Sample efficient reinforcement learning with active learning for molecular design.

Reinforcement learning (RL) is a powerful and flexible paradigm for searching for solutions in high-dimensional action spaces. However, bridging the gap between playing computer games with thousands of simulated episodes and solving real scientific problems with complex and involved environments (up to actual laboratory experiments) requires improvements in terms of sample efficiency to make the most of expensive information. The discovery of new drugs is a major commercial application of RL, motivated by the very large nature of the chemical space and the need to perform multiparameter optimization (MPO) across different properties. In silico methods, such as virtual library screening (VS) and de novo molecular generation with RL, show great promise in accelerating this search. However, incorporation of increasingly complex computational models in these workflows requires increasing sample efficiency. Here, we introduce an active learning system linked with an RL model (RL-AL) for molecular design, which aims to improve the sample-efficiency of the optimization process. We identity and characterize unique challenges combining RL and AL, investigate the interplay between the systems, and develop a novel AL approach to solve the MPO problem. Our approach greatly expedites the search for novel solutions relative to baseline-RL for simple ligand- and structure-based oracle functions, with a 5-66-fold increase in hits generated for a fixed oracle budget and a 4-64-fold reduction in computational time to find a specific number of hits. Furthermore, compounds discovered through RL-AL display substantial enrichment of a multi-parameter scoring objective, indicating superior efficacy in curating high-scoring compounds, without a reduction in output diversity. This significant acceleration improves the feasibility of oracle functions that have largely been overlooked in RL due to high computational costs, for example free energy perturbation methods, and in principle is applicable to any RL domain.

Unmodified rabies mRNA vaccine elicits high cross-neutralizing antibody titers and diverse B cell memory responses.

Licensed rabies virus vaccines based on whole inactivated virus are effective in humans. However, there is a lack of detailed investigations of the elicited immune response, and whether responses can be improved using novel vaccine platforms. Here we show that two doses of a lipid nanoparticle-formulated unmodified mRNA vaccine encoding the rabies virus glycoprotein (RABV-G) induces higher levels of RABV-G specific plasmablasts and T cells in blood, and plasma cells in the bone marrow compared to two doses of Rabipur in non-human primates. The mRNA vaccine also generates higher RABV-G binding and neutralizing antibody titers than Rabipur, while the degree of somatic hypermutation and clonal diversity of the response are similar for the two vaccines. The higher overall antibody titers induced by the mRNA vaccine translates into improved cross-neutralization of related lyssavirus strains, suggesting that this platform has potential for the development of a broadly protective vaccine against these viruses.

The effect of gum chewing on xerostomia and salivary flow rate in elderly and medically compromised subjects: a systematic review and meta-analysis.

BACKGROUND: Xerostomia negatively affects quality of life. Symptoms include oral dryness; thirst; difficulty speaking, chewing, and swallowing food; oral discomfort; mouth soft tissue soreness and infections; and rampant tooth decay. The objective of this systematic review and meta-analysis was to investigate if gum chewing is an intervention that results in objective improvements in salivary flow rates and subjective relief from xerostomia. METHOD: We searched electronic databases including Medline, Scopus, Web of Science, Embase, Cochrane Library (CDSR and Central), Google Scholar and the citations of review papers (last searched 31/03/23). The study populations included: 1) elderly people with xerostomia (> 60 years old, any gender, and severity of xerostomia), and 2) medically compromised people with xerostomia. The intervention of interest was gum chewing. Comparisons included gum chewing vs. no gum chewing. The outcomes included salivary flow rate, self-reported xerostomia, and thirst. All settings and study designs were included. We conducted a meta-analysis on studies where measurements of unstimulated whole salivary flow rate for both a gum chewing, and no gum chewing intervention (daily chewing of gum for two weeks or longer) were reported. We assessed risk of bias using Cochrane's RoB 2 and ROBINS-I tools. RESULTS: Nine thousand six hundred and two studies were screened and 0.26% (n = 25) met the inclusion criteria for the systematic review. Two of the 25 papers had a high overall risk of bias. Of the 25 papers selected for the systematic review, six met the criteria to be included in the meta-analysis which confirmed a significant overall effect of gum on saliva flow outcomes compared to control (SMD = 0.44, 95% CI: 0.22-0.66; p = 0.00008; I2 = 46.53%). CONCLUSIONS: Chewing gum can increase unstimulated salivary flow rate in elderly and medically compromised people with xerostomia. Increasing the number of days over which gum is chewed increases the improvement in the rate of salivation. Gum chewing is linked with improvements in self-reported levels of xerostomia (although it is noted that no significant effects were detected in five of the studies reviewed). Future studies should eliminate sources of bias, standardise methods to measure salivary flow rate, and use a common instrument to measure subjective relief from xerostomia. STUDY REGISTRATION: PROSPERO CRD42021254485.

