Sodium-Glucose Cotransporter-2 Inhibitors in Mood Disorders: A Narrative Review of Mechanisms, Evidence, and Challenges

Mood Disorder and Metabolic Syndrome

Mood disorders are a group of psychiatric diseases characterized primarily by significant and persistent disturbances in affect or mood, with major clusters including major depressive disorder (MDD) and bipolar disorder (BD). The 12-month prevalence of depressive disorder is estimated to be around 6% globally, with a lifetime risk of 15–18%.¹ The lifetime prevalence of bipolar I disorder worldwide is approximately 0.6%, while bipolar II disorder has a slightly higher prevalence, ranging from 0.4% to 1%.² Both conditions are strongly associated with reduced quality of life and increased mortality risk. Specifically, depression ranked 13th leading cause of disability-adjusted life years (DALYs) in 2019 and 2nd as a cause of years lived with disability (YLDs) globally.³ Bipolar disorder is associated with a potential reduction in life expectancy of approximately 10–20 years.² Among the excess mortality observed in patients with severe mental illnesses, including MDD and BD, about 60% is attributable to comorbid physical diseases, particularly cardiovascular conditions,⁴ indicating the possible relation between mood disorders and metabolic disorders.

Metabolic syndrome (MetS) directly increases the risk of cardiovascular disease, type 2 diabetes mellitus (T2DM), and all-cause mortality. Various criteria have been proposed for the clinical diagnosis of MetS.^5–7 Though varying across organizations, they typically represent a cluster of interrelated conditions including abdominal obesity, insulin resistance or hyperglycemia, dyslipidemia, and hypertension.⁸

Consistent with the elevated risk of cardiovascular mortality, patients with mood disorders show a significantly increased comorbidity rate with MetS, approximately 31.3% and 31.7% for MDD and BD, respectively.⁴ Conversely, individuals with metabolic disorders also exhibit a heightened risk of mood disorders. For instance, the prevalence of MDD in patients with T2DM is 14.5%, significantly higher than in the general population (OR = 1.73).⁹ These observations suggest a potential shared biological foundation between the two conditions.

This comorbidity exacerbates the severity of mood disorders and alters the disease course. The presence of MetS in older adults with depression is associated with greater symptom severity and chronicity.¹⁰ In BD, metabolic risk factors have been associated with poor treatment response, chronic course of illness, increased risk of rapid cycling, and worse global functioning.¹¹ Both metabolic dysfunction and mood disorders are also independently associated with cognitive deficits,¹² with each component of MetS negatively affecting cognitive performance.¹³ The most pronounced impairments manifest in working memory, followed by attention, executive function/processing speed, global cognition, verbal memory, and visual memory.¹²

Taken together, these findings suggest that the metabolic burden associated with mood disorders is not merely a comorbidity but an important factor that may worsen prognosis and complicate treatment. In this context, current antidepressants and mood stabilizers remain limited by incomplete response, relapse, residual cognitive symptoms, and, in some cases, adverse metabolic effects. Therefore, it may be a new direction to improve the prognosis of mood disorders by exploring treatment strategies from the perspective of metabolic mechanisms. Several metabolic modulating agents like lipid-lowering agents and anti-diabetic drugs, including statins, thiazolidinediones, glucagon-like Peptide 1 (GLP-1) agonists, and metformin, have been evaluated for the treatment of mood disorders.^14–16

Sodium-glucose co-transporter 2 inhibitors (SGLT-2i), as an emerging class of anti-diabetic agents, act through inhibiting SGLT-2 receptors in the proximal renal tubules, thereby reducing blood glucose levels by preventing glucose reabsorption from the urine and inducing a negative energy balance.¹⁷ They may be of particular interest because they exhibited the most significant association with the reduction of depression onset in diabetic patients among many agents including GLP-1 agonists and metformin.^18–20 Moreover, these agents have demonstrated therapeutic benefits across multiple organs, with robust clinical evidence supporting their efficacy in cardiovascular and renal diseases through mechanisms including ketogenesis which might also benefit in mood disorders, anti-oxidation and anti-inflammation, and increasing evidence suggesting their potential neuroprotective and cognitive improvement benefits in neurological conditions such as Alzheimer’s disease.^21–24

Overall, the rationale for examining SGLT-2 inhibitors in mood disorders rests on three considerations: the shared pathophysiological basis of metabolic and affective dysregulation, the persistent unmet therapeutic needs in mood disorders, and the possibility that SGLT-2i may offer a distinct translational profile among metabolic-modulating agents. Within this metabolic–psychiatric framework, SGLT-2i warrant discussion as a mechanistically relevant but still clinically preliminary approach in mood disorders. This review will summarize the established preclinical and clinical evidence regarding the potential effects of SGLT-2i in mood disorders, and then outline their underlying mechanisms targeting the frequent comorbidity of metabolic dysfunction in mood disorders, exploring their future clinical applications.

