Identification of Genes Induced by a Macrophage Activator, S-28463, Using Gene Expression Array Analysis

Abstract

S-28463 and imiquimod are imidazoquinoline compounds which stimulate microbicidal activity by inducing a local immune response at the site of application. Imiquimod-containing cream is an effective clinical treatment against cervical warts caused by human papillomavirus infection. Imiquimod also induces leishmanicidal activity both in vitro in macrophages and in vivo in a mouse model for cutaneous leishmaniasis. The major target cells of S-28463 and imiquimod are macrophages. To explore the molecular basis in which imidazoquinolines generate macrophage microbicidal activity, a cDNA gene array analysis was undertaken to identify genes induced by S-28463. Out of 588 genes screened in this assay, only 13 genes were significantly induced by S-28463. Remarkably, virtually all of the induced genes are involved in macrophage activation and inflammatory response. This experimental approach defines the mechanism of action of this clinically relevant compound in the induction of microbicidal activity in macrophages and also potentially identifies novel genes associated with microbicidal activity in this cell type.

Imiquimod-containing cream is a topically administered treatment which is widely used for the treatment of genital warts caused by human papillomavirus (HPV) infection (7). In the majority of cases, treatment results in a local inflammatory response, leading to wart regression without recurrence. This new treatment represents a major advance in the therapy of this viral infection. More recently, we have shown that imiquimod cream is also an effective treatment against experimental cutaneous leishmaniasis caused by Leishmania major infection in mice (2). Based on these observations, a clinical trial is presently under way to determine whether topical application of imiquimod-containing cream may also represent an effective treatment for human cutaneous leishmaniasis.

Imiquimod and its more active structurally related derivative, S-28463, are imidazoquinolines, the potent immune response modifiers, which act through stimulating a local immune response at the site of application. Stimulation of the local immune response has been shown to be important in the clearance of HPV-associated genital warts (4). The imidazoquinolines are active via their immunomodulatory activity on various cell types which are known to be involved in immune responses, including epidermal Langerhans cells and peripheral blood mononuclear cells, causing such cells to release a number of cytokines, such as alpha interferon (IFN-α) and tumor necrosis factor alpha (TNF-α). The major cell responsible for the cytokines produced by peripheral blood mononuclear cells is macrophages (9). This is consistent with our previous observation that imiquimod and S-28463 have no direct anti-Leishmania activity but rather exert their effect by activating macrophages to kill Leishmania donovani amastigotes in vitro in the absence of any other cell types (2). We also observed that macrophage leishmanicidal activity mediated by these compounds occurred through the stimulation of the expression of the inducible nitric oxide synthase (iNOS) gene and the release of nitric oxide (NO), which is the principal mediator of leishmanicidal activity in macrophages. It was found that the induction of iNOS gene expression is likely due in part through the induction of the AP-1- and NF-κB-associated signal transduction pathways (2).

Because these compounds are clinically relevant in the treatment of infectious disease and are capable of stimulating microbicidal activity in macrophages, we have undertaken to define their mechanism of action. A novel and effective approach to study drug effects on cells is to undertake a functional genomic approach to define cellular gene expression in response to drug treatment. This approach is well suited to study the mode of action of imidazoquinolines because appropriate, untreated (control) macrophages are available in such analysis. Using this approach, we have analyzed the gene expression profiles of quiescent and S-28463-stimulated bone marrow-derived macrophages (BMM). This experimental system was undertaken because we have previously established that S-28463 and its derivative, imiquimod, were effective at stimulating BMM to kill intracellular Leishmania amastigotes (2). Remarkably, S-28463 upregulated a relatively small number of cellular genes, the majority of which are directly associated with macrophage activation and inflammatory response. These two processes are important in the elimination of the invading microorganisms. S-28463 also induced the expression of an antiapoptotic gene, and future studies are required to determine whether this is associated with the normal macrophage activation process. This provides important understanding of the mechanism by which the imidazoquinoline drugs mediate microbicidal activity in macrophages and could further define novel genes involved in macrophage activation and the inflammatory response. These data also reveal that cDNA expression array analysis is effective in defining effector genes in response to drug stimulation of macrophages.

MATERIALS AND METHODS

Preparation of BMM.

