- Email: [email protected]
Gene expression profiling of pulpal tissue reveals the molecular complexity of dental caries
Biochimica et Biophysica Acta 1741 (2005) 271 – 281 http://www.elsevier.com/locate/bba
Gene expression profiling of pulpal tissue reveals the molecular complexity of dental caries Julia L. McLachlana, Anthony J. Smitha, Iwona J. Bujalskab, Paul R. Coopera,* a Oral Biology, School of Dentistry, University of Birmingham, Birmingham, B4 6NN, UK Institute of Biomedical Research, Endocrinology and Metabolism, The Medical School, University of Birmingham, Edgbaston, Birmingham, B15 2TT, UK
b
Received 10 November 2004; received in revised form 9 February 2005; accepted 8 March 2005 Available online 8 April 2005
Abstract High-throughput characterisation of the molecular response of pulpal tissue under carious lesions may contribute to improved future diagnosis and treatment. To identify genes associated with this process, oligonucleotide microarrays containing ¨15,000 human sequences were screened using pooled total RNA isolated from pulpal tissue from both healthy and carious teeth. Data analysis identified 445 genes with 2-fold or greater difference in expression level, with 85 more abundant in health and 360 more abundant in disease. Subsequent gene ontological grouping identified a variety of processes and functions potentially activated or down-modulated during caries. Validation of microarray results was obtained by a combination of real-time and semi-quantitative PCR for selected genes, confirming down-regulation of Dentin Matrix Protein-1 (DMP-1), SLIT2, Period-2 (PER2), Period-3 (PER3), osteoadherin, Glypican-3, Midkine, activin receptor interacting protein-1 (AIP1), osteoadherin and growth hormone receptor (GHR), and up-regulation of Adrenomedullin (ADM), Interleukin-11 (IL-11), Bone sialoprotein (BSP), matrix Gla protein (MGP), endothelial cell growth factor-1 (ECGF1), inhibin hA and orosomucoid-1 (ORM1), in diseased pulp. Real-time PCR analyses of ADM and DMP-1 in a panel of healthy and carious pulpal tissue and also in immune system cells highlighted the heterogeneity of caries and indicated increased expression of ADM in neutrophils activated by bacterial products. In contrast, DMP-1 was predominantly expressed by cells native to healthy pulpal tissue. This study has greatly extended our molecular knowledge of dental tissue disease and identified involvement of genes previously unassociated with this process. D 2005 Elsevier B.V. All rights reserved. Keywords: Pulpal tissue; Diagnosis; Treatment
1. Introduction A recent report by the World Health Organization highlighted dental caries as a growing problem in both industrialized and, increasingly, in developing countries affecting 5 billion people worldwide. The disease impacts on all age groups causing pain, suffering, impairment of function and a reduced quality of life. Treatment costs are estimated to account for between 5 and 10% of total health costs in industrialized countries and are beyond the resources of many developing countries [1,2]. Caries is multifactorial involving * Corresponding author. Tel.: +44 121 237 2785; fax: +44 121 237 2882. E-mail address: [email protected] (P.R. Cooper). 0925-4439/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.bbadis.2005.03.007
the interplay between genetics, diet, life-style and invading microflora and preventative disease management regimes have been targeted at several of these primary risk factors [3]. So far, however, these approaches have been insufficient for disease eradication. Tooth restorative treatments following caries are still fairly empirical and involve excavation of the decayed tissue and subsequent replacement with relatively inert filling materials. This procedure, however, requires critical placement and failure rates of up to 60% have been reported which necessitates either re-treatment or results in tooth loss [4]. The dentine –pulp complex of the tooth has inherent natural regenerative properties and two mechanisms for this process are described. Following relatively mild tooth injury,
272
J.L. McLachlan et al. / Biochimica et Biophysica Acta 1741 (2005) 271 – 281
odontoblasts lining the pulp chamber, localised beneath the lesion, are stimulated to up-regulate their synthesis and secretion of a tertiary dentine extracellular matrix (ECM), which subsequently protects the underlying cells and maintains the tissues vitality. Injury of greater intensity, however, can lead to local odontoblast necrosis and replacement with a new generation of odontoblast-like cells, recruited from the pulp core, which subsequently secrete new mineralized dentine ECM beneath the lesion [5]. Clearly, the latter process is more complex involving progenitor cell recruitment and induction of differentiation prior to new matrix secretion. Recently, studies have shown that the inflammatory process, which ensues due to carious infection, can negatively impact on these repair mechanisms [6,7]. Whilst this immune response is well characterized histologically [8 – 13], our knowledge of its molecular mediation is partial with the involvement of only a limited number of cytokines and other related immune regulatory molecules having been described [14 –19]. Genome sequencing projects and advances in microarray technologies now provide us with reliable tools to obtain a comprehensive picture of the molecular events that occur during disease progression. Such an analysis of caries could provide information on the recruitment and involvement of specific cell types, cellular activation mechanisms, expression of genes as a result of interactions between resident and immune cell populations and on repair associated molecules. Specifically, known and novel molecules previously not associated with the disease may be identified. Clinically, the description and classification of molecular changes will enable (a) dissection of the molecular circuitry of the disease process, (b) identification and validation of potential targets for therapeutic intervention and (c) identification of potential biomarkers for improved diagnosis and validation. Previously, we have developed a temperature-sensitive dental tissue extraction procedure for the selective isolation of core pulpal tissue whilst leaving the peripheral odontoblast population attached to the dentine ECM [20]. The aim of this present study was therefore to characterize gene expression profiles within this core pulpal tissue from clinically healthy and severely carious human teeth. Along with confirmation of differential expression profiles for selected genes, the differentially expressed datasets were subsequently grouped into biological and molecular ontological classes. In addition, as the gene expression data represented an average response for the heterogeneous population of cells present within the pulp, further specific cell and tissue expression profiling was performed.
2. Experimental procedures 2.1. Extraction of dental tissue Carious and sound mature human premolar and molar teeth, extracted for orthodontic purposes from patients in the
age range of 20 to 30 years, were obtained immediately postextraction from clinics at the Birmingham Dental Hospital following informed patient consent. Carious teeth exhibited disease ranging from enamel involvement only to deep dentinal lesions. Extracted teeth were immediately submerged in the RNA stabilising solution, RNA Later (Sigma, UK). Teeth were subsequently longitudinally sliced, using a segmented, diamond-edged rotary saw (TAAB Laboratories, UK) cooled with PBS, and the pulpal tissue carefully removed intact using a sterile dental probe and forceps. This technique has previously been shown to provide core pulpal tissue with the odontoblast layer being left intact on the dentine surface [20]. 2.2. Purification and activation of immune system cells Following informed patient consent, venous blood from medically healthy donors, who had no history of inflammatory disease, was collected under sterile conditions in vacutainer tubes containing either heparin sulphate (neutrophils) or sodium citrate (monocytes). Neutrophils were subsequently prepared from 7 mL venous blood using a previously described single gradient (d=1.079) Percoll density centrifugation technique [18]. In brief, blood was layered onto the 8-mL Percoll gradient (Amersham, UK) and centrifuged for 8 min at 150g and then for a further 10 min at 1200g to separate blood platelets and mononuclear cells. Neutrophils were subsequently harvested and contaminating red blood cells removed by addition of 30 mL ice-cold lysis buffer (0.83% NH4Cl, solution). Cells were then subjected to two 2-mL washes, firstly in lysis buffer then phosphate-buffered saline (PBS) for 6 min at 360 g, and finally resuspended in 1 mL PBS. Prior to stimulation, neutrophil cell counts and viability were determined using the trypane blue (Sigma, UK) exclusion method. A proportion of isolated neutrophil cells were stimulated by exposing 5106 cells to 1 mL GPBSS (PBS with the addition of 10 mM glucose, 1 mM CaCl2, 1.5 mM MgCl2) supplemented with 250 ng of Escherichia coli Lipopolysaccharide (LPS) (Sigma, UK) for 3 h at 37 -C. As control, 5106 neutrophils were incubated for 3 h at 37 -C in unsupplemented GPBSS. Following incubation, cells were centrifuged at 1200g for 2 min, the supernatant removed and the pellet used for RNA isolation. Monocytes were isolated from 200 mL venous blood using Ficoll Hypaque (Amersham-Pharmacia Biotech, UK) gradient centrifugation. Extracted blood was diluted at a ratio of 1:1 with Hanks’ Balanced Salt Solution (HBSS), pelleted by centrifugation and rinsed twice more in HBSS. Cells were resuspended in RPMI-1640 medium prior to cell counting. To enrich for monocytes, cells were plated at a density of 4106 cells/mL and incubated at 37 -C for 1 h. Plated cells were washed twice in HBSS to remove nonadherent lymphocytes. Monocyte RNA was subsequently harvested from adherent cells. Macrophages were matured by culturing monocytes for 5 days in RPMI-1640 plus 10%
J.L. McLachlan et al. / Biochimica et Biophysica Acta 1741 (2005) 271 – 281
fetal calf serum with 50 ng/mL GM-CSF (R&D Systems, USA). To confirm GM-CSF macrophage maturation osteopontin gene expression levels were assayed in RNA derived from the monocytes and macrophages [18]. 2.3. RNA isolation Total RNA was extracted from human immune cells and pulpal tissues using the RNeasy mini kit (Qiagen, UK) and eluted in a volume of 30 Al of sterile water as recommended by the manufacturer. Prior to RNA extraction, tissue was homogenized using an Ultra-Turrax T8 tissue disrupter (Fisher Scientific, UK). RNA concentrations were determined from absorbance values at a wavelength of 260 nm using a BioPhotometer (Eppendorf, UK). RNA samples used in this study had 260/280 nm ratios equal to or greater than 1.8 and RNA integrity was additionally verified by visual inspection of samples on 1% non-denaturing agarose gels stained with SYBR Gold (Molecular Probes, UK). 2.4. Microarray target preparation and hybridisation Gene expression using pooled RNA from 12 clinically healthy and 11 carious teeth, which exhibited only deep dentinal lesions/pulp exposure, were analysed using human Affymetrix HG_U133A oligonucleotide arrays, as described at http://www.affymetrix.com/products/arrays/specific/ hgu133.affx. The average amount of RNA isolated from pulps from clinically healthy teeth was 0.62 T 0.08 Ag and from carious teeth was 1.15 T 0.27 Ag. Total RNA from each sample was used to prepare biotinylated target RNA, according to the manufacturer’s instructions (http://www. affymetrix.com/support/technical/manual/expression _ manual.affx). Briefly, 5 Ag of DNase digested total RNA was used to generate double-stranded cDNA using SuperScript reagents (Life Technologies) and a T7-linked oligo(dT) primer. cRNAs were synthesized using the Enzo Bioarray High Yield RNA transcript labeling kit (Affymetrix, UK) and resulting biotinylated labeled cRNA was subsequently fragmented into 35- to 200-bp lengths using the Fragmentation buffer (Affymetrix, UK). As recommended by the manufacturers, RNA, cDNA and cRNA quality and size distribution were visually confirmed by agarose gel electrophoresis. Spike controls B2, bio-B, bio-C, bio-D and Cre-x were added to the hybridisation cocktail before overnight hybridisation at 45 -C for 16 h. Arrays were stained and washed on the Fluidics Station 400 (Affymetrix, UK) using the EukGE-WS2 protocol (dual staining) before being scanned twice on the GeneChip Scanner 3000 at an excitation wavelength of 488 nm. cRNA was initially hybridised to Test3 GeneChips to confirm sample integrity prior to hybridisation to U133A microarrays. Microarray analysis complied with MIAME standards (http:// www.mged.org/Workgroups/MIAME/miame.html) and hybridisation results have been submitted to the Gene Expression Omnibus (GEO) (see below).
