Immune checkpoint inhibitors (ICIs) have dramatically modified the prognosis of several advanced cancers, however many patients still do not respond to treatment. Optimal results might be obtained by targeting cancer cell metabolism to modulate the immunosuppressive tumor microenvironment. Here, we identify sphingosine kinase-1 (SK1) as a key regulator of anti-tumor immunity. Increased expression of SK1 in tumor cells is significantly associated with shorter survival in metastatic melanoma patients treated with anti-PD-1. Targeting SK1 markedly enhances the responses to ICI in murine models of melanoma, breast and colon cancer. Mechanistically, SK1 silencing decreases the expression of various immunosuppressive factors in the tumor microenvironment to limit regulatory T cell (Treg) infiltration. Accordingly, a SK1-dependent immunosuppressive signature is also observed in human melanoma biopsies. Altogether, this study identifies SK1 as a checkpoint lipid kinase that could be targeted to enhance immunotherapy.

Dose banding is a recently suggested dosing method that uses predefined ranges (bands) of body surface area (BSA) to calculate each patient's dose by using a single BSA-value per band. Thus, drugs with sufficient long-term stability can be prepared in advance. The main advantages of dose banding are to reduce patient waiting time and improve pharmacy capacity planning; additional benefits include reduced medication errors, reduced drug wastage, and prospective quality control. This study compares dose banding with individual BSA dosing and fixed dose according to pharmacokinetic criteria. Three BSA bands were defined: BSA<1.7 m(2), 1.7 m(2)≤ BSA<1.9 m(2), BSA ≥ 1.9 m(2) and each patient dose was calculated based on a unique BSA-value per band (1.55, 1.80, and 2.05 m(2), respectively). By using individual clearance values of six drugs (cisplatin, docetaxel, paclitaxel, doxorubicin, irinotecan, and topotecan) from 1012 adult cancer patients in total, the AUCs corresponding to three dosing methods (BSA dosing, dose banding, and fixed dose) were compared with a target AUC for each drug. For all six drugs, the per cent variation in individual dose obtained with dose banding compared with BSA dosing ranged between -14% and +22%, and distribution of AUC values was very similar with both dosing methods. In terms of reaching the target AUC, there was no significant difference in precision between dose banding and BSA dosing, except for paclitaxel (32.0% vs 30.7%, respectively; P<0.05). However, precision was significantly better for BSA dosing compared with fixed dose for four out of six drugs. For the studied drugs, implementation of dose banding should be considered as it entails no significant increase in interindividual plasma exposure.

Aims Both rituximab and plasmapheresis can be associated in the treatment of immune‐mediated kidney diseases. The real impact of plasmapheresis on rituximab pharmacokinetics is unknown. The aim of this study was to compare rituximab pharmacokinetics between patients requiring plasmapheresis and others without plasmapheresis. Methods The study included 20 patients receiving one or several infusions of rituximab. In 10 patients, plasmapheresis sessions were also performed (between two and six sessions per patient). Rituximab concentrations were measured in blood samples in all patients and in discarded plasma obtained by plasmapheresis using an enzyme‐linked immunosorbent assay method. Data were analysed according to a population pharmacokinetic approach. Results The mean percentage of rituximab removed during the first plasmapheresis session ranged between 47 and 54% when plasmapheresis was performed between 24 and 72 h after rituximab infusion. Rituximab pharmacokinetics was adequately described by a two‐compartment model with first‐order elimination. Plasmapheresis had a significant impact on rituximab pharmacokinetics, with an increase of rituximab clearance by a factor of 261 (95% confidence interval 146–376), i.e. from 6.64 to 1733 ml h⁻¹. Plasmapheresis performed 24 h after rituximab infusion decreased the rituximab area under the curve by 26%. Conclusions Plasmapheresis removed an important amount of rituximab when performed less than 3 days after infusion. The removal of rituximab led to a significant decrease of the area under the curve. This pharmacokinetic observation should be taken into account for rituximab dosing, e.g. an additional third rituximab infusion may be recommended when three plasmapheresis sessions are performed after the first rituximab infusion.

Pharmacokinetic (PK) parameter estimation is a critical and complex step in the model‐informed precision dosing (MIPD) approach. The mapbayr package was developed to perform maximum a posteriori Bayesian estimation (MAP‐BE) in R from any population PK model coded in mrgsolve. The performances of mapbayr were assessed using two approaches. First, “test” models with different features were coded, for example, first‐order and zero‐order absorption, lag time, time‐varying covariates, Michaelis–Menten elimination, combined and exponential residual error, parent drug and metabolite, and small or large inter‐individual variability (IIV). A total of 4000 PK profiles (combining single/multiple dosing and rich/sparse sampling) were simulated from each test model, and MAP‐BE of parameters was performed in both mapbayr and NONMEM. Second, a similar procedure was conducted with seven “real” previously published models to compare mapbayr and NONMEM on a PK outcome used in MIPD. For the test models, 98% of mapbayr estimations were identical to those given by NONMEM. Some discordances could be observed when dose‐related parameters were estimated or when models with large IIV were used. The exploration of objective function values suggested that mapbayr might outdo NONMEM in specific cases. For the real models, a concordance close to 100% on PK outcomes was observed. The mapbayr package provides a reliable solution to perform MAP‐BE of PK parameters in R. It also includes functions dedicated to data formatting and reporting and enables the creation of standalone Shiny web applications dedicated to MIPD, whatever the model or the clinical protocol and without additional software other than R.