Early-Stage Alzheimer’s Disease Is Associated with Simultaneous Systemic and Central Nervous System Dysregulation of Insulin-Linked Metabolic Pathways

Abstract

Background:

Brain insulin resistance is a well-recognized abnormality in Alzheimer’s disease (AD) and the likely mediator of impaired glucose utilization that emerges early and progresses with disease severity. Moreover, the rates of mild cognitive impairment (MCI) or AD are significantly greater in people with diabetes mellitus or obesity.

Objective:

This study was designed to determine whether systemic and central nervous system (CNS) insulin resistant disease states emerge together and thus may be integrally related.

Methods:

Insulin-related molecules were measured in paired human serum and cerebrospinal fluid (CSF) samples from 19 with MCI or early AD, and 21 controls using a multiplex ELISA platform.

Results:

In MCI/AD, both the CSF and serum samples had significantly altered mean levels of C-peptide (both elevated), Visfatin, incretins, PAI-1, ghrelin, and leptin, whereas only CSF showed a significant reduction in insulin and only serum had increased glucagon levels. Although the overall CSF and serum responses reflected insulin resistance together with insulin deficiency, the specific alterations measured in CSF and serum were different.

Conclusion:

In MCI and early-stage AD, CNS and systemic insulin-related metabolic dysfunctions, including insulin resistance, occur simultaneously, suggesting that they are integrally related and possibly mediated similar pathogenic factors.

INTRODUCTION

Alzheimer’s disease (AD) is manifested by progressive alterations in behavior and impairments in short-term memory, executive functions, and other cognitive abilities [1]. The histopathological hallmarks include co-accumulations of structural lesions caused by hyperphosphorylation of tau and aberrant cleavage of the amyloid-β protein precursor with formation of phospho-tau immunoreactive fibrillar deposits in neurofibrillary tangles, dystrophic neurites and neuropil threads, and amyloid-β (Aβ)_1–42 deposits in plaques and blood vessels. Consequently, the clinical diagnosis of AD is often aided by the use of positron emission tomography (PET) neuroimaging to detect isotopically-labeled tracers that mark brain accumulations of pTau and Aβ_1–42 [2, 3], and sensitive immunoassays that quantify pTau and Aβ_1–42 in cerebrospinal fluid (CSF) and serum [4–6]. Unfortunately, extension of these concepts to therapeutic strategies have not resulted in successful disease remediation [7–10]. On the other hand, growing evidence points to dysregulated insulin metabolism as a fundamental mediator of neurodegeneration [11–15] and another potential therapeutic target.

Glucose is the major energy source for brain metabolism. Impairments in brain glucose uptake and utilization are detectable in pre-symptomatic stages of AD [16]. Furthermore, in the prospective Baltimore Longitudinal Study, impairments in brain glucose uptake correlated with reduced expression of the GLUT3 glucose transporter and subsequent development of AD [17]. Mechanistically, since glucose uptake and utilization in the brain and neuronal cells are stimulated by insulin [18–21], insulin deficiency or insulin resistance could dysregulate energy metabolism and thereby contribute to the pathogenesis of AD [11, 15, 22]. Besides its role in glucose metabolism, insulin stimulates working memory and cognition [23–26], its receptors are abundantly expressed in brain regions most susceptible to AD neurodegeneration [27, 28], and experimental inhibition of related signaling networks causes neurodegeneration with AD features [29–31]. In addition, insulin stimulated brain energy metabolism may facilitate replacement aging cellular components and thereby help to prevent neurodegeneration.

Human postmortem [22, 32] and clinical [11, 33, 34] studies have demonstrated brain insulin deficiency and insulin resistance in AD. Insulin deficiency is mainly manifested by reduced insulin levels in brain and CSF, whereas insulin resistance is associated with reduced insulin receptor expression, tyrosine phosphorylation, and binding, activation of insulin receptor substrate signaling downstream through phosphoinositol-3-kinase (PI3K)-Akt pathways, and brain glucose levels [22, 28, 32, 35]. Disruption of these insulin-related networks adversely impacts neuronal survival, oligodendrocyte function, neuronal plasticity, and energy metabolism, and promotes neuro-inflammation, oxidative, nitrosative, and endoplasmic reticular stress, lipid peroxidation, and cell death [14, 15, 36, 37]. In addition, impairments in brain insulin signaling have been linked to increased pTau phosphorylation and Aβ_1–42 accumulation/toxicity [12, 14, 32, 38, 39]. Therefore, apart from the specific AD-associated outcomes, the molecular, biochemical, and cytopathological consequences of insulin deficiency/resistance in the brain closely resemble those that occur in diabetes mellitus and other insulin resistance diseases [15, 40].