A tool for evaluating heterogeneity in avidity of polyclonal antibodies.

Diversity in specificity of polyclonal antibody (pAb) responses is extensively investigated in vaccine efficacy or immunological evaluations, but the heterogeneity in antibody avidity is rarely probed as convenient tools are lacking. Here we have developed a polyclonal antibodies avidity resolution tool (PAART) for use with label-free techniques, such as surface plasmon resonance and biolayer interferometry, that can monitor pAb-antigen interactions in real time to measure dissociation rate constant (kd ) for defining avidity. PAART utilizes a sum of exponentials model to fit the dissociation time-courses of pAb-antigens interactions and resolve multiple kd contributing to the overall dissociation. Each kd value of pAb dissociation resolved by PAART corresponds to a group of antibodies with similar avidity. PAART is designed to identify the minimum number of exponentials required to explain the dissociation course and guards against overfitting of data by parsimony selection of best model using Akaike information criterion. Validation of PAART was performed using binary mixtures of monoclonal antibodies of same specificity but differing in kd of the interaction with their epitope. We applied PAART to examine the heterogeneity in avidities of pAb from malaria and typhoid vaccinees, and individuals living with HIV-1 that naturally control the viral load. In many cases, two to three kd were dissected indicating the heterogeneity of pAb avidities. We showcase examples of affinity maturation of vaccine induced pAb responses at component level and enhanced resolution of heterogeneity in avidity when antigen-binding fragments (Fab) are used instead of polyclonal IgG antibodies. The utility of PAART can be manifold in examining circulating pAb characteristics and could inform vaccine strategies aimed to guide the host humoral immune response.

Microglia regulate central nervous system myelin growth and integrity.

Myelin is required for the function of neuronal axons in the central nervous system, but the mechanisms that support myelin health are unclear. Although macrophages in the central nervous system have been implicated in myelin health1, it is unknown which macrophage populations are involved and which aspects they influence. Here we show that resident microglia are crucial for the maintenance of myelin health in adulthood in both mice and humans. We demonstrate that microglia are dispensable for developmental myelin ensheathment. However, they are required for subsequent regulation of myelin growth and associated cognitive function, and for preservation of myelin integrity by preventing its degeneration. We show that loss of myelin health due to the absence of microglia is associated with the appearance of a myelinating oligodendrocyte state with altered lipid metabolism. Moreover, this mechanism is regulated through disruption of the TGFß1-TGFßR1 axis. Our findings highlight microglia as promising therapeutic targets for conditions in which myelin growth and integrity are dysregulated, such as in ageing and neurodegenerative disease2,3.

Changing trends in hip fracture epidemiology in the Republic of Ireland: a follow-up study.

A national, population-based study of hip fracture epidemiology is reported following a previously published study for the same population to ascertain the accuracy of projected incidence rates. Results suggest a decline in age-specific incidence rates in females but not males, and an over-estimation of same in the previously reported projections. PURPOSE: Accurate estimates of hip fracture numbers and incidence rates are necessary for health service resource planning. A study in 2009 estimated projected figures for 2016 and 2026, for the Republic of Ireland. The purpose of the current study was to scrutinize these projections, based upon actual population and fracture statistics for the year 2016. METHODS: Population data was obtained from the Central Statistics Office, and hip fracture data was obtained from the National HIPE Database for the year 2016, using similar search criteria to the previously published study. Hip fracture numbers and incidence rates were calculated. The calculated figures were compared to projected figures from the previous report. RESULTS: Population projection models used in the 2009 paper accurately reflected the 2016 Irish population. For younger age groups, projected number of fractures closely estimated actual numbers, however, for those over 70 years of age, hip fracture projections overestimated the actual fracture numbers observed. There was a significant difference in the age-specific hip fracture rates in the over 70 s sub-groups compared to the published results of the 2009 study. CONCLUSIONS: Hip fracture numbers continue to increase, though the age-specific incidence rates in the Republic of Ireland are reducing in the older age groups. The previously published study over-estimated hip fracture projections for the year 2016 (and potentially 2026) due to a false assumption of a stable age-specific incidence rate in the Republic of Ireland, which is in fact reducing.

