Atherosclerosis and protection of the vascular wall

The link described in the previous paragraph between insulin resistance and endothelial dysfunction represents one of the first steps through which an initially metabolic alteration can translate, over time, into a true structural modification of the arterial wall. Atherosclerosis, in fact, does not appear suddenly with the formation of an already organized plaque, but develops through a long succession of events in which metabolic alterations, inflammation, oxidative stress, lipid accumulation, and changes in the response of vascular cells tend to progressively overlap. In this process, insulin resistance seems to play a particularly important role, so much so that its association with atherosclerosis can be detected when the disease is still subclinical, that is, before it produces evident cardiovascular manifestations.

One of the most extensive demonstrations comes from the analysis of coronary artery calcification, considered an indicator of coronary atherosclerotic burden. A meta-analysis including 15 studies and a total of 60,649 participants showed that higher HOMA-IR index values were associated with a greater prevalence of coronary calcification. Considering HOMA-IR as a continuous variable, the increase in insulin resistance was associated with an increased probability of presenting coronary calcifications, with an overall OR (a statistical indicator that measures how much more or less likely an event is) of 1.27. Even more interestingly, in the available longitudinal data, high HOMA-IR values were also associated with the progression of calcification, with an OR of 1.44. The quality of the evidence, however, was judged low for prevalence and very low for progression (Li et al., 2023) [13].

The same relationship emerges using another metabolic indicator, the triglyceride-glucose index, or TyG, obtained by combining fasting concentrations of triglycerides and glucose and frequently used as a surrogate marker of insulin resistance. A review of the available evidence compared TyG, HOMA, and triglycerides with different indicators of subclinical atherosclerosis, highlighting how TyG showed the most consistent association with coronary calcification. The relationship with arterial stiffness, assessed through pulse wave velocity (the pressure wave that propagates along the arteries every time the heart contracts), was also robust. Particularly significant is the fact that these alterations can be observed even before the onset of overt diabetes: insulin resistance therefore seems to accompany the atherosclerotic process in a phase in which traditional glycemic criteria might not yet be sufficient to identify its presence (Scott et al., 2024) [14].

The evidence becomes even more concrete by observing the carotid arteries directly. In a study conducted on 4,203 Chinese steelworkers, subjects belonging to the highest quartile of the TyG index had, compared to those in the lowest quartile, more than three times the probability of carotid atherosclerosis. The association concerned both the increase in intima-media thickness and, above all, the presence of an actual plaque: for the latter, the OR reached 3.7. The relationship was not completely linear, and the risk increased significantly beyond a TyG value of 8.7. The study also suggests that metabolic vulnerability may interact with other environmental factors, as the association appeared particularly evident in workers exposed to shift work, heat, noise, carbon monoxide, and dust (Cao et al., 2025) [15].

A consistent picture emerges in patients in whom coronary artery disease is already clinically manifest. In 1,516 subjects with symptomatic coronary artery disease, increasing TyG values were associated in parallel with a greater extent of coronary atherosclerosis and the presence of carotid plaques. Compared to the lower quartile, the one with the highest values had a 2.5 times greater probability of coronary lesions and 2.3 times greater probability of carotid plaque. It is important to emphasize that the association persisted after adjustment for various cardiovascular factors and independently of the presence of diabetes or hyperlipidemia. Furthermore, the ability of TyG to identify atherosclerotic lesions was superior to that obtained by considering glycemia or triglycerides separately (Li et al., 2023) [16].

Insulin resistance also seems to be associated not only with the presence of plaque but also with its characteristics. This aspect is particularly important because not all plaques present the same risk. A stable plaque can persist for a long time, while an unstable plaque, generally characterized by elements of greater vulnerability, can undergo complications with much more significant clinical consequences. In over 12,000 subjects who underwent carotid ultrasound, higher TyG values were significantly associated with the presence of unstable plaques. The association was present in both men and women but was more marked in men; it also became particularly evident after age 60 and in diabetic patients. The analysis showed a non-linear relationship, with a rapid increase in risk beyond a TyG of 8.63 (Huo et al., 2025) [17].

Epidemiological observations find important support in experimental studies, which allow for the isolation of the contribution of insulin signaling within the vascular wall. In mice genetically predisposed to atherosclerosis through the loss of apolipoprotein E, the selective elimination of the insulin receptor from endothelial cells produced a particularly significant result: despite systemic insulin sensitivity, glucose tolerance, insulinemia, plasma lipids, and blood pressure being substantially comparable to those of the controls, atherosclerotic lesions were more than doubled. The loss of insulin action in the endothelium was accompanied by lower activation of AKT and eNOS, impairment of vasodilation, increase in VCAM-1 (a protein present on the surface of cells lining blood vessels, which promotes the adhesion of white blood cells), and greater leukocyte adhesion to the vascular wall (Rask-Madsen et al., 2010) [18].

This experiment allows for an important conceptual step. The association between insulin resistance and atherosclerosis does not seem to depend exclusively on the systemic consequences that frequently accompany insulin resistance, such as hyperglycemia and dyslipidemia. The loss of proper insulin action directly in the endothelium can be sufficient, at least in the experimental model used, to strongly accelerate atherogenesis. In other words, physiologically functioning insulin does not simply appear as a metabolic hormone that controls glycemia, but also exerts vascular signals that contribute to maintaining the endothelium in a state less favorable to the development of atherosclerosis.

This interpretation is reinforced by an almost mirror-image experiment. If the loss of endothelial insulin signaling accelerates atherosclerosis, its enhancement can produce the opposite effect. The overexpression of IRS1 in the endothelial cells of mice predisposed to atherosclerosis has, in fact, increased the insulin response through PI3K/AKT and significantly reduced atherosclerotic lesions even in the presence of insulin resistance, hyperglycemia, and hyperinsulinemia. The improvement was accompanied by lower expression of VCAM-1 and greater production of NO. The study also identified an additional mechanism: through IRS1/PI3K/AKT, insulin increased the endothelial expression of EDNRB (endothelin receptor type B), promoting the increase of eNOS. The loss of EDNRB instead produced a reduction in NO and an acceleration of atherosclerosis (Park et al., 2016) [19].

These results also help to interpret the relationship between hyperinsulinemia and atherosclerosis with greater caution. The frequent association between high insulin concentrations, insulin resistance, and cardiovascular risk could, in fact, lead one to automatically consider insulin itself as directly atherogenic. The experimental studies provided show a more complex picture. In ApoE−/− mice made obese, insulin-resistant, hyperglycemic, and hyperlipidemic through a high-fat diet, chronic administration of insulin notably increased insulinemia but, rather than aggravating atherosclerosis, it reduced its extent and complexity, increasing the activation of AKT and eNOS and the production of NO. The reduction of glycemia through phlorizin (a natural substance present especially in the bark and roots of the apple tree), in the absence of these effects of insulin, did not determine a comparable reduction in atherosclerosis (Park et al., 2018) [20].

A fundamental pathophysiological distinction thus emerges. Insulin resistance and hyperinsulinemia cannot be considered equivalent terms, nor can their vascular effects be automatically superimposed. The available evidence indicates that the loss of proper insulin signaling, especially in the endothelium, can remove important protective vascular mechanisms and favor atherogenesis. Conversely, when the insulin signal is still able to effectively activate pathways such as IRS1/PI3K/AKT/eNOS, insulin can support the production of NO, reduce endothelial activation, and limit some processes involved in plaque progression. In the experimental models analyzed, even a marked increase in insulinemia can therefore be accompanied by a reduction in atherosclerosis when it determines an effective restoration of these actions.

Overall, epidemiological and experimental evidence therefore converge on a model in which insulin resistance represents an important metabolic terrain associated with atherosclerosis, from the subclinical phase up to the presence and, potentially, the instability of plaques. HOMA-IR and especially TyG show associations with coronary calcification, arterial stiffness, carotid and coronary atherosclerosis, and plaque instability characteristics. In parallel, experimental studies indicate that part of the link may depend directly on the loss of normal insulin action in the vascular wall. At the center of this relationship, the ability of insulin to activate the PI3K/AKT cascade and support endothelial nitric oxide production emerges once again. It is precisely this step that now allows us to shift the focus from the general association between insulin resistance and atherosclerosis to the molecular mechanism that could contribute to linking them: the relationship between PI3K/AKT and nitric oxide.

In the previous paragraph, it emerged how the loss of the correct vascular response to insulin can contribute to the progression of atherosclerosis. To understand the mechanism through which this occurs, it is necessary to delve deeper into the endothelial cell and follow one of the main pathways through which insulin transforms its signal into a vascular response: the PI3K/AKT/eNOS cascade. At the end of this sequence lies nitric oxide, or NO, a small gaseous molecule that, despite its extremely simple structure, plays a fundamental role in maintaining vascular homeostasis.

Under physiological conditions, as we have already seen in chapter 4 of the insulin pathway, the binding of the hormone to its receptor leads to the activation of IRS-1 and IRS-2 substrates, which allow for the recruitment of phosphatidylinositol 3-kinase, or PI3K. The activation of PI3K, in turn, allows the signal to be propagated to AKT, a protein kinase that represents one of the main intracellular nodes of the insulin response. In the endothelium, one of the most important consequences of AKT activation is the stimulation of endothelial nitric oxide synthase, eNOS, the enzyme responsible for NO production. Thus, a functional sequence is formed — insulin → IRS-1/2 → PI3K → AKT → eNOS → NO — through which a metabolic signal is converted into a vascular response. The same cascade is associated not only with vasodilation but also with numerous vasoprotective actions, while in insulin resistance, its selective impairment contributes to endothelial dysfunction and the vascular complications of diabetes (Fu et al., 2021) [21].

The final step of this chain deserves particular attention. AKT does not produce NO directly, but modifies the functional state of eNOS through its phosphorylation. Once activated, eNOS increases the endothelial synthesis of NO; the latter can then diffuse from the endothelial cell toward the smooth muscle cells of the vascular wall, inducing their relaxation and promoting vasodilation. It is therefore possible to imagine PI3K and AKT as components of a transmission system: the insulin signal is progressively transferred until it reaches eNOS, which represents one of the final “executors” capable of transforming molecular information into a concrete change in vessel behavior.