The steadily increasing prevalence of AD over the past several decades and across all age groups [41] indicate that factors other than genetic can mediate AD neurodegeneration. Furthermore, the parallel increases in rates of diabetes mellitus and other insulin resistant states, the higher rates of cognitive impairment and AD in people with obesity or type 2 diabetes mellitus [41], and the increased risk for developing MCI or AD in non-obese, non-diabetic people with elevated blood glucose [42, 43] suggest the drivers and mechanisms of peripheral insulin resistance and AD may be shared. Another way to consider the problem is that perhaps insulin resistant disease states are fundamentally related but differentially manifested due to variation in tissues, organs, and systems targeted. For example, atherosclerosis is a single pathologic process that causes different diseases based on compromised flow through specific arteries. If AD is truly one of the progressive insulin resistance/insulin deficiency diseases in which the brain is selective or prominently involved, then therapeutic interventions developed for other related diseases may be extendable to AD. Already this concept has some validity since cognitive impairment in MCI and AD are positively responsive to intranasal insulin, insulin sensitizers, incretins, and lifestyle modifications that enhance insulin responsiveness [15, 23, 26, 44–48]. Furthermore, intranasal insulin has been shown in humans to increase brain energy, including levels of ATP and phosphocreatine using 31-P magnetic resonance spectroscopy to assess cerebral energy metabolism [49]. To delve deeper into the overarching question concerning insulin network dysfunction versus insulin resistance/deficiency as mediators of neurodegeneration, in this study we measured broad indices of insulin-regulated metabolic integrity in paired serum and CSF samples from controls and MCI or early-stage AD human subjects. The primary objective was to assess the presence and characteristics of systemic and CNS dysregulated metabolic networks, and the degrees to which the abnormalities co-exist and whether they are independent of each other.

METHODS

Human subjects

The Lifespan Hospitals Institutional Review Board (IRB) approved the research plan to collect paired serum and CSF samples from human subjects for comparison biomarker expression and all subjects signed written consent forms to have their serum and CSF samples banked for future research [50, 51]. The patients with MCI or early AD were evaluated and followed clinically in the Rhode Island Hospital (RIH) Alzheimer’s Disease and Memory Disorders Center between 2010 and 2016. An AD diagnosis was rendered based on NINCDS/ADRDA criteria [1, 52], and MCI was diagnosed using consensus criteria [53]. Paired serum and lumbar puncture CSF samples were concurrently obtained in accordance with the Alzheimer’s Disease Neuroimaging Initiative (ADNI) protocol. Samples were collected from control subjects as part of a neurologic diagnostic evaluation. Biological fluids were obtained from patients with early AD or MCI in conjunction with procedures performed during a clinical trial or observational research study visit.

Controls were patients evaluated for headache in the Emergency Department of the Rhode Island Hospital between October 2014 and December 2015. The control subjects were 21 years of age or older and cognitively normal by standard neurological exam and review of the clinical records. Results of their diagnostic studies including CSF protein, cell counts, glucose, and Gram stain were negative for active or acute disease processes, and they were discharged from the emergency room to their homes on the day of admission. Additional inclusion criteria were that: 1) paired blood and CSF samples were collected in accordance with standard of care hospital practice; and 2) at least 100 μl each of undiluted serum and CSF were available for these studies. Although the subjects were not specifically screened for metabolic or endocrine disorders, the clinical records and results of the routine clinical assay provided no support for the existence of such confounders. The samples were immediately aliquoted into 1 mL sterile polypropylene screw capped tubes, frozen and stored at −80°C. All samples were hemoglobin-free and filtered (0.45 μM pore) prior to use in the multiplex immunoassays.

