Mechanisms Underlying Treatment‐Resistant Depression: Exploring Sex‐Based Biological Differences

ABSTRACT

Treatment‐resistant depression (TRD) represents a severe and complex subtype of major depressive disorder (MDD), affecting approximately 30% of patients who fail to respond adequately to multiple standard antidepressant therapies. While the pathophysiology of TRD remains incompletely understood, emerging evidence suggests that sex‐based biological differences might influence its onset, progression, and treatment response. Women are disproportionately affected by depression and are more likely to experience residual symptoms and treatment resistance, potentially due to hormonal fluctuations, immune system differences, and variations in brain circuitry and neuroplasticity. This narrative explores the current literature on the mechanisms underlying TRD, with a particular emphasis on sex‐specific biological factors. Key focus areas include dysregulation in neurotransmitters and neurotrophic pathways, inflammation, HPA axis alterations, mitochondrial dysfunction, as well as the influence of sex hormones such as estrogen and progesterone. By highlighting these differences, this review underscores the importance of personalized, sex‐informed approaches in the prevention and treatment of TRD and calls for further research to elucidate the biological underpinnings that contribute to sex disparities in treatment outcomes.

Treatment‐resistant depression (TRD) affects about 30% of people with major depressive disorder and presents significant clinical and biological challenges. This review consolidates current evidence on the complex causes of TRD, focusing on biological differences based on sex. We examine disruptions in monoaminergic signaling, neurotrophic support, immune activation, HPA axis function, and mitochondrial metabolism, and how these may be differently influenced by sex hormones like estrogen and progesterone. Special attention is given to structural and functional brain changes and new molecular pathways that could explain sex differences in treatment response. This overview highlights the importance of personalized medicine that considers sex as a key biological factor in preventing and treating TRD.

Abbreviations

17β‐E2

17β‐estradiol

3‐HAA

3‐hydroxyanthranilic acid

3‐HK

3‐hydroxykynurenine

5‐HT

5‐hydroxytryptamine (serotonin)

ACC

anterior cingulate cortex

ACTH

adrenocorticotropic hormone

ALFF

amplitude of low‐frequency fluctuations

BBB

blood–brain barrier

BDNF

brain‐derived neurotrophic factor

CNS

central nervous system

CRH

corticotropin‐releasing hormone

CRP

C‐reactive protein

DA

dopamine

DBS

deep brain stimulation

DMN

default mode network

DRD2

dopamine receptor D2

DTI

diffusion tensor imaging

ECT

electroconvulsive therapy

EREs

estrogen response elements

ERα

estrogen receptor alpha

ERβ

estrogen receptor beta

FA

fractional anisotropy

FGF21

fibroblast growth factor 21

fMRI

functional magnetic resonance imaging

GABA

gamma‐aminobutyric acid

GDF15

growth differentiation factor 15

GWAS

genome‐wide association study

HPA

hypothalamic–pituitary–adrenal

IL

interleukin

KYN

kynurenine

KYNA

kynurenic acid

MDD

major depressive disorder

mPTP

mitochondrial permeability transition pore

MRI

magnetic resonance imaging

MRS

magnetic resonance spectroscopy

NDRIs

norepinephrine‐dopamine reuptake inhibitors

NE

norepinephrine

NET

norepinephrine transporter

OFC

orbitofrontal cortex

PTSD

post‐traumatic stress disorder

ReHo

regional homogeneity

rs‐fMRI

resting‐state functional magnetic resonance imaging

rTMS

repetitive transcranial magnetic stimulation

SERT

serotonin transporter

SOD2

superoxide dismutase 2

SSRIs

selective serotonin reuptake inhibitors

TCAs

tricyclic antidepressants

TNF‐α

tumor necrosis factor‐alpha

TRD

treatment‐resistant depression

TRP

tryptophan

vmPFC

ventromedial prefrontal cortex

VNS

vagus nerve stimulation

XCI

X chromosome inactivation

XIST

X‐inactive specific transcript

1. Introduction

Major depressive disorder (MDD) is a leading cause of global disability and currently impacts more than 300 million people globally (Collaborators 2022; World Health Organization 2017). With genetic factors contributing up to 35% of the risk, it has a moderate degree of heritability, although it is also thought to be strongly influenced by adverse life events (Geschwind and Flint 2015; Otte et al. 2016). Despite various effective treatments for depression, such as antidepressants, psychotherapies, and brain stimulation, up to 30% of patients fail to achieve complete remission, a phenomenon that has been clinically defined as treatment‐resistant depression (TRD) (McIntyre et al. 2023). Compared to non‐resistant MDD, TRD is associated with greater disease severity, increased risk of chronicity, higher healthcare utilization, and more profound psychosocial impairment (Baig‐Ward et al. 2023). Moreover, individuals with TRD often endure heightened suffering, as evidenced by more severe manifestations of anxiety, depression, and cognitive impairment (Sun, Ma, et al. 2022). A growing number of interventions beyond traditional pharmacotherapy are being used in the management of TRD. FDA‐approved options include esketamine, electroconvulsive therapy (ECT), repetitive Transcranial Magnetic Stimulation (rTMS), and Vagus Nerve Stimulation (VNS), while DBS and psychedelic‐based therapies are under investigation in clinical trials.

Recent research has further emphasized the biological and genetic distinction between MDD and TRD, suggesting that TRD may constitute a unique clinical and neurobiological entity (Fabbri et al. 2021; Johnston et al. 2015; Machino et al. 2014). Moreover, sex consistently emerges as a robust effect modifier of treatment response, but whether the net effect is risk‐increasing or protective towards one or the other is modified by study design, with most literature attributing risk to female sex (Gronemann et al. 2018; Huang et al. 2020; Lähteenvuo et al. 2022; O'Connor et al. 2023). Genetic research further underscores this distinction, identifying distinct sex‐specific polymorphic genetic profiles, suggesting significant differences in epigenetic regulation, cell cycle pathways, inflammatory responses, and pharmacological targets (Kang et al. 2020; Silveira et al. 2023). These findings are supported by prior studies of antidepressant efficacy, which have demonstrated sex‐dependent treatment responses (LeGates et al. 2019). Despite this growing body of evidence, sex remains an underexamined variable in TRD research, with relatively few studies controlling for or explicitly analyzing outcomes by sex (Akil et al. 2018; Fettes et al. 2018; Williams et al. 2016). Understanding the pathophysiological mechanisms of TRD, especially across genders, is crucial for explaining the variations in its clinical presentation (Sun, Luo, Ma, et al. 2022). This study aims to provide a comprehensive review of the literature to identify the mechanisms underlying the pathophysiology of depression in individuals who do not respond to treatment. Additionally, it seeks to highlight the biological differences reported to date between men and women with TRD.

2. Mechanisms Underlying Sex Differences in Treatment‐Resistant Depression

As mentioned above, women experience depression at more than twice the rate of men, a disparity that may be influenced by biological, hormonal, and immune‐related factors (Bloch et al. 2003). Although sex differences in TRD are evident, the underlying reasons remain unclear. Closing this knowledge gap is crucial, as understanding the mechanisms and genetics that drive distinct susceptibility between TRD‐affected males and females informs personalized prevention and treatment strategies.

