FKBP5 is the gene; FKBP51 is the protein it encodes. The rs1360780 T allele is a variant of FKBP5 that produces more FKBP51 per cortisol pulse. Throughout this paper, the gene name is used when discussing genotype, transcription, and epigenetic regulation; the protein name is used when discussing what actually does the work — binding Hsp90, blocking GR translocation, stripping AMPA receptors.
The rest of this section is established science. Readers already familiar with FKBP51 feedback dynamics, FKBP51-Hsp90-AMPA receptor interactions, and cortisol's downstream neurotransmitter effects can skip to Part 2.
FKBP51 is a co-chaperone that sits in steroid receptor complexes. Its best-characterized target is the glucocorticoid receptor (GR). When cortisol binds GR in the cytoplasm, GR must translocate to the nucleus to activate gene transcription. Translocation requires the Hsp90 chaperone complex, and specifically which immunophilin is bound to Hsp90 determines whether it happens: FKBP52 allows translocation via dynein motor transport along microtubules, while FKBP51 blocks it. FKBP51 and FKBP52 compete for the same binding site. More FKBP51 means less GR reaching the nucleus, which means less cortisol-driven gene transcription — effectively reducing the cell's sensitivity to cortisol.
The rs1360780 T allele amplifies this. The T variant creates a TATA box adjacent to a glucocorticoid response element (GRE) in intron 2 of the FKBP5 gene, enhancing transcription. Each cortisol signal therefore produces more FKBP51 protein in T-allele carriers than in C-allele carriers, and more FKBP51 blocks more of the next cortisol signal. The result is an ultra-short intracellular negative feedback loop: cortisol activates GR → GR turns on FKBP5 → more FKBP51 produced → blocks next cortisol signal.
The standard psychiatric model calls the T-allele phenotype "GR resistance" and frames it as pathological: T/T carriers show impaired cortisol suppression on the dexamethasone suppression test, prolonged cortisol surges after psychosocial stress, and increased risk for depression and PTSD following childhood trauma (Binder et al., 2008; Klengel et al., 2013). This paper proposes that the same mechanism is better understood as a calibration system that adjusts both stress responsiveness and excitatory neurotransmission to match sustained environmental conditions — with the caveat that calibration designed for sustained input can produce pathology when the environment shifts faster than the system adjusts.
FKBP51 modulates several other steroid receptors through the same FKBP51/FKBP52 competition mechanism: the mineralocorticoid receptor (MR), progesterone receptor (PR), and androgen receptor (AR) (Jääskeläinen et al., 2011; Stechschulte & Sanchez, 2011). FKBP51 attenuates MR and PR activity (same direction as GR) but enhances AR signaling (opposite direction), with the effect on MR weaker than on GR (Baischew et al., 2023). This paper focuses primarily on GR because the rs1360780 variant's effects are best documented there, but MR modulation has consequences for the tonic signaling floor discussed in Part 2. Because MR has roughly tenfold higher cortisol affinity than GR and drives FKBP5 transcription in hippocampal neurons at baseline cortisol activity (Häusl et al., 2021), the FKBP51 feedback loop runs tonically — at resting cortisol levels — not only during stress-driven GR activation. In T-allele carriers, this baseline MR-driven transcription produces more FKBP51 (because of the enhanced TATA box), which then attenuates both MR and GR before any stressor arrives.
FKBP5 is expressed in every brain region this paper discusses. LacZ staining in knockout mice confirmed expression in the hippocampus, striatum, dorsal raphe, and locus coeruleus (Huang et al., 2016), while Fkbp5 mRNA has been detected in the nucleus accumbens, ventral tegmental area, amygdala, BNST, and PVN (Scharf et al., 2011; Brivio et al., 2016).
FKBP51 affects the brain through two distinct mechanisms that operate independently. The first is its effect on cortisol signaling through GR, described above, which has large downstream consequences when it fails. The second is a direct effect on glutamate receptors at the synapse, unrelated to cortisol.
Pathway 1: Cortisol signaling and the GSK3β cascade. When FKBP51 blocks GR translocation, it dampens both the slow genomic arm of GR signaling (nuclear translocation → gene transcription, minutes to hours) and the rapid non-genomic arm (cytoplasmic, seconds to minutes). The non-genomic arm matters because it is how GR suppresses glycogen synthase kinase 3 beta (GSK3β): cortisol-bound GR releases c-Src kinase, which activates PI3K and Akt, and Akt phosphorylates GSK3β at Ser9, inhibiting it.
GSK3β is constitutively active, always on unless something is actively suppressing it. When unopposed, it promotes neuronal apoptosis, degrades synaptic proteins, impairs BDNF signaling, disrupts circadian rhythms, promotes neuroinflammation, drives dendritic spine retraction, and commands the kinesin motor system to remove AMPA receptors from synaptic surfaces (Du et al., 2010), producing long-term depression of synaptic transmission (Peineau et al., 2007).
The cell must continuously suppress GSK3β to maintain normal synaptic architecture, and one of the primary endogenous pathways for doing so runs through cortisol: cortisol → GR → PI3K/Akt → GSK3β phosphorylation at Ser9 → inhibition. FKBP51 attenuates this pathway. Lithium inhibits GSK3β directly through competition with magnesium at the enzyme's catalytic site (Ryves & Harwood, 2001), bypassing the cortisol-GR pathway entirely.