273
2.5. Data analysis To extract significant changes (2-fold or higher) in gene expression between healthy and carious samples, the comparison data files generated by the Affymetrix Microarray Suite 4.0 (MAS 4.0) software were initially sorted by deleting transcripts (i) displaying no change in expression, (ii) having a detection P value of equal to or higher than 0.045 and (iii) having a signal log ratio of between 1 and +1. Based on previous experiences, Affymetrix recommendations and published literature [21] transcripts exhibiting over 2-fold change were subsequently classed as being differentially expressed between the healthy and carious pulpal samples. The data are available at the Gene Expression Omnnibus (GEO) website under the series number GSE1629 (http://www.ncbi.nlm.nih.gov/geo/). To classify genes according to standardized GeneOntology (GO), vocabulary for the categories of biological process and molecular function information from the NetAffx software (Affymetrix, UK) was utilised or entire datasets were imported into the Onto-Express software (http://vortex.cs.wayne.edu/projects.htm) [22]. The default binomial probability distribution and Bonferroni multiple experiment correction settings of the software were used for data analysis. Probability values were derived from the numbers of genes corresponding to each GO category among the differentially expressed datasets compared with the number of genes expected for each GO category, based on their representation on the Affymetrix U133A array [22]. 2.6. cDNA synthesis for PCR analysis For single-stranded cDNA synthesis for use in PCR analysis, 1 Ag of DNase digested total RNA extracted was used as templates for reverse transcription (Omniscript kit, Qiagen, UK). Either oligo-dT, for semi-quantitative RT-PCR analysis, or random hexamers, for real-time RT-PCR analysis, was used to prime the reverse transcription reactions, performed as recommended by the manufacturer (Qiagen, UK). Synthesized cDNA concentrations were determined from absorbance values at a wavelength of 260 nm using a BioPhotometer (Eppendorf, UK). 2.7. Semi-quantitative reverse transcriptase polymerase chain reaction (Sq-RT-PCR) analysis Sq-RT-PCR analysis was performed used the pooled RNA samples previously described for microarray analysis. Differential expression analysis was performed for the human genes SLIT2, Period-2 (PER2), Period-3 (PER3), osteoadherin, Glypican 3, Midkine, activin receptor interacting protein-1 (AIP1), osteoadherin and growth hormone receptor (GHR), Interleukin-11 (IL-11), Bone sialoprotein (BSP), matrix Gla protein (MGP), endothelial cell growth factor-1 (ECGF1), inhibin hA and orosomucoid-1 (ORM1)
274
J.L. McLachlan et al. / Biochimica et Biophysica Acta 1741 (2005) 271 – 281
with the housekeeping gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH) used as control. The primer sequences and cycling conditions used are shown in Supplementary Table S1. Fifty nanograms of cDNA were used to seed 50 AL PCRs, which were subjected to between 32 and 42 cycles. A typical amplification cycle of 95 -C for 20 s, 60/61 -C for 20 s and 72 -C for 20 s was performed using a Mastercycler thermal cycler (Eppendorf, UK). Following a designated number of cycles, 8 Al of the PCR mix was removed and the product separated and visualized on a 1.5% agarose gel containing 0.5 Ag/mL ethidium bromide. Primers were designed from the Affymetrix probe target ID sequences using the Primer3 program (http://frodo.wi.mit.edu/cgi-bin/ primer3/primer3_www.cgi). 2.8. Real-time RT-PCR analysis Real-time RT-PCR analysis was performed using RNAs obtained from pulps from clinically healthy and carious teeth exhibiting a range of carious lesions from enamel involvement only to pulpal exposure. Expression levels of the human genes Adrenomedullin (ADM) and Dentine Matrix Phosphoprotein-1 (DMP-1) were analysed. TaqMan probes and primers were designed to span intron– exon boundaries using Primer Express 1.5 software (Applied Biosystems, UK) (Table S1). HPLC purified primers and probes were labeled with the fluorescent reporter dye 6-FAM and the quencher dye 6-TAMRA, at the 5V and 3V ends, respectively (Thermo Hybaid, UK; Eurogentec, UK). Quantitative PCR was performed using the ABI Prism 7700 Sequence Detection System (Applied Biosystems, UK). Each individual 25 Al PCR was seeded with 25 ng of cDNA and contained final concentrations of 1 Universal TaqMan Master Mix (Applied Biosystems, UK), 0.9 AM of forward and reverse primer, 0.15 AM of 6-FAM probe, 1 18s rRNA eukaryotic endogenous control (5Vlabeled with 6VIC) (Applied Biosystems, UK). PCRs were performed in 96-well plates covered with MicroAmp Optical Caps (Applied Biosystems, UK). Amplification conditions were 50 -C for 2 min, 95 -C for 10 min, then 44 cycles of 95 -C for 15 s and 60 -C for 1 min. All PCR assays were performed in triplicate. ABI 7700 Prism SDS 2.1 software (Applied Biosystems, UK) recorded fluorescence intensity of reporter and quencher dyes and results were plotted versus time, represented by cycle number. To obtain a precise quantification of initial target, amplification plots were examined during the early log phase of product accumulation above background [the threshold cycle (Ct) number]. Ct values were subsequently used to determine DCt values (DCt = Ct of the gene minus Ct of the 18S rRNA control) and differences in Ct values were used to quantify the relative amount of PCR product, expressed as fold change by applying the equation 2 DCt.
3. Results 3.1. Microarray analysis From 5 Ag total RNA starting amounts, approximately 35 Ag of fragmented biotin labeled healthy and carious cRNA was generated. The high quality of each of the target samples was confirmed using Affymetrix Test3 GeneChips (see Table 1). HG_U133A microarrays were subsequently hybridised in duplicate with 15 Ag of each of the cRNA samples. Analysis of control parameter data (as described for Test3 GeneChip analysis) confirmed hybridisation success (data not shown) and scaling factors were within Affymetrix recommended guidelines, values of 0.732 and 1.116 for duplicate healthy samples, and 1.006 and 1.222 for carious samples. Of the 22,283 probes sets represented on the U133A GeneChips, healthy pulp replicate hybridisations detected 49.4% and 44.1% as present with 1.8% and 2.0% as being marginally present, respectively. For the carious replicate hybridisations, 41.4% and 40.4% were detected as being present with 2.1% and 2.2% detected as being marginally present, respectively. Pairwise analysis of duplicate hybridisation data indicated that of the 9766 genes detected as being present by both targets, 445 genes, 4.56% of detected genes, were 2fold or greater differentially expressed between healthy and carious pulpal samples. Of this dataset, 85 and 360 were more abundant in healthy and carious tissue, respectively. The majority of the differentially expressed genes (3.65%) were 2- to 10-fold differentially expressed with only 0.91% exhibiting greater than 10-fold differentially expression. Fig. 1 provides graphical representation of the differential expression distribution of the genes detected. Please refer to the Supplementary Material for this article for a compre-
Table 1 GeneChip Test3 cRNA hybridisation quality control Control measurement
Average background Noise (Raw Q) Signal value of eukaryotic hybridisation controls: bio-B bio-C bio-D cre-x Housekeeping gene signal value ratios (3Vto 5V): h-actin GAPDH
Observed value Healthy
Carious
90.52
83.65
2.75
2.51
61.24 159.62 671.13 2131.11
65.8 144.02 643.58 2047.12
1.67 1.29
1.55 1.41
Affymetrix recommendation Range 20 – 100 units Values within 10 – 15% Values within 10 – 15%
Probe detection and comparable signal intensities
Ratio less than 3
5 Ag aliquots of fragmented healthy and carious cRNA were hybridised to Test3 GeneChips. B2 oligo hybridisations enabled accurate grid alignment for GeneChip Test3 features.
J.L. McLachlan et al. / Biochimica et Biophysica Acta 1741 (2005) 271 – 281
275
correlating with the presence of infiltrating immune cells recruited in response to bacteria infection. Supportive of these data were significant molecular function GOs, which included chemokine, MHC class II receptor and manganese superoxide dismutase activities also previously shown to be associated with an immune response to bacteria infection. In contrast, the GO analysis of genes more abundantly expressed in healthy pulpal tissue identified a variety of processes and functions potentially lost or down-regulated due to disease progression. Of particular note were the biosynthetic pathways, which may be down-modulated due to the ensuing inflammation. 3.3. Individual gene expression analysis
Fig. 1. (a) Scatter plot of average healthy versus carious pulp sample signal intensity values for individual microarray features, plotted on a double logarithmic scale. Areas highlighted indicate: A—No differential expression; B—2- to 10-fold more abundantly expressed; C—>10-fold more abundantly expressed. (b) Histogram presenting frequency of fold change of the 445 2-fold or greater differentially expressed genes.
hensive list of genes increased (Supplementary Table S2) or decreased (Supplementary Table S3) in pulpal tissue due to carious disease. 3.2. Gene ontology analysis To translate the data into a more meaningful biological context and to characterize more thoroughly sets of functionally related genes, the differentially expressed datasets were subsequently organized into GO groupings using the publicly available software Onto-Express (OE). The OE software subsequently automatically translates the nonredundant datasets using standardized GO terms for biological process and molecular function providing statistical significance results for each category. GO groupings were determined by comparing the number of genes expected for each GO category based on their representation on the Affymetrix HG_U133A array. Significant differences from the expected were calculated with a two-sided binomial distribution. Figs. 2 and 3 show all GO classes with a Bonferroni-corrected significance of P < 0.05 for the two datasets analysed. Notably, the biological process GO groupings demonstrating the most significant representation in our set of genes more abundant in carious pulpal tissue, map to the inflammatory and immune response, although a variety of biological processes associated with cellular and tissue defence mechanisms were also identified, potentially
To validate the microarray data, semi-quantitative and real-time RT-PCR was performed on selected differentially expressed genes. Specific genes were chosen, based on criteria that included (i) their expression not previously having been described in mature pulpal tissue or during carious disease, (ii) their representation of a range of ontological classes/functions and (iii) their fold change and signal intensities spanning a range of values. Data confirmed that the 14 genes analysed by semi-quantitative RT-PCR (Fig. 4) were differentially expressed as predicted by the microarray data. No attempts were made to statistically correlate fold change with differential expression detected by PCR although a visible correlation was noted (Fig. 4). Real-time PCR analysis was also utilised to analyse expression levels of two genes, ADM1 and DMP1, in RNA extracted from individual human healthy and carious pulpal samples which exhibited a range of stages of disease severity from enamel involvement only through to pulpal exposure. Whilst overall expression level changes identified were in agreement with the microarray data, these analyses highlighted transcript level heterogeneity between samples, not only in carious, but also in healthy pulpal specimens (Fig. 5). No correlations were detected using regression analysis [18] between ADM and DMP1 expression levels or between expression levels and disease severity in the sample set used (data not shown). Real-time PCR analysis was also employed to further characterize the expression of ADM1 and DMP1 in a panel of immune system cells previously implicated in carious disease progression [10]. Expression profiling data indicated that the levels of ADM during carious disease were potentially attributable to infiltrating immune system cells, in particular activated neutrophils, whilst DMP1 was expressed predominantly by cells native to healthy pulps (Fig. 6).
