Axial stretch regulates rat tail collecting lymphatic vessel contractions

Lymphatic contractions play a fundamental role in maintaining tissue and organ homeostasis. The lymphatic system relies on orchestrated contraction of collecting lymphatic vessels, via lymphatic muscle cells and one-way valves, to transport lymph from the interstitial space back to the great veins, against an adverse pressure gradient. Circumferential stretch is known to regulate contractile function in collecting lymphatic vessels; however, less is known about the role of axial stretch in regulating contraction. It is likely that collecting lymphatic vessels are under axial strain in vivo and that the opening and closing of lymphatic valves leads to significant changes in axial strain throughout the pumping cycle. The purpose of this paper is to quantify the responsiveness of lympatic pumping to altered axial stretch. In situ measurements suggest that rat tail collecting lymphatic vessels are under an axial stretch of ~1.24 under normal physiological loads. Ex vivo experiments on isolated rat tail collecting lymphatics showed that the contractile metrics such as contractile amplitude, frequency, ejection fraction, and fractional pump flow are sensitive to axial stretch. Multiphoton microscopy showed that the predominant orientation of collagen fibers is in the axial direction, while lymphatic muscle cell nuclei and actin fibers are oriented in both circumferential and longitudinal directions, suggesting an axial component to contraction. Taken together, these results demonstrate the significance of axial stretch in lymphatic contractile function, suggest that axial stretch may play an important role in regulating lymph transport, and demonstrate that changes in axial strains could be an important factor in disease progression.

muscles are very sensitive to changes in the level of intracellular Ca 2+ , and circumferential stretch has been implicated as a modulator of calcium sensitivity 14,16,26,27 . There is also evidence that electrical activity of lymphatic vessels change with changes in circumferential stretch; higher stretches induce Ca 2+ sensitization that increases the responsiveness of key contractile molecules to Ca 2+28-30 . Similar to circumferential stretch, axial stretch may also be a modulator of lymphatic function. In some regions, such as the lung and diaphragm, axial stretch changes due to applied external forces; e.g., pleural and diaphragm lymphatics can be exposed to cyclic stretches which enhance lymph formation and transport 31,32 . The effect of axial load on the phasic contractions has been studied via axial load-controlled experiments on the bovine mesenteric lymphatics 20 . McCale et al. showed that, in contrast to pressure-induced stretch that changed both the frequency and the strength of contractions, axial load only increased the strength of contractions while the frequency of contraction remained unchanged; further, it was observed that axial load linearly increased the axial stretch, under fixed pressures 20 . An increase in the axial load with both pressure and axial stretch has been reported for the rat thoracic duct 33 ; however, there is no experimental evidence to show lymphatic vessels experience axial stretch in situ. Towards this end, the goal of this study is to quantify the in situ axial stretch of rat tail lymphatics and axial stretch-mediated modulation of lymphatic contractility in an ex vivo platform. Further, the distribution of collagen and LMCs's of rat tail lymphatics in the circumferential-axial plane were quantified to understand the axial contributions of these structural constituents.

Methods
Animal model and vessel isolation procedure. All animal experiments were approved by Georgia Institute of Technology's Institutional Animal Care and Use Committee and were performed in accordance with the principles outlined in the National Institutes of Health, Guide for the Care and Use of Laboratory Animals. A ~1-cm incision was made on one side of the rat tail, near the base, to gain access to the lateral tail vein of male Sprague Dawley rats (300-350 g), with care taken to avoid injury to the lateral vein or artery (Fig. 1A). Trypan Blue Solution 0.4% dye was injected upstream of the incision site to enhance the visibility of lymphatic vessels (Fig. 1B). Under a stereo microscope, a lymphatic chain that runs parallel to the lateral vein was visualized.
