Endurance exercise is widely recognized for its role in mitigating insulin resistance, yet the precise mechanisms remain unclear. In this Classics in Diabetes article, we revisit the article by Amati et al., “Skeletal Muscle Triglycerides, Diacylglycerols, and Ceramides in Insulin Resistance: Another Paradox in Endurance-Trained Athletes?” Published in the October 2011 issue of Diabetes, this article was among the first to highlight the nuanced roles of exercise-induced changes in bioactive lipids such as ceramide and diacylglycerol (DAG) in insulin signaling. The authors’ groundbreaking work challenged some existing paradigms, revealing a more complex relationship between DAGs and insulin resistance than previously thought. Their findings helped lay the foundation for further exploration to unravel the intricate biochemical pathways through which exercise influences insulin sensitivity and metabolic health.

Endurance exercise has long been known to be a powerful tool in the treatment or prevention of insulin resistance and type 2 diabetes. However, mechanisms underlying the beneficial effects of exercise on insulin sensitivity have been—and remain—elusive. The reduction in skeletal muscle glycogen after a session of exercise has long been linked to enhanced insulin action (1,2), with the timeline of glycogen resynthesis after exercise closely aligning with the time course of diminution of the exercise-induced improvement in insulin sensitivity (3). However, whether this link between muscle glycogen and insulin sensitivity is truly a causal relationship has not been clearly established. The increase in mitochondrial density in skeletal muscle that occurs with endurance training, and the subsequent increase in fatty acid oxidation, has also been associated with enhanced insulin sensitivity (4). But again, whether this is an important contributor to enhanced insulin action is questioned (5,6).

Excess intramyocellular lipid (IMCL) accumulation, considered by some to be caused by impaired mitochondrial fatty acid oxidation in skeletal muscle (7), has been known to correlate with insulin resistance for several decades, and the prospect that exercise-induced changes in lipid content and composition in skeletal muscle after training has garnered considerable attention. In this Classics in Diabetes article, we revisit the work of Amati et al. (8) published in 2011 that expanded understanding and helped set the foundation for the subsequent comprehensive examination of exercise-induced modifications of different muscle lipid species on insulin sensitivity. Early studies investigating the relationship between IMCL and insulin resistance suggested that insulin resistance may be a consequence of excessive triacylglycerol concentration in muscle, but these assumptions were based largely on correlational analyses (9), without a definitive underlying mechanism. The potential direct role of elevated accumulation of intramyocellular triacylglycerol (IMTG) on insulin resistance was essentially dismissed after the discovery that endurance-trained athletes are often keenly insulin sensitive despite having high IMTG concentrations (the “athlete’s paradox”) (10). Subsequently, accumulation of bioactive lipids such as ceramide and diacylglycerol (DAG) in skeletal muscle became key “suspects” linking muscle lipids and insulin resistance. Both ceramide and DAG have been identified as key factors inhibiting specific steps in the insulin signaling cascade (11–13). Ceramide primarily inhibits insulin action through inhibition of Akt, thereby disrupting the insulin-mediated signal to initiate the translocation of the GLUT4 glucose transporter to the plasma membrane. Mechanisms for the ceramide-induced inhibition of Akt include ceramide-mediated activation of protein phosphatase 2A (PP2A), which in turn dephosphorylates Akt (14), as well as through increasing activation of the serine/threonine kinase, protein kinase C ζ (PKCζ), which phosphorylates an inhibitory site of Akt (AktThr34) (15). DAG has also linked with a disruption in insulin signaling, largely via activation of PKCθ, resulting in phosphorylation of serine residue on insulin receptor substrate 1 (IRS-1) and a subsequent reduction in IRS-1 tyrosine phosphorylation that is part of the proximal insulin signaling pathway (12,13). Importantly, exercise has been found to modify the abundance and localization of these bioactive lipids, but it has taken many years to reach a clearer understanding—and work by Amati et al. (8) provided key clarifying evidence.