BMM were obtained from femurs of 6- to 8-week-old female BALB/c mice (Charles River Canada, St. Constant, Québec, Canada) as previously described (5, 16) by flushing femurs with RPMI 1640 complete medium. Bone marrow cells were incubated in tissue culture dishes (Nunc, Roskilde, Denmark) for 1 day at 37°C in 5% CO₂ in moist air in RPMI 1640 complete medium containing 15% (vol/vol) L929 cell-conditioned medium as a source of monocyte-macrophage colony-stimulating factor or colony-stimulating factor 1 (CSF-1). After 1 day in culture, the immature nonadherent cells were transferred into new polystyrene culture dishes (Falcon 1029), which were weakly adherent for macrophages and were cultured in 15% CSF-1 to induce macrophage differentiation for 7 days. The resulting BMM population was made quiescent by culturing in CSF-1-free medium for 18 h. Cell viability after scraping was determined by trypan blue exclusion assay, and live cells were counted with a hemocytometer. The quiescent BMM (10⁶ cells/ml) in polystyrene tubes were stimulated with 100 ng of S-28463/ml for 3 h. Total RNA were isolated from normal or S-28463-stimulated cells by Trizol reagent (Gibco BRL Lift Technologies, Burlington, Ontario, Canada) and were used as recommended by the manufacturer.

Gene array analysis.

Total RNA samples were reverse transcribed using the gene-specific cDNA synthesis (CDS) primer mix which will amplify all of the genes on the gene array membrane in the presence of reverse transcriptase (ClonTech Laboratories, Inc., Palo Alto, Calif.) and [α-³²P]dATP. The generated radiolabeled cDNA probes from stimulated and unstimulated cells had a specific radioactivity of ≈2.6 × 10⁶ cpm, were purified from unincorporated nucleotides, and were hybridized to identical membranes containing mouse cDNA arrays (Atlas Mouse cDNA Expression Array [PT3140–1]; ClonTech Laboratories, Inc.). Each cDNA array contained 588 previously characterized mouse genes. Each cDNA fragment is 200 to 600 bp long and is selected as a unique sequence without a poly(A) tail, repetitive elements, or highly homologous sequences to minimize cross-hybridization and nonspecific binding of cDNA probe. The amount of each cDNA fragment is 10 ng, and each cDNA fragment is immobilized in two adjacent dots in order to differentiate specific hybridization signal from nonspecific background signal. Following hybridization, high-stringency washes were performed and the membranes were subjected to autoradiography. The hybridization pattern was analyzed by using the AtlasImage 1.0 software package specifically designed for analyzing Atlas Array data.

Northern blot analysis.

Normal BMM (10⁷ cells per sample) were treated for 3 h with 100 ng of S-28463/ml, and control BMM incubated in parallel were left untreated. Total cellular RNA was prepared using Trizol reagent, and 10 μg was denatured by glyoxal at 50°C for 1 h and chilled on ice for 5 min. One microgram of ethidium bromide was added to each sample before electrophoresis in 1% agarose gel to fractionate RNA as described before (21). Following electrophoresis, RNA was blotted onto Hybond-N nylon membrane (Amersham Int., Amersham, United Kingdom) as recommended by the manufacturer. The membrane was UV cross-linked and prehybridized with a solution containing 20× SSPE (1× SSPE is 0.18 M NaCl, 10 mM NaH₂PO₄, and 1 mM EDTA [pH 7.7]), 50× Denhardt solution, 50% formamide, 10% sodium dodecyl sulfate, and 10 mg of denatured salmon sperm DNA/ml at 42°C for 3 h. Hybridization was performed at 42°C for 18 h with probes purified from agarose and nick translated in the presence of 125 μCi of [³²P]dCTP (ICN Biochemicals, Québec, Canada). The membrane was washed once at room temperature and twice at 55°C with 0.5× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate) for 30 min each and was autoradiographed at −70°C in cassettes on Kodak Bio Max MS films with Bio Max MS intensifying screens. The mouse iNOS probe was the 4.1-kb NotI fragment from pmmac-NOS, kindly supplied by Charles J. Lowenstein (Johns Hopkins University, School of Medicine, Baltimore, Md.). The interleukin-1β (IL-1β) probe was the 400-bp BamHI fragment from pBluescript, kindly supplied by A. Descoteaux. To ensure that equal amounts of RNA were analyzed, blots were stripped, rehybridized with a radiolabeled cDNA probe for actin (1.25-kb PstI of pBA-1), washed, and again subjected to autoradiography. When quantified by scanning densitometry, multiple exposures were used to ensure that all signals were within the linear response range of the film.

