Developmental Origins of Mechanical Homeostasis in the Aorta

Abstract

Background.

Mechanical homeostasis promotes proper aortic structure and function. Pathological conditions may arise, in part, from compromised or lost homeostasis. There is thus a need to quantify the homeostatic state and when it emerges. Here we quantify changes in mechanical loading, geometry, structure, and function of the murine aorta from the late prenatal period into maturity.

Results.

Our data suggest that a homeostatic set-point is established by postnatal day P2 for the flow-induced shear stress experienced by endothelial cells; this value deviates from its set-point from P10-P21 due to asynchronous changes in mechanical loading (flow, pressure) and geometry (radius, wall thickness), but is restored thereafter consistent with homeostasis. Smooth muscle contractility also decreases during this period of heightened matrix deposition but is also restored in maturity. The pressure-induced mechanical stress experienced by intramural cells initially remains low despite increasing blood pressure, then increases while extracellular matrix accumulates.

Conclusions.

These findings suggest that cell-level mechanical homeostasis emerges soon after birth to allow mechanosensitive cells to guide aortic development, with deposition of matrix after P2 increasingly stress shielding intramural cells. The associated tissue-level set-points that emerge for intramural stress can be used to assess and model the aorta that matures biomechanically by P56.

Introduction

The vascular tree forms early in development and is initially directed by genetic programming¹. As blood pressure and flow initiate and increase, the arteries continue to develop, in part, in response to these hemodynamic stimuli². This mechanosensitivity enables the wall to optimize biomechanical properties to facilitate efficient cardiovascular function over a lifetime of changing conditions. The lumen, for example, is controlled acutely by vasodilatation or vasoconstriction and chronically by remodeling of the wall in vaso-altered states^3,4. Importantly, endothelial cell (EC) mediated vasodilatation occurs through a flow-induced shear stress modulation of vasodilators, such as nitric oxide, which causes underlying smooth muscle cells (SMCs) to relax. Because of the initially isochoric changes in wall thickness upon dilatation, changes in flow also alter mean circumferential wall stress, which is proportional to the distending pressure and ratio of luminal radius to thickness. Wall thickness, and thus wall stress, appear to be controlled through a mechano-regulated deposition and degradation of extracellular matrix (ECM), mainly by intramural cells. Similar to the existence of a homeostatic set-point for flow-induced shear stress that is regulated by the ECs⁵, it has long been thought that a set-point exists for the wall stress that is sensed by the intramural cells though borne largely by the ECM^6,7.

Despite significant research on vascular development, cell biology and mechanobiology, and responses in health and disease in maturity^8–11, a detailed understanding of the onset of mechanical homeostasis and its relationship to both the developing vascular architecture and the dramatic changes that occur in mechanical stimuli (pressure, flow, and somatic growth) remains unclear. Cell culture studies provide valuable information on cellular responses to diverse loads and ECM properties (e.g.,^5,12), but cannot be translated directly to the physiological setting. Although many papers over the years have addressed tissue-level changes in development (e.g., from¹³ to¹⁰), lack of consistent information at a sufficient number of developmental ages has precluded overall understanding. In this paper, we seek to identify when homeostatic set-points are established for flow-induced luminal shear stress and pressure-induced intramural stress during aortic development and maturation. We also seek to correlate the emergence of these set-points with the evolving geometry and composition of the aortic wall as well as with EC, SMC, and wall function. To this end, we combine in vivo cardiovascular measurements (cardiac output, central blood pressures), quantitative polymerase chain reaction and immuno-histological studies (wall composition), ex vivo vascular biomechanical phenotyping (cell-dominated function), and nonlinear mechanics (ECM-dominated properties) to quantify development of the thoracic aorta in healthy wild-type mice through six different ages, ranging from the late fetal period well into maturity. We submit that the resulting data will aid in understanding many normal aortic adaptations and serve as a foundation for understanding better both congenital disorders and early versus late onset aortic diseases that cause or are a consequence of compromised mechanical homeostasis.

