All experiments and protocols were conducted according to protocols approved by the Animal Care and Use Committee, University of Missouri, United States. The studies used neonatal Sprague-Dawley rats at 3, 7, 11, 14, and 19 days and adults at 2 months and 2 years of age. Prior to surgical harvesting of vessels, rats were maintained in an accredited animal facility with 12:12 h light: dark cycle, controlled environmental conditions and food and water ad libitum. For vessel isolation neonatal rats were removed from their mothers immediately prior to an experiment and anesthetized with sodium pentobarbital (Nembutal, 35 mg/kg body weight) given by an intraperitoneal injection. Mature rats were similarly anesthetized with sodium pentobarbital (100 mg/kg body weight). On demonstration of a surgical plane of anesthesia mesenteric tissues were surgically removed and placed in a cooled (4°C) chamber containing dissection buffer (in mmol/L) 3 MOPS, 145 NaCl, 5 KCl, 2.5 CaCl₂, 1 MgSO₄, 1 NaH₂PO₄, 0.02 EDTA, 2 pyruvate, 5 glucose plus 1% bovine serum albumin). Following euthanasia by anesthetic overdose, a craniotomy was performed, the brain removed and placed in a cooled dissection chamber. Small cerebral arteries were isolated from the Circle of Willis, while mesenteric arteries were dissected from 2^nd and 3^rd order branching vessels. Isolated cerebral and mesenteric arteries were rapidly subjected to total RNA purification using a Melt Total Nucleic Acid isolation system kit (Life Technologies, CA, United States) according to the manufacturer’s instructions. Concentration and purity of RNA for each sample was determined by UV absorbance using a ND-1000 spectrophotometer (Nanodrop, Fisher Thermo, Wilmington, DE, United States). Equal amounts of total vessel RNA extract were then reverse-transcribed into single strand cDNA using a Superscript III First-Strand synthesis system (Life Technologies, CA) according to the manufacturer’s instructions. The resulting cDNA template from each sample was subsequently used for real-time quantitative PCR (qPCR) analysis.

Sex of rats used in the functional studies at ages 3 and 19 days was confirmed genetically via detection of the SRY gene. In mammals, the sex-determining region Y (SRY) gene, also known as Testis-determining factor (TDF) located on the Y chromosome, is the only gene necessary and sufficient for testis determination (Oh and Lau, 2006; Ngun et al., 2011). To detect the SRY gene, all cDNA samples were amplified with forward (5’-TGG GAT TCT GTT GAG CCA ACT-3’) and reverse (5’-GCG CCC CAT GAA TGC AT-3’) primers by end-point PCR. Using this approach, we determined 60% of the postnatal rats to be male and 40% female.

Relative mRNA expression levels for elastin, fibrillin-1, fibrillin-2, lysyl oxidase (LOX) and tissue transglutaminase 2 (t-TG) were initially determined using a Mastercycler EP Realplex² system (Eppendorf-North America, NY, United States) and the fluorescent dye SYBER Green (see Supplementary Table S1 for sequences of primer sets). In these assays, real-time PCR (qPCR) reactions were performed in triplicate on cDNAs prepared from each sample (n = 4–5) using KAPA SYBER FAST qPCR Kit Master mix (KAPA Biosystems, Woburn, MA, United States). SYBER Green reaction volume/well contained 20 µL: 10 µL of master mix, 1 µL of each forward and reverse primer (5 µM), 1 µL of cDNA templates and the remainder DNase-free water. PCR conditions were as follows: pre-heating at 95°C for 2 min, 40 cycles of two-step cycling of denaturation at 95°C for 3 s and anneal/extend steps of 25 s at 56°C. For each PCR run, no enzyme and no template controls were included to test for contamination of assay reagents. At the completion of PCR melt curve analysis was performed to demonstrate PCR product purity.