Quantification of in situ axial stretch. To approximate the in situ axial stretch, animals (n = 5, one vessel per animal) were placed under a ZEISS dissecting microscope equipped with an AxioCam MRc 5 (Zeiss, Inc.) camera. After removing the perivascular tissue, pins (0.2 mm diameter, Fine Science Tools, Inc.) were used to www.nature.com/scientificreports www.nature.com/scientificreports/ mark the length of the segment at the dissection locations (Fig. 1C). Using microsurgical scissors, the lymphatic vessel was dissected at the two pins (Fig. 1C). Digital images were acquired before and after excision and ImageJ software was used to determine the pre-and post-excision length; the in situ axial stretch was defined as L / z in situ i n situ u , λ =  , where in situ  is the pre-excision length and L u is the post-excision, unloaded length.
Ex vivo experimental setup, protocol, and data acquisition. For vessels used for ex vivo experiments, a ~2-mm segment of the collecting lymphatic vessels (n = 6, one vessel per animal) was excised parallel to the lateral vein and placed in Dulbecco's Phosphate Buffered Saline (DPBS, Corning) for the removal of the surrounding fat and connective tissues. The excised and cleaned lymphatic vessel (containing approximately 1-2 valves) was mounted on opposing cannulae in a custom-made vessel chamber 33 . A custom biomechanical testing device, with some modifications [33][34][35] , was employed to control transmural pressure and axial stretch, while monitoring diameter. Transmural pressure was adjusted using a feedback loop with a syringe pump (PHD 4400, Harvard Apparatus, Inc.) and pressure sensors (1 psi SSC series, Honeywell). Axial stretch was controlled via computer-controlled linear actuators and XYZ stages (Newport Precision, LTA series and M-461 series). This system was placed on an inverted microscope (2.5x magnification) and the vessel diameter was recorded via a digital camera (Allied Vision Technologies, Marlin F-033B, at 18 fps); recordings were post-processed to obtain diameter tracings (Fig. 2). In a temperature-controlled bath containing DMEM/F-12 (Dulbecco's Modified Eagle Medium/Nutrient Mixture F-12 plus Pen-Strep 1:100, pH = 7.4 at 37 °C) 36 , the vessels were equilibrated at 2 mmHg for ~1 hour to develop stable, repeatable contractions. For the axial stretch protocol, vessels were preconditioned by delivering three pressure steps (2 mmHg, 4 mmHg, 6 mmHg) over ~5 minutes before the pressure was held constant at 2 mmHg; this pressure was chosen based on a pilot study to determine the pressure at which contraction amplitude is maximum. The unloaded length (L) was defined as the minimum axial length required to the keep vessel straight at the (nearly) zero pressure; i.e., the pressure just above the pressure at which the vessel collapses, ~0-0.5 mmHg. The vessel length was adjusted using the precision micromanipulators to determine the length at which the vessel starts to bend. This length (L) was measured as the distance between the two mounting sutures on glass pipettes, while viewing the vessel on the inverted microscope, via the digital camera, with a custom-written LabView code. The axial length of the vessel was increased so that the axial stretch (λ = L / z  , where  is the loaded axial length, defined as the distance between the two sutures after stretching) was 10%, 20%, and 30% of the unloaded length. Vessels were held at each axial stretch for 5 minutes and contraction data collected during this time were recorded and analyzed. Based on our preliminary experiments and to prevent damage and plastic deformation that may occur during supra-physiological loading and would alter the unloaded length, diameter, and the overall response of subsequent cycles, we limited our experiments to a single pressure of 2 mmHg.

Figure 2.
A schematic diagram of ex-vivo measurement of axial stretch using ex vivo setup. The vessel was pressurized using a syringe pump at ~0.5 mmHg to determine the minimum length at which the vessel does not bend (unloaded length). Length was measured using a LabView program and axial stretch was determined with respect to unloaded length and actuators were used to change axial stretch. A custom-written LabView program was used to track contractions as a function of time. An image of a rat tail lymphatic vessel cannulated on glass micropipettes along with respective raw traces of contractions at different stretches are shown.
