Accelerated bio‐cognitive aging in Down syndrome: State of the art and possible deceleration strategies

Abstract Down syndrome (DS) has been proposed by George Martin as a segmental progeroid syndrome since 1978. In fact, DS persons suffer from several age‐associated disorders much earlier than euploid persons. Furthermore, a series of recent studies have found that DS persons display elevated levels of age biomarkers, thus supporting the notion that DS is a progeroid trait. Nowadays, due to the progressive advancements in social inclusion processes and medical assistance, DS persons live much longer than in the past; therefore, the early‐onset health problems of these persons are becoming an urgent and largely unmet social and medical burden. In particular, the most important ailment of DS persons is the accelerated cognitive decline that starts when they reach about 40 years of age. This decline can be at least in part counteracted by multi‐systemic approaches including early‐onset cognitive training, physical activity, and psychosocial assistance. However, no pharmacological treatment is approved to counteract this decline. According to the most advanced conceptualization of Geroscience, tackling the molecular mechanisms underpinning the aging process should be a smart/feasible strategy to combat and/or delay the great majority of age‐related diseases, including cognitive decline. We think that a debate is needed urgently on if (and how) this strategy could be integrated in protocols to face DS‐associated dementia and overall unhealthy aging. In particular we propose that, on the basis of data obtained in different clinical settings, metformin is a promising candidate that could be exploited to counteract cognitive decline in DS.

increase in the near future (Bittles & Glasson, 2004;Glasson et al., 2002). It is worth noting that such extension of the lifespan was of greater amplitude than that occurring in the same years in the general population, suggesting that the extremely short life expectancy of DS persons in the past was not only due to the absence of effective medical treatments but also due to sociocultural factors. In fact, the absence of biological knowledge and social repulsion pushed in the past the medical community to consider the extremely short life expectancy and poor health of these persons as inherent, nonmodifiable traits of the syndrome. On the contrary, when these persons are properly attended, not only does their life expectancy increase spectacularly, but also the alterations at molecular and cellular levels (e.g., immune alterations) result as less dramatic than expected (Franceschi et al., 1978). In the next sections, we will briefly summarize current knowledge for each of these layers. Starting from this evidence, we will discuss the possible strategies to hinder this accelerated aging, especially for cognitive functions, with a special focus on metformin as a candidate molecule.

| CLINICOPATHOLOGICAL FEATURES OF ACCELERATED AGING IN DS
Accelerated aging in DS is evident in particular at the level of the central nervous system, and by the age of about 50 years, most of DS persons suffer from an early-onset Alzheimer-like dementia.
Actually, dementia is the second more frequent medical condition after visual impairment present in adult DS persons (Capone et al., 2018), and by far the most relevant health problem, as it entails a loss of independence, a dramatic decrease in the quality of life, and it represents a risk factor of mortality, together with mobility restrictions and epilepsy (Coppus et al., 2008). Typical neuropathological hallmarks of Alzheimer's disease (AD), including deposition of senile plaques containing amyloid beta-peptide (Aβ), chronic oxidative stress, and neurofibrillary tangles composed of hyperphosphorylated tau protein are already present by the age of 40 years in DS persons (Lott, 2012). According to a recent study, the diagnosis of dementia is posed on average at 55.8 years (SD ±6.29) and median survival time after diagnosis is 3.78 years, with some differences between males and females (Sinai et al., 2018). The gene for amyloid precursor protein (APP) is triplicated in DS, being present in chromosome 21. Consequently, DS dementia is believed to be related to the overproduction of APP and consequent deposition of amyloid rather than to a reduced clearance of Aβ, as it occurs in sporadic AD. Therefore, DS dementia is considered more similar to familiar than sporadic forms of AD (SAD). This tenet is also supported by consistent differences between DS and SAD including Aβ deposition patterns, characterized by an increase in deposition of Aβ42 and diffused subcortical plaques in DS (Fukuoka, Fujita, & Ito, 1990;Kida, Wisniewski, & Wisniewski, 1995), and neurological symptoms such as seizures and incontinence, which tend to emerge earlier in DS (Zis & Strydom, 2018). Finally, from a clinical point of view, DS persons appear to be somewhat protected from cerebral amyloid angiopathy and consequent risk of hemorrhage and stroke (Zis & Strydom, 2018). When comparing DS dementia with AD forms due to APP gene microduplication, other differences in the clinical phenotype have been reported (Zis & Strydom, 2018). Despite these important differences between DS dementia and AD, typical signs of cognitive decline and neuropsychological features of dementia (lowered scores of battery tests for language and short memory skills, frontal lobe functions, visuospatial abilities, and adaptive behavior) appear early in DS persons (Arvio & Luostarinen, 2016;Ghezzo et al., 2014).
