Artificial intelligence enabled parabolic response surface platform identifies ultra-rapid near-universal TB drug treatment regimens comprising approved drugs

Radiation therapy with phenotypic medicine: towards N-of-1 personalization

In current clinical practice, radiotherapy (RT) is prescribed as a pre-determined total dose divided over daily doses (fractions) given over several weeks. The treatment response is typically assessed months after the end of RT. However, the conventional one-dose-fits-all strategy may not achieve the desired outcome, owing to patient and tumor heterogeneity. Therefore, a treatment strategy that allows for RT dose personalization based on each individual response is preferred. Multiple strategies have been adopted to address this challenge. As an alternative to current known strategies, artificial intelligence (AI)-derived mechanism-independent small data phenotypic medicine (PM) platforms may be utilized for N-of-1 RT personalization. Unlike existing big data approaches, PM does not engage in model refining, training, and validation, and guides treatment by utilizing prospectively collected patient's own small datasets. With PM, clinicians may guide patients' RT dose recommendations using their responses in real-time and potentially avoid over-treatment in good responders and under-treatment in poor responders. In this paper, we discuss the potential of engaging PM to guide clinicians on upfront dose selections and ongoing adaptations during RT, as well as considerations and limitations for implementation. For practicing oncologists, clinical trialists, and researchers, PM can either be implemented as a standalone strategy or in complement with other existing RT personalizations. In addition, PM can either be used for monotherapeutic RT personalization, or in combination with other therapeutics (e.g. chemotherapy, targeted therapy). The potential of N-of-1 RT personalization with drugs will also be presented.

Transfer learning predicts species-specific drug interactions in emerging pathogens

Machine learning (ML) algorithms are necessary to efficiently identify potent drug combinations within a large candidate space to combat drug resistance. However, existing ML approaches cannot be applied to emerging and under-studied pathogens with limited training data. To address this, we developed a transfer learning and crowdsourcing framework (TACTIC) to train ML models on data from multiple bacteria. TACTIC was built using 2,965 drug interactions from 12 bacterial strains and outperformed traditional ML models in predicting drug interaction outcomes for species that lack training data. Top TACTIC model features revealed genetic and metabolic factors that influence cross-species and species-specific drug interaction outcomes. Upon analyzing ~600,000 predicted drug interactions across 9 metabolic environments and 18 bacterial strains, we identified a small set of drug interactions that are selectively synergistic against Gram-negative (e.g., A. baumannii) and non-tuberculous mycobacteria (NTM) pathogens. We experimentally validated synergistic drug combinations containing clarithromycin, ampicillin, and mecillinam against M. abscessus, an emerging pathogen with growing levels of antibiotic resistance. Lastly, we leveraged TACTIC to propose selectively synergistic drug combinations to treat bacterial eye infections (endophthalmitis).

Efficacy and safety of an innovative short-course regimen containing clofazimine for treatment of drug-susceptible tuberculosis: a clinical trial

In preclinical studies, a new antituberculosis drug regimen markedly reduced the time required to achieve relapse-free cure. This study aimed to preliminarily evaluate the efficacy and safety of this four-month regimen, consisting of clofazimine, prothionamide, pyrazinamide and ethambutol, with a standard six-month regimen in patients with drug-susceptible tuberculosis. An open-label pilot randomized clinical trial was conducted among the patients with newly diagnosed bacteriologically-confirmed pulmonary tuberculosis. The primary efficacy end-point was sputum culture negative conversion. Totally, 93 patients were included in the modified intention-to-treat population. The rates of sputum culture conversion were 65.2% (30/46) and 87.2% (41/47) for short-course and standard regimen group, respectively. There was no difference on two-month culture conversion rates, time to culture conversion, nor early bactericidal activity (P > 0.05). However, patients on short-course regimen were observed with lower rates of radiological improvement or recovery and sustained treatment success, which was mainly attributed to higher percent of patients permanently changed assigned regimen (32.1% vs. 12.3%, P = 0.012). The main cause for it was drug-induced hepatitis (16/17). Although lowering the dose of prothionamide was approved, the alternative option of changing assigned regimen was chosen in this study. While in per-protocol population, sputum culture conversion rates were 87.0% (20/23) and 94.4% (34/36) for the respective groups. Overall, the short-course regimen appeared to have inferior efficacy and higher incidence of hepatitis but desired efficacy in per-protocol population. It provides the first proof-of-concept in humans of the capacity of the short-course approach to identify drug regimens that can shorten the treatment time for tuberculosis.