www.nature.com/scientificreports www.nature.com/scientificreports/ Ex vivo contractility metrics and statistics. A custom-written MATLAB (MathWorks) program was used to calculate end-systolic diameter (ESD), end-diastolic diameter (EDD), and contraction frequency (FREQ) based on the raw data obtained during experiments. From these raw measurements, contractile amplitude (AMP), ejection fraction (EF), and fractional pump flow (FPF) were calculated as: confocal microscopy for collagen organization. Multiphoton confocal images of unfixed rat tail collecting lymphatic vessels, loaded with a transmural pressure (p = 2 or 8 mmHg) and z λ = 10% on our custom mechanical testing device, were acquired using a Zeiss LSM 710 system with 40×/1.3 and Plan-Apochromat 63×/1.4 oil-immersion objectives. The second harmonic generation (SHG) of collagen fibers were obtained under 800 nm excitation wavelength. To compare differences in collagen organization, image stacks were collected in the valvular and tubular regions; the tubular region was defined as 100 µm distance from the valve. Note that a typical rat tail is 3.3 mm 37 . The image z-stacks were imported in ZEN lite (Zeiss, Inc.) software to obtain 3D reconstructions. Quantitative parameters, including fiber angle and fiber straightness, were obtained using CT_FIRE software (downloaded at https://loci.wisc.edu/software/ctfire) that automatically extracts individual fibers by applying the curvelet denoising filter followed by an automated fiber extraction method 38,39 . CT_Fire uses a two-step approach i) a curvelet filter is applied to denoise the image and enhance fibers edges ii) a fiber tracking algorithm (FIRE) detects fiber boundaries and center points and calculates length of fiber, angle and straightness. Straightness ranges from 0 to 1 and is defined as the linear length of the fiber divided by the distance along the fiber.
Whole-mount immunofluorescence staining for LMC organization. Rat tail lymphatic vessels were isolated and fixed, unloaded, for two hours in 4% PFA at room temperature. After fixing, the vessel was washed three times with PBS and blocked with 5% goat serum in PBS and 0.5% Triton X-100 for 1-2 hours, followed by addition of the primary antibody (1:100; α-SMA Sigma) and incubation for 48 hours overnight at 4 °C (primary antibody was replaced with fresh antibody after 24 hours). After washing in PBSTx (0.3% Triton X-100 in PBS) three times, the sample was incubated with the secondary antibody at room temperature for 2 hours (1:1000, Alexa Fluor ® 488, goat anti-mouse antibody). After finishing the staining, the vessel was washed in PBSTx three times and mounted in Prolong Gold for imaging. The orientation of the nuclei was obtained from DAPI staining and analyzing the respective images with CellProfiler software 40 .
Statistical analysis. Each experimental data point represents data from a unique animal. To determine statistical significance, a one-way ANOVA, with Bonferroni's post-hoc correction, was used to make multiple comparisons. Prism 5 (GraphPad Software) was used to perform all statistical analyses, with p < 0.05 considered as statistically significant, and results were reported as means (±SEM).

Results
Rat tail lymphatic vessels are under axial stretch in situ and increased axial stretch reduces pumping function. The measured in situ axial stretch, λ = . ± .
1 24 0 03 in situ (n = 5), demonstrates that rat tail lymphatic vessels are under significant axial stretch (Fig. 1). Representative diameter tracings illustrate the changes in the contractile response of rat tail collecting lymphatics under constant pressure, ex vivo, with sustained changes in axial stretch (Fig. 2). Ex vivo experiments on the isolated vessels (n = 6) showed that the amplitude of contractions decreased as the stretch increased, with significant reduction (50 µm equal to ~90% reduction) at 30% axial stretch, compared to 10% stretch (Fig. 3). Similar trends were observed for the frequency, which decreased from 15 min −1 to <5 min −1 (~80% reduction), the ejection fraction, which decreased from 46% to 9% (~80% reduction), and the fractional pump flow, which decreased from 6.8 min −1 to 0.58 min −1 (~92% reduction) from 10% to 30% stretch. Notably, these metrics of contractions changed significantly between 20% and 30% contraction, which represents the range of axial stretches measured in situ. Taken together, end systolic diameter, end diastolic diameter, frequency, contractile amplitude, ejection fraction, and fractional pump flow all decreased with increasing axial stretch, predominantly from 10% and 20% to 30% stretch, demonstrating that contractile function of rat tail collecting lymphatic vessels are sensitive to changes in axial stretch.