The relationship between PI3K/AKT, eNOS, and NO is not just a theoretical reconstruction. A dynamic mathematical model of endothelial insulin signaling experimentally reproduced the rapid activation of the cascade after insulin stimulation: phospho-AKT reached its peak after about 2 minutes, phospho-eNOS even more rapidly, while NO production increased to about 325 nM after 5 minutes: the model, therefore, predicted a clear vasodilatory response under physiological conditions. When, on the other hand, AKT activation was reduced to one-hundredth of the normal value, NO production decreased by about a hundredfold (Muniyappa et al., 2020) [22].

This observation helps to better understand the meaning of so-called selective endothelial insulin resistance. It is not necessary, in fact, for all pathways activated by insulin to become resistant simultaneously. The IRS/PI3K/AKT branch can be preferentially compromised. Consequently, the compensatory increase in insulin does not necessarily succeed in restoring NO production: the problem lies not only in the amount of hormone available, but in the cell's ability to correctly transfer the signal along the PI3K/AKT/eNOS branch. The result is the loss of a fundamental component of endothelial protection.

Some studies allow us to identify with molecular precision where this signal can be interrupted. In endothelial cells, the activation of protein kinase C, or PKC (an enzyme that functions as a cellular switch), compromises the insulin cascade by intervening at the level of IRS substrates and PI3K. In particular, a reduction in the tyrosine phosphorylation of IRS-2 and, simultaneously, a lower association between IRS-1 and PI3K have been observed. The resulting lower activation of PI3K is followed by a reduction in AKT and eNOS. Particularly significant is the fact that even angiotensin II (a hormone that helps regulate blood pressure) can produce similar alterations through the activation of PKC, thus linking metabolic and neuro-hormonal signals to the loss of endothelial NO production (Maeno et al., 2012) [23].

The functional importance of the cascade is also confirmed by observing it in the opposite direction: when PI3K/AKT/eNOS is reactivated, NO production increases. In a study conducted on human endothelial cells, aortic rings, and a mouse model of hypertension, the experimental compound YPX-C-05 increased the phosphorylation of AKT and eNOS and endothelial NO production, accompanied by marked vasodilation. The use of PI3K/AKT/eNOS cascade inhibitors significantly attenuated the vasodilatory response, indicating that the activation of the pathway contributed functionally, and not just by association, to the observed effect. In hypertensive mice, chronic treatment also resulted in a reduction of blood pressure and vascular remodeling (Pang et al., 2024) [24].

The same relationship emerges in a completely different biological context, that of endothelial senescence. Senescent endothelial cells showed a reduction in PI3K/AKT/eNOS activity accompanied by lower NO availability. Treatment with a hydrogen sulfide donor instead increased PI3K, the phosphorylation of AKT and that of eNOS, simultaneously increasing NO content and reducing some markers of the senescent phenotype. When PI3K was pharmacologically inhibited with LY294002 (a substance used in scientific research to block PI3K activity), these effects disappeared. The result therefore reinforces the idea that PI3K represents a necessary step to effectively propagate the signal toward AKT, eNOS, and finally NO (Niu et al., 2023) [25].

The PI3K/AKT/eNOS axis is also not exclusively dependent on insulin. Other physiological signals can converge on the same cascade. VEGF, for example, through its receptor VEGFR2, activates PI3K and AKT, leading to the phosphorylation of eNOS and increasing NO production. This convergence is particularly important because it demonstrates that PI3K/AKT/eNOS constitutes a true functional node of the endothelium, used by different extracellular signals. In endothelial cells stimulated with VEGF, the sequential inhibition of VEGFR2, PI3K, and AKT phosphorylation was indeed accompanied by a reduction in eNOS phosphorylation and NO production, along with the suppression of proliferation, migration, invasion, and the formation of tubular structures (Guo et al., 2023) [26]. The significance of this study, in the context of atherosclerosis, does not consist in attributing to the inhibition of the cascade a necessarily favorable or unfavorable effect in every situation: it demonstrates rather that PI3K/AKT/eNOS represents a common platform through which different stimuli can modulate NO production and the behavior of endothelial cells.

A further element emerges by observing the lymphatic endothelium. Even in these cells, acute exposure to insulin activates PI3K/AKT, increases eNOS phosphorylation, and increases NO production. Conversely, prolonged exposure to high concentrations of glucose and insulin experimentally induced a condition of insulin resistance in which eNOS phosphorylation and NO production were reduced. At the same time, an increase in endothelial permeability, alteration of cell junctions, impairment of mitochondrial respiration, and activation of NF-κB (a complex of proteins that regulates the activity of some genes involved in inflammation, immune response, and cell survival) were observed, with an increase in pro-inflammatory mediators. Although the study concerns the lymphatic endothelium and therefore cannot be automatically transferred to the arterial endothelium, it demonstrates how the alteration of the PI3K/AKT/eNOS insulin response can be associated with cellular consequences much broader than the simple loss of vasodilation (Lee et al., 2018) [27].

The production of NO through PI3K/AKT/eNOS also does not exclusively influence the diameter of the vessel. In endothelial cells exposed to angiotensin II, the experimental activation of the Ca²⁺/PI3K/AKT/eNOS cascade increased eNOS phosphorylation and NO production, while inflammatory mediators, adhesion molecules, monocyte adhesion, and NF-κB activation decreased. The inhibition of NOS, the removal of Ca²⁺, or the pharmacological blockade of PI3K abolished these effects, showing that the integrity of the entire cascade was necessary to obtain the observed anti-inflammatory response (Zhang et al., 2024) [28].

It is thus understood why the impairment of the PI3K/AKT pathway can have consequences much deeper than the simple reduction of vasodilation. If AKT is not adequately activated, eNOS receives an insufficient signal, NO availability decreases, and the endothelium progressively loses a part of its protective characteristics. At the same time, other components of insulin signaling may remain relatively functional. The result is therefore not just a quantitative reduction in the action of insulin, but a qualitative change in the message received by the arterial wall: the PI3K/AKT/eNOS/NO branch that supports vascular homeostasis weakens, while vasoconstrictor and pro-inflammatory signals gain greater relative weight.

In the context of atherosclerosis, this imbalance takes on particular importance. The NO produced by the endothelium should not, in fact, be considered only as a molecule capable of dilating the vessel. The evidence analyzed shows that its availability is part of a broader network that involves inflammation, leukocyte adhesion, cellular function, and protection against plaque formation. The loss of the PI3K/AKT insulin signal can therefore transform, through the reduction of eNOS activity and NO production, into a progressive loss of the endothelium's ability to maintain its surface in a physiologically protective state. At this point, it becomes necessary to understand what one of the most important consequences of this loss is: how the reduction of nitric oxide can favor the expression of endothelial adhesion molecules, and in particular VCAM-1, preparing the arterial wall for the recruitment of inflammatory cells.

In the previous paragraph, we saw how the activation of the PI3K/AKT pathway leads to the phosphorylation of eNOS and the production of nitric oxide. This step is particularly significant in atherosclerosis because NO does not act solely on vascular tone. A relevant part of its function, in fact, consists of helping to maintain the endothelium in a quiescent, poorly adhesive state that is not prone to recruiting inflammatory cells. When the bioavailability of NO decreases, not only is a vasodilator stimulus lost, but so is one of the systems through which the inner surface of the vessel limits its own inflammatory activation.

In this process, vascular cell adhesion molecule 1, or VCAM-1, assumes particular importance. Under physiological conditions, its expression on the endothelial surface is very low; in the presence of inflammatory signals, however, it can increase considerably. Its appearance functionally modifies the endothelium because it promotes interaction with circulating leukocytes (also called white blood cells, these are blood cells that defend the body against viruses, bacteria, and other harmful agents), contributing to their adhesion and subsequent recruitment into the vascular wall. The increase in VCAM-1 therefore represents one of the steps through which a normally low-adhesion endothelium progressively acquires pro-inflammatory and pro-atherogenic characteristics.

One of the most direct demonstrations of the relationship between NO and VCAM-1 comes from experiments conducted on human endothelial cells stimulated with TNF-α (a protein produced mainly by immune system cells that acts as a chemical messenger, promoting inflammation). Experimentally increasing the availability of NO resulted in a 65% reduction in the cytokine-induced surface expression of VCAM-1. The effect of NO therefore appeared at least partially selective and particularly evident specifically against VCAM-1. Even more importantly, the reduction did not only concern the amount of protein present on the cell surface: the accumulation of VCAM-1 mRNA and the transcriptional activation of its promoter also decreased (Khan et al., 1996) [29].

The relationship becomes even more significant when considering that the simple reduction of NO produced physiologically by the endothelium can be sufficient to change the cell's behavior even without strong external inflammatory stimulation. In HUVECs (human cells taken from the umbilical cord vein), the inhibition of endogenous NO synthesis using L-NMMA (a substance used in scientific research to block the activity of the nitric oxide synthase enzyme) resulted in a concentration-dependent increase in VCAM-1, which at the experimental concentration of 5 mmol/L reached approximately 4.8 times the baseline value. The increase in the adhesion molecule also had a concrete consequence: the endothelial cells became more capable of retaining monocyte-like cells (cells that exhibit characteristics similar to monocytes). The decrease in NO did not, therefore, only produce a biochemical change, but functionally transformed the endothelial surface, making it more adhesive (Lazzerini et al., 2009) [30].

This relationship has also been observed through experimental models in which endothelial NO production is increased. Xanthotoxin (a natural substance present in some plants), in human endothelial cells, increased the phosphorylation of AKT and eNOS and the production of NO; simultaneously, it reduced the expression of VCAM-1 and ICAM-1 (a protein present on the surface of various cells, especially those lining blood vessels). When eNOS was inhibited, the anti-inflammatory effect was attenuated (Cao et al., 2020) [31]. The relationship is important because it functionally links the two steps developed in these paragraphs: upstream AKT → eNOS → NO, and downstream a lower expression of the molecules that make the endothelium capable of interacting with leukocytes.

A similar dynamic emerges from studies on progranulin (a protein produced by the body that helps regulate inflammation). In human endothelial cells, this protein increased the phosphorylation of AKT and eNOS, leading to an increase in NO production. In parallel, it reduced VCAM-1, ICAM-1, and MCP-1 (a protein that attracts monocytes to areas of the body affected by inflammation). The final consequence was a decrease in monocyte adhesion to endothelial cells. Furthermore, the pharmacological blockade of AKT prevented the activation of eNOS induced by progranulin, confirming the functional continuity between AKT/eNOS signaling, NO availability, and the control of endothelial inflammatory activation (Hwang et al., 2013) [32].

