Neurobiology of Depression: Monoamine Systems, Neuroplasticity Deficits, HPA Axis Dysregulation, and Inflammatory Markers in Major Depressive Disorder

Major depressive disorder (MDD) is a heterogeneous, multisystem illness that affects approximately 280 million people worldwide according to the World Health Organization. The lifetime prevalence in the United States is estimated at 20.6% by the National Institute of Mental Health (NIMH), with a 12-month prevalence of approximately 8.3% among U.S. adults. MDD is the leading cause of disability globally as measured by years lived with disability (YLDs), and its economic burden in the United States exceeds $326 billion annually when accounting for direct healthcare costs, workplace productivity losses, and suicide-related costs.

For decades, the dominant explanatory framework was the monoamine hypothesis — the idea that depression results from a deficiency of serotonin (5-HT), norepinephrine (NE), or dopamine (DA) in key brain circuits. While this model drove the development of every major class of antidepressant from tricyclics to SSRIs, it is now understood to be profoundly incomplete. The therapeutic lag of 2–6 weeks despite immediate monoamine reuptake blockade, the fact that approximately one-third of patients fail to respond to multiple adequate antidepressant trials (as demonstrated in the landmark STAR*D study), and the rapid antidepressant effects of ketamine — which acts primarily on the glutamate system — all demand a more nuanced, systems-level understanding.

Contemporary neuroscience frames MDD as a disorder of neural circuit dysfunction driven by converging pathologies across at least four major domains: (1) monoamine neurotransmitter dysregulation, (2) impaired neuroplasticity and neurotrophic signaling, (3) hypothalamic-pituitary-adrenal (HPA) axis hyperactivity and glucocorticoid resistance, and (4) chronic low-grade neuroinflammation. These systems interact bidirectionally, creating self-reinforcing pathological loops. This article examines each domain with specificity regarding receptor pharmacology, neural circuits, genetic moderators, biomarker data, and treatment implications.

Serotonin (5-HT)

The serotonergic system originates primarily from the dorsal and median raphe nuclei of the brainstem, with projections to the prefrontal cortex (PFC), amygdala, hippocampus, hypothalamus, and basal ganglia. At least 14 serotonin receptor subtypes have been identified, but several are particularly implicated in MDD:

• 5-HT1A receptors: Somatodendritic autoreceptors on raphe neurons inhibit serotonin firing. Postsynaptic 5-HT1A receptors in the hippocampus and PFC mediate anxiolytic and antidepressant effects. PET imaging studies using [11C]WAY-100635 have demonstrated reduced 5-HT1A binding potential in the mesiotemporal cortex, raphe, and PFC of unmedicated depressed patients. Desensitization of presynaptic 5-HT1A autoreceptors during SSRI treatment is believed to partly explain the therapeutic lag.
• 5-HT2A receptors: Upregulated in the PFC of suicide victims in postmortem studies. These receptors are the primary target of psilocybin and other serotonergic psychedelics now under investigation for treatment-resistant depression (TRD).
• 5-HT1B receptors: Serve as terminal autoreceptors modulating serotonin release. Genetic variation in HTR1B has been associated with depression susceptibility in candidate gene studies, though genome-wide significance has not been consistently replicated.

The 2022 umbrella review by Moncrieff et al., published in Molecular Psychiatry, systematically evaluated evidence for the serotonin hypothesis and concluded that there is no consistent evidence of lowered serotonin activity or reduced serotonin levels in depression. This was widely misinterpreted in media as meaning that SSRIs "don't work." The distinction is critical: SSRIs demonstrate efficacy over placebo in randomized controlled trials (RCTs) with a pooled effect size (Cohen's d) of approximately 0.30 in the Cipriani et al. (2018) network meta-analysis. This modest effect size translates to a number needed to treat (NNT) of approximately 7–8 for response (≥50% symptom reduction). SSRIs work, but their mechanism of action likely extends far beyond correcting a simple serotonin deficit — downstream effects on neuroplasticity, BDNF signaling, and neurogenesis appear to be the more proximal mediators of clinical improvement.