SGLT-2 and SGLT-2i

SGLT-2 belongs to a group of sodium-dependent symporters that mediate the transport of various substances.²⁵ SGLT-2 itself is a glucose transporter that functions against a concentration gradient, transporting one Na⁺ and one D-glucose molecule, which is thought to be responsible for 80%–97% of renal glucose reabsorption.²⁶

SGLT-2 is predominantly located on the apical membrane of the brush border in renal proximal tubule cells, with emerging research revealing that SGLT-2 may also be expressed in tissues beyond the kidney. Notably, the brain depends heavily on glucose as its primary energy source, demanding glucose to cross several barriers linked with tight junctions via transporter proteins.²⁷ In vivo PET imaging in rats with Me-4-FDG revealed widespread functional expression of SGLTs in the brain, including hippocampal pyramidal neurons, cerebellar Purkinje cells, the frontal cortex, caudate-putamen, amygdala, parietal cortex, and hypothalamic paraventricular nucleus.²⁸ Among SGLTs, though SGLT-2 expression in the brain is lower than SGLT-1, co-expression of SGLT-1 and SGLT-2 has been reported in several CNS regions.²⁹ Human SGLT-2 mRNA has been detected in the cerebellum.²⁵ Evidence from animal models suggests that SGLT-2 may localize in the brain microvasculature of the blood–brain barrier (BBB),³⁰ the choroid plexus and ependymal cells facing the CSF,³¹ as well as in the hypothalamus, amygdala, periaqueductal gray (PAG), and the nucleus tractus solitarius (NTS) involved in autonomic control,³² such distribution patterns indicate that SGLT-2 may participate in the regulation of learning, food intake, energy, and glucose homeostasis, central cardiovascular and autonomic functions,³³ supporting the rationale for their potential involvement in metabolic and neuropsychiatric diseases.

SGLT-2i has become a cornerstone in the treatment of T2DM. Widely approved agents nowadays include empagliflozin, dapagliflozin, and canagliflozin. These agents differ in their pharmacokinetic profiles and selectivity. While none is absolutely selective for SGLT-2, empagliflozin exhibits the highest selectivity (SGLT-2/SGLT-1 about 2500), followed by dapagliflozin (about 1200), and canagliflozin (about 250).³⁴

Beyond glycemic control, accumulating evidence suggests broader therapeutic potential for SGLT-2i. Notably, due to their lipophilicity, SGLT-2i are capable of crossing the BBB and exert direct effects on the CNS. Among them, empagliflozin shows the highest brain-to-serum area under the curve (AUC) ratio at 0.5, compared to 0.3 for dapagliflozin and canagliflozin.³⁵ Such property supports their potential utility in the treatment of neuropsychiatric disorders. Their potential role in mood disorder treatment will be reviewed in the following section.

Application of SGLT-2i in Mood Disorders

Evidence from Preclinical Animal Models

The accumulating animal experimental data have demonstrated the role SGLT-2i might play in the treatment of mood disorders. As anti-diabetic agents, SGLT-2i exert regulatory effects on diabetes-induced depressive behaviors in diabetic animal models. In DM rats, systemic or local administration of dapagliflozin directly acts on the LHb, suppressing its neuronal activity thereby producing anti-depressant effects.³⁶ High-dose empagliflozin (10 mg/kg) daily oral administration alleviates depressive-like behaviors induced by olfactory bulbectomy in DM rats, which could be attributed to its antioxidant and brain-derived neurotrophic factors (BDNF) modulating properties.³⁷

The anti-depressant potential of SGLT-2i has also been demonstrated across various depression-related animal models. In rats subjected to chronic unpredictable stress (CUS), all the mentioned SGLT-2i have been found effective to reverse the resulting pathological changes in the hippocampus, such as upregulated inflammatory and oxidative markers, as well as improve behavioral performance, indicated by shortened immobility in tail suspension and forced swim tests and normalized sucrose preference.^38–40 In mice exposed to single-prolonged stress (SPS), dapagliflozin treatment reversed depressive-like behaviors and normalized SPS-induced increases Crh, Bax, Il1b, and Bdnf mRNA levels in brain, as well as corticosterone level in serum.⁴¹ As noted above, endocrine dysregulation can promote depression development. Ovariectomized rats exhibit depressive-like behaviors which could be reduced by empagliflozin, likely through regulation BDNF and AMPK pathways.⁴² Which is also suggested in reserpine-induced depressive animals.⁴³ These actions ultimately improve hippocampal neuroplasticity and restore the balance of monoamines, autophagy, and inflammation, contributing to the reversal of behavioral, biochemical, and histopathological alterations associated with depression.⁴³

Recent studies have also explored the effects of SGLT-2i on BD. In a rat model of mania-like behavior induced by paradoxical sleep deprivation (PSD), dapagliflozin is found to alleviate those behaviors accompanied by enhanced autophagic clearance in the hippocampus and promoted GABAergic transmission.⁴⁴

The studies investigating the effects of SGLT-2i on mood disorders are listed in Table 1. And the verified molecular pathways of SGLT-2i in mood disorders are summarized in Figure 1.

Table 1 Summary of Evidence for SGLT-2i in Animal Models Related to Mood Disorders

Figure 1 Proposed neurobiological mechanisms underlying the potential benefits of SGLT-2 inhibitors in mood disorder animal models. SGLT-2 inhibitors modulate multiple pathways implicated in mood disorders. These converging mechanisms help restore cellular homeostasis and ultimately improve mood-related behaviors, as validated in preclinical models. Upward arrows indicate relative increases or enhancement, whereas downward arrows indicate relative decreases or inhibition in the indicated molecules, pathways, or biological effects. Abbreviations: SGLT-2, Sodium-glucose cotransporter-2; NF-κB, Nuclear factor kappa-light-chain-enhancer of activated B cells; NLRP3, NOD-like receptor thermal protein domain associated protein 3; IL-1β/IL-18, Interleukin-1 beta/Interleukin-18; TNF-α, Tumor necrosis factor-alpha; BBB, Blood–brain barrier; ROS, Reactive oxygen species; iNOS, Inducible nitric oxide synthase; GSH, Glutathione; CAT, Catalase; AMPK, AMP-activated protein kinase; mTOR, Mammalian target of rapamycin; LC3-I/LC3-II, Microtubule-associated protein 1A/1B-light chain 3 isoforms I and II; BDNF, Brain-derived neurotrophic factor; CREB, cAMP response element-binding protein; IDO, Indoleamine 2,3-dioxygenase; GABAA/GABAB R, Gamma-aminobutyric acid A/B receptor; HPA axis, Hypothalamic–pituitary–adrenal axis; CORT, Cortisol; TJ, Tight junction.

