Background:Despite the large socioeconomic burden of alcohol use disorders（AUD）,therapeutic treatment options are limited. There is a need to characterise the underlying neurochemistry driving alcohol seeking to identify and evaluate novel targets. AUDs are characterised by a transition to compulsive seeking,which is hypothesized to involve a shift from ventral to dorsal striatum and a medial to lateral dorsal striatal shift in the transition from goaldirected to habitual alcohol-seeking behaviours. Muscarinic acetylcholine receptors（mAChRs） are a potential target for AUD;they are expressed within the mesocorticolimbic reward system,including dense expression in the dorsal striatum,where they modulate dopamine and glutamate release,which may regulate reward processing. Methods:To assess the role of mAChRs in AUD,we first conducted genome-wide RNA sequencing in the caudate/putamen of 10 human alcoholics and 10 healthy controls and concurrently examined muscarinic receptor expression in the corresponding regions in rat（dorsolateral and dorsomedial striatum） following chronic alcohol consumption/withdrawal using qPCR. Next we examined the role of select muscarinic and nicotinic receptor subtypes in alcohol consumption and seeking using selective allosteric modulators. Finally,we probed the role of select muscarinic and nicotinic receptor subtypes in the dorsal striatum in alcohol consumption and seeking. Results:In human alcoholics M4 receptor expression was significantly decreased in the putamen. In line with this,M4 receptor expression was decreased in the rat dorsolateral striatum. Further,administration of VU154（30 mg·kg-1,p.o.） reduced alcohol self-administration and cue-induced relapse,without effects on natural reward consumption or sedation. We also found that a centrally active and selective negative allosteric modulator（NAM） for the rat M5 muscarinic receptor（mAChR）,ML375,selectively decreases ethanol self-administration and attenuates cue-induced reinstatement of ethanol-seeking in iP rats. We further show that in iP rats with an extensive history of ethanol intake that intra-dorsolateral（DL）,but not intra-dorsomedial（DM）,striatal injections of ML375 reduced ethanol self-administration to a similar extent as the nicotinic acetylcholine receptor（nAChR） ligand varenicline,which can reduce the reinforcing effects of ethanol in humans with AUD. Conclusions:Collectively,our data show that specific mAChRs are potential novel target pharmacotherapies for the treatment of AUD. These data also implicate the DL striatum as a locus for the effects of cholinergic-acting drugs on ethanol-seeking in rats with a history of longterm ethanol use. Accordingly,we demonstrate in rats that selectively targeting mAChR’s can modulate both voluntary ethanol intake and cue-induced ethanol-seeking implicating mAChR’s as a potential novel target for pharmacotherapies aimed at treating alcohol use disorders

Abstract Background Cognitive deficits in individuals with schizophrenia present a significant barrier to maintaining interpersonal relationships, employment, and independent living. Moderate to severe impairments have been found in cognitive domains including working memory, attention, learning and executive function. Despite the clear impact on functional outcomes, there are currently no approved treatments for cognitive impairments associated with schizophrenia (CIAS). Extensive efforts to develop pro-cognitive medicines, both for CIAS and other neuropsychiatric disorders, have repeatedly led to clinical trials that do not reflect the efficacy predicted in preclinical models. Aims & Objectives The rodent touchscreen platform has begun to bridge the translational gap between preclinical and clinical assessments of cognitive function. However, to date there has been no pharmacological validation of whether these tests are stringent enough to be clinically predictive and can therefore improve translation of CIAS treatments. Method Here, we use an NMDA receptor antagonist mouse model (acute low-dose administration) that we have validated for cognitive deficits in working memory performance on the rodent trial-unique delayed nonmatch-to-location touchscreen task. We then measured the effect of 1) current standard of care (aripiprazole and olanzapine), 2) failed investigational new drugs (IND) for CIAS (atomoxetine and encenicline), and 3) a current IND for schizophrenia (xanomeline) on working memory and attentional impairments. Results For the first time, we demonstrate a translational method for cognitive assessment in mice that can distinguish drugs with clinical efficacy from those without. Discussion & Conclusion These results provide critical insight into the predictive validity of our preclinical assay and cognitive deficit model, and have significant implications for the development and translation of novel pro-cognitive medicines.

