Abstract
Little is known about the effect of salivary gland function during aging based on gene expression. Recently emerged DNA array technology provides a sensitive, quantitative, rapid approach to the monitoring of the global pattern of gene expression. In this study, we used high-density oligonucleotide arrays to monitor the changes of gene expression levels in the submandibular gland (SMG) by comparing adult mice with elderly adult mice. Of the 1328 genes screened, 160 genes (12.0%) showed more than two-fold changes; 154 (96.3%) of these genes, associated with transcription regulation, transport, signal transduction, and enzymes in the elderly mice, exhibited decreased expression levels. The remaining 6 genes (3.7%) in the elderly mice showed increased expression levels. In mouse SMG, analysis of these data suggests that aging may lead the gene expression to decrease than increase. Thus, DNA array technology can be a powerful tool for the identification of age-associated candidate genes for further analysis in aging.
INTRODUCTION
It is well-known that saliva plays an important role in the functions of taste, mastication, and speech. An age-related decrease in the protein synthesis and salivary flow rate has been reported, whereas other reports have shown no relationship to age. These discrepancies have been reported even in healthy subjects (non-medicated) and may be related to the method of saliva sample collection and the stimulus used (Vissink et al., 1996: Yeh et al., 1998). Tylenda et al. (1988) did not observe an age-related decrease in submandibular/sublingual saliva flow, but Pedersen et al. (1985) found a significant age-related reduction following mild stimulation. Wu et al. (1995) reported an age-related decrease following intense stimulation for an extended period of time in SMG. Moreover, Yeh et al. (1998) demonstrated significant reductions of both unstimulated and stimulated SMG flow rates in subjects with advancing age.
Recently, there has been great progress in techniques for the comparison of differentially expressed genes. DNA microarrays are divided into two classes, cDNA-based and oligonucleotide-based. Affymetrix (Santa Clara, CA, USA) oligonucleotide arrays contain 40 specific oligonucleotide probes for each gene. Twenty of these oligonucleotides represent perfect matches (PMs), and the rest of them contain single mismatches (MMs) as compared with the PMs. The hybridization intensity from each MM is subtracted from that of its PM for the calculation of an “Average Signal Intensity”. GeneChip Expression Analysis software (Affymetrix) examines the hybridization intensity data from one experiment to calculate a set of absolute metrics. Some of the metrics are used along with a detection matrix to determine an Absolute Call (Abs Call) judged as “Present”, “Absent”, or “Marginal” for each gene. In this study, we used oligonucleotide microarray analysis to detect age-associated changes in gene expression in mouse submandibular glands for screening a large number of candidate genes and understanding their interactions.
MATERIALS & METHODS
Animal Treatment
Senescence-accelerated mice (SAM), which have been under development at Kyoto University, were obtained from Kyoto University (Kyoto, Japan). Food and water were given ad libitum. The institutional review board reviewed and approved the animal use protocol. A feature of aging is observed after approximately 4 mos, and pregnancy is attained at 2 mos after birth in SAM (Takeda et al., 1997). Thus, we chose mice aged 2 to 3 mos (2M) as a model for adults, and 15-month-old (15M) males as a model for the elderly, because the average life span was 18 mos.
Oligonucleotide Microarray Analysis
One μg of mRNA was mixed with T7-(dT)24 primer containing a T7 RNA polymerase promoter sequence, reverse-transcribed to cDNA, and synthesized to the double-stranded cDNA with a SuperScript Choice System (Invitrogen, Carlsbad, CA, USA). Using cDNA as a template, we carried out in vitro transcription with Bio-11-CTP and Bio-16-UTP (Enzo Diagnostics, Farmingdale, NY, USA), using a T7 Megascript kit (Ambion, Austin, TX, USA). The biotin-labeled cRNAs were purified, fragmented, and hybridized to the array (GeneChip Mu6500; Affymetrix) spotted about 6500 known mouse genes or expression sequence tags (ESTs) (Sequence source: Unigene database, August, 1996). The arrays were washed and stained with streptavidin-phycoerythrin (Molecular Probes, Eugene, OR, USA) by means of an Affymetrix Fluidics Station. Fluorescence intensities were scanned with an Affymetrix GeneArray Scanner. The experiment was carried out in duplicate. Six of the 2M mice were randomly divided into 2 groups. The cRNA synthesized from individual groups was hybridized to each microarray. Six of the 15M mice were treated in the same fashion (Appendix Fig.: www.dentalresearch.org).