Before the publication of the work by Amati et al. (8), the general state of understanding regarding the effects of exercise on IMCLs and their potential impact on insulin action was that exercise increased intramuscular triacylglycerol, often with a concomitant reduction in DAG and ceramide (16,17). However, the impact of the exercise-induced changes in these IMCLs and muscle insulin resistance in human participants with obesity remained unclear. In their study, Amati et al. recruited older (60–75 years old) sedentary men and women with obesity (“OBS”) (BMI >30 kg/m2; n = 21), an age-matched cohort of “normal weight” sedentary adults (“NWS”) (BMI 18–25 kg/m2; n = 7), and an age-matched cohort of normal weight endurance-trained athletes (“NWA”) (n = 14). As anticipated, insulin sensitivity (measured with stable isotope dilution methods for assessment of the Rd of glucose during a 4-h hyperinsulinemic-euglycemic clamp) was greatest in NWA, followed by NWS and OBS. Furthermore, recapitulating the athlete’s paradox, despite the marked difference in insulin sensitivity, IMTG concentrations were elevated to a similar degree in both NWA and OBS compared with NWS. Their findings regarding muscle ceramide content also aligned with the prevailing contemporary view, demonstrating a nearly twofold greater muscle ceramide content in OBS versus both NWS and NWA (total ceramide content, as well as both saturated and unsaturated ceramide species). However, perhaps the most impactful findings from their study involved their measures of muscle DAG content: 1) muscle DAG content was nearly twofold greater in NWA than in OBS, and 2) total DAG content in skeletal muscle was positively correlated with insulin sensitivity. These findings were contrary to the common view of the time and suggested that the relationship between DAG and insulin resistance was far more complex than simply the gross accumulation of this metabolite within skeletal muscle. Through deeper analysis of distinct DAG species within skeletal muscle, Amati et al. were among the first to identify relationships between particular DAG species and insulin resistance. Intriguingly, they found that DAGs containing two unsaturated fatty acids appeared to be linked with deleterious effects on insulin action, whereas DAGs containing one saturated and one unsaturated fatty acid or two saturated fatty acids did not appear to be linked with insulin resistance. This finding was in subtle contrast to findings from a contemporary study by Bergman et al. (18), who reported that skeletal muscle from highly insulin-sensitive endurance athletes contained a higher abundance of unsaturated DAG species in comparisons with sedentary control participants. As Amati et al. noted at the time, this discrepancy is difficult to reconcile but may be related to the diet and/or age of the participants, or to the methods of quantifying the DAG species (Bergman et al. [18] measured total fatty acid species within the entire muscle DAG pool, while Amati et al. measured fatty acids within the intact DAG moiety). Regardless, this groundbreaking work of Amati et al. (along with the work of Bergman et al.) laid the foundation for a more comprehensive understanding regarding the complex relationship between muscle DAG and insulin action in response to exercise.

One of our authors, B.H.G., was corresponding author for the article by Amati et al. and, reflecting on the article in the context of what was known at the time they conducted the study, remarked:

Prior to the Amati et al. report, the original athlete’s paradox study (10) and other studies in model systems (13,19,20) led us to believe that excess ceramide and DAG accumulation in skeletal muscle were both the true culprits in causing insulin resistance. Moreover, some evidence suggested that specific molecular species of ceramides and DAG were responsible. Until this time, however, little data in humans were available to support this. Thus, our thinking was quite simple: repeating the athlete’s paradox study (10) with the additional measures of skeletal muscle ceramide and DAG would convincingly support our hypothesis that higher insulin sensitivity in the athletes would coincide with lower ceramides and DAG.