These findings also allow for a direct link between the insulin signal discussed in the previous paragraphs and the adhesive behavior of the endothelium. In a myocardial ischemia/reperfusion model, insulin activated AKT, increased eNOS phosphorylation, and increased NO production in endothelial cells. In parallel, P-selectin (a protein present on the surface of activated platelets) and ICAM-1 decreased, and leukocyte adhesion was reduced. The most mechanistically important aspect was represented by the inhibition experiments: by blocking AKT, the activation of eNOS, the increase in NO, and the anti-adhesive effect of insulin were abolished; similarly, the inhibition of eNOS or NO production prevented insulin from reducing leukocyte adhesion. Although VCAM-1 was not the molecule analyzed in this study, the results directly demonstrate that the sequence insulin → AKT → eNOS → NO can translate into a lower capacity of the endothelium to retain leukocytes (Li et al., 2009) [33].

The importance of the balance between NO, oxidative stress, and VCAM-1 emerges further in a DOCA-salt hypertension animal model (an experimental model of high blood pressure obtained through the administration of a synthetic hormone). Treatment with celiprolol (a beta-blocker drug used mainly to treat arterial hypertension) increased the expression and phosphorylation of eNOS and the phosphorylation of AKT; these effects were suppressed by the PI3K inhibitor wortmannin. The increase in NO and the reduction of oxidative stress can thus cooperate in shifting the endothelium toward a less adhesive and less pro-atherogenic phenotype (Kobayashi et al., 2003) [34].

The same principle is observable when the endothelium is subjected to a stimulus more directly attributable to atherogenesis. In murine aortic endothelial cells exposed to oxidized LDL lipoproteins, oxLDL caused a condition characterized by inflammation and endothelial dysfunction. The overexpression of the adipokine CTRP3 (a protein produced mainly by adipose tissue that helps regulate metabolism and inflammation) increased the phosphorylation of PI3K, AKT, and eNOS, restored the balance between NO and endothelin-1, and simultaneously reduced ICAM-1 and VCAM-1. The use of the PI3K inhibitor LY294002 or the NOS inhibitor L-NAME attenuated the protective effects, supporting the functional involvement of the PI3K/AKT/eNOS axis (Chen et al., 2019) [35].

A similar association between the recovery of PI3K/AKT/eNOS signaling and the reduction of adhesion molecules has also been observed in a diabetic brain injury model. In diabetic rats, eNOS and the p-PI3K/PI3K and p-AKT/AKT ratios were reduced, while ICAM-1 and VCAM-1 were increased. Treatment with 1,25-dihydroxyvitamin D3 reversed these modifications, increasing the activation of PI3K/AKT and eNOS and decreasing the two adhesion molecules, without a significant concomitant reduction in glycemia. Although the model concerns the diabetic microcirculation and not directly the atherosclerotic plaque, it reinforces the association between impaired PI3K/AKT/eNOS signaling and the acquisition of a more inflammatory vascular phenotype (Song et al., 2025) [36].

Experiments conducted in a cardiac cell model subjected to inflammatory stimulation also provide consistent indications, although caution is required in extrapolating to the arterial endothelium. Licochalcone C (a natural compound belonging to the flavonoid family, present in licorice root) increased the activation of the PI3K/AKT/eNOS pathway, simultaneously reducing the expression of ICAM-1 and VCAM-1. Blocking PI3K with LY294002 abolished most of these effects and again resulted in a reduction of eNOS activity and an increase in adhesion molecules (Franceschelli et al., 2017) [37].

Overall, these findings allow us to attribute a much broader function to NO than just the regulation of vascular caliber. Under physiological conditions, the tonic production of NO helps maintain the endothelial surface in a low-adhesion state. When, however, the bioavailability of NO decreases, a brake on redox signaling and the activation of transcription factors, especially NF-κB, is progressively removed. Thus, the transcription of VCAM-1 increases, and the endothelium becomes more capable of capturing and retaining circulating leukocytes. The sequence can therefore be schematized as: reduction of PI3K/AKT/eNOS → reduction of NO → greater pro-inflammatory activity of NF-κB → increase in VCAM-1 → greater endothelial adhesiveness.

This step represents a crucial point in the construction of the atherosclerotic lesion: the vessel is no longer characterized only by a lower vasodilatory capacity, but its surface progressively assumes the characteristics of an interface capable of promoting cellular recruitment and wall inflammation. However, it remains to be understood which of the stimuli is capable of sustaining and amplifying this transformation. Among the main candidates are oxidized LDLs, which have already emerged experimentally as being capable of simultaneously altering NO availability, adhesion molecules, and endothelial function. It is therefore on the relationship between insulin resistance and oxLDL that the reconstruction of the atherosclerotic process can now continue.

In the journey followed so far, we have seen how vascular insulin resistance can compromise the PI3K/AKT/eNOS pathway, reduce the availability of nitric oxide, and promote the endothelium's acquisition of a more inflammatory and adhesive phenotype. However, atherosclerosis does not depend solely on the response of vascular wall cells. Another fundamental step concerns what happens to the lipoproteins that transport cholesterol in the blood and, in particular, to LDL when they undergo oxidative modification.

Native LDL perform a normal lipid transport function. However, in an environment characterized by an excess of oxidizing species, the lipids and apolipoprotein B that make up these particles can undergo progressive modifications, generating oxidized LDL, commonly referred to as oxLDL. These are not simply LDL present in greater quantities, but particles that have acquired different biological characteristics. This aspect is particularly important because the metabolic alterations that accompany insulin resistance can simultaneously create favorable conditions for both the modification of LDL characteristics and an increase in oxidative stress.

Initial evidence of this relationship was obtained in non-diabetic overweight men. Lower insulin sensitivity was associated with higher concentrations of triglycerides and VLDL-TG and the presence of smaller LDL particles. Insulin sensitivity also showed an inverse correlation, albeit relatively modest, with oxLDL. However, the multivariate analysis was particularly interesting: when triglycerides and HDL-C were included in the model, insulin sensitivity lost its ability to independently predict LDL size. The data therefore suggest that part of the relationship between insulin resistance and LDL characteristics occurs through the alterations in lipid metabolism that accompany reduced insulin action, especially through the increase in triglycerides (Ho et al., 2002) [38].

The phenomenon does not appear to be confined to adulthood. In children and adolescents, oxLDL concentrations were found to be significantly higher in overweight and obese subjects compared to those of normal weight and were positively correlated with BMI (a value that relates a person's weight to their height), body fat percentage, waist circumference, trunk fat, and the amount of both visceral and subcutaneous adipose tissue. Even more significantly, oxLDL showed an inverse relationship with insulin sensitivity measured directly via euglycemic insulin clamp (a test used in scientific research to precisely measure the body's sensitivity to insulin): higher oxLDL levels corresponded to lower insulin sensitivity, and this association remained significant even after adjustment for adiposity. The relationship between insulin resistance and LDL oxidation, therefore, does not seem to be explained solely by the greater amount of adipose tissue (Kelly et al., 2010) [39].

It is important to emphasize this last point. Obesity and insulin resistance are closely associated, but they are not overlapping phenomena. If oxLDL increased solely as a consequence of body fat accumulation, the association with insulin sensitivity should attenuate completely once adiposity is statistically accounted for. The persistence of the association observed via clamp suggests instead that oxidative stress and reduced insulin sensitivity may represent interconnected metabolic components as early as childhood (Kelly et al., 2010) [39].

The relationship emerges with particular clarity, however, when insulin resistance and oxLDL are observed longitudinally. In obese men followed for a total of 32 months during a weight loss program, dietary counseling, exercise, and subsequent follow-up, changes in HOMA-IR showed a trend surprisingly parallel to those of oxLDL and the oxLDL/HDL-C ratio. During metabolic improvement, both HOMA-IR and oxidized LDL decreased; when insulin resistance subsequently increased again, oxLDL also began to rise. Subjects in the highest quartile of HOMA-IR had oxLDL and oxLDL/HDL-C values 23–51% higher than those in the lowest quartile (Linna et al., 2015) [40].

Even more interestingly, during follow-up, changes in oxLDL remained associated with fasting insulin and HOMA-IR even after controlling for BMI and waist circumference, while changes in fasting glycemia did not show a similar relationship. This does not prove that insulin is directly responsible for LDL oxidation, but it suggests that the association may be more closely linked to the alteration of insulin metabolism than to fasting glucose concentration alone. The authors emphasize, in fact, that the study design does not allow for the establishment of a mechanism or direct causality (Linna et al., 2015) [40].

The relationship becomes even more interesting when moving from fasting glycemia to metabolic oscillations following a glucose load. In 120 subjects with a BMI greater than 25 kg/m² who underwent an OGTT (a test that evaluates the body's ability to regulate glycemia) with 75 g of glucose, the progressive increase in HOMA-IR was accompanied by progressively higher concentrations of both oxLDL and TBARS, an indicator of lipid peroxidation (a harmful process in which free radicals attack fats). Subjects in the quartile with the highest insulin resistance therefore presented the highest oxidative load. The acute increase in glucose availability can also promote the production of ROS and processes of oxidation and glycoxidation of lipoproteins, indicating how postprandial metabolic excursions can occur in an environment already predisposed to oxidative stress (Bielawska et al., 2023) [41].

Data obtained in young people also suggest that the relationship is not necessarily identical in both sexes. In a population of 136 subjects, divided into individuals with and without insulin resistance, young women with IR had significantly higher concentrations of oxLDL compared to both women without IR and insulin-resistant men, while the same association did not emerge in men. The study also found a relationship between higher carbohydrate consumption and higher oxLDL concentrations. The authors hypothesize that compensatory hyperinsulinemia, hormonal differences, and alterations in lipid metabolism may contribute to the observed differences, but the cross-sectional nature of the study does not allow for the establishment of causal relationships (Marchán-Figueroa et al., 2024) [42].