Results

Wall stress and stiffness evolve at different rates in circumferential and axial directions. Data from standard biaxial mechanical testing reveal complex evolutions of the overall pressure-diameter and axial force-length (structural) behaviors of the aorta from embryonic day E18.5 through the early postnatal period (P2), prior to weaning (P10 and P21), and finally near (P42) and well into (P98) maturity (Fig 1A, Fig 2A), with distinct behaviors manifesting at E18.5 and P2 compared with those at and after P10. The near monotonic increase in blood pressure (Fig 1B) was not followed directly by the time course of changes in loaded inner radius and wall thickness, or their ratio (Fig 1C), particularly at P10 when loaded wall thickness reached an unexpectedly high value (Table S1). These different time-courses in mechanical loading and associated geometry resulted in passive values of mean circumferential wall stress (${\sigma }_{\theta }=Pa/h$, where P is blood pressure, a is luminal radius, and h is wall thickness) that evolved sigmoidally (Fig 1D), with early values (~20–25 kPa) dramatically lower than mature values (~225–250 kPa) and an evolution that lagged that of the driving pressure (with stress reaching its maximal rate of change at P25.2 versus P14.4 for pressure). Findings were similar in the axial direction, though with the near monotonic increase in thorax length (Fig 2B) preceded slightly by that for the in vivo value of axial length (Fig 2C), which exhibited its maximum rate of change at P10.7. Passive values of the mean axial stress (${\sigma }_z=f/\pi h\left(2a+h\right)$ where f is axial force) also evolved sigmoidally (Fig 2D), with its maximum rate of change (at P21.8) lagging that of thorax lengthening (at P14.4).

Figure 1. Evolving load, geometry, properties, and passive mechanics reveal key transitions.

A: Pressure-Outer diameter relationships of descending thoracic aortas from E18.5-P98, where red star-symbols mark near in vivo conditions at mean arterial blood pressure (MAP) and in vivo axial stretch (see Table S1) for each age group. B: MAP data of different age groups; early ages (denoted by x) were estimated based on data presented by Le et al⁵⁴. C: The ratio of loaded inner radius (a) and loaded wall thickness (h) displayed an unexpected decrease ^~P8.2 (red vertical dash-dotted line which indicates the calculated time at which the rate of change is zero). D: Evolution of the calculated total circumferential wall stress $\left({\sigma }_{\theta }=Pa/h\right)$ revealed a sigmoidal behavior with near a constant value between E18.5-P10 (see magnified box). E: Temporal evolution of circumferential stiffness. Vertical black dotted lines in B,D,E indicate the time at which the maximum rate occurs. Gray-shaded areas in B-E show the 95% prediction intervals of the fitted function of the mean.

Figure 2.

A: Axial force-stretch relationships of descending thoracic aortas from E18.5-P98 old C57BL/6 mice. B,C: Axial somatic lengthening (thorax length) from P0.5-P84 and the loaded aortic length $\left(l={\lambda }_{z,iv}L\right)$ from E18.5-P98. D: Evolution of the calculated total axial stress displayed a similar sigmoidal behavior as the circumferential stress but with a slight decrease after P42 that was not observed circumferentially. E: Axial stiffness evolved significantly faster (max rate ^~P15.6) and higher compared to circumferential stiffness (max rate ^~P29.5, Fig 1).

Importantly, the biaxial mechanical data were well described by a single nonlinear constitutive relation at each of the six ages (Table S2), which allowed consistent calculation of associated values of the evolving biaxial material stiffness. Axial (Fig 2E) material stiffness evolved at a rate (maximal at P15.6) similar to that for thorax length (at P14.4), and reached mature values (at P28.7) prior to circumferential stiffness (Fig 1E; maximal rate at P29.5, maturity at P52.5). Hence, the aortic wall is not only strongly anisotropic, it reaches mature values of Cauchy stress and material stiffness at different rates in the circumferential and axial directions.

Early cell-dominated wall mechanics gives way rapidly to matrix-dominated mechanics. The sigmoidal time-courses for wall stress and material stiffness reflected well the evolving aortic composition that was dominated early on by a collection of intramural cells but was eventually dominated structurally by ECM (Fig 3A-R, Fig S1). Intramural cell density peaked near P10 (Fig 3A’), reflecting early cell growth with marked increases in ECM deposition thereafter. Biaxial wall stress was nearly constant during the period E18.5 to P2, with only mild increases through P10, suggesting that the early values may represent a homeostatic cell-level set-point. As wall stress increased, cells began to deposit more ECM, thus allowing this matrix to bear the increasing mechanical loads due to both somatic growth and rising blood pressure. Whereas elastin transcription and deposition peaked before P10 (Fig 3B’,C’), the rate of fibrillar collagen production was highest early in the transitional period defined by the sigmoidal mechanical time-course (~P10 to P42), with histologically measured values of collagen peaking in the media at P15.6 and in the adventitia at P23.2 (Fig 3D’, Fig S1). These differences in the time of peak production of ECM components could suggest overlapping roles of genetically programmed and stress-mediated synthesis, though elastin production appears to be stimulated by the pulsatility of hemodynamic loading on the wall⁷. Regardless, collagen continued to accumulate within the adventitia after P42, up to the time (~P56) at the aortic wall appeared to be largely mature biomechanically (Fig 1D-E, Fig 2D-E). This continued, late production of adventitial collagen could be influenced, in part by, non-mechanical cues.

