regulatory mechanisms of hypoxia transcriptional response
Differential modulation by hypoxia of immune-related genes in Mn
As summarized in Table III⇑⇑⇑, a prominent set of novel HMGs have immunological relevance. These include genes encoding surface immunoregulatory signaling (IRS) receptors, such as early activation Ag (CD69), leukocyte membrane Ag CMRF-35H, low-affinity IgG receptors FcγRIIA,B (CD32), IgA receptor FcαR (CD89), and triggering receptor expressed on myeloid cells 1 (TREM1), that were up-regulated, and high-affinity IgG receptor FcγRIA (CD64), histocompatibility Ag class IG (HLA-G), leukocyte-associated Ig-like receptor 1 (LAIR1), leukocyte Ig-like receptor 9 (LIR9), and B1,B2,B3,B4 (LILRB1–4, CD85), that were down-regulated. Several scavenging and pattern recognition receptors were also selectively induced (complement component 1q receptor 1 (C1qR1); macrophage receptor with collagenous structure (MARCO); macrophage scavenger receptor 1 (MSR1); scavenger receptor-FI (SCARF1); scavenger receptor with C-type lectin 1 (COLEC12/SRCL)) or repressed (CD163 Ag; stabilin, STAB-1; TLR-5 and -7) under hypoxia (Table III⇑⇑⇑). Other differentially expressed genes coded for costimulatory and adhesion molecules involved in cell-cell and cell-matrix interaction, such as bone marrow stromal cell Ag (CD157), CD36, CD84, and CD86 Ags, integrin α X (CD11C), integrin β5 and β7 (ITGB5,7), ICAM3,5, semaphorin 4D (CD100), and sialoadhesin (CD169) (Table III⇑⇑⇑).
The hypoxic transcriptome was also characterized by the modulation of a cluster of genes coding for cytokines/chemokines and/or their receptors. Within the chemokine system, we identified for the first time CCL20, CXCL2, CXCL3, CXCL5, the fractalkine receptor CX3CR1, and the G protein-coupled chemokine orphan receptor (RDC1) as hypoxia-inducible genes, whereas CCL15, CCL18, CCL23, CCL8, CXCL6, CCR1, and CCR2 were the most highly hypoxia-repressed genes (Table III⇑⇑⇑). Finally, Mn hypoxic profile included various components of the IL-1 system (IL-1 family member 9 (IL1F9); IL-1 receptor accessory protein (IL1RAP); IL-1R associated kinase 3 (IRAK-3); IL-18R1) and members of the TNF receptor and ligand superfamilies, as well as IL-23A, IL-6 signal transducer (IL-6ST), IL-13RA1, IL-21R, CSF1,3, and CSF1,3Rs (Table III⇑⇑⇑).
Confirmation of microarray data by qRT-PCR analysis of selected hypoxia-modulated genes
To validate the microarray results using a different technique, mRNA levels for a subset of known and novel hypoxia-modulated genes were quantified by qRT-PCR on a new RNA pool (Fig. 3⇓), using the primer pairs listed in Table I⇑. For this analysis, we randomly selected 27 genes involved in immune regulation, inflammatory responses, and transcription. Three reference genes (Table I⇑⇓), although they were higher according to qRT-PCR for six genes (CCL23, FCGR1A, LIR9, BNIP-3, Mxl1, and VEGF), in agreement with previous findings showing that microarray can often underestimate the extent of gene regulation compared with qRT-PCR (Varesio et al., unpublished observations). For other genes, such as CMRF-35H, FCGR2B, STAB-1, and MIF, however, higher expression differences were detected by microarray. These results confirm hypoxia responsiveness of novel genes identified by microarray.
