Understanding the effects of oxytocin receptor variants on OXT-OXT receptor binding: A mathematical model.

Approximately half of U.S. women giving birth annually receive Pitocin, the synthetic form of oxytocin (OXT), yet its effective dose can vary significantly. This variability presents safety concerns due to unpredictable responses, which may lead to adverse outcomes for both mother and baby. To address the need for improved dosing, we developed a data-driven mathematical model to predict OXT receptor (OXTR) binding. Our study focuses on five prevalent OXTR variants (V45L, P108A, L206V, V281M, and E339K) and their impact on OXT-OXTR binding dynamics in two distinct cell types: human embryonic kidney cells (HEK293T), commonly used in experimental systems, and human myometrial smooth muscle cells, containing endogenous OXTR. We parameterized the model with cell-specific OXTR surface localization measurements. To strengthen the robustness of our study, we conducted a comprehensive meta-analysis of OXT- OXTR binding, enabling parameterization of our model with cell-specific OXT-OXTR binding kinetics (myometrial OXT-OXTR K d = 1.6 nM, kon = 6.8 × 10 5 M -1 min -1 , and koff = 0.0011 min -1 ). Our meta-analysis revealed significant homogeneity in OXT-OXTR affinity across experiments and species with a K d = 0.52 - 9.32 nM and mean K d = 1.48 ± 0.36 nM. Our model achieves several valuable insights into designing dosage strategies. First, we predicted that the OXTR complex reaches maximum occupancy at 10 nM OXT in myometrial cells and at 1 µM in HEK293T cells. This information is pivotal for guiding experimental design and data interpretation when working with these distinct cell types, emphasizing the need to consider effects for specific cell types when choosing OXTR-transfected cell lines. Second, our model recapitulated the significant effects of genetic variants for both experimental and physiologically relevant systems, with V281M and E339K substantially compromising OXT-OXTR binding capacity. These findings suggest the need for personalized oxytocin dosing based on individual genetic profiles to enhance therapeutic efficacy and reduce risks, especially in the context of labor and delivery. Third, we demonstrated the potential for rescuing the attenuated cell response observed in V281M and E339K variants by increasing the OXT dosage at specific, early time points. Cellular responses to OXT, including Ca 2+ release, manifest within minutes. Our model indicates that providing V281M- and E339K-expressing cells with doubled OXT dose during the initial minute of binding can elevate OXT-OXTR complex formation to levels comparable to wild-type OXTR. In summary, our study provides a computational framework for precision oxytocin dosing strategies, paving the way for personalized medicine.

A convolutional neural network to characterize mouse hindlimb foot strikes during voluntary wheel running.

Voluntary wheel running (VWR) is widely used to study how exercise impacts a variety of physiologies and pathologies in rodents. The primary activity readout of VWR is aggregated wheel turns over a given time interval (most often, days). Given the typical running frequency of mice (∼4 Hz) and the intermittency of voluntary running, aggregate wheel turn counts, therefore, provide minimal insight into the heterogeneity of voluntary activity. To overcome this limitation, we developed a six-layer convolutional neural network (CNN) to determine the hindlimb foot strike frequency of mice exposed to VWR. Aged female C57BL/6 mice (22 months, n = 6) were first exposed to wireless angled running wheels for 2 h/d, 5 days/wk for 3 weeks with all VWR activities recorded at 30 frames/s. To validate the CNN, we manually classified foot strikes within 4800 1-s videos (800 randomly chosen for each mouse) and converted those values to frequency. Upon iterative optimization of model architecture and training on a subset of classified videos (4400), the CNN model achieved an overall training set accuracy of 94%. Once trained, the CNN was validated on the remaining 400 videos (accuracy: 81%). We then applied transfer learning to the CNN to predict the foot strike frequency of young adult female C57BL6 mice (4 months, n = 6) whose activity and gait differed from old mice during VWR (accuracy: 68%). In summary, we have developed a novel quantitative tool that non-invasively characterizes VWR activity at a much greater resolution than was previously accessible. This enhanced resolution holds potential to overcome a primary barrier to relating intermittent and heterogeneous VWR activity to induced physiological responses.