3. Genetic and Chromosomal Context

Genetic sex is determined by the copies of sex chromosomes an organism carries, with females having duplicated X chromosomes, and males having a single X plus a Y chromosome. This chromosomal difference shapes core cellular programs such as cell‐cycle surveillance, DNA‐damage repair, chromatin remodeling, and telomere maintenance through dosage‐sensitive genes that reside on the sex chromosomes (Carrel and Brown 2017). These sex chromosomes have distinctive biology and epigenetics, which have caused their exclusion from disease studies. Nonetheless, the X chromosome accounts for approximately 5% of the genome and has therefore been omitted from GWAS and other genetic disease analyses (Balaton et al. 2018).

In females, X chromosomes undergo inactivation (XCI) through the expression of the XIST allele, which can be inherited from either the father or the mother. This process helps maintain balanced gene expression. However, 15% to 30% of X‐linked genes escape XCI from both the active and inactive X chromosomes, especially during chronic stress and aging (Balaton et al. 2018; Tukiainen et al. 2017). These genes include key regulators of the cell cycle and DNA damage response and are generally expressed at higher levels in females, with some exceptions, such as genes in PAR1 (Carrel and Brown 2017). A summary of X‐linked genes relevant to MDD is provided in Table 1.

TABLE 1.

X‐linked genes implicated in depression: functions, outcomes, and sex‐specific associations.

Gene	Function	Outcome	Link to depression	References
KDM6A/UTX (Xp11.3)	Histone H3K27 di/tri‐demethylase; co‐activator of p53 and NF‐κB pathways	XCI escape → ~2‐fold higher baseline levels in XX cells; depletion increases cell proliferation and alters immune signaling; drives sex‐dimorphic hypothalamic differentiation (mouse)	Higher expression (p < 0.05) in MDD female patients compared to healthy controls	Cabrera Zapata et al. (2021); Dupéré‐Richer et al. (2024); Ji et al. (2015)
KDM5C/JARID1C (Xp11.22)	Histone H3K4 di/tri‐demethylase → transcriptional repression of neuronal & synaptic genes	Overexpressed in lymphoblastoid cells and post‐mortem brains of women with MDD (p < 0.001) compared to healthy controls	Linked to apoptotic dysregulation and depression pathology	Ji et al. (2015)
TLR7 (Xp22.2)	Endosomal TLR for ssRNA mediation of type‐I IFN cascade	Activation in animal models raises IFN‐α and impairs synaptic plasticity	Significant difference after 4 weeks of antidepressants (p < 0.001)	Hung et al. (2016); O'Driscoll et al. (2017)
IRAK1 (Xq28)	Ser/Thr kinase downstream of MyD88 in TLR/IL‐1R signaling; phosphorylates TRAF6 → NF‐κB/MAPK activation	Partial XCI escape; female vs. male neonate cord‐blood leukocytes: ↑ IRAK1 mRNA & protein, plus enhanced cytosolic localization; Dosage likely amplifies TLR4 responses; Expression modulated by miR‐146a and sex steroids (hypothesized)	IRAK1 has been identified in depressed patients with Parkinson's Disease in cognitively preserved (p = 0.031), sex stratification (r = −0.4, p = 0.065); downregulation in MDD hippocampus	Gao et al. (2015); O'Driscoll et al. (2017); Su et al. (2025)

It has been proposed that increased expression of genes escaping XCI may confer functional advantages (O'Driscoll et al. 2017). However, skewed XCI, which involves the preferential silencing of one parental X chromosome, has been observed in approximately 50% of adult females (Shvetsova et al. 2019). Additionally, XIST expression was significantly higher in females with MDD compared to healthy controls (Ji et al. 2015), suggesting a disruption in dosage compensation mechanisms that may contribute to MDD pathophysiology.

In males, the Y chromosome contributes uniquely to mood regulation via SRY and other ampliconic genes that modulate catecholamine synthesis (e.g., tyrosine hydroxylase, monoamine oxidase A) in mid‐brain dopamine neurons. Knockdown of SRY reduces dopaminergic firing and blunts the corticosterone response to stress, whereas its overexpression enhances stress reactivity and induces depressive‐like behaviors, implicating a Y‐linked mechanism underlying male‐predominant susceptibility to TRD (Czech et al. 2012).

Sex‐stratified genome‐wide association studies (GWAS) support divergent genetic architectures. Eleven genome‐wide loci were identified in females with broad depression, compared to only one in males. Female loci were enriched for adaptive immunity and DRD2/GRM5 signaling, whereas male loci centered on epigenetic modifiers of neurotransmitter release, paralleling the X versus Y‐driven mechanisms outlined above (Silveira et al. 2023). In the largest GWAS of TRD to date, three genome‐wide loci were identified, all autosomal, and associated with zinc‐finger transcription factors (RNF219‐AS1, MECOM, ZNF48/25/33A) and complement receptor biology (CR1L) (Li et al. 2020). All loci map to autosomal regions enriched for zinc‐finger transcription factors (RNF219‐AS1, MECOM, ZNF48/25/33A) or complement receptor biology (CR1L). Gene set enrichment analysis highlighted innate immune mediators (LTB, LST1, NCR3) that overlap with TLR7‐linked mechanisms (Table 1). Notably, this study excluded sex chromosomes, underscoring a critical gap in understanding sex‐specific genetic risk for TRD.

Collectively, these findings outline a dosage‐sensitive, sex‐chromosome framework for TRD. First, escape and skewing on the X contribute to female‐specific neural mosaics of resilience or liability. Second, Y‐linked dopaminergic modulators may define a male‐biased stress reactivity axis. Third, both common and rare TRD risk alleles are enriched in genes governing chromatin accessibility, synaptic plasticity, and cell cycle control, many of which are regulated by X‐chromosome dosage. This chromosomal architecture provides a mechanistic substrate for the well‐documented disparities in depression outcomes between males and females and offers testable targets, such as XCI modifiers or SRY antagonists, for precision therapeutics in treatment‐resistant populations.

4. Sex‐Steroid Signaling

Hormonal differentiation, driven by the presence or absence of the SRY gene on the Y chromosome, initiates divergent endocrine environments. In males, testicular development leads to lifelong exposure to androgens, whereas in females, ovarian development establishes cyclical production of estrogen and progesterone. These steroid hormones are lipophilic and readily cross the blood–brain barrier (BBB), acting both genomically and non‐genomically within the central nervous system (CNS). In addition to passive diffusion from the periphery, neurons and glial cells can synthesize neurosteroids locally, enabling region‐specific hormonal regulation. These mechanisms, supported by neuroimaging, postmortem, and preclinical studies, underscore the central, not merely peripheral, role of sex hormones in modulating neural plasticity, stress responsivity, and affective processing (Diotel et al. 2018; Lafta et al. 2024).

Estrogens, particularly 17β‐estradiol, can bind to nuclear or membrane receptors and potentially stimulate many different interconnected signaling pathways, including PI3K/Akt and ERK/MAPK signaling, promoting neuronal survival, synaptogenesis, and dendritic branching (Fiocchetti et al. 2012; Sellers et al. 2015). Estrogen also enhances serotonergic tone, modulates glutamatergic plasticity, and facilitates brain‐derived neurotrophic factor (BDNF) expression, all of which have been implicated in antidepressant response (Hwang et al. 2020).

Progesterone often exhibits context‐dependent effects, attenuating proliferation or promoting differentiation in certain brain regions, where it can be locally converted to allopregnanolone, which modulates GABAA receptors (Guennoun 2020). These effects may confer transient resilience to stress or antidepressant response under specific conditions, although sustained progesterone elevation, as seen in luteal‐phase dysphoria, has been associated with affective destabilization (Schüle et al. 2014).