To verify and extend the initial elastin data, relative mRNA expression of elastin plus selected additional ECM proteins (collagen types 1, 2, 3, and 4, Emillin-1 were also determined using specific pre-loaded FAM-labeled TaqMan probes and primer sets (Life Technology, CA, United States) (see Supplementary Table S2 for details). QPCR reactions were performed for each sample (n = 4–5) in customized 96 well plates containing separate TaqMan assays in duplicate. Each well contained 10 µL TaqMan Fast advanced Master Mix (Life Technology, CA, United States), 1 µL of cDNA template and DNase-free water to reach a final volume of 20 µL. QPCR reactions were performed using FAM as the detector probe under the following conditions: pre-heating at 50°C for 2 min, denaturation at 95°C for 1 s and 40 cycles of amplification at 60°C for 20 s 18S rRNA was included in each qPCR experiment to verify the quality of RNA. β-actin and GAPDH TaqMan assays were also included to verify age-dependent stability of housekeeper genes expression. β-actin (Actb) expression used as a reference gene and changes in mRNA expression levels were calculated as the fold relative to elastin mRNA expression of a cerebral artery sample prepared from a 2-month-old rat. This sample was included as a calibrator for all SYBER and TaqMan qPCR assays. Data were collected and analyzed using Realplex software (Eppendorf-North America, NY, United States). Relative quantification was performed using the comparative threshold (Ct) method after determining the Ct values for reference gene (β-actin) and the target genes in each sample set according to the 2^−ΔΔCT method (Livak and Schmittgen, 2001).

Real-time PCR reactions using the fluorescent dye SYBER Green were performed to determine the age-dependent changes in endothelial nitric oxide synthase (eNOS), prostacyclin synthase, and Ca²⁺-activated K⁺ channels including small (SK_Ca), intermediate (IK_Ca) and large (BK_Ca) channels (See Supplementary Table S3 for details of primer sets). These mRNA targets were chosen due to their known involvement in coding for critical proteins underlying various vasodilator pathways. Experimental conditions for SYBER Green-based PCR were similar to those described earlier with the exception of using 42 cycles of two-step cycling of denaturation at 95°C for 3 s and anneal/extend steps of 25 s at 58°C. To calculate the relative fold mRNA expression for each of the chosen genes in cerebral and mesenteric vasculatures, 3- day-old samples have been chosen as calibrator separately. β-actin was used as a references gene for data normalization. Using GAPDH, as a second housekeeping gene, parallel calculations were performed to check validity of the β-actin data (see Supplementary, Figure 1).

In situ elastin fibers within the arterial wall were identified and analyzed from images collected using 3D fluorescence confocal microscopy as previously described (Clifford et al., 2011). Segments of superior cerebellar and small mesenteric arteries from 3, 7, 11, 14, 19 days, 2 months and 2 years old rats were dissected and cannulated onto glass micropipettes mounted in a pressure myograph (Living Systems Instrumentation, VT, United States). The cannulated arteries of the neonatal animals were pressurized to 70% of systemic pressure previously reported for rats aged from birth to 1 month (Litchfield, 1958). Vessels from 2 month to 2 year old rats were pressurized to 70 mmHg. Vessels remained pressurized during all steps of tissue processing and throughout imaging procedures. Pressurized small artery segments were fixed in 2% paraformaldehyde (10 min), rinsed twice in phosphate-buffered saline (PBS) and permeabilized with 0.5% Triton X-100 followed by extensive washing with PBS. Vessels were then stained with 0.2 µM Alexa Fluor 633 Hydrazide (Life Technology A30634; excitation 633/emission 700 nm). This fluorochrome has been previously verified to preferentially binds to elastin (Clifford et al., 2011), Fluorescent nuclear staining was performed with 1 µM Yo-Pro-1 iodide (Life Technology Y3603; excitation 491/emission 509 nm). All dyes, paraformaldehyde, and Triton X-100 were diluted in physiological saline solution (PSS) containing (in mmol/L) 145 NaCl, 4.7 KCl, 2 CaCl₂, 1 MgSO₄, 1.2 NaH₂PO₄, 4 glucose, 2 pyruvate with 20 min incubation time at room temperature. 3D image data sets were collected using a Leica TCS-SP5 confocal microscope, equipped with a Leica ×63 water objective lens (Numerical Aperture 1.2). Z-plane image slices were collected at 0.5 µm intervals as previously described (Clifford et al., 2011).