Hypoxia induces CCL20 expression and secretion by human Mn
We then selected CCL20, one of the novel immune-related genes most strongly up-regulated in hypoxic Mn, for further analysis. According to microarray and qRT-PCR data, CCL20 mRNA levels were increased by an average of 7.4- and 5.7-fold, respectively (Fig. 3⇑). To address the issue of donor-to-donor variability, we analyzed CCL20 mRNA expression by RT-PCR in a subset of samples comprising the pools used for microarrays (Fig. 4⇓A). CCL20 basal expression was detected in all Mn preparations cultured under normal pO2, although with some variations among individual donors. Consistent CCL20 transcript up-regulation was triggered by hypoxia in every donor, independently of the baseline levels (Fig. 4⇓A), indicating the general inducibility of the gene in primary Mn. qRT-PCR was also performed to quantify the magnitude of hypoxia-induced changes. As shown in Fig. 4⇓B, the extent of CCL20 mRNA increase ranged from 3.1- to 11.4-fold among the samples examined. mRNA up-regulation was paralleled by increased protein expression, as determined by immunocytochemistry (Fig. 5⇓A). Mn isolated from the three donors analyzed by qRT-PCR were cytocentrifuged and immunostained with a mAb directed to CCL20. Low levels of CCL20 immunoreactivity were detectable in the cytoplasm of normoxic Mn. Hypoxia exposure for 16 h resulted in a marked increase in intracellular protein content in all the samples. No staining was detected when an isotype-matched control Ab was used (Fig. 5⇓A).
Conditioned medium from normoxic and hypoxic Mn cultures was then analyzed for CCL20 content by ELISA to determine whether a parallel rise in chemokine secretion was triggered by hypoxia (Fig. 5⇑B). Normoxic Mn constitutively secreted variable amounts of CCL20, which accumulated in the culture medium ranging from 10 to 410 pg/106 cells/ml in six different donors. Consistent with the mRNA data, chemokine secretion increased by 2.4- to 3.5-fold upon incubation to 1% O2 for 16 h (Fig. 5⇑B) and was further augmented at longer time points, with almost 7-fold higher amounts detectable in the conditioned medium at 72 h of culture (Fig. 5⇑C, donor 12). CCL20 is a LPS-inducible chemokine (3) and the response of Mn to LPS was evaluated in parallel experiments as a positive control. Interestingly, the extent of CCL20 secretion elicited by hypoxia exceeded that triggered by stimulation with LPS at all the time points analyzed (Fig. 5⇑C), suggesting that hypoxia is a more potent stimulus than LPS in inducing CCL20 production by Mn.
Taken together, these results indicate that primary human Mn isolated from different donors produce variable baseline levels of CCL20 but respond to low pO2 with consistent CCL20 up-regulation, confirming CCL20 as a hypoxia-inducible gene.
Mn responses that ensue following recruitment at pathological sites begin in the setting of reduced pO2 (1). Hypoxia regulatory effects on Mn gene expression have not been characterized in detail, and only a limited number of hypoxia-inducible genes have been identified in these cells (17, 31, 41). In this study, we provide the first transcriptome analysis of hypoxic primary peripheral blood human Mn.
Microarray analysis was conducted on RNA pooled from different donor-derived Mn, whose response to hypoxia was demonstrated by VEGF up-regulation. Three pools, each composed by RNA from five different donors, were investigated per each experimental condition. RNA pooling is a powerful, cost-effective, and statistically valid mean of identifying common changes in a gene expression profile. Previous studies have established the feasibility of using this type of approach to overcome interindividual variability and provide reliable results that reflect gene expression in individual donors (42). The suitability of our experimental protocol for identifying bona fide hypoxia-responsive genes in primary cells was inferred by the demonstration that hypoxia modulated 74 genes known from the literature to be responsive to hypoxia. The majority of these genes, involved in glycolysis, apoptosis, cell cycle, and transcription, may constitute a gene cluster commonly induced by hypoxia in cells of different lineages. The general representativeness of the gene expression changes detected by microarray was further established by qRT-PCR, which confirmed the differential expression of 27 genes randomly selected from the microarray profile in a fourth set of pooled RNAs.