Multiplex Human Gut Hormone enzyme-linked immunosorbent assay (ELISA)

The Human Gut Hormone 10-Plex ^™ Assay (Bio-Rad, Hercules, CA), a magnetic bead-based multiplex ELISA, was used to measure serum and CSF immunoreactivity to insulin, leptin, ghrelin, gastric inhibitory polypeptide (GIP), glucagon like peptide-1 (GLP-1), pancreatic polypeptide (PP), and pancreatic peptide YY (PYY). The polypeptides measured help maintain energy balance both systemically and in the CNS (Table 1). The assays were performed in accordance with the manufacturer’s protocol. In brief, duplicate serum (diluted 1:4 in assay dilution buffer) and CSF (undiluted) samples were incubated with magnetic beads covalently coupled with capture antibodies. Captured antigens were detected with biotinylated secondary antibodies followed by a streptavidin-phycoerythrin reporter conjugate. Fluorescence intensity was measured in a MAGPIX (Bio-Rad, Hercules, CA) and hormone concentrations (pg/mL) were determined from standard curves using MAGPIX software.

Table 1.

The names and function of the 10 gut hormones or trophic factors assayed in paired serum and CSF samples from MCI, AD, and control subjects. Summary of Insulin-Related Metabolic Peptides and Their Functions

Polypeptide	Functions
Ghrelin	Ligand for growth hormone secretagogue receptor type 1; induces growth hormone release from the pituitary-regulates growth; stimulates appetite; induces adiposity; stimulates gastric acid secretion.
GIP1: Gastric inhibitory polypeptide	Incretin: Potent stimulator of insulin secretion; stimulates lipoprotein lipase; modulates fatty acid metabolism; poor inhibitor of gastric acid secretion
GLP-1: Glucagon-like peptide-1	Incretin: Potent stimulator of glucose-dependent insulin release; stimulates glucose disposal, independent insulin actions; suppresses plasma glucagon; modulates gastric motility; may suppress satiety; promotes growth of intestinal epithelium. neuroprotective.
Leptin	Important regulator of energy balance by inhibiting food intake and promoting energy expenditure; helps regulate fat depots.
Insulin	Reduces blood glucose; regulates metabolism by increasing cell permeability to monosaccharides, amino acids and fatty acids; accelerates the pentose phosphate cycle and glycogen synthesis in the liver
PAI-1: Plasminogen activator inhibitor-1	Serine protease inhibitor that acts as ‘bait’ for tissue plasminogen activator, urokinase, protein C and matriptase-3/TMPRSS7; regulates fibrinolysis.
C-peptide: Connecting peptide	Stable by-product of insulin production and cleavage; mediates efficient assembly, folding, and processing of insulin in the ER.
Glucagon	Regulates glucose metabolism and homeostasis by increasing gluconeogenesis and decreasing glycolysis and counterregulatory to insulin; raises plasma glucose in response to insulin-induced hypoglycemia; initiates and maintains hyperglycemic conditions in diabetes mellitus.
Resistin	Promotes insulin resistance; suppresses insulin-stimulated glucose uptake in adipocytes; potentially links obesity to diabetes; increases hepatic production of LDL and degradation of LDL receptors, increasing risk of cardiovascular disease; promotes cytokine inflammatory responses and DNA transcription.
Visfatin	Characteristically regulates circadian clock functions, promotes B-cell maturation, and inhibits neutrophil apoptosis; increases insulin sensitivity; promotes cytokine activation.

Statistics

Results are graphed using box plots to depict means (horizontal bars), 95% confidence interval limits (upper and lower box limits), and ranges (stems). Inter-group statistical comparisons were made using unpaired T-tests with 1% false discovery corrections. Statistical significance was defined as p < 0.05.

RESULTS

Study groups

Paired serum and CSF samples were available from 21 controls and 18 subjects with MCI or early AD. Demographic characteristics are provided in Supplementary Table 1. The mean age (±S.D.) of the control group was significantly lower than the MCI and early AD groups (p < 0.0001). However, the sex ratios were similar in all groups. The mean (±S.D.) Mini-Mental State Examination (MMSE) scores at the time of sample collection were 26.4 ± 3.1 (range 21–30) and 21.9 ± 5.5 (range 13–28) in the MCI and AD groups, respectively (p < 0.05). MMSE assessments were not performed on control subjects. Corresponding with previous reports [54, 55], the MCI and AD subjects in this study had significantly reduced serum/CSF ratios reflecting decreased Aβ_1–42 clearance from the CNS. Initial analysis of the multiplex ELISA results revealed generally small differences between the MCI and early AD groups [50, 51, 56]. Therefore, to simplify the data presentation and increase statistical power, data from the MCI and early AD groups were combined (MCI/AD) and compared to the control group.

