Hypoxia‐inducible transcription factors, HIF1A and HIF2A, increase in aging mucosal tissues

Summary

Hypoxia (i.e. oxygen deprivation) activates the hypoxia‐signalling pathway, primarily via hypoxia‐inducible transcription factors (HIF) for numerous target genes, which mediate angiogenesis, metabolism and coagulation, among other processes to try to replenish tissues with blood and oxygen. Hypoxia signalling dysregulation also commonly occurs during chronic inflammation. We sampled gingival tissues from rhesus monkeys (Macaca mulatta; 3–25 years old) and total RNA was isolated for microarray analysis. HIF1A, HIF1B and HIF2A were significantly different in healthy aged tissues, and both HIF1A and HIF3A were positively correlated with aging. Beyond these transcription factor alterations, analysis of patterns of gene expression involved in hypoxic changes in tissues showed specific increases in metabolic pathway hypoxia‐inducible genes, whereas angiogenesis pathway gene changes were more variable in healthy aging tissues across the animals. With periodontitis, aging tissues showed decreases in metabolic gene expression related to carbohydrate/lipid utilization (GBE1, PGAP1, TPI1), energy metabolism and cell cycle regulation (IER3, CCNG2, PER1), with up‐regulation of transcription genes and cellular proliferation genes (FOS, EGR1, MET, JMJD6) that are hypoxia‐inducible. The potential clinical implications of these results are related to the epidemiological findings of increased susceptibility and expression of periodontitis with aging. More specifically the findings describe that hypoxic stress may exist in aging gingival tissues before documentation of clinical changes of periodontitis and, so, may provide an explanatory molecular risk factor for an elevated capacity of the tissues to express destructive processes in response to changes in the microbial biofilms characteristic of a more pathogenic microbial challenge.

Introduction

We have recently implemented an investigation using a non‐human primate model of periodontitis that has been demonstrated to have extensive similarities in clinical, microbiological and immunological features of human periodontitis. The model enables us to document characteristics of the transcriptome in gingival tissues, as a representative mucosal tissue, obtained from animals representing young individuals (approximately 10‐year‐old humans) to aged individuals (approximately 70‐ to 80‐year‐old humans).1, 2, 3, 4 We have shown significant differences in apoptosis pathway gene expression profiles associated with aging, even in healthy gingival tissues.2, 4 Differences were also noted in inflammasome gene pathways, including both receptors critical for signalling, and downstream effector functions,5 in antigen processing and presentation pathways,6 and differential expression of immune system genes.7

The human subgingival ecology has been shown to accommodate over 700 species of bacteria8 that differ both qualitatively and quantitatively in health, gingivitis and periodontitis.9 These difference have been documented to relate to the capacity of pioneer microorganisms to bind to host surfaces through receptor–ligand interactions,10 asaccharolytic species of bacteria that develop in inflamed sites related to a different nutritional microenvironment,11 anaerobic species that emerge at diseased sites presumably reflecting environmental changes of low oxygen levels, probably resulting from inflammation‐induced processes generating reactive oxygen intermediates,12 and species of bacteria that produce a range of volatile sulphur compounds.13 However, much of the research in this area has focused on the physiology and metabolism of the bacterial species related to their ability to survive and expand, as well as new knowledge of their genomes delineating unique pathways enabling them a competitive edge and probably contributing to the virulence potential of the pathogenic biofilms.14

Generally overlooked in these investigations is that the cellular response to oxygen is a central process in mammalian cells. Oxygen is required for aerobic energy metabolism and is required by the cells to produce adequate amounts of ATP necessary for metabolic activities. Hypoxia (i.e. oxygen deprivation) is an event in a diverse range of human tissues and cells with disease states including neoplastic development, ischaemia and chronic inflammation.15, 16 Organisms have evolved a range of adaptive mechanisms for hypoxic conditions including activation or repression of certain homeostatic genes that enable survival of tissues and cells. Specifically, low oxygen conditions activate the hypoxia signalling pathway, primarily via the hypoxia‐inducible factor‐1 (HIF‐1). The gene for HIF‐1 is among the primary genes involved in the homeostatic processes in hypoxia by increasing vascularization.17, 18, 19 It is also a transcription factor for numerous target genes, which mediate multiple biological functions, such as angiogenesis, haematopoiesis and the maintenance of vascular tone to provide or replenish tissues with blood and oxygen. HIF‐1 further mediates cellular responses to hypoxia by regulating glucose uptake and anaerobic respiration in oxygen‐depleted environments.17, 20

