Kotoku Kurachi, Age Dimension Research Center, National Institute of Advanced Industrial Science and Technology, 1-1-1, Higashi, Tsukuba, Ibaraki, 305–8566, Japan. Tel.: +81 298 61 6528; fax: +81 298 61 2788; e-mail: firstname.lastname@example.org
Age is a critical risk factor for many diseases such as thrombosis, cardiovascular diseases, cerebral vascular diseases, diabetes, cancer and Alzheimer's disease [1–10]. Although its importance has been shown clearly by many epidemiological studies, literally nothing is known about its molecular mechanisms of action.
Physiological systems are controlled by homeostatic mechanisms, maintaining them within small tolerable fluctuation ranges. Various internal or external insults, stresses or pathological conditions affect the homeostatic conditions. Importantly, homeostatic conditions of physiological systems are not constant with age, but continue changing slowly with age [11–13]. Such physiological changes may play a critical role either additively or synergistically in the initiation and development of many age-dependent diseases.
To explore the little-understood molecular mechanisms of age-dimension homeostasis, we first focused our study on the blood coagulation system, our model physiological system for the study (Fig. 1). This led to the first discovery of the molecular mechanisms of age-related gene regulation and therefore of homeostatic regulation of physiological systems. In this article, we review recent advances in this newly emerging research field.
Blood coagulation system and age
The complex formation of factor VII with tissue factor, which becomes available upon tissue injury, initiates blood coagulation  (Fig. 1). This initial reaction triggers cascades of reactions involving nearly 20 pro- and anticoagulant factors, resulting in a sufficient amount of stable fibrin clot production in a matter of minutes. Most cascade steps of this system are of proteolytic reactions in combination with increasing levels of plasma concentration of procoagulant factors as they go downstream, thus making blood coagulation an extremely efficient amplification system. In healthy individuals, blood coagulation activity makes a rapid increase during the perinatal stage and reaches a level similar to that of young adults at around weaning, followed by a gradual and continuous increase with age [15–19]. As demonstrated by epidemiological evidence [11,17,18,20], this appears to be due to an age-dependent increase in an imbalance between procoagulant activity and anticoagulant activities. Plasma concentration and/or activity of most known procoagulant factors (shown with arrows in Fig. 1) increase with age, whereas anticoagulant factors such as antithrombin (ATIII), TFPI and protein C, and profibrinolytic factors including plasminogen and tPA do not increase with age or, in some cases, even decrease slightly with age [20–22]. Importantly, PAI-1, an inhibitor of tPA which catalyzes plasminogen activation to plasmin, also increases its plasma concentration with age , thus resulting in an enhancement of procoagulant tendency. Together, this evidence indicates that as age proceeds, the balance between the overall procoagulant activity including the antifibrinolytic activity and that of the anticoagulant activity combined with the fibrinolytic activity tips toward the procoagulant state. In combination with various environmental conditions, including diet consumed, such an imbalance may contribute to an increased thrombotic tendency, particularly in elderly people [16–21,23,24].
Circulatory levels of human factor IX (hFIX), a key coagulation factor, significantly increase with advancing age in normal human populations [11,16,25]. In contrast, circulatory levels of human PC (hPC), a key factor in the potent anticoagulant protein C pathway, show a stable pattern with only marginal, if any, age-associated fluctuations [11,16,21]. Therefore, we considered that our studies on the age-related regulatory mechanisms of hFIX and hPC genes may provide us with critical information on the fundamental molecular mechanisms involved in the age-dimension homeostasis of gene expression.
Natural gene sizes of hFIX and hPC are approximately 40 and 13 kb pairs, respectively. To make experimental manipulations feasible, we constructed a series of their minigenes [26–31] (Fig. 2). After verifying with HepG2 cells for appropriate construction of these minigene vectors and their expression activities, all the minigene vectors were subjected to systematic construction of transgenic mice. Circulatory levels of hFIX or hPC produced in individual transgenic animals were then monitored biweekly or monthly as needed for their life spans using serum samples and hFIX or hPC-specific enzyme-linked immunosorbent assays (ELISA).
Characteristics of the mouse blood coagulation system are similar to those of the human system, and we showed that age-related expression patterns of FIX and PC genes, age-related increase or stable, respectively, are also similar to those of the human counterparts [26,27]. This similarity justified mice as our animal model for analyzing the age-dimension homeostasis of gene expression.