COVID-19 acute respiratory distress syndrome: A simulation study of the effects of combination therapy with tocilizumab and siltuximab.

AIMS: To assess the potential of interleukin-6 (IL-6) signalling blockade in the lung to treat SARS-CoV-2 infection via model-based simulation by exploring soluble IL-6 receptor (sIL-6R) sequestration by tocilizumab (TCZ) and IL-6 sequestration by siltuximab (SIL). METHODS: Literature values of IL-6, IL-6 antagonist SIL, sIL-6R, IL-6R antagonist TCZ and their respective binding constants were used to develop a model to predict the impact of treatment on IL-6 signalling. Models were used to generate simulated bronchoalveolar lavage fluid concentrations for normal subjects, subjects at risk of developing acute respiratory distress syndrome (ARDS), and subjects with ARDS under 4 conditions: without treatment; treatment with TCZ; treatment with SIL; and treatment with TCZ + SIL. RESULTS: With TCZ intervention, IL-6 levels are unaffected and sIL-6R is reduced somewhat below the Normal case. IL-6:sIL-6R complex only slightly decreased relative to the no-intervention case. With SIL intervention, sIL-6R levels are unaffected and IL-6 is greatly reduced below the Normal case. IL-6:sIL-6R complex is greatly decreased relative to the no-intervention case. With TCZ + SIL intervention, IL-6 and sIL-6R levels are reduced below the Normal case and achieve suppression equivalent to monotherapy results for their respective targets. IL-6:sIL-6R complex reduction is predicted to be greater than that achieved with monotherapy. This reflects sequestration of both components of the complex and the nonlinear binding equilibrium. CONCLUSION: Coadministration of both IL-6 and IL-6R sequestering products such as SIL and TCZ may be necessary to effectively treat COVID-19 patients who have or are at risk of developing ARDS.

The impact of the SARS-CoV-2 pandemic on referral characteristics in a national tertiary spinal injuries unit.

BACKGROUND: The SARS-CoV-2 pandemic has had profound implications on healthcare institutions. AIMS: This study aims to assess and compare referral patterns during COVID-19 to corresponding dates for the preceding 3 years (2017-2019), in order to preemptively coordinate the logistics of the surgical unit for similar future experiences. METHODS: Retrospective review for our institution, a national tertiary referral centre for spine pathology. Two distinct time-points were chosen to represent the varied levels of social restriction during the current pandemic: (i) study period 1 (SP1) from 11 November 2020 to 08 June 2020 represents a national lockdown, and (ii) study period 2 (SP2) from 09 June 2020 to 09 September 2020 indicates an easing of restrictions. Both periods were compared to corresponding dates (CP1: 11 March-08 June and CP2 09 June-09 September) for the preceding 3 years (2017-2019). Data collected included age, gender, and mechanism of injury (MOI) for descriptive analyses. MOIs were categorised into disc disease, cyclist, road-traffic-accident (RTA), falls < 2 m, falls > 2 m, malignancy, sporting injuries, and miscellaneous. RESULTS: All MOI categories witnessed a reduction in referral numbers during SP1: disc disease (-29%), cyclist (-5%), RTAs (-66%), falls < 2 m (-39%), falls > 2 m (-17%), malignancy (-33%), sporting injuries (-100%), and miscellaneous (-58%). Four of 8 categories (RTAs, falls < 2 m, malignancy, miscellaneous) showed a trend towards return of pre-lockdown values during SP2. Two categories (disc disease, falls > 2 m) showed a further reduction (-34%, -27%) during SP2. One category (sporting injuries) portrayed a complete return to normal values during SP2 while a notable increase in cyclist-related referrals was witnessed (+ 63%) when compared with corresponding dates of previous years. CONCLUSION: Spinal injury continues to occur across almost all categories, albeit at considerably reduced numbers. RTAs and falls remained the most common MOI. Awareness needs to be drawn to the reduction of malignancy-related referrals to dissuade people with such symptoms from avoiding presentation to hospital over periods of social restrictions.

Population Cellular Kinetics of Lisocabtagene Maraleucel, an Autologous CD19-Directed Chimeric Antigen Receptor T-Cell Product, in Patients with Relapsed/Refractory Large B-Cell Lymphoma.

