Alzheimer’s disease (AD) has long been considered an irreversible neurodegenerative disorder, with available treatments offering only limited symptomatic relief. However, a recent study published in Cell Reports Medicine presents compelling evidence that advanced AD pathology may be pharmacologically reversible—at least in animal models. Using two complementary mouse models of Alzheimer’s disease, researchers demonstrated that treatment with the neuroprotective compound P7C3-A20 restored brain NAD⁺ homeostasis, improved cognitive and motor functions, and reversed multiple hallmarks of late-stage pathology, including synaptic dysfunction, neuroinflammation, oxidative stress, and blood–brain barrier damage. Importantly, the study also identified consistent molecular signatures in human AD brain samples that align with the mechanisms observed in mice, strengthening the translational relevance of the findings. These results suggest that restoring metabolic resilience—rather than targeting a single pathological protein—may represent a promising therapeutic strategy for Alzheimer’s disease. While further research and clinical validation are required, this work challenges long-standing assumptions about the irreversibility of AD and opens new avenues for treatment development.
Alzheimer’s disease (AD) is the most prevalent neurodegenerative disorder worldwide and the leading cause of dementia. It acts like a silent eraser, gradually and relentlessly wiping away a person’s memory, cognition, and even personality.
What is even more frustrating is that since the disease was first identified more than a century ago, the medical community has largely believed its progression to be irreversible. Existing medications can only temporarily alleviate symptoms or target a single pathological protein (such as β-amyloid), but their overall effectiveness is limited and they cannot stop or reverse disease progression.

However, the latest study we are discussing today, recently published in Cell Reports Medicine, brings a potentially paradigm-shifting piece of hope. The research team not only succeeded in reversing late-stage Alzheimer’s disease symptoms in two different mouse models using a pharmacological intervention—essentially allowing the animals to “recover”—but also identified corresponding therapeutic targets in the brains of human AD patients.
Before diving deeper, here are the key highlights of this study, which we will explain in detail below:
-
Challenging a dogma: The study provides evidence for the first time that even late-stage Alzheimer’s disease may possess reversible pathological and functional features.
-
Identifying a core mechanism: The progression of the disease is directly linked to disruption in the homeostasis of a key cellular metabolite—nicotinamide adenine dinucleotide (NAD⁺).
-
Precision intervention: A compound called P7C3-A20 restores NAD⁺ balance in the brain without exceeding physiological levels, thereby rebuilding neuronal resilience and enabling functional recovery.
-
Bridging mouse and human biology: Beyond animal experiments, multi-omics analyses revealed similar molecular alterations in the brains of human Alzheimer’s patients, significantly strengthening the potential for clinical translation.

Why Focus on NAD⁺?
Traditional Alzheimer’s disease research has primarily focused on two pathological hallmarks: β-amyloid (Aβ) plaques and tau neurofibrillary tangles. However, therapies targeting these features have produced disappointing results.
The researchers in this study approached the problem from a different perspective. Why do some individuals—such as those classified as NDAN (Non-Demented with Alzheimer’s Neuropathology)—harbor large amounts of amyloid plaques yet maintain normal cognition? The answer may lie in greater brain resilience.

This resilience depends on cellular health and repair capacity, and NAD⁺ plays a central role in these processes. NAD⁺ is an essential coenzyme involved in cellular energy metabolism (including mitochondrial respiration) and several key repair pathways, such as DNA repair and stress resistance.
Previous studies have shown that NAD⁺ levels decline in the brains of AD models. However, directly supplementing NAD⁺ precursors could potentially carry cancer risks. Therefore, the core goal of this research became determining whether pharmacologically restoring physiological NAD⁺ homeostasis—rather than simply increasing NAD⁺ levels—could prevent and even reverse Alzheimer’s pathology and cognitive deficits.

Two Complementary Mouse Models to Mimic Human Disease
To systematically test this hypothesis, the researchers designed a rigorous multi-layered, cross-species strategy. They used two genetically engineered mouse models with complementary mechanisms.
5xFAD mice:
These mice model amyloid-driven Alzheimer’s disease. They rapidly develop large amounts of Aβ plaques and exhibit multiple pathological changes including abnormal tau phosphorylation, blood–brain barrier disruption, neuroinflammation, and clear deficits in learning and memory.
PS19 mice:
These mice model tau-driven Alzheimer’s disease. They overexpress a mutant human tau protein (P301S), forming neurofibrillary tangles similar to those found in human patients, leading to neuronal death and cognitive decline.
The therapeutic compound used in the study was P7C3-A20, a neuroprotective molecule previously shown to activate the rate-limiting enzyme in the NAD⁺ biosynthesis pathway. Importantly, this compound restores intracellular NAD⁺ homeostasis without pushing levels beyond physiological limits.