Alexa Fluor 633 was excited with a HeNe laser at 633 nm to visualize fluorescent staining for elastin (Clifford et al., 2011). Analyses of elastin expression/distribution were performed to quantify and characterize IEL fenestrae using NIH Image (NIH, Bethesda, MD). Briefly, a representative region of interest (ROI) of approximately 60 × 120 µm of vessel image z-stacks was selected from each image to minimize the effects of vessel curvature. Cropped vessel stacks were then analyzed to identify and label the fenestrae. Total numbers of fenestrae and average area per fenestrae (µm²) within the ROI were computed for each vessel segment and compared across age-groups using GraphPad Prism.

Small mesenteric and superior cerebellar arteries were isolated from Sprague-Dawley rats at 3 and 19 days of age. Isolated vessels were cannulated onto glass micropipettes and mounted in a 7 mL chamber of a cannulation stage as previously described (Hill et al., 2000). The micropipettes and cannulated mesenteric arteries contained a modified Kreb’s buffer containing (in mM/L: NaCl 112, Glucose 10.5, NaHCO₃ 25.5, HEPES 10, MgSO₄.7H2O 1.2, CaCl₂.2H2O 2.5, KH₂PO₄ 1.2, KCl 4.7, adjust PH to 7.3). All solutions were maintained at 37°C during an experiment. The cannulation chamber was placed on an inverted microscope and continuously suffused with Kreb’s buffer (4 mL/min). As the physiological relevant range of mean arterial blood pressure for these two postnatal ages are between 20–80 mmHg (Litchfield, 1958), the cannulated arteries were initially pressurized to 20 mmHg without luminal flow. Pressurized arteries were gently stretched in the longitudinal direction such that while increasing intraluminal pressure from 5 to 50 mmHg lateral bowing did not occur. Arteries were then warmed to 37°C and equilibrated for approximately 1 h prior to functional assessment.

Following the equilibration period, pressure-diameter relationships were recorded across a pressure range of 5–50 mmHg in the absence of flow. After measurement of myogenic responsiveness, vessels were set to an intraluminal pressure of 30 mmHg and endothelial function assessed by responses to acetylcholine (ACh; 10⁻⁸M-10⁻⁵M). To evaluate components of ACh-induced dilation, concentration response curves were repeated in the presence of combinations of eNOS inhibitor (L-NAME, 100 µM), prostaglandin synthesis inhibitor (indomethacin, 10 µM), SK_Ca (apamin, 1 µM), and IK_Ca blockers (TRAM-34, 1 µM). Since the pressurized cerebral arteries developed differing levels of myogenic tone at the two time-points studied, vessels were exposed to U-46619 (a thromboxane receptor agonist, 1 µM) for 20 min as a pre-constrictor. To directly assess the function of vascular smooth muscle cells, sodium nitroprusside (SNP, 50 µM) was applied after the final ACh concentration-response curve in the continued presence of the inhibitors. Finally, vessels were suffused with 0 mM Ca²⁺ buffer containing 2 mM EGTA and passive pressure-diameter relationships were determined across a pressure range of 5–50 mmHg. Changes in intraluminal diameter in response to alterations in intraluminal pressures or dilator agents were monitored using an inverted microscope (Olympus IX70, St Louis, United States) and a video-based caliper together with an A-D data processing system (ADInstruments, Colorado Springs, United States).

Expression data for each vessel type were analyzed for age effects using one-way analysis of variance (ANOVA) together with Dunnett’s post-test. Day 3 was chosen to represent control. Comparisons between vessel types were performed using two-way ANOVA together with Bonferroni’s multiple comparison test (GraphPad Prism, CA, United States). Statistically significant differences were identified at probability values of p < 0.05. For functional studies all vessel diameters were normalized to the passive diameter at 30 mmHg (%D₃₀passive).

