Constriction of isolated collecting lymphatic vessels in response to acute increases in downstream pressure

Abstract

Collecting lymphatic vessels generate pressure to transport lymph downstream to the subclavian vein against a significant pressure head. To investigate their response to elevated downstream pressure, collecting lymphatic vessels containing one valve (incomplete lymphangion) or two valves (complete lymphangion) were isolated from the rat mesentery and tied to glass cannulae capable of independent pressure control. Downstream pressure was selectively raised to various levels, either stepwise or ramp-wise, while keeping upstream pressure constant. Diameter and valve positions were tracked under video microscopy, while intralymphangion pressure was measured concurrently with a servo-null micropipette. Surprisingly, a potent lymphatic constriction occurred in response to the downstream pressure gradient due to (1) a pressure-dependent myogenic constriction and (2) a frequency-dependent decrease in diastolic diameter. The myogenic index of the lymphatic constriction (−3.3 ± 0.6, in mmHg) was greater than that of arterioles or collecting lymphatic vessels exposed to uniform increases in pressure (i.e. upstream and downstream pressures raised together). Additionally, the constriction was transmitted to the upstream lymphatic vessel segment even though it was protected from changes in pressure by a closed intraluminal valve; the conducted constriction was blocked by loading only the pressurized half of the vessel with either ML-7 (0.5 mm) to block contraction, or cromakalim (3 μm) to hyperpolarize the downstream muscle layer. Finally, we provide evidence that the lymphatic constriction is important to maintain normal intraluminal valve closure during each contraction cycle in the face of an adverse pressure gradient, which probably protects the lymphatic capillaries from lymph backflow.

Key points

• Arterioles undergo a myogenic constriction, defined as a decrease in diameter in response to an increase in pressure, which serves to protect the downstream capillaries from changes in pressure and flow.

• A lymphatic constriction was recently identified but it remained unknown whether it reflected a true myogenic constriction.

• By selectively raising downstream pressure in isolated lymphatic vessels containing a single valve, we discovered that the upstream segment constricted, even when protected from increases in pressure by the closed valve. The constriction consisted of a myogenic component and a frequency component, which were blocked pharmacologically.

• The lymphatic constriction facilitated proper closure of the intraluminal valves in the face of a hydrostatic gradient, preventing lymph backflow.

• This work adds to our understanding of the lymphatic myogenic constriction by showing that it maintains a functioning valve system in lymphatic vessels and that it is mechanistically similar to the arteriolar myogenic constriction.

Introduction

A major function of the lymphatic vasculature is to preserve tissue fluid balance by absorbing the capillary filtrate and returning it to the bloodstream at the junction of the thoracic duct and subclavian vein. Due to the higher pressures within the blood vasculature relative to those of the lymphatic vasculature, lymphatic vessels transport lymph against a pressure gradient (Smith, 1949; Hargens & Zweifach, 1977; Muthuchamy & Zawieja, 2008). To perform this task, the lymphatic vasculature utilizes two distinct types of vessels: lymphatic capillaries and collecting lymphatics. Lymphatic capillaries are defined by a single layer of overlapping endothelial cells with discontinuous tight junctions (Baluk et al. 2007) that allow one-way entry of tissue fluid, much like a check-valve (Lynch et al. 2007). Collecting lymphatics are larger calibre vessels that possess intraluminal valves, a basement membrane and at least one muscle cell layer, in addition to continuous interendothelial junctions comparable to those present in blood vessels. Collectively, these features provide collecting lymphatics with unique contractile properties and a lower permeability to water and solute relative to lymphatic capillaries (Scallan & Huxley, 2010), which are appropriate for their role as transport vessels.

Due to vestment with muscle cells, collecting lymphatic vessels possess basal tone and are able to modulate their contractile state in response to stimuli in a manner similar to that of arterioles. For instance, arterioles and collecting lymphatics respond both to local changes in intraluminal flow (Kuo et al. 1991; Gashev et al. 2002; Gasheva et al. 2006; Bohlen et al. 2009, 2011) and to changes in circumferential stretch (Zhang et al. 2007; Davis et al. 2009b). The specialized musculature of collecting lymphatic vessels also allows them to contract spontaneously, rapidly and periodically to generate sufficient pressure to propel lymph towards the thoracic duct against a higher hydrostatic pressure (Smith, 1949; Zweifach & Prather, 1975).

Another characteristic of arterioles is that they exhibit myogenic responses to local changes in pressure, in which an increase in pressure induces arteriolar constriction and a decrease in pressure induces dilation (Johnson, 1981). Two functions that have been ascribed to the arteriolar myogenic response are to buffer downstream capillary beds against abrupt changes in pressure – which would alter net capillary filtration – and to contribute to the autoregulation of blood flow (Davis, 2012). We recently demonstrated that collecting lymphatic vessels undergo a myogenic constriction as a time-dependent decline in end-diastolic diameter in response to elevated intraluminal hydrostatic pressure (Davis et al. 2009a). Although the constriction was small in absolute magnitude, the calculated myogenic index (percentage constriction per change in intraluminal pressure) was comparable to that of arterioles. However, those lymphatic vessel experiments were performed in the absence of a trans-axial pressure gradient, i.e. during simultaneous elevation of both input and output pressures.

