High molecular weight gingipains from Porphyromonas gingivalis induce cytokine responses from human macrophage-like cells via a non-proteolytic mechanism

Abstract

Periodontal disease is an oral inflammatory disease affecting the supporting structures of teeth. Porphyromonas gingivalis, a major pathogenic agent for the disease, expresses a number of virulence factors, including cysteine proteases called the gingipains. The arginine- and lysine-specific gingipains, HRgpA and Kgp, respectively, are expressed as high molecular weight forms containing both catalytic and adhesin subunits. We examined the expression pattern of cytokines and their receptors in differentiated macrophages following exposure to active and inactive forms of the gingipains, using a cDNA array, quantitative PCR and ELISA analysis. Amongst other pro-inflammatory cytokines, results from the cDNA array suggested that interleukin-1β (IL-1β), granulocyte-macrophage colony stimulatory factor (GM-CSF) and interferon-γ were up-regulated after exposure of the macrophages to the gingipains. Quantitative PCR analysis substantiated these observations and indicated that active or inactive forms of the high molecular weight gingipains were able to up-regulate expression of transcripts for these cytokines. The strongly enhanced production of IL-1β and GM-CSF by differentiated macrophages in response to active or inactive forms of the high molecular weight gingipains was confirmed at the protein level by ELISA analysis. The results indicate that the adhesin subunits of the gingipains mediate strong up-regulation of the expression of pro-inflammatory cytokines in macrophages.

Introduction

Periodontal disease is an oral disorder characterized by inflammation of the gingiva and the supporting structures of the periodontium [1], most likely as a consequence of a complex immunopathological host response to microbial challenge. This eventually may result in destruction of the periodontal tissues, alveolar bone resorption around the tooth root surface and ultimately, exfoliation of the teeth [2]. More than 300 different species of bacteria have been identified as members of the periodontal environment [1] and homeostasis in the healthy periodontium is maintained by a complex multifactorial process involving interactions between multiple bacterial species, their products, many of which are immunomodulating, and host cells. This balance can be disturbed by colonization of dental biofilm by predominantly Gram-negative, anaerobic bacteria considered to be periodontopathogenic, with Porphyromonas gingivalis deemed to be one of the most important [3]. Interaction of these bacteria with the host is ultimately pathogenic, with the host cell releasing cytokines and enzymes capable of destroying the host periodontal tissue and producing conditions of chronic inflammation in the gingiva and periodontium [1, 4].

P. gingivalis is a non-motile, Gram-negative, obligate anaerobe that has been implicated as a major causative agent in the development and progression of chronic periodontitis [5-7). Experimentally, P. gingivalis has been demonstrated to be a major pathogen, initiating the progression of periodontitis in mono-infected primate and rodent virulence models of the disease [8, 9]. The bacterium has a number of important virulence factors, including outer membrane vesicles, fimbriae, hemagglutinins, lipopolysaccharides and the cysteine proteases (gingipains). In particular, the cysteine proteases of P. gingivalis have been implicated as the causal link in the progression of periodontitis [10-15]. P. gingivalis produces two arginine-specific cysteine proteases, RgpA and RgpB, which are the products of two related genes, rgpA and rgpB, respectively, and one lysine specific cysteine protease, Kgp, the product of the kgp gene. The rgpA and kgp genes are initially translated into large pre-pro-peptide sequences, which are proteolytically processed to contain a catalytic domain with a C-terminal Ig-like sub-domain (IgSF) and four hemagglutinin/adhesin subunits [16]. In contrast, RgpB consists of only the catalytic domain and is missing nearly the entire gene region that encodes the hemagglutinin/adhesin subunits, with the exception of a small C-terminal segment [16] implicated in a novel pathway of gingipain secretion [17].

The gingipains have been implicated in a wide variety of pathogenic activities in the host and these pathogenic mechanisms are biphasic, as the cysteine proteases can both activate and inactivate host proteins by very specific limited proteolysis at selected peptide bonds in substrates [18]. The gingipains interact with host cells to cause a number of responses that may be associated with disease. The arginine-specific gingipains have been found to activate protease-activated receptors on epithelial cells [19] and platelets [20], with the former causing the up-regulation of interleukin-6 production by the cells. Protease-activated receptor-2 (PAR-2) appears to play a significant role in the progression of periodontal disease since PAR-2 null mice are apparently protected against periodontal disease [21]. The gingipains have been demonstrated to up-regulate production of a number of cytokines and other pro-inflammatory genes from monocytes [22], monocyte-epithelial cell co-cultures [23] and stromal cells or osteoblasts [24]. The perturbation of the cytokine signaling networks and proinflammatory genes are two mechanisms through which P. gingivalis may be able to initiate tissue destruction and alveolar bone resorption in the host.

Under conditions of chronic inflammation, such as in periodontal disease, the mononuclear phagocyte or macrophage is one of the principle cell types involved in the inflammatory response against infection. The exact role that macrophages play in periodontitis has not been fully elucidated, but it is known that exposure to P. gingivalis and components such as LPS or fimbriae induces the expression of pro-inflammatory cytokines [25, 26]. The aim of this study was to determine whether any of these responses could be attributed to the gingipains. We have examined the responses of differentiated macrophage-like cells to the gingipains of P. gingivalis and show that there is marked up-regulation of expression of IL-1β in particular, but that this expression is induced independently of the proteolytic activity of the gingipains.

