Effects of Hypertriglyceridemia With or Without NEFA Elevation on β-cell Function and Insulin Clearance and Sensitivity

Abstract

The natural history of type 2 diabetes (T2D) is typically characterized by an early increase in insulin availability due to elevated insulin production and reduced insulin clearance, coupled with a parallel reduction in insulin responsiveness (1). The biological mechanisms responsible for maintaining chronic hyperinsulinemia and insulin resistance in the absence of significant hyperglycemia remain largely unknown. Key features of insulin resistance and metabolic syndrome are hypertriglyceridemia (2, 3) and increased circulating nonesterified fatty acids (NEFAs) (4). Mounting evidence suggests that elevated triglycerides and NEFAs are not innocent bystanders in the context of the metabolic syndrome but may further exacerbate insulin resistance and the attendant metabolic dysregulation (5-11) (reviewed in (12)). On the other hand, the effects of excess lipid availability on insulin secretion and clearance are still debated.

Most prior human studies focused primarily on the metabolic consequences of increased NEFA availability during infusion of both lipid emulsions and heparin (12-14), the latter to activate the conversion of the infused triglycerides into NEFAs and glycerol (15). Acute and chronic elevation of plasma NEFA concentration consistently induced insulin resistance (12, 16, 17). Furthermore, acute NEFA elevation had a stimulatory effect on insulin secretion, while prolonged NEFA elevation impaired glucose-stimulated β-cell function, particularly in subjects without diabetes (17, 18). Few available studies focused specifically on the effects of triglyceride elevations by employing low-grade lipid infusions without heparin. Acute hypertriglyceridemia induced by a single lipid bolus did not alter the mean blood glucose or insulin response to oral glucose in individuals with normal glucose tolerance (NGT) (19). However, low-grade Intralipid infusion induced insulin resistance, impaired β-cell function (20, 21), and reduced insulin clearance in the long-term (20). A single available study comparing the metabolic effects of lipid vs lipid plus heparin infusions demonstrated the potentiating effect of supraphysiological hypertriglyceridemia on the deterioration of peripheral and hepatic insulin resistance due to the accompanying NEFA elevation (22). While this study did not explore the effect of hypertriglyceridemia alone and only focused on insulin resistance, it supports the direct effect of hypertriglyceridemia on glucose metabolism independent of circulating NEFA. Interpretation of previous findings is often complicated by the use of supraphysiological NEFA and/or triglyceride levels, which are not always representative of the varying increases of these lipids occurring in the metabolic syndrome (23, 24). Moreover, results vary according to whether acute or chronic effects were studied, the degree of glucose tolerance of the participants enrolled, the use of surrogate indexes of β-cell function, and the heterogeneous experimental settings (25, 26). These complexities make it challenging to discern the isolated effects of elevated triglycerides from those of NEFA, especially in relation to β-cell function.

In this prespecified post hoc analysis, we pooled the data of 2 previous studies in subjects with NGT where we experimentally increased circulating triglyceride levels to equally high physiological concentrations by an intravenous infusion of lipids (27) or lipids and heparin (28). We compared the effects of the 2 infusions on insulin sensitivity, model-derived β-cell function, and insulin clearance (27, 28), assessed under fasting and postload conditions, to dissect the pathogenetic mechanisms of glucose intolerance induced by mild hypertriglyceridemia from those mediated by a parallel increase in NEFAs, which frequently occurs in the metabolic syndrome.

Materials and Methods

Study Participants and Design

Twenty-two metabolically healthy, young adults were enrolled in 2 cross-over, randomized controlled trials (Fig. 1). They were lean or mildly overweight (body mass index ≤28 kg/m²) and had normal glucose tolerance, plasma lipids, and blood pressure. None of them was taking any medications or dietary supplements. All study procedures were conducted in accordance with the Declaration of Helsinki and were approved by the local ethics committee (protocol numbers 23474 and 13053). Participants gave written informed consent before enrollment.

Figure 1.

Protocols of the 2 crossover, randomized controlled trials with 75 g oral glucose tolerance tests (OGTT) performed during continuous intravenous infusion of normal saline (0.9% NaCl) or either 20% Intralipid (Intralipid group) or 20% Intralipid plus heparin (I + H group) in healthy volunteers.