Two recent reports investigated the hypoxia transcriptome of hMDM and demonstrated induction of genes encoding angiogenesis and ECM regulators/components (20, 21), some of which were also present in our profile. Expression of ADM, F3, FLT1, FN1, MIF, INHBA, MMP-1, PTGS2, SPP1, and VEGF, among others, was detectable at significant levels in both primary human Mn (this report) and MDM (20), and exposure to hypoxia resulted in their strong up-regulation in both cell populations. Moreover, we demonstrated up-regulation of other genes previously shown to be inducible by hypoxia in mononuclear phagocytes and coding for immunoregulatory/proinflammatory proteins and transcription factors, such as ARG1, CXCR4, EGR1, FGF1, ID2, TIMP-1, and VLDR (17, 21, 32, 41, 43, 44). Interestingly, IL-1α, IL-6, and TNF-α, induction in Mn was observed in response to hypoxia alone, whereas up-regulation in Mf and MDM required LPS costimulation or a reoxygenation period (43). Furthermore, we confirmed in primary Mn hypoxic inhibition of a cluster of inflammatory genes found down-regulated in mMf and hMDM, such as CCL2 (16), CCR5 (18), CTSC (21), OASE (40), and RAB7 (21). These findings indicate that hypoxia is active on different types of mononuclear phagocytes and across species in regulating the expression of selected target genes, which may be critical for their adaptation to the hypoxic environment and ability to function in it and could be representative of the hypoxic transcriptome of cells belonging to the monocytic lineage. In contrast, the observation that other genes up-regulated in hMDM (20, 21) were not induced or were even down-regulated in primary Mn, e.g., the angiogenesis inducers endothelial cell growth factor-1 (ECGF1) and IL-18 binding protein (IL-18BP), suggests that hypoxia may also activate in mononuclear phagocytes a specific transcriptional response depending on their differentiation stage.
Several HMGs identified in our analysis were endowed with transcription regulatory activity and/or encoded important components of the HIF-1 transcription pathway. Hypoxia increased the expression of P4HA1, P4HA2, and EGLN1, three members of the prolyl-hydroxylase family which mediate HIF-α hydroxylation in well-oxygenated cells targeting the protein for proteosomal degradation by the von Hippel Lindau tumor suppressor protein (pVHL) (15), and of BHLHB2, a bHLH family member implicated in the regulation of pVHL/HIF pathways (15). These data are in agreement with previous findings in other cell types (15, 35, 36) and suggest an O2-driven HIF-1-dependent autoregulatory mechanism, required to ensure fast HIF-1α degradation upon reoxygenation, also in Mn. Of note is the novel evidence that hypoxia promotes the expression of ATF2 and ATF5, two members of the ATF family of transcription factors (45) required for efficient activation of gene transcription by hypoxia (15). Moreover, we demonstrated up-regulation of FosB and FRA-2, two components of the AP-1 complex (45), which is activated by hypoxia and was recently shown to cooperate with HIF-1α in the transactivation of hypoxia inducible genes (46), and of JEM-1, a novel transcriptional cofactor which enhances AP-1 activity (47). Up-regulation of these genes together with down-regulation of ZnF197, a novel VHL-interacting protein that functions as a repressor of HIF-1α transactivation (48), may represent a positive regulatory mechanism of hypoxia transcriptional response in Mn. Down-regulation of several HIF-1 transcriptional cofactors, including CBP, C/EBPα, MITF, and NCOA1/SRC-1, was also observed in hypoxic Mn. All these proteins physically interact with HIF-1α enhancing its transactivation function under hypoxia (9, 15, 49), and their down-regulation suggests the existence of a negative feedback mechanism to control Mn hypoxic response. Consistent with this hypothesis, we observed a parallel decrease of the mRNAs encoding HIF-1α and one of its dimerization partners, ARNT-2 (9)
GO data mining characterized a significant cluster of genes as being associated with immune regulation and inflammatory responses, chemotaxis, cell adhesion, and ECM remodeling, the majority of which has not been previously identified as responsive to hypoxia. Profound changes were observed in the expression of scavenger receptors. Of relevance is the up-regulation of MSR1 and SCARF1, because these molecules functions as endocytic receptors for acetylated LDL and may thus be implicated in lipid-loaded foam cell formation contributing to atherosclerotic plaques development (50). Consistent with the view that hypoxia may exert a pathogenetic role in atherosclerosis and Alzheimer’s disease (9, 10, 41) is also the down-regulation of CD163 and STAB1 scavenger receptors, which are endowed with atheroprotective activity (51, 52), and of various genes involved in the regulation of fatty acid and/or cholesterol biosynthesis/transport and acting as anti-atherogenic factors, e.g., LDLR, ApoE, and CYP27A1 (53, 54). Mn hypoxic profile also showed up-regulation of a number of other pattern recognition receptors critical to host defense, the most relevant of which are the C1qR1, COLEC12, and MARCO. These molecules bind specific Ags on bacteria facilitating their recognition and phagocytosis (55, 56, 57). Interestingly, the TLR family members TLR5 and TLR7, whose function is also to recognize pathogen or their products and hence initiate innate immune responses (58), were down-regulated by hypoxia. Differential modulation of receptors with similar functions is intriguing and is likely to contribute to the fine tuning of Mn antimicrobial activities at sites of infection.