In subsequent experiments testing the adverse (uphill) pressure gradient required to close intraluminal lymphatic valves (Davis et al. 2011), we noted that increasing the output pressure while maintaining a constant, low input pressure often led to constriction. This result prompted us to test the hypotheses that: (1) lymphatic vessel segments exposed to selective increases in output pressure consistently produce a constriction; (2) the magnitude of this constriction exceeds the constriction to combined elevation of input and output pressures; and (3) this constriction is indeed a myogenic response in the same sense used to describe the arteriolar myogenic constriction. To address these hypotheses, isolated collecting lymphatic segments were subjected to rapid step-increases or gradual ramps in output (downstream) pressure at a constant input (upstream) pressure. We used lymphatic vessels containing one or two valves for the various protocols on a microscope-based system designed to provide separate control of input and output pressure in the respective segments, while diameter and intralymphangion pressure were continuously measured. Our results show that there are two components to the lymphatic constriction when output pressure is selectively elevated, and that a signal originating in the output segment is conducted across a closed valve to induce a constriction in the input segment, which never senses a rise in pressure directly. We demonstrate that this constriction functions, at least in part, to preserve normal valve function when the vessel is pumping against an adverse pressure gradient. This new physiological mechanism may be important in preventing the onset or spread of lymphoedema by maintaining a patent valve system, and is probably required in dependent extremities where the lymphatic pump is subjected routinely to an elevated adverse hydrostatic gradient.

Methods

Vessel isolation

Male Sprague–Dawley rats were handled in accordance with the US Public Health Service Policy for the humane care and use of laboratory animals and all protocols were approved by the University of Missouri Animal Care and Use Committee. After anaesthetizing the animal (170–260 g body weight) with pentobarbital sodium (Nembutal, 100 mg kg⁻¹, i.p.), an abdominal midline incision was made, through which a small loop of intestine was exteriorized. Collecting lymphatic vessels were isolated from the mesentery near the duodenum. Immediately after their isolation, vessel segments were transferred to a dish containing albumin-supplemented physiological saline solution (APSS) for removal of fat and connective tissue. In keeping with our protocol, the animal was then killed with an overdose of Nembutal followed by cervical dislocation. Cleaned vessels were transferred to a 3 ml chamber, capable of temperature control, for subsequent cannulation followed by equilibration to 37°C.

Solutions

All chemicals were obtained from Sigma (St Louis, MO, USA), excluding bovine serum albumin (BSA), which was obtained from US Biochemicals (catalogue no. 10856; Cleveland, OH, USA). ML-7 (Sigma catalogue no. I2764) and cromakalim (Sigma catalogue no. C1055) were dissolved initially in 50% ethanol and further diluted in APSS to the stated concentrations. The standard solution, physiological saline solution (PSS), contained (in mm): 145.0 NaCl, 4.7 KCl, 2.0 CaCl₂, 1.2 MgSO₄, 1.2 NaH₂PO₄, 0.02 EDTA, 5.0 glucose, 2.0 sodium pyruvate and 3.0 Mops (pH 7.4 at 37°C). APSS was an identical solution, except for the addition of 0.5 g per 100 ml purified BSA. During the cleaning and temperature equilibration steps, the bath and pipettes both contained APSS, after which the external bath solution was changed to PSS. Ca²⁺-free PSS, used to obtain passive pressure–diameter curves at the end of the experiment, was identical to PSS, except 3.0 mm EDTA was substituted for CaCl₂.

Cannulation, pressure control and diameter tracking

For determination of lymphatic contractile responses to elevated downstream pressure, a vessel segment was cannulated at each end with a glass micropipette mounted on a Burg-style V-track system (Duling et al. 1981) and pressurized on a Zeiss inverted microscope. Pressure was increased briefly to 10–13 cmH₂O from a standing reservoir while adjusting the distance between the two pipettes to remove slack from the vessel; otherwise axial buckling at higher pressure levels led to inaccurate diameter tracking. Afterwards, pressure was reduced to 3 cmH₂O for the temperature equilibration period that lasted from 45 to 60 min. For the entire experiment, a peristaltic pump exchanged the bath solution continuously with fresh PSS exposed to room air at a rate of 0.4 ml min⁻¹. To enable computer control of the pipette pressures, the pipette connections were switched from the standing reservoir to a servo-controlled pressure system by which input (P_in) and output (P_out) pressures could be independently controlled (Davis et al. 2011). The microscope image was digitized at 30 Hz using a firewire camera (model A641FM, Basler, Ahrensburg, Germany) and recorded in AVI format (Davis et al. 2011). Internal diameter changes and valve leaflet movements were tracked with custom LabVIEW software (National Instruments Corp., Austin, TX, USA) online and/or offline.

Servo-null pressure measurements

Two-valve lymphatic segments required the use of a servo-null micropipette system for reliable pressure measurement within the central lymphangion. Borosilicate glass (1.0/0.5 mm outer diameter/inner diameter, Omega-dot; Frederick Haer, Bowdoin, ME, USA) was drawn to a fine point using a Sutter P-97 puller (Sutter Instruments, Novato, CA, USA) and broken to an outer diameter of ∼2 μm. Afterward, the micropipette was backfilled with 2 m NaCl and connected to a servo-null micropressure system (model 4A; IPM, La Mesa, CA, USA) that was calibrated before the start of each experiment using a standing reservoir water manometer system. Micropuncture was performed after cannulation but prior to equilibration of the vessel at 37°C, using the procedure previously described (Davis et al. 2011), to avoid excess trauma to the vessel. Once the bath temperature had reached 37°C, the micropipette calibration was tested by manipulating both the input and the output pressures and adjusted if necessary. The calibration was checked periodically throughout the experiment and data were not used if the calibration was unreliable.