Abstract Background Schizophrenia manifests as a broad and diverse symptomatology that creates a heterogenous patient population that responds to standard of care medicines to varying extents. The three characterised symptom domains; positive, negative and cognitive, represent, but are not limited to: hallucinations and delusions, negative affect, and impairments in learning and memory, respectively. Frontline drugs only effectively address the positive symptoms in ~70% of patients, yet it is the cognitive impairments associated with schizophrenia (CIAS) that pose the greatest hurdle to improve societal integration and quality of life. These issues culminate in a critical unmet medical need. There have been a number of clinical candidates and mechanisms that have sought to address CIAS, and all have failed. These failures point to two probable influences: 1. insufficient insight into the mechanisms capable of driving a change in disease symptoms; 2. the lack of stringent preclinical models and assays that derive endpoints similar to those tested in patients. Whilst schizophrenia standard of care medicines and investigational agents engage a number of G protein-coupled receptors (GPCRs), there remains an array of CNS-enriched orphan GPCRs that represent new opportunities as drug targets (Lu et al., 2023). Objective Our approach employs a holistic workflow to target validation and drug discovery through applying techniques that accelerates the process whilst increasing fidelity. We have applied structure- enabled drug design, disease-relevant pharmacology and advanced rodent models to multiple orphan GPCRs to create a global understanding of the target ranging from ligand-receptor interactions to in- depth behavioural insights. Method We used cryogenic electron microscopy to generate molecular models of our target orphan GPCRs in complex with their cognate G proteins to gain molecular insights into ligand binding and G protein coupling – facilitating structure-based drug design. We applied multi-endpoint pharmacology in recombinant and primary native cells to provide granularity to ligand-mediated signalling sequalae. Finally, we overlayed these findings with a comprehensive behavioural and cognitive battery (including rodent cognition touchscreens) to provide a correlation between our ligand-receptor mechanisms and psychosis- and cognition-relevant in vivo behavioural outcomes. Results Application of our technical workflow has provided molecular insights to accelerate ligand optimisation and deliver ligands with distinct, signal-biased pharmacological profiles. These ligands demonstrated antipsychotic activity (reversal of hyperlocomotion) and distinct cognition-enhancing profiles in mice, across working memory and attention tasks. Conclusions Development of this unique workflow has allowed the deconvolving of mechanisms of ligand-receptor interactions and signalling that accelerated development of ligands that target orphan GPCRs and display antipsychotic and pro-cognitive effects in mice. References Lu Y. 2023. Molecular insights into orphan G protein-coupled receptors relevant to schizophrenia. Br J Pharmacol. doi: 10.1111/bph.16221

Protein language models (pLMs) pre-trained on vast protein sequence databases excel at various downstream tasks but often lack the structural knowledge essential for some biological applications. To address this, we introduce a method to enrich pLMs with structural knowledge by leveraging pre-trained protein graph neural networks (pGNNs). First, a latent-level contrastive learning task aligns residue representations from pLMs with those from pGNNs across multiple proteins, injecting inter-protein structural information. Additionally, a physical-level task integrates intra-protein information by training pLMs to predict structure tokens. Together, the proposed dual-task framework effectively incorporates both inter- and intra-protein structural knowledge into pLMs. Given the variability in the quality of protein structures in PDB, we further introduce a residue loss selection module that uses a small model trained on high-quality structures to select reliable yet challenging residue losses for the pLM to learn. Applying our structure alignment method as a simple, lightweight post-training step to the state-of-the-art ESM2 and AMPLIFY yields notable performance gains. These improvements are consistent across a wide range of tasks, including substantial gains in deep mutational scanning (DMS) fitness prediction and a 59

