Dual PPAR delta/gamma agonists offer therapeutic potential for Alzheimer’s disease

The increasing incidence of Alzheimer’s disease (AD) in the aging population, indicates the critical need for the development of novel targeted molecular therapies for ameliorating AD pathology. Moreover, clinical and pre-clinical evidence demonstrates that peroxisomal proliferator activating receptors (PPAR) agonists regulate energy and lipid homeostasis in models of diabetes, as well as improve spatial memory in animal models of AD (Khan et al., 2019). Mechanistically, PPARs are nuclear transcription factors that form heterodimeric complexes with retinoid X receptors. PPARs are key mediators responsible for the activation of genes involved in cell metabolism, differentiation, and development. More specifically they regulate the transcription of genes associated with energy homeostasis e.g., glucose metabolism, lipid transport, insulin sensitivity, mitochondrial biogenesis, and thermogenesis. They exists in three highly conserved isoforms, gamma (γ), delta (δ/β), and alpha (α). These highly conserved isoforms are well known for their clinical importance for example; PPARα agonists include the fibrates class of drugs (hypercholesterolemia) and PPARα consist of the thiazolidinedione class of drugs for type 2 diabetes (Zhang et al., 2020). Considering these findings, extensive investigation on PPAR agonists, showed reduced levels of amyloid plaques and tau hyperphosphorylation i.e., pathological hallmarks of AD via exhibiting anti-inflammatory properties, regulating ATP metabolism by reducing oxidative stress in mitochondria and improving symptoms associated with behavioral deficits and cognitive decline (Zhen et al., 2023).

Therefore, PPARs agonist has been identified as a well-known therapeutic interest for neurodegenerative disease (AD). However, it is noteworthy to mention that PPAR agonists either display poor blood-brain-barrier permeability and/or are associated with severe side effects on human health including edema and myocardial infarction (Zhang et al., 2020). Thus, limiting the clinical success and further development of novel PPAR agonists. Although there have been volume of reports, describing the beneficial attributes of PPAR agonists towards improving synaptic plasticity, behavioral deficits, and improved pathology, but there remains a considerable negative bias towards the clinical application of this class of drugs for AD and consequently, other neurodegenerative diseases. To overcome and meet these challenges, research efforts should focus on the development of a rational-based dual PPAR agonist that improves therapeutic efficacy, while limiting the toxic side effects of these compounds. In the manuscript by Steinke et al, the investigators used an in-silico approach to design and develop compound AU9, a novel PPAR δ/γ dual agonist (Steinke et al., 2023). AU9 was designed to avoid interactions with tyrosine-473 residue in the PPAR activation function 2 (AF2) ligand binding domain. Moreover, this site is required for full PPAR and therefore, the avoidance of TYR-473 interaction in the PPAR AF2 ligand binding domain was shown to elude the unwanted side effects observed with full PPARγ agonist. Further, this form of molecular design also demonstrated an improvement in behavioral deficits, synaptic plasticity, and reduction in AD-related pathologies and markers of neural inflammation in aged 3×TgAD mice.

The recent evolution of PPAR modulators has focused on a distinct balance of the isoforms, thus leading to the recent evolution of PPAR modulators that are either dual or pan agonists. This emerging role presents a unique opportunity for the rational-based development of compounds for neurodegenerative diseases, as all three isoforms in the PPAR family are abundantly expressed in the brain. Whereas, PPAR δ/β is the most abundant in expression in the brain, specifically in neurons and microglial cells (Warden et al., 2016). PPARγ is lower in expression when compared to PPARγ in these cells, and PPAR is expressed predominantly in astrocytes. However, in-spite of that PPARγ agonists are the most extensively investigated form of PPAR agonist for AD therapy. Further, both PPARδ and γ are restricted to the hippocampus, and are enriched in the dentate gyrus, CA1 region, and CA3 regions respectively. Consequently, activation of these receptors offers regulation of lipid metabolism, energy regulation, reduction in neuro-inflammation, and enhanced neurotrophin expression. Conversely, loss of activity causes transcriptional transrepression of proteins and intra-neuronal accumulation of AD-related pathologies which limits its potency (D’Angelo et al., 2018).