Materials & Methods

Materials

The human monocytic cell line, THP-1, was purchased from the American Type Culture Collection (ATCC), USA. Fetal bovine serum (FBS), L-glutamine, penicillin, phorbol 12-myristate 13-acetate (PMA), Pro-Phe-Arg-chloromethylketone (PPACK), non-enzymatic dissociation solution, heat-denatured sheared salmon testes DNA and streptomycin sulphate were purchased from Sigma-Aldrich (Sydney, Australia). All primers for polymerase chain reaction (PCR) and quantitative PCR (qPCR) were synthesized by Geneworks (Adelaide, Australia). Reverse-transcription PCR (RT-PCR) reagents, such as deoxynucleoside triphosphates (dNTPs), PCR markers, PCR buffer and 25 mM MgCl₂, were purchased from Promega (Madison, USA). The RNeasy mini kit was purchased from Qiagen (Hilden, Germany). RPMI 1640, SuperScript III reverse transcriptase, Platinum® SYBR® Green qPCR SuperMix, SYBR Safe reagent and Sybr Green master mix were purchased from Invitrogen (Carlsbad, USA). Pyrotell® Limulus Amebocyte Lysate gel clot test was purchased from Associates of Cape Cod Inc. (MA, USA). The human cytokine/receptor Atlas nylon cDNA expression array, AtlasImage 2.01 software and ExpressHyb were purchased from Clontech (BD Biosciences, CA, USA). α-³³P dATP was purchased from Amersham Biosciences (Buckinghamshire, UK). Micro Bio Spin columns were purchased from Biorad (Hercules, CA, USA). The Quantikine® human GM-CSF ELISA kit was obtained from R & D Systems, (Minneapolis, MN, USA). The human IL-1β ReadySETGo! and human IFN-γ ReadySETGo! ELISA kits were obtained from eBioscience (San Diego, CA, USA). Human PAR-1 activating peptide (PAR-1 A/P; TFFLR) and human PAR-2 activating peptide (PAR-2 A/P: SLIGRL) were synthesized as carboxy amide derivatives and purified by high performance liquid chromatography to >95% purity by AusPep (Melbourne, Australia).

Tissue culture conditions and differentiation of THP-1 cells

The human monocytic cell line, THP-1, was maintained in RPMI 1640 medium, supplemented with 10% heat-inactivated fetal bovine serum, 2 mM L-glutamine, penicillin G (100 U/mL) and streptomycin sulphate (100 μg/mL) in 5% CO₂ at 37°C. For microarray and qPCR experiments, THP-1 cells were seeded at a density of 1 × 10⁷ cells/flask in 75cm² flasks, grown until they reached a density of 1 × 10⁶ cells/mL, washed twice in phosphate buffered saline (PBS) and treated with 10 nM PMA for 48 h in 5% CO₂ at 37°C to induce differentiation into macrophage-like adherent cells. For ELISA experiments, THP-1 cells were seeded at a density of 2 × 10⁶/well and were differentiated as described above in six well tissue culture plates (Techno Plastic Products, Zurich, Switzerland).

Purification and activation of the gingipains

The enzymes RgpB, HRgpA and Kgp were purified from culture supernatants of P. gingivalis strain HG66 as previously described [27, 28]. The gingipains were activated with 10 mM cysteine at 37°C for 10 min in 0.2 M Tris-HCl, 0.1 M NaCl and 5 mM CaCl₂, pH 7.6 and, when necessary, inactivated by incubating with 100 μM PPACK at 37°C for 20 min.

RNA isolation and synthesis of cDNA

For microarray and qPCR experiments, after PMA treatment the cells were washed twice in PBS, incubated overnight in serum-free RPMI media, then challenged with 2.5 nM active HRgpA, 2.5 nM inactive HRgpA and 2.5 nM active Kgp for 1 h at 37°C in 5% CO₂ (conditions that had been used in a previous study of cytokine responses in epithelial cells without any adverse effects on the cells [19]). For the microarray experiments, immediately after the 1 h treatment, the THP-1 cells were washed twice in PBS and incubated with non-enzymatic dissociation solution at 37°C in 5% CO₂ for 5 min to detach the cells from the flask. After detachment, the cells were lysed and the RNA was immediately extracted using the RNeasy mini kit according to the manufacturer’s instructions. For the qPCR experiments, after treatment for 1 h, the cells were washed twice in PBS and then either the RNA was extracted immediately as described above or the cells were incubated in serum-free RPMI media with the RNA subsequently extracted 3, 12 or 24 h after the initial incubation period. The quality and quantity of purified RNA was checked by agarose gel electrophoresis and UV spectroscopy at 260 nm and 280 nm. RNA samples with a ratio of A260/A280 ≥ 2.0 and without degradation as analysed by agarose gel electrophoresis were used for microarray assays and qPCR experiments. All samples were stored at −80°C until required. Reverse transcription was conducted with 1 μg of total cellular RNA using SuperScript III reverse transcriptase (200U/μL) for first strand cDNA synthesis in a 20 μL reaction volume in accordance with the manufacturer’s instructions. The synthesised cDNA was subsequently used as a template in qPCR experiments.

