**4. Discussion**

Iron homeostasis is altered in muscle of old rodents, possibly contributing to muscle fibre atrophy and loss via oxidative stress-mediated signalling pathways [18]. A specific form of non-apoptotic cell death, referred to as ferroptosis, seems to occur upon intracellular iron overload, causing oxidative injury which probably involves lipid peroxidation [46]. This iron-driven cell death may operate via mitochondrial and NADPH-dependent oxidases reactive oxygen species burst [46]. However, to the best of our knowledge, the relationship between iron status and physical function in old people was not previously explored.

Studies from our group showed increased levels of muscle non-haem iron, including labile fraction, with age in old rats following hind limb suspension [18]. Such changes were associated with elevated expression of ferritin and decreased TFR1 content [18]. Age-dependent iron accumulation was also reported in muscle subsarcolemmal mitochondria of rats [47]. Notably, mitochondrial iron levels were shown to impact organelle RNA damage as well as the susceptibility to opening of the mitochondrial permeability transition pore [47]. This prompted us to test the hypothesis of a relationship between iron status and age-related functional decline involving muscular mitochondrial damage.

Our finding of an age-dependent accumulation of iron in skeletal muscle (Figure 1) paralleled by decreased expression of two of the three metal importers assayed (i.e., TFR1 and ZIP14) in the LF group (Figure 2A–C) supports the idea of a link between iron dyshomeostasis in muscle and functional status. The analysis of iron-related circulating factors offered further insights into this association. Indeed, ferritin levels, an indicator of stored iron, were found to be higher in both HF and LF older adults (Figure 3A), which might arise from chronic inflammation [48]. This view is consistent with our observation of an age-dependent elevation of plasma IL6 and CRP, the levels of which were both higher in LF relative to HF participants (Figure 4A,B).

Although in apparent contrast to our original hypothesis, the measurement of circulating levels of hepcidin provided interesting information regarding such an association. This defensin-like hormone, produced mainly by the liver, plays a major role in modulating iron metabolism [11]. Indeed, via binding to the iron exporter ferroportin at the level of key iron sources [i.e., duodenal enterocytes

(absorption of dietary iron), splenic and hepatic macrophages (recycling iron from erythrophagocytosis), and hepatocytes (iron stores)], hepcidin induces its own endocytosis and lysosomal degradation as well as of ferroportin [12]. As a consequence, decreased intestinal absorption and release of iron from recycling macrophages occur, ultimately resulting in lower circulating iron concentrations [11].

In the present investigation, higher levels of hepcidin were found in older participants, especially in those classified as LF (Figure 3B), with a parallel increase in IL6 and CRP (Figure 4A,B). These results are in line with previous reports pointing to IL6 as a major hepcidin inducer in older adults [49–51], in whom it may be responsible for iron-limited erythropoiesis [52,53]. Whether inflammation reduces iron availability for myoglobin assembly, thereby contributing to impairing muscle function, is presently unknown. Further support to the link among inflammation, iron status, and functional impairment is lent by the strong negative correlation (*r* = −0.99, *p* = 0.04) between circulating IL6 levels and muscle expression of ZIP14 in LF older participants (Table 2). Although our experimental design does not allow inferring about a direct involvement of ZIP14 in muscle iron clearance, a link between ZIP14 expression and IL6 induction has previously been reported and a role for ZIP14 in iron uptake has been hypothesised [54].

A hepcidin-independent regulation of iron status with ageing cannot be excluded. Indeed, studies conducted in older adults with anaemia and chronic inflammation did not detect increased levels of hepcidin in urine or serum [55,56]. In this context, the co-occurrence of multiple age-related conditions may explain changes in the iron status [57,58]. This could be the case for higher circulating ferritin levels in HF participants, which may result, for instance, from the stimulation of ferritin expression by reactive species [59,60].

mtDNA content and damage (Figure 5A,B) as well as the expression of OGG1, one major enzymatic system of mtDNA base excision repair, and 3-NT (Figure 6A,B) showed an age-related association rather than changes dependent on functional status. These findings are in line with previous results in other aged post-mitotic tissues [61].

Taken as a whole, results from the present study sugges<sup>t</sup> that altered iron metabolism during ageing may predispose to oxidant generation and damage to cell components, including mitochondria. In particular, the association of iron dyshomeostasis with systemic inflammation might represent a kingmaker towards functional decline. Disruption of iron metabolism in myocytes might therefore represent a novel target for interventions aimed at preserving muscle health in old age.

#### **5. Limitations of the Study**

While reporting novel findings, our work is not devoid of limitations that need to be discussed. First of all, the study is exploratory in nature due to the small sample size and the limited amount of muscle tissue available for analyses. In addition, the cross-sectional design hampers inference about the time course of changes in analysed mediators and the development of functional decline. Also, only total iron levels were measured and no information is available about haem and non-haem iron. Likewise, neither haemoglobin levels nor mean corpuscular haemoglobin concentration in erythrocytes were measured. Furthermore, plasma iron levels, Tf a ffinity and saturation, and ferritin capacity were not assessed, which impeded a comprehensive appraisal of body iron homeostasis. Finally, the study did not include a group of actively exercising older people. Both categories of old participants were physically inactive and this did not allow appreciating the possible e ffect of physical activity on iron status in muscle in old age.

**Author Contributions:** Conceptualization, A.P. and R.T.M.; methodology, A.P. and R.T.M.; formal analysis, A.P., E.M., R.C., T.M.M., and T.W.B.; investigation, A.P., G.K., and S.K.S.; resources, C.L., F.L., and R.B.; data curation, R.T.M.; writing—original draft preparation, A.P., E.M., and S.D.A.; writing—review and editing, C.L. and R.C.; supervision, E.M. and C.L.; funding acquisition, C.L., F.L., and R.B.

**Funding:** This research was funded by the National Institute on Aging [R01AG17994], Innovative Medicines Initiative-Joint Undertaking [IMI-JU #115621], Intramural Research Grants from the Università Cattolica del Sacro Cuore [D3.2 2013 and D3.2 2015], and the nonprofit research foundation "Centro Studi Achille e Linda Lorenzon". **Acknowledgments:** The authors thank Robert J. Cousins (Dept. of Food Sciences and Human Nutrition, University of Florida. Gainesville, FL) for providing insightful scientific inputs on metal transporters.

**Conflicts of Interest:** The authors declare no conflict of interest.