Further research is needed to understand the mechanisms by which PPAR agonist induce their effect on the clearance of plaques and macrophage/microglial activation associated with neuroprotective action in AD (Saez-Orellana et al., 2021). To limit the wide scope of PPARs, here in this perspective, we discuss the far-reaching role of dual PPAR and partial agonist and its predominant role in selective amino acid residue interactions in the AF2 ligand binding domain, which helps to nurture the synaptic plasticity and blood-brain-barrier permeability. These partial dual agonists enhance the therapeutic window by reducing full transcriptional activation while maintaining conformational stability of the receptors relative to PPARs full agonists (Capelli et al., 2016). Various molecular modeling approaches have been studied and found successful to predict drug likelihood e.g., ligand-receptor affinity, molecular stability, and biodistribution. These successes allow insights into end-point selection before the initiation of expensive drug trials. Current rational in-silico modeling is used to obtain the potential leads on crystal structures of PPAR and PPAR ligands within the ligand binding domain (Kroker and Bruning, 2015; More et al., 2017). PPARs have four functional domains, the N-terminal domain, the DNA binding domain, ligand binding domain, and the AF2 domain are highly conserved i.e. hydrogen bonding (predominantly carboxylic group). PPARγ partial agonist selectively interacts with key amino acid residues PHE-347, HIS-323, HIS-413, LYS-367, and ARG-288, while avoiding TYR-473; whereas, full PPARγ agonism requires the tyrosine-473-based hydrogen bonding for adiposity and other biological properties (Capelli et al., 2016; Gim et al., 2018).

Taking together, several clinically approved PPAR agonists display a negatively charged nitrogen group that forms hydrogen bonding interaction with TYR-473 side-chain hydroxyl group in helix 12. Moreover, it is evident that TYR-473 is required for the activity of full PPARγ agonism. However, avoidance of TYR-473 interaction results in a decrease in the ability to induce full receptor conformation and intrinsic binding potencies as demonstrated by compound AU9. This understanding offers a key pathway for designing partial PPARδ/γ agonist and thus overcoming the side effects associated with TYR-473 (Capelli et al., 2016; Figure 1). Work in the Amin lab, utilized in-silico modeling to design PPARδ/γ dual agonist and the designed compound differing in interactions with key amino acid residues in the ligand binding domain when compared to GW0742 (full PPARδ agonist) (Steinke et al., 2023). More specifically, GW0742 forms key interactions with TYR437 and forms interactions with 23 amino acids in the AF2 ligand binding domain. Conversely, AU9 avoids the key interaction with TYR437 and forms interactions with 26 amino acids in the AF2 ligand binding domain. Interestingly, the trifluoro side group holds the ring moiety in the cavity and thus prevents the interaction with TYR-437 in the PPARδ AF2 ligand binding domain. Furthermore, AU9 was designed to avoid specific interactions in the AF2 ligand binding domain while forming strong interactions with PHE282, LYS 367, and TYR-327 via hydrogen bonding and GLU 343, ILE341, and ILE281 via hydrophobic contacts to provide greater flexibility in the AF2 ligand binding pocket, stability to the protein complex and improve the potential clinical application. This makes AU9 a novel partial agonist. The strong interactions that form are due to hydrogen bonds, water bridge-mediated hydrogen bonds, ionic bonds, and hydrophobic bonds. Subsequently, this was further verified in the AD mouse model (3×TgAD), which suggested that AU9 induces PPARδ receptor activity levels less than full PPARδ agonist GW0742. Full PPARδ agonism in the brain has been shown to have anti-inflammatory effects. AU9 has also been observed to display significant anti-inflammatory effects on microglia and in the brain as demonstrated by ELISA and transcriptome analysis (Steinke et al., 2023).

Figure 1.

Dual AU9 peroxisomal proliferator activating receptors delta/gamma (PPARδ/γ) partial agonist promotes neuroprotective action in Alzheimer’s disease (AD).