Hypoxia signalling dysregulation commonly occurs during chronic inflammation, at least in part through the activation and/or potentiation of nuclear factor‐κB (NF‐κB). Nuclear factor‐κB can be engaged through a classical pathway that is activated through the inhibitor of κB kinase (IKKα/β/γ) complex, and an alternative pathway that involves NF‐κB‐inducing kinase‐mediated activation of IKKα.21 Both sustained and intermittent hypoxia enhance NF‐κB activity using the classical pathway.22

The analyses in this report were designed to test the hypothesis that gingival tissues in aging animals demonstrate enhanced expression of genes in physiological pathways consistent with hypoxic stress in the tissues before a disease process. These changes might be anticipated to provide a risk for the tissues to express destructive processes in response to the microbial challenge, as occurs during periodontitis.

Methods

Non‐human primate model and oral clinical evaluation

Rhesus monkeys (Macaca mulatta) (n = 34; 14 female and 20 male) housed at the Caribbean Primate Research Center at Sabana Seca, Puerto Rico, were used in these studies. Healthy animals (five to seven per group) were distributed by age into four groups: ≤ 3 years (young), 3–7 years (adolescent), 12–16 years (adult) and 18–23 years (aged). Only adult (n = 5) and aged (n = 6) additional animals with periodontitis were used, as periodontitis does not occur naturally in younger animals. The non‐human primates were typically fed a 20% protein, 5% fat, and 10% fibre commercial monkey diet (diet 8773, Teklad NIB primate diet modified: Harlan Teklad). The diet was supplemented with fruits and vegetables, and water was provided ad libitum in an enclosed corral setting.

A protocol approved by the Institutional Animal Care and Use Committee of the University of Puerto Rico enabled anaesthetized animals to be examined for clinical measures of periodontal health including probing pocket depth and bleeding on probing as we have described previously.23 Periodontitis was defined as mean mouth values of probing pocket depth ≥ 3 mm and bleeding on probing ≥ 1.

Tissue sampling and gene expression microarray analysis

A buccal gingival sample from either healthy or periodontitis‐affected tissue from the premolar/molar maxillary region of each animal was taken using a standard gingivectomy technique, and maintained frozen in RNAlater solution. Total RNA was isolated from each gingival tissue using a standard procedure as we have described and tissue RNA samples were submitted to the microarray core to assess RNA quality and analyse the transcriptome using the GeneChip^® Rhesus Macaque Genome Array (Affymetrix, Santa Clara, CA).4, 24 Individual samples were used for gene expression analyses. Table 1 lists the hypoxia‐related gene set examined in this report.

Table 1.