RESULTS AND DISCUSSION

The objective of this study was to identify downstream gene targets of the imidazoquinolines in macrophages to define the mechanism in which they induce microbicidal activity. Primary BMM were prepared from BALB/c mice and treated with S-28463 for 3 h as described in Materials and Methods. The 3-h treatment was selected to limit the analysis predominantly to the primary response, as opposed to potential secondary responses caused by autocrine response to cytokines induced by S-28463. Untreated, quiescent BMM served as the control. Total RNA was isolated from the cultures and used to prepare cDNA probes, which were hybridized to parallel cDNA array filters. Each filter contained 588 cDNA fragments representing previously characterized mouse genes divided into six quadrants representing different functional categories, including (i) oncogenes and tumor suppressor genes and cell cycle regulators; (ii) stress response genes, ion channels, and transport genes and intracellular signal transduction modulators and effectors; (iii) apoptosis-related genes and genes involved in DNA synthesis, repair, and recombination; (iv) transcriptional factors and general DNA-binding proteins; (v) receptors, cell surface antigens, and cell adhesion molecules; and (vi) cell-cell communication factors. The hybridization pattern reflects the gene expression profiles in the control cells (Fig. 1, upper panels) and S-28463-stimulated cells (Fig. 1, lower panels). As shown in Fig. 1, several differences in gene expression were apparent and highlighted in the S-28463-stimulated cells. The uniform expression of the housekeeping genes on the membranes served as the internal controls.

FIG. 1.

Comparison of gene expression profiles in control and S-28463-stimulated BMM. Gene expression profiles of control and S-28463-stimulated cells are presented in upper and lower panels, respectively. Genes which were induced in the S-28463-stimulated cells by fivefold or greater are circled. Row g in panels D, E, and F includes the indicated housekeeping genes, which are used as internal references to compare control and S-28463-stimulated cells. As demonstrated, the expression of the housekeeping genes was the same in the two cell populations. Total RNA was isolated from either control or S-28463-stimulated mouse BMM and used as templates to synthesize α-³²P-radiolabeled cDNA probes of equal specific activity. The cDNA probes were hybridized to two identical gene array membranes, and the hybridization results are displayed by autoradiography. Each quadrant presents different functional groups of genes: oncogenes and tumor suppressor genes and cell cycle regulators (A); stress response genes, ion channels and transport genes and intracellular signal transduction modulators and effectors (B); apoptosis-related genes and genes involved in DNA synthesis, repair, and recombination (C); transcriptional factors and general DNA-binding proteins (D); receptors, cell surface antigens, and cell adhesion molecules (E); cell-cell communication factors (F).

We defined a threshold value for genes which were “significantly” induced by S-28463 as at least fivefold-higher expression in the treated than in the control macrophages. This serves to emphasize major changes in gene expression, as opposed to minor changes, and also helps to simplify the analysis of the large quantity of data generated in this type of study. As shown in Table 1, only 13 genes were significantly induced by S-28463 and the values ranged from 5.6- to 26.7-fold higher in treated than in untreated control macrophages. There were no genes that were expressed at higher levels in the control cells than in the treated cells, even when including values which were lower than the established threshold.

TABLE 1.

Genes upregulated in S-28473-stimulated BMM^a

GenBank accession no.	Gene code	Fold induction	Protein or gene
X15842	A2m	6.9	c-rel
U36277	B3m	8	IκB-α
U19799	B3n	5.6	IκB-β
U97076	C3h	6.4	FLIP-L
M87039	C3m	7	iNOS
Z31663	E1a	6.1	ActIBR
M83312	E1f	6.5	CD40
X52264	E7i	5.8	ICAM-1
X12531	F3e	9.7	MIP-1α
M35590	F3f	25.4	MIP-1β
X53798	F3g	26.7	MIP-2α
M15131	F4k	12.9	IL-1β
X16490	F7i	13	PAI-2

^aThe expression level of the indicated genes was determined by using AtlasImage 1.0, a software package specifically designed for analyzing Atlas Array data. 