Collagen fibers are predominantly oriented in the axial direction. A typical 3-D reconstruction of a vessel, based on the z-stack of images, illustrates predominantly axially oriented collagen fibers, both in the tubular region and valvular region of the rat tail collecting lymphatics (Fig. 4). The CT_FIRE code analysis of collagen fibers in the tubular region showed that the probability distribution for fiber angle (n=5) was maximum in the axial direction (defined as 0°) and the minimum of probability distribution occurred at ~±75°. Further, imaging of collagen fiber orientation at high pressure (p ≅ 8 mmHg), compared to those at lower pressure (p ≅ 2 mmHg) in tubular regions, revealed that an increase of pressure does not change the peak fiber angle of collagen, which was in the axial direction; however, the circumferential fiber angle increased (<5%) and fiber straightness increased suggesting a small percentage of collagen fibers are recruited at high pressures. At physiological pressure, the distribution was normal with a maximum near the axial direction (~70% of fibers, −30 < α < 30).

Scientific RepoRtS |
(2020) 10:5918 | https://doi.org/10.1038/s41598-020-62799-x www.nature.com/scientificreports www.nature.com/scientificreports/ LMcs are oriented in both the circumferential and axial directions. Whole mount α-SMA and DAPI staining revealed a dense single layer of LMCs; occasionally, a second layer of LMC's was observed (Figs. 5 and 6). Both DAPI and actin images illustrate that LMC bundles in the tubular region were predominantly oriented in the circumferential direction, with a second, smaller, more sporadic subset of LMC bundles oriented in the axial direction. In contrast, LMCs in the valvular region were more randomly organized, with thick bundles of LMCs primarily in the longitudinal direction. The analysis of nuclei orientation (n = 5) confirmed that most cells, as well as collagen fibers in the valvular region, were aligned in the axial direction, whereas the tubular region's nuclei distributions were circumferential and axial. Taken together, the local maximum of nuclei and actin fiber orientations occurred in both axial and circumferential directions (Figs. 5 and 6), but collagen fibers were predominantly oriented in the axial direction (Fig. 4).

Discussion
This study shows that rat tail collecting lymphatic vessels are under both circumferential and axial loading, under physiological conditions, and that changes in axial stretch have a significant effect on the contractile pumping function of lymphatics. These results suggest that axial stretch may play an important role in regulating lymph transport and changes in axial strains could be an important factor in disease progression. Further, these results reveal an important consideration that lymphatic experimentalist must consider when quantifying lymphatic contractile function; namely, consistency in axial strain is important when comparing results across experiments and across research groups.

Axial loading likely varies over the pumping cycle.