Hypoxia pathway genes and associated functions

Gene ID	Fxn	Gene title
ARNT (HIF1B)	T	Aryl hydrocarbon receptor nuclear translocator
COPS5	T	COP9 constitutive photomorphogenic homologue subunit 5
EPAS1 (HIF2A)	T	Endothelial PAS domain protein 1
HIF1A	T	Hypoxia inducible factor 1, α subunit (basic helix‐loop‐helix transcription factor)
HIF1AN	T	Hypoxia inducible factor 1, α subunit inhibitor
HIF3A	T	Hypoxia inducible factor 3, α subunit
HNF4A	T	Hepatocyte nuclear factor 4 , α
NCOA1	T	Nuclear receptor coactivator 1‐like
PER1	T	Period homologue 1
APEX1	I	APEX nuclease (multifunctional DNA repair enzyme) 1
EGLN1	I	Egl nine homologue 1
EGLN2	I	Egl nine homologue 2
NFKB1	I	Nuclear factor of κ light polypeptide gene enhancer in B‐cells 1
P4HA1	I	Prolyl 4‐hydroxylase, α polypeptide I
P4HB	I	Protein disulphide‐isomerase‐like
TP53	I	Tumour protein p53
ADORA2B	A	Adenosine A2b receptor
ANGPTL4	A	Angiopoietin‐related protein 4
ANXA2	A,C	Annexin A2
BTG1	A,AP,CP	B‐cell translocation gene 1, anti‐proliferative
EGR1	A,CP	Early growth response 1
EDN1	A	Endothelin 1
EPO	A	Erythropoietin
F3	A,C	Coagulation factor 3, thromboplastin
GPI	A,M	Glucose‐6‐phosphate isomerase
HMOX1	A	Haem oxygenase (decycling) 1
HMOX2	A	Haem oxygenase (decycling) 2
JMJD6	A	Jumonji domain containing 6
LOX	A	Lysyl oxidase
MMP9	A	Matrix metallopeptidase 9 (gelatinase B)
PGF	A,CP	Placental growth factor
PLAU (uPA)	A,C	Plasminogen activator, urokinase
SEMA7A	A	Semaphorin 7A, GPI membrane anchor
SERPINE1 (PAI1)	A,C	Plasminogen activator inhibitor 1
VEGFA	A	Vascular endothelial growth factor A
ALDOA	C,M	Aldolase A
F10	C	Coagulation factor X
LIS1 (PAFAH1)	C	Platelet‐activating factor acetylhydrolase 1b
SLC16A3	C,TR	Solute carrier family 16, member 3
ATR	D	Ataxia telangiectasia and Rad3 related
MIF	D,AP,CP	Macrophage migration inhibitory factor
NDRG1	D	N‐myc downstream regulated 1
RUVBL2	D	RuvB‐like 2 (Escherichia coli)
DDIT4	M,AP	DNA‐damage‐inducible transcript 4
CP	M	Ceruloplasmin
ENO1	M	Enolase 1, (α)
CTSA	M	Cathepsin A
EROL1	M	ERO1‐like (Saccharomyces cerevisiae)
GBE1	M	Glucan (1,4‐α‐), branching enzyme 1
GYS1	M	Glycogen synthase 1 (muscle)
HK2	M	Hexokinase 2
LDHA	M	Lactate dehydrogenase A
PDK1	M	Pyruvate dehydrogenase kinase, isozyme 1
PFKFB3	M	6‐phosphofructo‐2‐kinase/fructose‐2,6‐biphosphatase 3
PRKFB4	M	6‐phosphofructo‐2‐kinase/fructose‐2,6‐biphosphatase 4
PFKP	M	Phosphofructokinase, platelet
PGAM1	M	Phosphate glycerate mutase 1
PGK1	M	Phosphoglycerate kinase 1
PKM2	M	Pyruvate kinase, muscle
SLC2A1	M,TR	Solute carrier family 2 (facilitated glucose transporter), member 1
SLC2A3	M,TR	Solute carrier family 2 (facilitated glucose transporter), member 3
TPI1	M	Similar to Triosephosphate isomerase (TIM)
ADM	AP,CP	Adrenomedullin
BNIP3	AP	BCL2/adenovirus E1B 19 000 interacting protein 3‐like
IER3	AP	Immediate early response 3
NOS3	AP,CP	Nitric oxide synthase 3 (endothelial cell)
PIM1	AP,CP	Pim‐1 oncogene
BLM	CP	Bloom syndrome, RecQ helicase‐like
CCNG2	CP	Cyclin G2
IGFBP3	CP	Insulin‐like growth factor binding protein 3
MET	CP	Met proto‐oncogene (hepatocyte growth factor receptor)
MXI1	CP	MAX interactor 1
NAMPT	CP	Nicotinamide phosphoribosyltransferase
ODC1	CP	Ornithine decarboxylase 1
TXNIP	CP	Thioredoxin interacting protein
BULHE40	CD	Basic helix‐loop‐helix family, member e40
FOS	CD	FBJ murine osteosarcoma viral oncogene homologue
RBPJ	CD	Recombination signal binding protein for immunoglobulin κ J region
USF2	CD	Upstream transcription factor 2
TRFC	TR	Transferrin receptor (p90, CD71)
VDAC1	TR	Similar to voltage‐dependent anion‐selective channel protein 1
ANKRD37	OR	Ankyrin repeat domain 37
CA9	OR	Carbonic anhydrase IX
DNAJC5	OR	DnaJ (Hsp40) homologue, subfamily C, member 5
EIF4EBP1	OR	Eukaryotic translation initiation factor 4E binding protein 1
LGALS3	OR	Lectin, galactoside‐binding, soluble, 3
MAP3K1	OR	Mitogen‐activated protein kinase kinase kinase 1

A, angiogenesis; AP, regulation of apoptosis; C, Coagulation; CD, cell differentiation; CP, regulation of cell proliferation; D, DNA damage and repair; I, HIF interactors; M, metabolism; OR, other response genes; T, transcription/co‐transcription factors; TR, transporters, channels, receptors.