Cognitive impairment is a key feature of mood disorders and is often exacerbated by comorbid metabolic conditions. SGLT-2i has been shown to be useful in cognition improvement manifested in diabetes-associated cognitive impairment. Empagliflozin reduces cortical thinning, amyloid plaque burden, microglial activation, and oxidative stress, improving memory performance possibly through modulating neuroinflammation and neurotrophic factors.^45–47 But these findings are indirect with respect to mood disorders and should be interpreted as supportive mechanistic evidence rather than disorder-specific therapeutic data.

Clinical Evidence of SGLT-2i in Mood Disorders

While the growing animal studies have demonstrated the therapeutic potential and mechanisms of SGLT-2i for mood disorders, it’s important to note that the doses used in these experiments are typically much higher than clinical doses, and the metabolic profiles of animal models differ significantly from those of human patients. Therefore, additional clinical data are required before SGLT-2i can be formally adopted for mood disorder treatment.

Since SGLT-2i are primarily used as anti-diabetic medications, most related clinical studies were conducted in diabetic populations. A retrospective study from Taiwan, China, showed that patients with T2DM treated with SGLT-2i have a significantly reduced risk of developing psychiatric disorders (HR=0.80, 95% CI: 0.72–0.88).⁴⁸ Focusing on mood disorders, a retrospective cohort study published in 2019 has already found that among multiple anti-diabetic drugs, SGLT-2i was associated with a significantly lower risk of developing depression in diabetic patients (OR=0.09, 95% CI: 0.01–0.63) but was limited by the small number of T2DM patients with comorbid depression receiving SGLT-2i.⁴⁹ Subsequent clinical studies with larger sample sizes provide more reliable results. A nested case–control study finds that several anti-diabetic agents are related to reduced risk of depression onset in diabetic patients, among which SGLT-2i exhibits the most significant association (OR=0.55, 95% CI: 0.44–0.70).¹⁸ Two other large cohort studies report similar findings.^19,20

Another case–control study further investigates the broader neuropsychiatric effects of SGLT-2i in diabetic patients. Diabetic patients have significantly higher PHQ-9 scores than healthy controls, but among the diabetic cohort, it’s significantly lowered through the treatment of SGLT-2i compared to sulfonylureas.⁵⁰ Similar favoring effects are also observed in inflammatory biomarker levels, cognitive function tests, and quality of life (QoL) scores.⁵⁰

There is also evidence supporting the role of SGLT-2i for clinical symptoms of mood disorders, including cognitive defects. A recent meta-analysis assesses the relationship between anti-diabetic drugs and cognitive impairment in T2DM patients and reveals that, those taking SGLT-2i have a significantly reduced risk of dementia (OR=0.75, 95% CI: 0.64–0.87), which is the most robust among all anti-diabetic drugs studied.⁵¹

While these population-based studies provide encouraging evidence for the beneficial psychiatric effects of SGLT-2i, they primarily focus on comorbidity risk in diabetic populations, which are not designed to evaluate their therapeutic efficacy in individuals already diagnosed with mood disorders. Two recent studies have attempted to amend this gap. A randomized controlled trial in MDD patients with baseline HDRS scores ≥17 shows that those receiving citalopram plus 10 mg empagliflozin have a greater reduction in HDRS scores than the placebo group, suggesting a potential therapeutic effect of empagliflozin in MDD.⁵² Another cohort study focuses on patients diagnosed with BD and suggests that SGLT-2i may have a neutral to potentially protective effect on suicide-related outcome compared to DPP-4 (IR = 0.17 vs. 0.26 per 100 person-years; IRD = −0.09; 95% CI: −0.16–0.02).⁵³

However, not all studies report consistent findings. A cross-sectional study finds no statistically significant difference in depressive symptom severity between diabetic patients using and not using SGLT-2i.⁵⁴ This study underscores the need for further research to clarify the effects of SGLT-2i on mental health.

The clinical studies concerning SGLT-2i and mood disorders are summarized in Table 2.

Table 2 Summary of the Clinical Studies Evaluating the Effects of SGLT-2 Inhibitors on Mood Disorders

Mechanisms Linking Mood and Metabolic Disorders

SGLT-2i exert potential therapeutic effects on mood disorders through multiple mechanisms. To better understand their multi-target effects, we first explore the various pathways involved in the comorbidity between metabolic disorders and mood disorders. Several hypotheses linking these two disorders have been proposed, like genetic and environmental factors, psychotropic medications, and moreover, the interconnected framework including inflammatory processes, hypothalamic–pituitary–adrenal (HPA) axis, oxidative stress, neuroplasticity, and gut microbiota.⁵⁵

Genetic Factors

The causal links between depression and metabolic traits have been verified by Mendelian randomization studies: fat mass is a causal risk factor for depression,⁵⁶ while depression demonstrates a significant causal effect on T2DM and 36.5% of the effect was mediated by BMI.⁵⁷ Specifically, Amare et al identify 24 candidate pleiotropic genes and important biological pathways that are likely to be shared with mood and metabolic disorders through meta-GWAS analysis.⁵⁸

