Wound healing consists of a sequence of biological processes often grouped into different stages. Interventions applied to accelerate normal wound healing must take into consideration timing with respect to wound healing stages in order to maximize treatment effectiveness. Macrophage polarization from M1 to M2 represents a transition from the inflammatory to the proliferative stage of wound healing. Accelerating this transition may be an effective way to accelerate wound healing; however, it must be induced at the appropriate time. We search for an optimal spatio-temporal regime to apply wound healing treatment in a mathematical model of wound healing. In this work we show that to maximize effectiveness, treatment must not be applied too early or too late with respect to peak inflammation. We also show that the effective spatial distribution of treatment depends on the heterogeneity of the wound surface. In conclusion, this research provides a possible optimal regime of therapy that focuses on macrophage activity and a hypothesis of treatment outcome to be tested in future experiments. Finding optimal regimes for treatment application is a first step toward the development of intelligent algorithms for wound treatment that minimize healing time.

Background

Wound healing involves careful coordination among various cell types carrying out unique or even multifaceted functions. The abstraction of this complex dynamic process into four primary wound stages is essential to the study of wound care for timing treatment and tracking wound progression. For example, a treatment that may promote healing in the inflammatory stage may prove detrimental in the proliferative stage. Additionally, the time scale of individual responses varies widely across and within the same species. Therefore, a robust method to assess wound stages can help advance translational work from animals to humans.

Results

In this work, we present a data-driven model that robustly identifies the dominant wound healing stage using transcriptomic data from biopsies gathered from mouse and human wounds, both burn and surgical. A training dataset composed of publicly available transcriptomic arrays is used to derive 58 shared genes that are commonly differentially expressed. They are divided into 5 clusters based on temporal gene expression dynamics. The clusters represent a 5-dimensional parametric space containing the wound healing trajectory. We then create a mathematical classification algorithm in the 5-dimensional space and demonstrate that it can distinguish between the four stages of wound healing: hemostasis, inflammation, proliferation, and remodeling.

Conclusions

In this work, we present an algorithm for wound stage detection based on gene expression. This work suggests that there are universal characteristics of gene expression in wound healing stages despite the seeming disparities across species and wounds. Our algorithm performs well for human and mouse wounds of both burn and surgical types. The algorithm has the potential to serve as a diagnostic tool that can advance precision wound care by providing a way of tracking wound healing progression with more accuracy and finer temporal resolution compared to visual indicators. This increases the potential for preventive action.

The capacity of machine-learning methods to handle large and complex datasets makes them suitable for applications in precision medicine. Current methods automate data analysis and predict physiological outcomes of patients with various types of clinical data to inform treatment strategies. In this perspective, we propose ways in which machine learning can be leveraged even further to advance methods of optimizing patient treatment. Namely, machine learning can be used to expand applications of feedback control to direct the response of complex biological systems predictably and automatically. This paves the way for highly sophisticated treatments that continuously adapt to an individual patient's response. The elements of control that can be improved using machine learning include sensor data analysis, modeling, and methods of reconfiguring the control algorithm “on the fly.” We discuss the control challenges unique to the analysis/control of complex biological systems, existing work, and areas that remain underdeveloped.

Finding the optimal treatment strategy to accelerate wound healing is of utmost importance, but it presents a formidable challenge owing to the intrinsic nonlinear nature of the process. We propose an adaptive closed-loop control framework that incorporates deep learning, optimal control and reinforcement learning to accelerate wound healing. By adaptively learning a linear representation of nonlinear wound healing dynamics using deep learning and interactively training a deep reinforcement learning agent for tracking the optimal signal derived from this representation without the need for intricate mathematical modelling, our approach has not only successfully reduced the wound healing time by 45.56% compared to the one without any treatment, but also demonstrates the advantages of offering a safer and more economical treatment strategy. The proposed methodology showcases a significant potential for expediting wound healing by effectively integrating perception, predictive modelling and optimal adaptive control, eliminating the need for intricate mathematical models.

Shear-induced conformational changes of von Willebrand factor (VWF) play an important role in platelet activation. A novel approach describing VWF unfolding on the platelet surface under dynamic shear stress is proposed. Cumulative effect of dynamic shear on platelet activation via conformational changes of VWF is analysed. The critical condition of shear-induced platelet activation is formulated. The explicit expression for the threshold value of cumulative shear stress as a function of VWF multimer size is derived. The results open novel prospects for pharmacological regulation of shear-induced platelet activation through control of VWF multimers size distribution.

Bioelectronic devices can provide an interface for feedback control of biological processes in real-time based on sensor information tracking biological response. The main control challenges are guaranteeing system convergence in the presence of saturating inputs into the bioelectronic device and complexities from indirect control of biological systems. In this paper, we first derive a saturated-based robust sliding mode control design for a partially unknown nonlinear system with disturbance. Next, we develop a data informed model of a bioelectronic device for in silico simulations. Our controller is then applied to the model to demonstrate controlled pH of a target area. A modular control architecture is chosen to interface the bioelectronic device and controller with a bistable phenomenological model of wound healing to demonstrate closed-loop biological treatment. External pH is regulated by the bioelectronic device to accelerate wound healing, while avoiding chronic inflammation. Our novel control algorithm for bioelectronic devices is robust and requires minimum information about the device for broad applicability. The control architecture makes it adaptable to any biological system and can be used to enhance automation in bioengineering to improve treatments and patient outcomes.

Macrophages can exhibit pro-inflammatory or pro-reparatory functions, contingent upon their specific activation state. This dynamic behavior empowers macrophages to engage in immune reactions and contribute to tissue homeostasis. Understanding the intricate interplay between macrophage motility and activation status provides valuable insights into the complex mechanisms that govern their diverse functions. In a recent study, we developed a classification method based on morphology, which demonstrated that movement characteristics, including speed and displacement, can serve as distinguishing factors for macrophage subtypes. In this study, we develop a deep learning model to explore the potential of classifying macrophage subtypes based solely on raw trajectory patterns. The classification model relies on the time series of x-y coordinates, as well as the distance traveled and net displacement. We begin by investigating the migratory patterns of macrophages to gain a deeper understanding of their behavior. Although this analysis does not directly inform the deep learning model, it serves to highlight the intricate and distinct dynamics exhibited by different macrophage subtypes, which cannot be easily captured by a finite set of motility metrics. Our study uses cell trajectories to classify three macrophage subtypes: M0, M1, and M2. This advancement holds promising implications for the future, as it suggests the possibility of identifying macrophage subtypes without relying on shape analysis. Consequently, it could potentially eliminate the necessity for high-quality imaging techniques and provide more robust methods for analyzing inherently blurry images.

The ability to deliver drugs with precise dosages at specific time points can significantly improve disease treatment while reducing side effects. Drug encapsulation for gradual delivery has opened the doors for a superior treatment regimen. To expand on this ability, programming bioelectronic devices to deliver small molecules enables ad-hoc personalized therapeutic profiles that are more complex than gradual release. Here, a wearable bioelectronic device with an integrated electrophoretic ion pump that affords on-demand drug delivery with precise dose control is introduced. Delivery of fluoxetine to wounds in mice result in a 27.2% decrease in the macrophage ratio (M1/M2) and a 39.9% increase in re-epithelialization, indicating a shorter inflammatory phase and faster overall healing. Programmable drug delivery using wearable bioelectronics in wounds introduces a broadly applicable strategy for the long-term delivery of a prescribed treatment regimen with minimal external intervention.