Clofazimine for the treatment of tuberculosis

Shorter (6-9 months), fully oral regimens containing new and repurposed drugs are now the first-choice option for the treatment of drug-resistant tuberculosis (DR-TB). Clofazimine, long used in the treatment of leprosy, is one such repurposed drug that has become a cornerstone of DR-TB treatment and ongoing trials are exploring novel, shorter clofazimine-containing regimens for drug-resistant as well as drug-susceptible tuberculosis. Clofazimine's repurposing was informed by evidence of potent activity against DR-TB strains in vitro and in mice and a treatment-shortening effect in DR-TB patients as part of a multidrug regimen. Clofazimine entered clinical use in the 1950s without the rigorous safety and pharmacokinetic evaluation which is part of modern drug development and current dosing is not evidence-based. Recent studies have begun to characterize clofazimine's exposure-response relationship for safety and efficacy in populations with TB. Despite being better tolerated than some other second-line TB drugs, the extent and impact of adverse effects including skin discolouration and cardiotoxicity are not well understood and together with emergent resistance, may undermine clofazimine use in DR-TB programmes. Furthermore, clofazimine's precise mechanism of action is not well established, as is the genetic basis of clofazimine resistance. In this narrative review, we present an overview of the evidence base underpinning the use and limitations of clofazimine as an antituberculosis drug and discuss advances in the understanding of clofazimine pharmacokinetics, toxicity, and resistance. The unusual pharmacokinetic properties of clofazimine and how these relate to its putative mechanism of action, antituberculosis activity, dosing considerations and adverse effects are highlighted. Finally, we discuss the development of novel riminophenazine analogues as antituberculosis drugs.

Advances in the design of combination therapies for the treatment of tuberculosis

INTRODUCTION

Tuberculosis requires lengthy multi-drug therapy. Mycobacterium tuberculosis occupies different tissue compartments during infection, making drug access and susceptibility patterns variable. Antibiotic combinations are needed to ensure each compartment of infection is reached with effective drug treatment. Despite drug combinations' role in treating tuberculosis, the design of such combinations has been tackled relatively late in the drug development process, limiting the number of drug combinations tested. In recent years, there has been significant progress using in vitro, in vivo, and computational methodologies to interrogate combination drug effects.

AREAS COVERED

This review discusses the advances in these methodologies and how they may be used in conjunction with new successful clinical trials of novel drug combinations to design optimized combination therapies for tuberculosis. Literature searches for approaches and experimental models used to evaluate drug combination effects were undertaken.

EXPERT OPINION

We are entering an era richer in combination drug effect and pharmacokinetic/pharmacodynamic data, genetic tools, and outcome measurement types. Application of computational modeling approaches that integrate these data and produce predictive models of clinical outcomes may enable the field to generate novel, effective multidrug therapies using existing and new drug combination backbones.

Anti-tuberculosis treatment strategies and drug development: challenges and priorities

Despite two decades of intensified research to understand and cure tuberculosis disease, biological uncertainties remain and hamper progress. However, owing to collaborative initiatives including academia, the pharmaceutical industry and non-for-profit organizations, the drug candidate pipeline is promising. This exceptional success comes with the inherent challenge of prioritizing multidrug regimens for clinical trials and revamping trial designs to accelerate regimen development and capitalize on drug discovery breakthroughs. Most wanted are markers of progression from latent infection to active pulmonary disease, markers of drug response and predictors of relapse, in vitro tools to uncover synergies that translate clinically and animal models to reliably assess the treatment shortening potential of new regimens. In this Review, we highlight the benefits and challenges of 'one-size-fits-all' regimens and treatment duration versus individualized therapy based on disease severity and host and pathogen characteristics, considering scientific and operational perspectives.

The IDentif.AI-x pandemic readiness platform: Rapid prioritization of optimized COVID-19 combination therapy regimens

IDentif.AI-x, a clinically actionable artificial intelligence platform, was used to rapidly pinpoint and prioritize optimal combination therapies against COVID-19 by pairing a prospective, experimental validation of multi-drug efficacy on a SARS-CoV-2 live virus and Vero E6 assay with a quadratic optimization workflow. A starting pool of 12 candidate drugs developed in collaboration with a community of infectious disease clinicians was first narrowed down to a six-drug pool and then interrogated in 50 combination regimens at three dosing levels per drug, representing 729 possible combinations. IDentif.AI-x revealed EIDD-1931 to be a strong candidate upon which multiple drug combinations can be derived, and pinpointed a number of clinically actionable drug interactions, which were further reconfirmed in SARS-CoV-2 variants B.1.351 (Beta) and B.1.617.2 (Delta). IDentif.AI-x prioritized promising drug combinations for clinical translation and can be immediately adjusted and re-executed with a new pool of promising therapies in an actionable path towards rapidly optimizing combination therapy following pandemic emergence.