The Concise Guide to Pharmacology 2025/26 marks the seventh edition in this series of biennial publications in the British Journal of Pharmacology. Presented in landscape format, the guide provides a comparative overview of the pharmacology of drug target families. The concise nature of the Concise Guide refers to the style of presentation, being clear, accessible, and well-structured, rather than the scope of the content, which spans approximately 500 pages. The Concise Guide summarises the key pharmacological properties of around 1900 human drug targets, and nearly 7000 interactions, involving around 4400 ligands. While the content is a substantially condensed version of the more detailed information and links available at the www.guidetopharmacology.org website, the printed guide serves as a permanent, citable, point-in-time record, that remains stable despite ongoing updates to the online database. The full contents of this publication can be found at https://bpspubs.onlinelibrary.wiley.com/doi/10.1111/bph.70230. The Concise Guides provide expert-curated recommendations of 'Gold Standard' selective pharmacological tools, available either commercially or as donations, which enable the identification of individual drug targets or families of drug targets. While the Concise Guide offers a more streamlined overview, more comprehensive information, including detailed pharmacological profiles and links to multiple online databases, is available through the Guide to Pharmacology website. The 2025/26 edition of the Concise Guide is based on material current as of mid-2025, and supersedes all previous editions, including the 2023/24 Guide, and earlier Guides to Receptors and Channels. It is produced in close conjunction with the Nomenclature and Standards Committee of the International Union of Basic and Clinical Pharmacology (NC-IUPHAR), and as such provides official IUPHAR classification and nomenclature for human drug targets, where applicable. G protein-coupled receptors are one of the six major pharmacological targets into which the Guide is divided, with the others being: ion channels, nuclear hormone receptors, catalytic receptors, enzymes and transporters. Each section includes nomenclature guidance, concise summaries, information of the best available pharmacological tools, key references, and suggestions for further reading.

Background:The development of more effective treatments for schizophrenia targeting cognitive and negative symptoms has been limited, partly due to a disconnect between rodent models and human illness. Ketamine administration is widely used to model symptoms of schizophrenia in both humans and rodents. In mice, subchronic ketamine treatment reproduces key dopamine and glutamate dysfunction; however, it is unclear how this translates into behavioral changes reflecting positive, negative, and cognitive symptoms. Methods:In male and female mice treated with either subchronic ketamine or saline, we assessed spontaneous and amphetamine-induced locomotor activity to measure behaviors relevant to positive symptoms, and used a touchscreen-based progressive ratio task of motivation and the rodent continuous performance test of attention to capture specific negative and cognitive symptoms, respectively. To explore neuronal changes underlying the behavioral effects of subchronic ketamine treatment, we quantified expression of the immediate early gene product, c-Fos, in key corticostriatal regions using immunofluorescence. Results:We showed that spontaneous locomotor activity was unchanged in male and female subchronic ketamine-treated animals, and amphetamine-induced locomotor response was reduced. Subchronic ketamine treatment did not alter motivation in either male or female mice. In contrast, we identified a sex-specific effect of subchronic ketamine on attentional processing wherein female mice performed worse than control mice due to increased nonselective responding. Finally, we showed that subchronic ketamine treatment increased c-Fos expression in prefrontal cortical and striatal regions, consistent with a mechanism of widespread disinhibition of neuronal activity. Conclusions:Our results highlight that the subchronic ketamine mouse model reproduces a subset of behavioral symptoms that are relevant for schizophrenia.

Targeting allosteric sites of M1 muscarinic acetylcholine receptors (M1 receptors) is a promising strategy to treat neurocognitive disorders, such as Alzheimer's disease and schizophrenia. Indeed, the last two decades have seen an impressive body of work focussing on the design and development of positive allosteric modulators (PAMs) for the M1 receptor. This has led to the identification of a structurally diverse range of highly selective M1 PAMs. In preclinical models, M1 PAMs have shown rescue of cognitive deficits and improvement of endpoints predictive of symptom domains of schizophrenia. Yet, to date only a few M1 PAMs have reached early-stage clinical trials, with many of them failing to progress further due to on-target mediated cholinergic adverse effects that have plagued the development of this class of ligand. This review covers the recent preclinical and clinical studies in the field of M1 receptor drug discovery for the treatment of Alzheimer's disease and schizophrenia, with a specific focus on M1 PAM, highlighting both the undoubted potential but also key challenges for the successful translation of M1 PAMs from bench-side to bedside.