4. Discussion Vital pulp therapy aims to treat reversible pulpal injury whilst maintaining cell viability and function. Currently, two
276
J.L. McLachlan et al. / Biochimica et Biophysica Acta 1741 (2005) 271 – 281
Fig. 2. The significant (Bonferroni-corrected-P < 0.05) biological processes and molecular functions identified by the 360 genes more abundantly expressed in carious pulpal tissue. P value and the number of genes corresponding to each GO functional class are shown.
therapeutic approaches are frequently used, (i) indirect pulp capping in cases of deep dentinal lesions and (ii) direct pulp capping/pulpotomy in cases of pulp exposure. Successful treatment outcomes are dependent on the type and location of the injury, age of the tooth, the pulp capping material used and the integrity of the cavity restoration. In the future, it is hoped that biomimetic approaches will be developed based on a thorough understanding of the molecular and cellular events that occur during dental injury and disease. Such therapies aim to maintain the pulp’s vitality and stimulate its innate reparative and regenerative mechanisms. Previous molecular and biochemical studies characterizing caries have, however, so far been limited. In this current study, we have addressed this issue and now describe large-
scale gene expression analyses in human healthy and carious pulpal tissue. To date, most reported microarray studies have focused on cultured cells exposed to factors important for disease progression. Subsequently, the disease relevance of the gene expression changes identified required validation in vivo. For our approach, we have directly used pooled clinical samples from in excess of 10 healthy and 10 carious pulpal teeth. Pooling of RNA from such relatively high numbers of specimens was required to generate sufficient amounts of RNA to enable microarray analysis, providing an Faverage_ response of core pulpal tissues under deep carious lesions. Notably, the use of microarrays and complex tissues poses challenges as the pulp is composed of heterogeneous and
J.L. McLachlan et al. / Biochimica et Biophysica Acta 1741 (2005) 271 – 281
277
Fig. 3. The significant (Bonferroni-corrected-P < 0.05) biological processes and molecular function identified by the 85 genes more abundantly expressed in healthy pulpal tissue. P value and the number of genes corresponding to each GO functional class are shown.
changing cell populations with interactions between immune and non-immune cells being pivotal to disease pathogenesis [10]. Indeed, analysis of the expression of CD
antigens in carious disease (Supplementary Table S2) highlights this complexity. Gene expression changes identified using this approach will therefore represent an average
Fig. 4. (a) (i) Eight and (ii) six genes identified as more abundant in healthy (H) and carious (C) pulpal samples, respectively, were analysed by semiquantitative RT-PCR analysis. (b) Expression levels were normalised to GAPDH housekeeping gene levels. Superscript a or b indicates the GAPDH gel image used for normalisation for each differentially expressed gene analysed. Fold-change identified from microarray experiments are shown in parentheses.
278
J.L. McLachlan et al. / Biochimica et Biophysica Acta 1741 (2005) 271 – 281
Fig. 5. Real-time PCR analysis for ADM and DMP1 in (a) 5 healthy (nos. 1 – 5) and 5 carious (nos. 6 – 10) pulpal samples. Fold change was obtained by comparison with the sample exhibiting the lowest level of expression which was assigned a value of 1. (b) Average expression levels for carious and healthy samples from (a) is shown. Vertical bars indicate T one standard error of the mean. Asterisks indicate significant difference between carious and healthy gene expression levels ( P < 0.05).
of many different cell types and transcripts from cells native to the pulp may appear decreased relative to the total mRNA pool as a result of infiltrating immune cell populations. Since genes contained on the arrays were not selected specifically for the analysis of gene expression in caries, they do not reflect biases inherent to its pathogenesis. This approach, therefore, enhances the likelihood of identifying important expression changes that are not currently part of our understanding of caries. Pertinently, whilst results obtained (Supplementary Tables S2 and S3) concur with differential expression reported elsewhere, including increases in expression in caries of IL-1h, IL-6, IL-8, S100A8 and S100A9 [18] and decreases of DSPP [20], our data now identify a significant number of expression changes associated with chronic inflammation and tissue remodelling that have not previously been described. The translation of such high-throughput gene expression data into meaningful biological information and identification of associated pathways/networks is a challenge to biologists and bio-informaticians in this post-genomic era. As
manual literature searching is inefficient, extremely time consuming and does not provide a common language for researchers, we chose to use the Onto-Express software, which utilises the standardized GO vocabulary [23]. The GO groupings identified for biological process and molecular function from the genes more abundant in diseased pulps (Fig. 2) identify many groupings previously linked with an immune/inflammatory response associated with a bacterial infection. In conjunction with the TEST3 GeneChip hybridisation results (Table 1), which indicated the high quality of the labeled cRNA, this analysis provides additional confidence in the list of differentially expressed genes presented in this study. In contrast, the GO groupings identified by genes more abundant in healthy pulps identify potentially homeostatic biological processes and molecular functions down modulated during the disease process (Fig. 3), many of which warrant further investigation. Our confirmatory PCR analyses (Figs. 4 and 5) support our microarray findings and highlight the differential expression in pulpal tissue of several genes previously not
Fig. 6. Real-time PCR analysis of ADM and DMP1 gene expression in immune system cells. Mo=Monocytes; Mf=Macrophage; PBL=Lymphocytes; N=Neutrophil; N+=E. coli LPS stimulated neutrophils; H=Pooled healthy pulp; C=Pooled carious pulpal tissue.
J.L. McLachlan et al. / Biochimica et Biophysica Acta 1741 (2005) 271 – 281
associated with carious disease. Whilst genes down-regulated due to caries may correspond to pulpal transcripts representing a decreased proportion of the total mRNA pool, these analyses nevertheless provides novel data on their dental tissue expression. In particular, the presence of the circadian clock genes, PER2 and PER3, has not previously been reported. PER2 has, however, been described in murine bone marrow [24] and data indicated its expression to be potentially lineage or differentiation stage-dependent. As circadian rhythms have also been associated with the deposition of incremental dentine [25,26], the role of these genes in dental tissue repair warrants further investigation. Differential expression of members of the non-collagenous protein family termed SIBLING (Small Integrin-Binding Ligand, N-linked Glycoprotein), namely DMP1 and BSP, were also confirmed by PCR analyses (Figs. 4 and 5). This family also includes osteopontin (OPN), DSPP and matrix extracellular phosphoglycoprotein (MEPE), which play key biological roles in mineralization of bone and dentin ECM [27]. Data presented here and previously [18] now indicate, that whilst both DMP1 and DSPP (Supplementary Table S3) are down-regulated by core pulpal tissue during caries, the converse is true for the transcript levels of BSP (Fig. 4) and OPN (Supplementary Table S2). Interestingly, work by Decup et al. [28] has also shown that BSP is capable of stimulating pulpal cells to differentiation into odontoblasts which secrete an organized ECM. Hence, the increased expression of BSP may be indicative of potential repair processes ongoing within diseased pulp. IL-11 (Fig. 4) is a pleiotropic cytokine produced by a variety of stromal cells [29]. Recent work [30,31] has shown that IL-11 acts synergistically with BMP-2 to induce osteoblast differentiation and accelerate bone formation. As previous reports have shown BMP-2 capable of stimulating pulpal cells to differentiate into odontoblast [32 –35], our data now indicate that the role of IL-11 in this process should also be considered. Whilst the expression of SLIT2, GHR, Inhibin hA, midkine, osteoadherin and matrix Gla protein has been described during tooth development and mineralization [36 –41], our analyses now provide evidence of their differential expression in mature healthy and carious pulpal tissue (Fig. 4). Notably, another potentially interesting finding of our microarray analyses is the identification of the relatively dramatic downregulation, 29-fold, of zeta-crystallin in pulps from carious teeth (Supplementary Table S3). Whilst this represents novel data, other family members have been shown to play a role in mineralized tissue function and are potentially important in the pulp’s response to mechanical and thermal stresses [42,43]. Perhaps, our most significant finding is the increased ADM expression associated with caries. ADM is a recently discovered pluripotent peptide hormone important in a range of physiological and pathological processes. Many tissues of the body express it and it has a remarkable range of effects mediated by paracrine/autocrine and endocrine
279
mechanisms [44]. Active ADM is derived by proteolytic cleavage of pro(p)ADM which subsequently acts through two receptor subtypes, termed L1-receptor (L1-R) and calcitonin receptor-like receptor (CRLR). The CRLR functions as either a calcitonin gene-related peptide (CGRP) receptor or a selective ADM receptor depending on which member of a family of chaperones, named receptor-activitymodifying proteins (RAMPs), is expressed. RAMP1 generates CGRP receptors, whilst RAMP2 and RAMP3 produce L1-Rs [45]. ADM is important in embryonic tissue development, particularly in the tooth, where strong mesenchymal – epithelial interactions occur [45]. ADM can act as a mitogen, apoptosis survival factor, angiogenic stimulant and regulator of immune response [46,47] and is therefore a powerful candidate mediator of various cellular responses in the pulp. In addition, ADM is a potent osteoblast stimulator, in terms of proliferation and osteogenic activity, both in vitro and in vivo, and as a result has been proposed as an attractive candidate treatment for bone diseases and defects [48]. Our analyses now highlight its differential expression within the pulp and our microarray data also indicate that RAMP1, RAMP3, CRLR and CGRPreceptor component protein expression are all expressed in healthy pulps although their expressions were unchanged as a result of disease (data not shown). These findings indicate that further characterisation of the role of ADM, its receptors and RAMPs is warranted as it may provide a novel target for dental inflammation and regeneration control. In addition to the confirmation of differential expression of the 16 genes reported here (Figs. 4 and 5), we have also verified differential expression of 6 genes up-regulated and 4 genes down-regulated during caries which represent transcripts of unknown function (unpublished data). Work is currently underway to characterize the role of these novel genes in dental health and disease. A better understanding of the roles of the differentially expressed genes identified by our microarray analyses during the caries disease process will require identification of their cellular origin using techniques such as in situ hybridisation and immunohistochemistry in mature and developing dental tissue. Such analyses will determine whether expression of these molecules is by cells native to the pulp or due to infiltrating immune cells. Notably, comparison of genes up-regulated in pulpal tissue during caries with genes increased in neutrophils stimulated with the cariogenic pathogen, Fusobacterium nucleatum [49], has identified nine genes common to these two datasets (unpublished data). These analyses indicate that the metallothioneins 1L, 1X and 2A, the GRO1 oncogene, ORM1, RNase A k6, the MAFF transcription factor, HLA-DQB1 and superoxide dismutase 2, are potentially expressed by the infiltrating neutrophil population during carious disease. Technological improvements will soon facilitate interrogation of whole genome microarrays with RNA derived from single teeth. Thus, high-throughput gene expression
280
J.L. McLachlan et al. / Biochimica et Biophysica Acta 1741 (2005) 271 – 281
analyses, taking into account important disease factors, such as age of carious lesion, remaining dentine thickness, age of patient, disease severity and infecting microbial population, will enable a more comprehensive understanding of the molecular circuitry involved in carious disease and facilitate future targeted treatment regimes for early and end stage disease. Our study provides novel data regarding the molecular characterisation of the complex interplay between reparative and inflammatory events that occur under lesions from one of the most prevalent chronic diseases in the world. In addition, potential novel targets are identified for further research that may enable development of new treatment modalities targeted at controlling the inflammatory process and maintaining tooth vitality, whilst permitting essential defence and repair reactions to occur. Such approaches have the potential for immense impact on healthcare globally.
[8]
[9]
[10]
[11]
[12]
[13] [14]
[15]
Acknowledgements JLM is supported by a University of Birmingham, Scientific Projects Committee grant (EBX 1193). We would like to thank Miss Sally Kerr (Oral Surgery) for providing the teeth for this investigation. We thank Dr T. Stankovic for provision of cRNA labeling reagents. This work was supported in part by an Oral and Dental Research Trust Grant.
[16]
Appendix A. Supplementary data
[19]
Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.bbadis.2005. 03.007.
