Reduction in Inflammation and the Expression of Amyloid Precursor Protein and Other Proteins Related to Alzheimer's Disease following Gastric Bypass Surgery

Abstract

Objective:

Obesity and type 2 diabetes are associated with an increase in the incidence and prevalence of Alzheimer's disease (AD) and an impaired cognitive function. Because peripheral blood mononuclear cells (MNC) express amyloid precursor protein (APP), the precursor of β-amyloid, which forms the pathognomonic plaques in the brain, we hypothesized that APP expression diminishes after the marked caloric restriction and weight loss associated with Roux-en-Y gastric bypass (RYGB) surgery.

Research Design and Methods:

Fifteen type 2 diabetic patients with morbid obesity (body mass index, 52.1 ± 13 kg/m²) underwent RYGB, and the expression of inflammatory and AD-related genes was examined before and after 6 months in plasma and in MNC.

Results:

Body mass index fell to 40.4 ± 11.1 kg/m² at 6 months after RYGB. There was a significant fall in plasma concentrations of glucose and insulin and in homeostasis model of assessment for insulin resistance. The expression of APP mRNA fell by 31 ± 9%, and that of protein fell by 36 ± 14%. In addition, there was a reduction in the expression of other AD-related genes including presinilin-2, ADAM-9, GSK-3β, PICALM, SORL-1, and clusterin (P < 0.05 for all). Additionally, the expression of c-Fos, a subunit of the proinflammatory transcription factor AP-1, was also suppressed after RYGB. These changes occurred in parallel with reductions in other proinflammatory mediators including C-reactive protein and monocyte chemoattractant protein-1.

Conclusions:

Thus, the reversal of the proinflammatory state of obesity is associated with a concomitant reduction in the expression of APP and other AD-related genes in MNC. We conclude that obesity and caloric intake modulate the expression of APP in MNC. If indeed, this effect also occurs in the brain, this may have implications for the pathogenesis and the treatment of AD. It is relevant that cognitive function has been shown to improve with weight loss following bariatric surgery.

Obesity and type 2 diabetes mellitus (T2DM) are characterized by chronic low-grade inflammation and carry a significant increase in the risk of Alzheimer's disease (AD) (1, 2). AD is a chronic inflammatory condition of the brain, characterized by the deposition of amyloid plaques and neurofibrillary tangles. The amyloid plaques are formed by the deposition of β-amyloid, whereas the neurofibrillary tangles are formed by the deposition of hyperphosphorylated tau protein (3). β-Amyloid is derived from amyloid precursor protein (APP) through the action of β-secretase and γ-secretase, which has two subunits—presenilin (PS)-1 and PS-2 (4). Hyperphosphorylation of tau protein is induced by the enzyme glycogen synthase kinase (GSK)-3β (5). The formation of β-amyloid is prevented by the proteolytic action of some metalloproteinases including ADAM-9, ADAM-10 and ADAM-17, which are members of the α-secretase (6), whereas formed β-amyloid is degraded by insulin-degrading enzyme and neprilysin (7). Several other novel genes have also been found recently to be associated with AD through genome-wide association scans. They include PICALM, clusterin, TOMM40, and SORL-1 (8).

In view of our recent demonstration that marked weight loss after Roux-en-Y gastric bypass (RYGB) surgery induces a significant reduction in several proinflammatory mediators with an increase in insulin sensitivity and the fact that insulin suppresses AD-related genes in mononuclear cells (MNC) (9), we hypothesized that RYGB leads to a reduction in the expression of the AD-related proteins mentioned above in MNC.

Subjects and Methods

Subjects

Fifteen adult subjects with morbid obesity [mean body mass index (BMI), 52.1 ± 13.0 kg/m²] and T2DM on stable antidiabetic, antihypertensive, and cholesterol-lowering treatments, not on chronic antiinflammatory drugs, and scheduled to undergo RYGB were included in the study. Baseline characteristics for the subjects are presented in Table 1. Fasting blood samples were collected before and 6 months after the RYGB procedure. The study was approved by the Institutional Review Board of the Catholic Health System. Each participant signed an informed consent (www.ClinicalTrials.gov, no. NCT00960765).

Table 1.