The finding that muscle triacylglycerol was higher, and ceramides were lower, in the highly insulin-sensitive athletes corroborated our hypothesis. The results with DAG, however, seemed to be more complicated; some DAG species were positively associated with insulin resistance, while others, in contrast, were associated with better insulin sensitivity. So, while these associations in humans did not prove that ceramides and DAG cause insulin resistance, the study did prove that DAG as a general lipid class did not cause insulin resistance. One of the reviewers criticized the paper because we did not demonstrate the cause of insulin resistance. We agreed! This was actually a good point because it demonstrated to us—and I would argue to the field—that the role of these intramyocellular lipids in insulin resistance and type 2 diabetes was much more complicated than what many of us had previously thought. This study also provided the impetus for others to begin to interrogate the effects of exercise on these lipid species to better understand the complex links between exercise, muscle lipids, and insulin resistance.

Indeed, in the years since Amati et al. published their formative work in Diabetes in 2011, there has been considerable interest by many research groups in further advancing our understanding of the impact of exercise on muscle lipids and insulin sensitivity. Some of this important work focused on the effects of exercise on the cellular distribution of specific muscle lipid species, and the subsequent impact on insulin action. Although it is an oversimplification, lipid accumulation near the muscle cell membrane is often found to be more likely to disrupt insulin signaling compared with lipids stored in other myocellular compartments. Along these lines, total DAG abundance in the muscle membrane fraction has been found to be negatively correlated with insulin sensitivity (21–23). Furthermore, a relatively high proportion of these DAGs in insulin-resistant patients with type 2 diabetes were saturated DAGs (C18:0/C20:4, Di-C16:0, and Di-18:0). However, the interplay among exercise, the intracellular distribution of muscle lipids, and insulin sensitivity is far more complex. For example, Perreault et al. (24) reported that endurance athletes had an elevated abundance of 1,2-DAGs in the mitochondrial/endoplasmic reticulum (ER) fraction, and the abundance of 1,2 DAGs in this compartment was positively correlated with insulin sensitivity. This was unexpected, since 1,2-DAG is the predominant DAG isomer thought to interfere with insulin signaling. Whether there is a mechanistic link between accumulation of these DAGs within this mitochondrial/ER fraction and insulin signaling is not yet known. Similarly to DAGs, ceramides in the membrane fraction have also been linked with insulin resistance (25). However, in contrast to DAG accumulation in the mitochondria/ER, an accumulation of ceramides in the mitochondrial/ER fraction was inversely correlated with insulin sensitivity (24). Further adding to the complexity, characteristics of the lipid droplets that contain the IMCLs can impact insulin sensitivity, and many of these characteristics can be altered by exercise. Lipid droplet size, the family of proteins coating the lipid droplet (i.e., perilipins), and physical contact or proximity of the lipid droplet with other cellular organelles (e.g., mitochondria, ER) have all been linked with regulating insulin action, and all can be modified by exercise. In their 2011 article, Amati et al. touched on some of these factors, reporting no relationship between lipid droplet volume density and insulin resistance, but they found perilipin-5 (the only member of the perilipin protein family measured in their study) to be highest in their cohort of endurance athletes, and it correlated with insulin sensitivity (8). Several other research groups have conducted and are continuing to conduct cutting-edge experiments examining the effects of exercise on muscle lipid droplet characteristics, regulatory proteins, and physical characteristics, as well as cellular dynamics and distribution—and how these changes may impact insulin action (26–29).

Overall, the findings in the classic study by Amati et al. provided key evidence confirming that the effects of chronic exercise on IMCLs and insulin resistance are complex. Considering these associations in humans, while not inferring cause and effect, is nonetheless important to promote a reductionist view—to demonstrate what is not the cause of insulin resistance. The study by Amati et al. is still relevant, as newer studies are still investigating which muscle lipids, and their subcellular location, are playing a mechanistic role in insulin resistance. Exercise can be a powerful intervention to prevent or improve insulin resistance, and additional studies that use exercise as an experimental instrument will certainly help to unravel these complex associations.

The classic 2011 Diabetes article by Amati et al. can be found at https://doi.org/10.2337/db10-1221.

For more information on Classics in Diabetes and to read other articles in the collection, please visit https://diabetesjournals.org/collection/2685/Classics-in-Diabetes.