We conduct an extensive study on using near-term quantum computers for a task in the domain of computational biology. By constructing quantum models based on parameterised quantum circuits we perform sequence classification on a task relevant to the design of therapeutic proteins, and find competitive performance with classical baselines of similar scale. To study the effect of noise, we run some of the best-performing quantum models with favourable resource requirements on emulators of state-of-the-art noisy quantum processors. We then apply error mitigation methods to improve the signal. We further execute these quantum models on the Quantinuum H1-1 trapped-ion quantum processor and observe very close agreement with noiseless exact simulation. Finally, we perform feature attribution methods and find that the quantum models indeed identify sensible relationships, at least as well as the classical baselines. This work constitutes the first proof-of-concept application of near-term quantum computing to a task critical to the design of therapeutic proteins, opening the route toward larger-scale applications in this and related fields, in line with the hardware development roadmaps of near-term quantum technologies.

Public protein sequence databases contain samples from the fitness landscape explored by nature. Protein language models (pLMs) pre-trained on these sequences aim to capture this landscape for tasks like property prediction and protein design. Following the same trend as in natural language processing, pLMs have continuously been scaled up. However, the premise that scale leads to better performance assumes that source databases provide accurate representation of the underlying fitness landscape, which is likely false. By developing an efficient codebase, designing a modern architecture, and addressing data quality concerns such as sample bias, we introduce AMPLIFY, a best-in-class pLM that is orders of magnitude less expensive to train and deploy than previous models. Furthermore, to support the scientific community and democratize the training of pLMs, we have open-sourced AMPLIFY's pre-training codebase, data, and model checkpoints. ### Competing Interest Statement The authors have declared no competing interest.

Tunicates are evolutionary model organisms bridging the gap between vertebrates and invertebrates. A genomic sequence in Ciona intestinalis (CiOX) shows high similarity to vertebrate orexin receptors and protostome allatotropin receptors (ATR). Here, molecular phylogeny suggested that CiOX is divergent from ATRs and human orexin receptors (hOX1/2). However, CiOX appears closer to hOX1/2 than to ATR both in terms of sequence percent identity and in its modelled binding cavity, as suggested by molecular modelling. CiOX was heterologously expressed in a recombinant HEK293 cell system. Human orexins weakly but concentration-dependently activated its Gq signalling (Ca2+ elevation), and the responses were inhibited by the non-selective orexin receptor antagonists TCS 1102 and almorexant, but only weakly by the OX1-selective antagonist SB-334867. Furthermore, the 5-/6-carboxytetramethylrhodamine (TAMRA)-labelled human orexin-A was able to bind to CiOX. Database mining was used to predict a potential endogenous C. intestinalis orexin peptide (Ci-orexin-A). Ci-orexin-A was able to displace TAMRA-orexin-A, but not to induce any calcium response at the CiOX. Consequently, we suggested that the orexin signalling system is conserved in Ciona intestinalis, although the relevant peptide-receptor interaction was not fully elucidated.

Generative biology combines artificial intelligence (AI), advanced life sciences technologies, and automation to revolutionize the process of designing novel biomolecules with prescribed properties, giving drug discoverers the ability to escape the limitations of biology during the design of next-generation protein therapeutics. Significant hurdles remain, namely: (i) the inherently complex nature of drug discovery, (ii) the bewildering number of promising computational and experimental techniques that have emerged in the past several years, and (iii) the limited availability of relevant protein sequence-function data for drug-like molecules. There is a need to focus on computational methods that will be most practically effective for protein drug discovery and on building experimental platforms to generate the data most appropriate for these methods. Here, we discuss recent advances in computational and experimental life sciences that are most crucial for impacting the pace and success of protein drug discovery.

Biologic drug discovery pipelines are designed to deliver protein therapeutics that have exquisite functional potency and selectivity while also manifesting biophysical characteristics suitable for manufacturing, storage, and convenient administration to patients. The ability to use computational methods to predict biophysical properties from protein sequence, potentially in combination with high throughput assays, could decrease timelines and increase the success rates for therapeutic developability engineering by eliminating lengthy and expensive cycles of recombinant protein production and testing. To support development of high-quality predictive models for antibody developability, we designed a sequence-diverse panel of 83 effector functionless IgG1 antibodies displaying a range of biophysical properties, produced and formulated each protein under standard platform conditions, and collected a comprehensive package of analytical data, including in vitro assays and in vivo mouse pharmacokinetics. We used this robust training data set to build machine learning classifier models that can predict complex protein behavior from these data and features derived from predicted and/or experimental structures. Our models predict with 87% accuracy whether viscosity at 150 mg/mL is above or below a threshold of 15 centipoise (cP) and with 75% accuracy whether the area under the plasma drug concentration-time curve (AUC(0-672 h)) in normal mouse is above or below a threshold of 3.9 x 10(6) h x ng/mL.

