Two Diabetes Pills, One Poisoned Brain: Metformin and Empagliflozin Team Up Against Parkinsonism in Rats
Egyptian researchers gave rats a mitochondrial poison called rotenone to produce a rapid, laboratory version of Parkinson’s disease, then asked whether two widely prescribed diabetes drugs could blunt the damage. Metformin and empagliflozin each helped on their own. Given together, starting a week before the poison and continuing throughout, the pair produced the largest recovery of movement, the fewest dying neurons in the substantia nigra, and the lowest accumulation of alpha-synuclein, the protein that clumps in Parkinson’s brains. The combination also restored signalling through PI3K/AKT, a survival pathway, and lifted expression of PPAR-gamma, a metabolic transcription factor. On the single measurement that most directly reflects dopamine nerve function, striatal dopamine content, the combination was not statistically better than metformin alone. This is a short, young-animal, prevention-design toxin study, and it does not test whether either drug helps an already damaged brain.
There is a persistent and slightly awkward observation in neurology: people with type 2 diabetes get Parkinson’s disease more often than people without it, and they tend to progress faster once they have it. Insulin resistance, chronic inflammation, oxidative stress and failing mitochondria show up in both conditions. That overlap raises an obvious question. If a drug fixes the metabolic problem, does it also protect the brain?
A team at Kafrelsheikh University in Egypt has just published one of the more direct attempts to answer that question for two specific drugs. Metformin has been the first-line diabetes medicine for decades and is already a favourite of the longevity field. Empagliflozin belongs to a newer class, the SGLT2 inhibitors, which make the kidneys dump glucose into the urine and which have turned out to protect the heart and kidneys in ways nobody predicted from their glucose effects.
The researchers used fifty male Wistar rats and rotenone, a pesticide that blocks complex I of the mitochondrial electron transport chain. Rotenone is the workhorse toxin for modelling Parkinson’s in rodents because it reliably kills dopamine neurons and triggers alpha-synuclein clumping, the two defining features of the human disease. Rats got either nothing, rotenone alone, or rotenone plus metformin, plus empagliflozin, or plus both. Crucially, drug treatment began a week before the poison, so this is a prevention experiment rather than a rescue experiment.
The rotenone did what rotenone does. Untreated poisoned rats travelled 1.4 metres in a five minute open field test, against 30.7 metres for healthy controls. They could hang from an inverted wire grid for 9 seconds instead of 156. Their substantia nigra showed roughly nineteen times more degenerating neurons than healthy tissue, and their striatal dopamine fell by 79 per cent.
Both drugs pulled the animals back from that. Metformin was consistently the stronger of the two, particularly for dopamine, where it restored 71 per cent of the lost content compared with empagliflozin’s 43 per cent. The combination did better still on almost everything: it recovered 59 per cent of the lost grip endurance where metformin alone recovered only 19 per cent, and it cut degenerating neuron counts by 86 per cent relative to poisoned animals.
The mechanistic story the authors tell is that the two drugs converge from different directions on the same protective machinery, metformin through AMPK and PI3K/AKT, empagliflozin through PPAR-gamma and antioxidant signalling. It is a reasonable story. It is also, in this paper, largely untested: the team measured PI3K and PPAR-gamma and nothing else along those chains, ran no blocking experiments, and took no mitochondrial measurements at all in a study built on a mitochondrial poison.
Actionable Insights
Nothing here should change what anyone takes tomorrow. The reasons are specific and worth understanding.
First, dose. The rat metformin dose converts to roughly 1,700 mg per day in a 70 kg human, squarely inside normal clinical use. The empagliflozin dose converts to about 113 mg per day, which is four to eleven times the maximum approved human dose of 10 to 25 mg. The empagliflozin arm of this study is not a test of the dose anyone actually takes.
Second, timing. Drugs were started seven days before the poison. Humans arrive at a Parkinson’s diagnosis having already lost 50 to 70 per cent of their dopamine neurons. A prevention result in rats says nothing about whether a drug helps after damage has occurred.
Third, magnitude. Effect size is just the gap between two groups measured in units of how much the animals naturally vary. A gap of 0.8 standard deviations is conventionally called large. The gaps here run from 3 to 37 standard deviations. Numbers that extreme are a signature of a toxin model that nearly ablates a system, not of a subtle intervention, and they do not translate to human dose-response.
The defensible takeaway is narrow: metformin outperformed empagliflozin on the most meaningful endpoint, and adding empagliflozin to metformin bought measurable extra benefit on motor and histological measures but not on dopamine itself.
Context and Source
-
Full title: Neuroprotective effects of empagliflozin and metformin combined therapy on rotenone induced Parkinsonism in Wistar rats via PI3K and PPARγ modulation
-
Institution: Kafrelsheikh University, Faculty of Medicine (Departments of Medical Biochemistry and Molecular Biology, and Medical Physiology) and Faculty of Pharmacy, Kafrelsheikh, Egypt
-
Country: Egypt
-
Journal: Scientific Reports (Nature Portfolio), 2026, volume 16, article 28103
-
Impact evaluation: The impact score of this journal is 4.9 (2-year Journal Impact Factor, 2025 JCR data year, as listed by Nature Portfolio), evaluated against a typical high-end range of 0 to 60+ for top general science, therefore this is a Medium impact journal.
