Mito.newsMito.news
← All articlesEditorial brief · abstract-levelScore 87/100Confidence high
biorxiv2026-09-08neurobiologyfusiongliablood-brain barrier

Low testosterone opens the nucleus-accumbens blood-brain barrier by remodeling astrocyte mitochondria

In outbred male rats, low testosterone is not just a mood correlate. It raises anxiety by loosening the nucleus accumbens blood-brain barrier, and the physical failure is astrocytic: fewer endfeet, disordered endfoot mitochondria, fewer mitochondria-endoplasmic-reticulum contacts, less mitofusin 2. Restore testosterone, or put Mfn2 back in accumbens astrocytes, and anxiety falls.

Mito.news · at a glance

Signal profile (abstract-level)

neurobiology · fusion · glia · blood-brain barrier

Score 87/100BIORXIVhigh confidenceneurobiology
87
Importance
50
Mito signal
95
Dysfunction
75
Evidence
15
Translational

Editorial signal profile from the abstract (importance score, mito keywords, dysfunction tags, evidence density, translational cues). Not a figure reproduced from the preprint PDF.

Finding. Low testosterone makes male rats more anxious because it opens a hole in the nucleus accumbens blood-brain barrier, and the hole is an astrocyte-mitochondria problem. de Castro Abrantes, Sandi and colleagues sort outbred males by the anxiety they already have. The high-anxiety animals leak tracer in the accumbens, not as a whole-brain smear. Astrocytic endfeet pull back. Their mitochondria lose order and lose contacts with endoplasmic reticulum. Mitofusin 2 falls. Take testosterone away from a calm rat and you can buy that phenotype. Put physiological testosterone back, including in old low-testosterone animals, and the barrier, the endfeet, Mfn2, and the behavior recover. Knock down androgen receptor in the accumbens and those responses blunt. Overexpress Mfn2 only in accumbens astrocytes and anxiety drops.

Why this paper matters

Men with low testosterone report anxiety. The usual story is receptor-in-neuron, mood circuit, done. This paper puts a vascular-glial mitochondrion in the middle of the nucleus accumbens, the region that scores salience and stress. Barrier leak is region-selective. The organelle anatomy is endfoot-specific. A fusion GTPase rescue in astrocytes is enough to move behavior. That is a different therapeutic silhouette than another SSRI or another testosterone gel aimed at muscle.

The aged-rat arm matters because it keeps the axis from being a young-castration curiosity. Natural low testosterone in older animals sits on the same mitochondrial-barrier program.

The causal stack

Natural anxiety already carries the NAc leak and the Mfn2-low endfoot. Endocrine push-pull (suppress, restore) moves anxiety, transcription, coverage, and permeability together. AR in the accumbens is required for the full ride. Mfn2 in astrocytes is sufficient to cut anxiety. Sufficiency plus necessity-ish (AR knockdown) is why the score is not a methods paper.

How to read the score

High eighties. Causal hormone, named nucleus, BBB, astrocyte mitochondria, Mfn2 rescue. Confidence is high in the rat. Human translation is a separate paper.

Caveats

Anxiety-like tests are not diagnoses. Regional AR knockdown may still hit neurons. Systemic androgens do more than glaze NAc endfeet. Do not start or stop testosterone therapy from this brief.

What to do with it

If you work on Mfn2, MERCs, or glial mitochondria, this is a behavior-level phenotype at the blood-brain barrier. If you work on male anxiety or hypogonadism, look at accumbens barrier biology before you stop at amygdala serotonin. Pull the endfoot mitochondrial images and the astrocyte Mfn2 overexpression curves.

Free HTML is above. Bots pay for JSON at /api/v1/papers/10-64898-2026-09-03-749088. Optional wallet tester: MetaMask ($0.005).

Source preprint

Low testosterone promotes anxiety through astrocytic mitochondrial remodelling at the nucleus accumbens blood-brain barrier

10.64898/2026.09.03.749088

de Castro Abrantes H, Depaauw-Holt L, Ulgen DH, Di Giulio C, Barbetti M, Gebara E, Hollis F, Solano JL, Schlotterose L, Menard C, Salman M, Astori S, Sandi C.

Related briefs