Testosterone influences both synaptic transmission and synaptic remodeling, albeit in a more variable and region‐specific manner, potentially reflecting differential androgen receptor expression or conversion to estrogens via aromatase in specific neural circuits (Grassi et al. 2010). Sex hormones also influence drug metabolism, receptor expression, and barrier permeability, thereby modulating the pharmacodynamics and pharmacokinetics of antidepressants. These interactions likely contribute to the observed sex differences in treatment efficacy, tolerability, and remission rates, and may partly explain the sex bias in the prevalence and trajectory of TRD (Bosch et al. 2025; Romanescu et al. 2022).

5. The Role of Mitochondrial Dysfunction in Treatment‐Resistant Depression

A second axis of sex‐based biological divergence lies in the mitochondria. These organelles contain a maternally inherited 16.6‐kilobase genome that encodes key proteins essential to oxidative phosphorylation and cellular energy metabolism. Mitochondrial function is further shaped by nuclear sex‐linked gene expression and is tightly regulated by circulating steroid hormones. Sex differences in mitochondrial dynamics, including reactive oxygen species (ROS) generation, calcium buffering, and metabolic substrate preference, have been increasingly linked to stress sensitivity and antidepressant response (Belenguer et al. 2019; Brasanac et al. 2022; Chakrabarty et al. 2018).

Mitochondria are essential for energy generation through oxidative phosphorylation and are critical regulators of cellular metabolism, oxidative stress, and apoptosis. Recent findings suggest that mitochondrial dysfunction, such as impaired ATP production and elevated oxidative stress, may contribute to the onset of depression and resistance to treatment (Allen et al. 2018). Pan et al. (2023) have identified metabolic signatures specific to TRD, highlighting mitochondrial stress markers such as fibroblast growth factor 21 (FGF21) and growth differentiation factor 15 (GDF15), which demonstrate high diagnostic accuracy. Notably, up to 76% of the metabolic disruptions in TRD and suicidal ideation involve lipid pathways—including sphingolipid and phospholipid metabolism—alongside altered mRNA/tRNA turnover and dysregulated purine and purinergic signaling.

Hormonal fluctuations throughout a woman's life appear to exacerbate mitochondrial dysfunction. In fact, during periods of hormonal variability, such as premenstrual syndrome or menopause, mitochondrial function may be compromised, increasing susceptibility to depressive symptoms (Bansal and Kuhad 2016; Caruso et al. 2019; Khan et al. 2023). Estradiol (17β‐E2) enhances mitochondrial efficiency by upregulating oxidative phosphorylation complexes I, IV, and ATP synthase, increasing ATP production, and improving the respiratory control ratio (Irwin et al. 2008; Lejri et al. 2018; Nilsen et al. 2007; Olajide et al. 2024; Rettberg et al. 2014). Estrogens also reduce mitochondrial ROS production and lipid peroxidation by inducing antioxidant enzymes, notably superoxide dismutase (SOD2) and peroxiredoxins (Guo et al. 2012; Irwin et al. 2008; Mooga et al. 2018). Importantly, mitochondrial estrogen receptor beta (ERβ) mediates critical protective effects, influencing mitochondrial dynamics and biogenesis through PGC‐1α, NRF1 signaling, and proteins involved in mitochondrial fusion/fission, such as Drp1 and MFN2 (Arnold et al. 2008; Kalkhoran and Kararigas 2022; Yang et al. 2009).

Progesterone further complements these actions by maintaining mitochondrial coupling and stability, supporting cytochrome oxidase activity, stabilizing mitochondrial membrane potential, and limiting ROS leakage (Andrabi et al. 2017; Gaignard et al. 2016). These mitochondrial protective effects of progesterone are pronounced after ischemic or traumatic injury, where progesterone stabilizes ATP production and restricts mitochondrial permeability transition pore (mPTP) opening (Robertson and Saraswati 2015; Yousuf et al. 2016). Conversely, synthetic progestins such as medroxyprogesterone acetate (MPA) can antagonize the beneficial mitochondrial effects of estradiol, emphasizing distinct mitochondrial outcomes based on hormone specificity (Irwin et al. 2011). Overall, females exhibit greater mitochondrial resilience, largely attributed to the regulatory roles of estrogen and progesterone, an advantage that diminishes with age or following ovariectomy (Gaignard et al. 2015; Kalimon and Sullivan 2021; Olajide et al. 2024). These baseline differences emphasize the critical contribution of sex steroids to mitochondrial homeostasis and their potential role in shaping depressive vulnerability.

In males, physiological testosterone supplementation reverses age‐related and injury‐induced mitochondrial deficits by restoring activities of mitochondrial complexes I–IV, stabilizing mitochondrial membrane potential, and enhancing mitochondrial biogenesis and antioxidant defenses (Carteri et al. 2019; Yan et al. 2021). However, compared to estrogenic mechanisms, testosterone's mitochondrial actions are less thoroughly characterized, underscoring a need for further research.

Although mitochondrial dysfunction is strongly implicated in the pathophysiology of MDD, including in treatment‐resistant forms, direct experimental evidence linking sex steroid‐induced mitochondrial changes to depressive symptoms remains sparse. Despite theoretical links between reduced estrogenic states and increased mitochondrial dysfunction contributing to depression risk (Allen et al. 2018), empirical validation of these mechanisms in depressive phenotypes is critically needed. Moreover, mitokines like FGF21 and GDF15, although proposed as mitochondrial biomarkers, lack direct evidence within sex steroid contexts related to depression, representing important areas for future investigation.

In summary, sex steroids significantly modulate mitochondrial function, influencing neuronal resilience and potentially affecting vulnerability to depressive disorders, including TRD. Future studies should address existing gaps, specifically examining how sex‐specific mitochondrial mechanisms contribute directly to clinical outcomes in MDD, and guiding the development of novel therapeutic strategies based on sex‐informed mitochondrial modulation.

Together, these chromosomal, hormonal, and mitochondrial differences constitute a biologically grounded framework for understanding how sex shapes central nervous system vulnerability in TRD. These upstream influences modulate critical downstream processes, including neuroinflammation, neurotransmitter signaling, neurotrophic factor expression, and synaptic remodeling, all of which are explored in the sections that follow.

6. Neurotransmitter Mechanisms in Treatment‐Resistant Depression

The monoaminergic hypothesis of depression, centered on serotonin (5‐HT), norepinephrine (NE), and dopamine (DA) signaling, has long shaped the pharmacological approach to MDD (Schildkraut 1965). However, the persistence of depressive symptoms in a substantial proportion of patients despite adequate monoaminergic modulation suggests additional interplay, especially in TRD, where deeper disruptions in neurochemical balance and synaptic plasticity appear to be involved (Table 2).

TABLE 2.

Sex‐based neurotransmitter profiles and their implications for treatment‐resistant depression (TRD).

