A Computational Growth and Remodeling Framework for Adaptive and Maladaptive Pulmonary Arterial Hemodynamics

Introduction:: Pulmonary arterial hypertension (PAH) is a severe disease characterized by progressive narrowing of the distal pulmonary arteries (PA), which leads to an increased right ventricular afterload that can cause heart failure. Hemodynamic changes, such as increased blood flow to the lungs by an atrial or ventricular septal defect, can initiate disease progression, but the time course of changes are poorly understood. Changes to mechanobiological cues, including wall shear stress (WSS) and intramural stress (IMS), play key a role in altering cell phenotypes that drive maladaptive remodeling of the vessels¹. Computational models of the mechanical environment are necessary for understanding the effects of WSS and IMS, as they are difficult to directly measure. However, simulation of these cues has focused on single time points in disease and are not linked with functional changes driving the transition from adaptive to maladaptive remodeling. Here, we generated a predictive computational model for overcoming these challenges with the goal of identifying the functional stimuli necessary for PAH initiation and progression.


Materials and Methods:: To simulate hemodynamics of the pulmonary arterial tree in murine models, we used a Strahler diameter-defined morphometric tree with a Poiseuille flow assumption. Morphometric data on connectivity were adapted from the literature to provide the network of blood vessels for flow simulations2. Diameters for vessels of smaller orders were extrapolated from that of the branch pulmonary arteries and fit to match the murine pulmonary vascular resistance previously measured by catheterization3. Biomechanical data from biaxial testing were used to generate the material behavior of the branch pulmonary arteries4. We then used a constrained mixture theory growth and remodeling framework to model the loaded PA tree by averaging hemodynamic loads from vessels within the same order and optimizing the mechanical behavior of each order to allow for the homeostatic geometries to generate a resistance that is the same as that prescribed⁵. We then simulated adaptive and maladaptive remodeling scenarios over time using mass kinetics equations that include production and removal of load bearing constituents, including collagen, elastin, and smooth muscle. Finally, we modeled modulation of the maladaptive effects to assess their impact on outcomes of clinical interest.


Results, Conclusions, and Discussions:: To illustrate adaptive remodeling, we simulated an 80% increase in inlet flow, as found in left-to-right shunts, which led to dilatation of all vessels of the tree to reduce WSS back towards its homeostatic levels, though WSS values in small vessels remained 6-fold more elevated than those in the proximal vessels. PA tree resistance was decreased, which allowed for near-normal levels of proximal PA pressure as expected for an adaptive response. For maladaptation, multiple functional changes were sequentially added to understand the required modeling features to generate the pressure changes found in PAH. At minimum, smooth muscle hyperproliferation and stiffening were necessary to generate moderate pressure increases, while additional loss of regulation for mechanobiological mass kinetics and active tone further increased disease severity. As PAH can be initiated by hemodynamic changes, we trialed a functional form for smooth muscle hyperproliferation that depended on WSS, which allowed for demonstration of a positive feedback loop driven by increased flow that led to geometric changes similar to those observed in clinical patients. To simulate interventions that modulate hyperproliferation and sensitivity to mechanobiological loading, we inhibited these effects after initiation and progression of the disease and found that changing these effects, even after a period of maladaptation, could decrease pulmonary arterial pressures and restore more normal geometries.


In this work, we developed an approach to simulate the mechanobiological cues experienced across different vessel diameters with initiation and evolution of disease driven by hemodynamic perturbations and altered cellular behavior. Our model predicted distinct mechanical behavior and loading time courses in different sized vessels, which could contribute to long term differences in outcome between the large and small PAs. We further identified that certain combinations of functional changes to vascular wall constituents led to varying degrees of PAH severity and diverging WSS changes for different sized vessels. These methods offer promise for in silico testing of PAH treatments and linking biological pathways to organ-scale disease features.


Acknowledgements (Optional): : The authors would like to acknowledge the Parker B. Francis Fellowship and a training fellowship from the National Heart, Lung and Blood Institute (1T32HL098049) for funding for this work.

References (Optional): : [1] Yang, W., et al. Biomech model mechanobiol. 18.3(2019):779-796.
[2] Jiang, Z.L., et al. J Appl Physiol. 76.2(1994):882-892.
[3] Champion, H.C., et al. Am J Physiol. 278.1(2000):H8-H15.
[4] Ramachandra, A.B., et al. J Biomech. 84(2019):18-26.
[5] Humphrey, J.D., J Elast. 145(2021):49-75.