Mechanisms of Connexin-Related Lymphedema

Abstract

Rationale: Mutations in GJC2 and GJA1, encoding Cxs (connexins) 47 and 43, respectively, are linked to lymphedema, but the underlying mechanisms are unknown. Because efficient lymph transport relies on the coordinated contractions of lymphatic muscle cells (LMCs) and their electrical coupling through Cxs, Cx-related lymphedema is proposed to result from dyssynchronous contractions of lymphatic vessels.

Objective: To determine which Cx isoforms in LMCs and lymphatic endothelial cells are required for the entrainment of lymphatic contraction waves and efficient lymph transport.

Methods and results: We developed novel methods to quantify the spatiotemporal entrainment of lymphatic contraction waves and used optogenetic techniques to analyze calcium signaling within and between the LMC and the lymphatic endothelial cell layers. Genetic deletion of the major lymphatic endothelial cell Cxs (Cx43, Cx47, or Cx37) revealed that none were necessary for the synchronization of the global calcium events that triggered propagating contraction waves. We identified Cx45 in human and mouse LMCs as the critical Cx mediating the conduction of pacemaking signals and entrained contractions. Smooth muscle-specific Cx45 deficiency resulted in 10- to 18-fold reduction in conduction speed, partial-to-severe loss of contractile coordination, and impaired lymph pump function ex vivo and in vivo. Cx45 deficiency resulted in profound inhibition of lymph transport in vivo, but only under an imposed gravitational load.

Conclusions: Our results (1) identify Cx45 as the Cx isoform mediating the entrainment of the contraction waves in LMCs; (2) show that major endothelial Cxs are dispensable for the entrainment of contractions; (3) reveal a lack of coupling between lymphatic endothelial cells and LMCs, in contrast to arterioles; (4) point to lymphatic valve defects, rather than contraction dyssynchrony, as the mechanism underlying GJC2- or GJA1-related lymphedema; and (5) show that a gravitational load exacerbates lymphatic contractile defects in the intact mouse hindlimb, which is likely critical for the development of lymphedema in the adult mouse.

Keywords: calcium signaling; connexins; lymph; lymphatic system; lymphedema; vascular disease.

Figure 1.. Intercellular coupling through connexins in the endothelial cell layer is dispensable for the generation and coordination of the spontaneous contractions of lymphatic vessels.

(A) A Space Time Map (STM) was generated from high-speed, bright-field videos of the spontaneous contractions of a popliteal lymphatic vessel from a WT (C57BL/6) mouse; the conduction speed of the propagated contraction wave (see also Online Video I) was measured by edge detection of the wave front at each horizontal pixel followed by a linear fit of the detected points. Red open circle indicates approximate location of the initiation site for the contraction. The top blue trace represents the outside diameter from a single location specified by the blue dotted box. (B) A Space Frequency Map (SFM) showing the main contraction frequency components along the lymphatic vessel was generated from a Fourier transform of the corresponding STM. (C-E) Representative STMs of entrained, rapidly-conducting spontaneous contractions in popliteal lymphatics from Cx37^−/−, Cx47^−/− and Lyve1-Cre;Cx43^fx/fx mice, respectively. Yellow arrows indicate direction of propagation of the contraction waves. Bright-field images of each isolated lymphatic vessel preparation are shown at the left, aligned next to their corresponding STM. Lymphatic valves are indicated with black arrows. (F-I) Mean contractile parameters of popliteal lymphatic vessels from control mice and mice deficient in specific endothelial connexin isoforms (i.e. Cx37, Cx43, or Cx47) at 3 cmH₂O intraluminal pressure: conduction speed, number of initiation sites for contractions per unit length, percent conduction length associated with the propagation of contractions, and ejection fraction. No significant differences were found, when comparing all groups and their corresponding controls, using a 1-way ANOVA followed by Dunnett’s multiple comparison test with P<0.05 (n represents number of animals).

Figure 2.. L-type Ca²⁺ channels are essential for the large amplitude, coordinated Ca²⁺ events that precede and drive spontaneous contractions.

(A) A Space Time Map (STM) generated from GFP-fluorescence associated with the Ca²⁺ flashes that proceeded contractions in a popliteal lymphatic from a mouse expressing GCaMP6f in the muscle layer (Smmhc-Cre;GCaMP6f) (see Online Videos II and III). Conduction speed of the Ca²⁺ flash was nearly identical to that of the contraction wave when both were measured in the same vessel segment. (B) STM of Ca²⁺ flashes before and after their inhibition by the L-type Ca²⁺ channel blocker nicardipine (300 nM). After inhibition of Ca²⁺ flashes, various types of underlying intracellular Ca²⁺ events are revealed. (C) A single-cell STM showing at least two different types of Ca²⁺ events. See also Online Video IV.

Figure 3.. Smooth muscle and endothelial cell layers appear to be uncoupled.

