Non-clinical platform to predict exposure to maternal medicines via breastfeeding

Korte inhoud:Research shows the importance of breastfeeding for mother and infant. Human milk is perfectly tailored to support the growth and development of the baby. In addition, human milk contains antibodies against bacteria and viruses. Breastfeeding has a protective effect against infections, including otitis media, respiratory and gastrointestinal infections. There is also a smaller risk of developing diabetes, leukaemia or obesity. In addition, breastfeeding has many benefits for the mother, such as a lower risk of breast and ovarian cancer. Finally, breastfeeding also has socio-economic benefits. Human milk is not only environmentally friendly, but also provides an indirect economic benefit through the health benefits for mother and child. The World Health Organization recommendation is exclusive breastfeeding for 6 months and to combine breastfeeding with complementary foods for at least 2 years. At the same time, more than half of mothers need medication in the first weeks and months after giving birth. However, there is a huge lack of knowledge about the use of medicines in combination with breastfeeding. For many medicines, it is not clear whether they pose safety risks for the infant while taken during breastfeeding. Although many medicines have limited transfer into human milk and are poorly absorbed by the infant, there is often no evidence to support this. The package leaflet of most medicines contains little information about the concentrations in human milk and often refers to the doctor or pharmacist for advice. As a result, mothers are using medicines off-label potentially exposing the infant to (unknown) health risks, decide to stop breastfeeding or avoid taking medication. Although clinical trials can generate information, they also pose many ethical and practical challenges. Non-clinical methods and models (in vitro & in silico) can play an important role in rapidly generating data, as well as supporting clinical studies. This doctoral research fits within the "Innovation Medicines Initiative (IMI) project ConcePTION" (https://www.imi-conception.eu/). ConcePTION was carried out by a European consortium of public and private partners, with the aim of building a platform to generate information on the safety of medicines during pregnancy and breastfeeding. The overarching aim of this PhD thesis was to develop a non-clinical platform to predict the transfer of medicines into human milk. Knowledge about medicine concentrations in human milk, in combination with information about absorption, distribution, metabolism, excretion (ADME) and toxicity of medicines in infants, can be used to evaluate the safety of maternal medicines during lactation. The first objective was to develop and characterize an in vitro model for the human blood-milk barrier (Chapter 3). The underlying hypothesis was that human mammary epithelial cells are the major barrier to medicine secretion in human milk. Several human cell lines, as well as primary mammary epithelial cells were investigated. Primary human mammary epithelial cells (hMECs) were selected as the best model to evaluate medicine transfer into human milk. This model was extensively characterized, including morphology, barrier function and membrane transport protein expression. Next, this in vitro model was applied to determine permeability of the blood-milk barrier for a set of model medicines (Chapter 4). A regression model was developed based on the permeability representative of secretion towards human milk, also considering the unionized and unbound fractions in plasma and human milk. This model provides an accurate prediction of the observed milk-to-plasma ratio (M/P ratio) in the current dataset. In addition, an in vitro model based on Göttingen Minipig was also developed, considering that an animal model was developed based on the Göttingen Minipig within the IMI project ConcePTION. The combination of these models may provide mechanistic insights into the blood-milk barrier, for example about the extrapolation of in vitro data towards predictions about the in vivo concentrations. An in vitro model can be used to support decisions around medication use during lactation. In the future, additional experiments can clarify the expression of membrane transport proteins and confidence in the predictions can be gained through external validation. In the next part of this PhD thesis, the objective was to develop Physiologically-based pharmacokinetic (PBPK) models to predict in vivo medicine exposure in human milk (Chapters 5 & 6). The hypothesis was that these in silico predictions can be used to simulate medicine concentration in human milk when clinical data is lacking. PBPK models were developed in the Open Systems Pharmacology software for ten medicines, including amoxicillin, metformin and valproic acid. The first step was to develop and verify PBPK models for adults that are not breastfeeding, as clinical data is available in this population. The PBPK models were able to predict systemic exposure within 2-fold error. Subsequently, these models were extended to incorporate physiological changes during lactation. Here, a breast compartment, including human milk, was added to the standard PBPK structure. Initially, the Koshimichi method was used, predicting the bidirectional clearance between plasma and human milk based on physicochemical parameters (Chapter 5). Inclusion of this method into PBPK models gave a reasonable prediction of exposure in human milk for 80 % of the medicines. As an alternative method, the in vitro data (Chapters 3 & 4) was integrated into the PBPK models (Chapter 6). The advantage of this method is that it considers active medicine transport across the blood-milk barrier. The feasibility of this method was demonstrated, but the need for medicine-specific factors for permeability to obtain a good prediction shows that further refinement of the PBPK models (e.g., consideration of human milk pH and fraction ionized) and in vitro models (e.g., expression of membrane transport proteins) is needed. To assess how good non-clinical methods and models (in vitro models and PBPK models) are at predicting in vivo exposure of the patient to medicines in human milk, clinical data is used. Data can be obtained from literature, but there is a need for standardization of study design and reporting of the results. The hypothesis was that a generic protocol for clinical lactation studies in collaboration with UZ Leuven would allow to generate more reliable data. The feasibility of this study was verified through a case study for sildenafil and bosentan (Chapter 7). The patient was taking sildenafil (20 mg, 3x/day) and bosentan (125 mg, 2x/day) because of pulmonary arterial hypertension and was breastfeeding her 21-month-old infant. The concentrations of both medicines in human milk were low, especially compared to the maternal dose or a dose used in infants for therapeutic indications. Moreover, the mother reported no health problems in the child. The lessons learned from this case study were applied to establish a generic study protocol. The UmbrelLACT study allows evaluation of medicine concentration in plasma and human milk from breastfeeding mothers. In addition, safety in the infant is checked by means of a questionnaire, and (optionally) a plasma sample. This study directly generates information on medicine safety but the results can also be used to further evaluate non-clinical methods. This doctoral research contributes to the development of a non-clinical platform (in vitro and PBPK models) to investigate exposure of the infant to medicines via breastfeeding. Awaiting further refinement of the models, the relevance of a non-clinical platform was demonstrated. One of the important advantages of a nonclinical platform is that it can also be applied in an early stage of drug development and provide input to the package leaflet. Non-clinical methods and models can contribute to the field of pharmacotherapy during lactation and support risk assessment for medicines during lactation so that lactating women no longer must be uncertain about the safety of maternal medicines for their infants.

Jaar van publicatie:2025

Toegankelijkheid:Embargoed