Insulin and glucagon

Insulin and glucagon have opposing effects in regulating glucose metabolism. Insulin decreases blood glucose concentration and increases cell permeability to monosaccharides, while glucagon increases blood glucose by increasing gluconeogenesis. In serum, the mean insulin levels were similar in the control and MCI/AD groups (Fig. 1A), whereas the mean glucagon level was significantly elevated in MCI/AD relative to control (Fig. 1C). In CSF, insulin concentration was significantly lower in the MCI/AD group relative to control (Fig. 1B) but the mean levels of glucagon were similar in the two groups (Fig. 1D).

Fig. 1.

Insulin and glucagon expression. Bead-based multiplex ELISAs were used to measure (A, B) insulin and (C, D) glucagon in (A, C) serum and (B, D) CSF of control and MCI/early-stage AD subjects. Bar plots Bar plots depict the mean levels (horizontal bar) of immunoreactivity/ml (expressed in arbitrary fluorescence units), 95% confidence interval limits (upper and lower boundaries of the boxes), and ranges (stems). Inter-group comparisons were made using T-tests with 4% false discovery, and significant differences are shown over the graphs.

Incretins

Two incretins were measured, GIP-1 and GLP-1. Incretins are potent stimulators of glucose-dependent insulin secretion. For example, GLP-1 suppresses plasma glucagon and stimulates glucose disposal. In MCI/AD, serum GLP-1 and CSF GIP-1 were significantly elevated relative to control (Fig. 2A, B), whereas serum GIP-1 and CSF GLP-1 levels were similar to control (Fig. 2C, D). Therefore, although incretin levels were simultaneously increased in the CNS (CSF) and periphery (serum) of MCI/AD subjects, the specific responses activated to drive glucose-dependent insulin secretion were not identical.

Fig. 2.

Leptin and ghrelin. Bead-based multiplex ELISAs were used to measure (A, B) leptin and (C, D) ghrelin in (A, C) serum and (B, D) CSF of control and MCI/early-stage AD subjects. Bar plots depict the mean levels (horizontal bar) of immunoreactivity/ml (expressed in arbitrary fluorescence units), 95% confidence interval limits (upper and lower boundaries of the boxes), and ranges (stems). Inter-group comparisons were made using T-tests with 4% false discovery, and significant differences are shown over the graphs.

Ghrelin and leptin

Leptin and ghrelin have opposing actions in that leptin inhibits food intake and positively regulates energy expenditure while ghrelin increases food intake and reduces energy expenditure. The MCI/AD group had altered systemic expression of these energy balance regulators such that the mean serum level of leptin was significantly reduced and ghrelin was significantly increased relative to control (Fig. 3A, C). In contrast, there were no significant inter-group differences with respect to CSF levels of leptin or ghrelin (Fig. 3B, D).

Fig. 3.

Incretins. Bead-based multiplex ELISAs were used to measure (A, B) GIP-1 and (C, D) GLP-1 in (A, C) serum and (B, D) CSF of control and MCI/early-stage AD subjects. Bar plots depict the mean levels (horizontal bar) of immunoreactivity/ml (expressed in arbitrary fluorescence units), 95% confidence interval limits (upper and lower boundaries of the boxes), and ranges (stems). Inter-group comparisons were made using T-tests with 4% false discovery, and significant differences are shown over the graphs.

C-peptide and resistin

C-peptide is co-generated with insulin and often used to gauge insulin concentration and insulin resistance due to its stability. Resistin suppresses insulin’s ability to stimulate glucose uptake into adiopocytes and thereby contributes to insulin resistance. C-peptide levels were significantly elevated in both serum and CSF samples of MCI/AD relative to control subjects (Fig. 4A, B), despite normal serum insulin and reduced CSF insulin in MCI/AD. The mean serum and CSF levels of resistin did not differ significantly between the MCI/AD and control groups (Fig. 4C, D).

Fig. 4.

C-Peptide and resistin. Bead-based multiplex ELISAs were used to measure (A, B) C-peptide and (C, D) resistin in (A, C) serum and (B, D) CSF of control and MCI/early-stage AD subjects. Bar plots depict the mean levels (horizontal bar) of immunoreactivity/ml (expressed in arbitrary fluorescence units), 95% confidence interval limits (upper and lower boundaries of the boxes), and ranges (stems). Inter-group comparisons were made using T-tests with 4% false discovery, and significant differences are shown over the graphs.