Artificial-intelligence tools that enable companies to share data about drug candidates while keeping sensitive information safe can unleash the potential of machine learning and cutting-edge lab techniques, for the common good.

Protein engineers aim to discover and design novel sequences with targeted, desirable properties. Given the near limitless size of the protein sequence landscape, it is no surprise that these desirable sequences are often a relative rarity. This makes identifying such sequences a costly and time-consuming endeavor. In this work, we show how to use a deep Transformer Protein Language Model to identify sequences that have the most promise. Specifically, we use the model’s self-attention map to calculate a PROMISE SCORE that weights the relative importance of a given sequence according to predicted interactions with a specified binding partner. This PROMISE SCORE can then be used to identify strong binders worthy of further study and experimentation. We use the PROMISE SCORE within two protein engineering contexts— Nanobody (Nb) discovery and protein optimization. With Nb discovery, we show how the PROMISE SCORE provides an effective way to select lead sequences from Nb repertoires. With protein optimization, we show how to use the PROMISE SCORE to select site-specific mutagenesis experiments that identify a high percentage of improved sequences. In both cases, we also show how the self-attention map used to calculate the PROMISE SCORE can indicate which regions of a protein are involved in intermolecular interactions that drive the targeted property. Finally, we describe how to fine-tune the Transformer Protein Language Model to learn a predictive model for the targeted property, and discuss the capabilities and limitations of fine-tuning with and without knowledge transfer within the context of protein engineering.

BACKGROUND: Disrupted motivational control is a common-but poorly treated-feature of psychiatric disorders, arising via aberrant mesolimbic dopaminergic signaling. GPR88 is an orphan G protein-coupled receptor that is highly expressed in the striatum and therefore well placed to modulate disrupted signaling. While the phenotype of Gpr88 knockout mice suggests a role in motivational pathways, it is unclear whether GPR88 is involved in reward valuation and/or effort-based decision making in a sex-dependent manner and whether this involves altered dopamine function. METHODS: In male and female Gpr88 knockout mice, we used touchscreen-based progressive ratio, with and without reward devaluation, and effort-related choice tasks to assess motivation and cost/benefit decision making, respectively. To explore whether these motivational behaviors were related to alterations in the striatal dopamine system, we quantified expression of dopamine-related genes and/or proteins and used [18F]DOPA positron emission tomography and GTPg[35S] binding to assess presynaptic and postsynaptic dopamine function, RESULTS: We showed that male and female Gpr88 knockout mice displayed greater motivational drive than wildtype mice, which was maintained following reward devaluation. Furthermore, we showed that cost/benefit decision making was impaired in male, but not female, Gpr88 knockout mice. Surprisingly, we found that Gpr88 deletion had no effect on striatal dopamine by any of the measures assessed. CONCLUSIONS: Our results highlight that GPR88 regulates motivational control but that disruption of such behaviors following Gpr88 deletion occurs independently of gross perturbations to striatal dopamine at a gene, protein, or functional level. This work provides further insights into GPR88 as a drug target for motivational disorders.

We introduce a novel contrastive representation learning objective and a training scheme for clinical time series. Specifically, we project high dimensional EHR. data to a closed unit ball of low dimension, encoding geometric priors so that the origin represents an idealized perfect health state and the Euclidean norm is associated with the patient's mortality risk. Moreover, using septic patients as an example, we show how we could learn to associate the angle between two vectors with the different organ system failures, thereby, learning a compact representation which is indicative of both mortality risk and specific organ failure. We show how the learned embedding can be used for online patient monitoring, can supplement clinicians and improve performance of downstream machine learning tasks. This work was partially motivated from the desire and the need to introduce a systematic way of defining intermediate rewards for Reinforcement Learning in critical care medicine. Hence, we also show how such a design in terms of the learned embedding can result in qualitatively different policies and value distributions, as compared with using only terminal rewards.