Using a SYBER green-based approach, qPCR analysis demonstrated that cerebral artery elastin mRNA levels rise significantly from postnatal day 3 to a peak at 11–14 days after which expression appears to decline significantly throughout life (p < 0.05, one-way ANOVA, Figure 1A). Similarly, elastin mRNA expression for small mesenteric arteries peaked at postnatal day 11 (relative expression 22.4 ± 1.2) which was significantly greater than for cerebral vessels (relative expression 12.5 ± 1.1) (p < 0.05, two-way ANOVA). Although elastin mRNA levels for small mesenteric arteries declined throughout life, they remained elevated compared to the cerebral vessels (Figure 1A). Both the patterns of elastin mRNA expression and differences in magnitude of expression between vessel types were confirmed using a TaqMan assay system (see Supplementary Figure S2). As LOX is a key enzyme for the intramolecular cross-linking of both elastin fibers and collagen, thus contributing to the formation of elastin-rich sheet-like structure of the IEL (Rodriguez et al., 2008), its mRNA expression levels were similarly examined. LOX mRNA expression, in both vessel types, reached a peak around days 14–19 before declining between days 19 and 60 to a low level maintained through 2 years of age. During the peak expression period, LOX mRNA levels were greater in mesenteric vessels compared to cerebral arteries (Figure 1B). In addition to LOX, mRNA expression levels for t-TG2 were also examined because of its previously reported crucial role in the formation of the beaded outer microfibrils for construction of mature elastic fibers (Kielty et al., 2002). As shown in Figure 1C, the mRNA expression patterns for transglutaminase 2(t-TG2) were markedly different compared to that for elastin and LOX. t-TG2 mRNA rose slightly and gradually from postnatal day 3 through life in both vessel types with a significant increase at 2 years compared to control (3 day, p < 0.05 one-way ANOVA. t-TG2 mRNA levels were also significantly greater in mesenteric compared with cerebral arteries from day 11 through all later time points (p < 0.05, two-way ANOVA). Fibrillin-1 & 2, mRNA expression was similarly examined because of their reported role as major structural elements of fibrillin-rich microfibrils in ECM assemblies (Kielty et al., 2002; Wagenseil and Mecham, 2007). Although expressed at very low levels, fibrillin-1 mRNA in mesenteric arteries was higher compared to cerebral arteries in which expression levels were largely unchanged throughout life (Figure 2A). At days 3 and 7 fibrillin-2 mRNA levels were significantly (p < 0.05, two-way ANOVA) greater in mesenteric arteries compared to cerebral vessels while at later time-points expression was similar (Figure 2B). mRNA expression of Emilin1 (“elastin microfibril interface located protein”) showed a similar multiphasic expression pattern in and mesenteric arteries while remaining relatively constant in cerebral vessels (Figure 2C).

TaqMan assays were used to quantify mRNA levels for collagen types I, II, III, and IV and to determine developmental differences in expression patterns between mesenteric and cerebral arteries. Fibrillar types I and III collagens showed mRNA expression patterns similar to that of elastin although at lower absolute levels (Figure 3A compared to Figure 1A). Thus, mRNA for both proteins peaked at approximately 11–14 days with significantly higher levels being apparent in mesenteric vessels compared to cerebral arteries (p < 0.001, two-way ANOVA). In contrast, expression of collagen Type II mRNA was not detectable in either vessel type while mRNA for basement membrane type IV collagen was relatively constant throughout life and of similar magnitude in both vessels (Figure 3B).