Data Analysis
Data analysis was performed with the GeneChip Expression Analysis software (Affymetrix) and GeneSpringTM software (Silicon Genetics, Redwood, CA, USA). The differences in hybridization efficiency among arrays were equalized by intensities of spiked-in control cRNA (BioD). A further explanation of data mining is available on our Web site (http://www2.mascat.nihon-u.ac.jp/biochem/MADS/Mammalia/Mouse/SAA/smgaa1.html). In a set of GeneChip experiments comparing the same sample hybridized to two different arrays, method-associated experimental artifacts produced less than two-fold differences between two identical samples. Thus, genes displayed over two-fold expression change were subjected to further testing.
RT-PCR
One μg of total RNA was reverse-transcribed with random hexamer, oligo (dT)12-18 primer, dNTPs, and SuperScript II (Invitrogen) in a total volume of 20 μL. One μL of the reaction mixture was then mixed with Taq DNA polymerase and Platinum Taq Antibody (Invitrogen) for RT-PCR analysis by use of the gene-specific primers. Twenty-five cycles of the PCR were completed, each cycle consisting of 30 sec at 94°C, 30 sec at 60°C, and 30 sec at 72°C.
RESULTS
To examine the molecular events associated with aging in mouse SMG, we used oligonucleotide arrays to detect changes in gene expression. After data mining, we subjected 1328 genes to further analysis. The majority of genes displayed decreased expression levels in 15M compared with those in 2M (Fig. 1). The expression levels of 154 genes had greater than two-fold decreases in 15M with aging, whereas 6 genes showed more than two-fold increases of their expression levels in 15M, by statistical group comparison between 2M and 15M (Appendix Table: www.dentalresearch.org). The category of altered genes is shown in Table 1. Several protein biosynthesis-, enzyme-, and transport-associated genes are on the list. To verify the microarray data, we tested randomly selected genes from the altered gene expression table using semi-quantitative RT-PCR (Table 2). The expression levels of tested genes (Genebank ID; AA038677, U49393, AA024297, and W62742) significantly decreased in 15M and 23M in contrast to those in 2M (Fig. 2).
With regard to saliva secretion, PI 4 kinase alpha (6.0-fold), phospholipase C alpha (2.1-fold), sarco-endoplasmic reticulum Ca(2+) ATPase (4.9-fold), Cl-channel 1 (2.2-fold), and aquaporin 8 (2.3-fold) exhibited decreases in aging. As for protein synthesis and secretion, several transcription/translation factors also displayed decreases: initiation factor 5A (15.2-fold), 4AI (6.9-fold) and 4E binding proteins (2.2-fold); elongation 1 (2.5-fold), 2 (2.1-fold), and factor Tu (3.6-fold); and transcription factor 3 (2.6-fold). The RNA helicase II/Gu and RNA-dependent ATPase/RNA helicase exhibited a 4.9- and 3.2-fold decrease, respectively. Others-sec 61 alpha (11.1-fold), four kinds of clathrin-associated proteins (2.2- to 3.5-fold), dynein (2.8-fold), and protein-tyrosine phosphatase-associated proteins (2.5- to 6.5-fold)-showed decreases in the aged mice as well.
DISCUSSION
The 6500 genes spotted on the array were selected from a cDNA database of common organs such as liver, brain, and blood cells, but not from that of SMG. In addition, the number of genes whose “Abs Call” equaled “Present” in at least 1 sample was 1968, suggesting that not all SMG genes may be represented on the Mu6500 GeneChip. Of the 1328 genes screened, 160 genes (12.0%) showed more than two-fold changes; 154 (96.3%) of these genes, associated with transcription regulation, transport, signal transduction, and enzymes in the elderly mice, exhibited decreased expression levels as a function of age. The remaining 6 genes (3.7%) showed increased gene expression levels. These observations suggest that the aging process may lead to significantly decreased gene expression in SMG.