Environmental Influences

Unhealthy lifestyles such as smoking, excessive alcohol consumption, poor sleep hygiene, physical inactivity, and unhealthy nutritional patterns are highly prevalent among MDD/BD patients, and the limited access to quality healthcare services further exacerbates this risk of pathophysiological disturbances and developing MetS.^59–61 Considering the prevalence of unhealthy behaviors among patients with psychiatric disorders and its substantial influence on MetS, it is likely that adverse habits contribute remarkably to this comorbidity.⁵⁵ Intriguingly, it is suggested that novel environmental inputs, including artificial light at night and sustained refined sugar intake, may aggravate the metabolic and circadian dysregulation, driving the BD development.⁶²

Effects of Psychotropic Medications

Psychotropic medications, particularly antipsychotics, further elevate the risk of metabolic disturbances like obesity, dyslipidemia, and diabetes in these patients.⁶³ Individuals treated with any single antipsychotic agent, especially olanzapine, commonly used in mood disorders treatment, have a higher risk of developing MetS compared to those who received none.⁴ The prevalence of T2DM significantly increased in those treated with antipsychotics except for aripiprazole and amisulpride compared to non-users.⁶⁴ Most psychotropic medications also contribute to weight gain through mechanisms involving increased appetite, higher caloric intake, and delayed satiety signaling.⁶⁵ The effect of mood stabilizers on weight gain in BD patients is also significant, with a reported 77% patients experiencing an average weight gain of 4–6.3 kg.^65,66 Anti-depressants, particularly those with high affinity for norepinephrine re-uptake transporters, serotonergic 5-HT2C receptors, and histamine H1 receptors, also contribute to the increased risk of new onset diabetes, especially in high doses and long-term treatment.⁶⁵

However, it is important to note that even after adjusting for lifestyle covariates, the association between metabolic disorders and psychiatric illnesses persists,⁶⁷ and is also observed in patients receiving diverse treatment regimens and extends even to those who remained untreated.⁶⁸ These findings suggest that additional pathophysiological mechanisms inherent to psychiatric disorders themselves contribute to the development of metabolic disorders.

Inflammation and Oxidative Stress

CNS inflammation would affect synaptic plasticity, neurogenesis, and emotional behavior.⁶⁹ It is well recognized that a high-fat diet (HFD) related to inflammation of the hypothalamus as well as mitochondrial dysfunction and insulin resistance in the hippocampus, induces cognitive or affective dysfunction.^70–72 Obesity is a pro-inflammatory state, with adipose tissue functioning as an endocrine organ capable of secreting pro-inflammatory cytokines and hormones such as leptin.⁵⁵ An immune cell will further infiltrate into metabolic tissues, promoting insulin resistance and exacerbating systemic glucose metabolism dysregulation, thereby propagating the vicious cycle of inflammation.⁷³ In conclusion, inflammation associated with mood disorders predisposes individuals to metabolic dysregulation; metabolic dysfunction perpetuates inflammation, establishing a vicious cycle that reinforcing both metabolic and psychiatric pathology.

Disturbance of the HPA Axis

The HPA axis plays a key role in the body’s adaptation to environmental stress. Its hyperactivation is a hallmark of MDD, contributing to emotional and cognitive dysfunction and stress-related abnormalities through its effects on neurotransmitters, neuroplasticity, and inflammatory responses, especially in the amygdala and hippocampus.⁷⁴ The resulting chronic elevation of cortisol promotes visceral fat accumulation and insulin resistance through the activation of lipoprotein lipase and inhibition of lipid mobilization, thereby increasing the risk of MetS, obesity, and diabetes.⁵⁵

Neuroplasticity and Neurotrophic Factors

Neuroplasticity helps encode experiences and adapt to stimuli. Its mechanisms, mediated by regulators like BDNF and mammalian target of rapamycin (mTOR), encompass neurogenesis, programmed cell death, and activity-dependent synaptic plasticity.⁷⁵ The development of mood disorders is closely related to neuroplasticity, and some anti-depressants could regulate it to take their effect.⁷⁶ Neuroplasticity is also affected in obesity. The magnitudes of long-term potentiation and long-term depression were lower at the hippocampus of obese mice.⁷⁷ Application of mTOR inhibitor rapamycin in HFD-induced metabolic dysregulated and depressed mice could ameliorate both disorders,⁷⁸ implicating the role of neuroplasticity between mood and metabolic disorders.

Gut Microbiota

The gut microbiota has been proven to be related to MDD and BD, through the bidirectional gut-brain axis based on metabolic, endocrine, neural, and immunological pathways.⁷⁹ Meanwhile, gut microbiota and its metabolites may act directly or indirectly on the brain via vagal stimulation and food reward signals to regulate metabolism, body homeostasis, and energy balance, thus contributing to obesity.⁸⁰

Summary of Comorbid Mechanisms

To sum up, the interaction between mood disorders and metabolic disorders is bidirectional and involves multiple interrelated factors, as shown in Figure 2, suggesting that medications targeting metabolic dysfunction may also help mood disorders treatment. Moreover, considering the adverse effects of antipsychotic medications, combination therapy to alleviate patients’ metabolic burden could be beneficial. Against this background, SGLT-2i, whose effects target those overlapping pathways, has attracted growing attention for its potential neuropsychiatric benefits, which will be discussed in the following section.