Valve tracking

Intraluminal lymphatic valves default to an open position when there is no trans-valve pressure gradient. The valve leaflets close when P_out exceeds P_in, but the magnitude of the adverse pressure gradient required for closure increases 20-fold with increasing vessel diameter (Davis et al. 2011). To confirm that a valve remained closed in the face of an adverse pressure gradient in the present protocols, valve leaflet position measurements were routinely made during replay of the AVI files. Detection of the valve leaflet position was accomplished by placement of a rectangular window over the video image to measure mean pixel intensity (Davis et al. 2011). Typically the window was placed immediately upstream from the closed valve so that the base of each leaflet occupied the majority of the window. When the valve opened, the leaflets moved out of the window, causing the signal to increase. Afterwards, a threshold was applied to convert the raw densitometer output signal to ‘0’ when the valve was closed, and to ‘1’ when open.

Protocols

Both single-valve and two-valve segments were utilized for three different protocols (see Fig. 1). Two-valve segments contained one complete lymphangion that defined its own intralymphangion pressure pattern in response to changes in P_in and P_out. Single-valve segments enabled determination of the response of the input segment to P_out elevation in the absence of intraluminal pressure generation; in this configuration intraluminal pressure was fixed by the cannulas, i.e. pressure on the input side remained constant despite the occurrence of spontaneous contractions. Therefore, the use of the servo-null method to measure intraluminal pressure was unnecessary in single-valve segments, which did not contain a complete lymphangion.

Figure 1. Cartoon depicting the two types of protocols (B) run on isolated lymphatic vessel segments containing either one or two valves (A).

A, segments of collecting lymphatic vessels were cut to lengths containing a single valve (left) or two valves (right). The pressures in a single-valve segment were defined by the input (left) and output (right) pipettes, while a two-valve segment contained a central lymphangion capable of regulating its own pressure. To measure the intralymphangion pressure in these segments, a servo-null micropipette was used (top pipette). Valve gating was tracked by using densitometer windows (dotted rectangles). Arrows indicate the direction of flow in situ. Solid rectangles highlight the sites used for diameter tracking. B, protocols used for assessing the effects of pressure on lymphatic diameter and tone. Abrupt step increases in pressure tested the speed of the myogenic constriction and enabled its study over time, while pressure ramps allowed the vessel to adapt to a gradual rise in pressure. C, another protocol was employed to further investigate whether the output segment initiated the lymphatic constriction of the input segment in response to selective P_out elevation. For this protocol, the output half of a single-valve lymphatic segment (n= 6) was perfused with pharmacological inhibitors to alter the function of the muscle layer, while the input/output diameters and valve gating were measured. A third, drug-containing pipette (labelled) was advanced into the lumen of the output half of the vessel and pressure was raised to exchange the solution for the drug (shaded region). P_out was raised to close the valve before perfusion, preventing contamination of the input side with the drug.

For each protocol vessel diameters, pressures and valve leaflet positions were recorded. The sites where diameter and valve leaflets were tracked are depicted in the schematic in Fig. 1A. Arrows (Fig. 1A) indicate the normal direction of net lymph flow in vivo; therefore, the input side of the vessel is always described as upstream relative to the output, or downstream, side. Single-valve and two-valve collecting lymphatic vessel segments were each exposed to step-increases in output pressure to investigate the time course of the lymphatic constriction to rapid pressure changes, or to gradual ramps in P_out to assess adaptation of the lymphatic constriction to a constantly rising pressure (see Fig. 1B). Ramps were performed at a set rate of 4 cmH₂O min⁻¹. In addition to the P_in and P_out traces in Fig. 1B, the intralymphangion pressure (P_L), measured by the servo-null method, was included. To enable discrimination of each individual trace on the graphs, a common colour scheme was used (see Fig. 1A) in which the traces for pressure, diameter and valve position on the input side of the vessel were coloured blue; the same traces on the output side of the vessel were coloured red; and the traces for P_L and diameter measured between the valves were coloured black.

A third protocol (Fig. 1C) was performed on single-valve collecting lymphatic segments to investigate the hypothesis that the muscle layer of the output segment generated and transmitted a signal to the input segment to induce its constriction. For these experiments P_out was stepped to 6, 8, 12 and 14 cmH₂O before and after administering pharmacological inhibitors solely to the output half of the vessel segment. Input diameter was measured as before, while output diameter was measured offline. To prevent spillover of the perfused inhibitors into the input segment, the intraluminal valve was first closed by raising and holding P_out higher than P_in by 1 cmH₂O, a low pressure gradient that was insufficient to induce a constriction. A third, smaller bore cannula (tip size ∼20 μm outer diameter), containing an inhibitor diluted in APSS, was inserted inside the output cannula but advanced only as far as the middle of the pipette shank prior to vessel cannulation to avoid premature diffusion of the compound into the vessel segment. After the control pressure steps were completed, this inner pipette was advanced past the outer pipette tip inside the lumen of the output portion of the vessel and pressure was raised slightly above that of the output cannula for ∼1–2 min to exchange the intraluminal contents for the inhibitor (Fig. 1C, shaded region). The same pressure steps were then repeated. Two inhibitors were used to disrupt lymphatic smooth muscle function: ML-7 was administered at a maximal dose (0.5 mm) to inhibit myosin light chain kinase, thereby abolishing spontaneous contractions in only the downstream half of the vessel; and cromakalim (3 μm), a K_ATP channel agonist, was administered at a dose that hyperpolarized the muscle layer in the downstream half of the vessel, but did not inhibit spontaneous contractions.

At the end of every experiment, lymphatic vessels were equilibrated for ∼15–30 min in Ca²⁺-free PSS solution at 37°C. Hydrostatic pressure was lowered to less than 0.5 cmH₂O and held there for the duration of the Ca²⁺-free equilibration period. Afterwards, pressure was increased, simultaneously in both cannulae, to every pressure level used in that particular protocol and the diameters were recorded at each pressure. A passive pressure–diameter curve was then constructed for each vessel.