The possibility of simultaneously modifying the metabolic and oxidative profiles also emerges from intervention studies. In obese Latin American adolescents with prediabetes, a 12-week program based on lifestyle modification resulted in a 7% reduction in fat mass, an improvement in triglycerides, total cholesterol, VLDL, and non-HDL cholesterol, along with an increase in cardiorespiratory fitness. During the same period, oxLDL decreased significantly from 51.0 to 48.7 U/L (Rentería-Mexía et al., 2019) [43].

A similar picture, but in a condition of more severe obesity, was observed after sleeve gastrectomy (a surgical procedure for the treatment of obesity). Twelve months after the surgery, 93 patients showed a simultaneous reduction in BMI and waist-to-hip ratio, improvement in the lipid profile, decrease in HOMA-IR, and reduction in leptin, hs-CRP, and oxLDL. Significantly, carotid intima-media thickness, an indicator of subclinical atherosclerosis, also decreased. Although the absence of a control group prevents attributing these effects directly to the surgical procedure, the concomitance between the improvement in insulin resistance, the reduction in oxLDL, and the favorable modification of a structural vascular parameter reinforces the link between metabolism, oxidative stress, and atherosclerosis (Hany et al., 2026) [44].

The relationship could also be bidirectional. A prospective multicenter study including 3056 participants, 1521 of whom had prediabetes followed for five years, showed that higher concentrations of oxidized lipoproteins were associated with a progressively higher risk of progression to type 2 diabetes. In particular, oxLDL-C and oxHDL-C correlated positively with HOMA-IR and negatively with HOMA-IS (an index that estimates the body's sensitivity to insulin); higher concentrations were also associated with lower β-cell function (Gu et al., 2026) [45].

Taken together, this evidence suggests that insulin resistance and oxLDL are not two independent phenomena. Reduced insulin sensitivity is frequently associated with hypertriglyceridemia, increased VLDL production, LDL remodeling, visceral adiposity, and increased oxidative stress: a set of conditions that can increase the probability that LDL will be oxidatively modified. At the same time, high concentrations of oxidized lipoproteins are associated with a progressively more compromised insulin metabolism; the available data outline a metabolic-oxidative circuit in which insulin resistance, lipoprotein alterations, and oxidative stress can reinforce each other.

The presence of oxLDL, however, introduces an additional problem. Once modified by oxidation, LDL no longer behave as simple particles dedicated to cholesterol transport: they acquire properties capable of directly modifying the behavior of vascular cells. It therefore becomes necessary to return to the mediator that we identified in the previous paragraphs as one of the main protective systems of the endothelium, nitric oxide, to understand what happens when this balance encounters oxidized LDL.

In the previous paragraph, we saw how insulin resistance is frequently associated with an increase in oxidized LDL, creating a link between altered insulin metabolism, oxidative stress, and atherosclerotic risk. At this point, it is necessary to return to nitric oxide, because evidence shows that NO and LDL oxidation are not independent phenomena. The NO produced in the vascular wall can, in fact, directly interfere with the oxidative transformation of LDL, thereby helping to prevent a normal cholesterol-transporting particle from progressively acquiring atherogenic characteristics.

LDL oxidation can be imagined, in very simplified terms, as a self-sustaining reaction. Once the first lipid radicals are formed, they can propagate oxidation to other molecules, generating new intermediates and keeping the peroxidation chain active. NO can intervene precisely in this sequence. Experiments conducted using SNAP (a chemical substance used in scientific research to release nitric oxide), sodium nitroprusside (a vasodilator drug that causes rapid relaxation of blood vessels), and directly dissolved NO have shown that the molecule can reduce oxygen consumption and the formation of compounds that arise during fat oxidation. The proposed mechanism consists of the ability of NO to react with free radicals, interrupting the propagation of the reaction. In this sense, NO can act as an antioxidant capable of “breaking the chain” of lipid peroxidation (Hogg et al., 1993) [46].

There is also another decisive variable: the speed at which NO is produced. Using donors characterized by different release kinetics, a biphasic relationship between NO production and LDL protection was observed. Therefore, the principle that “more NO means greater protection” does not simply apply. When oxidation proceeded rapidly, a relatively fast release of NO was necessary to achieve maximum inhibition; when peroxidation was slower, a more gradual release was more effective. Antioxidant activity therefore depends on the kinetic balance between the speed at which radicals are generated and the speed at which NO becomes available to intercept them (Goss et al., 1997) [47].

This concept helps to understand why NO produced physiologically and continuously by the vascular wall may have a different significance compared to a high concentration administered suddenly. Slow and controlled endothelial production could constitute a sort of permanent antioxidant defense, while the loss of this balance or a concomitant excessive production of oxidizing species could completely alter its effects.

The ability of endogenous NO to limit LDL oxidation has been observed directly in macrophages. Activated murine macrophages showed a lower capacity to oxidize LDL compared to unstimulated cells and, at the same time, produced NO. When NOS was inhibited with NMMA (a molecule that inhibits nitric oxide production), the ability to oxidize LDL was restored. Even by depriving cells of arginine, a substrate necessary for NO synthesis, the protective effect disappeared; reintroducing arginine increased nitrites again and decreased LDL oxidation. These observations establish an inverse functional relationship between cellular NO production and the oxidative modification of lipoproteins (Jessup and Dean, 1993) [48].

A similar relationship has been observed in vascular smooth muscle cells. The induction of iNOS via IL-1β increased NO production and simultaneously reduced indicators of LDL oxidation. By inhibiting NO synthesis, the protection was attenuated. Furthermore, SNAP donors dose-dependently reduced copper-induced oxidation. Therefore, cells other than the endothelium can also locally produce NO and modify the oxidative environment of the arterial wall (Rikitake et al., 1998) [49].

The relevance of this mechanism becomes even more evident when moving from cellular systems to experimental atherosclerosis. In ApoE−/− mice (genetically modified laboratory animals lacking the gene that produces apolipoprotein E) fed a Western diet, angiopoietin-2 (a protein that regulates the stability and functioning of blood vessels) increased endothelial NO production and resulted in a reduction of approximately 40% in the atherosclerotic area, without significantly changing plasma lipids. In the plaques, oxLDL and macrophages decreased simultaneously. In endothelial cells, Ang-2 reduced LDL oxidation by 57%, while NOS inhibition attenuated this effect and abolished the protection against atherosclerosis observed in the animals (Ahmed et al., 2009) [50].

A particularly significant result was obtained using an aspirin capable of releasing NO. In hypercholesterolemic mice deficient in the LDL receptor, twelve weeks of treatment with NCX-4016 reduced the cumulative area of aortic lesions by 39.8%, while conventional aspirin did not produce a comparable effect. At the same time, the susceptibility of LDL to oxidation, oxLDL in the arterial wall, and foam cells derived from macrophages decreased. These effects occurred without a reduction in cholesterolemia, suggesting that part of the protection could depend precisely on the modification of the oxidative state of lipoproteins and the vascular wall (Napoli et al., 2002) [51].

There is also evidence in humans, although more limited. In patients with stable coronary artery disease, four weeks of supplementation with L-arginine, the physiological substrate used by NOS to produce NO, significantly increased flow-mediated dilation and prolonged the time required for LDL to begin oxidizing by 27.1%. Plasma LDL concentrations did not change, indicating a modification of their susceptibility to oxidation rather than their quantity (Yin et al., 2005) [52].

Overall, a particularly important bidirectional relationship emerges. A normal availability of NO can limit the oxidative modification of LDL; at the same time, when oxidative stress increases, superoxide can rapidly inactivate NO and compromise the mechanisms necessary for its production. oxLDL themselves can also interfere with eNOS function, contributing to further reducing the availability of the molecule. A circuit can thus be established in which less NO favors an environment more permissive to LDL oxidation, while the increase in oxLDL further contributes to endothelial dysfunction and the loss of NO (Davignon and Ganz, 2004) [53].

However, the picture must be interpreted without attributing a universal antioxidant function to NO. Its action depends on the concentration, the speed of production, and above all on the redox context. It is therefore the balance between NO and oxidizing species, rather than the simple absolute quantity of the molecule, that determines the final direction of the biological response.

This balance assumes a central significance in atherosclerosis. If the endothelium progressively loses the ability to produce NO and oxidative stress increases simultaneously, the LDL present in the vascular environment can more easily become susceptible to modification. Once oxidized, however, these particles must be recognized by the cells of the arterial wall. It is precisely here that the next step of our reconstruction comes into play: oxLDL acquire the ability to interact with specific scavenger receptors (proteins present on the surface of some cells, especially those of the immune system, which act as cellular scavengers), among which CD36 assumes particular importance.

When an LDL particle undergoes oxidative modification, it is not just its chemical composition that changes: the way it is recognized by cells also changes. This step represents a fundamental turning point in the progression of the atherosclerotic process. Native LDL particles are normally handled through finely regulated receptor systems, whereas oxidation generates new molecular determinants that make the particle a ligand for a different family of receptors, known as scavenger receptors. Among these, CD36 occupies a particularly important position because it recognizes oxidized LDL with high efficacy and allows for their uptake by various cell populations, especially macrophages.

One of the most significant experimental demonstrations of this selectivity was obtained by studying macrophages derived from human monocytes. Blocking CD36 with a specific antibody reduced the binding of oxLDL by approximately 50% and their degradation by 26%. The effect was even more evident when human CD36 was artificially expressed in cells that normally possessed a much lower capacity to interact with these lipoproteins: the presence of the receptor increased the binding, internalization, and degradation of oxLDL by approximately fourfold. The same increase was not observed with native LDL, indicating that CD36 does not simply recognize an LDL particle, but shows a particular capacity to recognize the modifications generated by oxidation (Nicholson et al., 1995) [54].

The preference of CD36 for oxLDL was subsequently observed through a completely different approach, using atomic force microscopy to directly visualize, at the nanoscale, the interaction between different lipoproteins and the CD36 and SR-B1 receptors (a protein that functions as a receptor allowing the absorption of cholesterol transported by HDL). Although all analyzed lipoproteins showed some capacity for interaction, a precise hierarchy emerged for CD36: oxLDL > LDL > VLDL. Above all, oxLDL possessed a binding capacity with CD36 that was more than twice as high as the other lipoproteins examined. Oxidation therefore modifies the receptor properties of the particle, transforming it into a particularly effective ligand for CD36 (Gan et al., 2018) [55].