Based upon the microarray outcomes we selected five genes and performed a quantitative PCR analysis using a standard technique in our laboratory employing a Roche 480 LightCycler.25 Quantitative PCR primers were designed using software primerquest at the Integrated DNA Technologies website (http://www.idtdna.com) and were synthesized by Integrated DNA Technologies, Inc. (Coralville, IA). Primers were prepared for LDHA (forward: AGATTCCAGTGTGCCTGTATG; reverse: ACCTCTTTCCACTGTTCCT TATC), PGAM1 (forward: CATCTGGAGGGTCTCTCTGAA; reverse: AACTGCA TGGGCTTGATAGG), PGK1 (forward: GATGATTATTGGTGGCGGAATG; reverse: GACAATCTTGGCTCCCTCTT), ENOL1 (forward: GAGGTTTACCACAACCTGAAGA; reverse: AGCTCCAGGCCTTCTTTATTC) and EPAS1 (forward: GAAGCGACAGC TGGAGTATG; reverse: TGAGGTTCTTCATCCGTTTCC) genes. The level of message was determined according to our previously published methods25 and those levels were compared across the RNA samples prepared from each of the healthy groups and the two periodontitis groups.

Data analysis

Normalization of values across the chips was accomplished through signal intensity standardization across each chip using the Affymetrix PLIER algorithm. The arrays contained matched and mismatched pairs allowing the MAS 5 algorithm to be used. For each gene we first determined differences in expression across the groups using analysis of variance (version 9.3, SAS Inc., Cary, NC). The healthy aged tissues were then compared with the other age groups using a t‐test and accepting a P‐value ≤ 0·05 for significance. Because of the cost of these types of non‐human primate experiments and the availability of primates of the various ages, we did not have sufficient samples to identify whether the relationship between age and gene expression could be treated using a linear model, so the subjects were classified and analysis of variance was used for analysis. The choice of least significant difference for multiple comparisons (analysis of variance followed by t‐tests) provided maximum power given our necessarily small sample sizes. We determined a correlation with aging in healthy tissues using a Spearman Rank correlation analysis that was fitted to the gene expression by age. A P‐value ≤ 0·05 was used to evaluate the significance of the correlation. Volcano plots were prepared to visualize outlier gene expression profiles in the aging animals compared with the other age group.26 JMP (version 10.0, SAS Inc.) was used to create metagenes independently of group classification using principal components based on the correlation matrix. The plots are of the first two principal components analysis scores across the healthy tissues. The variability is explained by each of the scores indicated on the plots. The data have been uploaded into the arrayexpress data base (http://www.ebi.ac.uk) under accession number: E‐MTAB‐1977.

Results

The results in Fig. 1 depict a principal components analysis of both healthy tissues from the four age groups, and a comparison of gene expression between healthy and periodontitis tissues in the adult and aged groups (young and adolescent animals do not exhibit naturally occurring periodontitis). The points denote a composite expression profile for all hypoxia genes for each individual animal. The results demonstrated a grouping of the aged healthy tissues with 26·1% of the variation in hypoxia gene expression related to the age of the animal (Fig. 1a). Similarly, the aged periodontitis group members (five of six) were clustered, with aging and disease accounting for 29·6% of the variation in hypoxia gene expression patterns (Fig. 1b). These results suggested that a profile of these genes related to the hypoxia pathway was indicative of healthy gingival tissues that were differentiated by aging, and that periodontitis significantly alters the environment contributing to the altered gene expression patterns.

Figure 1.

Principal components analysis of hypoxia gene expression in healthy gingival tissues from four age groups (a), or comparing expression in healthy to periodontitis tissues from adult and aged animals (b). Each point denotes the relative position using two principal component factors for each animal. The matching coloured triangles denote the ‘group means’ of the principal component factors.

Figure 2 provides a heat map of gene expression levels in the healthy and periodontitis tissues. The results depict a group of hypoxia pathway genes that were elevated in gingival tissues from healthy aged animals. These are identified by increases in Fold‐Δ in the figure and tended to cluster in transcription factors, angiogenesis/coagulation, metabolic (M) and other response genes. Expanding this analysis to the periodontitis tissues demonstrated an array of genes whose expression was even more elevated in periodontitis, compared with healthy tissues from similar age groups of animals. These genes clustered in the angiogenesis/coagulation, metabolic, cellular differentiation and other reponse gene groups.

Figure 2.

Heatmap of gene expression profiles in gingival tissues from different age groups and periodontitis tissues from adult and aged animals. The fold change in expression for each individual animal is based upon the median level in the adult healthy gingival tissues. Groups of genes are described as: T, transcription/co‐transcription factors; I, hypoxia‐inducible factor interactors; A,C, angiogenesis/coagulation; D, DNA damage and repair; M, metabolism; AP, regulation of apoptosis/regulation of cell proliferation; CD, cellular differentiation; TR, transporters, channels, receptors; and OR, other response genes.