Figure 2 Mechanisms contributing to the bidirectional influence of mood disorders and metabolic disorders. Multiple interacting pathways contribute to the close relationship between mood and metabolic disturbances. Together, these mechanisms form a vicious cycle linking emotional and cognitive dysfunction with metabolic dysregulation. Upward arrows indicate relative increases or enhancement, whereas downward arrows indicate relative decreases or inhibition in the indicated molecules, pathways, or biological effects. Abbreviations: HPA axis, Hypothalamic–pituitary–adrenal axis; CRF, Corticotropin-releasing factor; AMPK, AMP-activated protein kinase; mTOR, Mammalian target of rapamycin; BDNF, Brain-derived neurotrophic factor; TrkB, Tropomyosin receptor kinase B; PI3K/AKT, Phosphoinositide 3-kinase / protein kinase B; LTP / LTD, Long-term potentiation/Long-term depression; mPFC, Medial prefrontal cortex; NLRP3, NOD-like receptor thermal protein domain associated protein 3; MTHFR, Methylenetetrahydrofolate reductase; GNAS, Guanine nucleotide-binding protein G(s) subunit alpha; FTO, Fat mass and obesity-associated protein; POMC, Pro-opiomelanocortin; GSK3B, Glycogen synthase kinase 3 beta; TLR4, Toll-like receptor 4; APOE, Apolipoprotein E; CRY2, Cryptochrome circadian regulator 2; CACNA1D, Calcium voltage-gated channel subunit alpha1 D; IGF1, Insulin-like growth factor 1.

Mechanical Basis for Neuropsychiatric Benefits of SGLT-2i

Given the various overlapping mechanisms between mood and metabolic disorders, SGLT-2i may confer neuropsychiatric benefits through different biological pathways. This section outlines the major mechanistic domains through which SGLT-2i may modulate mood and cognitive function.

Metabolism

As an anti-diabetic medication, SGLT-2i effectively lower blood glucose levels and reduce insulin resistance, while concurrently improving lipid metabolism and promoting weight loss.⁸¹ These metabolic benefits alleviate peripheral metabolic burden, potentially mitigating mood disorder through previously described mechanisms shared between metabolic and mood disorders. However, except for such indirect metabolic effects, SGLT-2i directly regulates various metabolic pathways like ketogenesis, mTOR and central insulin signaling pathways also benefits multiple neuropathological processes.

Ketogenesis

SGLT-2i induces mild ketosis via a mechanism often described as an “accelerated starvation state”.⁸² During prolonged fasting when glycogen stores are depleted, the liver converts fatty acids into ketone bodies to serve as an alternative fuel.⁸² SGLT-2i induce a similar metabolic state by promoting renal glucose excretion in the absence of fasting, leading to metabolic adaptations including enhanced lipolysis and ketogenesis, mediated through AMPK activation and mTOR inhibition.^82,83

Ketone bodies, especially β-hydroxybutyrate (BHB), are a more efficient energy substrates that improve tissue energy metabolism.⁸⁴ Given the brain’s high energy demand, it may particularly benefit from such metabolic effects. Ketos also manifested neuroprotective, antioxidant, and neurotransmitter-regulatory effects in the brain.⁸⁵ Specifically, BHB could function as a histone deacetylase inhibitor, leading to transcriptional activation of genes responsible for oxidative stress defense and metabolic homeostasis.⁸⁶ It also modulates the activation of NLRP3 inflammasome and other inflammatory pathways, thereby suppressing inflammatory responses.⁸⁷ Ketogenic diet has been associated with changes in monoamine neurotransmitters, including dopamine, noradrenaline, and serotonin in epilepsy, and also inhibits glutamate decarboxylase, thus stimulating GABA synthesis.⁸⁵ Clinically, multiple randomized trials have already confirmed that the ketogenic diet does improve depressive mood, but its poor adherence and limited serum ketone maintenance pose challenges that SGLT-2i may help overcome pharmacologically, making them promising agents in mood disorder treatment.⁸⁸

Restoring the Rhythm of mTOR Activity

mTOR is another target of SGLT-2i concerning metabolism,⁸³ which integrates extracellular and intracellular signals to regulate cell homeostasis and metabolism,⁸⁹ thus acting as a potential link between metabolic diseases and cognitive impairment in psychiatric disorders.²⁷ Under conditions where sufficient insulin and amino acids are available, mTORC1 complex is activated and suppresses autophagy, whereas nutrient deprivation inhibits mTORC1, restoring autophagic and lysosomal activity to release amino acids into circulation.⁹⁰ However, in patients with metabolic diseases exhibiting insulin resistance, chronically elevated circulating amino acids sustain mTORC1 activation and impair autophagy.⁹¹ By promoting continuous glucose excretion, SGLT-2i helps re-establish normal circadian dynamics of mTOR activity—promoting anabolic activity during the day and catabolic activity at night⁹¹—thereby maintaining metabolic balance, mitigating inflammation, and ultimately contributing to improvements in cognitive function and mood regulation.

In addition to its role in metabolic regulation, mTOR regulates multiple neural processes, including autophagy-dependent neuronal survival, axonal regeneration and sprouting, neurogenesis and differentiation, myelination and oligodendrocyte maturation, as well as synaptic activity regulation downstream of NMDA receptor.⁹² Several of these mechanisms are also implicated in the modulatory effects of SGLT-2i in neuropsychiatric disorders.