Figure 3. Evolving composition reflects a transition from cell- to tissue-dominated mechanics.

A-F: Histological images of descending thoracic aortas from E18.5 to P98 stained with a fluorescent antibody to quantify smooth muscle 𝛼-actin (SM𝛼A) and counterstained with 4′,6-diamidino-2-phenylindole (DAPI) to quantify SMC nuclei; (G-L), Verhoeff-Van Gieson (VVG) to quantify elastin, and (M-R) picro-sirius red (PSR) to quantify medial and adventitial collagen. A’: Average number of SMC nuclei per cross section normalized to the unloaded circumference of P2-P98 aortas. B’: Temporal trends of RNA expression by qRT-PCR for markers of the ECM (Eln, Col3a1, Col1a1) and SMC (Acta2). C’,D’: Layer-specific histological quantification (elastin, medial and adventitial collagen area fractions, AF), consistent with B’. Note that elastin is deposited before (^~P5.7) medial (with max rate ^~P15.6) and adventitial (with max rate ^~P23.2) collagen. Vertical red dash-dotted lines in A’,C’ indicate the calculated day at which the rate of change is zero. Vertical black dotted lines in D’ indicate the day at which the maximum rate occurs. Gray-shaded areas in A’,C’,D’ show the 95% prediction intervals of the fitted function of the mean. All images are taken at the same magnification.

Functional elastin matures under low wall stress. Although elastin production begins in the late fetal period^14,15 and is increasingly evident under standard light microscopy at P2 and especially P10 (Fig 3G-L, Fig S1), the biaxial biomechanical data and constitutive modeling revealed that the elastin only became functionally mature near P21. That is, the well-known S-shaped diameter-pressure curve of the mature murine aorta began to manifest at P10 (Fig 1A), suggesting that elastic fibers were beginning to bear load, and the constitutive parameter associated with the elastin-dominated mechanical behavior rapidly matured between P10 and P21 (Fig S1J, Table S2). Importantly, because of the sigmoidal evolution of biaxial wall stress, the deposition and functional maturation of elastic fibers occurred while wall stress was well below mature values (which are >200 kPa passive), namely ~31 kPa at P10 and ~79 kPa at P21 (Table S1). Elastin deposition was thus maximal around the same time as the aforementioned increase in wall thickness and decrease in the in vivo axial stretch around P10 (Table S1). By contrast, collagen continued to be deposited and incorporated within the wall well into maturity and thus at elevated values of wall stress.

Early, rapid matrix accumulation associates with a transient functional modulation of SMC phenotype. Vasoregulation of the aorta contributes to hemodynamic control, but also impacts wall stress, which is an important stimulus to the highly mechano-sensitive intramural cells. The ability of SMCs to vasoconstrict the aorta as a function of maturation revealed an unexpected transient (Fig 4A,B; Table S4): the ability of the wall to constrict in response to exogenous phenylephrine (PE) was surprisingly high at P2, reduced significantly from P10-P21, but largely restored by P42. This trend in the evolution of SMC contractility in response to PE was preserved when normalizing contractility to the age-specific SMC density (Fig 4C) and remained when stimulating SMC contractility through membrane depolarization (i.e., high KCl), though to a slightly lower degree (Table S4). Importantly, the period of reduced contractile strength coincided with the aforementioned period of rapid accumulation of ECM within the media. Indeed, a strong negative correlation emerged between SMC contractile strength and the rate of accumulation of medial collagen (Fig 4D), consistent with a transient phenotypic modulation of the SMCs away from contractile during periods of high ECM synthesis. Interestingly, the onset of rapid matrix accumulation (with increased medial collagen accumulation ~P15.6) also corresponded with the beginning of a reduced ability of the ECs to vasodilate (with 10 μM acetylcholine) the aorta when pre-constricted with 1 μM phenylephrine (Fig 4E). This reduced vasodilation could signal reduced EC density or function (generation of nitric oxide) or an increased load carrying by the accumulated collagen that restricts the degree of possible acute distension, or both.