The immune system is also rapidly deteriorating in DS. It is known that adult DS persons display a series of early-onset changes that apparently recapitulate in a shorter timescale the normal aging process of the immune system, such as diminished NK cell activity, decreased number of T and B lymphocytes, erosion of telomeres in T lymphocytes, decreased response to mitogenic stimuli of blood leukocytes, and increased risk of autoimmune disorders, among others (Cuadrado & Barrena, 1996;Kusters, Verstegen, & Vries, 2011).
Many other age-associated diseases display an early onset in DS persons, including thyroid disorders (Prasher, 1999), osteopenia and

Key points
Down syndrome is the most common genetic cause of intellectual disability and appears to be characterized by an accelerated aging, affecting in particular the central nervous system, with a number of features that overlap with Alzheimer's disease.
Social inclusion and cognitive training programs have greatly improved DS cognitive performances over the last decades; however, this advancement is largely frustrated by an early-onset of an Alzheimer-like dementia.
No approved treatment exists so far to counteract such peculiar cognitive condition. A number of clinical trials have been performed but with unsatisfactory results.
Based on literature data, we propose that the use of metformin could be worth investigating.
In many cases, these pathologies occur simultaneously (Capone et al., 2018;Kinnear et al., 2018). A possible exception to this rule is the incidence of solid cancer. Actually, when testis and perhaps other male genital cancers and stomach cancers are excluded, the incidence of solid tumors in adult DS persons is markedly lower with respect to age-matched controls (Nižetić & Groet, 2012). This finding is paradoxical if we consider that DS has many factors predisposing to cancerogenesis, including chromosome instability, increased DNA damage, and defective DNA repair systems, as well as the presence on chromosome 21 of a number of oncogenes (Nižetić & Groet, 2012). This paradox likely reflects our lack of knowledge regarding anti-oncogenic mechanisms related to the trisomic condition and will not be further discussed here.

HIGH LE VELS OF MARKERS FOR BIOLOGICAL AGE
A biomarker of age is defined as a biological parameter (or a combination of biological parameters) able to predict age-related functional decline and lifespan better than chronological age (Baker & Sprott, 1988;Butler et al., 2004;Johnson, 2006). Different potential biomarkers of age have been proposed so far, and many of them have been shown to grasp age-acceleration effects (i.e., higher biological age), including cognitive decline and AD (Jylhävä, Pedersen, & Hägg, 2017).
So far, four different biomarkers of age (namely telomere shortening, GlycoAgeTest, Horvath's epigenetic clock, and brain predicted age) have been analyzed in DS persons. As described in the following paragraphs, the results of these independent studies concordantly point for an age-acceleration effect in DS, indicating that DS persons are biologically older than expected from their chronological age.
Telomere attrition is a well-established hallmark of aging. Despite some conflicting results (Sanders & Newman, 2013), telomere shortening has been generally associated with age-related physical/functional decline and with mortality. It has been demonstrated that DS persons have shorter telomeres than age-matched controls (Vaziri et al., 1993). Subsequent studies showed that telomere attrition is associated with development of dementia in DS persons (Jenkins et al., 2006(Jenkins et al., , 2010, also in longitudinally assessed cohorts (Jenkins et al., 2017), resembling the telomere shortening observed in AD patients (Forero et al., 2016).