Plasminogen activator inhibitor-1 (PAI-1) and Visfatin

PAI-1 is a serine protease inhibitor that has a broad range of functions, including regulation of fibrinolysis, which may be important for preventing vascular occlusions and ischemic injury. Visfatin, a peripheral blood adipokine that promotes maturation of vascular smooth muscle cells, has insulin-mimetic effects, and improves insulin sensitivity. In MCI/AD patients, PAI-1 was significantly increased in CSF (Fig. 5B) but not serum (Fig. 5A), and Visfatin was significantly reduced in both serum (Fig. 5C) and CSF (Fig. 5D) relative to control.

Fig. 5.

PAI-1 and visfatin. Bead-based multiplex ELISAs were used to measure (A, B) PAI-1 and (C, D) visfatin in (A, C) serum and (B, D) CSF of control and MCI/early-stage AD subjects. Bar plots depict the mean levels (horizontal bar) of immunoreactivity/ml (expressed in arbitrary fluorescence units), 95% confidence interval limits (upper and lower boundaries of the boxes), and ranges (stems). Inter-group comparisons were made using T-tests with 4% false discovery, and significant differences are shown over the graphs.

DISCUSSION

The results of this study provide exciting new information about the complexity of insulin-linked metabolic dysfunction among patients diagnosed with MCI/early-stage AD. The findings highlight the need for additional research to clarify the extent to which systemic versus CNS manifestations of insulin resistance differ, and whether those pathophysiological responses are distinct, related, or perhaps causal of one another. This study uniquely examined paired fresh frozen serum and CSF samples to assess the co-occurrence of systemic and CNS abnormalities linked to dysregulation of insulin-modulated metabolic pathways in MCI and early-stage AD. In addition, our secondary aim was to determine if the alterations in neuroendocrine polypeptide levels in CSF were present in serum and therefore could be used in non-invasive diagnostic and clinical monitoring assays. Moreover, the degrees to which systemic versus CNS manifestations of insulin resistance differ could help clarify whether those pathophysiological responses are distinct, related, or perhaps causal of one another. Despite potential shortcomings of the cross-sectional rather than longitudinal design, this study generated exciting new information about the complexity of insulin-linked metabolic dysfunction detected in CSF and serum of MCI/early-stage AD subjects. In addition, the findings could potentially expand the repertoire of insulin-related therapeutic targets for remediating CNS metabolic dysfunction in AD, and a new dialogue about the possible use of non-invasive peripheral blood biomarkers as surrogate indices to guide further evaluation of patients with MCI or suspected early AD.

Insulin and glucagon have opposing effects in regulating glucose metabolism. Insulin decreases blood glucose concentration, increases cell permeability to monosaccharides, amino acids and fatty acids, and accelerates glycolysis, the pentose phosphate cycle, and glycogen synthesis. Glucagon has a key role in glucose metabolism and homeostasis as it regulates blood glucose by increasing gluconeogenesis and decreasing glycolysis. Glucagon’s counterregulatory actions relative to insulin raise plasma glucose in response to insulin-induced hypoglycemia. In addition, glucagon plays an important role in initiating and maintaining hyperglycemic conditions in diabetes mellitus.

The significantly reduced mean level of insulin measured in CSF of MCI or early-stage AD subjects confirms previously reported results [57]. Insulin deficiency in the CNS compromises brain glucose uptake and utilization needed for neuronal plasticity, neuronal and glial homeostasis, myelin maintenance, cell survival, and inhibits pro-inflammatory, prooxidant, and pro-apoptosis mechanisms [13, 27, 36, 37, 58, 59]. Correspondingly, intranasal insulin delivery enhances working memory and glucose utilization in the brain [23, 25, 45, 60–62]. Furthermore, intranasal insulin was demonstrated to enhance brain neuroenergetics without corresponding increases in caloric consumption [49]. Therefore, a major potential benefit of using intranasal insulin to treat insulin deficiency/insulin resistance associated cognitive impairment as occurs in MCI and AD is the selective potentiation of cerebral metabolism independent of caloric intake, possibly circumventing problems stemming from obesity-linked cognitive impairment [49].