Figure 4. Reduced SMC contractile function associates with ECM deposition during maturation.

A: Normalized outer diameter over time (sec) in response to 1 μM phenylephrine (PE) in aortas between P2-P98. B: The steady-state change in normalized outer diameter (at 900 sec) reveal reduced SMC contractile responses around P10 and P21. C: When calculating a representative SMC contractility by normalizing contractile responses by age-matched SMC densities (Fig. 2A’), a similar trend manifested. D: Interestingly, SMC contractility was negatively correlated to the medial collagen (deposition) rate, suggesting that when SMCs switch to a less contractile phenotype around P10-P21, more medial collagen is deposited in the aorta. E: The vasodilatory function in response to 10 μM acetylcholine, when preconstricted with 1 μM PE, decreased around the time that medial collagen area fraction reached its maximum rate (see Fig. 2D’). F-H: In vivo measurements of volumetric blood flow (cardiac output, CO) (F) and ex vivo loaded inner radius (G) of aortas from mice between E18.5-P98 were used to calculate shear stress $\left({\tau }_w=4\mu Q/\pi a^3\right)$ (H) over developmental time. Vertical red dash-dotted lines in B,C,H indicate the days at which the rate of change is zero. Vertical black dotted lines in E-G represent the time when the absolute maximum rate occurs. Gray-shaded areas in A shows SEM for each age-group and the 95% prediction intervals of the fitted function of the mean in B-H.

Restoration of flow-induced luminal shear stress during maturation confirms an early set-point for ECs. In vivo measurement of volumetric blood flow (cardiac output, CO) confirmed its continuous increase after birth until maturity (Fig 4F). Ex vivo measurement of luminal radius at age-specific physiologic conditions, from E18.5 to P98, revealed continued increases up to maturity that lagged those of blood flow, with inner radius reaching its maximal rate of change at P18.1 versus P15.2 for cardiac output (Fig 4G). Importantly, there was a sharp increase in the calculated value of mean flow-induced shear stress (${\tau }_w=4\mu Q/\pi a^3$, where μ is the blood viscosity, Q blood flow, and a the loaded inner radius) at P10 that began to decrease at P21 and was restored to the early value by P42 (Fig 4H). Recall that this transient period, P10-P21, corresponds with that of rapid ECM deposition (Fig 3) and loss of smooth muscle contractility, both of which affect the evolution of luminal radius, wall thickness, and axial stretch (cf. Fig S2-S4), among other key biomechanical quantities. Hence, the transient in flow-induced shear stress can be regarded as an external perturbation due to abrupt changes in cell and ECM activity that manifest at the tissue level. Consistent with the concept of homeostasis, the flow-induced shear stress returned toward the early value ~6.5 Pa, which appeared to be preferred by the ECs following the perturbation. This finding also supports an interpretation that the near constant value of intramural wall stress observed during early maturation (Fig 1D), prior to marked ECM deposition, represents a homeostatic set-point at the level of the SMC.

Discussion

The perinatal murine aorta (E18.5–P2) consists primarily of contiguous layers of rounded SMCs that are able to vasoconstrict the aorta and withstand the low pressures of the developing cardiovascular system. Associated values of circumferential wall stress during this early period are on the order of 15–20 kPa (~16 kPa active and ~20 kPa passive, both calculated at mean arterial pressure). Interestingly, this range of stress is only slightly above that for both endogenous cell-generated stresses in compliant tissue equivalents, which are 5–10 kPa¹⁶, and preferred values of stress estimated at focal adhesions in cells studied in vitro, which are also on the order of 5–12 kPa¹⁷. This similarity in values of stress across these different scales and situations, coupled with our observations of a marked developmental deposition of ECM beginning around P7 to P10, suggests that a wall stress of 15–20 kPa is preferred, or at least well-tolerated, by aortic SMCs. Early development of a cell-level set-point for wall stress is supported by the equally early development of a set-point for the flow-induced luminal shear stress, at ~6.5 Pa. This calculated early and late value of shear stress is similar to that computed in the mature mouse aorta using computational fluid dynamics, ~8.5 Pa¹⁸, which provides increased confidence in this value as an in vivo aortic set-point. Collectively, these findings are also consistent with long-standing knowledge that flow-induced luminal shear stress is an important driver of vascular development^19–21 as well as subsequent adaptation or disease^22–24.

