Introduction
The immune system is of primary importance to control diseases throughout an individual’s life, and therefore crucial to its fitness. In Vertebrates, the immune system involves different immune functions which can be divided into innate and adaptive components (Hoebe et al., 2004). The innate immune functions are the first defence against pathogens, involving phagocytic cells (e.g. neutrophils, macrophages and dendritic cells) and molecules such as cytokines, also able to activate other parts of the immune system (Akira et al., 2006; Mantovani et al., 2011; Nathan, 2006; Vivier et al., 2011). The adaptive immune functions comprise a cell-mediated immune response, with the stimulation of T lymphocytes, and a humoral immune response, controlled by activated B lymphocytes that can produce immunoglobulins against specific antigens (Iwasaki & Medzhitov, 2010; Mantovani et al., 2011; Vivier et al., 2011).
Mounting an immune response carries costs (Graham, Allen, & Read, 2005; Lochmiller & Deerenberg, 2000; Maizels & Nussey, 2013) and trade-offs with other life-history traits are likely to emerge (Eraud, Jacquet, & Faivre, 2009; Graham et al., 2010; Hanssen, Hasselquist, Folstad, & Erikstad, 2004; Lemaitre et al., 2015; Viney, Riley, & Buchanan, 2005). Therefore, according to the theory of senescence (Medawar, 1952), and more particularly the disposable soma theory (Kirkwood & Rose 1991), a decrease in immune performance with age is expected (reviewed in Lavoie, 2006; Shanley, Aw, Manley, & Palmer, 2009; Simon, Hollander, & McMichael, 2015).
Immunosenescence was mainly studied in humans and laboratory animals (Bektas, Schurman, Sen, & Ferrucci, 2017; Frasca, Riley, & Blomberg, 2005; Gayoso et al., 2011; Larbi et al., 2008; Noreen, Bourgeon, & Bech, 2011; Solana et al., 2012), with the general pattern being a decline in adaptive immunity with age, while innate immunity remains unchanged and inflammatory markers increase (Bauer & De la Fuente, 2016; Franceschi et al., 2007; Frasca, Diaz, Romero, Landin, & Blomberg, 2011; Panda et al., 2009; Shaw, Goldstein, & Montgomery, 2013; Simon et al., 2015). In non-model organisms, a recent review found similar trends (Peters et al., 2019). These findings suggest that the decrease in the immune functions with age could occur and impaired survival (e.g. Froy et al., 2019; Schneeberger, Courtiol, Czirjak, & Voigt, 2014), but that a remodelling of the immune functions and ‘inflammaging’ (accumulation of pro-inflammatory factors, (Franceschi et al., 2018; Goto, 2008) characterized by changes in the proportion of the  different cells involved in the immune response could also take place. Such changes could lead immune changes with age to be adaptive (Fulop et al., 2018; Mueller et al., 2013; Nikolich-Zugich, 2018).
Because the immune system is complex, involving many cell types and pathways, its characterization in non-model organisms is challenging, thus limiting the study of immunosenescence in free-ranging animals (Boughton et al., 2011; Demas et al., 2011). Only few cross-sectional studies investigated the variations in the immune function with age (mammals: Abolins et al., 2018; Cheynel et al., 2017; Nussey, Watt, Pilkington, Zamoyska, & McNeilly, 2012; birds: Hill et al., 2016; Lecomte et al., 2010; Palacios, Cunnick, Winkler, & Vleck, 2007; Saino, Ferrari, Romano, Rubolini, & Moller, 2003; Vermeulen, Eens, Van Dongen, & Muller, 2017; reptiles: Massot et al., 2011; Ujvari & Madsen, 2011; Zimmerman et al., 2013), and even less with survival (Froy et al., 2019; Hanssen et al., 2004; Schneeberger et al., 2014). However, cross-sectional studies cannot disentangle whether the observed variations arise from within-individual changes or from processes like selective disappearance (van de Pol & Verhulst, 2006; van de Pol & Wright, 2009). Due to this shortcoming, immunosenescence can be either hidden when it occurs or observed when it does not (Nussey, Coulson, Festa-Bianchet, & Gaillard, 2008) leading to inappropriate conclusions regarding its evolutionary consequences.
The current lack of longitudinal studies investigating variations in immune functions with age (to the best of our knowledge, four studies: Beirne et al., 2016; Andrea L. Graham et al., 2010; Schneeberger et al., 2014; van Lieshout et al., 2020) is one of the biggest limitations to our understanding of immunosenescence in wild populations (Peters et al., 2019). In the present study, we recorded the age-specific leukocyte concentration and counts in 52 dominant individuals repeatedly sampled between 2011 and 2015 (for a total of 169 measurements) from a wild and long-term studied (1992-2018) population of Alpine marmots. We first tested whether, once controlled for a potential selective disappearance, individuals’ leukocyte concentration and counts only decrease as they age as expected from the disposable soma theory or whether more complex patterns involving changes in leukocyte counts do occur. We further tested whether age variations in leukocyte concentration and counts correlated with survival probabilities using a longitudinal approach.
Material and methods