System	Female profile	Male profile	TRD implication
Serotonin (5‐HT)	↑ 5‐HT1A receptor density ↑ SSRI response ↑ tryptophan sensitivity	↓ 5‐HT1A expression ↓ SSRI efficacy	SSRIs more effective in women
Dopamine (DA)	↓ D1 binding ↑ DAT availability	↑ DA toneDRD2 polymorphism modulates severity	Men may benefit from dopaminergic agents
Norepinephrine (NE)	Unknown sex‐specific NET differences	↑ NE/DA reuptake blockade may enhance response	NDRIs preferable in men
Glutamate (Glx)	↑ Glx in striatum & cerebellum ↑ hormonal sensitivity	↑ Glx in PFC	Region‐ and hormone‐specific modulation in TRD
GABA	↓ GAD65/67, SST expression in women	More stable GABAergic expression	GABA‐targeted interventions may benefit women

Sex‐based differences in neurotransmitter systems have emerged as a relevant factor shaping both affective symptomatology and treatment response. These differences are mediated by a combination of chromosomal, hormonal, and epigenetic influences. For instance, estrogen and progesterone modulate serotonin synthesis, receptor density, and reuptake dynamics, while testosterone affects dopaminergic tone and transporter expression (Kim et al. 2018; LeGates et al. 2019). Neuroimaging studies have demonstrated that women with MDD show increased central serotonin synthesis despite lower plasma tryptophan (TRP) levels and greater mood sensitivity to TRP depletion, suggesting heightened serotonergic vulnerability (Frey et al. 2010; Moreno et al. 2006). Paradoxically, women generally respond better to SSRIs than men, who show improved outcomes with norepinephrine‐dopamine reuptake inhibitors (NDRIs) or tricyclics (Keers and Aitchison 2010; Kornstein et al. 2000).

Beyond the serotonergic system, catecholamines such as dopamine and norepinephrine also play key roles in the pathophysiology of TRD and are differentially regulated by sex. Women have been shown to exhibit higher dopamine transporter (DAT) availability and lower D1 receptor binding in the caudate, with differential neurochemical responses to bupropion (Williams et al. 2021). In the prefrontal cortex (PFC), 5‐HT plays a crucial role in modulating dopamine activity while also suppressing noradrenergic and dopaminergic neuronal function observed in preclinical research (Barbon et al. 2011; Di Matteo et al. 2008). Specific serotonin receptors facilitate DA release, including 5‐HT1AR, 5‐HT1BR, 5‐HT2AR, 5‐HT3R, and 5‐HT4R, whereas 5‐HT2CR agonists suppress both dopamine and norepinephrine release (Alex and Pehek 2007; Howell et al. 2015). Additionally, DRD2 polymorphisms show sex‐specific associations with depression severity (Douillard‐Guilloux et al. 2017). Norepinephrine, through its interaction with dopamine via the norepinephrine transporter (NET), further complicates this picture, although direct sex comparisons in NET expression remain limited. Clinical data suggest that men may respond more favorably to noradrenergic or dopaminergic agents (Keers and Aitchison 2010; Kornstein et al. 2000). These intricate interactions challenge the conventional monoamine hypothesis, underscoring the multifaceted nature of serotonin regulation and its broader implications for depression pathophysiology.

Glutamate and GABA, the brain's principal excitatory and inhibitory neurotransmitters, respectively, have been increasingly implicated in TRD. MRS studies in TRD patients consistently demonstrate reduced Glx in the PFC and amygdala, and decreased GABA levels in occipital and frontal regions (Baeken et al. 2017; Knouse et al. 2022). Sex differences in glutamate distribution are region‐specific: men exhibit higher Glx in the prefrontal cortex, while women show increased levels in the striatum and cerebellum (O'Gorman et al. 2011; Zahr et al. 2013). These patterns are sensitive to hormonal fluctuations, with peripheral glutamate levels inversely correlated with estrogen and progesterone levels across the menstrual cycle (Zlotnik et al. 2011). Postmortem gene expression studies further reveal greater dysregulation of glutamate and GABA signaling in women, including reduced expression of somatostatin and GAD enzymes, key markers of GABAergic interneuron function (Gray et al. 2015; O'Gorman et al. 2011).

Together, these findings support the hypothesis that sex modulates the structure and function of neurotransmitter systems across multiple levels, from receptor distribution to synaptic clearance and neuroplasticity. These sex‐based differences not only influence baseline affective regulation but also shape differential responsiveness to antidepressant therapies. In TRD, where neurotransmitter adaptation to treatment is impaired, such differences may contribute to sex‐specific pathways of vulnerability. Incorporating sex as a biological variable in neurochemical studies and clinical trial design is thus essential for developing biomarker‐informed, individualized treatment strategies.

7. The HPA Axis: Stress and Resilience in Treatment‐Resistant Depression

The hypothalamic–pituitary–adrenal (HPA) axis plays a crucial role in the body's stress response, with important implications for depression and resilience. Activation of the HPA axis in response to stressors triggers a cascade of hormonal reactions culminating in cortisol release (Pariante and Lightman 2008). Dysregulation of this axis is common in depression, characterized by hyperactivity resulting in elevated cortisol levels (hypercortisolemia), reduced inhibitory feedback mechanisms, increased levels of corticotropin‐releasing hormone (CRH) and adrenocorticotropic hormone (ACTH), and enlargement of the pituitary and adrenal glands (M. F. Juruena 2014). Increased pituitary volume, specifically, has been identified as a marker of HPA axis activation (Kessing et al. 2011).

The relationship between stress, the HPA axis, and depression is complex and bidirectional. While HPA axis dysregulation may predispose individuals to depression, depressive symptoms themselves can further disrupt HPA axis function, creating a cycle of stress vulnerability and mood disturbance (M. Juruena 2015). Interestingly, these changes appear to be state‐dependent, often resolving with depressive symptom remission (M. Juruena 2015). Nevertheless, persistent HPA axis hyperactivity correlates with higher relapse rates, suggesting the importance of axis normalization for sustained recovery (Varghese and Brown 2001).

In TRD, studies have observed conflicting cortisol patterns. Some research describes a blunted cortisol awakening response and overall hypocortisolemia in TRD patients compared to healthy controls, while others report elevated salivary and plasma cortisol levels (Markopoulou et al. 2009; Wu et al. 2022). In contrast, other studies have found no significant differences in cortisol levels between TRD patients and healthy controls (Bauer et al. 2002, 2003). Juruena et al. (2013) showed that TRD patients failed to demonstrate cortisol increases in response to the mineralocorticoid receptor antagonist spironolactone, suggesting dysfunction in mineralocorticoid receptors, potentially due to receptor downregulation or altered receptor activity. Additionally, impaired conversion of spironolactone into its active metabolite, canrenone, indicates pharmacokinetic abnormalities linked to mineralocorticoid receptors, aligning with broader evidence of receptor alterations in depression‐associated cognitive deficits and stress dysregulation (Otte et al. 2015).

While mineralocorticoid receptor agonists, such as fludrocortisone, have been shown to improve cognitive function in non‐resistant depression, their efficacy in TRD appears limited, likely due to more pronounced mineralocorticoid receptor dysfunction in this subgroup (Otte et al. 2015). Notably, neuromodulatory interventions including intermittent theta‐burst stimulation and rTMS have demonstrated efficacy in normalizing HPA axis hyperactivity among TRD responders, potentially by enhancing prefrontal cortex‐limbic‐HPA axis connectivity, thereby restoring inhibitory feedback mechanisms (Duval et al. 2024; Pridmore 1999; Zwanzger et al. 2003). Additionally, Mickey et al. (2018) identified distinct cortisol trajectories preceding ECT, with cortisol patterns predicting treatment response with approximately 80% accuracy, suggesting potential for cortisol‐based biomarkers to guide intervention strategies in TRD.