References [1] K.J. Anusavice, Dental caries: risk assessment and treatment solutions for an elderly population, Compend. Contin. Educ. Dent. 23 (2002) 12 – 20. [2] P.E. Petersen, The World Oral Health Report 2003 Report. Continuous improvement of oral health in the 21st century—The approach of the WHO Global Oral Health Programme [Online]. Oral Health Programme, Noncommunicable Disease Prevention and Health Promotion, World Health Organization, Geneva, Switzerland. http://www.who. int/oral_health/publications/report03/en/ [2003]. [3] Y. Abiko, Passive immunization against dental caries and periodontal disease: development of recombinant and human monoclonal antibodies, Crit. Rev. Oral Biol. Med. 11 (2000) 140 – 158. [4] F.J. Burke, S.W. Cheung, I.A. Mjor, N.H. Wilson, Reasons for the placement and replacement of restorations in vocational training practices, Prim. Dent. Care 6 (1999) 17 – 20. [5] A.J. Smith, Pulpal responses to caries and dental repair, Caries Res. 36 (2002) 223 – 232. [6] K.R. Baumgardner, M.A. Sulfaro, The anti-inflammatory effects of human recombinant copper-zinc superoxide dismutase on pulp inflammation, J. Endod. 27 (2001) 190 – 195. [7] R.B. Rutherford, K. Gu, Treatment of inflamed ferret dental pulps with
[17]
[18]
[20]
[21]
[22]
[23] [24]
[25]
[26]
[27]
[28]
[29] [30]
recombinant bone morphogenetic protein-7, Eur. J. Oral. Sci. 108 (2000) 202 – 206. A. Fouad, Molecular mediators of pulpal inflammation, in: K.M. Hargreaves, H.E. Goodis (Eds.), Seltzer and Bender’s Dental Pulp, Quintesscence Publ. Co, Chicago, 2002, pp. 247 – 279. K.J. Heyeraas, O.B. Sveen, I.A. Mjor, Pulp – dentin biology in restorative dentistry. Part 3: pulpal inflammation and its sequelae, Quintessence 32 (2001) 611 – 625. T. Izumi, I. Kobayashi, K. Okamura, H. Sakai, Immunohistochemical study on the immunocompetent cells of the pulp in human non-carious and carious teeth, Arch. Oral Biol. 40 (1995) 609 – 614. C.R. Rauschenberger, J.C. Bailey, C.J. Cootauco, Detection of human IL-2 in normal and inflamed dental pulps, J. Endod. 23 (1997) 366 – 370. P. Stashenko, Interrelationship of dental pulp and apical periodontitis, in: K.M. Hargreaves, H.E. Goodis (Eds.), Seltzer and Bender’s Dental Pulp, Quintesscence Publ Co, Chicago, 2002, pp. 389 – 410. H. Trowbridge, Pathogenesis of pulpitis resulting from dental caries, J. Endod. 7 (1981) 52 – 60. R. D’Souza, L.R. Brown, J.R. Newland, B.M. Levy, L.B. Lachman, Detection and characterization of interleukin-1 in human dental pulps, Arch. Oral Biol. 34 (1989) 307 – 313. G.T. Huang, A.P. Potente, J.W. Kim, N. Chugal, X. Zhang, Increased interleukin-8 expression in inflamed human dental pulps, Oral Surg. Oral Med. Oral Pathol. Oral Radiol. Endo. 88 (1999) 214 – 220. Y. Jiang, T.R. Russell, H. Schilder, D.T. Graves, Endodontic pathogens stimulate monocyte chemoattractant protein-1 and interleukin-8 in mononuclear cells, J. Endod. 24 (1998) 86 – 90. K. Matsushima, E. Ohbayashi, H. Takeuchi, S. Hosoya, Y. Abiko, M. Yamazaki, Stimulation of interleukin-6 production in human dental pulp cells by peptidoglycans from lactobacillus case, J. Endod. 24 (1998) 252 – 255. J.L. McLachlan, A.J. Sloan, A.J. Smith, G. Landini, P.R. Cooper, S100 and cytokine expression in caries, Infect. Immun. 72 (2004) 4102 – 4108. M. Zehnder, N. Delaleu, Y. Du, M. Bickel, Cytokine gene expression—Part of host defence in pulpitis, Cytokine 22 (2003) 84 – 88. J.L. McLachlan, A.J. Smith, A.J. Sloan, P.R. Cooper, Gene expression analysis of genes of the dentine pulp complex in healthy and carious teeth, Arch. Oral Biol. 48 (2003) 273 – 283. L. Wodicka, H. Dong, M. Mittmann, M.H. Ho, D.J. Lockhart, Genome-wide expression monitoring in Saccharomyces cerevisiae, Nat. Biotechnol. 15 (1997) 1359 – 1367. S. Draghici, P. Khatri, R.P. Martins, G.C. Ostermeier, S.A. Krawetz, Global functional profiling of gene expression, Genomics 81 (2003) 98 – 104. D. Murphy, Gene expression studies using microarrays: principles, problems, and prospects, Adv. Physiol. Educ. 26 (2002) 256 – 270. Y.G. Chen, A. Mantalaris, P. Bourne, P. Keng, J.H. Wu, Expression of mPer1 and mPer2, two mammalian clock genes, in murine bone marrow, Biochem. Biophys. Res. Commun. 276 (2000) 724 – 728. M.C. Dean, A.E. Scandrett, The relation between long-period incremental markings in dentine and daily cross-striations in enamel in human teeth, Arch. Oral Biol. 41 (1996) 233 – 241. M. Ohtsuka-Isoya, H. Hayashi, H. Shinoda, Effect of suprachiasmatic nucleus lesion on circadian dentin increment in rats, Am. J. Physiol., Regul. Integr. Comp. Physiol. 280 (2001) R1364 – R1370. C. Qin, O. Baba, W.T. Butler, Post-translational modifications of sibling proteins and their roles in osteogenesis and dentinogenesis, Crit. Rev. Oral Biol. Med. 15 (2004) 126 – 136. F. Decup, N. Six, B. Palmier, D. Buch, J.J. Lasfargues, E. Salih, M. Goldberg, Bone sialoprotein-induced reparative dentinogenesis in the pulp of rat’s molar, Clin. Oral Investig. 4 (2000) 110 – 119. S.X. Leng, J.A. Elias, Interleukin-11, Int. J. Biochem. Cell Biol. 29 (1997) 1059 – 1062. K. Suga, M. Saitoh, S. Kokubo, S. Fukushima, S. Kaku, S. Yasuda, K. Miyata, Interleukin-11 acts synergistically with bone morphoge-
J.L. McLachlan et al. / Biochimica et Biophysica Acta 1741 (2005) 271 – 281
[31]
[32] [33]
[34]
[35]
[36]
[37]
[38]
[39]
netic protein-2 to accelerate bone formation in a rat ectopic model, J. Interferon Cytokine Res. 23 (2003) 203 – 207. K. Suga, M. Saitoh, S. Kokubo, K. Nozaki, S. Fukushima, S. Yasuda, M. Sasamata, K. Miyata, Synergism between interleukin-11 and bone morphogenetic protein-2 in the healing of segmental bone defects in a rabbit model, J. Interferon Cytokine Res. 24 (2004) 343 – 349. I. About, T.A. Mitsiadis, Molecular aspects of tooth pathogenesis and repair: in vivo and in vitro models, Adv. Dent. Res. 15 (2001) 59 – 62. C. Begue-Kirn, A.J. Smith, J.V. Ruch, J.M. Wozney, A. Purchio, D. Hartmann, H. Lesot, Effects of dentin proteins, transforming growth factor ß1 (TGFß1) and bone morphogenetic protein 2 (BMP2) on the differentiation of odontoblasts in vitro, Int. J. Dev. Biol. 36 (1992) 491 – 503. C. Begue-Kirn, A.J. Smith, M. Loriot, C. Kupferle, J.V. Ruch, H. Lesot, Comparative analysis of TGFßs, BMPs, IGF, msxs, fibronectin, osteonectin and bone sialoprotein gene expressions during normal and in vitro induced odontoblast differentiation, Int. J. Dev. Biol. 38 (1994) 405 – 420. T. Saito, M. Ogawa, Y. Hata, K. Bessho, Acceleration effect of human recombinant bone morphogenetic protein-2 on differentiation of human pulp cells into odontoblasts, J. Endod. 30 (2004) 205 – 208. C.W. Brown, D.E. Houston-Hawkins, T.K. Woodruff, M.M. Matzuk, Insertion of Inhbb into the Inhba locus rescues the Inhba-null phenotype and reveals new activin functions, Nat. Genet. 25 (2000) 453 – 457. R. Buchaille, M.L. Couble, H. Magloire, F. Bleicher, Expression of the small leucine-rich proteoglycan osteoadherin/osteomodulin in human dental pulp and developing rat teeth, Bone 27 (2000) 265 – 270. S. Loes, K. Luukko, I.H. Kvinnsland, P. Kettunen, Slit1 is specifically expressed in the primary and secondary enamel knots during molar tooth cusp formation, Mech. Dev. 107 (2001) 155 – 157. T.A. Mitsiadis, T. Muramatsu, H. Muramatsu, I. Thesleff, Midkine (MK), a heparin-binding growth/differentiation factor, is regulated by retinoic acid and epithelial – mesenchymal interactions in the deve-
[40]
[41] [42]
[43]
[44] [45]
[46]
[47]
[48] [49]
281
loping mouse tooth, and affects cell proliferation and morphogenesis, J. Cell Biol. 129 (1995) 267 – 281. A.L. Symons, A.C. Henry, S. Chang, T.J. Daley, D.J. Harbrow, B.K. Joseph, The effect of glucocorticosteroid treatment on dentine formation in the Lewis rat, a histological study, Growth Factors 18 (2000) 157 – 167. P.G. Robey, Vertebrate mineralized matrix proteins: structure and function, Connect. Tissue Res. 35 (1996) 131 – 136. K. Behnam, S.S. Murray, J.P. Whitelegge, E.J. Brochmann, Identification of the molecular chaperone alpha B-crystallin in demineralized bone powder and osteoblast-like cells, J. Orthop. Res. 20 (2002) 1190 – 1196. T. Thyagarajan, A.B. Kulkarni, Transforming growth factor-beta1 negatively regulates crystallin expression in teeth, J. Bone Miner. Res. 17 (2002) 1710 – 1717. M. Jougasaki, J.C. Burnett, Adrenomedullin: potential in physiology and pathophysiology, Life Sci. 66 (2000) 855 – 872. L.M. Montuenga, A. Martinez, M.J. Miller, E.J. Unsworth, F. Cuttitta, Expression of adrenomedullin and its receptor during embryogenesis suggests autocrine or paracrine modes of action, Endocrinology 138 (1997) 440 – 451. D.M. Smith, H.A. Coppock, D.J. Withers, A.A. Owji, D.L. Hay, T.P. Choksi, P. Chakravarty, S. Legon, D.R. Poyner, Adrenomedullin: receptor and signal transduction, Biochem. Soc. Trans. 30 (2002) 432 – 437. T.H. Elsasser, S. Kahl, Adrenomedullin has multiple roles in disease stress: development and remission of the inflammatory response, Microsc. Res. Tech. 57 (2002) 120 – 129. J. Cornish, D. Naot, I.R. Reid, Adrenomedullin—A regulator of bone formation, Regul. Pept. 112 (2003) 79 – 86. F.E. Martin, M.A. Nadkarni, N.A. Jacques, N. Hunter, Quantitative microbiological study of human carious dentine by culture and realtime PCR: association of anaerobes with histopathological changes in chronic pulpitis, J. Clin. Microbiol. 40 (2002) 1698 – 1704.