Norepinephrine (NE)

The noradrenergic system arises from the locus coeruleus (LC), with widespread projections to cortical, limbic, and cerebellar targets. NE modulates arousal, attention, executive function, and the stress response. Depression — particularly the melancholic subtype characterized by psychomotor retardation, anhedonia, and impaired concentration — has been associated with reduced NE turnover as measured by levels of the metabolite 3-methoxy-4-hydroxyphenylglycol (MHPG) in cerebrospinal fluid, though findings are inconsistent.

SNRIs (serotonin-norepinephrine reuptake inhibitors) such as venlafaxine and duloxetine block both SERT and NET. The Cipriani et al. (2018) network meta-analysis found venlafaxine among the more efficacious antidepressants (OR for response vs. placebo: 1.78), though this came at the cost of relatively higher dropout rates due to side effects. The dual-action NRI component may be particularly relevant for patients with comorbid chronic pain, fatigue, and cognitive symptoms — symptom domains mediated in part by noradrenergic circuits.

Dopamine (DA)

Dopaminergic projections from the ventral tegmental area (VTA) to the nucleus accumbens (NAc) and PFC — the mesolimbic and mesocortical pathways — are central to reward processing, motivation, and hedonic capacity. Anhedonia, a core feature of MDD per DSM-5-TR Criterion A, is strongly linked to dopaminergic hypofunction in this circuit. PET studies using [11C]raclopride have demonstrated blunted dopamine release in the ventral striatum in response to reward cues in depressed patients. Bupropion, which inhibits dopamine and norepinephrine reuptake (NDRI), is the principal antidepressant targeting this system and shows particular utility in patients with prominent anhedonia, fatigue, and hypersomnia. Pramipexole, a D2/D3 dopamine receptor agonist, has shown antidepressant effects in small RCTs, with effect sizes comparable to SSRIs.

The Neuroplasticity Hypothesis

A paradigm shift in depression neurobiology occurred with the recognition that MDD is associated with structural brain changes — not just neurochemical imbalances. Meta-analyses of structural MRI studies consistently demonstrate hippocampal volume reductions of approximately 4–6% in MDD patients compared to healthy controls, with greater reductions correlating with longer illness duration and number of depressive episodes (Schmaal et al., 2016 — ENIGMA consortium meta-analysis, N > 8,900). Volumetric reductions have also been identified in the PFC (particularly orbitofrontal and dorsolateral regions), anterior cingulate cortex (ACC), and caudate nucleus.

These structural changes reflect impaired neuroplasticity — the brain's capacity for synaptic remodeling, dendritic arborization, and adult neurogenesis. Postmortem studies of depressed patients reveal reduced dendritic spine density in the dorsolateral PFC (dlPFC) and reduced neuropil in the subgenual ACC (Brodmann area 25), a region consistently implicated as hyperactive in treatment-resistant depression.

Brain-Derived Neurotrophic Factor (BDNF)

BDNF, acting through the TrkB (tropomyosin receptor kinase B) receptor, is the principal neurotrophin supporting synaptic plasticity, long-term potentiation (LTP), and neuronal survival in the hippocampus and PFC. The neurotrophic hypothesis of depression posits that stress-induced reductions in BDNF contribute to neuronal atrophy and that effective antidepressant treatments restore BDNF signaling. Key evidence includes:

• Serum BDNF levels are reduced in depressed patients compared to controls, with a meta-analytic effect size of approximately d = −0.71 (Molendijk et al., 2014). Levels normalize with successful antidepressant treatment.
• Postmortem hippocampal BDNF mRNA and protein levels are reduced in suicide victims with depression.
• The Val66Met polymorphism (rs6265) of the BDNF gene — present in approximately 20–30% of Caucasian populations and up to 50% of East Asian populations — impairs activity-dependent secretion of BDNF. Met carriers show reduced hippocampal volume and episodic memory performance, though the polymorphism's direct relationship to MDD risk is modest and likely interacts with environmental stressors (gene × environment interaction).
• All established antidepressant modalities — SSRIs, SNRIs, ECT, exercise, and ketamine — increase BDNF-TrkB signaling, suggesting this may be a final common pathway of therapeutic action.

Glutamate and the Rapid-Acting Antidepressant Revolution

Glutamate is the brain's primary excitatory neurotransmitter, and its role in depression has been transformative for the field. Depressed patients show elevated glutamate levels in the PFC and limbic regions on magnetic resonance spectroscopy (MRS), and postmortem studies reveal altered expression of glutamate receptors and transporters.