Oligonucleotide cDNA microarray analysis

The human cytokine/receptor Atlas nylon cDNA expression array containing duplicate cDNAs from 268 human cytokine related genes and 9 housekeeping genes immobilized onto a nylon membrane was used in accordance with the manufacturer’s instructions. Briefly, radiolabeled cDNA probes were prepared from 5 μg of total RNA for each sample by a reverse transcription step in the presence of 3.5 μL (α-³³P) dATP, 1 μL CDS primer mix, 1 μL Superscript III (200 U), 2 μL dNTP (5 mM), 0.5 μL DTT (100 mM) and incubated at 42°C for 50 min. The probe was purified after synthesis using Micro Bio Spin columns. The incorporation of the radiolabel was checked using a scintillation counter and probes that had a count per minute (cpm) above 1 × 10⁶ were deemed to have satisfactory signal strength for the array analysis. The array membranes were washed with deionized H₂O and pre-hybridized using 5 mL pre-warmed ExpressHyb containing 0.5 mg of heat-denatured sheared salmon testes DNA at 68°C for 4 h. Before hybridization, the cDNA probe was denatured by the addition of 5.5 μL of 10 x denaturing solution and was incubated at 68°C for 20 min. After denaturation of the probe, 5 μg of Cot-1 DNA and 55.5 μL of neutralization solution were added to the cDNA probe and incubated at 68°C for 10 min. Subsequently, the cDNA probe was added to 5 mL pre-warmed ExpressHyb and the membranes were hybridized with the cDNA probes at 68°C for 16 h with continuous rotation in a hybridization oven. After hybridization, these membranes were washed four times with 2x standard saline citrate (SSC) and 1% (w/v) sodium dodecyl sulphate (SDS) at 68°C for 30 min, once in 0.1 X SSC-0.5% (w/v) SDS solution for 30 min at 68°C and finally with 2 x SSC at 68°C for 10 min. The membranes were exposed to X-Ray film and developed. Quantification and normalization was performed using AtlasImage 2.01 software. This software automatically averaged the two hybridization signals for a particular gene and compared the experimental intensity to the same gene in the control array. Genes showing a two-fold difference in signal intensity between experimental and control membranes were selected for confirmation of differential expression by qPCR.

Quantitative real-time PCR

To determine the relative levels of expression of specific cytokines expressed in differentiated THP-1 cells treated with 2.5 nM active HRgpA, 2.5 nM inactive HRgpA or 2.5 nM active Kgp, qPCR was performed using a Corbett Real Time PCR Rotorgene 3000 machine (Corbett Research, Sydney, Australia). Forward and reverse oligonucleotide primer pairs (Table 1) were designed specifically for qPCR using the web based software program, Primer3 [29] and designed to overlap exon-exon boundaries to ensure specific amplification of mRNA transcripts. qPCR reactions were optimized for each primer pair by varying concentrations of primer and cDNA. For each gene of interest and the housekeeping gene glyceraldehyde 3-phosphate dehydrogenase (GAPDH), qPCR assays were performed on three biological samples, in triplicate. Cycling was performed using 10 μL Platinum® SYBR® Green qPCR SuperMix, 500 nM of each forward and reverse primer and 1 μL cDNA template, using the following cycle profile: 55°C for 2 min, 95°C for 2 min, followed by 40 cycles consisting of 95°C for 30 s, 55°C for 30 s and 72°C for 30 s, followed by a thermal melt profile from 60°C to 95°C. Fluorescence readings were detected at the end of each extension step using the FAM/SYBR channel of the machine. Upon completion of cycling, a melt curve of fluorescence versus temperature was generated to ensure that a single amplicon was obtained for each primer set. Following amplification, cycle threshold (C_T) values for house keeping genes and genes of interest were exported to Microsoft Excel. Relative expression of each gene of interest in treated samples, expressed relative to untreated samples and normalized to the house keeping gene , along with statistical analysis using the non-parametric randomisation reallocation test and determination of the standard error for treatments was performed using the relative expression software tool (REST) [30].

Table 1. Sequences of primers used for quantitative PCR analysis of cytokines.

Target cDNA	Primer Name	Sequence (5′-3′)	Product Size (bp)
GAPDH	GAPDH_FP	CAATGCCTCCTGCACCAC	128
	GAPDH_RP	GGCCATCCACAGTCTTCTG
IL-1β	IL-1β_FP	GCTGAGGAAGATGCTGGTTC	126
	IL-1β_RP	TGCACATAAGCCTCGTTATCC
IL-10	IL-10_FP	CCAAGACCCAGACATCAAGG	138
	IL-10_RP	AAGGCATTCTTCACCTGCTC
GM-CSF	GM-CSF_FP	AGCCGACCTGCCTACAGAC	149
	GM-CSF_FP	AATCTGGGTTGCACAGGAAG
INF-γ	INF-γ_FP	TGTGGAGACCATCAAGGAAG	158
	INF-γ_RP	GCGACAGTTCAGCCATCAC
IRF1	IRF1_FP	AGACCAGAGCAGGAACAAGG	113
	IRF1_RP	TCTTAGCATCTCGGCTGGAC

ELISA analysis and cell normalisation

For ELISA experiments, after PMA treatment the cells were washed twice with PBS and challenged with 2, 20 or 200 nM active or inactive HRgpA, Kgp or RgpB, 100 μM PAR-1 AP or 100 μM PAR-2 AP for 18 or 36 hrs at 37°C in 5% CO₂. The conditioned culture medium was harvested, centrifuged at 150 x g for 5 min and the supernatant was aliquoted and stored at −20°C until required. After collection of the culture supernatant, cells were lysed in 500 μL 0.1 M Tris-HCl, pH 8, containing 0.5% (v/v) Triton-X-100 and the DNA content was measured as described previously [31] using the SYBR Safe reagent. DNA content was found to be essentially constant across all conditions tested and therefore results were not normalised to this parameter.