Pathway 2: Direct AMPA receptor stripping. Independently of its effects on cortisol, FKBP51 modulates AMPA receptor trafficking at the synapse through its interaction with Hsp90. Hsp90 is required for constitutive cycling of AMPA receptors between synaptic and nonsynaptic sites — the continuous maintenance trafficking that keeps glutamate receptors on the synaptic surface (Gerges et al., 2004). This cycling is mediated by TPR (tetratricopeptide repeat) domain-containing proteins binding Hsp90; FKBP51 contains a TPR domain and is a known Hsp90 co-chaperone. Bhatt et al. (2019) demonstrated that FKBP51 overexpression in the corticolimbic system increased the association of Hsp90 with GluR1-type AMPA receptors, accelerating their recycling off the synaptic surface. More FKBP51 means fewer AMPA receptors sitting on the synapse at any given moment, reducing glutamate signaling even when glutamate release is adequate. The behavioral consequence was impaired spatial reversal learning and disrupted long-term depression. The complementary finding from FKBP5 knockout mice confirms this: deletion of FKBP5 reduced expression of excitatory glutamate receptors (NMDAR1, NMDAR2B, and AMPAR) and reduced excitatory synaptic activity, while increasing GABA expression and inhibitory synaptic activity (Qiu et al., 2019).
This AMPA stripping is constitutive — it operates whenever FKBP51 protein is present, not in response to a specific signaling event. Because cortisol drives FKBP5 transcription, FKBP51 protein levels track cortisol exposure history: sustained stress produces sustained FKBP51 elevation, which produces sustained AMPA receptor depletion. FKBP51 turns over rapidly: isoform 1 has a half-life of approximately 8 hours, isoform 2 approximately 4 hours (Martinelli et al., 2024), which means FKBP51-mediated effects respond to changes in transcription rate within days, not weeks — with significant implications for recovery timelines after stress removal (see Part 3).
Additional evidence supports the convergence of these two pathways on glutamate signaling. In the prefrontal cortex, FKBP5 deletion abolished GR-induced reductions in excitatory transmission, indicating that FKBP5 is required for GR to modulate glutamate signaling in this region (Ryu et al., 2021). In human postmortem tissue from over 1,000 subjects, FKBP5 expression levels strongly and inversely correlated with dendritic mushroom spine density and BDNF levels in superficial layer neurons, across schizophrenia, major depression, and bipolar disorder (Sinclair et al., 2023).
A third point of convergence: FKBP51 directly inhibits GSK3β. The two pathways above converge through a direct interaction: FKBP51 itself binds and inhibits GSK3β (Gassen et al., 2015). During stress, this provides a third neuroprotective effect beyond GR blockade and AMPA stripping — elevated FKBP51 partially suppresses GSK3β directly. The implication for recovery is critical: when FKBP51 clears after stress removal (within days, given its short half-life), the GSK3β suppression it was providing disappears along with it. The post-stress cell loses three GSK3β brakes simultaneously — reduced cortisol (less GR→Akt→GSK3β suppression), reduced FKBP51 (less direct inhibition), and methylated NR3C1 (less GR expression to restart the Akt pathway). This triple withdrawal may explain why GSK3β becomes so severely unopposed during recovery, and why the system can tip into the bistable trap described in Part 3.
Glutamate is the brain's primary excitatory neurotransmitter and the excitatory signal required for virtually all neurotransmitter-producing neurons to fire. Glutamatergic inputs drive dopamine neurons in the VTA (Butts & Phillips, 2013), serotonin neurons in the dorsal raphe (Commons, 2015), and norepinephrine neurons in the locus coeruleus. GABA, the brain's primary inhibitory neurotransmitter, is synthesized directly from glutamate by glutamic acid decarboxylase. Endocannabinoid release is triggered by activation of postsynaptic metabotropic glutamate receptors. Glutamate is not one of several downstream systems — it is the layer between cortisol and everything else.
Cortisol regulates glutamate release directly: acute stress increases glutamate release in the prefrontal cortex through multiple glucocorticoid-dependent mechanisms (Popoli et al., 2012), and removal of cortisol through adrenalectomy reduces evoked glutamate release from hippocampal nerve terminals (Wang & Wang, 2009). GSK3β activation suppresses presynaptic glutamate release (Zhu et al., 2007); lithium's inhibition of GSK3β therefore increases glutamate release independently of cortisol. Chronic low glutamate produces anhedonia and cognitive dysfunction (Popoli et al., 2012). In adults with ADHD, lower striatal glutamate was associated with worse attentional impairment, and dopaminergic stimulants compensated for but did not reverse the underlying glutamatergic deficit (Maltezos et al., 2014). Dopamine hypofunction in ADHD has been proposed to result from impaired glutamate-stimulated dopamine release in the nucleus accumbens (Russell, 2003).
Cortisol also regulates some systems independently of glutamate. GR in the locus coeruleus directly modulates noradrenergic output: when GR is underactivated, norepinephrine release is disinhibited (Wang et al., 2015). PNMT, the enzyme that converts norepinephrine to epinephrine, is directly GR-dependent. Cortisol induces expression of TPH2, the rate-limiting enzyme for serotonin synthesis (Chen et al., 2006). Cortisol drives BDNF production through GR-mediated gene transcription, and BDNF in turn primes GR function (Jeanneteau et al., 2015). BDNF is downstream of both cortisol and glutamate, and its depletion degrades the synaptic architecture that all of these systems depend on.