Data analysis

The individual traces for every experiment were analysed with a custom-written LabVIEW program designed to detect end diastolic diameter (EDD), end systolic diameter (ESD), amplitude (AMP) and frequency (FREQ). These results were obtained for the trace preceding and during each P_out step or ramp. FREQ was calculated on a contraction-by-contraction basis. Other parameters were calculated as follows

(1)

where MaxD is the maximal passive diameter for a particular pressure in the protocol

(2)

where EDD_f is the average EDD during the P_out step, while EDD_i represents the initial EDD before the pressure step:

(3)

where FREQ_i and FREQ_f are the average frequencies taken for 2 min periods before and after the pressure steps in the third protocol:

(4)

(5)

where P_i and P_f represent the pressures before and after the P_out step (in mmHg), respectively.

Data sets were analysed with Excel, JMP 5.1 (SAS, Cary, NC, USA) or Prism 5 (Graphpad Software Inc., CA, USA). For most analyses a one-way ANOVA was used to test for differences between groups, with pressure as the independent variable. The Tukey–Kramer post hoc test was performed afterwards to test for significant within-group variation. For comparisons of ML-7 or cromakalim treatments to control datasets, paired t tests were used. In all cases data were reported as means ± SEM with significance defined as P < 0.05.

Results

Relationships of intraluminal pressure and valve gating to P_out elevation

The four types of pressure responses observed after a step-increase in P_out and their relationships to valve gating are summarized in Fig. 2. Valve leaflet position is passively regulated by the trans-axial pressure gradient (Davis et al. 2011); therefore, if P_in or P_L does not exceed P_out appreciably (∼0.2–2.0 cmH₂O), then the valve cannot open. For these protocols, only vessel segments with functional valves were used, meaning that when P_out was elevated while P_in was held at a lower level, pressure from the output portion did not leak through the output valve. Typically, when P_out was raised to a moderate level, the lymphangion of a two-valve collecting lymphatic segment matched its internally generated systolic pressure (P_L) to the imposed pressure so that the output valve opened near the peak of each systolic period, resulting in the consistent ejection of a portion of the lymphangion contents (Fig. 2A). With larger output pressure steps, the lymphangion pressure did not initially match the step-increase in P_out, preventing the opening of the output valve (Fig. 2B and C). Half of these vessels remained unable to open the valve at any time following the pressure step, even after several minutes (Fig. 2B). In the others, the lymphangion gradually developed a higher peak systolic pressure until it was able to consistently open the valve with each contraction cycle (Fig. 2C). In combination, these data emphasize that lymphangions are capable of increasing their internal pressure but there is a limit to the peak P_L that a lymphangion can generate at any given preload, or P_in.

Figure 2. Various contraction patterns evoked by P_out elevation and their relation to valve gating.

For simplicity, only intraluminal pressure and valve positions are shown. A, input and output valve positions (bottom and top, respectively) as a function of time are displayed on top. Shown below is a zoomed-in tracing of P_in (lower tracing), P_out (stepped tracing), luminal pressure measured by servo-null (P_L) and the average luminal pressure (avg P_L, dashed tracing) for a two-valve segment. After a step increase in P_out from 1 to 8 cmH₂O, the lymphangion immediately increases its P_L so that the valve is opened with each contraction. B, the opposite case is shown here, where a step increase in P_out from 1 to 10 cmH₂O prevents successful valve ejection, even over a ∼5 min time span. C, in ∼50% of these experiments, a two-valve segment cannot generate enough pressure to open the valve initially, but over time an apparent increase in lymphatic muscle contractility enhances the ability of the lymphangion to open the valve. D, for one-valve segments, all of the pressure generated by the input side (P_L) is shunted out of the input pipette so that the pressure matches that of P_in. For these segments the valve will not open if P_out exceeds P_in by more than ∼0.5 cmH₂O.

The single-valve collecting lymphatic segment, which contained only a partial lymphangion, allowed us to measure changes in diameter independent of fluctuations in intraluminal pressure, because any pressure increase that would have been generated by spontaneous contractions was shunted out of the input and output pipettes. A set of experiments (five vessels; 16 P_out steps and three P_out ramps) were performed where a servo-null micropipette was placed on the input side of the valve to verify the absence of a significant pressure increase arising from spontaneous contractions (Fig. 2D, P_L trace). The P_L trace of Fig. 2D shows this clearly, as the pressure spikes were greatly diminished in amplitude and mean pressure (P_L) did not deviate substantially from 2 cmH₂O, equal to P_in (lower trace). These measurements also confirmed that the valves were fully functional over the P_out range used in these protocols. Additionally, the pressure pattern shown in Fig. 2D dictates that any response detected in the input portion of single-valve segments during P_out elevation must originate from a cue derived from the output portion of the vessel, as the P_out increase is not transmitted upstream past the valve.

An increase in downstream pressure evokes lymphatic constriction

When a two-valve lymphatic vessel segment was subjected to step-increases in P_out (Fig. 3), a constriction occurred upstream of the output valve. Lymphatic constrictions shown in Fig. 3 were gradual declines in vessel diameter, sometimes preceded by a small passive dilation, such that the final EDD was less than the EDD prior to the pressure step. Additionally, the decline in EDD was accompanied by an increase in ESD, with the net effect of reducing contraction amplitude. However, not all vessel segments constricted gradually in response to a step-wise increase in P_out. Figure 4A and B provides two examples where lymphatic vessel segments constricted immediately to a new, smaller EDD in response to P_out elevation. This rapid constriction was always associated with an increase in FREQ in response to a P_out step. In the 27 vessels where P_out steps were performed, eight exhibited predominantly rapid constrictions with the remaining 19 vessels constricting gradually. Notably, Fig. 4B demonstrates that the frequency increase was greatest when the valve facing the opposing pressure remained closed – i.e. when P_L had not yet matched P_out– and tended to wane after P_L had matched P_out, thus opening the valve with each contraction cycle.