Open in new tabDownload slide

After an overnight fast (12 hours), participants were admitted to the Laboratory of Metabolism, Nutrition and Atherosclerosis at the University of Pisa. A 20-gauge polyethylene cannula was inserted into an antecubital vein for the infusion of saline or lipids. A second cannula was inserted retrogradely into an ipsilateral wrist vein for frequent blood sampling, and the hand was kept wrapped in a heated blanket to achieve the arterialization of venous blood. Subjects then underwent 2 5-hour intravenous infusions of either normal saline or 20% Intralipid, without (n = 12) or with heparin (I + H; n = 10) to activate the release of NEFAs. After 2 hours of lipid infusion (time 0 minute), participants consumed within 5 minutes an oral glucose drink consisting of 150 mL of 50% glucose solution (wt/vol) for the oral glucose tolerance test (OGTT). The paired metabolic studies were performed on 2 separate days 1 to 4 weeks apart.

Measurements

Bedside plasma glucose concentrations were measured using the glucose oxidase technique (Beckman Instruments, Fullerton, CA). Plasma insulin and C-peptide concentrations were measured by chemiluminescence on a COBAS e411 instrument (Roche, Indianapolis, IN). Plasma NEFAs were measured either by a standard colorimetric method (WAKO Chemicals, Neuss, Germany), or by standard spectrophotometric methods on a Synchron Clinical System CX4 (Beckman Instruments). Plasma triglycerides were assayed by standard spectrophotometric methods on a Synchron Clinical System CX4 (Beckman Instruments).

Calculations

Insulin sensitivity was assessed from the fasting plasma glucose and insulin levels using the inverse of the homeostatic model assessment of insulin resistance (HOMA-IR⁻¹) and by the Matsuda index during the OGTT. Insulin secretion rate (ISR) was estimated by C-peptide deconvolution as previously described (29). β-Cell function was assessed from the OGTT using a model that describes the relationship between insulin secretion and glucose concentration as the sum of 2 components. The first component represents the dependence of ISR on glucose concentration through a dose–response function (the slope of which is β-cell glucose sensitivity). The dose response is modulated by a potentiation factor which accounts from various mechanisms such as prolonged hyperglycemia, incretins, neural modulation, and nonglucose substrates. The potentiation factor was set to average 1 during the test and expresses relative potentiation or inhibition of insulin secretion; it was quantified by the ratio between the 3-hour value and the baseline value. The second component of insulin secretion is β-cell rate sensitivity, which represents the dependence of insulin secretion on the rate of change of glucose concentration and is related to early insulin release. Baseline insulin clearance was calculated as the ratio between ISR and plasma insulin measured at fasting. Insulin clearance during the OGTT was calculated as the ratio between the areas under the curves of ISR and plasma insulin, calculated using the trapezoid rule (30).

Statistical Analyses

Continuous data are presented as mean ± SD or median (interquartile range), as appropriate. Categorical variables are reported as count (percentage). Comparisons between groups were tested using Mann–Whitney U test for continuous variables and the Fisher exact test for categorical variables. Paired comparisons between lipid and saline infusion in all participants were performed using Wilcoxon signed rank test. A 2-tailed P < .05 was considered to be statistically significant. Analyses were performed using JMP Pro software version 17.0.0 (SAS Institute, Cary, NC).

Results

Participants Characteristics

The 2 groups who underwent the Intralipid or I + H infusion had similar clinical and metabolic characteristics (Table 1). Both lipid infusions were well tolerated without any adverse effect in all participants.

Plasma Triglyceride and NEFA Concentration

Fasting plasma concentrations of triglyceride and NEFAs were similar between the 2 study visits (Fig. 2). As for study design, the Intralipid and I + H infusions raised plasma triglyceride levels to the same extent (+337 [236-370] % and +368 [287-419] % compared with saline, respectively, P = .349 for group difference) and towards similar mean values over the OGTT (2.1 [1.4-3.3] and 2.9 [2.3-3.9] mmol/L, respectively, P = .291). In contrast, NEFAs showed a 12-fold smaller increase during Intralipid than I + H (+122 [31-238] % and +1543 [983-1955] % compared with saline, respectively, P < .0001 for group difference), reaching markedly lower levels during the OGTT in the Intralipid compared with the I + H group (453 [365-510] µmol/L and 3552 [2357-4268] µmol/L, respectively, P < .0001), which were even below fasting values.

