**1. Introduction**

Insulin signaling mediated by insulin receptor substrates 1 and 2 (IRS1 and IRS2) is involved in the regulation of growth, glucose homeostasis, energy metabolism, and lifespan [1–4]. The biological effects of insulin signaling are regulated by the modulation of IRS proteins through serine (Ser) and threonine (Thr) phosphorylation [5,6]. Notably, IRS1 is known to be abundantly phosphorylated at Ser and Thr residues regardless of insulin or IGF1 stimulation [5,7]. In vitro studies have demonstrated the relationship between IRS1 Ser/Thr phosphorylation and canonical downstream signaling components, including Akt/protein kinase B, glycogen synthase kinase 3 beta (GSK3β), and ribosomal protein S6 kinase (S6K) [5]. Under physiological and pathological conditions, Ser/Thr phosphorylation of IRS1 is potentially mediated by multiple kinases, including AMP-activated protein kinase (AMPK), conventional and novel protein kinase C (PKC), and c-Jun N-terminal kinases (JNKs), in response to intracellular energy status, nutritional conditions, and inflammatory stimulation [5,6,8].

Among the numerous Ser residues on IRS1, a few sites, including human(h)Ser312/mouse(m)Ser307, hSer616/mSer612, hSer636/mSer632, hSer639/mSer635, and hSer1101/mSer1097, have been studied because a limited number of phosphospecific antibodies were previously commercially available.

Nonetheless, hSer312/mSer307 has been widely investigated and implicated in insulin resistance under metabolic stress conditions and in inflammatory conditions such as obesity, hyperinsulinemia, and dyslipidemia [5,9,10]. However, IRS1 mSer307 knock-in mice display insulin resistance rather than increased insulin sensitivity, suggesting that IRS1 mSer307 is a positive regulatory site and is essential for normal insulin signaling [11]. IRS1 mSer612 and mSer632/635 have been cited as negative regulatory sites for IRS1 signaling through tyrosine phosphorylation [5], and mSer1097, regarded as a potential mammalian target of rapamycin (mTOR) /S6K signaling pathway, is activated in the liver of model animals of obesity [12]. However, the roles of these sites remain largely unknown because they have been examined in different context.

In the central nervous system, studies on postmortem brain tissues of patients with Alzheimer's disease (AD) have revealed increased phosphorylation of IRS1 at hSer312/mSer307, hSer616/mSer612, hSer636/mSer632, and hSer639/mSer635 compared with that in non-AD control subjects [13–16]. The phosphorylation levels of IRS1 at hSer312/mSer307 and hSer616/mSer612 are considerably elevated in the brains of patients with AD [15]. In animal studies, multiple AD mouse models display increased phosphorylation of IRS1 at mSer307 and/or mSer632 or mSer612 [16–18]. However, whether the modifications of IRS1 via Ser phosphorylation are involved in memory decline in amyloid precursor protein (APP) knock-in (KI)NL-G-F (APPKINL-G-F) mice, a novel AD mouse model, remains unclear [19]. Meanwhile, animals with diet-induced obesity (DIO), a model of type 2 diabetes mellitus (T2DM), that were fed a 40% or 60% high-fat diet (HFD) during different periods displayed cognitive impairment accompanied by increased phosphorylation of neural IRS1 at different Ser residues (mSer1097, mSer307, or mSer612) at different ages [20–22]. Few studies have reported the relationship between streptozotocin (STZ)-induced type 1 diabetes-related cognitive impairment and phosphorylation of neural IRS1 at Ser sites. Similarly, in aged animals, the phosphorylation of IRS1 at mSer307 has been shown to increase in the cortex; however, whether this alteration is correlated with age-related decline in cognitive function has not been explored [23].

In the present study, we investigated whether the modification of hippocampal IRS1 by Ser phosphorylation is commonly associated with different types of memory decline in DIO, STZ, aged, and APPKINL-G-F mice and whether it occurs before or after the onset of memory decline in APPKINL-G-F mice. We demonstrate that the concomitant activation of specific Ser sites on hippocampal IRS1 with amyloid-β (Aβ)-unrelated memory decline occurs in DIO and aged mice, whereas STZ mice exhibit memory deficits independent of IRS1 activity. We further show that increased phosphorylation of hippocampal IRS1 at Ser sites is already observed in young APPKINL-G-F mice showing normal memory function despite increased Aβ42 level. These data suggest that the activation of Ser residues on hippocampal IRS1 is associated with non-AD-related memory impairment in T2DM and aging and with Aβ42 level, which is related to the onset of cognitive decline in AD.

## **2. Materials and Methods**

#### *2.1. Animals*

C57BL/6J male mice (4 weeks of age) supplied by Japan SLC, Inc. (Shizuoka, Japan) were used to establish type 1 and 2 diabetes mellitus (T1DM, T2DM) mice and their respective control mice. Generation of high-fat-diet (HFD)-induced type 2 diabetes mice (DIO mice) was carried out as previously described [20]. Briefly, C57BL/6J mice were assigned to control wild-type (WT) mice] and HFD (DIO mice) groups, and were fed a normal diet (CE-2; CLEA Japan Inc., Tokyo, Japan) or a HFD (D12492, 60% kcal from fat; Research Diets, Inc., New Brunswick, NJ, USA) for 32 weeks, respectively. To generate Streptozotocin (STZ)-induced type 1 diabetes mice (STZ mice), eight-week-old C57BL/6J mice were intraperitoneally administered with 150 mg/kg STZ (Wako Pure Chemical Industries, Ltd.,

Osaka, Japan) dissolved in sodium citrate (pH 4.5) after overnight fasting, and control mice were injected with the sodium citrate buffer alone. Young and aged WT male mice, which were all pure C57BL/6J strains, were purchased from Charles River Inc., Kanagawa, Japan, or/and bred within our animal facility. APPKINL-G-F (Swedish (NL), Arctic (G), and Beyreuther/Iberian (F) mutations) homozygous mice were obtained from Dr. Saido at the Laboratory for Proteolytic Neuroscience, RIKEN Brain Science Institute, Saitama, Japan [19]. Age-matched WT mice of a similar strain (C57BL/6J) were used in the control experiments. All experiments were performed using DIO, STZ, and young and middle-aged APPKINL-G-F mice, and their respective age-matched control mice (DIO mice: 34–36 weeks, WT mice: 34–36 weeks; STZ mice: 10 weeks, WT mice: 10 weeks; young APPKINL-G-F mice: 12 weeks, WT mice: 12 weeks; middle-aged APPKINL-G-F mice: 34–36 weeks, WT mice: 34–36 weeks; young WT mice: 8–13 weeks, aged WT mice: 84 weeks). All mice were housed in a standard 12 h light–dark cycle with free access to water and food (room temperature: 25 ± 2 ◦C). Animal experiments were performed in compliance with the guidelines and with the approval of the ethics committee in Animal Care and Use of the National Center for Geriatrics and Gerontology in Obu, Aichi, Japan (Approval ID: 31-5).