Although it was previously reported that flow-induced shear stress decreases linearly during development²⁵, our finding is consistent with a report that this stress increases and then decreases nonlinearly in development²⁶. It appears that the transient offset in flow-induced shear stress near P10-P21 is due to a marked increase wall thickness, driven largely by the deposition of elastin but also GAGs (Fig 3, Fig S1), which prevents the inner radius from expanding commensurate with the progressive increase in flow at that time. Wall thickness is traditionally thought to be driven primarily by local blood pressure^27,28, yet early postnatal evolution of wall thickness (Fig S2) follows a different time-course than that of pressure (Fig 1B), particularly from P10-P21 when loaded wall thickness reaches unexpectedly high values (Table S1). This period, P10-P21, is distinguished by multiple unexpected developments, likely arising in part from the complex biaxial mechanics and differential rates therein as well as the differential medial and adventitial matrix production, the former of which includes an early production of elastin followed by that of collagen. Importantly, the time at which pressure increases the most (highest rate at P14.4) falls within this critical period, which is also characterized by a decrease in SMC contractile capacity (Fig 4B).

SMC contractile capacity correlates with passive circumferential wall stress (Fig S2C) and stiffness (Fig S2D), consistent with a need for increased actomyosin capability to mechano-sense and mechano-regulate ECM under increasingly greater matrix stress. This finding supports in vitro studies of smooth muscle cell–matrix interactions¹², although both ECM stiffness and the particular ECM ligand likely dictate SMC phenotype²⁹. Regardless, the intramural cells must use actomyosin activity to assess the local mechanical environment and both establish an appropriate ECM and determine whether to maintain or modify it^30–32. Despite the reduction in SMC nuclei density (per area) after P2 (Fig S4G), immunohistochemical evidence of increasing SMαA (Fig S4H) supports the importance of actomyosin activity during ECM development, not just following maturation. That is, although dogma has suggested that SMCs exhibit a synthetic phenotype in development and a contractile phenotype in maturity, consistent with an apparent switch from cells rich in synthetic organelles to cells rich in contractile proteins¹³, contractile capacity was strong early on. Interestingly, the unexpected reduction in vasoactive response to phenylephrine – which acts through G-protein coupled receptors³³ that have been reported to act as mechanical sensors in vascular SMCs³⁴ – during P10-P21 (Fig 4B) could be related to the increase in loading and vessel caliber occurring around the same time. We previously found acute adaptive changes of SMC contractile capacity in the aorta to different levels of circumferential stretch only when exposed to phenylephrine³⁵.

Vessel-level SMC contractility has multiple biomechanical effects, including modulation of luminal shear and intramural stresses (Table S4) and thus mechano-regulation of ECM production. Yet, perhaps most importantly the reduced contractility around P10-P21 appeared to reflect a transient phenotypic modulation of the SMCs during the period of rapid deposition of elastin, which occurred while the tissue-level stress was well below mature values. This interpretation further challenges the dogma that there is a linear transition from (fetal) proliferative/synthetic to (mature) contractile SMCs. Although deposited early, the elastin did not become mechanically functional until ~P21 (for reasons yet unclear), which appears to mark a transition from a cell-dominated to more of an ECM-dominated wall mechanics. The subsequent increase in collagen deposition appears to have increased stress-shielding of the intramural cells from the increasing hemodynamically induced stresses, consistent with a rule-of-mixtures interpretation (see Methods). A mean wall stress of ~15 kPa thus appears to be homeostatic for normal intramural cells and preferred for mechano-sensing, which suggests that these cells may need to exploit a mechanical advantage to sense and regulate the order of magnitude higher wall stresses ~150 kPa that emerge in maturity (~85 kPa when fully active and ~250 kPa when fully passive). Importantly, ECs continue to be exposed directly to the flow-induced luminal shear stress and they restore the homeostatic value once the lumen of the aorta, and thus wall stress, increases.