Sepsis is a potentially life threatening inflammatory response to infection or severe tissue damage. It has a highly variable clinical course, requiring constant monitoring of the patient's state to guide the management of intravenous fluids and vasopressors, among other interventions. Despite decades of research, there's still debate among experts on optimal treatment. Here, we combine for the first time, distributional deep reinforcement learning with mechanistic physiological models to find personalized sepsis treatment strategies. Our method handles partial observability by leveraging known cardiovascular physiology, introducing a novel physiology-driven recurrent autoencoder, and quantifies the uncertainty of its own results. Moreover, we introduce a framework for uncertainty aware decision support with humans in the loop. We show that our method learns physiologically explainable, robust policies that are consistent with clinical knowledge. Further our method consistently identifies high risk states that lead to death, which could potentially benefit from more frequent vasopressor administration, providing valuable guidance for future research

Background and Purpose Muscarinic acetylcholine receptors mediate alcohol consumption and seeking in rats. While M-4 and M-5 receptors have recently been implicated to mediate these behaviours in the striatum, their role in other brain regions remain unknown. The ventral tegmental area (VTA) and ventral subiculum (vSub) both densely express M-4 and M-5 receptors and modulate alcohol-seeking, via their projections to the nucleus accumbens shell (AcbSh). Experimental Approach In Indiana alcohol-preferring (iP) male rats, we examined Chrm4 (M-4) and Chrm5 (M-5) expression in the VTA and vSub following long-term alcohol consumption and abstinence using RT-qPCR. Using a combination of retrograde tracing and RNAscope, we examined the localisation of Chrm4 and Chrm5 on vSub cells that project to the AcbSh. Using selective allosteric modulators, we examined the functional role of M-4 and M-5 receptors within the vSub in alcohol consumption, context-induced alcohol-seeking, locomotor activity, and food/water consumption. Key Results Long-term alcohol and abstinence dysregulated the expression of genes for muscarinic receptors in the vSub, not in the VTA. Chrm4 was down-regulated following long-term alcohol and abstinence, while Chrm5 was up-regulated following long-term alcohol consumption. Consistent with these data, a positive allosteric modulator (VU0467154) of intra-vSub M-4 receptors reduced context-induced alcohol-seeking, but not motivation for alcohol self-administration, while M-5 receptor negative allosteric modulator (ML375) reduced initial motivation for alcohol self-administration, but not context-induced alcohol-seeking. Conclusion and Implications Collectively, our data highlight alcohol-induced cholinergic dysregulation in the vSub and distinct roles for M-4 and M-5 receptor allosteric modulators to reduce alcohol consumption or seeking.

Background Supervised learning from high-throughput sequencing data presents many challenges. For one, the curse of dimensionality often leads to overfitting as well as issues with scalability. This can bring about inaccurate models or those that require extensive compute time and resources. Additionally, variant calls may not be the optimal encoding for a given learning task, which also contributes to poor predictive capabilities. To address these issues, we present Harvestman, a method that takes advantage of hierarchical relationships among the possible biological interpretations and representations of genomic variants to perform automatic feature learning, feature selection, and model building. Results We demonstrate that Harvestman scales to thousands of genomes comprising more than 84 million variants by processing phase 3 data from the 1000 Genomes Project, one of the largest publicly available collection of whole genome sequences. Using breast cancer data from The Cancer Genome Atlas, we show that Harvestman selects a rich combination of representations that are adapted to the learning task, and performs better than a binary representation of SNPs alone. We compare Harvestman to existing feature selection methods and demonstrate that our method is more parsimonious-it selects smaller and less redundant feature subsets while maintaining accuracy of the resulting classifier. Conclusion Harvestman is a hierarchical feature selection approach for supervised model building from variant call data. By building a knowledge graph over genomic variants and solving an integer linear program , Harvestman automatically and optimally finds the right encoding for genomic variants. Compared to other hierarchical feature selection methods, Harvestman is faster and selects features more parsimoniously.