Rotenone-induced Parkinsonism in mice reverses within a year

Parkinson’s disease is the second most common neurodegenerative disorder in the world, yet every approved therapy, from levodopa to deep brain stimulation, only manages symptoms. None of them stop the relentless death of dopaminergic neurons in the substantia nigra pars compacta, the midbrain structure whose gradual failure drains dopamine from the striatum and robs patients of smooth movement. For decades, researchers have pointed to mitochondria as a central culprit: impaired complex I activity, surging oxidative stress, and bioenergetic collapse appear consistently in animal models and in postmortem human brains. Epidemiological studies have repeatedly linked exposure to pesticides containing mitochondrial toxins, most notably rotenone and paraquat, to elevated Parkinson’s risk, but a definitive causal relationship has remained stubbornly out of reach, because the longitudinal studies needed to establish causation in humans are simply not feasible.

Rotenone, a potent and widely used complex I inhibitor, has long been the workhorse for testing the mitochondrial hypothesis in animals. The landmark rat studies showed that chronic rotenone infusion over one to five weeks could selectively kill substantia nigra neurons, seed α-synuclein aggregates resembling Lewy bodies, and produce hypokinesia. Unlike faster toxins such as 6-hydroxydopamine or MPTP, which rely on the dopamine transporter to wipe out dopaminergic neurons within hours or days and effectively model only late-stage disease, rotenone exposure unfolds gradually, mimicking the insidious onset of human Parkinson’s and reproducing mitochondrial impairment, nigrostriatal dopamine loss, protein aggregation, and neuroinflammation. Rat models have been refined for decades, but mouse versions, which would allow powerful genetic manipulation, have produced inconsistent results, and nearly all studies in both species have focused on short-term outcomes, leaving the long-term consequences of exposure essentially unexplored.

Now a team has done what no one had done before in mice: a full 12-month longitudinal tracking of Parkinsonism after chronic systemic rotenone exposure, from the first signs of motor decline through an entire year of recovery. The researchers dissolved rotenone in DMSO, diluted it in the medium-chain triglyceride Miglyol 812, and injected it intraperitoneally at 2.5 milligrams per kilogram of body weight per day for 21 days into age-matched male mice, with vehicle-treated controls run in parallel. The dose was high enough to drive robust toxicity yet gentle enough that nearly all animals survived, a striking contrast to the 85 percent mortality previously reported within 56 days of oral rotenone exposure in mice.

During the acute phase, the fourth week of treatment, the rotenone-treated mice looked unmistakably parkinsonian. In the open field test, they traveled dramatically shorter total distances, moved more slowly, spent less time mobile, and froze more often than controls. Their center-zone distance collapsed selectively, a classic signature of anxiety-like behavior, which the elevated plus maze confirmed: treated mice entered the open arms less often and spent less time there. This non-motor deficit matters, because anxiety frequently precedes motor dysfunction in human patients. The rotarod test revealed severely impaired motor balance and coordination, with treated mice falling far sooner than controls, and while control animals improved their performance over three weeks of training, the rotenone group actually got worse, pointing to disrupted motor learning. The Y-maze showed spatial working memory remained intact, mirroring the selective profile of early Parkinson’s.

The cellular pathology in the midbrain was equally dramatic. Tyrosine hydroxylase, the rate-limiting enzyme of dopamine synthesis, showed a strong reduction in fluorescence across the whole substantia nigra pars compacta, with the most severe loss in the middle ventral region, a subregional selectivity that recapitulates the pattern of degeneration seen in human Parkinson’s disease. TH-positive cell number and density dropped by roughly 40 percent. Yet co-immunostaining for NeuN, a general neuronal marker, revealed that total neuronal density was completely preserved; only the TH-positive fraction within the neuronal population had fallen. This dissociation suggests the acute lesion is largely a phenotypic impairment, a stress-induced silencing of dopamine identity, rather than outright cell death. Intriguingly, striatal TH fluorescence was unchanged even in the dorsal lateral striatum, where high-resolution imaging confirmed intact TH-positive axons and varicosities, implying compensatory sprouting or multiple sources of striatal innervation rather than the axon loss seen in rat models.