GM-CSF, IL-1β or interferon-γ (IFN-γ) released by THP-1 cells when challenged by the gingipains or peptide agonists were detected using sandwich ELISAs. GM-CSF released by THP-1 was detected using the Quantikine® human GM-CSF ELISA kit in accordance with the manufacturer’s instructions. The GM-CSF ELISA plates, pre-coated with murine monoclonal antibodies against recombinant human GM-CSF, were incubated with 100 μL serially diluted recombinant human GM-CSF standards or conditioned media at room temperature (RT) for 2 h. Bound GM-CSF was detected using 200 μL anti-human GM-CSF antibody conjugated to horseradish peroxidase (HRP). The concentrations of GM-CSF in the samples were calculated from the linear range of the standard curve.

IL-1β or IFN-γ released by THP-1 cells were detected using the human IL-1β ReadySETGo! and the human IFNγ ReadySETGo! ELISA Kits, respectively, in accordance with the manufacturer’s instructions. The ELISA plates, pre-coated with monoclonal antibodies raised against recombinant human IL-1β (clone CRM56) or recombinant human IFN-γ (clone NIB42) were incubated with 100 μL serially diluted recombinant human IL-1β/ IFN-γ standards or conditioned media overnight at 4°C. After washing, 100 μL/well of biotin-conjugated detection antibody was added, incubated at RT for 1 h, following which 100 μL/well of avidin-HRP conjugate was added and incubated at RT for 30 min. Detection was carried out using tetramethylbenzidine (TMB) substrate solution which was read at 450 nm using a BioRad 3550 Microplate reader. The concentrations of IL-1β/IFN-γ in the samples were calculated from the linear range of the standard curve.

Limulus Amebocyte Lysate test

To detect and quantify the presence of Gram-negative bacterial endotoxin (lipopolysaccharide [LPS]) contamination of the purified RgpA, RgpB or Kgp samples, the Pyrotell® Limulus Amebocyte Lysate gel clot test was undertaken in accordance with the manufacturer’s instructions. In brief, a series of endotoxin standards were prepared and tested to confirm the sensitivity of the Pyrotell® reagent to be 0.03 endotoxin units (EU) per ml. Subsequently, two-fold serial dilutions (ranging from 2.5 nM to 78.125 pM) of heat-inactivated Kgp, RgpB or HRgpA were prepared and 200 μL of each dilution was added to single test vials of Pyrotell® which contained 200 μL of lyophilized Limulus amebocyte lysate. Tubes were incubated at 37°C for 1 hr. After the incubation period, if a gel clot had formed in the vials and remained intact after the vials were inverted by 180°, the sample was deemed positive for LPS contamination. No detectable amount of LPS contamination was seen for any of the gingipain preparations.

Results

Microarray analysis of gingipain-induced expression of cytokines and receptors in THP-1 cells

The Human Cytokine/Receptor Atlas Nylon cDNA Expression Array was used as an initial screen to examine cytokine/receptor expression in THP-1 cells when challenged with 2.5 nM active or inactive HRgpA or 2.5 nM active Kgp. The concentrations of enzymes used were based on previous experience with epithelial cells, where only small concentrations of the enzymes were tolerated by the cells without adverse effects such as lifting of the cells [19], although later in the study it was established that the differentiated THP-1 cells were resistant to much higher concentrations of enzyme without lifting of the cells from the surface occurring. Similarly, a time point of 1 h was used based on previous experiences, with RNA isolated immediately after exposure to the gingipains. The effect of time on expressed genes was investigated later using qPCR analysis. Each microarray analysis was performed twice using two biological samples for each treatment and the genes that were at least 2-fold differentially expressed across both microarrays are listed in Tables 2-4. Relative to untreated controls, differentiated THP-1 cells challenged with active HRgpA for 1 h showed up-regulation (≥2 fold ratio HRgpA/control) of the expression of 9 cytokine/cytokine receptor genes (3.4%) of the 268 genes spotted onto the array (Table 2). These genes included interleukin-4 receptor (alpha subunit), IFN-γ, interferon-gamma receptor (IFN-γR), interferon regulatory factor 1 (IRF1), CDW40 antigen, TRK-T3 oncoprotein, fibroblast growth factor 8 (FGF8) and epidermal growth factor (EGF). The most highly up-regulated gene was IFN-γ, which was up-regulated 7.7-fold.

Table 2. Genes on ATLAS cytokine array differentially regulated by active HRgpA.