Patients' demographic data at baseline and at 6 months after surgery

	Before surgery	At 6 months
Females	11 (73.3)
Age (yr)	44.9 ± 8.7
Duration of T2DM (yr)	7.5 ± 4.0
Systolic blood pressure (mm Hg)	132.9 ± 19.0	123.0 ± 11.1
Diastolic blood pressure (mm Hg)	71.4 ± 12.3	78.5 ± 9.9
Weight (pounds)	328.9 ± 72.8	255.3 ± 63.7a
BMI (kg/m2)	52.1 ± 13.0	40.4 ± 11.1a
Glucose (mg/dl)	148 ± 8	101 ± 4a
Insulin (μU/ml)	18.5 ± 2.2	8.6 ± 1.0a
HOMA-IR	7.1 ± 1.1	2.1 ± 0.3a
HbA1c (%)	7.9 ± 1.4	6.3 ± 0.8a
Total cholesterol (mg/dl)	175 ± 36	165 ± 46
LDL-cholesterol (mg/dl)	109.5 ± 30.4	96.2 ± 32.6a
HDL-cholesterol (mg/dl)	41.5 ± 8.2	48.5 ± 8.1a
Triglycerides (mg/dl)	209.7 ± 158.5	131.6 ± 58.9
Free fatty acids (mm)	0.68 ± 0.16	0.51 ± 0.17
Medications
ACEI/ARB	5 (33.3)	3 (20.0)
Statin	8 (53.3)	5 (33.3)
Exenatide	1 (6.7)	0 (0.0)
Insulin	5 (33.3)	3 (20.0)
Metformin	11 (73.3)	5 (33.3)
Sitagliptin	4 (26.7)	2 (13.3)
Sulfonylurea	4 (26.7)	0 (0.0)
Thiazolidinedione	4 (26.7)	1 (6.7)

Data are presented as mean ± se or number (percentage). HOMA-IR, Homeostasis model of assessment for insulin resistance; HbA1c, glycosylated hemoglobin; LDL, low-density lipoprotein; HDL, high-density lipoprotein; ACEI, angiotensin-converting enzyme inhibitor; ARB, angiotensin receptor blocker.

^aP < 0.05 by paired t test.

MNC isolation from the blood samples and the quantification of AD-related gene mRNA expression was carried out as previously described for APP, PS-1, PS-2, BACE, GSK-3β, PICALM, SORL-1, TOMM40, clusterin, c-Jun, c-Fos, and JNK-1. Gene-specific primers were obtained from Life Technologies (Grand Island, NY). All values were normalized to the expression of a group of housekeeping genes including actin, ubiquitin C, and cyclophilin A.

Western blotting

MNC total cell lysates were prepared and electrophoresis and immunoblotting conducted as previously described (9).

Plasma measurements

Glucose and insulin concentrations were measured as described before. Clusterin was measured by ELISA from R&D Systems (Minneapolis, MN), and amyloid-β-1-42 (p42) was measured in plasma using a high-sensitivity ELISA kit (Wako, Richmond, VA).

Statistical analysis

Statistical analysis was conducted using SigmaStat software (SPSS Inc., Chicago, IL). All data are represented as mean ± se. Change from baseline was calculated, and statistical analysis was carried out using paired t test.

Results

Anthropometric and metabolic changes after RYGB

BMI fell from 52.1 ± 13.0 to 40.4 ± 11.1 kg/m², and there were significant decreases in glycosylated hemoglobin (HbA1C), lipid profile, homeostasis model of assessment for insulin resistance, and the concentrations of insulin, glucose, and free fatty acids (Table 1).

Effect of RYGB on the expression of APP-1 and AD-related genes in MNC

The expression of APP mRNA and protein in MNC fell significantly by 31 ± 9 and 36 + 14%, respectively (Fig. 1, A–C; P < 0.05). β-Amyloid (p42) peptide levels did not alter (from 3.9 ± 0.3 to 3.6 ± 0.2 pm). There was a fall in PS-2 mRNA by 27 ± 10% and in ADAM-9 by 35 ± 12% (Fig. 1A; P < 0.05 for both), whereas PS-1 and BACE expression did not change. The expression of GSK-3β gene fell by 28% (Fig. 1A; P < 0.05). There were reductions by 29 ± 7% in PICALM, by 20 ± 8% in SORL-1, and by 22 ± 8% in clusterin expression (Fig. 1D; P < 0.05). Plasma clusterin concentration fell by 22 ± 8% from 268 ± 20 to 197 ± 14 μg/ml (Fig. 1F; P < 0.05). The mRNA expression of c-Fos subunit of the activator protein-1 (AP-1) transcription factor fell by 49 ± 12%; the expression of the other subunit (c-Jun) did not alter (Fig. 1E; P < 0.05). Furthermore, there was a trend toward a fall in JNK-1 expression (by 17 ± 8; P = 0.08) (Fig. 1E).

Fig. 1.