Figure 3. Two-valve collecting lymphatic vessel segments constrict gradually in response to stepwise elevation of P_out.

Tracings from a representative experiment show valve positions (top), pressures (middle) and vessel diameter (bottom) plotted as a function of time (x-axis). During each step, the lymphangion generated sufficient pressure from each contraction to open the output valve. For this vessel, every step produced a gradual decrease in EDD over time (arrows) such that the final EDD after the constriction was less than the EDD prior to the pressure step. In some steps a passive dilation was observed prior to the constriction.

Figure 4. Immediate constriction of a two-valve collecting lymphatic vessel segment in response to step-increases in P_out.

Two examples (A and B) of an immediate constriction are shown where EDD drops rapidly (arrow) from its initial value before the step (dashed line). The reduction in EDD is also accompanied by an increase in contraction frequency in both examples. However, in B, the contraction frequency is highest before P_L matches P_out. After this occurs the contraction frequency is reduced.

Lymphatic constrictions in response to pressure consist of two components

In some two-valve collecting lymphatic segments, constrictions occurred without an increase in contraction frequency (9 of 52 vessels in total; compare Fig. 5A and B). Interestingly, in these cases (Fig. 5B), the EDD constriction to pressure elevation appeared attenuated compared to instances when contraction frequency was increased (Fig. 5A). Furthermore, spontaneous diastolic pauses in the contraction frequency were sometimes observed during sustained P_out elevation (19 of 52 vessels in total; Fig. 5A). When these pauses occurred, the EDD of the lymphatic vessel rose to an intermediate level that was still constricted relative to the initial EDD, but markedly less than when a higher contraction frequency was maintained (compare dotted lines on diameter trace). Taken together, these data suggest that separate frequency-dependent and pressure-dependent effects on EDD exist.

Figure 5. Constrictions of two-valve collecting lymphatic vessel segments consist of two components.

Two examples (A and B) of a collecting lymphatic vessel that constricts immediately following the downstream pressure increase. A, after the pressure step, the two-valve collecting lymphatic segment is initially unable to open the output valve, but peak systolic PL rises over time until the valve opens; periodic pauses are evident subsequently, but unlike the previous figure, the pauses are not closely correlated to ejection. During the periodic pauses, EDD rises partially, but is still constricted relative to the EDD before the pressure step (compare dashed lines). B, during a pressure ramp, a separate vessel exhibits a gradual decline in EDD without an increase in frequency. When frequency did not also increase, the reduction in EDD was less marked. Together these results point to a myogenic reduction in EDD combined with a further reduction due to elevated FREQ, which reduces the time for filling (similar to negative lusitropy in vivo).

Study of single-valve collecting lymphatic vessel segments to dissect pressure and frequency effects

To resolve the separate effects of pressure and frequency on diameter, we next studied single-valve lymphatic segments in which the input portion of the vessel did not directly experience the increase in pressure from the output pipette due to the presence of the closed valve (see P_L trace, Figs 2D and 6); recall also that the input segment is unable to generate an increase in intraluminal pressure due to the open cannulation pipette. Single-valve experiments therefore allowed us to determine how an increase in frequency alters diameter in a system isolated from direct pressure-induced effects. When single-valve segments were exposed to step increases in P_out (Fig. 6), an immediate constriction of the input segment was always obtained with a concomitant increase in contraction frequency. The top diameter traces in Fig. 6 show the response of the output segment that was exposed to the pressure elevation. Both the input and the output segments exhibited synchronized contraction frequency increases with magnitudes generally proportional to the step increase in P_out. The degree of constriction for the input segment also appeared proportional to the magnitude of the P_out step. The principal difference between the responses of the two portions was that during P_out elevation, the EDD of the input side declined below the initial EDD prior to the pressure increase, while P_out elevation to pressures ≥8 cmH₂O caused an initial, near-maximal distention of the output segment so that even after subsequent constriction, the final EDD was still greater than the initial EDD prior to the pressure step (compare the top trace to the dotted line, Fig. 6). It is important to note that the input portions of the vessels used in these protocols exhibited spontaneous contractions that were synchronized with the contraction cycle of the output segments, even though a closed valve separated the two segments and only the output segment was exposed to an elevated pressure.

Figure 6. Study of single-valve collecting lymphatic vessel segments to dissect the two components of the lymphatic constriction in response to P_out elevation.

Single-valve segments were unable to generate sufficient pressure to open the valve against a step increase in pressure (see valve tracing), yet an immediate reduction in EDD still occurred upon P_out elevation (lower arrows). This response was accompanied by an increase in contraction frequency that varied with the magnitude of the pressure step. The diameter of the output segment was measured using offline video analysis and overlaid on the last two steps (upper diameter traces). Input and output diameter traces were graphed on the same y-axis scale. The output segment also constricted in response to the P_out steps (top arrow). The constriction of the input segment appears to be due to an increase in contraction frequency that is transmitted from the output segment in response to P_out elevation because pressure did not change in the input segment.

A signal derived from the output segment muscle layer initiates the constriction of the input segment

The results shown in Fig. 6 suggested that the frequency response was initiated by the output segment and transmitted upstream. We hypothesized that a signal was transmitted through either the lymphatic endothelium or the muscle cell layer from the output to the input segment across a closed valve when P_out was elevated.

To test the contribution of the endothelial versus muscle layers to the frequency-induced constriction, we aimed to denude the endothelium or selectively impair muscle function. We performed several experiments to attempt selective denudation of the lymphatic endothelium in the output segment by using mechanical approaches employed previously for arterioles, but the thin lymphatic muscle layer was additionally damaged, leading to an irreversible hyper-contractile state.