Figure 2.

Plasma levels of triglycerides (A) and NEFAs (B) in response to a 75-g oral glucose tolerance test (OGTT, time 0-180 minutes) during continuous intravenous infusion (time −120 to 180 minutes) of normal saline (dashed lines) or either 20% intralipid (white circles) or 20% intralipid plus heparin (black circles). The dotted line at time 0 minutes indicates glucose ingestion at the beginning of the OGTT. Data are mean ± SEM.

Open in new tabDownload slide

Effects on Glucose Homeostasis

Fasting plasma concentrations of glucose and insulin (Fig. 3A-3B) and ISRs (Fig. 3C) were similar between the groups and the 2 study visits. Furthermore, during the saline study there was no difference between groups in β-cell glucose sensitivity (ie, the slope of the dose–response function relating the observed plasma glucose concentrations and ISR over the OGTT, Fig. 3D).

Figure 3.

Plasma glucose (A), insulin (B), insulin secretion rate (C), and relationship between plasma glucose and insulin secretion rate (D) in response to a 75-g oral glucose tolerance test (OGTT, time 0-180 minutes) during continuous intravenous infusion (time −120 to 180 minutes) of normal saline (dashed lines) or either 20% intralipid (white circles) or 20% intralipid plus heparin (black circles). The dotted line at time 0 minutes indicates glucose ingestion at the beginning of the OGTT. Data are mean ± SEM.

Open in new tabDownload slide

At baseline (time 0), neither lipid infusion influenced plasma glucose levels (Fig. 4), while I + H only increased plasma insulin levels mostly by stimulating ISR and to a minor extent by reducing insulin clearance (Fig. 3 and 4). This resulted in a numerical decrease in the estimated insulin sensitivity (HOMA-IR⁻¹), which however did not reach statistical significance.

Figure 4.

Percent changes from saline to Intralipid infusion, the latter without or with heparin, in baseline and postglucose plasma glucose levels, plasma insulin, insulin secretion, insulin sensitivity, insulin clearance, and main β-cell function parameters. Baseline values were measured after 120 minutes of saline/lipid infusion before glucose ingestion. Postglucose values are the average of measurements during a 180-minute 75-g oral glucose tolerance test (OGTT) with saline/lipid infusion, with plasma glucose calculated as the incremental plasma glucose level from the baseline value. Insulin sensitivity was estimated by the HOMA-IR⁻¹ at baseline and by the Matsuda index during the OGTT. Data are median and interquartile range. *P < .05 for difference with saline infusion in all subjects (Wilcoxon signed rank test). °P < .05 for group differences (Mann–Whitney test).

Open in new tabDownload slide

Following glucose ingestion, the 2 lipid infusions similarly increased plasma glucose excursions and reduced insulin sensitivity (Matsuda index) (Figs. 2A and 3). Plasma insulin concentrations were also increased by the lipids, especially in the second half of the OGTT (Fig. 2B) due to the synergistic effect of enhanced ISR and lower insulin clearance, without significant group differences (Fig. 3). β-Cell glucose sensitivity significantly deteriorated during the infusion of Intralipid alone, while it was virtually unchanged during I + H (Figs. 3D and 4). Among the other β-cell function parameters, Intralipid enhanced potentiation, while this effect was not observed with I + H (Fig. 4). Both lipid infusions were neutral on the dynamic component of insulin secretion (ie, β-cell rate sensitivity).

Discussion

The primary objective of this study was to distinguish the effects of triglycerides alone or triglycerides and NEFAs on glucose tolerance and its key determinants, namely insulin sensitivity, secretion, and clearance. To this end, we infused a 20% triglyceride emulsion (Intralipid) with and without heparin in 2 groups of healthy individuals at rates calculated to reproduce the high physiological triglyceride levels that can be observed in the general population without diabetes (7, 8). Both infusions similarly reduced oral glucose tolerance, insulin sensitivity and insulin clearance. Isolated hypertriglyceridemia had opposite effects on insulin secretion, impairing β-cell glucose sensitivity and increasing potentiation, which were abolished by a concomitant rise in NEFA.