Whereas the importance of mechanobiology is well established in embryonic development^36–38, we focused on perinatal and early postnatal periods to identify when aortic wall structure and function begins to be dominated by ECM. Prior studies report pre- and post-natal changes in ECM gene expression during development of the aorta in mice¹⁵. Such information is vitally important but, given that hundreds of genes are differentially regulated during this period, it is currently not possible to correlate individual or grouped transcriptional changes with specific morphological or functional consequences. This should remain our collective goal, including increased attention to microRNAs, but was beyond the current scope as we focused on tissue-level changes that dictate the physiology. Given the importance of elastic fiber development in the aorta, others have appropriately used genetically modified mice (e.g., Eln^−/−, Eln^+/−, Fbn1^−/−, Fbn1^+/−, and Fbln5^−/−) and gestational pharmacological treatments to examine developmental consequences of particular structural contributors and defects^10,39–42. Again, such information is vitally important, but we focused on the natural history of geometric, compositional, and biomechanical metrics in normal thoracic aorta. Understanding normalcy is a critical first step toward understanding disease.

Biological maturation is often considered in terms of overall skeletal development, reproductive capability, or somatic growth, yet different tissues and organs mature at different rates⁴³. We focused on biomechanical maturation of the aorta, which was characterized by multiple emergent steady state (i.e., homeostatic) metrics. Homeostasis is a ubiquitous biological process by which key regulated variables are maintained near preferred target values (or set-points); it leads to stable, near optimal structure and function via negative feedback control⁴⁴. Compromised or lost homeostasis is often a cause or consequence of disease. The current findings provide new insight into developmental origins of mechanical homeostasis of the normal aorta, which promise to serve as a foundation for future studies that seek to understand differences in the structure, function, and properties of the aorta in congenital as well as early versus late onset diseases, particularly given that different metrics may be affected more or less in different diseases. We emphasize that many of the present inferences were possible only because of data collection at sufficient developmental times to enable parameterization of continuous, nonlinear functions that described the time-courses. Similar longitudinal data will be needed to study vascular diseases, congenital and acquired. We also emphasize that mechanical homeostasis requires consideration of states, sensors, and set-points⁴⁵. For example, homeostatic regulation of aortic function can include minimization of differences ∆ in the mechanical state of stress σ from its homeostatic set-point σ_h, as, for example, $\mathrm{\Delta }\sigma \propto K\left(\left(1-\delta \right)\sigma -{\sigma }_h\right)\to 0$, with K > 0 a gain parameter (capturing the sensitivity of the stress modulation) and δ ∈ [0,1] a measure of how well a cell can sense the state of stress (with δ = 0 for perfect sensing, δ < 1 compromised sensing, and δ = 1 lost sensing). Delineating possible alterations in K, δ, and σ_h in connective tissue disorders and diseases will likely require computational modeling cf.⁴⁶, which was beyond the current scope.

With data at six developmental times, judiciously selected based in part on findings by others, we were able to construct continuous nonlinear functions of time that helped uncover when normal steady state values emerge (Fig 5A) and, moreover, specific values of homeostatic targets at cell-versus tissue-levels (Fig 5B,C) that will be critical for allied studies of aortic mechanics in health and disease. Although the assumption that the homeostatic circumferential stress evolves with age allows certain computational models to be parameterized easily^10,20, our data suggest that normal cell-level homeostatic set-points are set during the perinatal period whereas tissue-level intramural set-points emerge only after the wall matures biomechanically, around P56 in male C57BL/6J mice.

Figure 5. Model of developmental origins of mechanical homeostasis.

A: Day at which the maximum rate of change occurred (y-axis) versus the day at which the chosen metric reached 90% of its steady-state value (x-axis), calculated as $f_{90\%}=f\left(P120\right)-0.1\left(f\left(P120\right)-f\left(P0\right)\right)$ of the studied metric f. Biomechanical maturation was reached by ^~P56 (grey vertical zone) for key metrics, including the ability to store elastic energy and the circumferential stiffness (Table S3). In contrast, the maximum rate of loading (cardiac output, blood pressure, thorax length) occurred ^~P16, which preceded the maximum rates of ECM stiffness (blue) and stress (green). Note that certain metrics (shear stress, SMC density and contractility; denoted with asterisks) appeared to be homeostatic near birth and were perturbed near ^~P12 (red squares display day at zero rate) but later returned toward near homeostatic values. B,C: Conceptual summary of the evolution of various mechanical metrics during maturation. Flow-induced luminal shear and pressure-induced intramural stress are low at early postnatal ages (timepoint 1, ^~P2), but reduction in a/h (^~P8.2), mainly driven by increased wall thickness (h) due to elastin deposition, perturbs shear stress (timepoint 2, ^~P12.9). This is associated with a decrease in SMC contractile function (^~P12.2) and later EC vasodilatory capacity (timepoint 3, ^~P17.7). Loss of vascular tone is followed by with an increase in loaded inner radius (a) and wall stress (timepoint 4, ^~P18.1-P25.2), followed by an increase in collagen and overall wall stiffness (timepoint 5, ^~P23.2-P29.5). As ECM constituents increase, SMC contractile function and density are restored (timepoint 6). These data suggest that SMCs are stress-shielded after the ECM matures and carries more hemodynamically increased loads (yellow dash-dotted curve).