To test whether the muscle layer was required for initiating a myogenic constriction, we filled the output portion of single-valve collecting lymphatics (n= 6) with ML-7 to inhibit myosin light chain kinase, and thus spontaneous contractions (Fig. 7; Supplemental Movie 1). A typical example of the control response to pressure steps is provided in Fig. 7A. Upon examination of the pauses that sometimes occurred during the control pressure steps, it becomes evident that the contractions of the output and input segments were synchronized initially (Fig. 7A). After selective application of ML-7 to the output portion of the vessel, where the inhibitor was confined, spontaneous contractions were abolished solely in the output segment (Fig. 7B, top trace), rendering the input portion of the lymphatic vessel intact (Fig. 7B, lower trace). Notably, the basal EDD and contraction frequency of the input portion did not change significantly from control levels after administering ML-7 (compare lower traces of Fig. 7A and B). When P_out was stepped to 6, 8, 12 and 14 cmH₂O while measuring the response of the input segment, EDD decreased with increasing P_out under control conditions, but did not change in response to any pressure step while the output segment was exposed to ML-7 (Fig. 7C, P= 0.02). Likewise, the frequency increase that correlated with increasing P_out control steps was absent during ML-7 treatment of the output segment (Fig. 7D, P= 0.01).

Figure 7. Treatment of half of a single-valve lymphatic vessel with ML-7 abolishes the constriction of the input segment.

A, example trace of four control P_out steps to 6, 8, 12 and 14 cmH₂O in a single-valve segment before perfusion of ML-7. Upper diameter trace is of the output segment (right y-axis) over time (x-axis). Lower diameter trace is of the input segment (left y-axis) over the same time scale. B, the same vessel as in A, after ML-7 was applied to only the output half of the vessel. Note that spontaneous contractions are inhibited only in the output segment (upper diameter trace), while the input segment contracts normally (lower diameter trace). The lymphatic constriction in response to P_out elevation is completely inhibited. C and D, quantification of the effect of ML-7 treatment on the lymphatic constriction of the input segment. For all vessels (n= 6), the percentage change in EDD and the percentage change in frequency in response to P_out elevation were inhibited by ML-7 at every pressure step (means ± SEM; *P < 0.05 for both graphs).

To investigate whether the output signal relied on a change in the membrane potential of the muscle cells, the muscle layer of only the output segment was hyperpolarized with cromakalim using the same approach (n= 6 vessels, Fig. 8, Supplemental Movie 2). The concentration of cromakalim used temporarily inhibited spontaneous contractions for ∼5 min in the entire vessel; after this brief period, synchronized contractions returned in both segments (compare diameter traces, Fig. 8B). We noted that before the application of cromakalim, the output portion of five out of the six vessels initiated a contraction wave that propagated towards the input segment. The speed and direction of this contraction wave were quantified by tracking vessel diameter at the two opposite ends of the vessel over a known distance and measuring the delay between the initiation of contraction in the output segment versus input segment during P_out steps to 6 cmH₂O. Under control conditions, the contraction wave travelled from the output to the input side of the vessel at an average velocity of 5.8 ± 0.8 mm s⁻¹ in these five vessels (Supplementary Movie 2). After cromakalim was added to the output half of the vessel, the contraction wave travelled at a velocity of −7.1 ± 1.3 mm s⁻¹ (i.e. in the reverse direction), appearing to be initiated by the input segment (Supplementary Movie 2). Additionally, cromakalim typically reduced the basal spontaneous contraction frequency from control (compare lower traces of Fig. 8A and B). These two observations indicate that cromakalim hyperpolarized the output portion of the vessel, but was not sufficient to inhibit spontaneous contractions. Application of cromakalim to the output segment abrogated any changes in EDD (Fig. 8C, P= 0.02) or frequency (Fig. 8D, P= 0.01) of the input segment in response to P_out elevation for every pressure step when compared to control responses.

Figure 8. Cromakalim treatment of half of a single-valve lymphatic vessel abolishes the constriction of the input segment.

A, example trace of four control P_out steps to 6, 8, 12 and 14 cmH₂O in single-valve segments (n= 6) before perfusion of cromakalim. Upper diameter trace is of the output segment (right y-axis) over time (x-axis). Lower diameter trace is of the input segment (left y-axis) over the same time scale. B, the same vessel as in A after luminal perfusion of cromakalim in the output half of the vessel. Note that spontaneous contractions remain in both halves of the vessel (top and bottom diameter traces). However, when P_out is stepped to each pressure, no constriction is observed for either the input or the output segments. C and D, quantification of cromakalim treatment on percentage change in EDD (C) or percentage change in contraction frequency (D) of the input segment. Cromakalim treatment of the output half of the lymphatic segment inhibited the constriction and frequency increase of the input segment at every pressure step (means ± SEM; *P < 0.05 for both graphs).

Comparison of two-valve and single-valve myogenic indices

To quantify lymphatic constrictions in response to ramps or step increases in output pressure, the percentage change in EDD was plotted against the change in pressure for two-valve and single-valve lymphatic segments in Fig. 9A and B, respectively. This analysis (eqn 5) enabled the quantification of the lymphatic constriction and facilitated comparison to the arteriolar myogenic response as a reference point. The percentage change in EDD of two-valve lymphatic segments was plotted against the change in intraluminal pressure (P_L) experienced by the vessel as well as against the change in P_out, as we hypothesized that the output segment initiated the constriction of the input segment. However, the graph of percentage change in EDD of single-valve segments (Fig. 9B) was necessarily plotted only against the change in P_out, as P_L for the input portion never deviated from the baseline value and a graph against P_L would be represented by a vertical line at x= 2.