Figure 3 provides Volcano plots of hypoxia gene expression patterns with a visualization of differential gene expression in the healthy aging tissues and how periodontitis alters the gingival tissue profiles. The points signify the level of expression of individual hypoxia pathway genes in aged healthy gingiva compared with levels expressed in the other age groups (Fig. 3a). Those genes significantly altered at least at P < 0·05 are provided with gene IDs. Points with a fold change < 0 are at lower levels in aged tissues and > 0 appear to be up‐regulated. As was predicted from the heat map of individual genes, in healthy aged tissues gene expression was altered. This was represented by some primary hypoxia regulatory factors, as well as those related to angiogenesis and metabolism. Interestingly, those genes most highly expressed in aging periodontitis tissues were generally unique compared with those altered in healthy aging, with substantially greater changes in expression levels versus those noted in healthy gingival tissues (Fig. 3b).

Figure 3.

Volcano plot of gene expression in healthy tissues of aging animals compared with other age groups (a) and comparison of aging periodontitis compared with all healthy age groups (b). The points denote each gene comparing the expression values as statistically different and as a fold difference (Fold Δ) in expression. The red dashed line denotes a P‐value of < 0·05.

Figure 4 focuses on primary biomolecules involved in sensing and response to low oxygen conditions, including the various hypoxia‐inducible transcription factors. The results demonstrate significant elevations in both HIF1A and HIF1B gene expression in the gingival tissues from aged animals, even though the tissues were deemed clinically healthy. Interestingly, HIF3A and HIF4A were also elevated, albeit, they did not reach statistical significance. In contrast, HIF2A (EPAS1) was significantly decreased in the aged gingival tissues. An additional analysis also evaluated up‐regulation of the 88 genes, focusing on gene expression that was at least twofold in aged animals compared with the median value of the adult group of animals. These altered genes included, c‐Fos, MMP9, HNF4A and CP (ceruloplasmin) (see Fig. 3).

Figure 4.

Expression of hypoxia transcription factor genes in aging healthy gingival tissues compared with levels in healthy gingival tissues from all other age groups. Bars denote the means for aged (AG, n = 6) and other age groups (ALL, n = 17). The vertical brackets enclose one SD. The asterisk (*) identifies a significant difference at least at P < 0·05.

Table 2 provides a summary of the genes that were altered and helped to characterize the age and disease status of the gingival tissues. Of the 88 genes examined, approximately 30% of the genes demonstrated some differences with aging. In health, aging gene expression differences or correlations were observed with various HIFs and HIF interactors (e.g. PH4B) (9 of 18), as well as numerous angiogenesis (13 of 20) and metabolic (13 of 28) genes.

Table 2.

Alterations in gene expression in healthy and periodontitis tissues

Gene ID	Hypoxia pathway function	Aging healtha	Correl health	Correl PD	Health versus PD	Aged PD versus All
ARNT (HIF1B)	HIF/co‐transcription	+			+
HIF1A	HIF/co‐transcription	+	+		+	+
HIF3A	HIF/co‐transcription		+
EPAS1 (HIF2A)	HIF/co‐transcription	+	−	−
PER1	HIF/co‐transcription					+
HNF4A	HIF/co‐transcription
EGLN2	HIF interactors			−	+
APEX1	HIF interactors		−
P4HB	HIF interactors	+	+
ADORA2B	Angiogenesis	+
EDN1	Angiogenesis	+
MMP9	Angiogenesis	+	+		+	+
HMOX1	Angiogenesis		−			+
HMOX2	Angiogenesis	+	+
ANXA2	Angiogenesis, coagulation	+	+
PLAU	Angiogenesis, coagulation		+		+	+
EGR1	Angiogenesis, cell proliferation		+			+
F10	Coagulation				+
LIS1 (PAFAH1)	Coagulation	+	−		+
SLC16A3	Transporter					+
ENO1	Metabolism	+				+
ERO1L	Metabolism		+
GYS1	Metabolism	+
LDHA	Metabolism	+	+		+
PGK1	Metabolism	+	+
GPI	Metabolism				+	+
TPI1	Metabolism				+
GBE1	Metabolism				+	+
DDIT4	Metabolism, apoptosis		+
TXNIP	Cell proliferation	+	−	−
MET	Cell proliferation				+	+
MXI1	Cell proliferation					+
ODC1	Cell proliferation					+
FOS	Apoptosis, transcription factors				+	+
MIF	DNA damage/repair, apoptosis	+
RUVBL2	DNA damage/repair			+		+
LGALS3	Response genes		+
CTSA	Response genes		+
DNAJC5	Response genes		+
EIF4EBP1	Response genes	+		+

^aAging Health denotes expression values in aged tissues compared with all other age groups; Correl Health denotes correlation in gene expression in healthy tissues across all age groups; Correl PD denotes correlation in gene expression in periodontitis tissues with age; Health versus PD denotes comparison of expression values in healthy adult and aged tissues with adult and aged periodontitis tissues; and Aged PD versus All denotes comparison of expression values in aged periodontitis tissues with healthy tissues from non‐aged groups. (+) denotes significant difference (t‐test at P < 0·05 or lower), or significant positive correlation (Spearman Rank Correlation, P < 0·05 or lower); (−) denotes significant negative correlation.