Ameliorating Brain Insulin Resistance

The brain is an insulin-sensitive tissue and is also exposed to the deleterious effects of insulin resistance. Brain insulin resistance disrupts neuronal energy metabolism, aggravates oxidative stress, and promotes neuroinflammation.⁹³ The resulted abnormal neuronal activities are considered a critical factor underlying AD, MDD/BD, and the accompanying cognitive deterioration.^94–97

SGLT-2i have been shown to enhance peripheral insulin sensitivity via enhanced fat utilization and browning, M2 macrophage polarization,⁹⁸ and mTOR pathway modulation.⁹⁵ Similarly in CNS, SGLT-2i enhances brain insulin signaling and improves hippocampal-dependent learning, memory, and cognitive functions in T2DM-AD mouse.⁹⁹ Dapagliflozin increases the expression of insulin receptor and improves hippocampal insulin signaling, rescuing the HFD-induced cognitive impairment.^72,100 As for humans, empagliflozin ameliorates insulin resistance in the hypothalamus, mediating the decrease in fasting glucose and liver fat.¹⁰¹ However, another study in HFD-fed female mice found that such improvement is only peripheral.¹⁰² More studies are needed to further validate the influence of SGLT-2i on central insulin signaling.

Inflammation

Elevated peripheral inflammatory markers in mood disorder patients may disrupt the blood–brain barrier, promoting neuroinflammation and microglial activation, as observed in both MDD and BD patients.^103–105 These changes, observed in brain regions involved in reward and cognition, are associated with neuronal network disruption, synaptic dysfunction, and altered neurotransmission, including the kynurenine pathway, glutamate, and dopaminergic signaling.^105,106

Firstly, SGLT-2i mitigates inflammation through the regulation of metabolic dysfunction, as shown in conditions such as myocardial infarction and renal diseases.¹⁰⁷ SGLT-2i also induce the polarization of macrophages toward M2 phenotype through multiple pathways,^107,108 and suppress NLRP3 inflamamsome.^109,110 Systemic inflammatory reduction is the result of these anti-inflammatory responses, supported by preclinical studies and population-based RCTs, which demonstrate that SGLT-2i significantly reduces multiple inflammatory as well as metabolic markers.^111–114 Alleviation of peripheral inflammation mitigates BBB disruption, thus indirectly slowing the progression of neuronal damage.

Moreover, SGLT-2i could directly alleviate neuroinflammation. The regulation of microglia by SGLT-2i may represent a critical mechanism underlying their therapeutic effects in mood disorders. In vitro studies have shown that empagliflozin reduced the expression of inflammatory mediators in LPS-activated rat primary microglia, which might be mediated by NHE-1 and by the inhibition of the ERK1/2 and NFκB pathways.¹¹⁵ In vivo, SGLT-2i are reported to reduce microglial activation and promote a shift in microglial to the anti-inflammatory M2 phenotype in the LPS-induced rat neuroinflammation model, as well as inhibit NLRP3 inflammasome expression by disrupting the pTau-CX3C1 interaction.^33,116 Inhibition of NLRP3 inflammasome activation and its downstream signaling also serves as a key therapeutic target of CNS-related pathologies,¹¹⁷ through which SGLT-2i exert their neuroprotective effects, as shown in CUS-induced depressed rats,³² as well as other various animal models, including Aluminium Chloride-induced AD rats,¹¹⁸ rotenone-intoxicated rats with Parkinson disease (PD)¹¹⁹ and methotrexate-induced cognitive impairment rats.¹²⁰

Mitochondrial Function and Cell Homeostasis

Mitochondria act as the hub of energy metabolism and autophagy, essential for maintaining neuronal homeostasis. Mitochondrial dysfunction has emerged as a key etiological factor in mood disorders, manifesting through: (i) structural abnormalities in morphology and distribution, (ii) impaired energy metabolism process with reduced ATP production and enhanced anaerobic glycolysis, and (iii) compromised quality control due to dysregulated biogenesis and mitophagy clearance pathways.^121,122 Oxidative stress and autophagy, both closely linked to mitochondrial function, contribute greatly to the pathogenesis of mood disorders and metabolic disorders. The ability of SGLT-2i to modulate these processes may contribute to its therapeutic effects for mood disorders.

Anti-Oxidative Properties of SGLT-2i

The brain is highly vulnerable to oxidative stress due to its high metabolic rate, lipid content, and limited antioxidant defenses.¹²¹ Excessive ROS disrupts cellular communication, induces lipid and protein peroxidation, and triggers neuronal apoptosis via Ca²⁺ overload.¹²³ Elevated oxidative stress is consistently observed in mood disorders, correlates with disease severity, and some treating medications like fluoxetine, citalopram, lithium, and valproate may exert benefits by modulating redox balance.^124–127

SGLT-2i reduces oxidative stress through multiple mechanisms: direct improvement of mitochondrial function, downregulation of oxidative enzymes, and enhancement of antioxidant systems.¹²⁸ Indirect effects include improving glycemic control, alleviating insulin resistance, modifying hemodynamics, and ameliorating renin-angiotensin system (RAS) activation.¹²⁸ By suppressing inflammatory pathways, SGLT-2i further reduce oxidative stress.¹²⁹ Additionally, ketone bodies promoted by SGLT-2i can serve as antioxidants and support mitochondrial function.^82,88

Many preclinical studies have confirmed such antioxidative and neuroprotective effects, including CUS-induced depressive rats in which empagliflozin reversed the increased oxidative markers.³³ In vitro studies reveal that empagliflozin mitigates dendritic and spine alterations induced by hydrogen peroxide in primary neurons and inhibits neuronal ferroptosis induced by oxygen–glucose deprivation/reoxygenation.^130,131 In rotenone-induced PD rat models, both empagliflozin and dapagliflozin are able to reverse the elevated oxidative stress through AMPK/SIRT-1/PGC-1α pathway or DJ-1/Nrf2 pathway.^132,133