The experimental design included several key stages:
-
Preventive treatment: Administration beginning before symptom onset (at 2 months of age) in 5xFAD mice until early disease stage (6 months) to determine whether disease development could be prevented.
-
Reversal treatment: Administration starting at the “late-stage” phase (6 months), when mice already displayed extensive pathology and cognitive deficits, continuing until old age (12 months) to assess whether disease progression could be reversed.
-
Late-stage tau model validation: Short-term treatment beginning in severely ill PS19 mice at 11 months of age, close to end-of-life, to determine whether the effects were broadly applicable.
Throughout the experiments, researchers conducted comprehensive assessments, including behavioral tests (novel object recognition, Morris water maze, rotarod) to evaluate cognitive and motor function; biochemical analyses of NAD⁺ levels, Aβ, and tau; histological staining to examine plaques, neuronal survival, neuroinflammation, blood–brain barrier integrity, DNA damage, and oxidative stress; proteomic analyses of hippocampal protein changes; and cell-based experiments examining protection of human brain microvascular endothelial cells from oxidative stress.
NAD⁺ Homeostasis as a “Gauge” of Alzheimer’s Severity
The researchers discovered that both 5xFAD mice and human AD patients exhibited significantly reduced NAD⁺/NADH ratios, a key indicator of NAD⁺ homeostasis.
For example, compared with controls, the NAD⁺/NADH ratio in brain tissue from Alzheimer’s patients was approximately 30% lower. Moreover, the degree of reduction strongly correlated with the severity of tau pathology, oxidative damage, neuroinflammation, blood–brain barrier disruption, and neuronal loss.
More importantly, individuals with Alzheimer’s-type pathology but preserved cognition (NDAN) displayed normal expression levels of NAD⁺-synthesis enzymes, similar to healthy individuals, whereas AD patients showed clear dysregulation.

This finding strongly suggests that loss of NAD⁺ homeostasis may be a key driver of Alzheimer’s clinical manifestations, while maintaining this balance may represent a fundamental mechanism underlying brain resilience.
P7C3-A20 Prevented and Reversed Late-Stage Alzheimer’s Pathology
In late-stage 5xFAD mice already exhibiting severe pathology and cognitive deficits, treatment with P7C3-A20 produced striking reversal effects.
At the behavioral level, impairments in novel object recognition memory, spatial learning ability, and motor coordination were restored to levels comparable to those of healthy mice.
Further analyses revealed that these functional improvements corresponded with significant recovery in multiple pathological processes. Abnormal tau phosphorylation was suppressed. Transmission electron microscopy showed structural repair of previously damaged blood–brain barrier components, including astrocytic endfeet and perivascular spaces. As a result, immunoglobulin leakage into brain tissue was reduced. Markers of oxidative stress, DNA damage, and neuroinflammation were also substantially diminished.

Importantly, electrophysiological recordings from hippocampal slices confirmed that previously impaired long-term potentiation (LTP)—a cellular correlate of learning and memory—was fully restored, demonstrating functional reconstruction of neural circuits.
Another key discovery was that these broad therapeutic effects were not achieved by reducing production of amyloid precursor protein (APP) or Aβ peptides. Their levels remained unchanged. Instead, the treatment significantly reduced the burden of already-formed pathological amyloid plaques.
This strongly supports the idea that the core mechanism of P7C3-A20 is not simply removing pathogenic substrates, but rather restoring NAD⁺ homeostasis to enhance the brain’s intrinsic capacity for cellular cleanup, repair resilience, and metabolic health, thereby reversing downstream toxic cascades.
Notably, this therapeutic strategy also showed effectiveness in the tau-driven model. In late-stage PS19 mice close to death, just 30 days of treatment significantly improved cognitive performance while simultaneously restoring NAD⁺ homeostasis and reducing tau hyperphosphorylation and oxidative damage.
A Molecular Network Associated with Disease Reversal
To explore the molecular mechanisms underlying these large-scale functional recoveries, the researchers performed proteomic analyses of hippocampal tissue from treated 5xFAD mice and compared the results with a human Alzheimer’s brain proteomic database integrating 38 independent studies.

Through this cross-species comparison, they identified 46 proteins that showed consistent dysregulation in both human and mouse AD brains but returned toward normal levels following P7C3-A20 treatment.
These proteins formed a clear functional network primarily associated with:
-
Mitochondrial function and metabolism (e.g., MRPS5, AUH)
-
Protein quality control and autophagy (e.g., SQSTM1/p62, USP10)
-
RNA processing and translation (e.g., RPL7A)
-
Synaptic structure and signaling
Among them, 17 proteins also displayed corresponding mRNA changes in human AD brains, including known AD risk genes (such as PTK2B) and critical synaptic proteins (such as ANK3). This multi-layer molecular consistency greatly strengthens their reliability as potential biomarkers or therapeutic targets.
Conclusion
This study offers far more than the discovery of a promising compound—it proposes a fundamentally different therapeutic concept.
The findings suggest that even in advanced stages of Alzheimer’s disease, the brain retains remarkable potential for repair. The key lies in restoring metabolic resilience, particularly through the regulation of NAD⁺ homeostasis.
By precisely re-establishing this balance, the compound P7C3-A20 appears capable of rebooting the brain’s metabolic “power supply,” enabling widespread recovery from cognitive behavior to molecular networks in multiple animal models.
Perhaps most excitingly, the researchers identified corresponding therapeutic nodes in human Alzheimer’s brain tissue, bringing laboratory discoveries closer to clinical reality.
Of course, significant work remains ahead, including validation in more complex sporadic AD models and rigorous testing of the safety and efficacy of P7C3-A20 or related compounds in humans. Nevertheless, this research provides a powerful new source of optimism: in the future, we may not only slow the aging of the brain—but perhaps even turn back the clock.