But which part of the oxidized LDL is actually recognized? Early experiments had already provided an important indication: eliminating the lipid component of oxLDL caused the increase in CD36-mediated binding to disappear. This suggested that the signal through which the receptor distinguishes an oxidized LDL from a native LDL resided at least in part in the lipids modified by oxidation (Nicholson et al., 1995) [54]. Subsequent studies have allowed for a better definition of this mechanism, showing that CD36 can recognize determinants present in both the lipid fraction and the protein component of oxLDL. The EO6 antibody, directed against specific oxidized phospholipid epitopes (parts of fats present in cell membranes modified by oxidation), reduced the binding of intact oxLDL to CD36 by approximately 70% and that of their lipid fraction by 85%. More than half of the recognition of oxLDL by CD36 could thus be attributed to oxidized phospholipids present freely in the lipid component (Boullier et al., 2000) [56].

More recent structural insights have pushed this interpretation even further. CD36 does not, in fact, seem to behave simply as a sort of "hook" present on the cell surface. Analyses based on mass spectrometry, molecular docking, and dynamic simulations have identified specific receptor residues distributed from the apex of the extracellular domain to the interior of a true lipid transport tunnel. CD36 might therefore not just recognize oxidized phospholipids, but contribute to their progressive translocation toward the cell membrane (Gao et al., 2025) [57].

This model is particularly interesting because it allows us to understand how oxidation can literally change the "receptor destiny" of LDL. A particle initially handled primarily through normal lipoprotein metabolism pathways acquires molecular structures that are recognized with great efficacy by scavenger receptors. This does not mean, however, that CD36 possesses an exclusively pathological function. Experiments conducted in vivo have shown that its role depends on the tissue and the metabolic context. In mice lacking CD36, in fact, the clearance of oxLDL was slowed, demonstrating that the receptor physiologically contributes to their elimination. In hepatic cells, this process can have a potentially protective significance because it promotes the removal of modified lipoproteins from circulation. The same receptor, when expressed in macrophages of the arterial wall, can instead promote their intracellular accumulation. CD36 must therefore be considered a system for the recognition and management of modified lipids whose biological significance depends strongly on the cell in which it is expressed (Luangrath et al., 2008) [58].

An additional element profoundly modifies the interpretation of CD36: the receptor does not only represent a gateway through which oxLDL can enter the cell. Their binding can trigger actual intracellular signaling cascades, and receptor recognition is considered an active event: the particle binds to CD36, and the receptor transforms this extracellular information into a signal capable of modifying the behavior of the macrophage (Rahaman et al., 2006) [59].

The study demonstrates that oxidized LDL (oxLDL) lipoproteins promote the formation of atherosclerotic plaques through the activation of macrophages. oxLDL binds to the CD36 receptor, which, thanks to the NKA protein (a protein that uses energy to expel sodium and introduce potassium), activates specific intracellular signals. This process promotes the accumulation of cholesterol in macrophages, transforming them into foam cells that remain trapped in the arterial walls. In experimental models, reducing NKA limited cholesterol accumulation and the development of atherosclerosis, suggesting new therapeutic possibilities to counteract plaque formation (Chen et al., 2015) [60].

The study demonstrates that oxidized LDL (oxLDL) lipoproteins promote atherosclerosis by increasing oxidative stress in macrophages. By binding to the CD36 receptor, they activate the PKM2 protein (an enzyme that helps cells derive energy from glucose through glycolysis), which reaches the mitochondria and increases the production of oxidizing substances (ROS). This process stimulates macrophages to absorb more material, promoting the accumulation of fats and the formation of foam cells in the arteries. The researchers suggest that limiting PKM2 activity and ROS production could help counteract atherosclerosis while preserving normal cellular defenses (Zhang et al., 2025) [61].

The significance of this interaction becomes even more evident under altered metabolic conditions. In human macrophages exposed simultaneously to high concentrations of glucose and oxLDL, a greater intracellular accumulation of lipids was observed, mediated primarily by CD36. The combination of the two stimuli also produced a more intense inflammatory response, with an increase in inflammatory messengers and a decrease in IL-10 (a protein that performs an important anti-inflammatory function). However, the authors highlighted an important distinction: while CD36 was central to lipid accumulation, it did not appear to be the main driver of the entire inflammatory response (Sanjurjo et al., 2023) [62].

CD36 is also expressed on platelets. In these cells, too, the recognition of oxLDL produces functional consequences: exposure to oxidized lipoproteins increases the formation of extracellular vesicles when platelets are subsequently stimulated with thrombin (an enzyme essential for blood coagulation) and collagen. Blocking CD36 largely suppresses this phenomenon. The same oxLDL-CD36 axis can therefore link the metabolism of oxidized lipoproteins not only to macrophages but also to the communication between platelets, coagulation, and immune cells (Maaninka et al., 2023) [63].

Overall, the oxidation of LDL represents much more than simple chemical damage to the lipoprotein. It generates new epitopes, especially at the level of oxidized phospholipids, which transform the particle into a high-affinity ligand for CD36. The receptor recognizes it, can facilitate its internalization, and simultaneously converts this recognition into intracellular signals capable of modifying metabolism, oxidative stress, phagocytosis, and cellular behavior. The sequence can therefore be summarized as: LDL → oxidation → formation of oxidized lipid epitopes → recognition by CD36 → uptake and activation of cellular signaling.

It is precisely the location where this process occurs, however, that determines most of its consequences. If the recognition and removal of oxLDL occur through systems dedicated to their clearance, CD36 can participate in the elimination of modified particles. When, instead, oxLDL are retained in the arterial wall and encounter monocytes and macrophages that express CD36, the same recognition system can take on a completely different meaning. It therefore becomes necessary to understand how the oxLDL-CD36 interaction, once transferred to the microenvironment of the arterial wall, contributes directly to the development and progression of atherosclerosis.

The recognition of oxLDL by CD36, described in the previous paragraph, takes on its most significant pathological meaning when it occurs within the arterial wall. Here, modified lipoproteins are not simply material to be eliminated: they encounter monocytes and macrophages, are taken up through scavenger receptors, and can fuel a process in which lipid accumulation, inflammation, and the persistence of cells in the lesion reinforce one another. CD36 occupies a central position in this step because it does not merely facilitate the entry of oxLDL, but converts their recognition into signals capable of profoundly modifying the macrophage's behavior. In this way, the oxLDL–CD36 axis can transform from a system for recognizing altered lipoproteins into a true mechanism for amplifying atherogenesis (Park, 2014) [64].

One of the clearest experimental pieces of evidence comes from models in which CD36 has been genetically eliminated. In mice lacking both apoE (a protein that helps transport and dispose of fats) and CD36 and fed a Western diet, the atherosclerotic area of the entire aortic tree was reduced by 76.5% compared to apoE-null animals (which do not produce apolipoprotein E) that retained CD36. Even with a normal diet, lesions in the aortic sinus were 45% smaller. In parallel, macrophages lacking CD36 showed a reduction of over 60% in the binding and internalization of modified LDL, accompanied by a decrease of at least 50% in lipid accumulation and foam cell formation. It is particularly significant that this protection occurred despite some potentially more atherogenic lipoprotein parameters. This suggests that the mere presence of lipoproteins susceptible to modification in circulation is not enough to explain the entire process: the ability of the lesion's cells to recognize and internalize them through CD36 also constitutes a determining step (Febbraio et al., 2000) [65].

This observation was subsequently reinforced using more genetically homogeneous models and by comparing CD36 with another important scavenger receptor, SRA I/II. After twelve weeks of a Western diet, the absence of CD36 reduced the area of aortic lesions by 61% in males and 74% in females. The simultaneous elimination of CD36 and SRA did not produce greater protection than the absence of CD36 alone, indicating that the two systems do not simply contribute additively to atherogenesis. Furthermore, macrophages lacking CD36 exhibited a less inflammatory phenotype, lower production of reactive oxygen species, and a different migratory response. CD36 thus emerges not only as the receptor responsible for lipid uptake, but as a component of a broader cellular response capable of contributing to the inflammatory environment of the lesion (Kuchibhotla et al., 2008) [66].

Within this process, a true self-amplification circuit can be established. Oxidized lipids internalized through CD36 can activate PPARγ (a protein that regulates the activity of certain genes involved in fat and sugar metabolism), which in turn increases the transcription of CD36 itself. The macrophage that has already encountered oxLDL can therefore acquire an even greater capacity to recognize and take up more. Consistently, tamoxifen (a drug used primarily to treat and prevent the recurrence of certain breast cancers) has been shown experimentally to interrupt this circuit: in ApoE−/− mice, it reduced atherosclerosis along with the expression of CD36 and PPARγ in the lesions, while in human monocytes and macrophages, it decreased CD36 and the intracellular accumulation of oxLDL. The effect depended on the functional inactivation of PPARγ and was lost in macrophages deficient in this transcription factor. The study therefore provides an indirect but important demonstration that modulating the PPARγ–CD36 axis simultaneously modifies oxLDL uptake and the atherosclerotic process (Yu et al., 2016) [67].

Further confirmation comes from direct interference with CD36. Salvianolic acid B (a natural compound belonging to the polyphenol family) has been identified as a non-peptide antagonist capable of binding directly to the receptor. The compound reduced the uptake of modified LDL in macrophages in a dose-dependent manner and also attenuated the increase in CD36 expression induced by oxLDL. When CD36 was silenced, the additional effect of the compound on lipid uptake substantially disappeared; conversely, artificial expression of the receptor conferred sensitivity to the treatment (Bao et al., 2012) [68].

The control of CD36 expression also appears much more complex than it might seem. Recent data identify PIM1 kinase (an enzymatic protein that regulates cell growth, survival, and multiplication) as an important upstream regulator of the receptor. In macrophages lacking PIM1, CD36 protein expression decreased by up to 96.7%, foam cell formation after exposure to oxLDL was reduced by 49.6%, and cellular cholesterol by 25.5%. Pharmacological inhibition of PIM kinases also reduced oxLDL binding by 64.5% and uptake by 57.9%. The most significant data, however, emerged in vivo: in Apoe−/− mice with myeloid deficiency of Pim1, plaque formation decreased by 69.4%. Molecular analyses indicated PPARγ as an important mediator of this relationship, outlining a PIM1–PPARγ–CD36 axis capable of modulating oxLDL uptake and atherogenesis (Beg et al., 2026) [69].

