BCAAs acutely drive glucose dysregulation and insulin resistance: role of AgRP neurons

Abstract

Background: High-protein diets are often enriched with branched-chain amino acids (BCAAs) known to enhance protein synthesis and provide numerous physiological benefits, but recent studies reveal their association with obesity and diabetes. In support of this, protein or BCAA supplementation is shown to disrupt glucose metabolism while restriction improves it. However, it is not clear if these are primary, direct effects of BCAAs or secondary to other physiological changes during chronic manipulation of dietary BCAAs.

Methods: Three-month-old C57Bl/6 mice were acutely treated with either vehicle/BCAAs or BT2, a BCAA-lowering compound, and detailed in vivo metabolic phenotyping, including frequent sampling and pancreatic clamps, were conducted.

Results: Using a catheter-guided frequent sampling method in mice, here we show that a single infusion of BCAAs was sufficient to acutely elevate blood glucose and plasma insulin. While pre-treatment with BCAAs did not affect glucose tolerance, a constant infusion of BCAAs during hyperinsulinemic-euglycemic clamps impaired whole-body insulin sensitivity. Similarly, a single injection of BT2 was sufficient to prevent BCAA rise during fasting and markedly improve glucose tolerance in high-fat-fed mice, suggesting that abnormal glycemic control in obesity may be causally linked to high circulating BCAAs. We further show that chemogenetic over-activation of AgRP neurons in the hypothalamus, as present in obesity, significantly impairs glucose tolerance that is completely normalized by acute BCAA reduction. Interestingly, most of these effects were demonstrated only in male, but not in female mice.

Conclusion: These findings suggest that BCAAs per se can acutely impair glucose homeostasis and insulin sensitivity, thus offering an explanation for how they may disrupt glucose metabolism in the long-term as observed in obesity and diabetes. Our findings also reveal that AgRP neuronal regulation of blood glucose is mediated through BCAAs, further elucidating a novel mechanism by which brain controls glucose homeostasis.

Fig. 1. BCAA infusion acutely raises blood glucose.

A Experimental design, B blood glucose, C plasma BCAAs, and D plasma insulin following either saline (n = 8) or BCAA infusion (n = 10) through jugular catheter during 30 min period in regular chow-fed male mice. E Western blot data. A separate cohort of male mice that were pre-treated with acute saline or BCAAs were injected with insulin (0.75 IU/kg i.p.; n = 4 saline; n = 5 BCAA) and euthanized 1 h later to measure protein expression of phosphorylated AKT (pAKT) as a marker of insulin signaling in liver, muscle, and epididymal white adipose tissue. F Blood glucose, G plasma BCAAs, and H plasma insulin after either saline (n = 8) or BCAA infusion (150 mM; n = 8) in regular chow-fed females. Mean ± SEM; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

Fig. 2. Pre-treatment with BCAAs does not affect glucose tolerance or insulin sensitivity.

A Experimental design. B Blood glucose during GTT (1.5 g/kg i.p.). C Glucose AOC after i.p. injection of either saline (n = 6) or BCAAs (225 mM; n = 6) in regular chow-fed male mice. D Blood glucose during GTT (1.5 g/kg i.p.). E Glucose AOC after i.p. injection of either saline (n = 7) or BCAAs (n = 8) in regular chow-fed female mice. F Blood glucose during ITT (0.75 IU/kg i.p.) following i.p. injection of either saline (n = 5) or BCAAs (225 mM; n = 5) in male mice or G female mice (n = 6/group). Mean ± SEM; *P < 0.05; **P < 0.01.

Fig. 3. BCAAs acutely impair whole-body insulin sensitivity during pancreatic clamps.

A Experimental design of hyperinsulinemic–euglycemic clamps. B High plasma BCAAs during constant BCAA infusion. C Blood glucose achieving euglycemia during steady state. D Confirmation of hyperinsulinemia during clamps. E Glucose infusion rate (GIR) during clamps and constant infusion of either saline (n = 7) or BCAAs (n = 7) in 4 h-fasted male mice. Mean ± SEM; *P < 0.05.

Fig. 4. BT2, a BCAA-lowering compound, acutely prevents rise in circulating BCAAs in male mice.

Vehicle or BT2 (40 mg/kg) was administered in mice fasted for 5 h. A Plasma BCAAs before and after treatment. B Changes in plasma BCAAs. C Western blots for BCKDH protein and its phosphorylation state in the liver, quadriceps muscle, hypothalamus, cortex, and brainstem. D Western blot analysis in male mice following i.p. injection of either vehicle (n = 4) or BT2 (n = 4). E Plasma BCAAs. F Changes in plasma BCAAs in male mice after intravenous infusion of either saline (n = 8) or BT2 (n = 9) in 3 h-fasted male mice. G Plasma BCAAs (n = 8/group), H western blots for BCKDH protein in tissues, and I western blot analysis in female mice after i.p. injection of either vehicle (n = 4) or BT2 (n = 4). Mean ± SEM; *P < 0.05; **P < 0.01; ***P < 0.001.

Fig. 5. A single injection of BT2 is sufficient to improve glucose homeostasis in diet-induced obese mice.

Mice were fed a high-fat diet for 8 weeks. A Blood glucose during GTT following pre-treatment with either vehicle (n = 5) or BT2 (40 mg/kg i.p.; n = 5) in 5 h-fasted male mice. B Blood glucose AOC. C Plasma BCAAs before and after treatments. D Blood glucose during GTT following pre-treatment with either vehicle (n = 5) or BT2 (40 mg/kg i.p.; n = 5) in 5 h-fasted, diet-induced obese female mice. E Blood glucose AOC. F Plasma BCAAs before and after treatments. G Blood glucose during ITT after pre-treatment with either vehicle (n = 5) or BT2 (n = 5) in diet-induced obese male mice or H female mice. Mean ± SEM; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

Fig. 6. Acute stimulation of AgRP neurons disrupts glucose homeostasis primarily due to BCAAs.

A Schematic shows unilateral injection of AAV encoding Cre-dependent hM3Dq-mCherry (DREADD; 400 nl) into the arcuate nucleus (ARH) of C57Bl/6J or AgRP-IRES-Cre male mice. B Representative image showing hM3Dq-mCherry-expressing AgRP neurons in AgRP-hM3Dq mouse at ×20 magnification. C Food intake measurement as a functional readout of DREADD expression in Control (n = 14) vs. AgRP-hM3Dq mice (M3; n = 13). D Plasma BCAAs, E plasma insulin, and F plasma corticosterone between mice without AgRP neuronal stimulation and with AgRP neuronal stimulation (n = 6–11/group). G Experimental design showing CNO injection to stimulate AgRP neurons, followed by vehicle or BT2 injection (40 mg/kg i.p.) and ipGTT (1.5 g/kg) or ipITT (0.75 IU/kg). H Blood glucose, I blood glucose AOC, and J Plasma BCAAs during GTT between Control vs. M3 mice (n = 6–7/group). K Blood glucose during ITT after vehicle or BT2 pre-treatment between Control vs. M3 mice (n = 5–8/group). L Inverted AOC of blood glucose during ITT. Mean ± SEM; *P < 0.05; **P < 0.01; ***P < 0.001.