SGLT-2i Modulates Autophagy

Dysregulated autophagy is observed in both mood disorders and neurodegenerative diseases characterized by abnormal protein aggregation, like AD.^134,135 Recent studies demonstrate that acute stress activates, while chronic stress suppresses autophagy in the lateral habenula (LHb), and the latter is associated with maladaptation to chronic stress.¹³⁶ Several anti-depressants and mood stabilizers have been shown to modulate autophagy, including fluoxetine, sertraline, paroxetine, desvenlafaxine, lithium, and valproate.¹³⁷

But it’s noteworthy that both excessive and insufficient autophagy have been implicated in neurological disorders, with some depression-like animal models exhibiting increased autophagic markers.¹³⁷ AKT1/mTOR has been found to be reduced in BD patients during depressive episodes,¹³⁸ and is also compromised in the prefrontal cortex of MDD patients during postmortem studies.¹³⁹ But the anti-depressant effect of mTOR inhibitor rapamycin has also been reported in animal models.¹⁴⁰

SGLT-2i can modulate autophagy through both mTOR-dependent and independent mechanisms, as verified in heart and kidneys,⁸⁹ and also animal models of mood disorders. Empagliflozin reverses the depressive behaviors and hinders autophagy in the hippocampus of depressed rats induced by reserpine through AMPK/mTOR/Beclin1/LC3B machinery.³⁷ Canagliflozin also modulates TYR/KYN and AMPK/mTOR autophagic signaling in rats subjected to chronic unpredictable mild stress.³⁴ In manic mouse models, dapagliflozin influences autophagic flux and ameliorates mania-like behaviour.³⁸ Similar regulation of the autophagy pathway is also found in AD rat models,¹⁴¹ rotenone-lesioned rats,¹¹⁹ vigabatrin-induced neurotoxicity,¹⁴² and PD rats.¹⁴³ But the exact effects of SGLT-2i on autophagy and thus mood disorders need further clarification.

Brain-Derived Neurotrophic Factor

Neurotrophic factors, particularly BDNF, are important targets of SGLT-2i. BDNF plays a critical role in neuronal survival and growth, acting as a mitogenic and chemotactic factor that promotes the proliferation, migration, and maturation of neural progenitor cells, ultimately contributing to synaptic transmission and plasticity, especially in hippocampus.¹⁴⁴ BDNF also exerts neuroprotective effects through the suppression of apoptosis, oxidation, and autophagy.¹⁴⁵

Decreased serum levels of mature BDNF (m-BDNF) are consistently observed in MDD patients, while pro-BDNF remains unchanged.¹⁴⁶ In BD patients, the reduction is more pronounced, and the m-BDNF/pro-BDNF ratio differs significantly from MDD, potentially serving as a discriminatory biomarker between the two disorders.¹⁴⁷ BDNF mediates the effects of many anti-depressants like ketamine, which triggers BDNF release via AMPA receptor activation, enhances BDNF synthesis by inhibiting eEF2K, and promotes BDNF-TrkB binding, ultimately activating mTOR and stimulating synaptogenesis.⁷⁶

BDNF contributes greatly to the interaction between SGLT-2i and psychiatric disorders, and acts synergically with other mechanisms listed above. The verified anti-depressant mechanisms of SGLT-2i in various studies involve BDNF/CREB signaling, highlighting its significance in mood disorder treatment.^{31,32,35–37} Besides, SGLT-2i have also manifested their benefits in AD cognitive impairment by increasing BDNF,^148–150 with similar effects observed in other neuropsychiatric conditions like PD and epilepsies.^132,151,152

Additional Effects

Beyond mood disorders, other pathways have also been implicated in the therapeutic effects of SGLT-2i in a range of neurological and neurodegenerative conditions, including AD, PD, HD, and metabolic-related cognitive impairment. Although some mechanisms have not yet been directly validated in mood disorders, they provide important insights for understanding the potential intersections between SGLT-2i and mood pathology.

Cholinergic Modulation

SGLT-2i has also been found to possess potential acetylcholinesterase (AChE) inhibitory property in AD.¹⁵³ Animal studies have shown that SGLT-2i can improve memory deficits in scopolamine-induced amnesia rat models by dampening AChE activity and increasing monoamine neurotransmitter levels.¹⁵⁴

Cerebrovascular Protection

The neurovascular unit (NVU), composed of neurons, glial cells, vascular cells, the extracellular matrix, and the functional coupling of these diverse components, is crucial for maintaining CNS homeostasis.¹⁵⁵ Damage to both micro- and macro-vascular structures can lead to NVU dysfunction and neurovascular remodeling, which play critical parts in the onset and progression of neurodegenerative diseases and cognitive impairment.¹⁵⁶ Empagliflozin has been shown to ameliorate microvascular damage and NVU remodeling caused by diabetes¹⁵⁷ and also enhance endothelial cell function through various mechanisms, thus hindering the progression of atherosclerotic disease.¹⁵⁸

Summary

Evidence suggests that the neuroprotective effects of SGLT-2i arise from a convergence of multiple mechanisms summarized in Figure 3. SGLT-2i attenuates systemic inflammation and oxidative stress, thereby reducing harmful signaling to the central nervous system. And by improving metabolic status through ketogenesis, regulation of mTOR activity, and restoration of insulin sensitivity, these agents benefit the CNS. Within the brain, SGLT-2i enhances mitochondrial bioenergetics, restores autophagy balance, and suppresses neuroinflammation through inhibition of microglial activation and inflammasome pathways. In addition, SGLT-2i upregulates BDNF, promotes synaptic plasticity, supporting neuronal resilience. Collectively, these mechanisms interact under the treatment of SGLT-2i, providing a plausible biological basis for the observed improvements in affective and cognitive symptoms. This multifaceted mode of action highlights the potential of SGLT-2i as a bridge between metabolic regulation and neuropsychiatric benefits.