Table 3 provides a summary of the fold‐difference in selected gene expression levels (LDHA, PGAM1, PGK1, ENOL1, EPAS1) determined using the microarray analysis compared with results from a quantitative PCR analysis. The results of this comparison identified similar directions for the altered gene expression, although the magnitude varied between these two independent analyses.

Table 3.

Comparison of gene expression profiles using quantitative PCR and microarray analyses. Values represent fold‐difference compared with Adult Healthy tissue message levels assigned a value of 1·0. Perio denotes periodontitis tissues from aged animals

Gene ID	Young	Adolescent	Aged	Perio
LDHA
qPCR	1·33	1·58	1·47	1·44
GeneChip	1·19	1·44	1·24	1·02
PGAM1
qPCR	1·24	1·65	1·96	1·54
GeneChip	1·05	1·38	1·78	1·06
PGK1
qPCR	1·93	2·23	2·03	1·94
GeneChip	1·11	1·01	1·25	1·08
ENOL1
qPCR	1·56	1·38	1·60	1·71
GeneChip	1·46	1·28	1·79	1·06
EPAS1
qPCR	3·38	1·54	1·69	0·81
GeneChip	1·96	1·57	0·81	1·17

Discussion

Periodontal disease manifests as a persistent inflammatory response of the local tissues that has been suggested to reflect changes in the characteristics of the subgingival microbial ecology at diseased sites.27, 28, 29 Additional findings in studies of periodontitis report the increased frequency and severity of disease with aging,30, 31, 32 leading to the consideration that periodontitis is a disease of aging related to altered immune functions that occur with increasing prevalence coincident with decades of life in the general population,33 or potentially a reflection of changing oral environments that select for a microbial ecology with greater pathogenic potential.

Using a non‐human primate oral model to explore the responses of mucosal tissues constantly exposed to a polybacterial challenge of mucosal tissues with responses in these tissues we have shown significant differences in apoptosis pathway gene expression profiles associated with aging, even in healthy gingival tissues.2, 4 This report evaluated gene expression patterns that would reflect changes in the gingival microenvironment in healthy gingival tissues of aging non‐human primates, to address aging effects on this milieu that could presage susceptibility to tissue destruction, a hallmark of periodontitis. We specifically targeted gene expression patterns related to hypoxia, as genes in this pathway could reflect alterations in the subgingival area that could enhance the emergence of pathogenic anaerobic bacteria in the ecology, as well as the recognition that chronically inflamed tissues are increasingly hypoxic.17, 34, 35, 36 The results demonstrated that sets of hypoxia pathway genes are altered in healthy tissues from aged animals. These included all of the major HIF transcription factors [HIF1A, HIF1B (ARNT), HIF2A (EPAS1), HIF3A] that are critical for altering gene expression profiles in cells under hypoxic stress.17, 37 Additionally, at least one member of the PHD (prolyl hydroxylase)–HIF signalling pathway, i.e. P4HB was increased in aged healthy tissues. This gene produces a protein that senses the external hypoxic environment and hydroxlates and activates the HIF molecules enabling nuclear translocation.38, 39, 40 These findings support that clinically healthy gingival tissues from aging individuals demonstrated a response to an altered gingival microenvironment with lower oxygen availability, without gross inflammation. One interpretation of these findings is that the aging gingival tissues, although clinically healthy, may be reflecting a profile of gene expression alterations that are indicative of enhanced tissue susceptibility to bacteria that emerge in more anaerobic pathogenic biofilms. Alternatively, these changes could reflect altered characteristics of the microbial biofilms in the aging animals that continually stress the tissues contributing to increased risk for disease initiation and progression. This is consistent with recent suggestions that proposed periodontopathogens, in particular Porphyromonas gingivalis, have metabolic characteristics that enable the pathogen to propel the development of more pathogenic biofilms in which the commensal bacteria may play a role in inflammation and tissue destruction.41 Additionally, these transcription factors were further increased in periodontitis versus healthy gingival tissues.