It is known that the stimulation of an acinar cell activates phospholipase C to hydrolyze PI 4,5 bisphosphate to inositol 1,4,5 trisphosphate (IP3) and diacylglycerol. The IP3 binds to the receptor on the endoplasmic reticulum (ER), and releases Ca(2+) from the internal Ca(2+) stores in ER to the cytosol. The internal Ca(2+) is taken up by sarco/endoplasmic reticulum Ca(2+)-ATPases; the expressions of PI 4 kinase alpha (catalyzed the synthesis of PI 4 phosphate), phospholipase C alpha (hydrolyzed IP 4,5 bisphosphate to IP3), and the Ca(2+)-ATPases displayed decreases. Analysis of these data suggests that the IP3-dependent Ca(2+) release from ER to cytosol may not be enough for saliva secretion in the elderly mice in terms of the influence of aging in the SMG as a function of saliva secretion, because Ca(2+) in ER is not completely discharged by the lack of production of IP3 and reticulum Ca(2+)-ATPases. In addition, it is equally likely that decreased flow may be caused by loss of acinar cells with aging.
The decreased gene expression of transcription/translation factors and RNA helicase, including 'DEAD/H', by aging suggests that these protein reductions may directly affect protein synthesis. One of the genes that showed a significant decrease with aging was a protein transport protein, “Sec 61 alpha”. The misfolded secretory proteins are primarily degraded by cytosolic proteasomes. The protein-translocation channel formed by the Sec61 complex is responsible for both forward and retrograde transport of proteins across the ER membrane (Romisch, 1999). Therefore, the significant reduction of Sec 61 alpha may lead to secretion of the misfolded proteins.
It is known that the assembly of clathrin coats on specific membrane organelles leads to the formation of vesicles, and that cytoplasmic dynein plays an important role in maintaining the integrity, intracellular location, and function of the Golgi apparatus, as well as in the translocation of membrane between the endoplasmic reticulum and Golgi apparatus. Thus, the reductions of clathrin-associated protein and dynein may affect selective intracellular protein transportation.
In the salivary gland, tyrosine kinases and phosphatases play important roles in inducing the transition from stasis to active proliferation, but they have potential roles in mediating the secretory function of the glands. In endocrine cells, protein tyrosine phosphorylation and dephosphorylation are required for secretory vesicle release from the trans-Golgi network (Austin and Shields, 1996). Inhibitors of protein tyrosine phosphatase prevented amylase secretion from both untreated control and isoproterenol-stimulated parotid acinar cells (Purushotham et al., 1995).
A better understanding of the molecular effects of aging in the mouse SMG may help to reveal important aspects of the age-related process. In this study, we have demonstrated the oligonucleotide microarray as a new tool for the detection of changes in global gene expression patterns in the SMG of adult and elderly mice. This is the first investigation where GeneChip technology has been used to identify gene expression in salivary glands in aging. We analyzed the whole “submandibular gland”, which consists of mixed cells such as acinar, ductal, stroma, fatty, and so on, because the clinical phenomenon in elderly mice described in the “Introduction” has been associated with the function of the “whole gland” but not of each cell type. From this study, many differentially expressed genes were identified. The genes of previously reported salivary-flow- and secretion-associated proteins were consistent with the phenomena in aging, suggesting that aging may lead to significantly decreased gene expression of transcription factor, RNA process, protein/ion transport, signal intermediate, and enzyme. Also, gene expression associated with post-translational modification displayed a decrease, indicating the secretion of misfolded proteins in saliva with aging.
In conclusion, analysis of these data suggests that the function of SMG may be depressed by the aging process. Our results provided systematic analysis of the aging process in murine salivary glands. This study also demonstrates that microarray analysis can be a useful, sensitive, and quantitative technology for studying a complex and multigenetic phenomenon. In our ongoing investigation, human submandibular gland cell lines (HSG) have been differentiated in vitro into acinar cells, duct cells, and so on, so that the location of each gene expression can be studied by real-time PCR, in situ PCR, or in situ hybridization.
Scatter plot analysis of 1,328 genes.
RT-PCR analysis. Total RNAs from two-month-old (2M), 15-month-old (15M), and 23-month-old (23M) were reverse-transcribed. The resulting cDNAs were amplified with gene-specific primers. AA038677, Ca(2+)-transporting ATPase; U49393, sarcoendoplasmic reticulum Ca(2+) ATPase; AA024297, translocating chain-associating membrane protein; W62742, sec 61 alpha subunit; and X03672, mouse beta-actin, as a negative control. Each predicted PCR size is indicated.