The long march towards testing the adenosine receptor agonism hypothesis for human cerebroprotection: The story of the A₁R/A₃R receptor agonist AST-004

Abstract

AST-004 is an adenosine A₁R/A₃R agonist and promising cerebroprotectant currently in Phase II clinical development. Clinical evaluations have been initiated in concussion, with trials in preparation for mild complicated traumatic brain injury (TBI) and acute ischemic stroke (AIS). AST-004 was discovered as the result of basic research on the P2Y₁R-activated homeostatic roles of astrocytes in cerebroprotection, including control of edema, excitotoxicity and oxidative stress. AST-004 was identified as the active nucleoside metabolite of the P2Y₁R agonist MRS2365. Subsequent research demonstrated that astrocyte mitochondrial ATP production could be activated by AST-004, resulting in significant efficacy in rodent models of permanent and transient occlusion AIS, porcine models of TBI and a non-human primate model of AIS as well as models of neurodegeneration and addiction. The purpose of this review is to highlight the research that led to the discovery of AST-004, summarize its widespread preclinical efficacy in a variety of disease models, and describe how this compound’s excellent safety and pharmaceutical profile is differentiated from prototypical A₁R and A₃R agonists formerly or currently in the clinic.

Cerebroprotection for conditions such as acute ischemic stroke (AIS) or hemorrhagic stroke, traumatic brain injury (TBI) and concussion is a major global unmet medical need. Current estimates indicate that stroke is the second leading cause of death globally, with approximately 12 million new strokes each year [1]. The global incidence of stroke has risen by 70% over the last three decades. The global economic impact of stroke is estimated to be 890 billion dollars annually with an estimated increase to 1.7 trillion dollars by 2050. The incidence of TBI in the United States alone is estimated to be 1.7 million new cases per year, with an additional 1.6-3.8 million concussions (mild TBI) that are not treated, resulting in medical costs of approximately 77 billion dollars annually [2]. The most recent reported statistics for global TBI incidence were 27 million new cases in 2019 [3], resulting in an estimated global economic impact of 400 billion dollars annually.

Despite these significant medical needs and their associated economic and social impacts, current stroke therapy is limited to pharmacological thrombolysis (approximately 5% of patients), mechanical endovascular thrombectomy (10-20% of patients) or palliative care (80 + % of patients) [4–6]. Therapy for TBI and concussion is mainly limited to palliative care, except for invasive interventions such as craniotomy to reduce intracranial pressure in severe TBI [2]. After a series of failed clinical trials of both stroke and TBI/concussion therapies in the 1990s and 2000s, most major pharmaceutical companies exited these therapeutic areas in the early 2000s to pursue more tractable indications. However, stroke, TBI and concussion constitute critical therapeutic gaps in global healthcare.

Foundational research and the creation of Astrocyte Pharmaceuticals

Astrocyte Pharmaceuticals was formed in 2014 based on extensive preclinical research at the University of Texas Health Sciences Center at San Antonio, San Antonio, Texas (now UT Health San Antonio) in the laboratory of our colleague James Lechleiter. That initial research arose from his interest in calcium (Ca²⁺) signaling and mitochondria in astrocytes. Robust Ca²⁺ signaling was well described from work by other investigators [7–9]. However, mitochondrial physiology in astrocytes was very understudied and frankly, relegated to a minor metabolic role relative to glycolysis [10, 11]. In fact, mitochondria were considered largely dormant under normal conditions [12]. Lechleiter’s group discovered the homeostatic role of astrocytes could be greatly enhanced via their mitochondria, especially during times of metabolic stress and injury [13]. Lechleiter’s group demonstrated astrocyte-specific mitochondrial ATP production could be activated, leading to energy-dependent reductions in the size of brain lesions in mouse photothrombosis and middle cerebral artery occlusion (MCAO) models of stroke [14–16]. This work originally stemmed from in vitro experiments where cultured mouse astrocytes were stimulated with purinergic G protein-coupled receptor (GPCR) agonists [13]. The experiments targeted the P2Y₁R receptor pathway, which strongly couples to α_q-subtype G-proteins, phospholipase Cβ activation, IP₃ production and ultimately, intracellular Ca²⁺ release. Earlier work by multiple groups had demonstrated that in many cell types, including neurons, intracellular Ca²⁺ release was sequestered by mitochondria and stimulated cell death (necrosis and apoptosis) [17, 18]. Surprisingly, Lechleiter’s group discovered P2Y₁R agonism stimulated cell survival [13]. They were able to map out the pathway leading from agonism of P2Y₁Rs to astrocyte-specific mitochondrial ATP production. Thus, in contrast to neurons, stimulation of astrocyte Ca²⁺ signaling was protective [13]. Astrocytes were acting in a manner similar to cardiac myocytes, where the release of intracellular Ca²⁺ and sequestration by mitochondria signals these organelles to generate more ATP [19–21]. Ablation of any of the steps in the purinergic Ca²⁺ signaling pathway cascade led to loss of efficacy not only in vitro, but in mouse models of stroke and TBI [13, 22, 23].

