Page 1 of 20
Diabetes
1
Evaluation of the effect of enteral lipid sensing on endogenous glucose
2
production in humans
3 4
Short Running Title: Xiao et al., Duodenal Lipid Sensing and Glucose Production
5 6
Changting Xiao 1,2*, Satya Dash 1,2*, Cecilia Morgantini 1,2, Khajag Koulajian 1,2, Gary F. Lewis 1,2
7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
1
Departments of Medicine and Physiology, Division of Endocrinology and Metabolism, University of Toronto, Toronto, ON, Canada 2
Banting and Best Diabetes Centre, University of Toronto, Toronto, ON, Canada
*
These authors contributed equally to this work.
Corresponding Author: Gary F. Lewis, MD, FRCPC Toronto General Hospital, 200 Elizabeth Street, EN12-218 Toronto, Ontario, M5G 2C4 Tel: 416.340.4270; Fax: 416.340.3314; Email:
[email protected] Total Word Count: 1998 Number of Tables: 1 Number of Figures: 3
28
Diabetes Publish Ahead of Print, published online March 9, 2015
Diabetes
29 30
Abstract:
31 32
Administration of lipids into the upper intestine of rats has been shown to acutely decrease
33
endogenous glucose production in the pre-absorptive state, postulated to act through a gut-
34
brain-liver axis involving accumulation of long-chain fatty acyl CoA, release of cholecystokinin
35
and subsequent neuronal signaling. It remains unknown, however, whether a similar gut-brain-
36
liver axis is operative in humans. Here we infused 20% Intralipid (a synthetic lipid emulsion) or
37
saline intraduodenally for 90 min at 30 ml/h, 4 to 6 weeks apart, in random order, in nine
38
healthy men. Endogenous glucose production was assessed under pancreatic clamp conditions
39
with stable isotope enrichment techniques. Under these experimental conditions,
40
intraduodenal infusion of Intralipid, compared with saline, did not affect plasma glucose
41
concentration or endogenous glucose production throughout the study period. We conclude
42
that Intralipid infusion into the duodenum at this rate does not elicit detectable effects on
43
glucose homeostasis or endogenous glucose production in healthy men, which may reflect
44
important interspecies differences between rodents and humans with respect to the putative
45
gut-brain-liver axis.
46 47 48 49 50 51
Key Words: intestine, lipid sensing, glucose, humans
Page 2 of 20
Page 3 of 20
Diabetes
52 53
Introduction:
54 55
Endogenous glucose production (EGP), mainly by the liver, plays an important role in regulating
56
glucose homeostasis. In patients with type 2 diabetes, impaired insulin action in insulin
57
sensitive tissues such as the liver, skeletal muscle and adipose tissue contributes to the
58
hyperglycemia.
59
determinant of fasting hyperglycemia in this condition (1).
Hepatic glucose production is inappropriately elevated and is a major
60 61
Studies in rodents suggest that EGP is subject to neuronal regulation involving nutrient sensing
62
in the hypothalamus (2;3) and in the small intestine (4). Intraduodenal administration of lipids,
63
particularly long-chain fatty acids (LCFA), has been shown to reduce food intake in both rodents
64
and humans (5;6) and to suppress EGP profoundly and rapidly in rats under experimental
65
conditions. EGP was suppressed by >50% during a 50-min intraduodenal infusion of Intralipid
66
under conditions of a pancreatic insulin clamp, and plasma glucose concentration was lowered
67
by ~20% 15 min after intraduodenal infusion of Intralipid in non-clamped conditions (4). This
68
effect occurred prior to significant absorption of the lipids, as evidenced by the absence of
69
elevations in plasma free fatty acid (FFA) or triglyceride (TG) levels (4). Bases on studies in rats,
70
upper intestinal lipid sensing lowers EGP via a gut-brain-liver axis (4;7;8). Since this occurred at
71
a time when insulin levels at the liver were lower than would be present during a meal, the
72
physiological role of this pathway in regulating glucose homeostasis during food consumption
73
remains unclear.
Diabetes
74
75
The role of upper intestinal lipid sensing and the potential existence of a gut-brain-liver axis in
76
the regulation of glucose homeostasis has not previously been investigated in humans, hence
77
the aim of the current study. We assessed EGP during infusion of either normal saline or
78
Intralipid into the duodenum in healthy men. Intralipid is an emulsion consisting of
79
predominantly long chain polyunsaturated fatty acids, which has been used to demonstrate a
80
gut-brain-liver axis in rats (4). The choice of infusion rate could not easily be extrapolated from
81
previous rat studies. We chose to infuse Intralipid at a rate that has been demonstrated in
82
humans to inhibit food intake and induce cholecystokinin (CCK) release (9), which is believed to
83
mediate the gut-brain signaling effect in the regulation of satiety in humans (10;11). Higher
84
doses of Intralipid could cause nausea and raise plasma FFA concentrations. As in previously
85
published rat studies, in order to control for potential fluctuations of hormones in response to
86
enteral nutrient infusion, we performed our studies under conditions of a pancreatic clamp
87
with the infusion of somatostatin and insulin. Since somatostatin also inhibits the secretion of
88
glucagon and growth hormone, in our study these hormones were also replaced.