Sex differences in HPA axis function are well‐documented, with females generally exhibiting higher baseline cortisol levels and enhanced ACTH and cortisol responses during acute stress compared to males (Goel et al. 2014; Heck and Handa 2019; Sheng et al. 2021). These differences are largely attributed to the modulatory effects of sex hormones on HPA axis regulation and receptor sensitivity (Carey et al. 1995; Roca et al. 2003; Sharma et al. 2014). Furthermore, females generally demonstrate weaker glucocorticoid receptor‐mediated negative feedback, prolonging HPA axis activation post‐stress. Conversely, males show more efficient negative feedback, facilitating quicker HPA recovery (Bangasser and Valentino 2014; Heck and Handa 2019). These differences may explain the higher prevalence of stress‐related disorders among females. Early‐life stress exposure also appears to impact sexes differently, with females being more susceptible to sustained HPA dysregulation and greater vulnerability to depression and anxiety disorders, whereas males tend to exhibit higher resilience to similar stressors (An et al. 2022; Hodes and Epperson 2019). Evidence indicates that females exhibit a more active HPA axis even under baseline, non‐stress conditions, and during acute stress, they show an exaggerated release of ACTH and corticosterone, reflecting a heightened physiological stress response relative to males (Goel et al. 2014; Heck and Handa 2019; Sheng et al. 2021). In contrast, males tend to exhibit a blunted stress response, with lower ACTH and corticosterone secretion following similar stressors (Bangasser and Cuarenta 2021; Bangasser and Valentino 2014). Another key difference lies in glucocorticoid receptor (GR)‐mediated feedback inhibition. Females typically demonstrate weaker GR‐mediated negative feedback, which may contribute to prolonged HPA axis activation during stress. Conversely, males exhibit stronger feedback regulation, allowing for a more rapid recovery of HPA axis activity post‐stress (Bangasser and Valentino 2014; Heck and Handa 2019).

In summary, significant dysregulation of the HPA axis typifies TRD, with robust evidence of sex‐specific hormonal influences affecting receptor functionality and stress responses. These mechanistic insights underline the importance of personalized therapeutic strategies tailored to sex‐based biological differences. Moving forward, future research should prioritize comparative studies between TRD and non‐TRD cohorts utilizing dynamic HPA measures, integrating genetic, epigenetic, and inflammatory markers. Such comprehensive approaches promise enhanced precision in treatment selection, ultimately improving outcomes for patients struggling with TRD.

8. Role of Inflammation in Treatment‐Resistant Depression

Accumulating evidence underscores the role of neuroinflammation in the pathophysiology of TRD, highlighting elevated pro‐inflammatory cytokines, microglial activation, and immune system dysregulation. Patients with TRD exhibit persistently higher levels of interleukin‐6 (IL‐6), tumor necrosis factor‐alpha (TNF‐α), and C‐reactive protein (CRP) compared to treatment‐responsive patients and healthy controls (Chamberlain et al. 2019; Haroon et al. 2018; Osimo et al. 2020). Conversely, findings on interleukin‐1β (IL‐1β) levels remain varied. Uint et al. (2019) reported elevated IL‐1β plasma levels in TRD patients relative to healthy controls, while other studies observed lower IL‐1β levels among those with TRD (Kalkman and Feuerbach 2016; Zincir et al. 2016). Strawbridge et al. (2019) found that elevated inflammatory proteins, such as IL‐6 and TNF‐α, predicted poorer treatment response in patients with TRD. Patients with elevated inflammation tended to have more severe symptoms, cognitive impairment, and greater treatment resistance, highlighting the complex interplay between inflammation, the neurobiology of depression, and clinical outcomes. These biomarkers correlate with the severity of depressive symptoms and predict poor treatment response to conventional antidepressants, suggesting that an “inflammatory subtype” of MDD may be characteristic of patients with TRD (Strawbridge et al. 2017, 2018). This subtype is characterized by obesity, cognitive deficits, a history of early life stress, and a more significant presence of systemic inflammation (Strawbridge et al. 2018).

A systematic review conducted by Mancuso et al. (2023) identified TNF‐α as one of the most relevant biomarkers in discriminating patients with TRD. The study showed that patients with TRD have significantly higher levels of TNF‐α compared to controls and patients who respond to antidepressant treatment. Furthermore, the reduction in TNF‐α levels after treatment was correlated with clinical improvement, suggesting that modulation of inflammation may be a promising therapeutic target for patients with TRD. Rengasamy et al. (2022) demonstrated that inflammation is associated with reduced functional connectivity between the ventral striatum and the ventromedial PFC in patients with TRD, suggesting that inflammation may directly impact brain function in regions related to mood regulation. This effect is even more pronounced in individuals with a history of childhood trauma, indicating that environmental factors and adverse experiences may exacerbate inflammatory responses and influence treatment resistance.

Studies have also described changes in the kynurenine pathway in TRD. Zhou et al. showed that patients had significantly different serum kynurenine (KYN) metabolites (including TRP and kynurenic acid (KYNA) levels, KYN/TRP ratio, and KYNA/KYN ratio) at baseline compared to controls. In the same study, the authors observed that ketamine increased serum KYNA levels and the KYNA/KYN ratio at 24 h following the first infusion in responders compared to nonresponders, and this elevation lasted up to 14 days following the sixth infusion, which was related to symptomatic improvement. Another study found that treatment with ECT induced significant changes in peripheral kynurenine metabolites in patients with MDD. Thus, plasma levels of TRP, kynurenine, and quinolinic acid significantly decreased after a series of three ECT administrations, whereas plasma levels of KYNA did not change (Schwieler et al. 2016). Guloksuz et al. (2015) showed that levels of KYNA, the KYN/TRP ratio, the KYNA/KYN ratio, and the KYNA/3‐hydroxykynurenine ratio rise over the duration of the ECT treatment and continue to increase for up to 6 weeks after the final ECT session (Guloksuz et al. 2015). In the same line, Ryan et al. (2020) showed that improvements in mood scores exhibited a correlation with elevated levels of KYN, 3‐hydroxykynurenine, 3‐hydroxyanthranilic acid, and quinolinic acid, as well as the KYN/TRP ratio following ECT. Changes in the KYN pathway were also noted in patients with TRD following rTMS treatment, indicating that rTMS significantly elevated plasma TRP levels and reduced plasma serotonin levels. However, plasma KYN and kynurenic acid levels and the KYN/TRP ratio remained unchanged (Tateishi et al. 2021).

Sex‐based differences in immune responses might also contribute to the development of TRD, particularly in females, who exhibit a heightened pro‐inflammatory state (Kropp and Hodes 2023). This heightened immune response is linked to an increased risk of immune‐mediated diseases such as systemic lupus erythematosus, rheumatoid arthritis, and multiple sclerosis, commonly comorbid with depression (Stumper et al. 2020). These differences become more pronounced after puberty, where estrogen regulates immune signaling pathways by increasing cytokine‐producing leukocytes and elevating pro‐inflammatory gene expression (Kovats 2015; Shepherd et al. 2020). The resulting heightened immune activation in females is believed to contribute to the higher prevalence of depression in women. A retrospective cohort study exploring the relationship between six types of somatic comorbidities found that individuals with TRD had a 22% increased likelihood of developing autoimmune diseases. However, the temporal relationship, whether TRD preceded these conditions, remains unclear (Dome et al. 2021; Lauden et al. 2021). In a separate cross‐sectional study, individuals with TRD exhibited a 52% higher risk of allergic and autoimmune diseases compared to those whose depression responded to treatment (Lauden et al. 2021). Furthermore, a nested case–control study revealed that women with TRD were approximately 67% more likely to develop autoimmune conditions than women without TRD (Chan et al. 2023).