Since the seminal work by Randle and coworkers, the prevailing concept has been that excess NEFAs induce insulin resistance due to substrate competition at the level of the skeletal muscle that is evident in condition of hyperinsulinemia (6). Given that most subsequent studies in humans have used the combination of hyperinsulinemia and Intralipid infusion with heparin, the effect of triglycerides per se on fasting insulin sensitivity and secretion has been overlooked. In the present study, the infusion of lipids did not significantly affect plasma glucose values or insulin sensitivity under fasting conditions, while ISR and plasma insulin levels were increased only when both triglycerides and NEFA were raised. Under physiologic conditions, postabsorptive plasma glucose levels are sustained by endogenous glucose production (EGP) from gluconeogenesis and glycogenolysis. In subjects with normal hepatic insulin sensitivity, an enhanced ISR should suppress EGP and lower plasma glucose. Here, during I + H infusion the increased ISR was not associated with changes in plasma glucose concentration. This finding may indicate a direct stimulatory action of NEFAs on the β-cell (not mediated by hyperglycemia) that counterbalances the concomitant primary increase in EGP driven by a larger substrate (glycerol) and energy (NEFA) availability (22, 31). With regard to the mechanisms, β-cells express the G-protein–coupled receptor 40 (GPR-40), also known as free fatty acid receptor 1, which is activated by medium- and long-chain NEFAs to acutely enhance insulin secretion (32). There are several lines of evidence suggesting the importance of this receptor in the control of insulin secretion (33, 34). Human islets from T2D donors have reduced expression of GPR-40 compared with islets from donors without T2D, and there is a positive correlation between glucose-stimulated insulin secretion and GPR-40 expression (34). Under normal conditions, triglycerides would also be expected to indirectly activate GPR-40 following hydrolysis to NEFAs and glycerol. Furthermore, β-cells express receptors for triglyceride-rich lipoproteins, whose triglyceride content and circulating levels correlate with insulin secretion (35). However, there is a 5.5-fold concentration gradient between plasma and interstitial fluid in triglyceride-rich lipoprotein levels (36), and sustained hypertriglyceridemia has been associated to reduced activity of the lipoprotein lipase (37), a rate-limiting enzyme that hydrolyzes circulating triglycerides.

After glucose loading, both lipid infusions increased ISR and glucose levels and reduced insulin sensitivity to similar extents. This finding of triglyceride-induced deterioration of insulin sensitivity is in line with the results of Storgaard et al (20), who also assessed the acute effects of triglycerides increase by employing the gold standard method for the assessment of insulin sensitivity (ie, the euglycemic hyperinsulinemic clamp) and showed that Intralipid infusion decreased insulin action by ∼25% in subjects with impaired glucose tolerance and in healthy controls. The effect of both lipids on inducing systemic insulin resistance was evident also during the OGTT, and is in agreement with previous studies showing fat-induced inhibition of glucose disposal (38, 39), and increased intramyocellular lipid content (40). Of note, insulin sensitivity was not further reduced when triglycerides and NEFAs were simultaneously elevated suggesting either a null effect of mildly elevated NEFAs or a flat NEFA-insulin sensitivity dose response.

In our study, the increase in overall ISR was similar with both lipid infusions and largely driven by hyperglycemia. However, our study design and our methods allowed to reveal 2 direct and opposite effects of triglycerides on insulin secretion, namely a mild depression of β-cell glucose sensitivity and an enhanced potentiation, that were hidden by the concomitant rise in NEFAs. Rate sensitivity was not affected by either lipid, suggesting that the readily available insulin pool is not sensitive to acute changes of lipids, in line with previous reports on healthy subjects (20). We therefore can conclude that NEFAs simultaneously protects from the negative effects of triglycerides on β-cell glucose sensitivity, but also prevent their stimulatory effect on potentiation. Regarding a potential direct effect of triglycerides on the β-cell, studies using clonal INS-1 cells or freshly isolated normal islets have shown the role of lipolysis mediated by the hormone-sensitive lipase in glucose-stimulated insulin secretion (41). With respect to potentiation, our data agree with the evidence that NEFAs impair GLP-1 signaling in isolated β-cells (42), while long-chain triglycerides stimulate GIP secretion from K cells (43). Large prospective studies have shown that whereas elevated triglycerides increase the risk for incident T2D substantially (11), elevated NEFAs have only a marginal effect on predicting future dysglycemia (44). Our findings align with these observations, and further suggest that in acute settings NEFAs may attenuate the negative effects of the accompanying hypertriglyceridemia on β-cell glucose sensitivity.