Gene expression profiling of pulpal tissue reveals the molecular complexity of dental caries Julia L. McLachlana, Anthony J. Smitha, Iwona J. Bujalskab, Paul R. Coopera,* a Oral Biology, School of Dentistry, University of Birmingham, Birmingham, B4 6NN, UK Institute of Biomedical Research, Endocrinology and Metabolism, The Medical School, University of Birmingham, Edgbaston, Birmingham, B15 2TT, UK
b
Received 10 November 2004; received in revised form 9 February 2005; accepted 8 March 2005 Available online 8 April 2005
Abstract High-throughput characterisation of the molecular response of pulpal tissue under carious lesions may contribute to improved future diagnosis and treatment. To identify genes associated with this process, oligonucleotide microarrays containing ¨15,000 human sequences were screened using pooled total RNA isolated from pulpal tissue from both healthy and carious teeth. Data analysis identified 445 genes with 2-fold or greater difference in expression level, with 85 more abundant in health and 360 more abundant in disease. Subsequent gene ontological grouping identified a variety of processes and functions potentially activated or down-modulated during caries. Validation of microarray results was obtained by a combination of real-time and semi-quantitative PCR for selected genes, confirming down-regulation of Dentin Matrix Protein-1 (DMP-1), SLIT2, Period-2 (PER2), Period-3 (PER3), osteoadherin, Glypican-3, Midkine, activin receptor interacting protein-1 (AIP1), osteoadherin and growth hormone receptor (GHR), and up-regulation of Adrenomedullin (ADM), Interleukin-11 (IL-11), Bone sialoprotein (BSP), matrix Gla protein (MGP), endothelial cell growth factor-1 (ECGF1), inhibin hA and orosomucoid-1 (ORM1), in diseased pulp. Real-time PCR analyses of ADM and DMP-1 in a panel of healthy and carious pulpal tissue and also in immune system cells highlighted the heterogeneity of caries and indicated increased expression of ADM in neutrophils activated by bacterial products. In contrast, DMP-1 was predominantly expressed by cells native to healthy pulpal tissue. This study has greatly extended our molecular knowledge of dental tissue disease and identified involvement of genes previously unassociated with this process. D 2005 Elsevier B.V. All rights reserved. Keywords: Pulpal tissue; Diagnosis; Treatment
1. Introduction A recent report by the World Health Organization highlighted dental caries as a growing problem in both industrialized and, increasingly, in developing countries affecting 5 billion people worldwide. The disease impacts on all age groups causing pain, suffering, impairment of function and a reduced quality of life. Treatment costs are estimated to account for between 5 and 10% of total health costs in industrialized countries and are beyond the resources of many developing countries [1,2]. Caries is multifactorial involving * Corresponding author. Tel.: +44 121 237 2785; fax: +44 121 237 2882. E-mail address: [email protected] (P.R. Cooper). 0925-4439/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.bbadis.2005.03.007
the interplay between genetics, diet, life-style and invading microflora and preventative disease management regimes have been targeted at several of these primary risk factors [3]. So far, however, these approaches have been insufficient for disease eradication. Tooth restorative treatments following caries are still fairly empirical and involve excavation of the decayed tissue and subsequent replacement with relatively inert filling materials. This procedure, however, requires critical placement and failure rates of up to 60% have been reported which necessitates either re-treatment or results in tooth loss [4]. The dentine –pulp complex of the tooth has inherent natural regenerative properties and two mechanisms for this process are described. Following relatively mild tooth injury,
272
J.L. McLachlan et al. / Biochimica et Biophysica Acta 1741 (2005) 271 – 281
odontoblasts lining the pulp chamber, localised beneath the lesion, are stimulated to up-regulate their synthesis and secretion of a tertiary dentine extracellular matrix (ECM), which subsequently protects the underlying cells and maintains the tissues vitality. Injury of greater intensity, however, can lead to local odontoblast necrosis and replacement with a new generation of odontoblast-like cells, recruited from the pulp core, which subsequently secrete new mineralized dentine ECM beneath the lesion [5]. Clearly, the latter process is more complex involving progenitor cell recruitment and induction of differentiation prior to new matrix secretion. Recently, studies have shown that the inflammatory process, which ensues due to carious infection, can negatively impact on these repair mechanisms [6,7]. Whilst this immune response is well characterized histologically [8 – 13], our knowledge of its molecular mediation is partial with the involvement of only a limited number of cytokines and other related immune regulatory molecules having been described [14 –19]. Genome sequencing projects and advances in microarray technologies now provide us with reliable tools to obtain a comprehensive picture of the molecular events that occur during disease progression. Such an analysis of caries could provide information on the recruitment and involvement of specific cell types, cellular activation mechanisms, expression of genes as a result of interactions between resident and immune cell populations and on repair associated molecules. Specifically, known and novel molecules previously not associated with the disease may be identified. Clinically, the description and classification of molecular changes will enable (a) dissection of the molecular circuitry of the disease process, (b) identification and validation of potential targets for therapeutic intervention and (c) identification of potential biomarkers for improved diagnosis and validation. Previously, we have developed a temperature-sensitive dental tissue extraction procedure for the selective isolation of core pulpal tissue whilst leaving the peripheral odontoblast population attached to the dentine ECM [20]. The aim of this present study was therefore to characterize gene expression profiles within this core pulpal tissue from clinically healthy and severely carious human teeth. Along with confirmation of differential expression profiles for selected genes, the differentially expressed datasets were subsequently grouped into biological and molecular ontological classes. In addition, as the gene expression data represented an average response for the heterogeneous population of cells present within the pulp, further specific cell and tissue expression profiling was performed.
2. Experimental procedures 2.1. Extraction of dental tissue Carious and sound mature human premolar and molar teeth, extracted for orthodontic purposes from patients in the
age range of 20 to 30 years, were obtained immediately postextraction from clinics at the Birmingham Dental Hospital following informed patient consent. Carious teeth exhibited disease ranging from enamel involvement only to deep dentinal lesions. Extracted teeth were immediately submerged in the RNA stabilising solution, RNA Later (Sigma, UK). Teeth were subsequently longitudinally sliced, using a segmented, diamond-edged rotary saw (TAAB Laboratories, UK) cooled with PBS, and the pulpal tissue carefully removed intact using a sterile dental probe and forceps. This technique has previously been shown to provide core pulpal tissue with the odontoblast layer being left intact on the dentine surface [20]. 2.2. Purification and activation of immune system cells Following informed patient consent, venous blood from medically healthy donors, who had no history of inflammatory disease, was collected under sterile conditions in vacutainer tubes containing either heparin sulphate (neutrophils) or sodium citrate (monocytes). Neutrophils were subsequently prepared from 7 mL venous blood using a previously described single gradient (d=1.079) Percoll density centrifugation technique [18]. In brief, blood was layered onto the 8-mL Percoll gradient (Amersham, UK) and centrifuged for 8 min at 150g and then for a further 10 min at 1200g to separate blood platelets and mononuclear cells. Neutrophils were subsequently harvested and contaminating red blood cells removed by addition of 30 mL ice-cold lysis buffer (0.83% NH4Cl, solution). Cells were then subjected to two 2-mL washes, firstly in lysis buffer then phosphate-buffered saline (PBS) for 6 min at 360 g, and finally resuspended in 1 mL PBS. Prior to stimulation, neutrophil cell counts and viability were determined using the trypane blue (Sigma, UK) exclusion method. A proportion of isolated neutrophil cells were stimulated by exposing 5106 cells to 1 mL GPBSS (PBS with the addition of 10 mM glucose, 1 mM CaCl2, 1.5 mM MgCl2) supplemented with 250 ng of Escherichia coli Lipopolysaccharide (LPS) (Sigma, UK) for 3 h at 37 -C. As control, 5106 neutrophils were incubated for 3 h at 37 -C in unsupplemented GPBSS. Following incubation, cells were centrifuged at 1200g for 2 min, the supernatant removed and the pellet used for RNA isolation. Monocytes were isolated from 200 mL venous blood using Ficoll Hypaque (Amersham-Pharmacia Biotech, UK) gradient centrifugation. Extracted blood was diluted at a ratio of 1:1 with Hanks’ Balanced Salt Solution (HBSS), pelleted by centrifugation and rinsed twice more in HBSS. Cells were resuspended in RPMI-1640 medium prior to cell counting. To enrich for monocytes, cells were plated at a density of 4106 cells/mL and incubated at 37 -C for 1 h. Plated cells were washed twice in HBSS to remove nonadherent lymphocytes. Monocyte RNA was subsequently harvested from adherent cells. Macrophages were matured by culturing monocytes for 5 days in RPMI-1640 plus 10%
J.L. McLachlan et al. / Biochimica et Biophysica Acta 1741 (2005) 271 – 281
fetal calf serum with 50 ng/mL GM-CSF (R&D Systems, USA). To confirm GM-CSF macrophage maturation osteopontin gene expression levels were assayed in RNA derived from the monocytes and macrophages [18]. 2.3. RNA isolation Total RNA was extracted from human immune cells and pulpal tissues using the RNeasy mini kit (Qiagen, UK) and eluted in a volume of 30 Al of sterile water as recommended by the manufacturer. Prior to RNA extraction, tissue was homogenized using an Ultra-Turrax T8 tissue disrupter (Fisher Scientific, UK). RNA concentrations were determined from absorbance values at a wavelength of 260 nm using a BioPhotometer (Eppendorf, UK). RNA samples used in this study had 260/280 nm ratios equal to or greater than 1.8 and RNA integrity was additionally verified by visual inspection of samples on 1% non-denaturing agarose gels stained with SYBR Gold (Molecular Probes, UK). 2.4. Microarray target preparation and hybridisation Gene expression using pooled RNA from 12 clinically healthy and 11 carious teeth, which exhibited only deep dentinal lesions/pulp exposure, were analysed using human Affymetrix HG_U133A oligonucleotide arrays, as described at http://www.affymetrix.com/products/arrays/specific/ hgu133.affx. The average amount of RNA isolated from pulps from clinically healthy teeth was 0.62 T 0.08 Ag and from carious teeth was 1.15 T 0.27 Ag. Total RNA from each sample was used to prepare biotinylated target RNA, according to the manufacturer’s instructions (http://www. affymetrix.com/support/technical/manual/expression _ manual.affx). Briefly, 5 Ag of DNase digested total RNA was used to generate double-stranded cDNA using SuperScript reagents (Life Technologies) and a T7-linked oligo(dT) primer. cRNAs were synthesized using the Enzo Bioarray High Yield RNA transcript labeling kit (Affymetrix, UK) and resulting biotinylated labeled cRNA was subsequently fragmented into 35- to 200-bp lengths using the Fragmentation buffer (Affymetrix, UK). As recommended by the manufacturers, RNA, cDNA and cRNA quality and size distribution were visually confirmed by agarose gel electrophoresis. Spike controls B2, bio-B, bio-C, bio-D and Cre-x were added to the hybridisation cocktail before overnight hybridisation at 45 -C for 16 h. Arrays were stained and washed on the Fluidics Station 400 (Affymetrix, UK) using the EukGE-WS2 protocol (dual staining) before being scanned twice on the GeneChip Scanner 3000 at an excitation wavelength of 488 nm. cRNA was initially hybridised to Test3 GeneChips to confirm sample integrity prior to hybridisation to U133A microarrays. Microarray analysis complied with MIAME standards (http:// www.mged.org/Workgroups/MIAME/miame.html) and hybridisation results have been submitted to the Gene Expression Omnibus (GEO) (see below).