Figure 9. Analysis of end diastolic diameter as a function of pressure.

A, for two-valve collecting lymphatic segments (n= 18) subjected to step increases in output pressure, the change in EDD (%) was plotted against the change in intralymphangion pressure (P_L, cmH₂O) measured by the servo-null micropipette. Therefore, P_L data were necessarily grouped as they were not controlled, but measured. Fifteen of the 18 vessels constricted in response to an elevation in pressure. As shown, the relationship was linear to 3.5 cmH₂O, but the constriction of most vessels waned at a P_L of 4 cmH₂O, which corresponds to a P_out of 16 cmH₂O. Thus, data from this pressure were excluded from the line fit. B, change in EDD (%) from single-valve collecting lymphatic segments was plotted against the change in output pressure (P_out), as there was no change in the intraluminal pressure. This relationship was curvilinear and less pronounced than that shown in A, presumably due to the fact that single-valve segments experienced an increase only in the frequency of contractions, but not in intraluminal pressure, as this was shunted out of the input pipette. Where standard error bars appear to be missing, they are contained within the individual data point.

For the two-valve segments exposed to pressure steps in this study (n= 15), a negative change in EDD was observed at each P_L generated by the lymphangion. The magnitude of the constriction was linearly and inversely proportional to P_L (Fig. 9A, r²= 0.83). Only the two-valve experiments yielded a change in intraluminal pressure for the vessel, due to the presence of a complete lymphangion, thereby allowing the myogenic index to be calculated. The slope of this linear relationship yielded a lymphatic myogenic index of −2.4 ± 0.4 (in units of cmH₂O), or −3.3 ± 0.6 in units of mmHg, in response to selective output pressure elevation, exceeding the lymphatic myogenic index of −0.75 (in units of mmHg) in response to combined pressure steps by more than 4-fold (Davis et al. 2009a; Souza-Smith et al. 2010). When the same data were plotted against P_out, a curvilinear relationship yielded a lower magnitude constriction compared to the plot against P_L, reaching a plateau at ∼4 cmH₂O.

The data from single-valve lymphatic segments plotted in Fig. 9B followed a similar curvilinear trend (r²= 0.99, n= 12), but appeared to reach a plateau at ∼10 cmH₂O. Because we found that the one-valve lymphatic vessel constriction occurs as a result of an increase in contraction frequency (i.e. there is no rise in pressure in the input portion), the plateau of the curve appeared consistent with the frequency increase reaching its limit.

Discussion

The data presented here support our hypothesis that collecting lymphatic vessels exposed to selective elevation in downstream pressure consistently produce a constriction that is greater in magnitude than that of lymphatic vessels exposed to combined increases in input and output pressure. The myogenic index calculated for collecting lymphatics exposed to physiological pressure gradients also exceeds that of the arteriolar myogenic constriction, probably as a consequence of the relatively lower working pressures in the lymphatic vasculature and ability to increase contraction frequency. Furthermore, pharmacological inhibition of the output half of single-valve segments revealed a requirement for a signal to travel through the muscle layer, across the valve, to trigger a constriction. That this newly identified lymphatic constriction requires an active muscle layer further supports the hypothesis that it is similar to the arteriolar myogenic response.

Comparison of lymphatic constriction to selective output pressure elevation versus combined pressure elevation

Supporting our hypothesis, the responses of rat mesenteric collecting lymphatics exposed to downstream pressure elevation exceed the constriction of vessels exposed to combined pressure elevation (myogenic indices of −3.3 ± 0.6 vs. −0.75, respectively, calculated in units of mmHg). However, a strict calculation of the lymphatic myogenic index fails to take into account that the lymphatic constriction possesses two components, a pressure-dependent myogenic reduction in EDD and an increase in contraction frequency that decreases the time for diastolic relaxation and further reduces EDD (i.e. equivalent to negative lusitropy in vivo). If we make the assumption that these two components are additive, then we can subtract the slope of the line fit to the first three data points in Fig. 9B (dashed line, r²= 0.99, slope =−0.88 in units of mmHg) from that in Fig. 9A to determine the true ‘pressure-sensitive’ myogenic index. When such an analysis is performed, we obtain a corrected lymphatic myogenic index of −2.4 (in units of mmHg), which still represents a more than 3-fold increase over the response of lymphatic vessels exposed to combined pressure elevation.

Nature of the downstream signal that triggers the lymphatic constriction

To explore the mechanisms underlying the initiation of the lymphatic myogenic constriction, we selectively treated just the output half of a single-valve lymphatic segment with ML-7 to inhibit myosin light chain kinase, or cromakalim to hyperpolarize the muscle layer. This experimental approach stemmed from the prior observation that raising pressure in the output half of a one-valve lymphatic segment induced a constriction in the input half that was not exposed to any pressure change. Both treatments prevented the constriction of the input segment in response to P_out elevation, in support of the notion that the constriction of one-valve segments arises only from an increase in contraction frequency. From the traces in Figs 7 and 8, we know that the drugs did not contaminate the input segment because it continued to contract. However, abolishing spontaneous contractions in the output segment per se did not prevent the lymphatic myogenic constriction, as cromakalim treatment prevented the constriction without inhibiting spontaneous contractions. The dose of ML-7 used here is known to inhibit myosin light chain kinase completely (Nepiyushchikh et al. 2011), the activity of which is required for maintaining basal tone; as a result the output segment was maximally dilated. This in turn probably rendered the output portion insensitive to stretch upon further P_out elevation. Additionally, we conclude that the signal transmitted from the output segment required depolarization of the muscle layer, as the hyperpolarizing action of cromakalim completely blocked the lymphatic myogenic constriction. One may argue that the signal could have travelled through the endothelial layer, rather than the muscle layer. This is unlikely here because a relatively high concentration of cromakalim (3 μm) was required to block the constriction, which probably diffused across the endothelium to the muscle layer. Also, a dose of 1 μm cromakalim, which is sufficient to maximally hyperpolarize endothelial cells (Savineau & Marthan, 1993), did not inhibit transmission of the constriction to the input portion in response to P_out elevation, even though it was in direct contact with the endothelial layer (data not shown). Finally, a study of guinea pig mesenteric lymphatics (von der Weid & Van Helden, 1997) found that the endothelial layer possessed a resting membrane potential of −70 mV, much more hyperpolarized that the muscle layer, and that K_ATP channels did not contribute to this high potential. Taking these data together, we propose that P_out elevation stretches the output muscle layer to induce a depolarization that is transmitted upstream to the input segment, leading to a synchronized constriction of both segments.