Hormonal modulation of immune responses is central to observed sex differences. Animal models provide critical mechanistic insights into these interactions. Studies have shown that administering progesterone to female mice exposed to lipopolysaccharide (LPS) reduces hypothalamic levels of IL‐6 and TNF‐α, thereby reversing depressive‐like behaviors and establishing a causal, central, and sex‐specific inflammatory mechanism (Xu et al. 2023). Similarly, progesterone administration in male mice exposed to maternal separation stress reduces hippocampal TNF‐α, thereby reversing depressive behaviors and indicating the potent anti‐inflammatory and antidepressant effects of progesterone (Nouri et al. 2019). Moreover, testosterone exerts immunosuppressive effects, reducing pro‐inflammatory cytokines (TNF‐α, IL‐6, IL‐1β) and promoting anti‐inflammatory responses via IL‐10, supported by evidence from animal and clinical studies (Mohamad et al. 2019). Testosterone's anti‐inflammatory and neuroprotective actions, demonstrated in male models of hormone deficiency, reveal that testosterone and estradiol replacement therapy reverse cytokine elevations and depressive‐like phenotypes mediated through microglial activation and Traf6/TAK1 signaling pathways (Peng et al. 2022). In contrast, ovarian hormone deficiencies modeled by ovariectomy in female rodents highlight estrogen's protective role, as decreased estrogen/progestin levels correlate with elevated IL‐6 and TNF‐α, impaired progesterone receptor signaling, and reduced synaptic plasticity (Yang et al. 2019).

Evidence suggests that lower testosterone levels in men are associated with elevated systemic inflammation, which may increase vulnerability to depressive symptoms (Morsink et al. 2007). Testosterone replacement therapy has shown antidepressant potential, likely through its anti‐inflammatory effects and promotion of neuroplasticity (Walther et al. 2019). In contrast, estrogens play a more complex, dual role in immune regulation. At low concentrations, estrogens promote immune activation by upregulating pro‐inflammatory cytokines such as IL‐6, IL‐1β, and TNF‐α (Engler‐Chiurazzi et al. 2022; Galea et al. 2017). At higher levels, however, estrogens exert anti‐inflammatory effects by suppressing cytokine production and enhancing regulatory T cell activity (Salem 2004). In women with TRD, somatic symptoms, including sleep disturbances, fatigue, and appetite changes, are more frequently reported and are strongly associated with elevated inflammatory markers such as CRP and IL‐6 (Hazeltine et al. 2022; Piepenburg et al. 2019). Further supporting sex‐specific inflammatory profiles, IL‐8 exhibits divergent predictive value. In females, lower baseline IL‐8 and post‐treatment increases are associated with favorable outcomes, whereas in males, similar changes predict poorer responses (Kruse et al. 2021, 2020). Additionally, clinical trials of minocycline have identified CRP as a female‐specific predictive biomarker, while IL‐6 appears to have broader predictive validity across sexes (Lombardo et al. 2022).

In summary, chronic inflammation distinctly interacts with sex‐specific biological mechanisms, emphasizing the necessity of personalized, sex‐informed therapeutic strategies for TRD. Integrating insights from animal models and clinical studies will be essential to refine targeted interventions and enhance treatment outcomes for both sexes.

9. Role of BDNF in Treatment‐Resistant Depression

BDNF plays a crucial role in the pathophysiology of depression and the response to antidepressant treatment (Meshkat et al. 2022). However, its involvement in TRD is particularly complex, especially regarding sex differences in its regulation and function. BDNF can also be influenced by pharmacological interventions. Although direct studies specifically on sex differences in BDNF within TRD are limited, substantial evidence from broader depression research underscores sex‐specific variations in BDNF signaling.

Human and animal data together indicate that the interplay between sex hormones and the BDNF system shapes vulnerability to, and expression of, depressive symptoms in women. In a well‐characterized study, Kreinin et al. (2015) showed that, although serum BDNF was globally lower in drug‐free MDD patients than in controls, only untreated women with severe depression (HAM‐D > 24) displayed a positive, “paradoxical” correlation between BDNF concentration and symptom severity; this relationship was absent in men and disappeared after 2 weeks of antidepressant treatment.

Preclinical studies reinforce these findings, demonstrating higher baseline BDNF levels in female rodents compared to males, potentially explaining more robust antidepressant responses in females (Scharfman and MacLusky 2006). Fluctuations in estrogen during menstrual cycles, postpartum periods, and menopause have been associated with increased depression risk, correlating with reductions in BDNF levels (Begliuomini et al. 2007; Chhibber et al. 2017; Kundakovic and Rocks 2022; Pluchino et al. 2009). Estradiol and testosterone levels positively correlate with serum BDNF in perimenopausal women, negatively correlating with depression severity, although BDNF alone does not independently predict depressive symptoms (Hui et al. 2016). Menstrual cycle phases and hormone replacement therapy (HRT) robustly influence plasma BDNF, with levels rising during high estrogen and progesterone phases and with HRT in postmenopausal women (Begliuomini et al. 2007; Pluchino et al. 2009).

Mechanistically, estrogen significantly enhances BDNF expression in the hippocampus and PFC, supporting neuroplasticity and resilience (Luine and Frankfurt 2013; Scharfman and MacLusky 2006). Estradiol binds to ERα and ERβ, which interact with estrogen response elements (EREs) in the BDNF promoter, upregulating BDNF transcription (Galea et al. 2013). Moreover, estradiol can rapidly activate the ERK/MAPK and PI3K/Akt pathways, thereby enhancing BDNF–TrkB signaling and promoting neuronal survival (Abraham et al. 2004; Bryant et al. 2005; Kuroki et al. 2000). On the other hand, testosterone enhances BDNF expression through androgen receptor (AR) signaling, influencing hippocampal neurogenesis and stress resilience (Hung et al. 2019; Li et al. 2012; Zhang et al. 2024). A 2019 systematic review and meta‐analysis of randomized controlled trials on testosterone therapy found no evidence supporting its effectiveness in treating depression. However, it showed positive outcomes in specific subpopulations, including men with TRD or low testosterone levels. Additionally, the meta‐analysis indicated that higher testosterone doses could improve depressive symptoms in eugonadal or older men (Walther et al. 2019).

Clinical implications remain complex. Classical antidepressants, particularly SSRIs and SNRIs, have long been proposed to exert their therapeutic effects in part through modulation of BDNF expression. A recent meta‐analysis evaluating the impact of antidepressants and antipsychotics on BDNF levels reported significant increases in serum, but not plasma, BDNF concentrations following treatment. Despite the increases in serum BDNF after antidepressant treatment, no significant correlation was found between BDNF elevations and clinical symptom improvement, as measured by HAM‐D scores (Merabtine et al. 2024). These findings suggest that while serum BDNF may serve as a biomarker of treatment exposure or neuroplasticity, it may not reliably reflect clinical response, particularly in TRD, where underlying neurobiological dysfunctions are more heterogeneous and multifaceted.