Along with the acute deterioration of insulin sensitivity, we found that insulin clearance was reduced to a similar extent by the 2 lipid infusions, both at baseline and under postglucose load conditions (45). Although reduced insulin clearance has been traditionally regarded as a compensatory mechanism against dysglycemia, a pathogenetic role for lower insulin clearance and the resulting chronic hyperinsulinemia in the development of T2D has been recently proposed (46-48). Insulin clearance mainly reflects the insulin extraction by the liver, and quickly changes during OGTT within 15 to 30 minutes from glucose ingestion (49). In vitro studies on hepatocytes and in vivo experiments in dogs have demonstrated that both NEFAs and triglycerides decrease insulin clearance (50, 51). In a study in dogs, the delivery of NEFAs via the portal route resulted in greater peripheral hyperinsulinemia during euglycemic hyperinsulinemic clamps than NEFAs given peripherally, suggesting that NEFAs increase in the portal vein reduces hepatic insulin extraction (52). Alternatively, the effect on insulin clearance could be driven by the negative effects of lipids on insulin sensitivity, as suggested by the lack of differences between the 2 infusions.

Strengths of the present investigation are the use of experimental conditions that enabled the simultaneous assessment of the major determinants of glucose tolerance (ie, insulin sensitivity, secretion, and clearance) and the elevation of the circulating lipids in the high physiological range. Limitations include that the 2 interventions (Intralipid or I + H) were not performed in the same participants. However, the 2 groups were matched for age, sex, body mass index, and glucose tolerance, thereby surmising that the observed differences may be truly attributed to the different experimental exposures. Also, the groups were relatively small and included only subjects with NGT, which warrants caution in the interpretation of negative findings and limits generalizability of the present results to other groups of glucose tolerance. Further studies are also needed to elucidate the molecular mechanisms responsible for the differences observed.

In conclusion, acute physiologic elevations of triglycerides impair glucose tolerance directly by reducing insulin sensitivity and by decreasing β-cell glucose sensitivity despite their positive effect on potentiation. The combination of NEFAs and triglycerides is not synergistic in deteriorating glucose tolerance because mildly elevated NEFAs do not have a negative additive effect on insulin sensitivity and prevent both the negative and the positive effects of triglycerides on insulin secretion. These findings may provide novel insight into plausible biological signals that generate and sustain insulin resistance and chronic hyperinsulinemia during progression to diabetes.

Acknowledgments

The authors are grateful to Dr. Silvia Frascerra and Dr. Silvia Pinnola (Laboratory of Metabolism, Nutrition, and Atherosclerosis, University of Pisa, Pisa, Italy) for their assistance in the preparation and conduction of metabolic tests. The authors are also grateful to the study volunteers for their generous contribution to this project.

Funding

European Foundation for the Study of Diabetes (EFSD)/Sanofi Collaborative Program 2015 award to E.M. EFSD/AstraZeneca Future Leaders Mentorship Program for Clinical Diabetologists to D.T. EFSD/Novo Nordisk Rising Star Fellowship to D.T. Italian Society of Diabetology Research Grant to A.N. European Medical Information Framework (EMIF) grant (IMI JU GA 115372-2) to E.F. PERIS program 2016-2020 (LT017/20/000033) from the Departament de Salut de la Generalitat de Catalunya to B.A.

Author Contributions

D.T. performed the studies and wrote the manuscript. E.R. wrote the manuscript. B.A. and E.M. performed the studies. T.S. and S.B. performed the lab analyses. A.M. analyzed the β-cell function data. E.F. and A.N. conceived the study designs.

Disclosures

The authors declare that there is no duality of interest associated with this manuscript.

Data Availability

The datasets generated and analyzed during the current study are available from the corresponding author on reasonable request.

References

Abbreviations

 
• EGP

  endogenous glucose production

 
• GPR-40

  G-protein–coupled receptor 40

 
• HOMA-IR

  homeostatic model assessment of insulin resistance

 
• I + H

  Intralipid with heparin

 
• ISR

  insulin secretion rate

 
• NEFA

  nonesterified fatty acid

 
• NGT

  normal glucose tolerance

 
• OGTT

  oral glucose tolerance test

 
• T2D

  type 2 diabetes

Author notes