273
2.5. Data analysis To extract significant changes (2-fold or higher) in gene expression between healthy and carious samples, the comparison data files generated by the Affymetrix Microarray Suite 4.0 (MAS 4.0) software were initially sorted by deleting transcripts (i) displaying no change in expression, (ii) having a detection P value of equal to or higher than 0.045 and (iii) having a signal log ratio of between 1 and +1. Based on previous experiences, Affymetrix recommendations and published literature [21] transcripts exhibiting over 2-fold change were subsequently classed as being differentially expressed between the healthy and carious pulpal samples. The data are available at the Gene Expression Omnnibus (GEO) website under the series number GSE1629 (http://www.ncbi.nlm.nih.gov/geo/). To classify genes according to standardized GeneOntology (GO), vocabulary for the categories of biological process and molecular function information from the NetAffx software (Affymetrix, UK) was utilised or entire datasets were imported into the Onto-Express software (http://vortex.cs.wayne.edu/projects.htm) [22]. The default binomial probability distribution and Bonferroni multiple experiment correction settings of the software were used for data analysis. Probability values were derived from the numbers of genes corresponding to each GO category among the differentially expressed datasets compared with the number of genes expected for each GO category, based on their representation on the Affymetrix U133A array [22]. 2.6. cDNA synthesis for PCR analysis For single-stranded cDNA synthesis for use in PCR analysis, 1 Ag of DNase digested total RNA extracted was used as templates for reverse transcription (Omniscript kit, Qiagen, UK). Either oligo-dT, for semi-quantitative RT-PCR analysis, or random hexamers, for real-time RT-PCR analysis, was used to prime the reverse transcription reactions, performed as recommended by the manufacturer (Qiagen, UK). Synthesized cDNA concentrations were determined from absorbance values at a wavelength of 260 nm using a BioPhotometer (Eppendorf, UK). 2.7. Semi-quantitative reverse transcriptase polymerase chain reaction (Sq-RT-PCR) analysis Sq-RT-PCR analysis was performed used the pooled RNA samples previously described for microarray analysis. Differential expression analysis was performed for the human genes SLIT2, Period-2 (PER2), Period-3 (PER3), osteoadherin, Glypican 3, Midkine, activin receptor interacting protein-1 (AIP1), osteoadherin and growth hormone receptor (GHR), Interleukin-11 (IL-11), Bone sialoprotein (BSP), matrix Gla protein (MGP), endothelial cell growth factor-1 (ECGF1), inhibin hA and orosomucoid-1 (ORM1)
274
J.L. McLachlan et al. / Biochimica et Biophysica Acta 1741 (2005) 271 – 281
with the housekeeping gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH) used as control. The primer sequences and cycling conditions used are shown in Supplementary Table S1. Fifty nanograms of cDNA were used to seed 50 AL PCRs, which were subjected to between 32 and 42 cycles. A typical amplification cycle of 95 -C for 20 s, 60/61 -C for 20 s and 72 -C for 20 s was performed using a Mastercycler thermal cycler (Eppendorf, UK). Following a designated number of cycles, 8 Al of the PCR mix was removed and the product separated and visualized on a 1.5% agarose gel containing 0.5 Ag/mL ethidium bromide. Primers were designed from the Affymetrix probe target ID sequences using the Primer3 program (http://frodo.wi.mit.edu/cgi-bin/ primer3/primer3_www.cgi). 2.8. Real-time RT-PCR analysis Real-time RT-PCR analysis was performed using RNAs obtained from pulps from clinically healthy and carious teeth exhibiting a range of carious lesions from enamel involvement only to pulpal exposure. Expression levels of the human genes Adrenomedullin (ADM) and Dentine Matrix Phosphoprotein-1 (DMP-1) were analysed. TaqMan probes and primers were designed to span intron– exon boundaries using Primer Express 1.5 software (Applied Biosystems, UK) (Table S1). HPLC purified primers and probes were labeled with the fluorescent reporter dye 6-FAM and the quencher dye 6-TAMRA, at the 5V and 3V ends, respectively (Thermo Hybaid, UK; Eurogentec, UK). Quantitative PCR was performed using the ABI Prism 7700 Sequence Detection System (Applied Biosystems, UK). Each individual 25 Al PCR was seeded with 25 ng of cDNA and contained final concentrations of 1 Universal TaqMan Master Mix (Applied Biosystems, UK), 0.9 AM of forward and reverse primer, 0.15 AM of 6-FAM probe, 1 18s rRNA eukaryotic endogenous control (5Vlabeled with 6VIC) (Applied Biosystems, UK). PCRs were performed in 96-well plates covered with MicroAmp Optical Caps (Applied Biosystems, UK). Amplification conditions were 50 -C for 2 min, 95 -C for 10 min, then 44 cycles of 95 -C for 15 s and 60 -C for 1 min. All PCR assays were performed in triplicate. ABI 7700 Prism SDS 2.1 software (Applied Biosystems, UK) recorded fluorescence intensity of reporter and quencher dyes and results were plotted versus time, represented by cycle number. To obtain a precise quantification of initial target, amplification plots were examined during the early log phase of product accumulation above background [the threshold cycle (Ct) number]. Ct values were subsequently used to determine DCt values (DCt = Ct of the gene minus Ct of the 18S rRNA control) and differences in Ct values were used to quantify the relative amount of PCR product, expressed as fold change by applying the equation 2 DCt.
3. Results 3.1. Microarray analysis From 5 Ag total RNA starting amounts, approximately 35 Ag of fragmented biotin labeled healthy and carious cRNA was generated. The high quality of each of the target samples was confirmed using Affymetrix Test3 GeneChips (see Table 1). HG_U133A microarrays were subsequently hybridised in duplicate with 15 Ag of each of the cRNA samples. Analysis of control parameter data (as described for Test3 GeneChip analysis) confirmed hybridisation success (data not shown) and scaling factors were within Affymetrix recommended guidelines, values of 0.732 and 1.116 for duplicate healthy samples, and 1.006 and 1.222 for carious samples. Of the 22,283 probes sets represented on the U133A GeneChips, healthy pulp replicate hybridisations detected 49.4% and 44.1% as present with 1.8% and 2.0% as being marginally present, respectively. For the carious replicate hybridisations, 41.4% and 40.4% were detected as being present with 2.1% and 2.2% detected as being marginally present, respectively. Pairwise analysis of duplicate hybridisation data indicated that of the 9766 genes detected as being present by both targets, 445 genes, 4.56% of detected genes, were 2fold or greater differentially expressed between healthy and carious pulpal samples. Of this dataset, 85 and 360 were more abundant in healthy and carious tissue, respectively. The majority of the differentially expressed genes (3.65%) were 2- to 10-fold differentially expressed with only 0.91% exhibiting greater than 10-fold differentially expression. Fig. 1 provides graphical representation of the differential expression distribution of the genes detected. Please refer to the Supplementary Material for this article for a compre-
Table 1 GeneChip Test3 cRNA hybridisation quality control Control measurement
Average background Noise (Raw Q) Signal value of eukaryotic hybridisation controls: bio-B bio-C bio-D cre-x Housekeeping gene signal value ratios (3Vto 5V): h-actin GAPDH
Observed value Healthy
Carious
90.52
83.65
2.75
2.51
61.24 159.62 671.13 2131.11
65.8 144.02 643.58 2047.12
1.67 1.29
1.55 1.41
Affymetrix recommendation Range 20 – 100 units Values within 10 – 15% Values within 10 – 15%
Probe detection and comparable signal intensities
Ratio less than 3
5 Ag aliquots of fragmented healthy and carious cRNA were hybridised to Test3 GeneChips. B2 oligo hybridisations enabled accurate grid alignment for GeneChip Test3 features.
J.L. McLachlan et al. / Biochimica et Biophysica Acta 1741 (2005) 271 – 281
275
correlating with the presence of infiltrating immune cells recruited in response to bacteria infection. Supportive of these data were significant molecular function GOs, which included chemokine, MHC class II receptor and manganese superoxide dismutase activities also previously shown to be associated with an immune response to bacteria infection. In contrast, the GO analysis of genes more abundantly expressed in healthy pulpal tissue identified a variety of processes and functions potentially lost or down-regulated due to disease progression. Of particular note were the biosynthetic pathways, which may be down-modulated due to the ensuing inflammation. 3.3. Individual gene expression analysis
Fig. 1. (a) Scatter plot of average healthy versus carious pulp sample signal intensity values for individual microarray features, plotted on a double logarithmic scale. Areas highlighted indicate: A—No differential expression; B—2- to 10-fold more abundantly expressed; C—>10-fold more abundantly expressed. (b) Histogram presenting frequency of fold change of the 445 2-fold or greater differentially expressed genes.
hensive list of genes increased (Supplementary Table S2) or decreased (Supplementary Table S3) in pulpal tissue due to carious disease. 3.2. Gene ontology analysis To translate the data into a more meaningful biological context and to characterize more thoroughly sets of functionally related genes, the differentially expressed datasets were subsequently organized into GO groupings using the publicly available software Onto-Express (OE). The OE software subsequently automatically translates the nonredundant datasets using standardized GO terms for biological process and molecular function providing statistical significance results for each category. GO groupings were determined by comparing the number of genes expected for each GO category based on their representation on the Affymetrix HG_U133A array. Significant differences from the expected were calculated with a two-sided binomial distribution. Figs. 2 and 3 show all GO classes with a Bonferroni-corrected significance of P < 0.05 for the two datasets analysed. Notably, the biological process GO groupings demonstrating the most significant representation in our set of genes more abundant in carious pulpal tissue, map to the inflammatory and immune response, although a variety of biological processes associated with cellular and tissue defence mechanisms were also identified, potentially
To validate the microarray data, semi-quantitative and real-time RT-PCR was performed on selected differentially expressed genes. Specific genes were chosen, based on criteria that included (i) their expression not previously having been described in mature pulpal tissue or during carious disease, (ii) their representation of a range of ontological classes/functions and (iii) their fold change and signal intensities spanning a range of values. Data confirmed that the 14 genes analysed by semi-quantitative RT-PCR (Fig. 4) were differentially expressed as predicted by the microarray data. No attempts were made to statistically correlate fold change with differential expression detected by PCR although a visible correlation was noted (Fig. 4). Real-time PCR analysis was also utilised to analyse expression levels of two genes, ADM1 and DMP1, in RNA extracted from individual human healthy and carious pulpal samples which exhibited a range of stages of disease severity from enamel involvement only through to pulpal exposure. Whilst overall expression level changes identified were in agreement with the microarray data, these analyses highlighted transcript level heterogeneity between samples, not only in carious, but also in healthy pulpal specimens (Fig. 5). No correlations were detected using regression analysis [18] between ADM and DMP1 expression levels or between expression levels and disease severity in the sample set used (data not shown). Real-time PCR analysis was also employed to further characterize the expression of ADM1 and DMP1 in a panel of immune system cells previously implicated in carious disease progression [10]. Expression profiling data indicated that the levels of ADM during carious disease were potentially attributable to infiltrating immune system cells, in particular activated neutrophils, whilst DMP1 was expressed predominantly by cells native to healthy pulps (Fig. 6).