Name	Differentially regulated genes	GenBank accession number	Relative fold change*
IL4R	interleukin-4 receptor alpha subunit precursor	X52425	2.00
IFN-γ	interferon gamma precursor	X01992	7.72
IFN-γR	interferon-gamma receptor	J03143	2.31
IRF1	interferon regulatory factor 1	X14454	2.34
CDW40	CDW40 antigen; CD40L receptor precursor	X60592	4.62
trk-T3	trk-T3; P68 TRK-T3 oncoprotein	X85960	2.13
FGF8	fibroblast growth factor 8; androgen-induced growth factor precursor ; HBGF8	U36223	2.14
EGF	kidney epidermal growth factor; urogastrone	X04571	2.37

^*Each value is the mean of duplicate hybridisations for one representative microarray.

Table 4. Genes on ATLAS cytokine array differentially regulated by active Kgp.

Name	Differentially regulated genes	GenBank accession number	Relative fold change*
IL-1R1	interleukin-1 receptor type I precursor	M27492	184.9
IL4R	interleukin-4 receptor alpha subunit precursor	X52425	4.52
IL5RA	interleukin-5 receptor alpha subunit precursor	M75914	2.43
IL6R-alpha	interleukin-6 receptor alpha subunit precursor	M20566	5.40
IRF1	interferon regulatory factor 1	X14454	2.92
BCGF1	B-cell growth factor 1 precursor	M15530	4.15
EPH-3	tyrosine-protein kinase receptor EPH-3	L41939	9.13
NT4	neurotrophin-4	M86528;	3.78
ERBB2	ERBB2 receptor protein-tyrosine kinase	M95667	10.95
VEGFB	vascular endothelial growth factor B precursor + VEGF-related factor isoform VRF186	U48801; U43369	2.92
Wnt-5A	Wnt-5A	L20861	2.41
JAG2	jagged2	AF003521	3.41
NOTCH1	neurogenic locus notch protein homolog 1 precursor	M73980	2.31
HGF activator	hepatocyte growth factor activator	D14012	3.15
PHB	prohibitin	S85655	3.57

^*Each value is the mean of duplicate hybridisations for one representative microarray.

Inactive HRgpA challenge of differentiated THP-1 cells induced the up-regulation of 27 (10.1%) of the 268 cytokine/receptor genes (Table 3). Significantly up-regulated genes included IL-1β, interleukin-10 (IL-10), interferon-γ antagonist, interferon-γ receptor (IFNR-gamma), IRF1, B-cell growth factor 1 precursor (BCGF1), GM-CSF, macrophage inflammatory protein 1α precursor (MIP1-alpha), macrophage colony stimulating factor I receptor (MCSF-1-R) and TNFα. The most highly up-regulated gene was interferon-alpha/beta receptor beta subunit, which was up-regulated 172-fold. IFN-γR, IRF1, CDW40 antigen and TRK-T3 oncoprotein were up-regulated by both active and inactive HRgpA, implying that the proteolytic activity of HRgpA was not required for the induction of expression of these genes.

Table 3. Genes on ATLAS cytokine array differentially regulated by inactive HRgpA.

Name	Differentially regulated genes	GenBank accession number	Relative fold change*
IL-1β	interleukin-1 beta precursor	K02770	4.51
IL-10	interleukin-10 precursor	M57627	2.07
IL-1RA	interleukin-1 receptor antagonist protein precursor	M63099	2.66
IL-1R1	interleukin-1 receptor type I precursor	M27492	112.34
IL-5RA	interleukin-5 receptor alpha subunit precursor	M75914	4.66
IL-6RA	interleukin-6 receptor alpha subunit precursor	M20566;	17.72
IFN-alpha-RA	interferon-alpha/beta receptor beta subunit precursor	X77722	172.35
IFN-γ antagonist	interferon gamma antagonist	A25270	2.46
IFNR-γ	interferon-gamma receptor	J03143	3.16
IRF1	interferon regulatory factor 1	X14454	2.54
BCGF1	B-cell growth factor 1 precursor	M15530	8.93
GM-CSF	granulocyte-macrophage colony stimulating factor	M11220	7.92
MIP1-alpha	macrophage inflammatory protein 1 alpha precursor	M23452	4.37
C5AR	C5a anaphylatoxin receptor	M62505	2.94
CDW40	CDW40 antigen; CD40L receptor precursor	X60592	4.22
CC CKR1	CC chemokine receptor type 1; macrophage inflammatory protein 1 alpha receptor; RANTES R; LD78 receptor; CMKBR1	D10925	3.72
CSF-1-R	macrophage colony stimulating factor I receptor precursor; fms proto-oncogene; CD115	X03663	2.57
trk-T3	trk-T3; P68 TRK-T3 oncoprotein	X85960	4.22
TFAR15	apoptosis-related protein TFAR15	AF022385	2.36
TNF-alpha	tumor necrosis factor precursor; cachectin	X01394	2.11
PDGF AP	PDGF associated protein	U41745	3.07
GAF	glia-activating factor precursor; fibroblast growth factor 9	D14838	2.42
FGFR1	fibroblast growth factor receptor1 precursor; basic fibroblast growth factor receptor;	X66945;	4.31
VEGFB	vascular endothelial growth factor B precursor + VEGF-related factor isoform VRF186	U48801; U43369	2.72
Wnt-5A	Wnt-5A	L20861	2.74
NOTCH1	neurogenic locus notch protein homolog 1 precursor; translocation-associated notch protein	M73980	3.29
BSG	basigin precursor; leukocyte activation antigen M6; collagenase stimulatory factor; extracellular matrix metalloproteinase inducer; CD147 antigen	L20471	3.80

^*Each value is the mean of duplicate hybridisations for one representative microarray.

