Industrializing AI-powered drug discovery: lessons learned from the Patrimony computing platform.

INTRODUCTION: As a mid-size international pharmaceutical company, we initiated 4 years ago the launch of a dedicated high-throughput computing platform supporting drug discovery. The platform named 'Patrimony' was built up on the initial predicate to capitalize on our proprietary data while leveraging public data sources in order to foster a Computational Precision Medicine approach with the power of artificial intelligence. AREAS COVERED: Specifically, Patrimony is designed to identify novel therapeutic target candidates. With several successful use cases in immuno-inflammatory diseases, and current ongoing extension to applications to oncology and neurology, we document how this industrial computational platform has had a transformational impact on our R&D, making it more competitive, as well time and cost effective through a model-based educated selection of therapeutic targets and drug candidates. EXPERT OPINION: We report our achievements, but also our challenges in implementing data access and governance processes, building up hardware and user interfaces, and acculturing scientists to use predictive models to inform decisions.

Aging increases circulating BH2 without modifying BH4 levels and impairs peripheral vascular function in healthy adults.

Little is known about the mechanisms of aging on vascular beds and its relationship with tetra and di-hydrobiopterin (BH4 and BH2) levels. This observational clinical study analyzed the impact of aging on plasma and platelet biopterins, cutaneous blood flow (CBF), and coronary flow reserve (CFR) in healthy adults. The study enrolled healthy adults in 3 age groups: 18-30, 50-59, and 60-70 years (n = 25/group). Biopterins were assessed by LC-MS/MS using newly defined pre-analytical conditions limiting BH4 oxidation and improving long-term stability. CBF was measured by Laser Speckle Contrast Imaging coupled with acetylcholine-iontophoresis and CFR by adenosine stress cardiac magnetic resonance. In healthy adults, aging (60-70 years vs 18-30 years) significantly increased platelet BH2 (+75%, P = 0.033) and BH2 + BH4 (+31%, P = 0.033), and to a lesser extent plasma BH2 (+29%, P = 0.009) without affecting BH4 and BH4/BH2. Simultaneously, CBF was decreased (-23%, P = 0.004) but not CFR, CBF being inversely correlated with platelet BH2 (r = -0.42, P = 0.001) and BH2 + BH4 (r = -0.41, P = 0.002). The proportion of adults with abnormal platelet BH2 increased with age (+28% in 60-70y). These abnormal BH2 levels were significantly associated with reduced CBF and CFR (-16%, P = 0.03 and -26%, P = 0.02). In conclusion, our study showed that age-related peripheral endothelial dysfunction was associated with an increase in circulating BH2 without decreasing BH4, the effect being more marked in platelets, the most relevant blood compartment to assess biopterin bioavailability. Peripheral but not coronary vascular function is progressively impaired with aging in healthy adults. All these findings support biopterins as therapeutic targets to improve vascular function.

A new molecular classification to drive precision treatment strategies in primary Sjögren's syndrome.

There is currently no approved treatment for primary Sjögren's syndrome, a disease that primarily affects adult women. The difficulty in developing effective therapies is -in part- because of the heterogeneity in the clinical manifestation and pathophysiology of the disease. Finding common molecular signatures among patient subgroups could improve our understanding of disease etiology, and facilitate the development of targeted therapeutics. Here, we report, in a cross-sectional cohort, a molecular classification scheme for Sjögren's syndrome patients based on the multi-omic profiling of whole blood samples from a European cohort of over 300 patients, and a similar number of age and gender-matched healthy volunteers. Using transcriptomic, genomic, epigenetic, cytokine expression and flow cytometry data, combined with clinical parameters, we identify four groups of patients with distinct patterns of immune dysregulation. The biomarkers we identify can be used by machine learning classifiers to sort future patients into subgroups, allowing the re-evaluation of response to treatments in clinical trials.

Shallow Whole-Genome Sequencing from Plasma Identifies FGFR1 Amplified Breast Cancers and Predicts Overall Survival.

Background: Focal amplification of fibroblast growth factor receptor 1 (FGFR1) defines a subgroup of breast cancers with poor prognosis and high risk of recurrence. We sought to demonstrate the potential of circulating cell-free DNA (cfDNA) analysis to evaluate FGFR1 copy numbers from a cohort of 100 metastatic breast cancer (mBC) patients. Methods: Formalin-fixed paraffin-embedded (FFPE) tissue samples were screened for FGFR1 amplification by FISH, and positive cases were confirmed with a microarray platform (OncoscanTM). Subsequently, cfDNA was evaluated by two approaches, i.e., mFAST-SeqS and shallow whole-genome sequencing (sWGS), to estimate the circulating tumor DNA (ctDNA) allele fraction (AF) and to evaluate the FGFR1 status. Results: Tissue-based analyses identified FGFR1 amplifications in 20/100 tumors. All cases with a ctDNA AF above 3% (n = 12) showed concordance for FGFR1 status between tissue and cfDNA. In one case, we were able to detect a high-level FGFR1 amplification, although the ctDNA AF was below 1%. Furthermore, high levels of ctDNA indicated an association with unfavorable prognosis based on overall survival. Conclusions: Screening for FGFR1 amplification in ctDNA might represent a viable strategy to identify patients eligible for treatment by FGFR inhibition, and mBC ctDNA levels might be used for the evaluation of prognosis in clinical drug trials.