The mechanisms of CNS insulin deficiency in MCI and early-stage AD are not known, although roles for increased expression or activity of insulin degrading enzyme (IDE) have been suggested [63], and supported by the finding of higher IDE expression AD relative to aged control brains [64]. Although IDE may provide neuroprotection by degrading and enhancing clearance of Aβ [65–69], its increased expression would likely enhance Aβ clearance and reduce insulin immunoreactivity. Instead, our studies showed that in the MCI/AD group, both Aβ clearance and CSF insulin were significantly reduced [50, 51, 56], which could account for insulin deficiency but not Aβ accumulation. Another potential mechanism of CNS insulin deficiency in MCI/AD is increased expression of the regulator of calcineurin 1 (RCAN1) gene, which inhibits insulin secretion [70]. The RCAN1 gene was mapped to the critical Down locus on Chromosome 21, and in addition to type 1 diabetes due to reduced insulin levels, its overexpression causes mitochondrial dysfunction and oxidative stress [70–72].

Although the mean insulin level in MCI/AD serum was similar to control, other abnormalities detected indicate that systemic insulin regulatory mechanisms were perturbed. Elevated glucagon in serum suggests a need to support glucose homeostasis via gluconeogenesis, possibly as a compensatory response to the low CSF insulin and predicted reductions in brain glucose uptake and utilization. However, since fasting blood glucose was not measured in the research samples, the actual consequences of elevated serum glucagon could not be evaluated. In future studies, it would be of interest to characterize shifts in insulin and glucagon levels in paired CSF and serum samples over the time course of AD progression.

Incretins, including GIP-1 and GLP-1 are potent stimulators of glucose-dependent insulin secretion and modulators of fatty acid metabolism [73]. GLP-1 suppresses plasma glucagon, stimulates glucose disposal, and has neuroprotective actions that may benefit individuals with AD [74, 75]. The multiplex ELISA results revealed significantly higher incretin levels in MCI/AD CSF and serum. However, in MCI/AD CSF, GIP-1 was elevated whereas in serum, GLP-1 and not GIP-1 was increased relative to control. Therefore, although the responses in MCI/AD CSF and serum were fundamentally similar with respect to the anticipated effects of increased incretin expression on insulin levels, they nonetheless differed regarding the specific incretin that was modulated.

Significantly elevated incretin levels in CSF and serum should have resulted in increased concentrations of insulin, but instead, CSF insulin was reduced while serum insulin was similar to control. The explanation for these seemingly discordant results may rest upon the findings with respect to C-peptide expression, which was significantly elevated in both CSF and serum of MCI/AD subjects. C-peptide, a by-product generated with insulin synthesis, is considerably more stable than insulin, and therefore its levels are often used to more accurately gauge insulin production, concentration and resistance. The higher serum and CSF levels of C-peptide and incretin (GLP-1 or GIP-1) in MCI/AD indicate that systemic and CNS insulin production and resistance were both increased. However, reduced CSF levels of insulin vis-à-vis elevated GIP-1 and C-peptide indicate higher levels of insulin degradation in MCI/AD, corresponding with previous reports of elevated IDE expression in AD [64].

The “normal” serum levels of insulin despite elevated C-peptide and GLP-1 also suggest that insulin degradation was also increased in the periphery, and that the MCI/AD group also had evidence of systemic insulin resistance and compensated relative insulin deficiency. Furthermore, the significantly reduced insulin concentration in CSF but not in serum suggests that insulin degradation may have been more pronounced in MCI/AD CNS compared with the periphery. In light of the evidence that the blood-brain barrier integrity was impaired in the MCI/AD group [50, 51], it is unlikely that insulin deficiency in the CNS was due to reduced uptake from the periphery. Therefore, while the results indicate the co-existence of CNS and systemic insulin resistance and deficiency in MCI and early-stage AD, the more pronounced effects in the CNS could account for cognitive impairment in the absence of clinically suspected systemic dysregulation of insulin pathways. Potential roles for differentially increased expression of IDE and RCAN1 should be investigated as potential early biomarkers of insulin signaling network abnormalities in people with MCI who are at risk for progressing to AD. Although IDE represents one mechanism of amyloid clearance, our findings indicate that the MCI/AD subjects had reduced CNS amyloid clearance despite evidence of insulin degradation. This apparent discrepancy is likely due to the fact that other molecules such as neprilysins and endothelin-converting enzymes, which have roles in amyloid beta degradation and were anticipated to be reduced in AD [66, 68], are in fact inconsistently modulated [63] or even elevated in AD despite reduced clearance with disease progression [64].