Beyond the more direct reflection of cellular responses to hypoxic stress, we identified a substantial array of altered gene expression associated with metabolic changes in the cells, as well as a pattern of genes consistent with increases in angiogenesis activities consistent with downstream signalling of the hypoxia pathway. These processes are coordinated by HIF‐1 in concert with regulation by the Von Hippel–Lindau tumour suppressor protein (pVHL).42, 43 Under hypoxic conditions the HIF‐1α subunit in the cytoplasm is not recognized by the pVHL protein, allowing it to accumulate and form a heterodimer with HIF‐1β enabling translocation to the nucleus. Here the active HIF‐1 interacts with a range of cofactors, including CBP (CREB Binding Protein) and the Pol II (DNA polymerase II) complex to bind to hypoxia‐responsive elements that activate transcription of various target genes.19, 38, 44, 45 In the presence of oxygen, HIF is destroyed in the cytoplasm by ubiquitination through a ubiquitin ligase.42, 46, 47, 48 HIF‐1 is a transcription factor that transactivates genes encoding proteins controlling metabolism, cell proliferation and vascularization [e.g. erythropoietin, lactate dehydrogenase, vascular endothelial growth factor (VEGF)].43, 49, 50, 51 Overexpression of HIF‐1 in vivo results in increased localized inflammation, whereas loss of the HIF‐1 gene decreases immune cell functions, in these tissues related to their physiological anaerobic respiration for energy production.34, 52, 53 Hence, an inability to produce HIF‐1 minimizes emigration to injured tissues and destruction of infectious microbes.54, 55, 56, 57 HIF‐1 also impacts the differentiation of haematopoietic cells into monocytes and macrophages.52

An important function of the HIF‐1 pathway is to promote angiogenesis and so increase oxygenation of the hypoxic tissues. The HIF‐1 transcription factor activates genes to direct migration of mature endothelial cells toward a hypoxic environment through VEGF transcription.58, 59, 60 These endothelial cells ultimately help to form new vessels, enhancing delivery of oxygenated blood to the hypoxic tissues.61 Additionally, HIF‐1 activates other genes related to angiogenesis including GLUT1 (for glucose transporter‐1), which activates glucose transport; and EPO (for erythropoietin), which induces erythropoiesis; activation of transcription of nitric oxide synthase also promotes angiogenesis and vasodilatation. ARNT (HIF1B) is another protein that forms heterodimers with HIF‐1α to activate other angiogenesis genes with some tissue specificity.46 Our findings demonstrated a range of angiogenesis genes that were significantly altered or correlated with aging in healthy gingival tissues. These include, as examples, HMOX2, (haem oxygenase‐2 isozyme) a constitutive enzyme essential in haem catabolism,62 MMP9 (matrix metalloprotease 9), which degrades type IV and V collagens and enhances release of VEGF,63 EDN1 (endothelin 1) as a potent vasoconstrictor produced by vascular endothelial cells,64 PLAU (urokinase‐type plasminogen activator) a serine protease involved in degradation of extracellular matrix and converts plasminogen to plasmin,65 and ANXA2 (annexin 2), which is involved in diverse cellular processes, such as cell motility and fibrinolysis.66 Fewer of the angiogenic genes were further up‐regulated with periodontitis in the aged animals.

HIF‐1 activation can also regulate cellular anaerobic metabolism. In the presence of oxygen, most cells produce ATP via oxidative phosphorylation. However, under hypoxic conditions anaerobic metabolism is engaged to provide cellular energy production. HIF‐1 is a crucial factor in this metabolic shift. Nuclear translocation induces a variety of glycolytic enzymes and glucose transporters [e.g. ENO1 (enolase 1), PGK1 (phosphoglycerate kinase 1), LDHA (lactate dehydrogenase), GYS1 (glycogen synthase)], which support efficient energy production in a hypoxic environment.58, 67 Beyond increasing the expression of these enzymes to generate energy, elevated HIF‐1 can decrease mitochondrial oxygen consumption [e.g. ERO1L (ERO1‐like protein α, oxidoreductase), PDK1 (pyruvate dehydrogenase kinase isozyme 1)] as was noted in samples from the aged animals. Within those genes related to altered metabolism under hypoxic conditions we also observed an increased level of CP (ceruloplasmin) in the aged healthy tissues. This acute‐phase reactant is the major serum ferrioxidase enzyme that converts toxic ferrous iron (Fe²⁺) to ferric iron (Fe³⁺) in the presence of oxygen.68 Hence, up‐regulation of this protein enhances protection of cells from toxic iron accumulation.