Challenge of the differentiated THP-1 cells with active Kgp for 1 h resulted in up-regulation of 15 (5.6%) of the 268 cytokine/receptor genes (Table 4). Up-regulated genes included interleukin-1 receptor type I (IL-1R1), IL-4 receptor (alpha subunit), IL-6 receptor (alpha subunit), IRF1 and BCGF1. The most highly up-regulated gene was IL-1R1, which was up-regulated by 185-fold. Interestingly, IRF1 was up-regulated by all three treatment conditions. Cumulatively, the initial screen using the cytokine gene array indicated that IL-1β, IFN-γ, GM-CSF, IL-10 and IRF1 were interesting candidate genes to investigate further using qPCR.

Quantitative PCR analysis reveals that the gingipains induce the expression of cytokines and a modulator in a time-dependent manner

Expression of the cytokine and cytokine receptor genes IL-1β, IFN-γ, GM-CSF, IL-10 and IRF1 identified by gene arrays was further examined using qPCR. The relative levels of expression of these genes in differentiated THP-1 cells over time when treated with active and inactive forms of the gingipains are shown in Figure 1. THP-1 cells were treated with 2.5 nM active or inactive HRgpA, or 2.5 nM active Kgp for 1 hr and the RNA was extracted either immediately or 3, 12 or 24 h after the initial incubation period. The pro-inflammatory cytokine genes IL-1β and IFN-γ were significantly up-regulated in THP-1 cells 24 h after being treated with active or inactive HRgpA, or active Kgp for 1 h. Active HRgpA up-regulated the expression of IL-1β by 27-fold, inactive HRgpA up-regulated the expression of IL-1β 23-fold, while stimulation with active Kgp resulted in the 30-fold up-regulation of IL-1β. IFN-γ was less strongly up-regulated by the gingipains: challenge with active or inactive HRgpA induced a 10.7-fold increase in the expression of IFN-γ by THP-1 cells 24 h after initial challenge, while active Kgp induced a 15-fold up-regulation in the expression of IFN-γ after 24 h.

Figure 1.

Analysis of IL-1β, IFN-γ, GM-CSF, IL-10 and IRF1 expression in THP-1 cells in response to stimulation with 2.5 nM active HRgpA (dark bars), 2.5 nM inactive HRgpA (grey bars) or 2.5 nM active Kgp (clear bars) for 1 h. cDNA synthesized from RNA extracted immediately (1 h) or 3, 6 or 24 h later from THP-1 cells challenged with each treatment was used as a template for qPCR reactions using IL-1β, IFN-γ, GM-CSF, IL-10, IRF1 and GAPDH primers. The C_T values were used to calculate IL-1β, IFN-γ, GM-CSF, IL-10 and IRF1 transcript levels in treated cultures normalised to GAPDH relative to the normalised level of the transcripts for the same molecules in untreated samples. Data is presented as the mean ± SEM (n=3), ***P≤0.001. Comparisons between two groups were made by unpaired Student’s t-tests.

The expression of GM-CSF was very powerfully and significantly up-regulated in THP-1 cells 24 h after being treated with the active or inactive forms of the gingipains for 1 h. Active and inactive HRgpA induced an up-regulation in GM-CSF expression in differentiated THP-1 cells by 1291- and 1432-fold, respectively, while active Kgp induced a 2475-fold up-regulation in this gene. The anti-inflammatory cytokine IL-10 gene was significantly up-regulated by treatment with active (2.3-fold) or inactive HRgpA (2.3-fold) 3 h after the initial incubation period. IRF1 was up-regulated in THP-1 cells 24 h after stimulation with active or inactive HRgpA, or active Kgp for 1 h. Treatment with active HRgpA or Kgp up-regulated the expression of IRF1 by 3.8-fold in THP-1 cells, whereas treatment with inactive HRgpA up-regulated the expression of IRF1 by 4.6-fold.

ELISA analysis reveals significant induction of IL-1β and GM-CSF, but not IFN-γ

To verify the up-regulation in expression of IL-1β, IFN-γ and GM-CSF gene products in differentiated THP-1 cells when challenged with active or inactive forms of the gingipains, ELISA analysis was undertaken. THP-1 cells were stimulated with 2, 20 or 200 nM active or inactive HRgpA, active or inactive Kgp or active or inactive RgpB in comparison to 100 μM PAR-1 AP or 100 μM PAR-2 AP; the conditioned medium was recovered at 18 or 36 h, following which the levels of IL-1β, IFN-γ and GM-CSF protein in each treatment condition were analysed by ELISA (Fig. 2). Active or inactive HRgpA potently induced the secretion of both IL-1β and GM-CSF in a dose-dependent manner at both 18 and 36 h after treatment. Similarly, active or inactive Kgp significantly up-regulated the secretion of IL-1β and GM-CSF in a dose-dependent manner with the exception of the treatment with 200 nM active Kgp and IL-1β, where the amount of detectable IL-1β at 18 h was lower than that seen for 20 nM active Kgp at the same time point. However, dose-dependent stimulation of IL-1β by active Kgp was observed at 36 h.