As we reported in earlier studies (Clifford et al., 2011), elastin fibers in the walls of small arteries from adult rats (2 months) exhibit an intricate 3D arrangement within the vascular wall and show marked differences between various small arteries in elastin content and spatial distribution. To extend this study, 3D confocal microscopy was performed on cannulated pressurized cerebral and mesenteric arteries from all age groups described above to examine developmental changes in the expression and distribution of small artery wall matrix proteins. Z-stack images from cerebral arteries at postnatal ages of 3, 7, and 19 day-old (Figures 4A,B, respectively) show the absence of adventitial elastin layer in these vessels confirming our earlier finding and showing that this occurs throughout life. A progressive development of cerebral artery IEL was also apparent across postnatal ages of 3, 7, 11, 14, 19 days and 2 months. This was evident as a progression from a fibrous sheet like appearance towards more well-defined and solid appearing sheet with fenestrae (Figure 4, left to right, respectively). Similarly, the IEL of small mesenteric arteries showed a postnatal development towards a more uniformly fenestrated sheet (Figure 4, bottom images). The adventitial elastin component of mesenteric arteries appeared to become denser with age (Figure 5; top images (A) 3 day, (B) 14 day, and (C) 19 day. Z-stack images at the level of the media showed accumulation of higher elastin content at the adventitial layers of mesenteric arteries from 3, 7, 11, 14, 19 day-old and 2 months rats (Figure 4A from left to right, younger to older accordingly). This occurred in parallel with the development of IEL from a fibrous appearing form into a continuous fenestrated sheeted from (Figure 4B; Figure 5).

In an attempt to provide a quantitative analysis for elastin expression at the protein level, we determined fluorescence intensity (of Alexa Fluor 633 hydrazide) within regions of interest (ROI) while doing this at defined locations within the z-stack set of images. As shown in Figure 1A, throughout life, total elastin content of cerebral arteries was significantly lower than for mesenteric vessels. This is consistent with cerebral vessels lacking adventitial elastin. As shown in Figure 4B, the content of elastin in the IEL was relatively constant throughout postnatal life whereas total expression increased supporting the suggestion of elastin accumulation in the adventitia in older ages. The numbers of fenestrae (within a ROI of 60 × 120 µm) in rat mesenteric small arteries decreased from an apparent maximum at day 7 to a minimum at 2 years of age (Figure 6A). Note that we were not able to conduct this measurement at 3 day-old mesenteric arteries because of open mesh network characteristic of fibers within the IEL. Thus, identification and definition of clearly distinguishable fenestrae was not possible. The mean size of mesenteric fenestrae and IEL thickness increased significantly but only after the postnatal period (Figures 6B,C). Collectively these studies indicate that there is postnatal development of elastin networks in the walls of small arteries which raises questions as to the functional significance of these anatomical changes.

To determine whether small artery function showed developmental changes in the postnatal period we compared vasomotor responses of cannulated mesenteric and cerebral arteries at postnatal days 3 and 19. Emphasis was placed on myogenic responsiveness and endothelial-dependent vasodilation. As shown in Figure 7, although differences in patterns were apparent, both the cerebral and mesenteric arteries showed myogenic responsiveness at, postnatal day 3 and 19 consistent with development of a level of vascular smooth muscle cell-mediated contractile activity. Of potential significance was the observation that myogenic constriction of day 3 cerebral arteries were left-shifted on the pressure scale and occurred at lower intraluminal pressures (5–20 mmHg; Figure 7A) compared to day 19 at which time the myogenic response curve was more right-shifted with myogenic constriction expressed at intraluminal pressures over 30 mmHg (Figure 7B). Pressure-induced vasoconstriction was, however, nearly identical between day 3 and 19 mesenteric arteries (Figures 7C,D). Thus, these data suggest that mechanotransduction processes underlying myogenic function of mesenteric and cerebral arteriolar VSMCs may be fully developed in the immediate postnatal period. However, full expression of the vascular myogenic response develops in cerebral vessels as aging occurs shifting right ward on the pressure scale, perhaps paralleling and adapting to developmental increases in systemic pressure.