Funding. J.F.H. is funded by National Institutes of Health grants R01 DK13172, P30 DK089503, and R56 DK133298. B.H.G. is funded by National Institutes of Health grants R01 AG069476, U01 AR071133, and R01 AG059416.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

1.
Richter
EA
,
Garetto
LP
,
Goodman
MN
,
Ruderman
NB.
Muscle glucose metabolism following exercise in the rat: increased sensitivity to insulin
.
J Clin Invest
1982
;
69
:
785
793
2.
Richter
EA
,
Mikines
KJ
,
Galbo
H
,
Kiens
B.
Effect of exercise on insulin action in human skeletal muscle
.
J Appl Physiol (1985)
1989
;
66
:
876
885
3.
Cartee
GD
,
Young
DA
,
Sleeper
MD
,
Zierath
J
,
Wallberg-Henriksson
H
,
Holloszy
JO.
Prolonged increase in insulin-stimulated glucose transport in muscle after exercise
.
Am J Physiol
1989
;
256
:
E494
E499
4.
Goodpaster
BH
,
Katsiaras
A
,
Kelley
DE.
Enhanced fat oxidation through physical activity is associated with improvements in insulin sensitivity in obesity
.
Diabetes
2003
;
52
:
2191
2197
5.
Han
D-H
,
Hancock
C
,
Jung
S-R
,
Holloszy
JO.
Is “fat-induced” muscle insulin resistance rapidly reversible?
Am J Physiol Endocrinol Metab
2009
;
297
:
E236
E241
6.
Pino
MF
,
Stephens
NA
,
Eroshkin
AM
, et al
.
Endurance training remodels skeletal muscle phospholipid composition and increases intrinsic mitochondrial respiration in men with type 2 diabetes
.
Physiol Genomics
2019
;
51
:
586
595
7.
Petersen
KF
,
Dufour
S
,
Befroy
D
,
Garcia
R
,
Shulman
GI.
Impaired mitochondrial activity in the insulin-resistant offspring of patients with type 2 diabetes
.
N Engl J Med
2004
;
350
:
664
671
8.
Amati
F
,
Dubé
JJ
,
Alvarez-Carnero
E
, et al
.
Skeletal muscle triglycerides, diacylglycerols, and ceramides in insulin resistance: another paradox in endurance-trained athletes?
Diabetes
2011
;
60
:
2588
2597
9.
Pan
DA
,
Lillioja
S
,
Kriketos
AD
, et al
.
Skeletal muscle triglyceride levels are inversely related to insulin action
.
Diabetes
1997
;
46
:
983
988
10.
Goodpaster
BH
,
He
J
,
Watkins
S
,
Kelley
DE.
Skeletal muscle lipid content and insulin resistance: evidence for a paradox in endurance-trained athletes
.
J Clin Endocrinol Metab
2001
;
86
:
5755
5761
11.
Summers
SA
,
Birnbaum
MJ.
A role for the serine/threonine kinase, Akt, in insulin-stimulated glucose uptake
.
Biochem Soc Trans
1997
;
25
:
981
988
12.
Szendroedi
J
,
Yoshimura
T
,
Phielix
E
, et al
.
Role of diacylglycerol activation of PKCθ in lipid-induced muscle insulin resistance in humans
.
Proc Natl Acad Sci U S A
2014
;
111
:
9597
9602
13.
Itani
SI
,
Ruderman
NB
,
Schmieder
F
,
Boden
G.
Lipid-induced insulin resistance in human muscle is associated with changes in diacylglycerol, protein kinase C, and IκB-α
.
Diabetes
2002
;
51
:
2005
2011
14.
Dobrowsky
RT
,
Kamibayashi
C
,
Mumby
MC
,
Hannun
YA.
Ceramide activates heterotrimeric protein phosphatase 2A
.
J Biol Chem
1993
;
268
:
15523
15530
15.
Powell
DJ
,
Turban
S
,
Gray
A
,
Hajduch
E
,
Hundal
HS.