A, Percentage change in the mRNA expression of APP, PS-1, PS-2, ADAM-9, and GSK-3β. B and C, Representative Western blot (B), and percentage change in APP-1 protein levels (C) by Western blotting from total cell lysates. D, Percentage change in the mRNA expression of PICALM, SORL-1, TOMM40, and clusterin. E, Change in the mRNA expression of c-Jun, c-Fos, and JNK-1 from MNC. F, Change in plasma concentration of clusterin before and at 6 months after RYGB in obese T2DM patients (n = 15). Data are presented as mean ± se. *, P < 0.05 by paired t test.

Discussion

Our data show clearly that RYGB leads to a significant reduction in the expression of APP mRNA and protein in MNC. In addition, there was a significant reduction in the expression of PS-2, a subunit of the γ-secretase complex that leads to the formation of β-amyloid from APP. There was also a reduction in the expression of clusterin and PICALM, both of which are involved in the formation and the deposition of β-amyloid. In addition, there was a reduction in the expression of ADAM-9 and SORL-1, which is a paradox because they reduce the formation and the deposition of β-amyloid, respectively (6, 10). Thus, whereas other genes promoting amyloid deposition fall after RYGB, these two that reduce the deposition also fall. It is possible that the balance of these processes determines the progress of amyloid deposition and the related inflammation. However, APP and the presenilins are probably the more important ones in the pathogenesis of AD. Consistent with these observations, we have recently shown that RYGB also exerts a significant reduction in proinflammatory mediators (11) and that the intake of macronutrients induced an increase in nicotinamide adenine dinucleotide phosphate oxidase-dependent reactive oxygen species generation and intranuclear nuclear factor-κB binding and the expression of TNFα, IL-1β, and other proinflammatory genes (12).

We also investigated the expression of c-jun and c-fos, the two proteins that heterodimerize to form AP-1. β-Amyloid activates AP-1 to induce inflammation and to up-regulate clusterin expression (13). There was a marked fall in the expression of c-fos, although there was no significant change in the expression of c-jun. The expression of the proinflammatory JNK-1, which phosphorylates c-jun and activates AP-1, was reduced but not significantly after RYGB.

Our findings are consistent with recent reports that RYGB is associated with an improvement in cognitive function (14) and the fact that obesity and T2DM are important risk factors for AD. It is also relevant that obesity is associated with changes in the brain as observed with magnetic resonance imaging. The fact that the expression of AD-related proteins is reduced and cognitive function improves after RYGB indicates that this area requires further investigation.

Recent work has demonstrated that APP acts as a proinflammatory mediator in endothelial cells, THP-1 monocytic cell lines, and microglia. Its distribution in the cell membrane is close to β-integrins, and thus its activation by elements of extracellular matrix induces cytokine generation and increases the adhesion characteristics of these cells (15). Thus, APP may exert its proinflammatory effects in various organs including the brain, independently of its product, β-amyloid. It is of interest that endothelial nitric oxide (NO) synthase deletion results in an enhanced expression of APP in brain microvasculature (16). Furthermore, β-amyloid has been shown to suppress cGMP-mediated signaling through the receptors CD36 and CD47 (17). Thus, whereas NO from endothelial cells may suppress APP expression, diminished bioavailability of NO due to oxidative stress in obesity (18) may lead to an increase in APP. The restriction of caloric intake and weight loss may restore the bioavailability of NO and thus reduce the expression of APP. In addition, with the increase in the sensitivity to insulin after RYGB and weight loss, the inhibitory effect of insulin on APP expression may be restored (9).

The absence of a change in plasma concentration of β-amyloid (p42) after RYGB in the face of the reductions in APP and other genes is intriguing. Equally intriguing is the association of lower plasma concentrations of this peptide with the occurrence of AD (19), especially when the deposition of this protein in the brain leads to the formation of the pathognomonic amyloid plaques.

The limitation of this study is that our observations were made on peripheral blood MNC and not on neurons. Because a study on neurons is not possible in the human, surrogate targets are necessary. It is of interest that AD-related factors in MNC are reduced in association with weight loss after RYGB. In this context, it is relevant that a low-dose infusion of insulin that exerts an antiinflammatory effect also suppresses the expression of APP, PS-1, PS-2, and GSK-3β, the proteins related to AD in MNC (9), and that intranasal insulin administration, which provides a direct access for insulin into the brain, leads to an improvement of cognitive function in patients of AD (20). The second weakness of this study is the absence of a control group. Because patients with morbid obesity referred to bariatric surgeons who meet the requirements for surgical treatment are anxious to proceed with surgery, it is difficult to study these patients with randomized controls.

In conclusion, RYGB leads to a reduction in the expression of APP and other AD-related genes in addition to exerting a generalized reduction of inflammatory mediators. If, indeed, the reversal of obesity leads to a reduction in the risk of AD in the morbidly obese, it will provide a therapeutic approach and also a better understanding of the pathogenesis of AD.