Patients with TRD typically present with reduced peripheral BDNF levels, which correlate with illness severity (Hong et al. 2014; Rana et al. 2021). In TRD specifically, the link between peripheral BDNF and treatment response appears highly variable. Nevertheless, antidepressant treatments, including ketamine and ECT, show inconsistent effects on peripheral BDNF despite clinical improvement. In some cases, serum BDNF increases following ketamine treatment or ECT have been reported (Duncan Jr et al. 2013; Haile et al. 2014), while other studies find no such changes (Allen et al. 2015; Caliman‐Fontes et al. 2023; Vanicek et al. 2019). In a meta‐analysis conducted by Meshkat et al. (2022), no significant correlation was found between changes in BDNF levels and symptomatic changes in patients with TRD.

Understanding the role of BDNF in TRD through a sex‐specific lens highlights the need for personalized treatment approaches. The significant link between BDNF levels and depression severity in women suggests that therapies aimed at boosting BDNF might be especially effective for female patients. On the other hand, the absence of such correlations in men indicates the need to investigate alternative pathways to develop effective treatments for male patients (Psomiades et al. 2022). The observed sex differences in BDNF signaling and regulation suggest personalized therapeutic strategies, particularly hormone modulation, might be beneficial, especially in females. Future research should directly investigate how sex hormones influence central BDNF signaling and explore potential mechanisms underlying observed sex differences to advance precision treatments for TRD.

10. Neuroimaging Correlates of Treatment‐Resistant Depression: Structural, Functional, and Sex‐Specific Signatures

Neuroimaging studies have significantly advanced our understanding of structural and functional brain abnormalities specific to TRD. Structural MRI investigations commonly identify volumetric reductions in key brain regions shared by both treatment‐responsive MDD and TRD (Table 3), including the ventromedial PFC (vmPFC), orbitofrontal cortex, middle frontal gyrus, superior frontal gyrus, and hippocampus (Klok et al. 2019; Shah et al. 2002). Notably, reductions in the caudate nucleus volume and specific alterations in the insula appear uniquely associated with TRD (Johnston et al. 2015; Shah et al. 2002). Diffusion tensor imaging (DTI) has demonstrated widespread white matter disruptions in TRD patients compared to healthy controls, although explicit sex‐specific differences in fractional anisotropy (FA) are typically not reported (Diego‐Adeliño et al. 2013; Ho et al. 2021). Structural asymmetry, such as a smaller left amygdala compared to the right, has been documented, though not specifically analyzed by sex (Mervaala et al. 2000).

TABLE 3.

Structural and functional brain alterations in treatment‐resistant depression (TRD).

Brain region/network	Structural differences (TRD‐specific)	Functional differences (TRD‐specific)	Shared or subtle differences (TRD and MDD)	References
Striatum (caudate)	Reduced gray matter volume	Altered connectivity	Reduced putamen volume	Miola et al. (2023); Klok et al. (2019)
Insula	Reduced gray matter volume	Altered connectivity, increased ReHo	—	Miola et al. (2023); Klok et al. (2019)
Cerebellum (vermis/crus)	Reduced gray matter volume	Increased regional activity	—	Miola et al. (2023); Klok et al. (2019); Yamamura et al. (2016)
Thalamus	—	Increased regional activity (fALFF)	—	Miola et al. (2023)
Visual network	—	Reduced ALFF values	—	Miola et al. (2023)
Prefrontal cortex	Reduced volume (superior and medial frontal gyri)	Altered connectivity	Broadly observed volume reduction in depression	Miola et al. (2023); Klok et al. (2019)
Temporal cortex	Reduced volume (superior and middle temporal gyri)	—	Increased ReHo	Miola et al. (2023); Klok et al. (2019)
Parietal cortex	—	—	Reduced connectivity, altered ALFF/ReHo	Miola et al. (2023)
Amygdala	—	—	Increased gray matter volume	Miola et al. (2023)
White matter integrity	Reduced fractional anisotropy (corpus callosum, cingulum, fronto‐occipital fasciculus, superior longitudinal fasciculus)	—	—	Miola et al. (2023); Klok et al. (2019)

Abbreviations: ACC, anterior cingulate cortex; ALFF, amplitude of low‐frequency fluctuations; ERC, entorhinal cortex; FA, fractional anisotropy; GM, gray matter; HC, healthy controls; OFC, orbitofrontal cortex; ReHo, regional homogeneity; WM, white matter.

Recent studies in MDD have elucidated sex‐specific anatomical variations relevant to TRD (Table 4). Female patients with depression exhibit significantly higher myelin content (indexed by R1) in the uncinate fasciculus and corpus callosum genu compared to healthy females; this effect is absent in males (Ho et al. 2021). Additionally, female patients show reduced cortical surface areas across several right‐hemisphere regions, including the superior frontal gyrus, medial orbitofrontal cortex, superior and middle temporal gyri, lateral occipital cortex, and inferior parietal lobule relative to male patients (Mou et al. 2023). Other structural MRI comparisons with same‐sex healthy controls reveal that males with depression show larger gray‐matter volume in the left cerebellum but reduced volumes in the bilateral middle temporal gyri and ventromedial prefrontal cortex, whereas female patients exhibit gray‐matter reductions chiefly in the lingual gyrus (visual cortex) and dorsomedial PFC (Yang et al. 2017). These findings from broader MDD cohorts provide crucial insights into structural sex‐based vulnerabilities that may also inform the neurobiology of TRD.

TABLE 4.

Unified summary of neuroanatomical and functional sex‐based differences in TRD.

Brain region/network	Male‐specific findings	Female‐specific findings	References
Striatal regions (caudate, putamen)	Reduced GM density and volume; decreased ALFF in caudate	Increased ALFF; higher connectivity in caudate, nucleus accumbens	Mohammadi et al. (2023)
Insular cortex	Elevated connectivity in subgenual and posterior cingulate; decreased connectivity in anterior insula	Increased connectivity (subcallosal cingulate)	Mohammadi et al. (2023)
Cerebellum (vermis/crus)	Increased gray matter volume	—	Mohammadi et al. (2023)
Anterior cingulate cortex (ACC)	Smaller inferior ACC normalized volume	Larger rostral ACC volume	Mohammadi et al. (2023)
Habenula	—	Increased habenula WM volumes	Mohammadi et al. (2023)
Entorhinal cortex (ERC)	Larger ERC volume	Smaller ERC volume	Mohammadi et al. (2023)
Prefrontal cortex	Reduced ventromedial prefrontal volume	Reduced dorsomedial prefrontal volume	Miola et al. (2023); Klok et al. (2019); Mohammadi et al. (2023)
Temporal cortex (superior/middle)	Reduced GM volume	—	Miola et al. (2023); Klok et al. (2019); Mohammadi et al. (2023)
Thalamus	Reduced right thalamus GM volume	Increased activation during emotional tasks	Miola et al. (2023); Mohammadi et al. (2023)
Amygdala	—	Smaller bilateral amygdala volumes	Mohammadi et al. (2023)
White matter integrity	Reduced FA in parahippocampal cingulum, increased LD in fronto‐occipital fasciculus	Increased FA in genu of corpus callosum and bilateral cerebral peduncle	Miola et al. (2023); Klok et al. (2019); Mohammadi et al. (2023)
Hippocampus	Left–right asymmetry, smaller left hippocampus volume	—	Mohammadi et al. (2023)
Orbitofrontal cortex (OFC)	Negative correlation between depressive symptoms and left lateral OFC	Positive correlation between depressive symptoms and left medial OFC	Mohammadi et al. (2023)

Abbreviations: ACC, anterior cingulate cortex; ALFF, amplitude of low‐frequency fluctuations; ERC, entorhinal cortex; FA, fractional anisotropy; GM, gray matter; LD, longitudinal diffusivity; OFC, orbitofrontal cortex; WM, white matter.