The relevance of CD36 becomes even more interesting when atherosclerosis is placed in the context of the metabolic alterations already discussed in the previous paragraphs. In ten healthy subjects, a 24-hour lipid infusion doubled circulating free fatty acids and reduced systemic glucose disposal by approximately 20% during a hyperinsulinemic-euglycemic clamp. Simultaneously, surface expression of CD36 on monocytes increased by approximately 25% and the internalization of oxidized LDL by 50%. CD36 expression was also inversely correlated with glucose disposal. The results show how a lipid-induced metabolic perturbation can be rapidly accompanied by the acquisition of a monocyte phenotype characterized by higher CD36 and a greater capacity to take up oxLDL (Kashyap et al., 2009) [70].

Glucose metabolism can also contribute to the regulation of this system. In monocytes from healthy subjects, high glucose concentrations transiently increased CD36 mRNA and non-glycated protein, while in normoglycemic, prediabetic, and diabetic atherosclerotic patients, more complex differences emerged, especially in the post-translational modifications of the receptor. The authors also observed signs of altered CD36 expression as early as the prediabetic condition. Clinical data indicate that CD36 could represent one of the points at which alterations in glucose and lipid metabolism converge on the cellular mechanisms of atherogenesis (Lopez-Carmona et al., 2017) [71].

A particularly significant example of environmental amplification of the oxLDL-CD36 axis is represented by nicotine. At concentrations comparable to those observed in smokers, it increased CD36 in monocytes/macrophages through a sequence associated with increased ROS and increased PPARγ. In the presence of oxLDL, nicotine accentuated lipid accumulation; silencing CD36 prevented these effects. Even more relevantly, in ApoE−/− mice, nicotine aggravated plaque accumulation, while this effect was not observed in mice simultaneously lacking CD36. The receptor thus appears as a point through which an external risk factor can amplify the atherogenic effects of oxidized lipoproteins (Zhou et al., 2013) [72].

The role of CD36 is also not limited to macrophages. Vascular smooth muscle cells can also accumulate lipids and contribute to the foam cell component of lesions. In rat aortic smooth muscle cells, exposure to oleic acid resulted in an approximately five-fold increase in intracellular triglycerides and a phenotype compatible with transformation into foam cells; inhibition of CD36 significantly attenuated both phenomena. In ApoE-null mice, the increase in free fatty acids induced by a diet containing oleic acid was accompanied by a greater extent of lesions, while their reduction via acipimox (a drug derived from nicotinic acid) partially attenuated atherosclerosis. Although not directly concerning oxLDL uptake, these results broaden the biological significance of CD36 by showing how the same receptor can integrate different atherogenic lipid signals in different cells of the vascular wall (Ma et al., 2011) [73].

Taken together, these findings allow us to move beyond a simplistic view of atherosclerosis as a mere passive deposit of cholesterol. The availability of oxLDL certainly represents an important prerequisite, but their capacity to produce pathological consequences also depends on how the cells of the arterial wall recognize them and respond to their presence. CD36 constitutes one of the main points of contact between these two levels. Interaction with oxLDL increases lipid uptake, can activate inflammatory signals, and can fuel amplification circuits that further increase receptor expression. Simultaneously, metabolic and environmental factors—insulin resistance, glycemic alterations, elevated free fatty acids, and nicotine—can modulate the same system and make monocytes and macrophages more predisposed to respond to oxidized lipoproteins.

The available evidence indicates with remarkable consistency that the oxLDL–CD36 axis represents one of the main mechanisms through which the oxidative alteration of lipoproteins is translated into lipid accumulation and the progression of the atherosclerotic lesion; furthermore, CD36 is also present on another cell population fundamental to cardiovascular complications: platelets. When oxLDL encounter platelet CD36, the meaning of their recognition changes again, because the signal can be transferred from the lipid accumulation processes of the vascular wall to those that regulate platelet activation and aggregation.

In the previous paragraph, we saw how the interaction between oxLDL and CD36 contributes to the formation and progression of atherosclerotic lesions. However, this same molecular pair takes on a different significance when CD36 is expressed on platelets. In this case, oxidized lipoproteins do not merely promote lipid accumulation; they can transform a state of dyslipidemia and oxidative stress into a greater predisposition to thrombosis. This step is particularly important because many of the most serious clinical consequences of atherosclerosis do not result simply from the progressive narrowing of an artery, but from the sudden formation of a thrombus at the site of a vulnerable plaque.

Initial experimental evidence showed that CD36 can act as a true platelet sensor for oxidative lipid stress. Under conditions of hyperlipidemia, specific oxidized phospholipids, known as oxPC_CD36, increase and are capable of binding to the receptor with high affinity. In hyperlipidemic mice, these lipids reached plasma concentrations up to approximately 40 times higher than in normolipidemic animals. Their binding to CD36 triggered the activation of the αIIbβ3 integrin and the exposure of P-selectin, two characteristic events of platelet activation. In mice lacking CD36, however, hyperlipidemia did not produce the same acceleration of thrombosis. Significantly, under normolipidemic conditions, CD36-deficient platelets maintained a substantially normal response to traditional agonists, suggesting that CD36 acquires particular importance precisely when the availability of oxidized lipids increases (Podrez et al., 2007) [74].

This means that oxLDL can behave more like a sensitization system than a simple on-off switch. Intuitively, a platelet exposed to oxidized CD36 ligands can be compared to a system in which the alarm threshold has been lowered: a stimulus that would produce a limited response under normal conditions can become sufficient to trigger a much more intense activation.

The study demonstrates that oxidized LDL (oxLDL) promotes thrombus formation by increasing platelet activation. oxLDL binds to the CD36 receptor, which activates a series of proteins, including FcRγ and Syk. These signals promote platelet aggregation and blood clot formation. In experimental models, the absence of FcRγ prevented these effects. The results suggest that blocking specific signals activated by CD36 could reduce the risk of thrombosis without interfering with the receptor's other functions (Wraith et al., 2025) [75].

The study demonstrates that oxidized LDL (oxLDL) increases the risk of thrombosis by promoting platelet activation. By binding to the CD36 receptor, oxLDL activates the NOX2 protein, which increases the production of oxidizing substances (ROS). These substances reduce the effectiveness of natural mechanisms that prevent excessive platelet activation, thereby favoring clot formation. In experimental models, blocking NOX2 restored these protective mechanisms, suggesting a possible strategy to reduce the risk of thrombosis associated with excess blood lipids (Magwenzi et al., 2015) [76].

The study demonstrates that oxidized LDL (oxLDL) promotes thrombus formation by reducing the mechanisms that prevent excessive platelet activation. oxLDL binds to the CD36 receptor and activates PDE3A (an enzyme that regulates certain signals within cells), which reduces levels of cAMP, a molecule that limits platelet activity. Consequently, platelets become less sensitive to the protective action of prostacyclin (PGI2, a substance produced by cells lining blood vessels that promotes vasodilation), increasing the risk of clots. In experimental models, blocking PDE3A countered these effects, suggesting new possibilities for preventing thrombosis (Berger et al., 2020) [77].

The significance of this system is particularly important in type 2 diabetes. Platelets from diabetic patients show not only increased expression of numerous glycoproteins involved in activation but also increased N-glycosylation of CD36. Plasma from diabetic patients, rich in oxLDL, was found to be capable of activating even healthy platelets and increasing their adhesion to collagen, while CD36 inhibition attenuated these phenomena. The increased glycosylation of the receptor has therefore been proposed as a possible factor capable of increasing platelet sensitivity to oxLDL, although the specific causal relationship requires further confirmation (Agarwal et al., 2024) [78].

Overall, the available evidence outlines a model in which oxLDL acts on platelets through multiple converging mechanisms. CD36 activates pro-aggregatory pathways, stimulates ROS production, and simultaneously weakens important physiological control systems. The end result is not necessarily a continuously aggregated platelet, but a hyperreactive platelet, in which the threshold required to respond to collagen, thrombin, and other agonists is progressively lowered. This interpretation explains why the interaction between oxidized LDL and platelets can contribute as much to the progression of atherosclerosis as to its most dangerous complication: arterial thrombosis (Gąsecka et al., 2021) [79].

However, the recognition of oxLDL by CD36 does not only produce these platelet pathways. As already observed in macrophages, the receptor can connect to numerous intracellular systems and profoundly modify the metabolism and fate of the cell. It therefore becomes necessary to shift the focus back from the platelet to the molecular signaling of CD36 to understand through which mechanisms the interaction with oxLDL can also converge on the major PI3K–AKT–mTOR metabolic axis.

In the previous paragraphs, CD36 emerged as much more than a simple receptor responsible for the uptake of oxidized LDL. Its binding with oxLDL can, in fact, transform into a true intracellular signal, modifying the behavior of the cell expressing the receptor. Among the molecular networks that may be involved, the PI3K/AKT/mTOR pathway is of particular interest. This connection deserves special attention because it allows for linking the recognition of a modified lipid present in the extracellular environment with one of the primary systems through which the cell regulates metabolism, growth, survival, and adaptation to environmental conditions.

A first particularly relevant demonstration comes from platelets. As observed in the previous paragraph, oxLDL interact with CD36 and increase the platelet response. Even when using stimulation levels below the threshold normally required to produce an intense response, activation resulted in an increase in phosphorylation along the PI3K/AKT axis. Functionally, this cooperation increased granule secretion, integrin activation, aggregation, phosphatidylserine exposure, and fibrin generation. Therefore, oxLDL did not act simply as an additional stimulus, but amplified a shared intracellular network in which PI3K/AKT represented one of the convergence points (Zheng et al., 2023) [80].

The relationship between oxLDL, CD36, and PI3K/AKT becomes even more evident in macrophages, which are central cells in the development of atherosclerotic plaque. In this case, however, the mechanism appears more complex because oxLDL contain different oxidized lipids capable of interacting simultaneously with multiple receptors. In human macrophages, it has been demonstrated that some phospholipids generated during LDL oxidation can activate the platelet-activating factor receptor, PAFR. The stimulation of PAFR and certain proteins leads to the activation of PI3K/AKT. The inhibition of PI3K reduced not only these intracellular responses but also the expression of CD36 and the ability of macrophages to take up further oxLDL. This creates a possible amplification circuit: oxLDL activate PAFR–PI3K/AKT, this signaling increases CD36, and greater receptor availability facilitates further uptake of oxidized lipoproteins (Rios et al., 2012) [81].