Lucitanib for the Treatment of HR+/HER2- Metastatic Breast Cancer: Results from the Multicohort Phase II FINESSE Study.

PURPOSE: The FGFR1 gene is amplified in 14% of patients with HR + /HER2 - breast cancer. Efficacy and safety of lucitanib, an inhibitor of VEGFR1-3, FGFR1-3, and PDGFRα/ß, were assessed. PATIENTS AND METHODS: Patients with HR + /HER2 - metastatic breast cancer (MBC) received oral lucitanib in three centrally confirmed cohorts: (i) FGFR1 amplified, (ii) FGFR1 nonamplified, 11q13 amplified, and (iii) FGFR1 and 11q13 nonamplified. Key inclusion criteria included Eastern Cooperative Oncology Group Performance Status ≤2, ≥1 line of anticancer therapy, but ≤2 lines of chemotherapy. Primary endpoint was overall response rates (ORR) by RECIST1.1. Simon's two-stage design was used: If ≥2 patients responded among 21 patients, 20 additional patients could be enrolled in each cohort. FGFR1 copy-number variation (CNV) was determined by FISH and droplet digital PCR, whereas FGFR1 expression was determined by IHC. RESULTS: Seventy-six patients (32/18/26 in cohorts 1/2/3) from nine countries were enrolled. The prespecified primary endpoint was met in cohort 1 with ORR of 19% [95% confidence interval (CI), 9%-35%], but not in cohorts 2 and 3 with ORR of 0% (95% CI, 0%-18%) and 15% (95% CI, 6%-34%), respectively. Frequent adverse events included hypertension (87%), hypothyroidism (45%), nausea (33%), and proteinuria (32%). Exploratory biomarker analyses suggested higher ORR in patients with high FGFR1 amplification (≥4 CNV) than those without high amplification (22% vs. 9%). ORR in patients with FGFR1-high tumors (IHC, H-score ≥50) was 25% versus 8% in FGFR1-low cancers. CONCLUSIONS: Lucitanib had modest antitumor activity and significant hypertension-related toxicity in patients with HR + /HER2 - MBC. Although based on small sample sizes, exploratory biomarker analyses suggested that patients with high FGFR1 amplification or expression might derive greater benefit.

2P2I HUNTER: a tool for filtering orthosteric protein-protein interaction modulators via a dedicated support vector machine.

Over the last 10 years, protein-protein interactions (PPIs) have shown increasing potential as new therapeutic targets. As a consequence, PPIs are today the most screened target class in high-throughput screening (HTS). The development of broad chemical libraries dedicated to these particular targets is essential; however, the chemical space associated with this 'high-hanging fruit' is still under debate. Here, we analyse the properties of 40 non-redundant small molecules present in the 2P2I database (http://2p2idb.cnrs-mrs.fr/) to define a general profile of orthosteric inhibitors and propose an original protocol to filter general screening libraries using a support vector machine (SVM) with 11 standard Dragon molecular descriptors. The filtering protocol has been validated using external datasets from PubChem BioAssay and results from in-house screening campaigns. This external blind validation demonstrated the ability of the SVM model to reduce the size of the filtered chemical library by eliminating up to 96% of the compounds as well as enhancing the proportion of active compounds by up to a factor of 8. We believe that the resulting chemical space identified in this paper will provide the scientific community with a concrete support to search for PPI inhibitors during HTS campaigns.

Coordinate Transcriptomic and Metabolomic Effects of the Insulin Sensitizer Rosiglitazone on Fundamental Metabolic Pathways in Liver, Soleus Muscle, and Adipose Tissue in Diabetic db/db Mice.

Rosiglitazone (RSG), developed for the treatment of type 2 diabetes mellitus, is known to have potent effects on carbohydrate and lipid metabolism leading to the improvement of insulin sensitivity in target tissues. To further assess the capacity of RSG to normalize gene expression in insulin-sensitive tissues, we compared groups of 18-day-treated db/db mice with increasing oral doses of RSG (10, 30, and 100 mg/kg/d) with untreated non-diabetic littermates (db/+). For this aim, transcriptional changes were measured in liver, inguinal adipose tissue (IAT) and soleus muscle using microarrays and real-time PCR. In parallel, targeted metabolomic assessment of lipids (triglycerides (TGs) and free fatty acids (FFAs)) in plasma and tissues was performed by UPLC-MS methods. Multivariate analyses revealed a relationship between the differential gene expressions in liver and liver trioleate content and between blood glucose levels and a combination of differentially expressed genes measured in liver, IAT, and muscle. In summary, we have integrated gene expression and targeted metabolomic data to present a comprehensive overview of RSG-induced changes in a diabetes mouse model and improved the molecular understanding of how RSG ameliorates diabetes through its effect on the major insulin-sensitive tissues.