Comparison of the arteriolar and lymphatic constrictions

Traditionally, the arteriolar myogenic constriction is defined as a decrease in arteriolar diameter following an increase in intraluminal pressure. Resistance arterioles in vivo can be exposed to increases in mean blood pressure (from the arterial side of the vasculature) or conversely to increases in venous pressure. Venous pressure elevation elicits a potent arteriolar myogenic constriction and can result in as much as a 40% increase in tone (Davis & Gore, 1985). The arteriolar myogenic index has been reported for most tissue beds across a variety of species (e.g. hamster, rat, cat, rabbit, pig) with values typically falling between −0.1 and −0.85 (in units of mmHg) when assessed in vitro (Davis, 1993). Interestingly, arterioles in the rat intestinal muscle possess a myogenic index ranging from −1.4 to −2.6 (mmHg) when venous pressure was elevated in vivo (Davis & Gore, 1985). For the rat mesenteric collecting lymphatic vessels assessed in this study, we obtained a lymphatic myogenic index of −3.3 ± 0.6 (in mmHg), similar to 2nd and 3rd order arterioles from rat intestinal muscle, but exceeding the in vitro responses of many other types of arterioles.

The well-espoused function of the arteriolar myogenic response is to buffer the capillary bed against increases in blood pressure or flow that would otherwise pathologically alter capillary filtration and microvascular exchange (Johnson, 1981; Scallan et al. 2010; Davis, 2012). Does the lymphatic myogenic constriction fulfil a similar role? In two-valve lymphatic vessels, where P_out was selectively elevated in a ramp-wise fashion, we discovered that the output valve would sometimes become dysfunctional by failing to close at the end of systole (Fig. 10). Immediately preceding this phenomenon, the lymphatic myogenic constriction waned and the diastolic intralymphangion pressure rose (see dashed line). At a critical point, the output valve failed to close and pressure in the lymphangion equilibrated with P_out. Hence, one possible function of this constriction is to permit proper valve closure after each spontaneous contraction in the face of an adverse pressure gradient. Maintenance of normal valve gating probably protects the lymphatic capillaries from the consequences of lymph backflow, which would occur as a chain of collecting lymphatic valves failed in series. Thus it appears that the lymphatic myogenic constriction protects the lymphatic vasculature from an elevation in pressure, not by modulating intersegmental resistance to flow but by preserving normal valve gating.

Figure 10. Waning of the lymphatic constriction predisposes collecting lymphatics to valve dysfunction.

When the output pressure of a two-valve lymphatic segment was elevated progressively in a ramp-wise fashion, a constriction occurred (see arrow). While the vessel was constricting, normal valve function persisted (not shown). After the constriction reached a maximum, the vessel diameter began to dilate in the face of a larger pressure gradient (after dotted vertical line). As myogenic tone was lost, the intralymphangion pressure rose (see pressure trace) until it reached a point where the valve could not close (densitometer trace). At this point pressure equilibrated across the valve. The valve did not close again until the experimenter lowered output pressure a significant amount.

Physiological significance

The present study shows for the first time that collecting lymphatic vessels undergo a relatively powerful constriction when exposed to an adverse pressure gradient. This lymphatic constriction is composed of two components: a pressure-dependent myogenic component and a frequency-dependent component. Furthermore, the lymphatic muscle layer conducts a signal across the valve in response to downstream pressure elevation. In conclusion, we propose that the lymphatic constriction protects individual collecting lymphatic valves from dysfunction, which would otherwise increase the pressure experienced by the next upstream lymphangion. As a consequence it protects the lymphatic capillary network against increases in pressure associated with lymph backflow. Increased hydrostatic pressure in the lymphatic capillaries would be expected to disrupt the Starling forces that usually favour absorption of interstitial fluid, leading to oedema formation.

Glossary

AMP

contraction amplitude

APSS

albumin-supplemented physiological saline solution

BSA

bovine serum albumin

EDD

end diastolic diameter

ESD

end systolic diameter

FREQ

contraction frequency

MaxD

maximum passive diameter

P_f

final pressure after a step

P_i

initial pressure preceding a step

P_in

input cannula pressure

P_L

intralymphangion pressure

P_out

output cannula pressure

PSS

physiological saline solution

Author contributions

M.J.D. performed experiments investigating the response of collecting lymphatic vessels to pressure gradients, analysed data and critically revised each version of the final manuscript. J.P.S. designed and performed experiments, analysed data and wrote each version of the manuscript. J.H.W. performed some of the initial experiments revealing a lymphatic myogenic constriction, analysed data and approved the final version of the manuscript.

Supplementary material

Supplemental Movie 1

Supplemental Movie 2