Another study demonstrates that the protein PKCδ (an enzyme that regulates cell activity through chemical signals) promotes the formation of foam cells, which are involved in the development of atherosclerosis. In macrophages, PKCδ regulates PI3K/AKT signaling pathways, increasing the expression of CD36 receptors. These receptors facilitate the uptake of oxidized LDL (oxLDL), promoting the accumulation of cholesterol in the cells. In experimental models, reducing PKCδ activity limited oxLDL uptake and cholesterol accumulation, suggesting a possible therapeutic target to counteract the formation of atherosclerotic plaques in the arteries (Lin et al., 2012) [82].

These findings are particularly important because they directly show, in cells relevant to atherosclerosis, a link between oxLDL, CD36, and PI3K/AKT. A particularly clear demonstration comes from hepatocellular carcinoma. In these cells, the experimental increase of CD36 increased the phosphorylation of Src (a protein that transmits signals within cells) and sequentially activated PI3K, AKT, and mTOR. Simultaneously, glycolytic flux and lactate production increased, while the pharmacological inhibition of PI3K, AKT, or mTOR significantly attenuated the metabolic and proliferative effects produced by CD36. In this model, a functional sequence CD36 → Src → PI3K → AKT → mTOR can therefore be reconstructed, through which a receptor traditionally associated with fatty acid metabolism becomes capable of reprogramming the cell's glucose metabolism as well (Luo et al., 2021) [83].

A consistent result emerges in gastric carcinoma, although through a ligand different from oxLDL. In this model, apolipoprotein C-II (APOC2, a protein present in the blood that helps the body use fats) interacted functionally with CD36 and caused an increase in the phosphorylation of PI3K, AKT, and mTOR. The experimental reduction of APOC2 or CD36 attenuated the activation of the entire cascade and simultaneously decreased proliferation, migration, invasion, and metastatic progression. In this case, too, CD36 appears capable of transferring information from the lipid environment to the PI3K/AKT/mTOR system (Wang et al., 2021) [84].

The relationship does not seem limited to tumor cells. In synovial fibroblasts obtained from patients with frozen shoulder (a condition characterized by inflammation and stiffening of the capsule surrounding the shoulder), CD36 was strongly overexpressed, and its experimental overexpression increased AKT phosphorylation. Analysis of the molecular pathways also indicated the involvement of the PI3K/AKT/mTOR axis. Inhibition of AKT using MK2206 (an experimental drug that blocks AKT activity) counteracted the modifications induced by CD36, functionally demonstrating that at least the PI3K/AKT component was necessary for the profibrotic response (Yan et al., 2023) [85].

Further confirmation of CD36's ability to control PI3K/AKT comes from even different systems. In HC11 mammary epithelial cells (mouse mammary cells used in scientific research), oleic acid increased CD36 and the intracellular concentration of Ca²⁺, followed by the activation of PI3K/AKT. Silencing CD36 prevented the increase in calcium and the activation of the cascade, while Ca²⁺ chelation in turn blocked PI3K/AKT. Finally, PI3K inhibition eliminated the proliferative effect. A sequence of oleic acid → CD36 → Ca²⁺ → PI3K/AKT was thus outlined, further demonstrating that CD36 can function as the starting point of a true signaling response and not just as a lipid transporter (Meng et al., 2018) [86].

Even in PC12 cells, used as a neuronal model, CD36 overexpression increased the phosphorylation of AKT and mTOR. In this particular model, the phenomenon was associated with exposure to aluminum, which interacts with CD36. The experimental increase of the receptor primarily intensified AKT phosphorylation, once again supporting the existence of a functional CD36–PI3K/AKT/mTOR link. The biological significance, however, was different from that observed in tumors, as intense activation of the cascade was associated with apoptosis. This data reminds us of an important principle: the activation of PI3K/AKT/mTOR does not necessarily produce the same result in every cell, because the intensity, duration of the stimulus, and metabolic context can profoundly modify its consequences (Li et al., 2023) [87].

A study on intestinal epithelial cells also identified CD36 as a possible upstream component in the regulation of PI3K/AKT. In an experimental model of colitis, a polysaccharide extracted from the eggs of Strongylocentrotus nudus modified CD36 expression and PI3K/AKT signaling, associating with the regulation of autophagy and an improvement in the intestinal barrier. The authors therefore proposed a CD36–PI3K/AKT–autophagy axis, while acknowledging that the precise molecular interactions require further investigation (Zhou et al., 2023) [88].

Taken together, these studies allow us to reconstruct a more precise picture. In atherosclerosis, there is direct evidence that oxLDL can involve CD36 and converge on PI3K/AKT signaling, both in platelets and in macrophages. Furthermore, in macrophages, PI3K/AKT can contribute to the regulation of CD36 expression itself, generating a potential reinforcement circuit in which greater exposure to oxidized lipoproteins favors cellular systems that facilitate further uptake. In other biological models, it has been shown that signaling associated with CD36 can extend to mTOR, outlining a complete CD36 → PI3K → AKT → mTOR axis.

What emerges most solidly is a broader concept: CD36 represents a connection point between the extracellular lipid environment and intracellular metabolic signaling programs. When the cell encounters modified lipids, the receptor does not merely allow for their recognition or entry, but can modify PI3K/AKT and mTOR. In the atherosclerotic wall, this connection takes on particular significance in macrophages, because the same signaling that accompanies oxLDL uptake can increase the expression of scavenger receptors and promote further lipid acquisition. The next step therefore becomes inevitable: understanding how this CD36–PI3K/AKT network, and its connection with mTOR, can accompany the progressive transformation of the lipid-laden macrophage into the characteristic foam cell of atherosclerotic plaque.

The transformation of cells within the arterial wall into foam cells represents one of the most characteristic steps of atherosclerosis. Their appearance, due to the abundance of lipid droplets accumulated in the cytoplasm, is the visible manifestation of a deeper imbalance: the amount of lipids entering the cell progressively exceeds its capacity to utilize, degrade, or eliminate them. For a long time, this phenomenon was attributed almost exclusively to macrophages derived from monocytes. Today, however, we know that a substantial portion of foam cells can also originate from vascular smooth muscle cells. In mouse models, this component can reach up to 70%, while in human lesions, it has been estimated at around 50%. The foam cell should therefore not be considered a metabolically uniform population: its ability to manage lipid overload also depends on the cell from which it derives (Miller & Zhang, 2021) [89].

In the case of macrophages, however, CD36 represents one of the main entry points through which oxidized LDL are recognized and internalized. The quantitative importance of this system is considerable: CD36 and SR-A have been indicated as being responsible for approximately 75–90% of the uptake of modified lipoproteins by macrophages. The experimental reduction of CD36 can, in fact, translate directly into reduced foam cell formation. An example is provided by intermedin, an endogenous vasoactive peptide that, in ApoE−/− mice, reduced the uptake of oxLDL, intracellular cholesterol, and foam cell formation. The mechanism did not depend on the reduction of CD36 transcription, but on the lower stability of its mRNA: less mRNA meant less CD36, lower oxLDL entry, and, consequently, less lipid accumulation (Dai et al., 2014) [90].

The amount of CD36 available on the macrophage surface can, however, also be regulated by intracellular signaling. Another important study for the central theme of this chapter demonstrated that insulin can increase the protein expression of CD36 in human monocyte-derived macrophages by approximately twofold. This modification was associated with an approximately 80% increase in oxLDL uptake. In just 16 hours, the treated macrophages accumulated two to three times more cholesterol. The inhibition of PI3K prevented the insulin-induced increase in CD36, indicating that at least this component of the phenomenon depended on PI3K/AKT signaling (Park et al., 2012) [91].

This finding takes on particular significance in the context of insulin resistance and compensatory hyperinsulinemia. Insulin resistance, in fact, does not necessarily imply that every intracellular response to insulin is abolished to the same extent. The study therefore suggests the possibility that, while some metabolic effects of insulin are compromised, residual signaling through PI3K may continue to promote CD36 expression in macrophages.

Another supporting study demonstrates that the protein zDHHC1 (an enzyme that modifies certain cellular proteins by adding fat molecules to them) promotes the development of atherosclerosis through the PI3K/AKT/mTOR pathway. In macrophages, zDHHC1 activates PI3K, contributing to the activation of intracellular signals. This process promotes the uptake of oxidized LDL and the formation of foam cells, which are involved in the development of arterial plaques. In experimental models, eliminating zDHHC1 reduced AKT and mTOR activation, foam cell accumulation, and the formation of atherosclerotic plaques, suggesting a possible therapeutic target to combat atherosclerosis (Zhou et al., 2025) [92].

Convergent results were obtained with rutin (a natural flavonoid belonging to the polyphenol family) in RAW264.7 macrophages (immune cells derived from mice, cultured in the laboratory) exposed to oxLDL. The study demonstrates that rutin can counteract some processes involved in atherosclerosis. In macrophages exposed to oxidized LDL, rutin reduces the activation of the PI3K/AKT pathway, promoting autophagy (a cellular process that eliminates damaged components). This mechanism limits lipid accumulation and foam cell formation. Furthermore, rutin reduces inflammatory signals and promotes the polarization of macrophages toward the M2 phenotype, suggesting a possible protective effect against atherosclerosis (Li et al., 2022) [93].

Another study by the author demonstrates that the CXCR4 receptor (a receptor involved in inflammatory processes and the spread of some tumors) promotes the progression of atherosclerosis through the activation of the PI3K/AKT/mTOR pathway. In macrophages, this activation reduces autophagy, a process that allows cells to eliminate damaged components and regulate lipid accumulation. In experimental models, reducing CXCR4 expression promoted autophagy, limited the formation of atherosclerotic plaques, and improved cardiac tissue structure. These results suggest that CXCR4 could represent a possible therapeutic target to combat atherosclerosis and its cardiovascular consequences (Li et al., 2023) [94].