In order to realize the significance of the data, it was important to verify that the values obtained in the expression array analysis accurately reflected the mRNA levels in the cells. Therefore, Northern blot analysis was carried out on two of the genes, which were significantly induced following treatment. These included the iNOS and IL-1β genes. As shown in Fig. 2, the Northern blot analysis of the iNOS and IL-1β mRNA levels in the control and treated cells is consistent with the cDNA array analysis in confirming that S-28463 significantly induced the expression of the iNOS and IL-1β genes in the BMM. In comparison, the parallel Northern blot performed with the actin probe showed that the expression of this housekeeping gene was equal in the two cell populations. Notably, the level of induction of iNOS and IL-1β genes detected by these two methods was very similar, as both methods showed that the iNOS gene was induced by 7-fold and that the IL-1β gene was induced by 13-fold. These data show that the cDNA array approach does accurately reflect changes in gene expression following S-28463 treatment of BMM.

FIG. 2.

Northern blot analysis of the iNOS and IL-1β mRNAs. Total RNA was isolated from control or S-28463-stimulated BMM. The RNA samples were fractionated on agarose gel and then subjected to Northern blot analysis using α-³²P-radiolabeled cDNA probes corresponding to the indicated genes. Similar results were obtained from two independent experiments.

The major advantage of the gene array analysis is that hundreds of genes can be examined, in comparison to Northern blot analysis, which examines one or only several at a time. Therefore, the gene array approach provides a great deal more information than conventional Northern blot analysis. Northern blot analysis is, however, more rapid and requires shorter exposure times and less RNA than does the gene array analysis. Nevertheless, as demonstrated within this work, the data are comparable with respect to quantitation.

Based on these data, it was of particular interest to consider the genes in Table 1 that were induced following treatment with S-28463. Remarkably, the majority of these genes play significant roles in macrophage activation and inflammatory response. For example, S-28463 induced the expression of the iNOS gene. The product of this gene is an enzyme responsible for the synthesis of NO, a highly reactive free radical. It has been established that NO produced by activated macrophages is a major effector molecule in the host defense mechanism against intracellular pathogens, including Leishmania (reviewed in reference 10). This result is consistent with our previous study showing that imidazoquinolines induce the expression of the iNOS gene and synthesis of NO, which resulted in leishmanicidal activity by these compounds (2).

S-28463 also induced the expression of the IL-1β gene, whose product plays a major role in the inflammatory response. Cells known to express IL-1β include monocytes/macrophages. IL-1β also induces the expression of other cytokine genes involved in the inflammatory response, such as IL-6, IL-8, and TNF-α (reviewed in reference 6). Another example of inflammatory genes induced by S-28463 was genes encoding the macrophage inflammatory proteins MIP-2α, MIP-1α, and MIP-1β. Interestingly, the macrophage inflammatory protein genes were induced to a greater extent by S-28463 than were any other genes in this analysis. MIP-2α is produced by activated macrophages and is a potent chemoattractant for neutrophils (24). Sources of MIP-1α and MIP-1β are identified as monocytes/macrophages. MIP-1α is a primary chemotactic for monocytes, B lymphocytes, and activated CD8⁺ T cells, whereas MIP-1β is chemotactic for monocytes and activated CD4⁺ T cells (13, 17, 20). MIP-1α also induces intercellular adhesion molecule (ICAM)-1 expression (22); mast cell degranulation (1); and production of TNF-α, IL-1, and IL-6 (8). Taken together, S-28463-mediated induction of IL-1β, MIP-2α, MIP-1α, and MIP-1β plays a major role in the recruitment of immune cell populations to the site of the inflammation. This is consistent with the observation that imiquimod-containing cream induces a local inflammatory response at the site of application, resulting in the antiviral activity displayed against HPV infection (7).

S-28463 also induced the expression of the CD40 gene, which is a member of the TNF-receptor family proteins. The professional antigen-presenting cells, such as monocytes, dendritic cells, and follicular dendritic cells, all bear CD40 (reviewed in reference 19). CD40-stimulated macrophages display enhanced antigen-presenting capacity through the upregulation of major histocompatibility complex class II expression and also secrete IL-1, IL-6, IL-8, IL-10, IL-12, TNF-α, and MIP-1α in response to CD40-mediated signaling. CD40-stimulated macrophages display increased tumoricidal activity and increased production of NO. Furthermore, lack of CD40 signaling during Leishmania infection resulted in increased susceptibility to the infection due to a lack of NO production (18). Thus, CD40-mediated signaling enhances the antigen-presenting capacity and effector functions of macrophages associated with Leishmania killing.