The GlycoAgeTest derives from the relative amounts of two plasma N-glycans, measured by means of a DNA sequencer-aided, fluorophore-assisted carbohydrate electrophoresis (DSA-FACE) technique. This biomarker reliably increases with age after 40 years (Dall'Olio et al., 2013;Vanhooren et al., 2009Vanhooren et al., , 2007, and it is likely to be indicative of the biological age of an individual, as patients with Cockayne syndrome and with dementia were shown to have higher GlycoAgeTest values than age-matched healthy controls (Vanhooren et al., 2010). We recently investigated N-glycomic profiles in 76 DS persons of different ages, as compared with their mothers and siblings (Borelli et al., 2015), and demonstrated that GlycoAgeTest values resulted increased in DS persons as compared to their siblings, in particular at young age. Moreover, GlycoAgeTest values were negatively correlated with Performances IQ score (Borelli et al., 2015).
Epigenetic aging biomarkers, which rely on the widespread DNA methylation changes occurring with age, have proven to be associated with a variety of age-related conditions, including cognitive decline and AD, and with mortality (Field et al., 2018;Horvath & Raj, 2018). One of the most popular and widely used epigenetic biomarkers of age, Horvath's epigenetic clock (Horvath, 2013), has been applied to four independent sets of specimens from DS persons, including whole blood, total peripheral blood leukocytes (PBL), buccal epithelial cells, and brain postmortem biopsies. According to this study, DS persons resulted as significantly older than their calendar age, with an age acceleration ranging from 2.8 years in buccal cells to 11.5 years in brain .
Notably, the same trend of accelerated aging in DS brains has been recently reported by using a totally different approach, that is, magnetic resonance imaging (MRI; . This aging biomarker, based on a machine-learning guided analysis of MRI data , was previously associated with mild cognitive impairment and AD (Franke & Gaser, 2012;Gaser, Franke, Klöppel, Koutsouleris, & Sauer, 2013). By this method, Cole et al.
observed a brain predicted age difference (brain-PAD) of 7.69 years between adult DS persons (mean age 42.3 ± 8.7 years) and agematched controls (Cole & Franke, 2017). Moreover, in DS persons brain-PAD correlated with evidence of beta-amyloid deposition (assessed with Positron Emission Tomography imaging) and with cognitive impairment evaluated as CAMCOG score .
Serum and urine metabolites can be useful biomarkers of age as well. In a study on serum metabolomics, six serum metabolites were identified as part of a putative signature of aging and measured also in 53 DS persons (average age 28.3 years; Collino et al., 2013). Of these six metabolites, 1-O-alkyl-2-acylglycerophosphocholine 32:0, sphingomyelins 24:1 and 16:0 did not change in concentration as compared to age-matched controls, while tryptophan and lysophosphatidylcholines 18:2 and 20:4 had levels closely matching those of elderly and centenarian subjects (Collino et al., 2013).

ARE ALTERED IN DS PE RSONS: THE SEVEN PILLARS OF AGING
As discussed in the previous paragraph, different biomarkers of age are concordantly altered in DS persons. Moreover, several observations suggest that the main molecular mechanisms involved in the aging process are markedly altered in DS. These mechanisms, also FRANCESCHI ET AL.
| 3 of 11 referred as "the seven pillars of aging" (Kennedy et al., 2014), include metabolism, stem cells and regeneration, macromolecular damage, inflammation, adaptation to stress, proteostasis, and epigenetics. DS persons display remarkable alterations for each of these "pillars," thus supporting the concept of DS as an accelerated aging syndrome (Table 1). However, it is to note that in the majority of these studies, the different pillars were not evaluated in cohorts of different ages, or in longitudinal settings. Therefore, it is difficult to determine whether the alterations observed in DS are inherent to the trisomy rather than a sign of accelerated aging, or whether there is an interaction between the syndrome and the aging process. Further studies are needed to specifically address this issue.