Discovery of the first age-related regulatory mechanisms of gene expression
After systematic analyses of many lines of hFIX minigenes in transgenic mice, we identified two genetic elements, ASE and AIE (renamed from the initial names, AE5′- and AE3′, respectively; see [26,27]) to be essential for generating age-related stable and increase patterns, respectively, of hFIX gene expression  (Fig. 3). ASE is located to a small 5′ upstream region, nucleotide (nt) −802 to nt −784, of the hFIX gene sequence by footprinting assay and electrophoretic mobility shift assay (EMSA). ASE has a nucleotide core sequence, GAGGAAG, matching the Ets element consensus sequence (GGAAT) for the binding of Ets family transcriptional factors such as PEA-3 [32,33]. AIE, an element necessary for age-related increase in gene expression, was identified in the middle of the hFIX 3′-untranslated region (UTR) . AIE is composed of a 102 base pair (bp) dinucleotide repeat (mostly AT, GT, CA), and has the potential to form three distinct stem loop structures in its RNA form . Mouse factor IX (mFIX) shows an age-related increase in both its plasma level and gene expression level in the liver [34,35], and its gene contains an ASE sequence identical to the functional ASE in the hFIX gene and an AIE-like stretch of 106 bp dinucleotide repeats in the mFIX 3′-UTR . Through testing in transgenic mice, we showed that the mFIX repeat actually functions as an active AIE . These studies demonstrated that the age-related increase in expression of the FIX gene is regulated by a combination of two unique genetic elements, ASE and AIE (Fig. 3). In the absence of ASE, minigenes with or without AIE show an age-unstable hFIX gene expression pattern with a rapid decline in expression over the puberty period, and in the subsequent 3–4 months decline to low but stable levels or very low basal levels, respectively. In the presence of both ASE and AIE, the age-related expression increase pattern of the hFIX gene is reproduced. Importantly, the clearance rate (t½) of human factor IX from the blood circulation does not change significantly over age [26,27]. This is the first discovery of the age-related regulatory mechanisms of gene expression.
What molecular mechanisms function in regulation of age-stable hPC gene expression? As mentioned above, we first confirmed that the mouse PC gene expression pattern is age-stable, similar to that of the hPC gene . hPC and hFIX share significant similarities in protein structure and therefore coding regions of their genes, whereas the 5′-flanking regions and 3′-UTRs are grossly dissimilar, indicating unrelated evolutional origins [28,30,37,38]. For example, we have shown previously that the 5′-flanking region of the hFIX gene beyond approximately nt −350, including the region containing ASE, was derived from the retrotransposable element LINE-1 inserted into the 5′ UTR  (Fig. 2). In contrast, the 5′-flanking region of the hPC gene has no LINE-1 or its remnant sequences, indicating that no retrotransposition events took place. In addition, 3′-UTR sequences of hFIX and hPC genes share no similarity except the minimal local sequences required for polyadenylation at the 3′ end regions. The 3′-UTR sequence of the hFIX gene is about 1.4 kb in length, whereas that of the hPC gene is only 295 bp, and has no AIE or AIE-like element or dinucleotide repeats [28,30]. These differences between the hFIX and hPC genes correlated with the dissimilarities in their age-related expression patterns as well as tissue-specific expression patterns.
As shown in Fig. 4, age-related stable patterns of circulatory hPC were observed with transgenic mice carrying hPC minigene −1462hPCm1 or −849hPCm1, recapitulating the natural age-stable expression patterns of mPC and hPC genes .
These minigenes had 5′ flanking sequences where their promoter regions are contained, extending to either nt −1462 or −849, respectively. However, expression patterns of minigene −802hPCm1 were age-unstable, thus identifying CAGGAAG, present in the region spanning nt −832 to nt −826, as the functional hPC ASE responsible for the age-stable expression pattern of the hPC gene. The functional ASE sequences of both hPC and hFIX genes bind the same liver nuclear protein, but other sequences do not function even with their close similarity, as shown by EMSA DNA-protein binding assays . It is important to emphasize that hFIX ASE and hPC ASE have independent evolutionary origins. hFIX ASE was derived through mutational changes of a LINE-1 sequence inserted originally by a retrotransposition, while the hPC gene never experienced LINE-1 retrotransposition, and its ASE was derived through a different mechanism. This is a case of function-driven convergent evolution generating a critical genetic element required for homeostatic regulation of genes.