The landmark Zarate et al. (2006) study demonstrated that a single subanesthetic dose of intravenous ketamine (0.5 mg/kg over 40 minutes) produced rapid antidepressant effects within hours, with 71% of TRD patients meeting response criteria at 24 hours versus 0% on placebo. Ketamine blocks NMDA (N-methyl-D-aspartate) receptors on GABAergic interneurons, disinhibiting glutamate release onto AMPA receptors on pyramidal neurons. This triggers a signaling cascade involving mTORC1 (mechanistic target of rapamycin complex 1) activation, rapid BDNF release, and synaptogenesis — literally rebuilding synaptic connections within hours. This mechanism is fundamentally different from — and far faster than — the gradual monoamine-mediated neuroplasticity induced by traditional antidepressants.

The FDA approved esketamine (Spravato, the S-enantiomer of ketamine) as a nasal spray for TRD in 2019. In the TRANSFORM-2 trial, esketamine plus a new oral antidepressant showed a statistically significant advantage over placebo plus a new oral antidepressant at day 28 (MADRS score difference: −4.0 points, p = 0.020), with response rates of approximately 65% vs. 52%.

The hypothalamic-pituitary-adrenal (HPA) axis is the neuroendocrine system most consistently implicated in MDD. The cascade begins with corticotropin-releasing hormone (CRH) secretion from the paraventricular nucleus (PVN) of the hypothalamus, which stimulates adrenocorticotropic hormone (ACTH) release from the anterior pituitary, which in turn drives cortisol secretion from the adrenal cortex. Under normal conditions, cortisol exerts negative feedback via glucocorticoid receptors (GR) in the hippocampus, PFC, and hypothalamus to suppress further CRH and ACTH release.

In approximately 40–60% of patients with moderate-to-severe MDD, this negative feedback loop is impaired, resulting in:

• Elevated basal cortisol levels: Hypercortisolemia is documented across meta-analyses of salivary, urinary, and plasma cortisol measures. The effect is most pronounced in melancholic and psychotic depression subtypes.
• Dexamethasone non-suppression: The dexamethasone suppression test (DST) is abnormal in approximately 40–50% of patients with melancholic depression. The combined DEX/CRH test shows even higher sensitivity (~80%) for detecting HPA axis dysregulation and has been proposed as a biomarker of relapse risk — persistent non-suppression after treatment predicts higher relapse rates.
• Elevated CRH in cerebrospinal fluid (CSF): Postmortem studies show increased CRH-expressing neurons in the PVN and locus coeruleus of depressed suicide victims. CRH is anxiogenic and depressogenic when injected centrally in animal models.
• Glucocorticoid receptor (GR) resistance: Impaired GR function means cortisol is less effective at shutting down its own production. The FKBP5 gene, which encodes a co-chaperone protein modulating GR sensitivity, has been identified as a moderator. Specific alleles of FKBP5 interact with childhood adversity to predict MDD risk and altered cortisol reactivity — a well-replicated gene × environment finding (Binder et al., 2004; Klengel et al., 2013).

Chronic hypercortisolemia has direct neurotoxic effects on the hippocampus, reducing dendritic branching, suppressing neurogenesis in the dentate gyrus, and contributing to the hippocampal volume loss observed in neuroimaging studies. This creates a vicious cycle: hippocampal damage further impairs GR-mediated negative feedback, perpetuating HPA axis hyperactivity.

Pharmacological attempts to target the HPA axis directly have yielded mixed results. Mifepristone (a GR antagonist) showed promising early results in psychotic depression but failed to achieve consistent efficacy in larger trials. CRH receptor (CRHR1) antagonists have similarly underperformed in clinical development, possibly because targeting a single node in a complex feedback system is insufficient. Nevertheless, HPA axis normalization is observed with successful antidepressant treatment regardless of modality, supporting its role as a state marker of active depression.