Myeloid cell immune functions are regulated by a balance of inhibitory and activating signals transduced by multiple families of cell surface IRS receptors belonging to the Ig superfamily of inhibitory/stimulatory pairs of molecules (59). These receptors are either characterized by ITIM/ITAM regulatory components in their cytoplasmic domain or pair with ITAM-containing transmembrane adapter proteins (59). Mn exposure to hypoxia resulted in the selective modulation of various members of these receptor families. Of particular interest is the up-regulation of FcγRIIA,B and down-regulation of FcγRIA. FcγRIIA contains a functional ITAM motif, thus triggering cell activation, whereas FcγRIIB encodes an ITIM component, leading to repression of cellular responses (60). FcγRIA, which associates for signaling with an ITAM-containing γ-chain, also belongs to the activatory FcγR subclass (60). FcγRs involvement in various autoimmune inflammatory states was reported (60), and our findings suggest the potential role of hypoxia in influencing the pathogenesis of these diseases through the modulation of distinct FcγR genes.
Hypoxic modulation of other ITIM/ITAM Ig family members, specifically up-regulation of FCAR, CMRF-35H, and TREM-1 and down-regulation of LIR9 and LAIR1, was also observed. These molecules mediate a variety of effector functions vital to the adaptative immune response, and their ligation can have both proinflammatory and immunosuppressive consequences by differentially modulating the secretion of pro- or anti-inflammatory mediators (59, 60, 61, 62, 63). The differential expression of inhibitory and activating isoforms of a given receptor family with similar specificities on the same cell is another example of the tight regulatory role of hypoxia on Mn inflammatory responses. In particular, the concomitant down-regulation of the Ig-like inhibitory receptors LILRB1–4 and of their ligand HLA-G, a nonclassical inhibitory HLA class I Ag, is noteworthy given the role of these molecules in immune tolerance and immune escape (64). It is conceivable that inhibition of these molecules may decrease the activation threshold of Mn retained at hypoxic sites. Mn activation under conditions of reduced O2 availability may also be controlled by the selective down- and up-regulation of two C-type lectin receptors, the ITIM-containing inhibitory molecule CLECSF6 (65) and CD69, a member of the NK receptor family and a potent inducer of Mn inflammatory mediator production and cytotoxic activity (66). Although a ligand for CD69 has not been identified, previous studies suggested a pathogenetic role for CD69 in certain inflammatory states characterized by hypoxia, such as rheumatoid arthritis, chronic inflammatory liver diseases, and asthma (66).