Functional neuroimaging using resting‐state fMRI similarly underscores significant sex‐specific connectivity patterns in depressive disorders. For instance, males exhibit hyperconnectivity within the default mode network (DMN), while females demonstrate region‐specific patterns of altered connectivity (Talishinsky et al. 2022). Other studies highlight sex‐specific reductions in regional homogeneity (ReHo) and amplitude of low‐frequency fluctuations (ALFF), with females exhibiting specific decreases in regions like the dorsolateral PFC and calcarine cortex, while males show contrasting connectivity changes in areas such as the caudate and posterior cingulate cortex (Mei et al. 2022; Sun, Ma, Guo, et al. 2022; Tu et al. 2022). These functional differences suggest distinct neural mechanisms underlying depressive pathology in men and women, which may critically influence TRD progression and treatment response.

Longitudinal research is still scarce, but one prospective rs‐fMRI study has shown that first‐episode patients who subsequently develop TRD already exhibit distinct baseline ALFF abnormalities in somatosensory, auditory, and default‐mode regions relative to those who later respond to treatment (Zhang et al. 2019). However, the current absence of comprehensive sex‐stratified longitudinal analyses substantially limits our understanding of how these sex‐specific neurobiological pathways influence TRD trajectory and prognosis.

The habenula has recently emerged as a key region implicated in TRD. Deep brain stimulation (DBS) targeting the habenula effectively alleviates depressive symptoms in TRD patients, demonstrating its central role in mood regulation and highlighting its potential as a therapeutic target (Wang et al. 2024). Importantly, sex‐specific habenular changes have been identified, with females showing increased habenula white matter volume and connectivity compared to males when contrasting TRD with treatment‐sensitive depression (Cameron et al. 2024; Mohammadi et al. 2023). These sex‐specific habenular differences underscore the necessity of explicitly addressing sex as a biological variable in neuroimaging and clinical studies on depression.

In summary, integrating findings from broader MDD studies highlights critical structural and functional sex‐based differences that may inform our understanding of TRD neurobiology (Figure 1). However, explicit sex‐based analyses remain inadequately represented in the TRD literature. Future research should prioritize clearly delineating these differences, supporting more precise, personalized clinical strategies tailored to the unique neurobiological profiles of male and female TRD patients.

FIGURE 1.

Anatomical and functional brain differences associated with treatment‐resistant depression (TRD) and sex‐based differences. (A) Brain regions exhibiting structural and functional alterations associated with TRD, including differences in cortical and subcortical connectivity, gray matter volume, and cortical surface area. (B) Sex‐specific anatomical and functional variations observed in TRD patients. Arrows indicate increased (↑) or decreased (↓) structural volume or functional connectivity in females and males, respectively. Symbols represent alterations in connectivity, volume, and cortical surface area. These sex‐based distinctions highlight differential vulnerabilities and potential mechanisms underlying TRD. L, left (hemisphere/side). Created with BioRender.com.

11. Conclusion

TRD is a complex and multifactorial disorder characterized by persistent symptoms that resist standard treatments. Increasing evidence emphasizes the important role of sex‐based biological differences in influencing TRD susceptibility, clinical features, and treatment responses. This review highlights that TRD results from a complex interplay of neuroendocrine abnormalities, inflammatory processes, mitochondrial dysfunction, neurotransmitter imbalances, and genetic and epigenetic factors, all of which may manifest and interact differently in men and women (Figure 2). Targeting a single node in this network is unlikely to yield significant or sustained therapeutic benefits unless it represents a convergent point among multiple pathophysiological pathways.

FIGURE 2.

Sex‐based biological differences relevant to treatment‐resistant depression (TRD). This schematic summarizes key molecular, hormonal, immune, and neuroanatomical distinctions between males (left, blue) and females (right, pink) that may contribute to differential risk and expression of TRD. ACC, anterior cingulate cortex; ACTH, adrenocorticotropic hormone; BDNF, brain‐derived neurotrophic factor; GR, glucocorticoid receptor; OXPHOS, oxidative phosphorylation. Created with BioRender.com.

Although biological sex‐based differences in depression are often discussed in terms of hormones, it is essential to recognize the broader sociocultural context in which they emerge. Factors such as sex‐based violence, economic disparities, and caregiving burdens disproportionately affect women, increasing their vulnerability to depression (Kuehner 2017). In addition, female patients with depression are more likely to be unemployed, have lower educational attainment, and experience more recurrent or severe episodes (Kang et al. 2020), all of which are independently associated with a higher risk of treatment resistance (Gronemann et al. 2020). While these sociodemographic factors are critically important, they are difficult to isolate in biological or mechanistic studies.

Despite decades of rigorous research, the precise biological mechanisms underlying the pathogenesis of depression and the variability in antidepressant treatment response remain incompletely understood. Current evidence suggests that depression is a multifactorial and heterogeneous disorder, driven by a constellation of interacting molecular processes. These processes are influenced by individual differences in genetics, environmental exposures, stress history, and notably, sex‐specific biological factors.

This complexity contributes to the diverse clinical presentations of depression and complicates efforts to identify reliable, universal therapeutic targets. Consequently, the field has remained heavily reliant on trial‐and‐error prescribing approaches. Effective treatment will likely require personalized, subtype‐specific approaches that target multiple or converging pathophysiological pathways. The integration of evidence‐based multimodal pharmacotherapies with psychotherapy and other non‐invasive strategies offers a promising avenue for more precise, effective, and tolerable treatments. Advancing this goal will require extensive investigation into multimodal and convergent mechanisms, including sex‐specific considerations, and the development of agents capable of expediting adaptive neuroplastic responses necessary for therapeutic success.

Author Contributions

Francisco E. R. da Silva: writing – original draft, conceptualization. Atakan Yucel: writing – original draft. Antonio P. M. Menezes: writing – original draft. Ana C. Ruiz: writing – review and editing. Marcela C. Carbajal Tamez: writing – review and editing. Tatiana Barichello: writing – review and editing. Giselli Scaini: conceptualization, funding acquisition, writing – review and editing, supervision, project administration. Joao Quevedo: conceptualization, funding acquisition, writing – review and editing, supervision.

Conflicts of Interest

J.Q. has a clinical research support relationship with LivaNova; is a member of the speaker bureau with Myriad Neuroscience and AbbVie; is a consultant for EMS, Libbs, and Eurofarma; is a stockholder at Instituto de Neurociencias Dr. Joao Quevedo; and receives copyrights from Artmed Editora, Artmed Panamericana, and Elsevier/Academic Press. The other authors declare no conflicts of interest.

Peer Review

The peer review history for this article is available at https://www.webofscience.com/api/gateway/wos/peer‐review/10.1111/jnc.70215.

da Silva, F. E. R. , Yucel A., Menezes A. P. M., et al. 2025. “Mechanisms Underlying Treatment‐Resistant Depression: Exploring Sex‐Based Biological Differences.” Journal of Neurochemistry 169, no. 9: e70215. 10.1111/jnc.70215.

Data Availability Statement

The authors have nothing to report.