Leptin and ghrelin have opposing effects in that leptin regulates fat depots by inhibiting food intake and regulating energy expenditure while ghrelin stimulates appetite and induces adiposity in addition inducing growth hormone release from the pituitary gland [76]. Therefore, the significantly reduced serum leptin and increased ghrelin in MCI/AD together would drive food intake and conserve energy, i.e. reduce its expenditure. This finding is consistent with previous reports of altered nutrient intake [77, 78] and sedentariness in people with early-stage neurodegeneration [40]. Down-regulation of leptin may represent an adaptive response to insulin resistance since, although leptin inhibits tyrosine phosphorylation of the insulin receptor substrate 1 (IRS-1) protein and IRS-1’s interactions with growth factor receptor-bound protein 2 (GRB2), it does stimulate signaling through phosphoinositol-3-kinase (PI3K) which promotes metabolism, cell survival, neuronal plasticity, and other critical neuronal and glial functions in the brain [79]. In contrast to the serum response, CSF leptin and ghrelin were not significantly altered in MCI/AD. However, pertinent CNS responses may be largely mediated by targeting of peripherally derived leptin and ghrelin [80].

PAI-1 is a serine protease inhibitor that serves as “bait” for several enzymes including tissue plasminogen activator, urokinase, Protein C, and matriptase-3/TMPRSS7, and inhibits fibrinolysis [81]. Elevated levels of PAI-1 are strongly associated with insulin resistance states, including diabetes mellitus [82, 83], and have been deemed independent risk factors for type 2 diabetes [83]. Elevated PAI-1 is associated with increased risk for thrombotic cardiovascular [84] and cerebrovascular [81] diseases and ischemic stroke [85]. The significantly increased CSF level of PAI-1 in MCI/AD corresponds with the other neuroendocrine manifestations of insulin resistance demonstrated herein, but also suggests a potential mechanism for the frequent co-development of Aβ-independent micro-ischemic brain injury observed in AD studies [86, 87]. Mechanistically, pro-inflammatory cytokine activation has been linked to increased PAI-1 expression [82], and in earlier reports we demonstrated increased pro-inflammatory and reduced anti-inflammatory cytokine/chemokine levels in MCI/AD serum and CSF [50, 51].

Visfatin and resistin are adipokines with opposing effects on insulin responsiveness such that Visfatin increases insulin sensitivity and resistin promotes insulin resistance [76], yet both activate pro-inflammatory cytokines [88]. Visfatin enhances insulin responsiveness by activating the insulin receptor, producing insulin-mimetic effects, and thereby lowers blood glucose [88]. Visfatin is also known as ‘nicotinamide phosphoribosyltransferase’ (Nam-PRTase or Nampt), the rate-limiting enzyme in the nicotinamide adenine dinucleotide (NAD+) salvage pathway which converts nicotinamide to nicotinamide mononucleotide, enabling NAD+ biosynthesis [88]. Visfatin-mediated increases in NAD+ have anti-aging and neuroprotective actions [89–95]. On the other hand, pro-inflammatory cytokine inhibition of Visfatin/NAD+ reduces glucose uptake and Akt phosphorylation via activation of protein tyrosine phosphatase 1B (PTP1B) [96, 97]. Thus, the pro-inflammatory cytokine activation in serum and CSF [50] could have mediated reductions in serum and CSF Visfatin and thereby contributed to systemic and CNS insulin resistance through downstream PI3K-Akt pathways in MCI/AD.

Conclusions

This study provides strong evidence that insulin-related endocrine pathways are simultaneously dysregulated in the CNS and periphery in MCI/early-stage AD. However, due to the cross-sectional nature of the study design, we still cannot conclude whether the CNS preceded or followed the systemic abnormalities. The results suggest that CNS and peripheral insulin deficiency and resistance are not simply mono-molecular abnormalities, but instead are associated with altered expression of multiple polypeptides in manners that seem to conspire insulin-linked metabolic derangements. The co-existence of similar but non-identical systemic and CNS neuroendocrine abnormalities suggests their pathologies may be linked and might stem from the same etiopathic factors. Finally, since multiple neuroendocrine abnormalities contributing to insulin deficiency and insulin resistance were detected, eventual remediation of metabolic derangements in early-stage AD will likely require multi-pronged treatment approaches that address the spectrum of abnormalities that mediates insulin resistance, insulin deficiency, and attendant impairments in energy metabolism, targeting both CNS and non-CNS organ systems.