89 90
Page 4 of 20
Page 5 of 20
Diabetes
91 92
Research Design and Methods:
93 94
Participants:
95
Nine healthy men participated in the study. Their demographic and biochemical parameters are
96
shown in Table 1. Participants had no existing medical illnesses, were not taking any
97
medications, and had normal glucose tolerance. Each participant was studied on 2 occasions
98
(receiving intraduodenal infusion of Intralipid or normal saline), in random order, 4-6 weeks
99
apart, in a single-blind, cross over design.
100 101
Study Outline:
102
The study outline is depicted in Figure 1. Participants were admitted to the Metabolic Test
103
Centre of the Toronto General Hospital after an overnight fast. A radio-opaque polyvinyl
104
feeding tube was inserted into the first part of the duodenum under fluoroscopic guidance.
105
Two intravenous catheters were inserted into superficial forearm veins in opposite arms, one
106
for infusion of hormones and isotopically enriched glucose and the other for blood sampling.
107
Blood sampling from a forearm vein may systematically underestimate arterial plasma glucose.
108
Since the same protocol was used for both study arms, this is not expected to affect the
109
conclusion of the study. At 1300h (referred to as t = -120 min), a pancreatic clamp, with infusion
110
of somatostatin (30 ug/h), insulin (0.05 mU/kg/min), glucagon (0.65 ng/kg/min) and growth
111
hormone (3.0 ng/kg/min), was started and continued for the remainder of the study, as
112
previously described (12). Concurrently, a primed, constant infusion (22.5 µmol/kg bolus
Diabetes
113
followed by 0.25 μmol/kg/min) of d-[6,6-2H2]glucose (D2-glucose; Cambridge Isotope
114
Laboratories) was started and continued until the end of the study. Two hours later, i.e. at
115
1500h (referred to as t = 0 min), infusion into the duodenum was started for delivery of either
116
Intralipid (20%, DF) or normal saline (NS) at the rate of 30 ml/h and continued for 90 min.
117
Blood samples were drawn at baseline (t = -120 min), every 30 min before intraduodenal
118
infusion (t = -90, -60, -30 min), and every 10 min thereafter until the end of the study (t = 90
119
min). A dextrose (20%) solution was infused to maintain euglycemia if blood glucose declined to
120
< 4.0 mmol/L. Plasma was immediately separated from blood samples into tubes containing
121
aprotinin, sodium azide and tetrahydrolipstatin and stored for future analysis.
122 123
Laboratory Methods:
124
Plasma was deproteinized, delipidated, dried and derivatized with acetic anhydride.
125
Derivatized samples were analyzed with GC/MS (Agilent 5975/6890N) with electron impact
126
ionization using helium as the carrier gas. Selective ion monitoring at m/z = 242 and 244 was
127
performed. Atom percent excess (APE) fraction was calculated for each sample as APE = tracer /
128
(tracer + tracee). Commercial kits were used to measure TG (Wako), FFA (Wako), insulin
129
(Millipore) and glucagon (Millipore).
130 131
Calculation of endogenous glucose production (EGP):
132
EGP was estimated as described previously (13). Briefly, during steady state, rate of glucose
133
appearance (Ra) = rate of glucose disappearance (Rd), where Ra = tracer infusion rate / APE
134
fraction, and EGP = Ra – glucose infusion rate.
Page 6 of 20
Page 7 of 20
Diabetes
135 136
Statistics:
137
Results are presented as mean ± SEM. Paired t-test was used to compare mean plasma
138
concentrations of glucose, TG, FFA, insulin and glucagon and EGP. A p value < 0.05 was
139
considered significant.
140
Diabetes
141 142
Results:
143 144
Plasma concentrations of TG, FFA, insulin and glucagon
145
Baseline (before clamp, t=-120 min) levels of plasma TG were lower in DF compared to NS
146
(p=0.02), which may have contributed to a trend towards lower mean TG in DF vs NS during the
147
30 min prior to intraduodenal infusion (p=0.078) and lower mean TG in DF vs NS 30 min
148
following intraduodenal infusion (p=0.047) (Figure 2A). Intraduodenal infusion of Intralipid did
149
not elevate plasma TG. Plasma FFA levels in DF and NS were similar at baseline, declined
150
similarly in both arms prior to intraduodenal infusion possibly due to mild arterial
151
hyperinsulinemia, and were not significantly different between DF and NS following
152
intraduodenal infusion (Figure 2B). The absence of TG and FFA elevation indicates no significant
153
absorption of luminal lipids in DF. Plasma concentrations of insulin and glucagon were similar
154
between DF and NS, both before and following intraduodenal infusion (Figure 2C and 2D).
155 156
Plasma glucose concentrations, APE and EGP
157
Plasma glucose concentrations were not significantly different between treatments at baseline.
158
There was a small but significant increase in both groups (p