BACKGROUND AND OBJECTIVES: Lisocabtagene maraleucel (liso-cel) is a CD19-directed, defined composition, 4-1BB chimeric antigen receptor (CAR) T-cell product administered at equal target doses of CD8+ and CD4+ CAR+ T cells. Large between-subject variability has been noted with CAR T-cell therapies; patient characteristics might contribute to CAR T-cell expansion variability. We developed a population cellular kinetic model to characterize the kinetics of the liso-cel transgene, via quantitative polymerase chain reaction assessment after intravenous infusion of liso-cel, and to understand covariates that might influence liso-cel kinetics in individual patients. METHODS: We employed nonlinear mixed-effects modeling to develop a population cellular kinetic model for liso-cel. The population cellular kinetic analysis was performed using 2524 post-infusion transgene observations from 261 patients with relapsed/refractory large B-cell lymphoma who were treated with a single dose of liso-cel in TRANSCEND NHL 001. Covariates for the analysis included baseline intrinsic factors such as age, baseline disease characteristics, and liso-cel and coadministration factors. RESULTS: Liso-cel cellular kinetics were well described by a piecewise model of cellular growth kinetics that featured lag, exponential growth, and biexponential decay phases. Population means (95% confidence interval) of lag phase duration, doubling time, time to maximum levels, initial decline half-life, and terminal half-life were 3.27 (2.71-3.97), 0.755 (0.667-0.821), 9.29 (8.81-9.70), 5.00 (4.15-5.90), and 352 (241-647) days, respectively. The magnitude of effect on liso-cel expansion metrics demonstrated that the covariate associations were smaller than the residual between-subject variability in the population. CONCLUSIONS: The covariates tested were not considered to have a meaningful impact on liso-cel kinetics. CLINICAL TRIAL REGISTRATION: NCT02631044.

The Use of Halo Gravity Traction in Severe, Stiff Scoliosis.

PURPOSE: The correction of severe, stiff scoliosis in children is challenging. One method used to reduce the risk is preoperative halo gravity traction (HGT). In this study, the authors sought to define the efficiency and safety of HGT and characterize the chronology of the correction seen. METHOD: A consecutive group of pediatric patients with severe spinal deformities was treated with HGT before definitive correction. A standard protocol with the daily addition of weight to 50% of body weight at 3 weeks was used. Traction remained in place until signs of impending neurological complication or 6 weeks, whichever was sooner. RESULTS: Twenty-four patients were included with a mean age of 11.8 years. The mean coronal deformity was 123 degrees, with a T1-L5 height of 234 mm. The mean duration of traction was 42 days with a mean improvement in height of 72 mm with 82% occurring over the first 3 weeks. Hundred percent of the angular and 98% of T1-L5 height correction was reached by 6 weeks.One patient showed early signs of a cranial nerve palsy prompting early surgery and 8 patients showed pin loosening, 1 of which required revision of their halo. One patient underwent a slower progression of traction because of transitory urinary disturbance. Following fusion, angular correction of the major curve was 49%. CONCLUSION: HGT is a safe treatment for severe, stiff scoliosis because it can respond to early signs of impending neurological impairment. The first 3 weeks of treatment, reaching 50% of body weight as a traction force accounts for 80% of correction, with the remaining 20% in the following 2 weeks. At least 4 weeks of traction is recommended when following this protocol.

Model-informed drug repurposing: A pharmacometric approach to novel pathogen preparedness, response and retrospection.

During a pandemic caused by a novel pathogen (NP), drug repurposing offers the potential of a rapid treatment response via a repurposed drug (RD) while more targeted treatments are developed. Five steps of model-informed drug repurposing (MIDR) are discussed: (i) utilize RD product label and in vitro NP data to determine initial proof of potential, (ii) optimize potential posology using clinical pharmacokinetics (PK) considering both efficacy and safety, (iii) link events in the viral life cycle to RD PK, (iv) link RD PK to clinical and virologic outcomes, and optimize clinical trial design, and (v) assess RD treatment effects from trials using model-based meta-analysis. Activities which fall under these five steps are categorized into three stages: what can be accomplished prior to an NP emergence (preparatory stage), during the NP pandemic (responsive stage) and once the crisis has subsided (retrospective stage). MIDR allows for extraction of a greater amount of information from emerging data and integration of disparate data into actionable insight.