Intracellular ceramide synthesis and protein kinase Czeta activation play an essential role in palmitate-induced insulin resistance in rat L6 skeletal muscle cells
.
Biochem J
2004
;
382
:
619
629
16.
Dubé
JJ
,
Amati
F
,
Stefanovic-Racic
M
,
Toledo
FGS
,
Sauers
SE
,
Goodpaster
BH.
Exercise-induced alterations in intramyocellular lipids and insulin resistance: the athlete's paradox revisited
.
Am J Physiol Endocrinol Metab
2008
;
294
:
E882
E888
17.
Schenk
S
,
Horowitz
JF.
Acute exercise increases triglyceride synthesis in skeletal muscle and prevents fatty acid-induced insulin resistance
.
J Clin Invest
2007
;
117
:
1690
1698
18.
Bergman
BC
,
Perreault
L
,
Hunerdosse
DM
,
Koehler
MC
,
Samek
AM
,
Eckel
RH.
Increased intramuscular lipid synthesis and low saturation relate to insulin sensitivity in endurance-trained athletes
.
J Appl Physiol (1985)
2010
;
108
:
1134
1141
19.
Erion
DM
,
Shulman
GI.
Diacylglycerol-mediated insulin resistance
.
Nat Med
2010
;
16
:
400
402
20.
Adams
JM
II,
Pratipanawatr
T
,
Berria
R
, et al
.
Ceramide content is increased in skeletal muscle from obese insulin-resistant humans
.
Diabetes
2004
;
53
:
25
31
21.
Bergman
BC
,
Perreault
L
,
Hunerdosse
D
, et al
.
Novel and reversible mechanisms of smoking-induced insulin resistance in humans
.
Diabetes
2012
;
61
:
3156
3166
22.
Nowotny
B
,
Zahiragic
L
,
Krog
D
, et al
.
Mechanisms underlying the onset of oral lipid–induced skeletal muscle insulin resistance in humans
.
Diabetes
2013
;
62
:
2240
2248
23.
Jocken
JWE
,
Goossens
GH
,
Boon
H
, et al
.
Insulin-mediated suppression of lipolysis in adipose tissue and skeletal muscle of obese type 2 diabetic men and men with normal glucose tolerance
.
Diabetologia
2013
;
56
:
2255
2265
24.
Perreault
L
,
Newsom
SA
,
Strauss
A
, et al
.
Intracellular localization of diacylglycerols and sphingolipids influences insulin sensitivity and mitochondrial function in human skeletal muscle
.
JCI Insight
2018
;
3
:
e96805
25.
Chung
JO
,
Koutsari
C
,
Blachnio-Zabielska
AU
,
Hames
KC
,
Jensen
MD.
Intramyocellular ceramides: subcellular concentrations and fractional de novo synthesis in postabsorptive humans
.
Diabetes
2017
;
66
:
2082
2091
26.
Zacharewicz
E
,
Hesselink
MKC
,
Schrauwen
P.
Exercise counteracts lipotoxicity by improving lipid turnover and lipid droplet quality
.
J Intern Med
2018
;
284
:
505
518
27.
Seibert
JT
,
Najt
CP
,
Heden
TD
,
Mashek
DG
,
Chow
LS.
Muscle lipid droplets: cellular signaling to exercise physiology and beyond
.
Trends Endocrinol Metab
2020
;
31
:
928
938
28.
Gemmink
A
,
Schrauwen
P
,
Hesselink
MKC.
Exercising your fat (metabolism) into shape: a muscle-centred view
.
Diabetologia
2020
;
63
:
1453
1463
29.
Stokie
JR
,
Abbott
G
,
Howlett
KF
,
Hamilton
DL
,
Shaw
CS.
Intramuscular lipid utilization during exercise: a systematic review, meta-analysis, and meta-regression
.
J Appl Physiol (1985)
2023
;
134
:
581
592
Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at https://www.diabetesjournals.org/journals/pages/license.