Then came the surprise that transforms this study from a model characterization into a story about biological resilience. Twelve months after the toxin was withdrawn, the rotenone-treated mice were behaviorally indistinguishable from controls. Open field locomotion, center-zone exploration, elevated plus maze performance, and rotarod latency had all normalized. The longitudinal analysis showed locomotion fully recovered within about four months, and rotarod performance converged with controls by seven months. Both groups declined with age, but the acute divergence between them vanished entirely. The Y-maze, already intact acutely, remained so. For a disease defined by irreversible neurodegeneration, the complete reversal of Parkinsonism in a mammalian brain is a genuinely unexpected result.

The cellular story at 12 months is subtle and revealing. TH fluorescence in the substantia nigra pars compacta of treated mice remained lower than in age-matched controls, but TH-positive cell number and density had returned to control levels, and the TH-positive to NeuN-positive ratio was indistinguishable between groups. Even more striking, TH-positive cell density in the control group itself fell by roughly 45 percent over the year, consistent with the age-dependent decline documented in rats and humans, while this decline was nearly abolished in the rotenone-treated group. The treated dopamine neurons appear to have been stabilized, protected from the attrition that normally accompanies aging, even as their dopamine-producing machinery continued to run below normal. Whether chronic rotenone stress reprogrammed certain dopaminergic neurons toward regenerative capacity or altered their cellular identity remains an open question.

Neuroinflammation, long considered a driving force in Parkinson’s pathology, was conspicuously absent. Confocal imaging of Iba1, a microglial marker, and GFAP, an astrocytic marker, showed normal microglial morphology and no significant astrocytic activation in the substantia nigra or striatum at either the acute or the late stage. P2RY12, a homeostatic microglial receptor lost upon microglial dystrophy, was preserved in the substantia nigra pars compacta, with only a partial reduction in the pars reticulata at 12 months. Phosphorylated α-synuclein at Ser129, the post-translational modification found in most Lewy body deposits, showed no significant accumulation in either region at any time point, although a non-significant increasing trend appeared in three of five treated mice at 12 months, a hint that larger samples might eventually reveal delayed aggregation. These findings contrast sharply with rat rotenone models and with lipopolysaccharide or α-synuclein fibril injection models, underscoring how context dependent neuroinflammation and synucleinopathy are across species, routes, doses, and subregions.

The implications cut both ways. On one hand, the progressive motor deficits and TH-positive neuron loss during the acute phase support a critical role for mitochondrial dysfunction in Parkinson’s pathogenesis, aligning with the epidemiological correlation between pesticide exposure and disease risk. On the other hand, the complete recovery after rotenone withdrawal challenges the long-held assumption that pesticide exposure history is a direct cause of Parkinson’s, since even chronic, systemic, high-dose rotenone produced only transient deficits in this model, consistent with the reversibility of rotenone binding to complex I. The authors are careful about limitations: this is a toxin-based model designed to probe the resilience of the nigrostriatal system, not a genetic recapitulation of progressive human disease, and it lacks the α-synuclein-centered pathology and lysosomal mechanisms characteristic of human Parkinson’s. Systemic rotenone may also induce muscle weakness and malaise that contribute to behavioral scores, and mice differ from humans in dopamine metabolism, neuromelanin, α-synuclein and LRRK2 properties, and glial gene profiles. Still, the demonstration that the adult dopaminergic system can absorb a serious mitochondrial insult, silence its dopamine identity, and then restore both function and cell density over a year opens a tantalizing therapeutic question: if the brain can do this on its own after toxin withdrawal, can we learn to trigger that same resilience in patients whose disease is still progressing?

Subject of Research: Longitudinal rotenone mouse model shows reversible Parkinsonism and dopaminergic resilience

Article Title: A longitudinal study of rotenone-induced Parkinsonism in mice

Article References: Joshi, P., Fan, F., & Lou, X. (2026). A longitudinal study of rotenone-induced Parkinsonism in mice. iScience, 29(11), Article 117996. https://doi.org/10.1016/j.isci.2026.117996

Image Credits: AI Generated

DOI: 10.1016/j.isci.2026.117996

Keywords: Parkinson's disease, rotenone, dopamine neurons, mitochondrial dysfunction, substantia nigra, neurodegeneration, animal models, pesticide exposure, longitudinal, rotenone-induced, Parkinsonism, mice