The inflammatory hypothesis of depression has emerged as one of the most active research frontiers. A substantial body of evidence links peripheral and central immune activation to depressive symptomatology:

• Meta-analytic evidence of elevated inflammatory markers: Depressed patients show elevated levels of C-reactive protein (CRP), interleukin-6 (IL-6), interleukin-1β (IL-1β), and tumor necrosis factor-alpha (TNF-α) compared to healthy controls (Dowlati et al., 2010). The meta-analytic effect sizes are moderate (e.g., d ≈ 0.50 for CRP), but the findings are remarkably consistent across studies.
• Sickness behavior parallels: Proinflammatory cytokines induce a behavioral syndrome in animals and humans — fatigue, anhedonia, social withdrawal, psychomotor slowing, appetite changes, and sleep disturbance — that closely mirrors the neurovegetative symptoms of MDD.
• Interferon-alpha (IFN-α) model: When IFN-α was used therapeutically for hepatitis C, approximately 20–30% of patients developed clinically significant major depression. Pretreatment with SSRIs substantially reduced this incidence, providing a naturalistic proof-of-concept for the inflammation-depression link.
• Bidirectional relationship with medical illness: MDD rates are 2–3 times higher in patients with inflammatory conditions (rheumatoid arthritis, inflammatory bowel disease, psoriasis, cardiovascular disease), and depression independently worsens outcomes in these conditions.

Mechanisms of Inflammation-Induced Depression

Peripheral inflammatory signals reach the brain through multiple pathways: afferent vagal nerve signaling, circumventricular organs lacking a blood-brain barrier, active transport of cytokines across the BBB, and activation of brain endothelial cells. Once in the CNS, these signals activate microglia — the brain's resident immune cells — which release additional inflammatory mediators and become neurotoxic in their activated (M1) state.

Inflammatory cytokines disrupt monoamine systems through several specific mechanisms:

• Indoleamine 2,3-dioxygenase (IDO) activation: IL-6 and TNF-α upregulate IDO, which shunts tryptophan away from serotonin synthesis and toward the kynurenine pathway. This simultaneously reduces serotonin availability and increases production of quinolinic acid, an NMDA receptor agonist and neurotoxin. Quinolinic acid has been found elevated in the CSF and subgenual ACC of depressed suicide victims.
• Reduced tetrahydrobiopterin (BH4) availability: Inflammatory signaling depletes BH4, a critical cofactor for tryptophan hydroxylase (serotonin synthesis) and tyrosine hydroxylase (dopamine synthesis).
• Increased glutamate excitotoxicity: Activated microglia release glutamate and reduce astrocytic glutamate reuptake, contributing to excitotoxic synaptic damage.
• Impaired BDNF signaling: NF-κB activation by cytokines suppresses BDNF expression, directly linking inflammation to the neuroplasticity deficit.

Clinical Implications: Inflammation-Stratified Treatment

CRP has emerged as a potential treatment-selection biomarker. In the landmark GENDEP and subsequent trials, patients with elevated CRP (>1 mg/L) responded preferentially to nortriptyline (a noradrenergic tricyclic) over escitalopram (an SSRI). A secondary analysis of the STAR*D data found that patients with elevated inflammatory markers had poorer outcomes across all treatment steps. Conversely, the Raison et al. (2013) RCT of infliximab (a TNF-α antagonist) found that while infliximab did not outperform placebo in the overall TRD sample, patients with baseline CRP ≥5 mg/L showed a significant antidepressant response (response rate: ~62% vs. ~33% in the high-CRP placebo group). This points toward the possibility of immunologically defined depression subtypes — a concept being actively investigated in precision psychiatry.

Neuroimaging has revealed consistent patterns of circuit dysfunction in MDD, moving the field from neurochemistry to network neuroscience:

• Default Mode Network (DMN) hyperconnectivity: The DMN — comprising the medial PFC, posterior cingulate cortex (PCC), precuneus, and angular gyrus — is active during self-referential processing and mind-wandering. Meta-analyses of resting-state fMRI studies show increased DMN connectivity in MDD, which correlates with rumination severity. The DMN fails to deactivate appropriately during cognitive tasks, potentially reflecting the intrusive self-focused negative thinking that characterizes depression.
• Amygdala hyperreactivity: The amygdala shows exaggerated responses to negative emotional stimuli (fearful and sad faces) in depressed patients, with reduced top-down inhibition from the dlPFC. This bias toward threat detection and negative valence may underlie the cognitive distortions described in Beck's cognitive model.
• Subgenual anterior cingulate cortex (sgACC / Brodmann area 25): This region is metabolically hyperactive in treatment-resistant depression, as demonstrated in Helen Mayberg's seminal PET studies. Successful antidepressant treatment — whether by medication, CBT, or deep brain stimulation (DBS) — is associated with reduced sgACC activity. Mayberg's small open-label DBS trial targeting the sgACC showed sustained remission in approximately 60% of TRD patients at 12 months, though larger controlled trials have been slower to confirm these results.
• Reward circuit hypoactivation: Reduced activation of the ventral striatum and orbitofrontal cortex during reward anticipation tasks, as shown in the Pizzagalli et al. body of research, predicts anhedonia severity and poorer treatment response.
• Prefrontal hypofunction: Reduced dlPFC activation during executive tasks is a consistent finding and underlies the cognitive deficits (impaired concentration, indecisiveness) that are part of DSM-5-TR Criterion A for MDD. Left dlPFC hypoactivity is the therapeutic target of repetitive transcranial magnetic stimulation (rTMS).