Figure 2.

IL-1β and GM-CSF secretion in THP-1 cells stimulated with 2-200nM active HRgpA (HA), inactive HRgpA (HI), active Kgp (KA), inactive Kgp (KI), active RgpB (RA), inactive RgpB (RI), 100 μM PAR-1 AP (P1AP) or 100 μM PAR-2 AP (P2AP), compared to a control (CTL). The conditioned medium was recovered at 18 or 36 h (dark and light bars, respectively) and analysed for levels of IL-1β and GM-CSF and compared to controls. Data is presented as the mean ± SEM (n=3), *P≤0.05, **P ≤0.01 ***P≤0.001. Comparisons between two groups were made by unpaired Student’s t-tests.

In contrast to HRgpA and Kgp, active or inactive forms of RgpB were not potent stimulators of IL-1β and GM-CSF production in THP-1 cells. However, 20 nM active RgpB was able to stimulate the secretion of 168.9 ± 9.9 pg/ml of IL-1β at 18 h, which was a 9.6-fold increase when compared to control IL-1β levels (17.6 ± 1.0 pg/ml). At 36 h, 20 nM active RgpB was still able to stimulate secretion of 132.9 ± 2.2 pg/ml of IL-1β, 7-fold higher than the control IL-1β levels (19.1±0.65 pg/ml). Similarly, active or inactive forms of RgpB were weak stimulators of GM-CSF secretion in THP-1 cells, with 20 nM active RgpB at 18 h yielding the most potent stimulation observed for this enzyme (13.5 ± 0.6 pg/ml GM-CSF). However, this stimulatory effect was not evident at 36 h.

PAR-1 activating peptide was a weak stimulator of both IL-1β and GM-CSF levels when compared to controls at both 18 and 36 h. The PAR-2 activating peptide did not stimulate THP-1 cells to secrete either IL-1β or GM-CSF above control levels at either time point. The conditioned media from THP-1 cells treated with all activators for 18 or 36 h did not contain detectable amounts of IFN-γ (data not shown).

Discussion

Periodontal disease results from chronic inflammatory conditions induced in the gums and supporting structures for teeth, which may result in tooth loss and systemic sequelae due to the prolonged inflammatory conditions [1]. Porphyromonas gingivalis has been implicated as one of the major bacterial species involved in the pathogenesis of the disease [5-9] and the gingipain cysteine proteases have been demonstrated to be major virulence factors in models of periodontal disease [12-15]. The inflammation caused in the oral environment probably results from a complex interplay between the bacterium and human immune cells [2-4], with the paradoxical requirement of these cells to prevent the bacterial infection offset by their responses to the pathogen being strong contributors to the disease [4]. It is very likely that the cytokines produced by human immune cells strongly influence the outcome of the inflammatory response and in this regard it is also probable that macrophages play a crucial role, as they generally orchestrate many of the responses in inflammatory events surrounding pathogen invasion. While the importance of these cells in periodontal disease and their response to some components of the bacterium (LPS and fimbriae) have been mapped [26,32,33], their response to the gingipains has not yet been examined. In this study, we have examined the response of model macrophage cells to three forms of the gingipains in terms of the profile of cytokines and cytokine receptors expressed by the cells.

Previously, we have shown that activation of protease-activated receptors (PARs) by arginine-specific gingipains causes cytokine responses in oral epithelial cells [19] and therefore our null hypothesis at the outset of this study was that effects of the gingipains on the macrophages would most likely be mediated via the PARs. With this in mind, active and inactive forms of the high molecular weight arginine-specific gingipain, HRgpA, were particularly employed to reveal the differences between the responses most likely due to PAR-mediated effects (from active HRgpA) and those mediated by other effects (from inactive HRgpA). We had established previously that the gingipain preparations were not contaminated with LPS and thus responses derived from the cells were most likely to result from interaction with the protease-adhesin subunits of these molecules. Our first set of experiments involved the use of a cytokine/cytokine receptor array to probe which genes were being markedly up- or down-regulated following exposure to the gingipains. To our surprise, the inactive form of HRgpA up-regulated a much greater number of genes on the array (29 genes) than the active form (8 genes) and the level of the increase was also much greater on average following exposure to inactive HRgpA (several genes above 5-fold enhancement in expression, only one above 5 for active HRgpA). Interestingly, more genes were commonly regulated by inactive HRgpA and active Kgp (7 genes) than were commonly regulated by active and inactive forms of HRgpA (3 genes), suggesting that the adhesin domains of HRgpA and Kgp, which have 90% similarity, are mediating much of the response generated in cells as this would be the major commonality between the proteases. The inactive HRgpA and Kgp induced a similarly large up-regulation of the interleukin-1 receptor type I precursor (over 100-fold in each case). Overall, most cytokines and their receptors up-regulated by the gingipains would be considered pro-inflammatory, suggesting that the interaction of the bacterial enzymes with macrophages would promote a strong pro-inflammatory response from these cells. In order to further investigate and validate the responses from the cytokine array, the responses of 4 genes noted to be up-regulated in the array experiment (IL-1β, IFN-γ, GM-CSF and IRF-1) were further investigated using quantitative PCR. The expression of an anti-inflammatory gene, IL-10, was also investigated to provide further insight into the balance between pro- and anti-inflammatory responses of the macrophages in response to the gingipains.