Orthogonal array composite designs for drug combination experiments with applications for tuberculosis

The aim of this article is to provide an overview of the orthogonal array composite design (OACD) methodology, illustrate the various advantages, and provide a real-world application. An OACD combines a two-level factorial design with a three-level orthogonal array and it can be used as an alternative to existing composite designs for building response surface models. We compare the D $$ D $$ -efficiencies of OACDs relative to the commonly used central composite design (CCD) when there are a few missing observations and demonstrate that OACDs are more robust to missing observations for two scenarios. The first scenario assumes one missing observation either from one factorial point or one additional point. The second scenario assumes two missing observations either from two factorial points or from two additional points, or from one factorial point and one additional point. Furthermore, we compare OACDs and CCDs in terms of I $$ I $$ -optimality for precise predictions. Lastly, a real-world application of an OACD for a tuberculosis drug combination study is provided.

Systematic measurement of combination-drug landscapes to predict in vivo treatment outcomes for tuberculosis

Lengthy multidrug chemotherapy is required to achieve a durable cure in tuberculosis. However, we lack well-validated, high-throughput in vitro models that predict animal outcomes. Here, we provide an extensible approach to rationally prioritize combination therapies for testing in in vivo mouse models of tuberculosis. We systematically measured Mycobacterium tuberculosis response to all two- and three-drug combinations among ten antibiotics in eight conditions that reproduce lesion microenvironments, resulting in >500,000 measurements. Using these in vitro data, we developed classifiers predictive of multidrug treatment outcome in a mouse model of disease relapse and identified ensembles of in vitro models that best describe in vivo treatment outcomes. We identified signatures of potencies and drug interactions in specific in vitro models that distinguish whether drug combinations are better than the standard of care in two important preclinical mouse models. Our framework is generalizable to other difficult-to-treat diseases requiring combination therapies. A record of this paper's transparent peer review process is included in the supplemental information.

Pharmacokinetics and Target Attainment of SQ109 in Plasma and Human-Like Tuberculosis Lesions in Rabbits

SQ109 is a novel well-tolerated drug candidate in clinical development for the treatment of drug-resistant tuberculosis (TB). It is the only inhibitor of the MmpL3 mycolic acid transporter in clinical development. No SQ109-resistant mutant has been directly isolated thus far in vitro, in mice, or in patients, which is tentatively attributed to its multiple targets. It is considered a potential replacement for poorly tolerated components of multidrug-resistant TB regimens. To prioritize SQ109-containing combinations with the best potential for cure and treatment shortening, one must understand its contribution against different bacterial populations in pulmonary lesions. Here, we have characterized the pharmacokinetics of SQ109 in the rabbit model of active TB and its penetration at the sites of disease—lung tissue, cellular and necrotic lesions, and caseum. A two-compartment model with first-order absorption and elimination described the plasma pharmacokinetics. At the human-equivalent dose, parameter estimates fell within the ranges published for preclinical species. Tissue concentrations were modeled using an “effect” compartment, showing high accumulation in lung and cellular lesion areas with penetration coefficients in excess of 1,000 and lower passive diffusion in caseum after 7 daily doses. These results, together with the hydrophobic nature and high nonspecific caseum binding of SQ109, suggest that multiweek dosing would be required to reach steady state in caseum and poorly vascularized compartments, similar to bedaquiline. Linking lesion pharmacokinetics to SQ109 potency in assays against replicating, nonreplicating, and intracellular M. tuberculosis showed SQ109 concentrations markedly above pharmacokinetic-pharmacodynamic targets in lung and cellular lesions throughout the dosing interval.

Development of New Tuberculosis Drugs: Translation to Regimen Composition for Drug-Sensitive and Multidrug-Resistant Tuberculosis

Tuberculosis (TB) kills more people than any other infectious disease. Challenges for developing better treatments include the complex pathology due to within-host immune dynamics, interpatient variability in disease severity and drug pharmacokinetics-pharmacodynamics (PK-PD), and the growing emergence of resistance. Model-informed drug development using quantitative and translational pharmacology has become increasingly recognized as a method capable of drug prioritization and regimen optimization to efficiently progress compounds through TB drug development phases. In this review, we examine translational models and tools, including plasma PK scaling, site-of-disease lesion PK, host-immune and bacteria interplay, combination PK-PD models of multidrug regimens, resistance formation, and integration of data across nonclinical and clinical phases.We propose a workflow that integrates these tools with computational platforms to identify drug combinations that have the potential to accelerate sterilization, reduce relapse rates, and limit the emergence of resistance.

Biology of antimicrobial resistance and approaches to combat it

Insufficient development of new antibiotics and the rising resistance of bacteria to those that we have are putting the world at risk of losing the most widely curative class of medicines currently available. Preventing deaths from antimicrobial resistance (AMR) will require exploiting emerging knowledge not only about genetic AMR conferred by horizontal gene transfer or de novo mutations but also about phenotypic AMR, which lacks a stably heritable basis. This Review summarizes recent advances and continuing limitations in our understanding of AMR and suggests approaches for combating its clinical consequences, including identification of previously unexploited bacterial targets, new antimicrobial compounds, and improved combination drug regimens.