Our initial body of research focused on astrocytic P2Y₁R agonism as the potential cerebroprotective mechanism. Purinergic P2Y₁Rs are primarily expressed on astrocytes, affording a targeted in vivo therapeutic intervention. Studies utilizing MRS2365, a P2Y₁R agonist synthesized by the Ken Jacobson group at the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK), and 2-methylthioadenosine-5’-diphosphate, demonstrated significant efficacy in preclinical mouse models of permanent and transient MCAO, as well as mouse models of TBI and concussion. We found astrocytes to be highly attractive and novel therapeutic targets, and a distinct departure from the traditional neuron-centric view of the brain, which has led to repeated clinical failures in finding neuroprotective drugs. Astrocytes are the natural “caretakers” in the brain. Our serendipitous discovery that P2Y₁R agonism could enhance mitochondrial ATP production after injury or stress, uncovered a novel therapeutic pathway that could boost homeostatic mechanisms critical for neuronal and brain function, including protection against edema, glutamate excitotoxicity, oxidative stress, and preservation of blood–brain barrier integrity. Based on these findings and the collective cerebroprotective data of P2Y₁R agonists, Astrocyte Pharmaceuticals was established to advance MRS2365 as a lead therapeutic candidate for stroke and TBI.

Adventures in target and lead discovery and the discovery of AST-004 (aka MRS4322)

The initial studies at Astrocyte focused on characterizing the stability, pharmacokinetics and brain distribution of MRS2365. Theoretically, this compound had several potential impediments that could limit its utility in cerebroprotection. First, a double negatively charged compound such as MRS2365 should not be able to cross the blood–brain barrier. Second, a diphosphorylated compound could be prone to rapid dephosphorylation by ectonucleotidases. For these reasons, the first studies conducted at Astrocyte were in vitro mouse and human plasma and blood stability studies, and in vivo mouse pharmacokinetic studies including determination of brain concentrations of MRS2365 at doses demonstrating efficacy in various mouse models of stroke and TBI.

The initial plasma and blood stability studies were conducted in Ethylenediaminetetraacetic acid (EDTA)-treated matrices (a foreshadowing!) and demonstrated excellent stability of MRS2365 [24]. These results triggered pharmacokinetic studies in mice. Surprisingly, these in vivo pharmacokinetic studies demonstrated no circulating MRS2365 in plasma or brain tissue across a range of doses using a sensitive Liquid Chromatography–Tandem Mass Spectrometry (LC–MS/MS) bioanalytical assay. A close reevaluation of study protocols and literature revealed that the only substantive difference in sample handling between the in vitro and in vivo studies was the use of EDTA as an anticoagulant in the in vitro stability studies; EDTA is an effective chelator of divalent cations that are essential cofactors for ectonucleotidases [25]. Repetition of the in vitro stability studies with heparinized matrices revealed a very rapid (seconds) loss of MRS2365 in plasma and blood of mice and humans.

So, if MRS2365 was subject to rapid dephosphorylation resulting in no measurable in vivo drug concentrations, what was the source of its significant efficacy in mouse stroke and TBI models? Structural evaluation of potential MRS2365 metabolites led to metabolite scouting and reevaluations of mouse pharmacokinetic plasma and brain samples, focusing on the molecular weights associated with the monophosphorylated metabolite, and a completely dephosphorylated nucleoside metabolite. Whereas no monophosphorylated metabolite was detected in any sample, significant signals associated with a completely dephosphorylated nucleoside were detected in both plasma and brain samples. Was this dephosphorylated nucleoside the source of the mouse efficacy? An exhaustive search of the NIH PubChem database revealed a single close-in analog of the dephosphorylated metabolite, with reported biological activity as a receptor agonist of a purinergic P1 adenosine receptor, A₃R! Thus, these studies drove a major shift in focus at Astrocyte Pharmaceuticals from the purinergic P2 to P1 family, from P2Y₁R to A₃R agonism. Not only a shift in the company’s lead candidate, but a shift in the company’s target receptor, over the span of a few weeks! We named this dephosphorylated nucleoside AST-004 [24]. The next step was to synthesize enough of this new lead candidate to confirm its structure and evaluate its stand-alone efficacy and pharmaceutical properties.