Taken together, these data support George Martin's pioneering intuition that DS persons suffer from accelerated aging, affecting in particular the nervous system. In the next paragraph, we will discuss how this knowledge could be exploited in the search for new DS therapeutic targets.

DETERIORATION OF DS PERSON S?
Considering that DS persons become dependent on other people's support at a relatively young age because of dementia, and that the parents and families of DS persons also grow old and in some cases cannot take care of them anymore, dementia in adult DS persons is becoming not only an individual clinical problem but also a huge family and public health emergency. It is now clear that integrated interventions including social inclusion processes, schooling, and cognitive training can be of great importance to improve cognitive capabilities of DS persons. In particular, early programs of cognitive training have been demonstrated to counteract effectively cognitive deficits, by improving intellectual skills of DS persons (Connolly, Morgan, Russell, & Fulliton, 1993;Couzens, Haynes, & Cuskelly, 2012).
No pharmacological intervention is at present approved for the amelioration of cognitive deficits of DS (Dierssen, 2012;de la Torre & Dierssen, 2012); however, a number of trials have been performed by using drugs for AD, like acetylcholinesterase inhibitors (donepezil, rivastigmine, galantamine), GABAergic antagonists (pentetrazol), N-methyl-D-aspartate receptor antagonists (memantine), or microaliments (minerals, vitamins, antioxidants) that had little or no success (Lott et al., 2011;de la Torre et al., 2016). More recently, another promising molecule has been tested, that is, epigallocatechin gallate (EGCG), a natural polyphenol present in green tea leaves. EGCG is a potent inhibitor of DYRK1A, one of the triplicated genes believed to be involved in the brain alterations observed in DS. EGCG has been tested in vitro and in clinical trials, alone or in combination with cognitive training programs (de la Torre et al., 2016Torre et al., , 2014. However, the beneficial effects of EGCG appear largely inconsistent Stringer, Goodlett, & Roper, 2017) and other studies are granted to confirm the efficacy of this approach. Another possible pharmacological approach is based on inhibitors of GABA A receptors or other modulators of GABA A -mediated transmission, such as NKCC1 (Contestabile, Magara, & Cancedda, 2017). In fact, studies on DS animal models have demonstrated that GABA A antagonists such as pentylenetetrazole (PTZ) are able to improve many cognitive functions of the Ts65Dn mice (novel object recognition, Morris water maze test; Braudeau et al., 2011;Martínez-Cué et al., 2013), while NKCC1 inhibitors such as bumetanide can improve discriminative memory, spatial and associative memory (Deidda et al., 2015). On these premises, a number of clinical trials have been performed with this kind of drugs. Some of them are still ongoing, and the results are eagerly awaited, while others (NCT02024789 and NCT02484703) showed a good tolerance but a lack of global efficacy (Contestabile et al., 2017).
Despite these somehow disappointing results, the strategy of combining together cognitive training and pharmacological therapies is appealing and certainly deserves attention. However, it is emerging that the rationale used so far (i.e., targeting presumed AD molecular pathways) is not effective, and an alternative approach should be urgently pursued. ↑ oxidative damage (Cenini et al., 2012;Franceschi et al., 1992) ↑ sensitivity to DNA damaging agents (Morawiec et al., 2008) Inflammation ↑ chronic inflammation (Zhang et al., 2017) ↑ accumulation of immune cells with memory phenotype (Cossarizza et al., 1990(Cossarizza et al., , 1991 Adaptation to stress

Epigenetics
Age-related changes in epigenetic machinery (Ciccarone et al., 2018) ↑ DNA methylation age  We propose that the new vision of Geroscience could be implemented in DS management. According to this approach, aging and age-related diseases share the same common molecular mechanisms (the seven aging pillars mentioned above), and thus, tackling these mechanisms should in principle counteract both aging and age-related diseases at a time (Franceschi et al., 2018;Kennedy et al., 2014). Considering that DS persons are likely suffering from an accelerated aging that in turn impinges on the same pillars that are at the basis of physiological aging, as we have discussed thus far, it can be predicted that DS could effectively benefit from the emerging anti-aging approaches. In this regard, a number of strategies able to impact on the molecular mechanisms involved in the aging process are currently under investigation, including dietary interventions mimicking chronic dietary restriction, drugs that inhibit the growth hormone/IGF-I axis or the mTOR-S6K pathway, drugs that activate AMPK or specific sirtuins, and anti-inflammatory drugs (Longo et al., 2015), some of them being already in clinical trial.