S-28463 also induced the expression of the ICAM-1 gene, the plasminogen activator inhibitor 2 (PAI-2) gene, and the c-rel gene. ICAM-1 functions as a costimulatory molecule on antigen-presenting cells, such as macrophages, to activate major histocompatibility complex class II-restricted T cells (reviewed in reference 23). PAI-2 is synthesized by macrophages (3) and is believed to play a role in initiating the healing of inflammatory lesions. The c-rel gene encodes a member of a family of transcription factors which regulate the expression of a variety of genes involved in inflammatory response, including IL-1, TNF-α (15), and IL-8 (12). S-28463 also induced genes encoding IκB-α and IκB-β. These proteins are implicated in the regulation of NF-κB/Rel proteins. Recently, it has been argued that IκB-β acts as a chaperone for NF-κB and mediates its persistent activation, whereas IκB-α can bind to and inactivate free NF-κB (reviewed in reference 14). Thus, regulation of IκB-α and IκB-β proteins is critical for modulating NF-κB-directed gene expression, which is a key regulator of the cellular inflammatory and immune response.

It is particularly interesting that S-28463 induced the expression of the long form of FLICE-inhibitory protein (FLIP-L) gene, whose product is a potent inhibitor of apoptosis (11). To date, little is known about the role of FLIP-L in macrophages; however, it will be interesting to carry out further studies to explore this association. For example, the expression of the FLIP-L gene may be required to inhibit apoptosis during the activation process, which results in the release of various toxic compounds.

Based on the data obtained in this study and our previous study (2), we propose the following model in which S-28463 modulates macrophage functions to induce the antimicrobial response (Fig. 3). As shown previously, imiquimod and S-28463 induce AP-1- and NF-κB-mediated gene expression. This results in expression of the iNOS gene and the production of NO, which mediates Leishmania killing (2). In this study, we further define genes which are induced following treatment with S-28463, which results in the secretion of molecules associated with the inflammatory response (macrophage inflammatory proteins, IL-1β, PAI-2); the expression of surface receptors (CD40, ICAM-1); and the expression of transcription-regulatory molecules (IκB-α, IκB-β, c-rel) associated with initiating an immune response. The identification of these gene targets defines the molecular basis in which these compounds work in vivo to mediate a local immune response at the site of application, resulting in antimicrobial effects (2, 4, 7, 9).

FIG. 3.

Mechanism in which S-28463 mediates leishmanicidal activity in macrophages and modulates macrophage activity associated with inducing a local immune response at the site of application. The internal circle represents a phagolysosome containing Leishmania amastigotes which become targeted by NO following the expression of the iNOS gene mediated through activation of AP-1 and NF-κB (2). Another 13 genes induced in macrophages and identified in this study have a pleiotropic effect on the local immune response. These include genes involved in the inflammatory response (macrophage inflammatory proteins, IL-1β, PAI-2), the expression of surface receptors (CD40, ICAM-1), and the expression of transcription regulatory molecules (IκB-α, IκB-β, c-rel) which subsequently mediate cytokine gene expression. At the origin of the arrows are induced mRNAs. The product of each gene and its location are at the termini of the arrows, located inside the cell (iNOS, c-rel, IκB-α, IκB-β, and FLIP-L), secreted outside the cell (IL-1β, MIP-2α, MIP-1α, MIP-1β, and PAI-2), or located on the cell surface (CD40 and ICAM-1).

In summary, imiquimod-containing cream has been clinically shown to be safe and effective against HPV-associated genital warts and is presently in clinical trials for the treatment of cutaneous leishmaniasis. The present study helps to define the molecular basis in which macrophages respond to this important class of immunomodulating drug. The data obtained are consistent with the effect of this compound in vivo with respect to mediating antiviral and anti-Leishmania effects and define with considerable fidelity the macrophage target genes of this novel class of immunomodulating drugs. This study also underlines the effectiveness of this functional genomic approach in defining the mechanisms of the drug activity on macrophages and may also help to define novel genes associated with macrophage activation and killing of infectious agents.