Figure 3 Proposed mechanisms of SGLT-2 inhibitors in mood disorders. SGLT-2 inhibitors exert benefits to psychological disorders through multifaceted mechanisms, both directly in the brain, and indirectly peripherally. Collectively, SGLT-2 inhibitors contribute to mood and cognitive improvements. Upward arrows indicate relative increases or enhancement, whereas downward arrows indicate relative decreases or inhibition in the indicated molecules, pathways, or biological effects. Abbreviations: SGLT-2, Sodium-glucose cotransporter-2; HPA axis, Hypothalamic–pituitary–adrenal axis; NLRP3, NOD-like receptor thermal protein domain associated protein 3; BDNF, Brain-derived neurotrophic factor.

Overall, the mechanistic rationale linking SGLT-2 inhibitors to mood disorders remains biologically plausible but it should also be interpreted with caution. Much of the available evidence is derived from animal studies, in vitro experiments, or disease models not specific to mood disorders. So accordingly, the pathways discussed above should be regarded primarily as proposed or potentially relevant mechanisms rather than established therapeutic mechanisms in mood disorders. Further translational studies are needed to clarify which of these effects are reproducible in human populations and whether they are sufficient to translate into meaningful psychiatric benefit.

Potential and Challenges of SGLT-2i in Mood Disorders

Initially developed for T2DM, SGLT-2 inhibitors are gaining more recognition for their pleiotropic effects that may benefit mood regulation. Preliminary data from animal and population studies suggest a multifaceted framework of its biological mechanism, though translation to clinical use in mood disorders is still constrained by insufficient evidence and several practical challenges.

Safety Considerations of SGLT-2i in Mood Disorder Treatment

The application of SGLT-2i in patients with mood disorders calls for the consideration of potential side effects and complex interactions with existing psychotropic medications. Known adverse effects include common genitourinary infections, volume depletion-related events including orthostatic hypotension and hypovolemia, acute kidney injury, bone fractures, euglycemic or hyperglycemic diabetic ketoacidosis, elevated low-density lipoprotein and cholesterol, lower-limb amputation, and even malignancy.^20,159,160 Several clinical cases have also presented the potential negative outcomes when SGLT-2i are combined with commonly used antipsychotics or mood stabilizers.

For instance, empagliflozin could be associated with valproate toxicity when used in combination, likely due to changes in drug distribution related to weight loss.¹⁶¹ Co-administration of lithium and empagliflozin may result in serum lithium concentrations, possibly through increased renal excretion by inhibiting proximal tubular reabsorption of lithium-glucose co-transport.¹⁶² These cases highlight the need for careful consideration of potential direct or indirect interactions between SGLT-2i and psychotropic medications when applied to mood disorders. Dose adjustments and close monitoring of serum drug concentration are required to ensure safety and efficacy in such combined treatment regimens.

Direction of Future Research

Despite the growing interest in SGLT-2i as potential therapeutic agents for mood disorders, current evidence remains preliminary and fragmented, warranting further validation across multiple research levels.

At the basic research level, although studies have indicated low-level expression of SGLT-2 in the central nervous system,^24–26 its functional activity and regulatory pathways in the brain remain poorly characterized. Meanwhile, current investigations on SGLT-2i in mood disorders primarily employ CUS-induced depression-like models, with scarce research involving manic or phase-switching models.

Clinically, existing evidence mainly derives from secondary analyses or retrospective observational studies in T2DM patients, limiting causal inference. The absence of controlled trials in normoglycemic populations makes it difficult to distinguish mood improvements resulting from direct effects of SGLT-2i and those driven by glycemic normalization. Importantly, while some findings support a reduced risk of depression or cognitive impairment onset, evidence for the reversal of established symptoms remains insufficient. To date, only a small number of clinical trials have investigated the long-term effects of SGLT-2i use, so questions remain regarding the durability of therapeutic efficacy,¹⁶³ particularly in psychiatric populations where long-term safety and interactions with psychotropic medications are important. Given the marked clinical heterogeneity observed in patients with mood disorders, individual responses to SGLT-2i are likely to vary. Stratified studies are essential to delineate patient subgroups most likely to benefit from this type of medication.

More trials are needed to validate this hypothesis, with two already underway (NCT05757791, NCT05792540). In summary, while current findings are limited in scope and quality, SGLT-2 inhibitors represent a promising class of metabolic modulators with potential pleiotropic neuropsychiatric benefits. Future mechanistic studies integrating neuroimaging, multi-omics, and immune-metabolic profiling are needed to bridge translational gaps and identifying predictive biomarkers for precise therapy. Future clinical research should prioritize well-characterized patient subgroups, especially individuals with mood disorders and coexisting metabolic dysfunction, to identify the patient populations and biological pathways most relevant to future clinical application. In addition to broad diagnostic categories, studies should also consider stratification by illness stage, symptom profile, cognitive impairment, and exposure to metabolically adverse psychotropic medications, as these factors may influence both treatment response and tolerability. Moreover, well-designed RCTs are needed to clarify whether SGLT-2i improves core mood symptoms, cognition, and functioning, and whether such effects are independent of improvements in metabolic status. At the same time, longer-term studies should evaluate safety and feasibility in psychiatric populations, including side effects and potential interactions with commonly used psychotropic agents.