Mononuclear phagocyte migratory activity is a highly regulated process which depends on a defined repertoire of chemokines/receptors and adhesion molecules, and dysregulated expression of these proteins may alter their recruitment and activation (67). Various studies have investigated the mechanisms whereby mononuclear phagocytes are retained/concentrated at hypoxic pathological sites (1). One possibility is that hypoxia inhibits their migration in response to chemokines by decreasing the expression of specific chemokine receptors, as demonstrated for CCR5 in Mf (18). In agreement with this hypothesis and with previous findings showing impaired Mn migratory ability to CCL2 under conditions of low pO2 (1), we observed CCR5, CCR1, and CCR2 down-regulation. This study also suggests other potential mechanisms for Mn entrapment in hypoxic areas, such as the up-regulation of RGS1, a member of a new class of G protein-signaling deactivators which inactivates several chemotactic receptors inhibiting chemoattractant-induced Mn migration (68). Furthermore, repression of the secreted cell motility-promoting factor, ENPP2 (69), together with induction of the antichemotactic cytokine, MIF (13), and the GRO family chemokines, CXCL2 and CXCL3, which are specialized Mn-arrest chemokines (70), may also provide a “stop” signal to Mn within hypoxic tissues.
The complexity of the regulation of Mn migratory behavior by hypoxia is further emphasized by the demonstration of dysregulated expression of several other migration-related genes. Mn hypoxic profile was associated with down-regulation of the adhesion molecules CD11C, CD57, and ITGB5,7, which mediate Mn adhesion to the endothelium and/or to the ECM (67), and with up-regulation of both the fractalkine receptor, which binds to the CX3C chemokine fractalkine expressed on endothelial cells functioning as a potent adhesion molecule (71), and the RDC1 receptor, which share with CXCR4 the chemokine CXCL12/SDF-1 as a natural ligand (72). A critical role in the control of mononuclear phagocyte motility is also exerted by MMPs, a group of secreted enzymes that trigger ECM degradation facilitating leukocyte movement in tissues (73). Recent reports have shown up-regulation by hypoxia of MMP-1, MMP-7, and MMP12 in hMDM (20, 21). Interestingly, only MMP1 was induced also in primary Mn that showed specific up-regulation of MMP16 and MMP19 and down-regulation of MMP25. Moreover, the MMP inhibitors, TIMP1 and TFPI2, which reduce ECM degradation inhibiting cell migratory activity (73, 74), were also modulated in hypoxic Mn. Collectively, these studies indicate that regulation of MMPs and their inhibitors is a common denominator of mononuclear phagocyte response to hypoxia, but that different components of these families are controlled depending on the cell differentiation stage. Altered expression of MMPs has been associated with a variety of acute and chronic inflammatory states (73), and hypoxia can probably play a role in the pathogenesis of these diseases by modulating MMP production by infiltrating Mn
Various cytokines/chemokine and/or receptors are modulated by hypoxia in mononuclear phagocytes (1, 43). The Mn hypoxic transcriptome confirmed and extended those findings showing differential expression of other components of the cytokine/chemokine system. Various members of the IL-1 and the TNFR/ligand superfamilies were selectively up- or down-regulated in hypoxic Mn, including molecules associated with and mediating signal transduction from their receptors (e.g., IL1RAP, IRAK3, and the TNFR-associated factor, TRAF). Because of their pleiotropic effects on almost every types of cells, coordinated regulation of the TNF and IL-1 systems is likely to represent an important mechanism to control the amplitude and the duration of inflammatory responses. The demonstration that hypoxia induces CSF1 and CSF3, while inhibiting their receptors, is noteworthy given the role of these factors in the regulation of myeloid cell production, differentiation, and function (75) and is consistent with previous findings suggesting a reciprocal and divergent action of hypoxia on receptor vs ligand expression (16, 18). Accordingly, IL-4 up-regulation and concomitant inhibition of its receptor IL-13RA1 was observed. This interplay is likely to serve as a negative feedback mechanism to control the autocrine activation of producing Mn. However, not all the data presented in this study are consistent with this scenario. Concomitant down-regulation of the Mn chemoattractants CCL2 and CCL8 and of their receptors CCR2 and CCR5 (2) was in fact observed in response to hypoxia. Similarly, we found down-regulation of CCL15, a chemoattractant for neutrophils, monocytes, and lymphocytes (76), and CCL23, a chemokine mediating resting T cell and Mn chemotaxis (77), and of their common receptor CCR1. Collectively, these data indicate that a dynamic change in chemokine/receptor expression profile occurs in Mn within hypoxic tissues. This tight and complex level of control exerted by low O2 tension is clearly of pathophysiologic relevance, representing an important mechanism of regulation of leukocyte trafficking and function at sites of inflammation. Because some of the modulated chemokines have angiogenic activity (1, 2, 76), their altered expression under low pO2 may also influence neoangiogenesis in pathological tissues.