4. Discussion Vital pulp therapy aims to treat reversible pulpal injury whilst maintaining cell viability and function. Currently, two
276
J.L. McLachlan et al. / Biochimica et Biophysica Acta 1741 (2005) 271 – 281
Fig. 2. The significant (Bonferroni-corrected-P < 0.05) biological processes and molecular functions identified by the 360 genes more abundantly expressed in carious pulpal tissue. P value and the number of genes corresponding to each GO functional class are shown.
therapeutic approaches are frequently used, (i) indirect pulp capping in cases of deep dentinal lesions and (ii) direct pulp capping/pulpotomy in cases of pulp exposure. Successful treatment outcomes are dependent on the type and location of the injury, age of the tooth, the pulp capping material used and the integrity of the cavity restoration. In the future, it is hoped that biomimetic approaches will be developed based on a thorough understanding of the molecular and cellular events that occur during dental injury and disease. Such therapies aim to maintain the pulp’s vitality and stimulate its innate reparative and regenerative mechanisms. Previous molecular and biochemical studies characterizing caries have, however, so far been limited. In this current study, we have addressed this issue and now describe large-
scale gene expression analyses in human healthy and carious pulpal tissue. To date, most reported microarray studies have focused on cultured cells exposed to factors important for disease progression. Subsequently, the disease relevance of the gene expression changes identified required validation in vivo. For our approach, we have directly used pooled clinical samples from in excess of 10 healthy and 10 carious pulpal teeth. Pooling of RNA from such relatively high numbers of specimens was required to generate sufficient amounts of RNA to enable microarray analysis, providing an Faverage_ response of core pulpal tissues under deep carious lesions. Notably, the use of microarrays and complex tissues poses challenges as the pulp is composed of heterogeneous and
J.L. McLachlan et al. / Biochimica et Biophysica Acta 1741 (2005) 271 – 281
277
Fig. 3. The significant (Bonferroni-corrected-P < 0.05) biological processes and molecular function identified by the 85 genes more abundantly expressed in healthy pulpal tissue. P value and the number of genes corresponding to each GO functional class are shown.
changing cell populations with interactions between immune and non-immune cells being pivotal to disease pathogenesis [10]. Indeed, analysis of the expression of CD
antigens in carious disease (Supplementary Table S2) highlights this complexity. Gene expression changes identified using this approach will therefore represent an average
Fig. 4. (a) (i) Eight and (ii) six genes identified as more abundant in healthy (H) and carious (C) pulpal samples, respectively, were analysed by semiquantitative RT-PCR analysis. (b) Expression levels were normalised to GAPDH housekeeping gene levels. Superscript a or b indicates the GAPDH gel image used for normalisation for each differentially expressed gene analysed. Fold-change identified from microarray experiments are shown in parentheses.
278
J.L. McLachlan et al. / Biochimica et Biophysica Acta 1741 (2005) 271 – 281
Fig. 5. Real-time PCR analysis for ADM and DMP1 in (a) 5 healthy (nos. 1 – 5) and 5 carious (nos. 6 – 10) pulpal samples. Fold change was obtained by comparison with the sample exhibiting the lowest level of expression which was assigned a value of 1. (b) Average expression levels for carious and healthy samples from (a) is shown. Vertical bars indicate T one standard error of the mean. Asterisks indicate significant difference between carious and healthy gene expression levels ( P < 0.05).
of many different cell types and transcripts from cells native to the pulp may appear decreased relative to the total mRNA pool as a result of infiltrating immune cell populations. Since genes contained on the arrays were not selected specifically for the analysis of gene expression in caries, they do not reflect biases inherent to its pathogenesis. This approach, therefore, enhances the likelihood of identifying important expression changes that are not currently part of our understanding of caries. Pertinently, whilst results obtained (Supplementary Tables S2 and S3) concur with differential expression reported elsewhere, including increases in expression in caries of IL-1h, IL-6, IL-8, S100A8 and S100A9 [18] and decreases of DSPP [20], our data now identify a significant number of expression changes associated with chronic inflammation and tissue remodelling that have not previously been described. The translation of such high-throughput gene expression data into meaningful biological information and identification of associated pathways/networks is a challenge to biologists and bio-informaticians in this post-genomic era. As
manual literature searching is inefficient, extremely time consuming and does not provide a common language for researchers, we chose to use the Onto-Express software, which utilises the standardized GO vocabulary [23]. The GO groupings identified for biological process and molecular function from the genes more abundant in diseased pulps (Fig. 2) identify many groupings previously linked with an immune/inflammatory response associated with a bacterial infection. In conjunction with the TEST3 GeneChip hybridisation results (Table 1), which indicated the high quality of the labeled cRNA, this analysis provides additional confidence in the list of differentially expressed genes presented in this study. In contrast, the GO groupings identified by genes more abundant in healthy pulps identify potentially homeostatic biological processes and molecular functions down modulated during the disease process (Fig. 3), many of which warrant further investigation. Our confirmatory PCR analyses (Figs. 4 and 5) support our microarray findings and highlight the differential expression in pulpal tissue of several genes previously not
Fig. 6. Real-time PCR analysis of ADM and DMP1 gene expression in immune system cells. Mo=Monocytes; Mf=Macrophage; PBL=Lymphocytes; N=Neutrophil; N+=E. coli LPS stimulated neutrophils; H=Pooled healthy pulp; C=Pooled carious pulpal tissue.
J.L. McLachlan et al. / Biochimica et Biophysica Acta 1741 (2005) 271 – 281
associated with carious disease. Whilst genes down-regulated due to caries may correspond to pulpal transcripts representing a decreased proportion of the total mRNA pool, these analyses nevertheless provides novel data on their dental tissue expression. In particular, the presence of the circadian clock genes, PER2 and PER3, has not previously been reported. PER2 has, however, been described in murine bone marrow [24] and data indicated its expression to be potentially lineage or differentiation stage-dependent. As circadian rhythms have also been associated with the deposition of incremental dentine [25,26], the role of these genes in dental tissue repair warrants further investigation. Differential expression of members of the non-collagenous protein family termed SIBLING (Small Integrin-Binding Ligand, N-linked Glycoprotein), namely DMP1 and BSP, were also confirmed by PCR analyses (Figs. 4 and 5). This family also includes osteopontin (OPN), DSPP and matrix extracellular phosphoglycoprotein (MEPE), which play key biological roles in mineralization of bone and dentin ECM [27]. Data presented here and previously [18] now indicate, that whilst both DMP1 and DSPP (Supplementary Table S3) are down-regulated by core pulpal tissue during caries, the converse is true for the transcript levels of BSP (Fig. 4) and OPN (Supplementary Table S2). Interestingly, work by Decup et al. [28] has also shown that BSP is capable of stimulating pulpal cells to differentiation into odontoblasts which secrete an organized ECM. Hence, the increased expression of BSP may be indicative of potential repair processes ongoing within diseased pulp. IL-11 (Fig. 4) is a pleiotropic cytokine produced by a variety of stromal cells [29]. Recent work [30,31] has shown that IL-11 acts synergistically with BMP-2 to induce osteoblast differentiation and accelerate bone formation. As previous reports have shown BMP-2 capable of stimulating pulpal cells to differentiate into odontoblast [32 –35], our data now indicate that the role of IL-11 in this process should also be considered. Whilst the expression of SLIT2, GHR, Inhibin hA, midkine, osteoadherin and matrix Gla protein has been described during tooth development and mineralization [36 –41], our analyses now provide evidence of their differential expression in mature healthy and carious pulpal tissue (Fig. 4). Notably, another potentially interesting finding of our microarray analyses is the identification of the relatively dramatic downregulation, 29-fold, of zeta-crystallin in pulps from carious teeth (Supplementary Table S3). Whilst this represents novel data, other family members have been shown to play a role in mineralized tissue function and are potentially important in the pulp’s response to mechanical and thermal stresses [42,43]. Perhaps, our most significant finding is the increased ADM expression associated with caries. ADM is a recently discovered pluripotent peptide hormone important in a range of physiological and pathological processes. Many tissues of the body express it and it has a remarkable range of effects mediated by paracrine/autocrine and endocrine
279
mechanisms [44]. Active ADM is derived by proteolytic cleavage of pro(p)ADM which subsequently acts through two receptor subtypes, termed L1-receptor (L1-R) and calcitonin receptor-like receptor (CRLR). The CRLR functions as either a calcitonin gene-related peptide (CGRP) receptor or a selective ADM receptor depending on which member of a family of chaperones, named receptor-activitymodifying proteins (RAMPs), is expressed. RAMP1 generates CGRP receptors, whilst RAMP2 and RAMP3 produce L1-Rs [45]. ADM is important in embryonic tissue development, particularly in the tooth, where strong mesenchymal – epithelial interactions occur [45]. ADM can act as a mitogen, apoptosis survival factor, angiogenic stimulant and regulator of immune response [46,47] and is therefore a powerful candidate mediator of various cellular responses in the pulp. In addition, ADM is a potent osteoblast stimulator, in terms of proliferation and osteogenic activity, both in vitro and in vivo, and as a result has been proposed as an attractive candidate treatment for bone diseases and defects [48]. Our analyses now highlight its differential expression within the pulp and our microarray data also indicate that RAMP1, RAMP3, CRLR and CGRPreceptor component protein expression are all expressed in healthy pulps although their expressions were unchanged as a result of disease (data not shown). These findings indicate that further characterisation of the role of ADM, its receptors and RAMPs is warranted as it may provide a novel target for dental inflammation and regeneration control. In addition to the confirmation of differential expression of the 16 genes reported here (Figs. 4 and 5), we have also verified differential expression of 6 genes up-regulated and 4 genes down-regulated during caries which represent transcripts of unknown function (unpublished data). Work is currently underway to characterize the role of these novel genes in dental health and disease. A better understanding of the roles of the differentially expressed genes identified by our microarray analyses during the caries disease process will require identification of their cellular origin using techniques such as in situ hybridisation and immunohistochemistry in mature and developing dental tissue. Such analyses will determine whether expression of these molecules is by cells native to the pulp or due to infiltrating immune cells. Notably, comparison of genes up-regulated in pulpal tissue during caries with genes increased in neutrophils stimulated with the cariogenic pathogen, Fusobacterium nucleatum [49], has identified nine genes common to these two datasets (unpublished data). These analyses indicate that the metallothioneins 1L, 1X and 2A, the GRO1 oncogene, ORM1, RNase A k6, the MAFF transcription factor, HLA-DQB1 and superoxide dismutase 2, are potentially expressed by the infiltrating neutrophil population during carious disease. Technological improvements will soon facilitate interrogation of whole genome microarrays with RNA derived from single teeth. Thus, high-throughput gene expression
280
J.L. McLachlan et al. / Biochimica et Biophysica Acta 1741 (2005) 271 – 281
analyses, taking into account important disease factors, such as age of carious lesion, remaining dentine thickness, age of patient, disease severity and infecting microbial population, will enable a more comprehensive understanding of the molecular circuitry involved in carious disease and facilitate future targeted treatment regimes for early and end stage disease. Our study provides novel data regarding the molecular characterisation of the complex interplay between reparative and inflammatory events that occur under lesions from one of the most prevalent chronic diseases in the world. In addition, potential novel targets are identified for further research that may enable development of new treatment modalities targeted at controlling the inflammatory process and maintaining tooth vitality, whilst permitting essential defence and repair reactions to occur. Such approaches have the potential for immense impact on healthcare globally.
[8]
[9]
[10]
[11]
[12]
[13] [14]
[15]
Acknowledgements JLM is supported by a University of Birmingham, Scientific Projects Committee grant (EBX 1193). We would like to thank Miss Sally Kerr (Oral Surgery) for providing the teeth for this investigation. We thank Dr T. Stankovic for provision of cRNA labeling reagents. This work was supported in part by an Oral and Dental Research Trust Grant.
[16]
Appendix A. Supplementary data
[19]
Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.bbadis.2005. 03.007.