A mechanism even closer to the relationship between oxLDL and CD36 involves CD147 (a protein that contributes to cell communication by regulating various processes). In human and murine atherosclerotic lesions, this protein was particularly abundant in regions rich in macrophages and foam cells. oxLDL increased its expression through PI3K/AKT/mTOR: the inhibition of PI3K, AKT, or mTOR counteracted the increase in CD147, and rapamycin produced the most pronounced effect. CD147, in turn, promoted oxLDL uptake and the accumulation of cholesterol esters, likely through the regulation of CD36. The specific deletion of CD147 in macrophages, in fact, reduced foam cell formation. This configures a potential pro-atherogenic circuit: oxLDL → PI3K/AKT/mTOR → CD147 → CD36 → increased oxLDL uptake, capable of self-sustaining the macrophage's lipid overload (Lv et al., 2021) [95].

Overall, the evidence allows us to interpret foam cell formation as the result of an altered balance between lipid entry, accumulation, degradation, and efflux. CD36 acts primarily on the first aspect, facilitating oxLDL uptake. PI3K/AKT can promote CD36 expression in certain contexts, while mTOR activation can limit autophagy, reducing the macrophage's ability to mobilize and dispose of its lipid load. In this way, the cell can simultaneously find itself in a condition of acquiring more lipids and eliminating them with less efficiency.

It is precisely this convergence that makes the PI3K/AKT/mTOR node particularly interesting in atherosclerosis. It is not simply a matter of establishing whether the pathway is "on" or "off," but of understanding which of its components are active, in which cell, with what intensity, and for how long. In macrophages chronically subjected to oxLDL, signaling that promotes CD36 while simultaneously limiting autophagy can contribute to foam cell transformation; in other contexts, however, PI3K may utilize effectors other than AKT/mTOR. The origin of the foam cell also profoundly modifies the result, as demonstrated by the lysosomal differences between macrophages and smooth muscle cells. The atherosclerotic plaque thus emerges not as a simple passive deposit of cholesterol, but as an environment in which insulin signaling, scavenger receptors, lipid metabolism, and cellular recycling systems intersect continuously. This framework lays the foundation for the final step of our journey: understanding how the oxLDL–CD36–PI3K/AKT/mTOR axis can, in turn, intertwine with insulin resistance, closing the circuit between systemic metabolic alteration and cellular dysfunction of the arterial wall.

The path followed so far leads to a particularly important point: the relationship between insulin resistance, oxidized lipids, and CD36 does not appear to proceed in only one direction. Insulin resistance can modify the quantity and function of CD36, favoring greater lipid exposure in certain cells; at the same time, excessive activation of CD36 by oxLDL or fatty acids can interfere with the very cascade through which insulin exerts its metabolic effects. This outlines a possible circuit in which impaired insulin signaling and lipid dysregulation tend to progressively fuel each other.

One of the clearest demonstrations of the first direction of this relationship comes from macrophages. In obese ob/ob mice, characterized by severe insulin resistance, macrophages showed reduced expression and signaling of the insulin receptor, accompanied by an increase in the CD36 protein and greater uptake of oxLDL. A particularly interesting aspect was that the increase in CD36 did not stem from greater production of the relative mRNA, but from a post-transcriptional mechanism that slowed its catabolism, increasing its cellular availability. Inhibition of PI3K using LY294002 in normal macrophages reproduced the increase in CD36, as did the genetic loss of the insulin receptor. Conversely, improving insulin signaling in ob/ob mice was accompanied by a reduction in the CD36 protein. The proposed sequence is therefore extremely significant: defect in insulin signaling → reduction in PI3K-dependent activity → altered CD36 turnover → increase in CD36 → greater uptake of oxLDL (Liang et al., 2004) [96].

However, the relationship can also proceed in the opposite direction. In adipocytes, in fact, oxLDL are capable of directly compromising the insulin response, and CD36 participates in this process. In 3T3-L1 adipocytes (cells of murine origin), exposure to oxLDL reduced glucose uptake by approximately 40% and the translocation of GLUT4 to the plasma membrane by approximately 70%. The defect did not depend on a reduction in GLUT4 expression, but on the alteration of the mechanisms that normally allow its recruitment in response to insulin. Blocking CD36 countered the reduction in glucose uptake, indicating that the interaction of oxLDL with this receptor contributes directly to metabolic impairment (Scazzocchio et al., 2009) [97].

Another study demonstrates that the CD36 receptor promotes inflammation and insulin resistance associated with obesity and hyperlipidemia. In adipocytes, oxidized LDL binds to CD36, activating specific proteins that alter the function of IRS-1, compromising insulin signaling and glucose uptake. Furthermore, CD36 promotes the inflammatory interaction between adipocytes and macrophages. In experimental models, the absence of CD36 reduced inflammation and improved insulin sensitivity, suggesting a possible therapeutic target against metabolic alterations (Kennedy et al., 2011) [98].

The relevance of this mechanism is not limited to murine models. Human visceral adipocytes obtained from both non-obese subjects and people with severe obesity have shown the ability to internalize oxLDL and progressively accumulate esterified cholesterol. This result suggests a possible amplification mechanism: the more oxLDL modify the adipocyte phenotype, the more the cell increases some of the systems through which it can recognize modified lipoproteins, while its ability to respond to insulin decreases. The study demonstrates the coexistence of oxLDL accumulation, increased scavenger receptors, inflammation, and loss of insulin response in human adipocytes (Santiago-Fernández et al., 2020) [99].

Recent data obtained in childhood obesity also reinforce the link between CD36 and the alteration of the insulin pathway. A study conducted on 50 children with obesity and 50 healthy controls highlights an association between the CD36 receptor, inflammation, and insulin resistance. In adipocytes, CD36 promotes lipid accumulation and activates a specific inflammatory pathway. This process is associated with a reduction in IRS1 and GLUT4, compromising the insulin response. In cellular experiments, reducing CD36 increased AKT activation, restored IRS1 and GLUT4, and decreased inflammation, suggesting a possible therapeutic target in childhood obesity (Yao et al., 2026) [100].

The relationship between CD36 and insulin sensitivity becomes even more interesting when the focus shifts to skeletal muscle, one of the main tissues responsible for insulin-stimulated glucose utilization. In mice fed a Western diet, CD36 expressed by endothelial cells favored the transfer of fatty acids to the muscle, contributing to the intramuscular accumulation of FFA (free fatty acids present in the blood) and lipids, mitochondrial dysfunction, and reduced PI3K/AKT signaling. Selective deletion of endothelial CD36, on the other hand, attenuated hyperglycemia, hyperinsulinemia, increased HOMA-IR, and glucose intolerance and, simultaneously, restored PI3K/AKT, increased total and membrane GLUT4, and improved glucose utilization. In this case, CD36 therefore also acts indirectly on insulin sensitivity by controlling the amount of lipids that cross the endothelial interface and reach the muscle tissue (Dada et al., 2024) [101].

A similar mechanism was recently identified in the same vascular compartment. In the presence of obesity or excessive exposure to palmitic acid, CD36 promotes the mitochondrial production of reactive oxygen species, which subsequently activate endoplasmic reticulum stress (a cellular laboratory that produces, modifies, and transports proteins and fats). This sequence compromises insulin-stimulated AKT and eNOS phosphorylation and reduces the insulin-mediated vasodilatory response. Reducing CD36 in mice fed a high-fat diet attenuated ROS, endoplasmic reticulum stress, fibrosis, and vascular thickening, while in human endothelial cells, its silencing countered the effects of palmitate and restored the AKT/eNOS response. Experiments with mitochondrial antioxidants and an endoplasmic reticulum stress inhibitor also allowed for the definition of the order of events: CD36 → mitochondrial ROS → endoplasmic reticulum stress → impairment of AKT/eNOS insulin signaling (Xu et al., 2026) [102].

These results might suggest that reducing CD36 is always metabolically advantageous. However, the available evidence shows that such a conclusion would be overly simplistic. In skeletal muscle, CD36 also performs a physiological function necessary for proper insulin action. Its specific deletion actually reduced glucose uptake and utilization, while in human myotubes (immature muscle cells formed by the fusion of multiple precursor cells), CD36 depletion attenuated insulin signaling. Mechanistically, CD36 interacts with the insulin receptor through PI3K. Saturated fatty acids, particularly palmitate, rapidly interrupted this interaction. CD36 can therefore physiologically participate in the propagation of the CD36-PI3K signal, while an environment characterized by an excess of saturated lipids can transform this same metabolic interface into an element of dysfunction (Samovski et al., 2018) [103].

Context-dependence was confirmed by directly comparing normal conditions and lipid overload. In specific myotubes, silencing CD36 in the absence of excess palmitic acid reduced insulin-stimulated AKT phosphorylation and was associated with mitochondrial dysfunction and endoplasmic reticulum stress. In the presence of palmitate overload, however, the same reduction in CD36 produced the opposite effect: it limited the excessive entry of fatty acids, reduced lipid accumulation, and improved insulin signaling. These data therefore suggest that there may be a functional range in which CD36 contributes to metabolic homeostasis, while its activity becomes unfavorable when lipid supply persistently exceeds the cell's ability to utilize it correctly (Sun et al., 2023) [104].

The scientific research seen so far shows that fat metabolism and the response to insulin are closely linked. An important role is played by CD36, a protein that facilitates the entry of fatty acids into cells and recognizes oxidized LDL (oxLDL).

This relationship can develop in both directions. When insulin signaling is reduced, some cells increase the amount of CD36, favoring the absorption of oxLDL. Conversely, excessive exposure to oxLDL can compromise the insulin response, reducing the transfer of GLUT4 to the membrane and glucose entry.

In adipocytes, CD36 can also promote inflammation and alter IRS-1, a protein essential for transmitting the insulin signal; studies on human cells and childhood obesity confirm the link between lipid accumulation, inflammation, and reduced activity of IRS-1, AKT, and GLUT4.

Even in muscle, excessive fat entry favored by CD36 can compromise PI3K/AKT, reducing glucose utilization. In vascular cells, CD36 can also promote oxidative stress and alterations in the insulin response.

However, CD36 is not necessarily a harmful protein. In muscle, it also performs important physiological functions, contributing to the correct transmission of the insulin signal. The problem emerges primarily when there is exposure to oxidized LDL: under these conditions, limiting CD36 can improve the insulin response, whereas reducing it under normal conditions can worsen it.

Overall, the balance between lipid availability, CD36 activity, and PI3K/AKT signaling is fundamental to maintaining insulin sensitivity.