Simple mechanical analysis shows that over a typical pumping cycle, the axial stress and stretch in the lymphatic wall may change significantly. Static equilibrium of two adjacent lymphangions (Fig. 7) requires that where T zz L1 and T zz L2 are the axial components of the Cauchy stress, A w L1 and A w L2 are the cross-sectional area of the vessel wall, A L1  and A L2  are the luminal cross-sectional area, and P L1 and P L2 are the luminal pressure in lymphangion 1 and lymphangion 2, respectively. For an elastic lymphatic wall, T zz L1 and T zz L2 are a one-to-one function of the in-plane circumferential and axial stretches, L1 λ θ and z L1 λ and L2 λ θ and λ z L2 , respectively, where r R / where R L1 and R L1 are the loaded and unloaded radii of L1 and  L1 and L L1 are www.nature.com/scientificreports www.nature.com/scientificreports/ the loaded and unloaded lengths of L1, respectively. Consider a simple case where A A r , where r i and r o are the luminal and outer radius and h is the wall thickness. For this simple case, which says that, for a difference in transmural pressure − P P ( ) L L 2 1 across the valve between two lymphangions, the axial stress (and stretch) will be greater in the lymphangion under higher pressure and lower in the lymphangion under lower pressure, due to the differential pressure P P ( ) − applied across the closed valve. This differential pressure increases the stretch in L2 and decreases the stretch in L1. This analysis suggests that throughout the pumping cycle there is likely significant relative changes in axial stress (T zz L1 versus T zz L2 ) that arises due to transvalvular pressure differences caused by the combination of contraction and the opening and closing of valves. Of course, throughout pumping, as the vessel contracts and dilates, the cross-sectional areas of the lumen and wall also change differently in time between the two lymphangions, which in turn alter the axial stress and stretch of each lymphangion. Further, the lymphangions that are proximal and distal to L1 and L2 also affect T zz L1 and T zz L2 . We have observed significant axial motion of isolated lymphangions, in both in situ and ex vivo settings over the pumping cycle, supporting these analyses (data not shown). The vessels were excised from the base of the tail and mounted on glass pipettes in a custom-made vessel chamber compatible with the microscope. The vessels were unfixed, the pressure was ~2 mmHg, and the axial stretch was ~10%. The axial direction corresponds to 0° and the circumferential direction corresponds to ±90°. (2020) 10:5918 | https://doi.org/10.1038/s41598-020-62799-x www.nature.com/scientificreports www.nature.com/scientificreports/ We observed that the stretch that yields the maximum contractile response is below the measured in situ axial stretch. It should first be noted that the in situ axial stretch and the ex vivo axial stretch may not be directly comparable measurements; e.g., the ex vivo measurements were made after mechanical preconditioning and isolation in a bathing solution of PSS, whereas the in situ measurements were made while the vessel was under in vivo loading, with lymph fluid in the lumen, but bathed in PSS, without any mechanical preconditioning. Methodological differences aside, given that the axial stretch likely changes throughout the pumping cycle, the observation that maximum pumping efficiency occurs at stretches below the mean in situ stretch may suggest that these cyclic changes in axial stretch could play a role in regulating and even orchestrating timing of contractions; e.g., as the axial stretch in a lymphangion decreases over the pumping cycle, perhaps this triggers that lymphangion to initiate contraction. Similarly, as the axial stretch increases over the pumping cycle, perhaps this triggers the LMCs to relax.
Equation (4) states that, for equivalent radii in L1 and of L2 (and thus, λ λ = θ θ L L 1 2 ), the axial stretches in each lymphangion, z L1 λ and λ z L2 , are governed by the constitutive relation (e.g., strain-energy function) that relates stress to strain (or stretch) for the vessel wall. For a stiffer material, the differences in axial stretch (λ z L1 versus z L2 λ ),for a given geometry and given difference in pressure will be smaller, compared to that of a more compliant vessel. Anisotropy, too, will play an important role in controlling differences in the circumferential and axial stiffnesses. Therefore, it is important to quantify the biaxial mechanical behavior of lymphatics and the microstructural content and organization to quantify material behavior and anisotropy; see Caulk et al. 33 .