One of the most important findings of this study was the demonstration that hypoxia strongly induced the expression of the CCL20-encoding gene in primary Mn. Marked differences in the basal mRNA expression were detected among individual donors. However, mRNA up-regulation under hypoxia was consistently demonstrated in all Mn preparation analyzed, and gene expression results were associated to a parallel augmentation of protein expression and secretion. Previous reports have shown that the CCL20 gene is markedly up-regulated in PBMC by LPS and inducible in other cell types in response to various mediators of inflammation, including cytokines, growth factors, bacterial, viral and plant products, whereas it is poorly expressed in the absence of inflammatory stimuli (3). This study is the first to identify hypoxia as a new CCL20 inducer. Interestingly, the extent of CCL20 up-regulation by hypoxia exceeded that triggered by LPS, suggesting that hypoxia is a more potent stimulus than LPS for CCL20 production by Mn. Given its role in the recruitment of iDC, effector/memory T lymphocytes, and naive B cells, CCL20 has been proposed as an important mediator for both the initiation and effector phases of the inflammatory reactions, linking innate and acquired immunity (3). Hence, by producing CCL20, Mn may control the kinetics and composition of the cellular infiltrate under various inflammatory conditions and at tumor sites. The identification of CCL20 as a hypoxia-inducible gene may explain, in part, the high levels of this chemokine present in areas of inflammation and in various chronic inflammatory conditions, such as rheumatoid arthritis, inflammatory skin disorders, and tumors (for a review, see Refs. 3 and 78), as these sites are known to be hypoxic and infiltrated by Mn, and is indicative of a pathogenetic role for this molecule in these diseases. Further studies are ongoing in the laboratory to elucidate the molecular mechanisms underlying CCL20 induction by hypoxia.
In summary, we have described the hypoxia transcriptome of primary human Mn and identified a large number of genes not previously known to change as a result of reduced O2 concentrations. Our findings contribute to the definition of the gene cluster commonly induced by hypoxia in cells of different lineage. This study provides novel insights into the molecular responses to the hypoxic stress and the mechanisms linking low pO2 to the regulation of immune and inflammatory responses, leading to new perspectives of the role of hypoxia in programming Mn functions within pathological conditions and identifying potential molecular targets for the development of rational therapeutic approaches.
We thank S. Delfino for technical assistance and C. Dabizzi for secretarial work.
The authors have no financial conflict of interest.
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.
↵1 This work was supported by grants from the Italian Association for Cancer Research, Fondazione Italiana per la Lotta al Neuroblastoma, San Paolo Company, Italian Health Ministry, and Ministero Istruzione Universita’ e Ricerca.
↵2 Address correspondence and reprint requests to Dr. Maria Carla Bosco, Laboratorio di Biologia Molecolare, Istituto Giannina Gaslini, Padiglione 2, L.go Gerolamo Gaslini 5, 16147 Genova Quarto, Italy. E-mail address:
↵3 Abbreviations used in this paper: Mn, monocyte; Mf, macrophage; iDC, immature dendritic cell; pO2, partial oxygen pressure; HIF-1, hypoxia-inducible factor-1; GO, Gene Ontology; EASE, Expression Analysis Systematic Explorer; qRT-PCR, real-time quantitative PCR; VEGF, vascular endothelial growth factor; HMG, hypoxia-modulated gene; IRS, immunoregulatory signaling; ARNT, aryl hydrocarbon receptor nuclear translocator; ECM, extracellular matrix; hMDM, human monocyte-derived macrophage; MMP, matrix metalloproteinase; bHLH, basic helix-loop-helix.
- Received November 4, 2005.
- Accepted May 18, 2006.
- Copyright © 2006 by The American Association of Immunologists