Emerging work using functional connectivity-based biomarkers has identified potential depression "biotypes." Drysdale et al. (2017) used resting-state connectivity patterns to define four neurophysiological subtypes of depression that differentially predicted response to rTMS. While this study requires replication and has faced methodological critiques, it represents the direction of biomarker-guided treatment selection.

MDD has a heritability of approximately 30–40%, based on twin studies — lower than bipolar disorder (~80%) or schizophrenia (~80%), but still substantial. However, the genetic architecture is highly polygenic. The largest genome-wide association study (GWAS) to date, conducted by the Psychiatric Genomics Consortium (PGC) with over 800,000 participants (Howard et al., 2019), identified 102 independent genetic loci associated with MDD. Individually, each variant contributes a tiny fraction of risk (OR typically 1.02–1.05). Polygenic risk scores (PRS) derived from these data currently explain approximately 1.5–3.2% of variance in MDD liability — enough for research stratification but insufficient for clinical prediction at the individual level.

Notable implicated genes and pathways include:

• SLC6A4 (serotonin transporter): The 5-HTTLPR polymorphism was the subject of the influential Caspi et al. (2003) gene × environment study showing that the short (s) allele moderated the effect of childhood maltreatment on depression risk. This finding has been both replicated and contested; a large collaborative meta-analysis by Border et al. (2019) failed to find support, while others argue that methodological heterogeneity accounts for discrepancies. The field now views single-gene × environment interactions with appropriate skepticism, favoring polygenic approaches.
• FKBP5: Moderates HPA axis reactivity and interacts with early adversity via epigenetic mechanisms — specifically, childhood abuse induces demethylation of glucocorticoid response elements in intron 7 of FKBP5, leading to persistent GR resistance (Klengel et al., 2013). This is one of the best-characterized epigenetic mechanisms in psychiatric genetics.
• GABBR1, DRD2, GRM5: Loci implicating GABAergic, dopaminergic, and metabotropic glutamate receptor pathways, consistent with the neurobiological models described above.
• Synaptic and neuroplasticity genes: Multiple GWAS hits fall within genes involved in synaptic adhesion, axon guidance, and neuronal development, reinforcing the neuroplasticity hypothesis.

Epigenetic mechanisms — DNA methylation, histone modification, and non-coding RNA regulation — are increasingly recognized as mediators between environmental adversity (particularly childhood maltreatment, which approximately doubles MDD risk) and gene expression. These modifications can be long-lasting but are also potentially reversible, offering therapeutic targets. Antidepressants and psychotherapy have both been shown to alter DNA methylation patterns at specific loci, though the functional significance of these changes is still being mapped.

Pharmacotherapy

The STAR*D study (Sequenced Treatment Alternatives to Relieve Depression), the largest prospective antidepressant trial ever conducted (N = 4,041), demonstrated the challenge of achieving remission in real-world depression treatment:

• Step 1 (citalopram monotherapy): Remission rate ≈ 33%; response rate ≈ 47%
• Step 2 (switch or augmentation): Cumulative remission ≈ 50%
• Step 3: Cumulative remission ≈ 55%
• Step 4: Cumulative remission ≈ 67%

Critically, those requiring more treatment steps had higher relapse rates: Step 1 remitters had a 12-month relapse rate of approximately 33%, versus approximately 71% for Step 4 remitters. This underscores the treatment-resistant depression burden — approximately 30% of patients with MDD meet criteria for TRD (failure of ≥2 adequate antidepressant trials).