Validation of a universal and highly sensitive two-dimensional liquid chromatography-tandem mass spectrometry methodology for the quantification of pyrazinamide, ethambutol, protionamide, and clofazimine in different biological matrices

A novel and potent anti-tuberculosis drug combination pyrazinamide (PZA), ethambutol (EMB), protionamide (PTO), and clofazimine (CFZ) that rapidly kills Mycobacterium tuberculosis (Mtb) in the lungs has been identified using the artificial-intelligence-enabled parabolic response surface approach. A universal and highly sensitive two-dimensional liquid chromatography-tandem mass spectrometry (2D-LC-MS/MS) method for the simultaneous determination of PZA, EMB, PTO, and CFZ in various biological samples in different states (liquid samples: plasma, bile, and urine; solid samples: tissue and feces) using simple pretreatment was established and validated. For the first dimension of this column-switching arrangement, the automated purification and enrichment of the drugs were achieved on a Polar-RP column. The subsequent analytical separation was performed on an Agilent Zorbax SB-Aq column, and the total loop time was 7.5 min. The positive-ionization mode with multiple reaction monitoring was used for detection. The sensitivity was good with no carry-over detected, and the lower limit of quantification ranged from 100 to 500 pg/mL. This quantification method was fully validated and proved to be robust in accordance with US Food and Drug Administration guidelines. High recoveries (85.3-111.4%) and accuracies (92.1-109.3%), together with high precision values (0.5-13.8%), were verified in all matrices. All standard curves showed favorable linearities with r2 > 0.995. This validated method was applied to study plasma pharmacokinetics, tissue distribution, and excretion in Sprague-Dawley rats after oral administration of the drug combination.

Advances in clinical trial design: Weaving tomorrow's TB treatments

Christian Lienhardt and co-authors discuss the conclusions of the PLOS Medicine Collection on advances in clinical trial design for development of new tuberculosis treatments.

CURATE.AI: Optimizing Personalized Medicine with Artificial Intelligence

The clinical team attending to a patient upon a diagnosis is faced with two main questions: what treatment, and at what dose? Clinical trials' results provide the basis for guidance and support for official protocols that clinicians use to base their decisions upon. However, individuals rarely demonstrate the reported response from relevant clinical trials, often the average from a group representing a population or subpopulation. The decision complexity increases with combination treatments where drugs administered together can interact with each other, which is often the case. Additionally, the individual's response to the treatment varies over time with the changes in his or her condition, whether via the indication or physiology. In practice, the drug and the dose selection depend greatly on the medical protocol of the healthcare provider and the medical team's experience. As such, the results are inherently varied and often suboptimal. Big data approaches have emerged as an excellent decision-making support tool, but their application is limited by multiple challenges, the main one being the availability of sufficiently big datasets with good quality, representative information. An alternative approach-phenotypic personalized medicine (PPM)-finds an appropriate drug combination (quadratic phenotypic optimization platform [QPOP]) and an appropriate dosing strategy over time (CURATE.AI) based on small data collected exclusively from the treated individual. PPM-based approaches have demonstrated superior results over the current standard of care. The side effects are limited while the desired output is maximized, which directly translates into improving the length and quality of individuals' lives.

AI‐Enabled Parabolic Response Surface Approach Identifies Ultra Short‐Course Near‐Universal TB Drug Regimens

Tuberculosis (TB) is a major health problem that causes more deaths worldwide than any other single infectious disease. Current multidrug therapy for tuberculosis is exceedingly lengthy, leading to poor drug adherence, and consequently the emergence of drug resistance. Hence, much more rapid treatments are needed. Experimentally identifying the most synergistic drug combinations among available drugs is complicated by the astronomical number of possible drug-dose combinations. This problem is dealt with by the use of an artificial-intelligence-enabled parabolic response surface platform in conjunction with an in vitro Mycobacterium tuberculosis–infected macrophage cell culture assay amenable to high-throughput screening. This strategy allows rapid identification of the most effective drug-dose combinations by testing only a small fraction of the total drug-dose efficacy response surface. The same platform is then used to optimize the in vivo doses of each drug in the most potent regimens. Thus, regimens are identified that are dramatically more effective than the Standard Regimen in treating TB in a mouse model—a model broadly predictive of drug efficacy in humans. The most effective regimens reported herein shorten the duration of treatment required to achieve relapse-free cure by 80% and are suitable for treating both drug-sensitive and most drug-resistant cases of tuberculosis.