Quantitative PCR (qPCR) analysis was carried out with the same set of gingipain forms as used for the array experiments and also the same concentrations. It was found that IL-10 was moderately up-regulated in its expression by all forms of gingipains at the 3 h time point following exposure to the gingipains. In contrast, the other, pro-inflammatory genes were up-regulated much more strongly, but at 24 h after the initial exposure of the cells to the gingipains. It is possible that the difference in timing required to see upregulation of the gene expression using qPCR relates to the different sensitivities of the techniques employed. The most strongly up-regulated gene was GM-CSF, in response to exposure to all gingipain forms, with IL-1β and IFN-γ markedly up-regulated and IRF only up-regulated approximately 4-fold. These results indicated that pro-inflammatory cytokine genes were up-regulated by all forms of the gingipains, including inactive HRgpA, once again suggesting that the adhesin domains common between HRgpA and Kgp might be mediating the majority of the interactions leading to the response in the cells. The results also suggested that the pro-inflammatory genes were up-regulated late in the macrophages following exposure to the gingipains.

In order to confirm that changes seen in gene expression were also seen in terms of protein levels of the relevant cytokines, ELISA analysis was carried out for the 3 cytokines, IL-1β, IFN-γ and GM-CSF. The ELISA analysis was carried out with higher concentrations of the gingipains since initial results did not yield marked regulation of the protein levels using the low concentrations of the gingipains utilized in the gene expression analyses. This might simply reflect the different sensitivities of the techniques involved. Only low responses were seen for 2 nM levels of the HRgpA and Kgp for all cytokines. Using ELISA analysis, no responses were seen for any treatment for IFN-γ. This result, contradictory to the gene expression analysis, might reflect the different sensitivity once again as the IFN-γ was a gene that yielded a response on a very late cycle in the qPCR, indicating that very small quantities of the mRNA for the gene were present. This might indicate that while the gene was indeed up-regulated in its expression, the actual levels of protein were in reality very low. GM-CSF protein levels were strongly up-regulated at the protein level by using 20 or 200 nM of active/inactive HRgpA or Kgp, with active HRgpA prompting the highest response, closely followed by the inactive form of the enzyme. Use of a low molecular weight form of the arginine-specific gingipain, RgpB, which lacks the adhesin subunits, but is a reasonably potent activator of PAR-2 at the levels used, gave rise to quite small responses in GM-CSF relative to the high molecular weight forms of the enzymes, indicating again the importance of the adhesin subunits of the enzymes in mediating the effects of the gingipains on macrophages. Additionally, agonist peptides for PAR-1 and PAR-2 were found to yield much lower responses for GM-CSF, indicating that the PARs were not playing a major role in mediating the responses to the gingipains in differentiated macrophages. Interestingly, a recent paper has shown that in the undifferentiated form of the monocytic cell line being used here, PARs played a major role in the generation of the IL-8 response prompted by exposure to the gingipains [22]. This suggests that the undifferentiated versus differentiated forms of these cells apparently respond quite differently to the gingipains. It should be noted here that RT-PCR analysis revealed that PAR-1, -2 and -3 were expressed by the differentiated THP-1 cells (Fitzpatrick et al., unpublished results) and therefore should have been available to contribute to the responses to the arginine-specific gingipains. Indeed, it is likely that the difference in the response between active and inactive HRgpA might be attributable to PAR activation, as might be the responses to RgpB.

Interleukin-1β levels were massively up-regulated following exposure to either of the high molecular weight gingipain forms, HRgpA and Kgp. This cytokine is seen as an early responder in inflammatory situations, with subsequent response of cells including the up-regulation of the receptor for the interleukin, which was seen in the array analysis presented here, and subsequent release of many other cytokines [34]. This has previously been characterized to occur for oral epithelial cells in response to P. gingivalis or its products [34] and thus it is of interest to note here that the gingipains were also able to up-regulate this pivotal cytokine. Once again, the predominant response of the cells appears to be mediated via the adhesin subunits of HRgpA and Kgp, with only minor contributions occurring via the activity of the enzymes.

The profile of cytokine responses induced by the gingipains appears to indicate that these enzymes induce the majority of their response from differentiated macrophages through their adhesin subunits. At this stage, the receptors for these adhesin subunits on cells generally are not well characterized. It is conceivable that the adhesin subunits might be interacting with the Toll-like receptors which are found on the surface of the differentiated macrophages, but this has never been shown. The fimbriae and LPS of P. gingivalis have been shown to bind TLR-2 and TLR-7 and induce cytokine responses in macrophages [33] and therefore it would appear likely that the adhesin subunits of the gingipains may act to further augment such responses. The molecular targets of the gingipains on the surface of the cells remain to be characterized, but it seems clear that antagonizing such interactions might be one potential avenue to prevent pro-inflammatory responses from tissue macrophages which might mediate much of the unwanted chronic pro-inflammatory responses found in oral tissues following infection with P. gingivalis.