During later discussions with Ken Jacobson at the NIDDK, we learned that his group had previously synthesized AST-004 as a structural intermediate. AST-004 had an official NIDDK name, MRS4322. The Jacobson group had archived a sample of the compound but there were no reports of its pharmacology or receptor interactions. Jacobson kindly provided a sample of AST-004 (MRS4322) that enabled definitive proof of the structure of the dephosphorylated nucleoside metabolite detected in mouse samples. Subsequent pharmacokinetic studies with MRS2365 and NIDDK-generated AST-004 demonstrated that MRS2365 was rapidly and quantitatively converted to AST-004 in the mouse. Thus, preclinical efficacy was likely due to AST-004. But what were its receptor targets?

Collaborations with Christa Müller’s group at the University of Bonn and Lauren May’s group at Monash University revealed that AST-004 was a low micromolar affinity agonist at both A₁R and A₃R with no appreciable affinity for A_2AR and A_2BR [24]. The compound had lower A₁R affinity relative to the endogenous ligand adenosine, but higher A₃R affinity than that reported for adenosine [26]. The increased A₃R affinity is a function of the (N)-methanocarba bicyclic substitute for the ribose ring, established by the Jacobson lab to conformationally favor the A₃R (Fig. 1). This discovery was consistent with literature indicating the cerebroprotective effects of A₁R and A₃R agonism. In addition, both brain A₁R and A₃R are primarily expressed on glial cells, in particular astrocytes and microglia. Thus, although our primary focus has been on astrocytes, there is a possibility of microglial involvement in the efficacy of AST-004. Extensive evaluations of potential off-target pharmacology, such as the Eurofins Bioprint and SafetyScan assay panels, revealed no other potential targets for AST-004 other than A₁R and A₃R.

Fig. 1.

Changes in target specificity by N-methanocarba bicyclic substitution of riboside, and ectonucleotidase-mediated dephosphorylation. N-methanocarba bicyclic substitution increases the P2YR specificity of MRS2365 (P2Y₁R) compared to 2-MeS-ADP (P2Y₁R/P2Y₁₂R/P2Y₁₃R). AST-004 (A₁R/A₃R agonist) is the dephosphorylated nucleoside metabolite of MRS2365 (P2Y₁R agonist). Likewise, dephosphorylation of 2-MeS-ADP results in the A₁R/A_2AR/A₃R agonist 2-MeS-adenosine [7]

Subsequent studies with AST-004 have consistently demonstrated significant efficacy in mouse and rat models of stroke and TBI, porcine models of TBI, non-human primate models of stroke as well as additional new indications such as spinal-cord injury, Parkinson’s and Alzheimer’s diseases and drug addiction (Table 1) [14, 22, 27–29]. Moreover, we went back and demonstrated that the cerebroprotective efficacy of MRS2365 in mouse stroke models could be blocked by pharmacological antagonism of the A₃R [14]. Our working hypothesis is that by targeting astrocyte function, AST-004 enhances a fundamental homeostatic process that is of central importance in many brain injuries and disorders.

Table 1.

Summary of preclinical efficacy of AST-004

Species	RoA	Model	Neuroprotective Effects
Mouse	IV	Closed skull CCI moderate TBI	Reduced GFAP by 23-54%[30]
Mouse	IV	Moderate or single vessel ischemia	Reduced brain lesions by 34-72%[7]
Mouse	Oral	hAPP (J20) Alzheimer’s	Reduced Aβ by 66%
Mouse	IV	MPTP-Induced Parkinson’s Model	Increased substantia nigra [dopamine] [31] by 80%a
Mouse	IP	Spinal Cord Secondary Injury	Improved hind limb movement by 50%b
Mouse	PO	Cocaine Addiction	Reduced cocaine self-administration by 60%
Rat	IV	tMCAO stroke	Reduced brain lesions by 50%[7]
Pig	IV	CCI severe TBI	Reduced brain edema by 28%[32]
Cynomolgus macaque	IV	tMCAO stroke	Reduced brain lesions by 45%[33]
Human (in vitro)	In vitro	Fetal astrocyte sheer stretch injury	Lowered cell death rates by fourfold[34]

RoA Route of Administration, tMCAO transient middle cerebral artery occlusion, Aβ beta amyloid

^aUnpublished Data

^bManuscript in Preparation

Agonism of central adenosine receptors: A hypothesis waiting patiently to be tested in humans

Agonism of adenosine receptors, particularly A₁R and A₃R, has long been hypothesized to have significant cerebroprotective effects in ischemic stroke and TBI [30, 33, 35–41]. Soon after the discovery of A₃R in the early 1990s, studies were conducted highlighting the potential efficacy of A₃R agonism in stroke. There is also a wealth of published data on the potential cerebroprotective effects of A₁R agonism. Thus, through serendipity, our company had “rediscovered” a drug candidate that targeted both adenosine receptors hypothesized to be important targets for cerebroprotection.