Experimental Procedures and Theoretical Methods

In Vivo Measurements.

All animal procedures were approved by the Institutional Animal Care and Use Committee of Yale University. Following isoflurane anesthesia, a Millar SPR-1000 catheter was introduced through the right common carotid artery and advanced to the mid-section of the ascending aorta to measure blood pressure at multiple postnatal ages. Non-invasive ultrasound (PW doppler) data were collected using a high-frequency Vevo 2100 System (Visualsonics, Toronto) to quantify aortic dimensions.

Aortic Samples.

The proximal descending thoracic aorta was excised from male wild-type C57BL/6J mice, from the left subclavian artery to the third pair of intercostal branches, at one of six developmental times: embryonic day E18.5 or postnatal day P2, P10, P21, P42, or P98. Excised vessels were cannulated on custom-drawn glass micropipettes, secured with sutures at each end, and mounted within a custom computer-controlled biaxial testing system^47,48.

Mean Wall Stresses.

Values of circumferential (σ_θ) and axial (σ_Z) stresses averaged across the wall, which depend on the mean values of circumferential λ_θ and axial λ_Z stretches, were calculated using standard formulae⁶

$${\sigma }_{\mathrm{\theta }}\left({\lambda }_{\mathrm{\theta }},{\lambda }_z\right)=\frac{Pa}{h},{\sigma }_z\left({\lambda }_{\mathrm{\theta }},{\lambda }_z\right)=\frac{f_T+\pi a^{2}P}{\pi h\left(2a+h\right)},$$ (1, 2)

where P is the transmural pressure measured with standard transducers, f_T the axial load measured directly by a force transducer, a the inner deformed radius, and h the deformed thickness. Both a and h were calculated, assuming incompressibility, using the unloaded volume coupled with on-line measurements of outer diameter and axial length, which allowed calculation of biaxial stretches using standard formulae (i.e., ratios of current to unloaded mean radius and length, respectively).

Active wall mechanics.

The testing chamber was filled with a phosphate-buffered Krebs-Ringer solution that was maintained at 37°C while bubbled with 95% O₂ / 5% CO₂ to maintain pH at 7.4. After recording the unloaded configuration, vessel viability was assessed as responsiveness of smooth muscle contraction by twice increasing the KCl to 100 mM at a modest age-dependent luminal pressure and axial stretch, followed by smooth muscle relaxation in a normal Krebs solution. After setting the vessel at the in vivo (or preferred) passive value of axial stretch, ${\lambda }_Z^{iv}$, smooth muscle contractions to 100 mM KCl were elicited at age-specific pressure and axial stretches for 15 minutes followed by 15 minutes of relaxation (i.e., washout with normal Krebs-Ringer). This process was repeated for 1 μM phenylephrine (PE) and 1 μM angiotensin II (AngII). After stimulation with PE, smooth muscle relaxation was assessed by adding 10 μM acetylcholine (Ach) to the solution. Additional details can be found elsewhere^49,50.

Passive wall mechanics.

After completing the active testing, the vessel was washed three times in a calcium-free Krebs-Ringer solution. All passive biaxial mechanical testing and data analysis were then performed consistent with methods described previously⁴⁸. Vessels were mechanically preconditioned via cyclic pressurization while held at their preferred (in vivo) axial stretch, ${\lambda }_Z^{iv}$, then subjected to a series of seven biaxial protocols: cyclic age-specific pressurization at ${\lambda }_Z^{iv}$ and repeated at ±5% of this stretch, as well as cyclic axial stretching at four constant age-appropriate luminal pressures. The associated passive pressure-diameter and axial force-length data were fit with a validated four-fiber family constitutive model⁴⁸ via nonlinear regression of a data set consisting of the final cycle of unloading data from all seven protocols. Using data during unloading reveals the elastic energy stored during deformation that would be available to work on the distending fluid. This “four-fiber family” constitutive relation includes eight model parameters within a Holzapfel-type nonlinear stored energy function W, namely