Pseudo-ASE (GAGGAAA) present at the 5′ upstream close to the functional hPC ASE binds a nuclear protein different from that bound to the functional ASEs, and it does not function as an age-related regulatory element . Furthermore, two other ASE-like known Ets consensus elements, one in the first intron (GAGGATG) and the other in the last exon (CAGGATG), bind an identical nuclear protein, but again different from the protein which binds to the functional ASEs (G/CAGGAAG). These elements also do not function for age-related gene regulation. Together, these results indicate that specific single base differences in the motif G/CAGGAA/TG/A facilitate strictly selective binding of specific Ets family nuclear proteins, thus conferring distinctly different functions. As we tested in transgenic mice, hFIX ASE could functionally substitute hPC ASE in the regulation of hPC gene expression . ASE functions through binding a unique, probably Ets, family transcription factor. Its identification and characterization studies are in progress.
The hPC gene does not have any AIE-like element. However, as demonstrated by animal testing with minigene −1462hPCm1/AIE, which contains a unit of hFIX AIE, the age-stable expression pattern of hPC was dramatically changed to an age-related increase pattern, similar to that of the hFIX gene . The mechanism of action of AIE remains to be established. These findings, together with other observations, support the functional universality of ASE and AIE. Animal testing of the mouse gene counterpart of the hFIX AIE in combination with hPC ASE has proved its functionality. Thus, the molecular mechanisms involving ASE and AIE function across different species, at least within human and mouse physiologies.
Through differential or combined usages of ASE and AIE most, if not all, anti- or procoagulation factor genes may achieve age-related increase or stable gene expression patterns, respectively. Our observations that there are no significant differences in the rate of protein clearance from the circulation between young and old animals, either for hFIX or hPC, support this further [26,27]. Together, these observations explain how the age-related increase pattern of the overall blood coagulation activity is generated (Fig. 5). Nature may also use the mechanisms for many other genes involved in different physiological systems for their age-related regulation. Further studies in this regard are under way.
Development of age-dimension technology
New knowledge of the molecular mechanisms of age-dimension homeostasis allows us to develop a new research field, age-dimension technology (ADT). Through manipulating ASE and AIE activities, for instance, by using new drugs designed to affect the functions of ASE and/or AIE, we may be able to modify gene expression along the age axis, thus leading to the development of novel preventive and therapeutic methods for age-related diseases. Some anticipated examples of ADT applications may include the following.
One of the challenging difficulties facing the current efforts in developing effective and safe gene therapy approaches is how to secure a long-term effective gene expression in vivo after the initial single procedure of gene delivery without repeating it. ADT, which allows manipulation of expression vectors with ASE and AIE, has the great potential to make it feasible, generating optimal gene delivery systems.
Another example of ADT application is described as follows. During our studies involving a large number of transgenic mice, we found that animals overexpressing hFIX at higher than certain levels (approximately 1500 ng mL−1 serum) die at young ages as early as 3–5 months, while control animals or animals expressing hFIX at low levels (at less than 200 ng mL−1 levels) live a normal life span as expected  (Fig. 6). In these transgenic animals, hFIX produced was additive to the mouse intrinsic FIX.
Analyses of animals dying at much younger ages than the expected age showed thrombi present in blood vessels of various tissues including the brain, lungs and heart. Some animals developed myocardial fibrosis in the left ventricle presumably induced by thrombotic occlusions of blood vessels, thus mimicking a type of human myocardial infarction. These findings indicated that even a relatively small elevation in the circulatory level of FIX, which occupies a position in the middle phase of the blood coagulation cascade, significantly shifts the blood coagulation–anticoagulation balance toward prothrombosis. Such a thrombotic tendency, however, may be eliminated effectively by suppressing the AIE function, for instance, by designing a drug antagonistic to AIE, while keeping the ASE function intact, ensuring the needed coagulation activity.
In summary, we have discovered the very first molecular mechanisms of age-related gene regulation, explaining the basic mechanisms of the age-dimension homeostasis of the blood coagulation system. This has also opened the door for discovering novel molecular mechanisms in action for age-related homeostasis of many other physiological systems.