To test whether endothelial cell (EC) function changes during the postnatal period, ACh-concentration-response curves were examined in cerebral and mesenteric arteries at postnatal days 3 and 19. Experiments were performed in the absence or presence of L-NAME (NOS inhibitor), indomethacin (prostaglandin synthase inhibitor), apamin (SK_Ca blocker), and TRAM-34 (IK_Ca blocker) to examine mechanisms contributing to EC-dependent vasodilation. ACh-induced vasodilation in 3 day-old mesenteric arteries was markedly attenuated compared to day 19 arteries (Figure 8A, maximal relaxation: 25.8% ± 4.8% vs. 66.8% ± 10.5% respectively) suggesting developmental changes in EC-mediated vasodilation during the postnatal period. After blockade of two major paracrine vasodilators (NO and PGI₂), ACh-mediated vasodilation in 19 day-old mesenteric arteries reduced significantly to a maximal dilation of 34.9% ± 6.1% vs. control whereas EC-dependent vasodilation was almost completely abolished in 3 day-old vessels (Figure 8B). To examine any contribution of hyperpolarization mediated by small (SK_Ca) and intermediate (IK_Ca) conductance Ca²⁺-activated K⁺ channels, the 3 and 19 day-old mesenteric vessels were also treated with a combination of L-NAME, indomethacin, TRAM-34 (IK_Ca blocker), and apamin (SK_Ca blocker). The combination of all four inhibitors caused complete inhibition of ACh-induced vasodilation in day 19 mesenteric arteries (Figure 8C) suggesting that a hyperpolarization component of EC-dependent dilation develops during the postnatal period.

We similarly expanded the above studies of EC-dependent vasodilator function to cerebral arteries at postnatal days 3 and 19. As shown in Figure 9A, in the absence of inhibitors ACh-induced relaxation was evident at both ages but at a considerably lower magnitude in the younger animals (maximal relaxation 40.9% ± 10.0% in 19 day vs. 12.1% ± 9.9% in 3 day-old cerebral vessels. The combination of L-NAME and indomethacin totally abolished Ach-induced dilation 3 day-old cerebral arteries while at day 19 the vessels dilated to 19.1% ± 6.6% (Figure 9B). Consistent with the mesenteric data, at day 19, ACh-induced vasodilation was abolished in the presence of the four inhibitors (Figure 9C).

To determine whether the age-dependent differences in vasodilation reflected a smooth muscle component, responsiveness to the NO donor, SNP (50 µM), was examined. Robust and equivalent dilator responses were observed to SNP in both vascular beds and at both time-points (Figures 8C, 9C). Collectively, the data suggest an age-dependent maturation in an EDH-dependent component of vasodilation.

While argument could be made that the differences in the functional data indicate maturation of myoendothelial signaling, including development of the IEL, alterations in the expression of various ion channels and enzymes could also contribute. As such, we performed additional qPCR experiments on cerebral and mesenteric arteries to determine the relative mRNA expression levels of endothelial nitric oxide synthase (eNOS), prostaglandin-endoperoxide synthase 2 (Ptgs2), endothelial small (SK_Ca), intermediate (IK_Ca), and large (BK_Ca) conductance Ca²⁺-activated K⁺ channels at postnatal days 3 and 19. In cerebral arteries, SK_Ca, eNOS3, and α-BK_Ca mRNA levels were significantly increased at day 19 compared to the day 3 groups (Supplementary Figure S3A). No significant change in mRNA expression was detected for Ptgs2 while a small, but significant, decrease in IK_Ca mRNA expression was detected at day 19 compared to the day 3 (Supplementary Figure S3A). As shown in Supplementary Figure S3B, qPCR analysis for mesenteric arteries demonstrated that the abundance of eNOS3 and IK_Ca mRNA was similar in both age groups studied while there were significant increases in levels of mRNA expression for α-BK_Ca, SK_Ca, and Ptgs2 at day 19 compared to day 3. A comparison between cerebral and mesenteric expression levels is shown in Supplementary Figure S3C.

As development could conceivably alter the expression of the house keeping genes used for normalization, we examined both β-actin and GAPDH expression levels. In these experiments β-actin and GAPDH show consistent levels of expression at days 3 and 19 for both vasculatures (Supplementary Figure S1). Parallel qPCR analyses using GAPDH, confirmed gene expression levels with the exception of an age-dependent difference in expression for IK_Ca in cerebral vessels.