Although synergism has not been conclusively demonstrated between A₁R and A₃R agonism in the brain [33, 42], there is considerable evidence for synergism between these receptors in the protection of cardiomyocytes from hypoxic injury [43–45]. Dual A₁R/A₃R agonists have been considered promising candidates for treating ischemic heart disease but have never been clinically evaluated [46]. Although we have not yet specifically assessed AST-004 in models of cardioprotection, this is a promising future avenue of research [47, 48]. There is also evidence for enhanced efficacy of dual A₁R/A₃R agonists in murine models of nociceptive pain [49, 50]. The evidence for A₁R/A₃R synergism in models of cardioprotection and pain leads us to hypothesize that a similar synergism may exist in the brain leading to the significant efficacy of AST-004 we have observed to date.

Despite three decades of preclinical research demonstrating meaningful efficacy of A₁R and A₃R agonism in AIS and TBI, these mechanisms have never been evaluated in human clinical trials. This is likely due to a variety of reasons for each of these receptors. For the A₁R, candidate drugs advanced to human trials were high-affinity agonists of the receptor, administered at dose levels associated with significant cardiovascular side effects due to A₁R agonism [51]. For the A₃R, agonists advanced to clinical trials to date have physicochemical properties that could limit distribution into the brain central compartment at sufficient unbound drug concentrations to elicit a pharmacological effect. For example, high lipophilicity resulting in very high plasma and brain protein binding and P-glycoprotein mediated efflux at the blood brain barrier (BBB) [52]. Other agonists evaluated preclinically have never advanced to clinical trials, potentially due to the properties cited above.

In a receptor antagonist program, micromolar affinity would be considered a distinct liability, because high affinity as well as a long half-life is required to block a target’s active site from binding and activation by the endogenous ligand [53]. Importantly, this is not necessarily the case for receptor agonism. There are many examples in pharmacology of drug agonists eliciting EC₅₀ activity at very low percentages of receptor occupancy [54]; this has also been shown in previous studies with adenosine A₁ partial agonists [55]. Many GPCRs have huge receptor reserves leading to significant pharmacological effects at low receptor occupancies. Our research shows this is the also the case for AST-004. Estimated brain receptor occupancies associated with efficacy in early mouse models of AIS and TBI were in the low single digits, while estimated efficacious central receptor occupancy in a primate model of AIS ranged from 3-15% [29]. The relationship between efficacy and receptor occupancy in the primate stroke model was in the linear range across the assessed doses, leading to the possibility of even higher efficacy at higher administered doses.

Other attributes of AST-004 more than compensate for its “low” receptor affinity. Our research program focused on balancing target affinity with other key drug-like properties that are frequently as, or more, important than affinity. For example, its physicochemical properties result in very high unbound drug fractions in both plasma and brain compared to other adenosine agonists, meaning that there is substantial unbound drug available to interact with target receptors [52]. AST-004 is also a substrate for the Concentrative Nucleoside Transporter 3 (CNT3), but is not a substrate for P-glycoprotein, resulting in more than sufficient transport across the BBB. Moreover, it does not activate β-arrestin recruitment that can lead to adenosine receptor desensitization, as we have observed with other prototypical high-affinity A₃R receptor agonists [52]. AST-004 has very low potential for metabolic or pharmacodynamic drug-drug interactions (internal data). There is also no evidence of metabolic or pharmacodynamic interactions with plasminogen activators (alteplase, tenecteplase) that could be potentially coadministered to stroke patients [56]. AST-004 structurally contains no chemical moieties incorporated in some prototypical A₃R agonists that could potentially lead to toxic reactive metabolites, such as halogenated furans or thiophenes, and alkynes [57, 58]. This has been confirmed by its very high therapeutic index (TI) in preclinical studies, with TIs ranging from 27 (canines) to 400 (rodents). Overall, AST-004 has an excellent pharmaceutical profile making it amenable to a variety of routes of administration in both acute and chronic settings.