References [1] K.J. Anusavice, Dental caries: risk assessment and treatment solutions for an elderly population, Compend. Contin. Educ. Dent. 23 (2002) 12 – 20. [2] P.E. Petersen, The World Oral Health Report 2003 Report. Continuous improvement of oral health in the 21st century—The approach of the WHO Global Oral Health Programme [Online]. Oral Health Programme, Noncommunicable Disease Prevention and Health Promotion, World Health Organization, Geneva, Switzerland. http://www.who. int/oral_health/publications/report03/en/ [2003]. [3] Y. Abiko, Passive immunization against dental caries and periodontal disease: development of recombinant and human monoclonal antibodies, Crit. Rev. Oral Biol. Med. 11 (2000) 140 – 158. [4] F.J. Burke, S.W. Cheung, I.A. Mjor, N.H. Wilson, Reasons for the placement and replacement of restorations in vocational training practices, Prim. Dent. Care 6 (1999) 17 – 20. [5] A.J. Smith, Pulpal responses to caries and dental repair, Caries Res. 36 (2002) 223 – 232. [6] K.R. Baumgardner, M.A. Sulfaro, The anti-inflammatory effects of human recombinant copper-zinc superoxide dismutase on pulp inflammation, J. Endod. 27 (2001) 190 – 195. [7] R.B. Rutherford, K. Gu, Treatment of inflamed ferret dental pulps with
[17]
[18]
[20]
[21]
[22]
[23] [24]
[25]
[26]
[27]
[28]
[29] [30]
recombinant bone morphogenetic protein-7, Eur. J. Oral. Sci. 108 (2000) 202 – 206. A. Fouad, Molecular mediators of pulpal inflammation, in: K.M. Hargreaves, H.E. Goodis (Eds.), Seltzer and Bender’s Dental Pulp, Quintesscence Publ. Co, Chicago, 2002, pp. 247 – 279. K.J. Heyeraas, O.B. Sveen, I.A. Mjor, Pulp – dentin biology in restorative dentistry. Part 3: pulpal inflammation and its sequelae, Quintessence 32 (2001) 611 – 625. T. Izumi, I. Kobayashi, K. Okamura, H. Sakai, Immunohistochemical study on the immunocompetent cells of the pulp in human non-carious and carious teeth, Arch. Oral Biol. 40 (1995) 609 – 614. C.R. Rauschenberger, J.C. Bailey, C.J. Cootauco, Detection of human IL-2 in normal and inflamed dental pulps, J. Endod. 23 (1997) 366 – 370. P. Stashenko, Interrelationship of dental pulp and apical periodontitis, in: K.M. Hargreaves, H.E. Goodis (Eds.), Seltzer and Bender’s Dental Pulp, Quintesscence Publ Co, Chicago, 2002, pp. 389 – 410. H. Trowbridge, Pathogenesis of pulpitis resulting from dental caries, J. Endod. 7 (1981) 52 – 60. R. D’Souza, L.R. Brown, J.R. Newland, B.M. Levy, L.B. Lachman, Detection and characterization of interleukin-1 in human dental pulps, Arch. Oral Biol. 34 (1989) 307 – 313. G.T. Huang, A.P. Potente, J.W. Kim, N. Chugal, X. Zhang, Increased interleukin-8 expression in inflamed human dental pulps, Oral Surg. Oral Med. Oral Pathol. Oral Radiol. Endo. 88 (1999) 214 – 220. Y. Jiang, T.R. Russell, H. Schilder, D.T. Graves, Endodontic pathogens stimulate monocyte chemoattractant protein-1 and interleukin-8 in mononuclear cells, J. Endod. 24 (1998) 86 – 90. K. Matsushima, E. Ohbayashi, H. Takeuchi, S. Hosoya, Y. Abiko, M. Yamazaki, Stimulation of interleukin-6 production in human dental pulp cells by peptidoglycans from lactobacillus case, J. Endod. 24 (1998) 252 – 255. J.L. McLachlan, A.J. Sloan, A.J. Smith, G. Landini, P.R. Cooper, S100 and cytokine expression in caries, Infect. Immun. 72 (2004) 4102 – 4108. M. Zehnder, N. Delaleu, Y. Du, M. Bickel, Cytokine gene expression—Part of host defence in pulpitis, Cytokine 22 (2003) 84 – 88. J.L. McLachlan, A.J. Smith, A.J. Sloan, P.R. Cooper, Gene expression analysis of genes of the dentine pulp complex in healthy and carious teeth, Arch. Oral Biol. 48 (2003) 273 – 283. L. Wodicka, H. Dong, M. Mittmann, M.H. Ho, D.J. Lockhart, Genome-wide expression monitoring in Saccharomyces cerevisiae, Nat. Biotechnol. 15 (1997) 1359 – 1367. S. Draghici, P. Khatri, R.P. Martins, G.C. Ostermeier, S.A. Krawetz, Global functional profiling of gene expression, Genomics 81 (2003) 98 – 104. D. Murphy, Gene expression studies using microarrays: principles, problems, and prospects, Adv. Physiol. Educ. 26 (2002) 256 – 270. Y.G. Chen, A. Mantalaris, P. Bourne, P. Keng, J.H. Wu, Expression of mPer1 and mPer2, two mammalian clock genes, in murine bone marrow, Biochem. Biophys. Res. Commun. 276 (2000) 724 – 728. M.C. Dean, A.E. Scandrett, The relation between long-period incremental markings in dentine and daily cross-striations in enamel in human teeth, Arch. Oral Biol. 41 (1996) 233 – 241. M. Ohtsuka-Isoya, H. Hayashi, H. Shinoda, Effect of suprachiasmatic nucleus lesion on circadian dentin increment in rats, Am. J. Physiol., Regul. Integr. Comp. Physiol. 280 (2001) R1364 – R1370. C. Qin, O. Baba, W.T. Butler, Post-translational modifications of sibling proteins and their roles in osteogenesis and dentinogenesis, Crit. Rev. Oral Biol. Med. 15 (2004) 126 – 136. F. Decup, N. Six, B. Palmier, D. Buch, J.J. Lasfargues, E. Salih, M. Goldberg, Bone sialoprotein-induced reparative dentinogenesis in the pulp of rat’s molar, Clin. Oral Investig. 4 (2000) 110 – 119. S.X. Leng, J.A. Elias, Interleukin-11, Int. J. Biochem. Cell Biol. 29 (1997) 1059 – 1062. K. Suga, M. Saitoh, S. Kokubo, S. Fukushima, S. Kaku, S. Yasuda, K. Miyata, Interleukin-11 acts synergistically with bone morphoge-
J.L. McLachlan et al. / Biochimica et Biophysica Acta 1741 (2005) 271 – 281
[31]
[32] [33]
[34]
[35]
[36]
[37]
[38]
[39]
netic protein-2 to accelerate bone formation in a rat ectopic model, J. Interferon Cytokine Res. 23 (2003) 203 – 207. K. Suga, M. Saitoh, S. Kokubo, K. Nozaki, S. Fukushima, S. Yasuda, M. Sasamata, K. Miyata, Synergism between interleukin-11 and bone morphogenetic protein-2 in the healing of segmental bone defects in a rabbit model, J. Interferon Cytokine Res. 24 (2004) 343 – 349. I. About, T.A. Mitsiadis, Molecular aspects of tooth pathogenesis and repair: in vivo and in vitro models, Adv. Dent. Res. 15 (2001) 59 – 62. C. Begue-Kirn, A.J. Smith, J.V. Ruch, J.M. Wozney, A. Purchio, D. Hartmann, H. Lesot, Effects of dentin proteins, transforming growth factor ß1 (TGFß1) and bone morphogenetic protein 2 (BMP2) on the differentiation of odontoblasts in vitro, Int. J. Dev. Biol. 36 (1992) 491 – 503. C. Begue-Kirn, A.J. Smith, M. Loriot, C. Kupferle, J.V. Ruch, H. Lesot, Comparative analysis of TGFßs, BMPs, IGF, msxs, fibronectin, osteonectin and bone sialoprotein gene expressions during normal and in vitro induced odontoblast differentiation, Int. J. Dev. Biol. 38 (1994) 405 – 420. T. Saito, M. Ogawa, Y. Hata, K. Bessho, Acceleration effect of human recombinant bone morphogenetic protein-2 on differentiation of human pulp cells into odontoblasts, J. Endod. 30 (2004) 205 – 208. C.W. Brown, D.E. Houston-Hawkins, T.K. Woodruff, M.M. Matzuk, Insertion of Inhbb into the Inhba locus rescues the Inhba-null phenotype and reveals new activin functions, Nat. Genet. 25 (2000) 453 – 457. R. Buchaille, M.L. Couble, H. Magloire, F. Bleicher, Expression of the small leucine-rich proteoglycan osteoadherin/osteomodulin in human dental pulp and developing rat teeth, Bone 27 (2000) 265 – 270. S. Loes, K. Luukko, I.H. Kvinnsland, P. Kettunen, Slit1 is specifically expressed in the primary and secondary enamel knots during molar tooth cusp formation, Mech. Dev. 107 (2001) 155 – 157. T.A. Mitsiadis, T. Muramatsu, H. Muramatsu, I. Thesleff, Midkine (MK), a heparin-binding growth/differentiation factor, is regulated by retinoic acid and epithelial – mesenchymal interactions in the deve-
[40]
[41] [42]
[43]
[44] [45]
[46]
[47]
[48] [49]
281
loping mouse tooth, and affects cell proliferation and morphogenesis, J. Cell Biol. 129 (1995) 267 – 281. A.L. Symons, A.C. Henry, S. Chang, T.J. Daley, D.J. Harbrow, B.K. Joseph, The effect of glucocorticosteroid treatment on dentine formation in the Lewis rat, a histological study, Growth Factors 18 (2000) 157 – 167. P.G. Robey, Vertebrate mineralized matrix proteins: structure and function, Connect. Tissue Res. 35 (1996) 131 – 136. K. Behnam, S.S. Murray, J.P. Whitelegge, E.J. Brochmann, Identification of the molecular chaperone alpha B-crystallin in demineralized bone powder and osteoblast-like cells, J. Orthop. Res. 20 (2002) 1190 – 1196. T. Thyagarajan, A.B. Kulkarni, Transforming growth factor-beta1 negatively regulates crystallin expression in teeth, J. Bone Miner. Res. 17 (2002) 1710 – 1717. M. Jougasaki, J.C. Burnett, Adrenomedullin: potential in physiology and pathophysiology, Life Sci. 66 (2000) 855 – 872. L.M. Montuenga, A. Martinez, M.J. Miller, E.J. Unsworth, F. Cuttitta, Expression of adrenomedullin and its receptor during embryogenesis suggests autocrine or paracrine modes of action, Endocrinology 138 (1997) 440 – 451. D.M. Smith, H.A. Coppock, D.J. Withers, A.A. Owji, D.L. Hay, T.P. Choksi, P. Chakravarty, S. Legon, D.R. Poyner, Adrenomedullin: receptor and signal transduction, Biochem. Soc. Trans. 30 (2002) 432 – 437. T.H. Elsasser, S. Kahl, Adrenomedullin has multiple roles in disease stress: development and remission of the inflammatory response, Microsc. Res. Tech. 57 (2002) 120 – 129. J. Cornish, D. Naot, I.R. Reid, Adrenomedullin—A regulator of bone formation, Regul. Pept. 112 (2003) 79 – 86. F.E. Martin, M.A. Nadkarni, N.A. Jacques, N. Hunter, Quantitative microbiological study of human carious dentine by culture and realtime PCR: association of anaerobes with histopathological changes in chronic pulpitis, J. Clin. Microbiol. 40 (2002) 1698 – 1704.