The Cipriani et al. (2018) network meta-analysis of 522 RCTs (N = 116,477) comparing 21 antidepressants found that all active agents were superior to placebo. In terms of efficacy (response), agomelatine, amitriptyline, escitalopram, mirtazapine, paroxetine, venlafaxine, and vortioxetine ranked highest. Fluoxetine and fluvoxamine ranked lowest among active agents. In terms of acceptability (dropout), agomelatine, citalopram, escitalopram, fluoxetine, sertraline, and vortioxetine were best tolerated. The overall pooled OR for response versus placebo was approximately 1.66 (95% CI: 1.58–1.75).

Psychotherapy

Cognitive behavioral therapy (CBT) has the largest evidence base, with meta-analytic effect sizes of d ≈ 0.71 versus control conditions and NNT of approximately 4–5 for response. Behavioral activation (BA), interpersonal therapy (IPT), and psychodynamic psychotherapy also demonstrate efficacy. The DeRubeis et al. (2005) RCT found CBT equivalent to paroxetine for moderate-to-severe depression, and the Hollon et al. (2005) follow-up showed that CBT conferred a significant prophylactic advantage: relapse rates over 12 months were approximately 31% for prior CBT versus 76% for discontinued medication versus 47% for continued medication.

Neuromodulation

ECT remains the most effective acute treatment for severe depression, with response rates of 70–90% and remission rates of 50–65% in adequately powered trials. The NNT for ECT versus sham or pharmacotherapy comparators is approximately 3–4. rTMS targeting the left dlPFC shows response rates of approximately 50–55% and remission rates of approximately 30–35% in TRD populations, with an NNT of approximately 5–8 versus sham.

Prognostic Factors

Consistent predictors of poorer treatment outcomes include:

• Longer duration of current episode (>2 years reduces response probability by approximately 40–50%)
• Greater number of prior episodes
• Comorbid anxiety disorders (present in approximately 50–60% of MDD patients)
• Comorbid personality disorders (particularly borderline PD, present in approximately 10–30%)
• Childhood adversity and early onset (<18 years)
• Persistent insomnia
• Elevated inflammatory markers (CRP >3 mg/L)
• Suicidal ideation at baseline

Predictors of better outcomes include shorter illness duration, later onset, absence of comorbid personality pathology, higher baseline functioning, strong therapeutic alliance (in psychotherapy), and early response (≥20% improvement by week 2 predicts eventual response with ~70% accuracy).

MDD rarely occurs in isolation. Comorbidity is the rule, not the exception, and profoundly affects diagnosis, treatment selection, and prognosis:

• Anxiety disorders: 50–60% of MDD patients have a comorbid anxiety disorder (generalized anxiety disorder, social anxiety disorder, panic disorder, or PTSD). The anxious depression phenotype is associated with greater functional impairment, higher suicide risk, and poorer response to antidepressant monotherapy. The DSM-5-TR "anxious distress" specifier captures this clinically important subgroup.
• Substance use disorders (SUDs): Approximately 20–30% of individuals with MDD have a comorbid SUD. Alcohol use disorder is the most common, with odds ratios of approximately 2.0–3.0 for co-occurrence. Shared neurobiological substrates (reward circuit dysfunction, HPA axis dysregulation) likely mediate the association.
• Chronic pain conditions: 30–50% of chronic pain patients meet criteria for MDD, and the conditions share overlapping neurobiology (descending serotonergic and noradrenergic pain-modulating pathways, neuroinflammation). Duloxetine and amitriptyline have demonstrated dual efficacy for both conditions.
• Cardiovascular disease: Post-myocardial infarction depression prevalence is approximately 20%. Depression is an independent risk factor for cardiovascular mortality (RR ≈ 1.6–2.0), mediated in part by HPA axis-driven sympathetic activation, platelet hyperreactivity, and inflammatory mechanisms.
• Type 2 diabetes: Bidirectional relationship with MDD; each condition approximately doubles the risk of the other. Shared mechanisms include insulin resistance, inflammation, and HPA axis dysfunction.
• Personality disorders: Comorbid personality pathology — present in approximately 40–50% of patients with chronic or recurrent MDD — substantially reduces treatment response rates and increases relapse risk.