Status and next steps

AST-004 has been steadily advanced through preclinical development and clinical Phase I trials [59]. AST-004 is characterized by excellent safety margins in both rat and dog GLP toxicology studies. High oral bioavailability in preclinical species have led us to develop oral formulations for treatment of conscious patients, while intravenous administration will be the mainstay for more serious cases of stroke and TBI. Phase 1 human clinical trials in normal human volunteers demonstrated that at plasma and cerebral spinal fluid (CSF) concentrations higher than our preclinical efficacious target concentrations, the only notable observed side effect was potentially a mild headache, which was also observed in volunteers administered placebo. There were no cardiovascular side effects as might be expected for a high affinity A₁R agonist. AST-004’s safety profile is consistent with the side-effect profile of other high affinity A₃R agonists currently in clinical trials [60]. Potential backup candidates have equally promising characteristics coupled with higher A₁R and A₃R affinity.

In total, the AST-004 Phase I clinical studies demonstrated that both plasma and CSF concentrations associated with significant efficacy in preclinical studies can be safely achieved and exceeded in humans (EUDRA CT#s 2022-000819-32, 05/02/2022 and 2023-000028-12, 03/23/2023). This greatly increases our confidence that in current and planned Phase II studies, we will be able to evaluate the hypothesis that central A₁R/A₃R agonism has cerebroprotective efficacy in AIS, TBI and concussion.

Currently, AST-004 has begun Phase II clinical trials for concussion in Australian rules football players (STARFAST; Australian New Zealand Clinical Trails Registration ACTRN12625000657426, 06/20/2025) and is scheduled to initiate Phase II trials in mild complicated TBI as well as ischemic stroke in 2026 and 2027.

Final thoughts

Over 30 years ago, investigators conducted initial studies that suggested that agonism of either the adenosine A₁ or A₃ receptors could have significant cerebroprotective efficacy in the treatment of stroke and TBI. Yet, these compelling data have never been tested or translated to human patients. Despite a mounting body of research crying out for clinical validation, no good therapeutic candidates had been developed and advanced into clinical trials in these indications. With AST-004, our goal is to finally evaluate the efficacy of A₁R/A₃R agonism, leading to the maintenance of astrocyte ATP production and cerebroprotective mechanisms, in human stroke, TBI and concussion patients as well as new indications still in preclinical development. If a significant efficacious signal is observed, the result will be a revolutionary advance in the field of neurology as well as an exciting development in the field of purinergic pharmacology. The results of AST-004 in these key Phase II clinical trials are expected over the next 2—3 years.

Theodore E. Liston

received his Ph.D. in Pharmacology from Vanderbilt University School of Medicine with a focus on discovery and pharmacological characterization of novel arachidonic acid metabolites. He was a Vice-President in Research at Pfizer Inc, where he led pioneering research in early discovery high-throughput screening for Absorption, Distribution, Metabolism, and Excretion (ADME) and toxicology, computational ADME and the prediction of human pharmacokinetics from preclinical data, as well as leading the Pharmacokinetics, Dynamics and Metabolism unit at Pfizer’s largest research site. He is currently Vice-President and Head of Research at Astrocyte Pharmaceuticals Inc where he has led the discovery and characterization of AST-004 and its advancement into clinical trials.

Author contributions

TEL wrote and reviewed the main manuscript text, JDL wrote and reviewed the manuscript text, WSK and KAJ reviewed and offered revision suggestions for the manuscript text.

Funding

Astrocyte Pharmaceuticals has received support from U.S. government grants, including NIH/NINDS STTR Phase I (2015, Award No. R41NS094327), NIH/NINDS SBIR Phase II (2018, Award No. R44NS102398), MTEC/USAMRDC (2021, Award No. W81XWH-21-9-0001), and NIH/NINDS SBIR (2025, Award No. R44NS135820). KAJ thanks Intramural Research Program (ZIADK031117), National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK), National Institutes of Health (NIH). NIH author’s contribution is considered Works of the United States Government. The findings and conclusions presented in this paper are those of the author(s) and do not necessarily reflect the views of the NIH or the U.S. Department of Health and Human Services.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethical approval

All in vivo preclinical studies described in this article were conducted following review and approval by relevant IACUC committees. All clinical studies described were conducted in accordance with applicable regulatory requirements and were reviewed and approved by the relevant national regulatory authorities and independent ethics committees.

Competing interests

TEL and WSK are employees of Astrocyte Pharmaceuticals Inc. with multiple patents. JDL is a cofounder of Astrocyte Pharmaceuticals Inc. with multiple patents.