A detailed description of the experimental evidence supporting these strategies has been provided elsewhere (Longo et al., 2015); however, that some of these interventions, in particular regimens of dietary restriction, have been proven to be protective against neurodegenerative pathologies in experimental models. In particular, periodic protein restriction cycles have been shown to promote changes in circulating growth factors and tau phosphorylation associated with protection against age-related neuropathologies in mice (Parrella et al., 2013). Similarly, calorie restriction has been . However, it has to be considered that rapamycin has important side effects, which can limit its offlabel use as an anti-aging drug. These include, among others, metabolic dysregulation (e.g., hyperglycemia, hyperinsulinemia, and insulin resistance) and proliferative defects in hematopoietic lineages (Soefje, Karnad, & Brenner, 2011), which could be particularly severe in DS persons.
Another strategy that appears particularly promising is the use of drugs able to selectively kill senescent cells in vivo. Treatments with these "senolytic" drugs are capable of exerting beneficial effects on old animals and delay or block the onset and progression of age-related diseases (Jeon et al., 2017;Takata et al., 2013). Such an approach could be of great interest for DS persons. A specific investigation on the occurrence of cell senescence in adult DS persons has not yet been conducted; however, it is known that signs of cell Ciani, Contestabile, Guidi, & Bartesaghi, 2010;Contestabile et al., 2013;Faundez et al., 2018;Guidi et al., 2017). These studies indicated that the administration of lithium can restore the neurogenesis in adulthood thus rescuing the synaptic plasticity of newborn neurons, that in turn leads to the recovery of behavioral performance in fear conditioning, object location, and novel object recognition tests (Contestabile et al., 2013), as well as olfactory functions . Studies on human cells are lacking; however, we reported years ago that lithium chloride was able to increase the proliferative capability of PBMC from DS subjects after a sub-optimal stimulation with mitogenic lectins (Licastro et al., 1983). These studies suggest that lithium-based therapies should be further explored as a potential kinase crucial for the regulation of lipid metabolism, cellular glucose uptake, and mitochondria biogenesis, which is believed to mediate the majority of metformin effects on insulin resistance and metabolism. Beside these effects, metformin has also senomorphic activities, that is, the capability to inhibit cell senescence and its related deleterious secretory phenotype (Moiseeva et al., 2013;Noren Hooten et al., 2016). Accordingly, metformin increases lifespan on animal models such as mice (Anisimov et al., 2005(Anisimov et al., , 2011 and Caenorhabditis elegans (Cabreiro et al., 2013). A number of data have shown that metformin offers a protection against AD, likely through different mechanisms, including the inhibition of Aβ fibril deposition (Markowicz-Piasecka et al., 2017), a protein phosphatase 2-mediated reduction of tau phosphorylation (Kickstein et al., 2010), and a promotion of neurogenesis through the activation of an atypical PKC-CBP pathway (Wang et al., 2012). Moreover, it is known that AMPK can regulate mTOR, the main inhibitor of autophagy (Ravikumar et al., 2010).