$$W\left(\mathrm{C},{\mathrm{M}}^i\right)=\frac{c}{2}\left(I_{\mathrm{C}}-3\right)+\sum _{i=1}^4\frac{c_1^i}{4c_2^i}\left\{\exp \left[c_2^i{\left(I{V}_{\mathrm{C}}^i-1\right)}^2\right]-1\right\},$$ (3)

where c (kPa), $c_1^i\left(\mathrm{k}\mathrm{P}\mathrm{a}\right)$, and $c_2^i\left(-\right)$ are material parameters (i = 1,2,3,4 denote the four predominant fiber family directions). $I_C=tr\left(\mathrm{C}\right)$ and $I{V}_{\mathrm{C}}^i={\mathrm{M}}^i\cdot \mathrm{C}{\mathrm{M}}^i$ are coordinate invariant measures of the finite deformation, with the right Cauchy-Green tensor C = F^TF computed from the deformation gradient tensor $\mathrm{F}=\mathrm{d}\mathrm{i}\mathrm{a}\mathrm{g}\left[{\lambda }_r,{\lambda }_{\theta },{\lambda }_z\right]$, with $det\mathrm{F}=1$ because of assumed incompressibility. The direction of the i^th family of fibers was identified by ${\mathrm{M}}^i=\left[0,\sin {\alpha }_0^i,\cos {\alpha }_0^i\right]$, with ${\alpha }_0^i$ denoting a fiber angle relative to the axial direction in the traction-free reference configuration. Based on prior microstructural observations from multiphoton microscopy, and because of the yet unknown effects of cross-links amongst the multiple families of fibers, the four predominant families were: axial $\left({\alpha }_0^1=0\right)$, circumferential $\left({\alpha }_0^2=\pi /2\right)$, and symmetric diagonal $\left({\alpha }_0^{3,4}=\pm {\alpha }_0\right)$. Values of mean biaxial wall stress and material stiffness were computed from the stored energy function and calculated at individually measured values of blood pressure and at a common pressure. Additional details can be found elsewhere⁴⁸.

Stress-Shielding Postulate.

It is not possible to measure how stress partitions amongst different constituents; indeed, this is one of the greatest challenges in the continuum theory of mixtures. Yet, a constrained mixture theory, with a rule-of-mixtures description of stress, allows one conceptually to write σ = σ^smc + σ^eln + σ^coll + σ^act, with the first and last contributions capturing passive and active contributions by smooth muscle cells (SMCs). Thus, focusing on passive behavior (σ^act = 0): if stress is ~15 kPa when elastin and collagen are functionally nonexistent, then σ ≅ σ^smc = 15 kPa. If the passive smooth muscle continues to contribute similarly when stress is ~150 kPa, then σ − σ^smc = σ^eln + σ^coll ≅ 135 kPa, which is to say that most of the load is carried by the matrix, which need only be sensed, not carried, by smooth muscle cells of a matrix phenotype⁷, which seems energetically favorable. That the load carrying capability of passive SMCs is low was demonstrated years ago⁵¹.

Histology and Immunohistochemistry.

Following biomechanical testing, specimens were unloaded and fixed overnight in 10% neutral buffered formalin, then stored in 70% ethanol at 4°C for histological examination. Fixed samples were dehydrated, embedded in paraffin, sectioned serially (5-μm thickness), and stained with Verhoeff-Van Gieson (VVG), Movat Pentachrome (MOV), or picro-sirius red (PSR) for standard histology, and primary smooth muscle α-actin (SMαA) antibodies counterstained with 4,6-Diamidino-2-phenylindole dihydrochloride (DAPI) for immunohistochemistry. Detailed analyses were performed on three biomechanically representative vessels for each age group. Custom MATLAB scripts^48,52 extracted layer-specific cross-sectional areas and calculated positively stained pixels and area fractions (AF). Fig S1 in Supplementary Information shows extracted masks of the four primary individual constituents: elastic fibers, medial GAGs, and medial and adventitial collagen at multiple ages. Sequences of the cross-sectional AFs for these primary vascular constituents were fit individually with a common sigmoidal-exponential function (Table S1) to enable rates of change to be computed and compared, thereby revealing the time at which these rates are maximal. Histological images were acquired on an Olympus BX/51 microscope (under normal or polarized imaging) using an Olympus DP70 digital camera (CellSens Dimension) and a 20x magnification objective. Binary images from VVG-stained sections were analyzed with a curvelet-denoising filter followed by an automated fiber-tracking algorithm^52,53 was used to quantify elastin lamellar undulation.