(A) Neurodegeneration is observed in neurons from AD. (B) Generic PPARδ/γ linear structure of PPAR. The hypervariable blue box contains the activation function-1 (AF-1) domain. The peach box contains the DNA binding domain (DBD). The hinge region (green box) allows for conformational change following ligand binding to promote co-regulator (coactivator or corepressor) docking. The red and yellow box contains the ligand binding domain (LBD) of PPARδ/γ and the activation function-2 (AF-2) domain and is necessary for ligand-induced PPARδ/γ partial agonist transcriptional activity. Neuroprotective action exerted via PPARδ/γ partial agonist. (C) Molecular modeling: In silico modeling of PPARγ interactions between AU9 (partial agonist) and a full PPAR agonist, either thiazolidinedione (TZD) or GW0742 (full PPARδ agonist). AU9 lowest energy conformation and amino acid binding interactions with distances. (D) AU9 avoids a key interaction at TYR473 for PPARγ and TYR437 for PPARδ (AF-2 domain). The figure is based on data from (Steinke et al., 2023) using Schrodinger software suite for in silico modeling. (E) Diagram of the brain showing the various mechanisms by which AU9 offers neuroprotection for AD. PUFA: Polyunsaturated fatty acids. (F) Healthy neuron with dense spine formation is observed in neurons. Created with BioRender.com.

Selective activation of key amino acids in the ligand binding domain offers the potential for partial agonism. This design may offer a mechanism by which partial agonists may improve outcomes by inducing the receptor to favor different coactivators for interactions when compared to full agonists. Recent work by Hughes’s group helps illuminate the biased agonism of full agonists (Nemetchek et al., 2022). They explain the different mechanisms by which coactivators bind to the nuclear receptors, thus allowing different agonists to favor selective coactivator recruitment when compared to clinically available agonists. For example, they observed that the interaction between coactivators and a C-terminal residue on PPARγ helix 4 is highly significant. This residue makes the coactivator affinity with the receptor more dependent on the C-terminal helix of the receptor, helix 12. In contrast, full agonists that stabilize helix 12, such as thiazolidinediones, favor peptide that induces bond formation. Currently, the Amin lab is exploring how AU9 influences coactivator preference in cells. These findings will help explain ongoing single-cell genomic analysis comparing full PPAR agonist vs. AU9. These studies offer that the structural explanation of partial agonism for both PPARδ/γ will add to the current physiological findings and will provide a foundation for the rational development of PPAR agonists which favor select coactivators.

AU9 has been observed to improve cognitive deficits in (12-month-old) triple transgenic (3×TgAD) mice, which show improved discrimination index, recognition memory, and preference over new objects. One of the major losses in AD is loss of spine density, morphogenesis, and neurotrophin (brain-derived neurotrophic factor) which results in the loss of synaptic efficacy and stability in response to chronic stress. Thus, with AU9 and vehicle treatment studies, we observed improved neurotrophin levels, which may help explain the elevated spine density levels. Further gene expression profiling using Nanostring technology, determined improvement in markers for inflammation and stress.

The manuscript by Steinke et al. (2023) suggests that AU9 improves behavioural deficits and synaptic plasticity via indirect and direct mechanisms. Indirectly, AU9 was observed to reduce Aβ plaques, and the soluble form of Aβ, as well as reduced β secretase activity. In AD deposition of amyloid deposits leads to the formation of insoluble plaques in the neocortex and progresses to the hippocampus over a period of time. Thus, the cleaving efficiency of the Aβ peptides and time during the disease progression is critical for the treatment of AD. However, a direct connection is that the observed increased neurotrophin levels are responsible for the increased spine density observed in the CA1 and CA3 regions following AU9 treatment in 3×TgAD mice. Rather, these findings may be an accumulative result of increased neurotrophin levels, reduced Aβ levels, and as well as inflammation markers such as cytokines. Further work on steady-state levels as well as pharmacokinetic and pharmacodynamics evaluations are currently under investigation. We observed no toxicity in the liver nor in the heart following 3 months of drug treatment. However, for larger animals studies are under investigation. The novelty of the findings from the work concerning AU9 towards AD and the field of PPAR research by Steinke et al. (2023) offers a new perspective on the application of dual agonists for the treatment of neurodegenerative diseases. Further that selective amino acid interactions in the AF2 ligand binding domains of both PPAR-gamma and delta open a new window for potential therapeutic applications.

In conclusion, the innovative in silico design of a PPARδ/γ agonist, AU9, may offer new perspectives for this class of agonists for AD. The design and development of AU9 offer a unique perspective on PPAR drug development as it serves as a dual partial agonist while lacking many of the deleterious effects observed in traditional full PPAR agonists.