Therefore, metformin acts as an anti-aging drug also by activating autophagy, a process that seems to be deranged in neurodegenerative disorders (Boland et al., 2008;Cataldo, Hamilton, Barnett, Paskevich, & Nixon, 1996). Treatment with metformin is associated with a 51% reduced risk of cognitive impairment (defined by modified Mini-Mental Status Exam score ≤23) (Ng et al., 2014) and lowers the risk of dementia in T2D patients as compared with other diabetes medications (Cheng et al., 2014;Orkaby, Cho, Cormack, Gagnon, & Driver, 2017). Another study examining the effect of diabetes treatment on specific cognitive domains (verbal learning, working memory, and executive functions) over 4 years showed that only participants who used metformin alone had better cognitive function compared to participants who used other anti-diabetic drugs (Herath, Cherbuin, & Eramudugolla, 2016). A recent meta-analysis that considered these and other studies concluded that cognitive impairment is significantly less prevalent in diabetic patients treated with metformin (odds ratio = 0.55, 95% CI 0.38-0.78), and dementia incidence is also significantly reduced (hazard ratio = 0.76, 95% CI 0.39-0.88; Campbell et al., 2018). As a whole, these studies clearly suggest that metformin can effectively counteract neuronal progressive degeneration and dementia; furthermore, it appears that metformin can offer protection also toward other age-related diseases such as cardiovascular diseases and some types of cancer; therefore, life expectancy of metformin-treated T2D patients can be higher than DS persons could also benefit from anti-inflammatory properties of metformin, due to its effects on N-glycan biomarkers (de Kreutzenberg et al., 2015), on metabolic parameters such as hyperglycemia, insulin resistance, and atherogenic dyslipidemia, but also on the inhibition of NF-κB activation, ROS, and advanced glycation end-products (AGEs) formation (Saisho, 2015). Indeed, it has been reported that DS brain displays signs of neuroinflammation (Wilcock et al., 2015), and that DS persons are characterized by chronically elevated levels of some inflammatory markers such as IL-1β, IL-6, and TNF-α since young age (Iulita et al., 2016). Such markers are elevated also in AD patients and correlate with low brain functions. In particular, IL-6 and C-reactive protein resulted associated with a decline over time of cerebral function, assessed as blood flow with PET, in regions important for cognition such as the orbitofrontal cortex and hippocampus (Warren et al., 2018). Another astrocyte cytokine, S100B, is elevated in both DS and AD, and it has been reported to induce the synthesis of APP, which can activate microglia that in turn produces IL-1β (Barger & Harmon, 1997;Li et al., 1998). Thus, excess levels of neural IL-1β and S100B can influence the neuropathogenesis of AD in DS (Mrak & Griffin, 2004). The progressive, age-related increase in circulating pro-inflammatory mediators has been conceptualized as inflammaging, and it fuels/triggers many age-related diseases (Franceschi et al., 2000;Franceschi & Campisi, 2014). It is therefore conceivable that DS persons are characterized by a peculiar form of inflammaging and that the AD-like dementia occurring in adult DS persons is at least in part sustained by inflammatory cytokines leading to the production and maturation of amyloid plaques.
Finally, it is known that hypothyroidism is one of the most frequent conditions present in adult DS persons, and to this regard, metformin has been reported to decrease the levels of TSH in hypothyroid patients, thus offering a potential protection against enlargement of thyroid gland, goiter, and nodules (Meng, Xu, Chen, Derwahl, & Liu, 2017).

| CONCLUSIONS
Past attempts to face cognitive decline and dementia in DS persons have largely failed. We think it is time for a change in strategy, and, while waiting for specific targets of the AD-like dementia, we should try to tackle accelerated aging, under the assumption that this approach will also combat related ailments such as dementia. This strategy should possibly include, other than lifestyle and personalized approaches, also dietary interventions and drugs that have proven effective in (or at least good candidate for) delaying/combating the aging process. We have briefly discussed experimental evidence that in particular metformin could be a candidate molecule. In fact, at variance with past approaches, metformin has proven to be effective in preventing a number of phenomena related to both AD dementia and aging at a time. This does not exclude other molecules and therapeutic approaches, and our aim with the present paper was to stimulate an international debate on this topic, in order to obtain a broad consensus finalized to promote appropriately designed rigorous trials, capable of taking into account the clinical heterogeneity of DS persons, particularly when they grow old (Lott, 2012). This scientific debate is also of critical urgency to protect DS persons and their families from a "do-it-yourself" approach to this new generation of anti-aging treat-

CONF LICT OF I NTEREST
The authors have nothing to disclose.