Altered expression of DDIT4 (DNA‐damage‐inducible transcript 4, REDDI) interfaces the metabolic changes with the apoptosis pathway.69 These changes observed in both healthy aged and periodontitis tissues were also supported by alterations in expression of TXNIP (thioredoxin‐interacting protein); c‐MET (hepatocyte growth factor receptor, proto‐oncogene); c‐FOS (proto‐oncogene); MIF (macrophage migration inhibition factor) that can protect from redox‐stress‐induced apoptosis; and, RUVBL2 (RuvB‐like 2) for DNA repair, as examples.

Within the intracellular hypoxia pathway HIF‐1α can be regulated by extracellular signal‐regulated kinase 2 (i.e. MAPK1) as a biochemical signal transducer in the cells, which phosphorylate HIF‐1α.70 HIF‐1α can associate with heat‐shock protein 90, which appears to enhance HIF‐1α transcriptional activity, increasing both hypoxia‐induced accumulation of VEGF and hypoxia‐dependent angiogenesis.42 More recently, a factor inhibiting HIF‐1α activation, FIH or HIF1AN (hypoxia‐inducible factor 1α inhibitor) has been described that provides another level of regulation of this important transcription factor, although it was not altered in the aging tissues.71

In periodontitis, the local availability of oxygen and consumption by gingival tissues has been suggested to be decreased, reflecting the chronic inflammatory process, and an environment that appears to select for more anaerobic bacteria. The microbial ecology at disease sites also appears to physiologically confer a low oxygen tension in juxtaposition to the periodontal tissue lesion. This oxygen shortage would be predicted to lead to stabilization of HIF‐1α transcription factor and to control specific downstream genes that modulate varied cellular functions that influence the process of periodontitis. In this regard, HIF‐1α has been shown to be up‐regulated in tissues from periodontal pockets.72 Ng et al.73 demonstrated HIF1A‐containing nuclei in both epithelial and endothelial cells in gingival tissues accompanied by clear increases in this gene expression in fibroblasts and leucocytes of periodontitis lesions. They also identified increased levels of HIF‐1α, VEGF and tumour necrosis factor‐α proteins in the diseased tissues. Extending these observations are an array of in vitro studies that demonstrate the ability of P. gingivalis lipopolysaccharide under hypoxic conditions to alter the responses of periodontal ligament cells related to levels of VEGF, interleukin‐1β (IL‐1β), and MMP‐1. These alterations occurred through effects on NF‐κB and HIF‐1α activation in the cells.74 Hypoxia also augmented P. gingivalis lipopolysaccharide induction of tumour necrosis factor‐α, IL‐1β and IL‐6 expression by periodontal ligament cells,75 all markers of tissue destructive inflammation and apparently functioning through TLR4 engagement.76 Interestingly, HIF‐1α is up‐regulated by IL‐1β.77 Hypoxia was shown to regulate periodontal ligament stem cell functions by altering their osteogenic potential and mineralization capacity through alteration of ERK1/2 and p38 kinase activities78 and during culture of peripheral blood mononuclear cells and osteoblasts it lead to the formation of functional osteoclasts following up‐regulation of HIF, VEGF and RANKL.79 Similarly, HIF‐2α, a key regulator of cartilage destruction, was up‐regulated in periodontal ligament cells from diseased tissues. In vitro, HIF‐2α appeared to contribute to nicotine and lipopolysaccharide induction of inflammatory mediators through multiple transcription factors.80 We also observed altered ceruloplasmin mRNA levels. As ceruloplasmin is induced by hypoxia and inflammation, Iwata et al.81 noted that polymorphonuclear cells from patients with aggressive periodontitis were primed for production of ceruloplasmin that would enhance hypoxia‐mediated O₂ generation and increased oxidative stress.

Hence, an interpretation of the results of this study suggests that gingival tissue characteristics in aging are significantly different even in health. Hypoxia gene pathways appeared to be up‐regulated in healthy gingival tissues of aging non‐human primates. Major characteristics of the gene changes in health were related to angiogenesis and metabolism responses to hypoxia, in addition to the primary hypoxia transcription genes. Generally, hypoxia pathway gene expression profiles in both adult and aged periodontitis tissues were substantively different from those that were noted to change in healthy tissues as a reflection of an aging process. These molecular alterations may presage mucosal tissues in the aging individual to an enhanced risk of dysregulated responses, breakdown in homeostasis, and subsequent destruction from the chronic periodontal infections. The changes in gingival tissues with aging may signal a predisposition to a subgingival environment that can enrich for anaerobic species. It remains unclear whether a hypoxic/anaerobic tissue environment, particularly with inflammation, selects for characteristic pathogenic biofilms, or members of these biofilms ‘trigger’ these hypoxic changes.

Disclosures

The authors claim no conflict of interest regarding the conduct of the study or any results presented in the report.