Multimodal imaging of composite carbon fiber-based implants for orthopedic spinal fixation.

OBJECTIVE: As cancer treatments continue to improve, the incidence of spinal metastases and the need for surgical management of these with fixation procedures are growing rapidly. Traditionally metallic implants, composed of titanium alloy, have been used in surgical fixation of unstable or symptomatic vertebral metastases or traumatic injuries. Metallic implants, however, cause significant artifact on post-operative imaging, degrading image quality and limiting interpretation, and can also impair the planning and delivery of radiotherapy. Composite carbon fiber-based materials, such as carbon fiber-reinforced polyetheretherketone (PEEK), have been developed to overcome these issues and are now available for spinal fixation procedures. We aimed to review the multimodal imaging features of these new implants. MATERIALS AND METHODS: Current literature and a case example from our institution were used to describe the multimodal imaging characteristics and considerations of new carbon fiber-based spinal fixation implants. RESULTS: New carbon fiber-based spinal implants allow far greater visualization of surrounding structures on post-operative cross-sectional imaging, significantly improving diagnostic accuracy and precision of radiotherapy planning, and do not significantly absorb or scatter X-ray photons during radiotherapy delivery. There are, however, important surgical and radiologic considerations associated with the use of carbon fiber-based implants which radiologists must be aware of, such as implications for surgical planning and intra-operative fluoroscopic and post-operative plain radiographic imaging. CONCLUSION: The use of carbon fiber-based implants, rather than traditional metallic implants, for spinal fixation offers significant advantages for post-operative diagnostic imaging and radiotherapy planning and delivery.

Model-informed drug repurposing: Viral kinetic modelling to prioritize rational drug combinations for COVID-19.

AIM: We hypothesized that viral kinetic modelling could be helpful to prioritize rational drug combinations for COVID-19. The aim of this research was to use a viral cell cycle model of SARS-CoV-2 to explore the potential impact drugs, or combinations of drugs, that act at different stages in the viral life cycle might have on various metrics of infection outcome relevant in the early stages of COVID-19 disease. METHODS: Using a target-cell limited model structure that has been used to characterize viral load dynamics from COVID-19 patients, we performed simulations to inform on the combinations of therapeutics targeting specific rate constants. The endpoints and metrics included viral load area under the curve (AUC), duration of viral shedding and epithelial cells infected. Based on the known kinetics of the SARS-CoV-2 life cycle, we rank ordered potential targeted approaches involving repurposed, low-potency agents. RESULTS: Our simulations suggest that targeting multiple points central to viral replication within infected host cells or release from those cells is a viable strategy for reducing both viral load and host cell infection. In addition, we observed that the time-window opportunity for a therapeutic intervention to effect duration of viral shedding exceeds the effect on sparing epithelial cells from infection or impact on viral load AUC. Furthermore, the impact on reduction on duration of shedding may extend further in patients who exhibit a prolonged shedder phenotype. CONCLUSIONS: Our work highlights the use of model-informed drug repurposing approaches to better rationalize effective treatments for COVID-19.

Using in silico viral kinetic models to guide therapeutic strategies during a pandemic: An example in SARS-CoV-2.

AIMS: We propose the use of in silico mathematical models to provide insights that optimize therapeutic interventions designed to effectively treat respiratory infection during a pandemic. A modelling and simulation framework is provided using SARS-CoV-2 as an example, considering applications for both treatment and prophylaxis. METHODS: A target cell-limited model was used to quantify the viral infection dynamics of SARS-CoV-2 in a pooled population of 105 infected patients. Parameter estimates from the resulting model were used to simulate and compare the impact of various interventions against meaningful viral load endpoints. RESULTS: Robust parameter estimates were obtained for the basic reproduction number, viral release rate and infected-cell mortality from the infection model. These estimates were informed by the largest dataset currently available for SARS-CoV-2 viral time course. The utility of this model was demonstrated using simulations, which hypothetically introduced inhibitory or stimulatory drug mechanisms at various target sites within the viral life-cycle. We show that early intervention is crucial to achieving therapeutic benefit when monotherapy is administered. In contrast, combination regimens of two or three drugs may provide improved outcomes if treatment is initiated late. The latter is relevant to SARS-CoV-2, where the period between infection and symptom onset is relatively long. CONCLUSIONS: The use of in silico models can provide viral load predictions that can rationalize therapeutic strategies against an emerging viral pathogen.