Collagen fibers are orientated primarily in the axial direction, suggesting material anisotropy and increased stiffness in the axial direction. Collagen fibers play a key role in structural support and www.nature.com/scientificreports www.nature.com/scientificreports/ mechanics of lymphatic vessels. The mechanical properties of a load bearing tissues depend on the anisotropic architecture of collagen fibers, with the predominant collagen fiber direction correlating to increased stiffness in that direction. The current observation suggest that the predominant axial orientation of collagen fibers provide structural support for axial loads due to in vivo axial stretch, perhaps serving to regulate change in axial stretch throughout the pumping cycle. Collagen organization and fiber recruitment are important factors that not only determine the passive function but also influence the contractile function of lymphatic vessels. Recent multiphoton imaging of collagen fibers within the wall of the pelvic lymphatic vessel has shown that the fibers are predominately oriented in the axial direction 41 . Similar results were found for the bovine 42 and rat lymphatics, although  there are some variations 33 . Data on lymphatic vessel biaxial mechanics and compositions are limited to a few measurements from rat thoracic duct, rat mesentery, and human pelvis 33,41,43 . Observations from rat thoracic duct have shown the non-linear response and vessel stiffening occurs in response to both axial stretch and circumferential stretch 33 . It is believed that collagen, elastin, and LMCs are the main structurally-significant components that constitute the mechanical behavior of lymphatic vessels 33,41-43 . Axially-oriented LMcs may contribute to axial-oriented contractions and changes in axial stretch. Another way that axial stretch may change over the pumping cycle is through the direct contraction of LMC's oriented in the axial direction. Ohtani and Ohtani, report α-SMA positive cells in rat diaphragm run longitudinally between intraluminal valves, but circumferentially in the valve regions 44 . In guinea pig mesenteric lymphatics, von der Weid showed that some LMCs are longitudinally orientated while others are orientated at approximately a 45° angle 45 . An investigation of lymphatic vessels in the diaphragm, Moriondo et al. suggest that active vessels showing rhythmic contractions have a dense mesh of actin fibers, but non-contracting vessels have longitudinal actin fibers 46 . F-actin staining in human thoracic ducts has suggested that non-contracting thoracic vessels have an irregular organization of lymphatic LMCs and muscle bundles are heterogeneously oriented in the circular, oblique, and longitudinal directions 47 . Micrograph observation in monkeys has revealed cervical sections of the thoracic duct are less muscular compared to the abdominal and thoracic sections, suggesting the passive function of the cervical part 48 . In the current study, immunostaining and imaging of fibers revealed that the orientation of lymphatic LMCs is heterogeneous. In the tubular region, most fibers have circular orientation, with a subset of longitudinally-oriented LMCs. In the valvular region, actin fibers were predominantly in the axial orientation.
One limitation of this study is the fact that ex vivo and in situ axial stretch measurements are not quite comparable. To determine in vivo stretch, vessels were cleaned in situ and then stretch was quantified as described in the Methods section. In contrast, to determine ex vivo stretch, vessels were dissected, cleaned first, and subsequently mounted in a vessel chamber. Once preconditioned in temperature-controlled conditions, the stretch was calculated with respect to unloaded length. Further, the goal of ex vivo experiments was to study the effect of axial stretch on the contractions in a pressure-controlled condition. while lymphatics are mechanically supported by the surrounding perivascular tissue in vivo, ex vivo testing requires lymphatic vessels to be isolated from this perivascular support. While ex vivo experiments have the benefit of 'isolating' the observed mechanical response to the tissue of interest alone, thereby removing other confounding factors that may arise due to the perivascular tissue, the disadvantage is that the applied loading, ex vivo, does not simulate the exact in vivo loading condition delivered from the perivascular tissue.
Another limitation of this study is that we did not measure changes in axial force with changes in axial stretch. Our group has previously measured the axial force of lymphatic vessels during passive mechanical testing 33 . The passive axial forces under physiological pressure and axial stretch are very small (~0.1 mN) and differences in axial force due to contractions is, perhaps, an order of magnitude lower than the passive axial force. While it is possible to resolve such differences over acute mechanical testing (over a few minutes) with our experimental device, over longer experiments of 20-30 minutes, even the highest resolution load cells experience a drift on the order of 10 th 's or 100 th 's of mN's. Therefore, we could not, with confidence, report nominal values for measured changes in axial force versus changes in contractile function.
In closure, this work shows that the contractile function of rat tail collecting lymphatics are sensitive to changes in axial stretch. The microstructural organization of these lymphatics show that collagen fibers are aligned predominantly in the axial direction and a subset of LMCs are also axially-aligned, both of which may serve as a conduit for this mechanical signal (axial stretch) to be transduced to a biological response (altered contractility). These findings highlight the important role of axial loading in lymphatic pumping function, which may play a role in lymphatic disease.