(A-B) Membrane potential recordings from sharp electrodes for a LMC and a LEC, respectively, showing rhythmic firing of action potentials in an LMC but a much more hyperpolarized and quiescent LEC, despite residual contractions in the muscle layer. Contractions were blunted by wortmannin to permit stable V_m recording without dislodging the electrode. (C-D) Representative single-cell Ca²⁺ STMs in a LMC and a LEC from popliteal lymphatics from Smmhc-CreER^T2;GCaMP6f and Prox1-CreER^T2;GCaMP6f mice, respectively. LMCs show rhythmic action potentials that trigger the cascade-like influx of Ca²⁺ (flashes) through L-type Ca²⁺ channels, which drive spontaneous contractions. Ca²⁺ flashes and other intracellular Ca²⁺ events in the LMC layer are not transmitted into the LEC layer. (E-G) Endothelium specific intracellular Ca²⁺ events in popliteal lymphatic vessels from Prox1-CreER^T;GCaMP6f mice. STMs from Ca²⁺ imaging under (F) basal conditions, showing very rare spontaneous activity and (G) after incubation with 10 nM ACh, which induces Ca²⁺ events in every cell. See also Video V.

Figure 4.. Similar connexin expression in human and mouse lymphatics.

Representative gels showing the mRNA expression for a panel of 9 different vascular connexin isoforms using RT-PCR on intact isolated human mesenteric (A and B, n=3) and mouse popliteal (C and D, n=3) lymphatic vessels, as well as their corresponding positive controls (human jejunum wall and mouse brain, respectively). (E-G) Immunostaining for Cx45, smooth muscle actin (SMA), and CD31 (PECAM) in transverse sections of mouse popliteal lymphatic vessels. (H-J) Cx45 appears to be expressed exclusively in LMCs as evident by its colocalization with SMA but not CD31.

Figure 5.. Human mesenteric lymphatics express Cx45 in LMCs and exhibit strong, entrained, rapidly-conducting spontaneous contractions.

(A-C) Immunofluorescence images showing Cx45 expression (co-staining with SMA and DAPI) in the LMC layer of human mesenteric lymphatic vessels (whole-mount). (D) Representative STM and SFM (at 3 cmH₂O intraluminal pressure) showing entrained, rapidly-conducting spontaneous contractions of human mesenteric lymphatic vessels (see also Online Video VIII). (E) Side-by-side comparison of the mean conduction speed of propagated contraction waves (from bright-field videos of contractions) for human and mouse lymphatics pressurized to 3 cmH₂O. (F-H) Contractile parameters (number of contraction initiation sites, contraction frequency, and ejection fraction) as a function of intraluminal pressure. Significant differences (*) were assessed using 1-way ANOVA followed by Dunnett’s multiple comparison test with P<0.05 (n represents number of individuals/patients and animals respectively). Additional contractile parameters (i.e. contraction amplitude, tone, end diastolic diameter, and fractional pump flow) are shown in Online Figure VII.

Figure 6.. Lymphatic vessels from Cx45-deficient mice exhibit loss of coordination of the spontaneous contractions.

STMs and SFMs showing entrained, rapidly-conducting spontaneous contractions from (A) a control (WT) popliteal lymphatic, in contrast to those showing non-coordinated, slowly-conducting, aberrantly-propagating contractions from popliteal lymphatics isolated from (B and C) Cx45 deficient mice (Nestin-Cre;Cx45^fx/fx and Smmhc-CreER^T2;Cx45^fx/fx, respectively). See also Online Video IX.

Figure 7.. Impaired ejection fraction and slower, aberrantly propagating contraction waves in popliteal lymphatic vessels lacking Cx45 in smooth muscle cells.

(A) Conduction speed, (B) number of initiation sites for spontaneous contractions per unit length, and (C) percent conduction length associated with the propagation of initiated contractions for popliteal lymphatic vessels from Cx45-deficient mice (Nestin-Cre;Cx45^fx/fx and Smmhc-CreER^T;Cx45^fx/fx[TMX](6–11 days post tamoxifen-induction)) and their corresponding controls. The spontaneous contractions of lymphatics from Cx45-deficient mice are uncoordinated, as evident by the increased number of contraction initiation (pacemaking) sites. The impaired electrical coupling between LMCs resulted in a significantly reduced number of LMCs that can be recruited by each pacemaker, leading to: contraction waves with decreased percent conduction length and decreased (D) ejection fraction. Electrical coupling between LMCs appears to alter (E) contraction frequency and (F) tone. These contractile parameters were obtained in vessels pressurized to 3 cmH₂O. Mean values and corresponding SEMs are reported in Online Table I. Significant differences were evaluated at P<0.05 using 1-way ANOVA followed by Dunnett’s multiple comparison test. The symbol * indicates a significant difference between the specified group(s) and the corresponding direct controls (Cx45^fx/fx or Cx45^fx/fx[TMX] respectively). When comparing different time-points (6–11 days, 2–4 weeks, and 9-week groups) the symbol ^# indicates that the specified group was significantly different when compared to a second specified group or groups from all other time-points.

Figure 8.. The ability of lymphatic vessels to pump fluid in the presence of an adverse pressure gradient is impaired in vessels from LMC-specific Cx45-deficient mice.

Assessment of the contractile function and lymphatic ability to transport fluid forward when pumping against an adverse pressure gradient in popliteal vessels from control (Cx45^fx/fx[TMX]) and Cx45-deleted (Smmhc-CreER^T2;Cx45^fx/fx[TMX]) mice. All vessels tested 6–11 days post induction with tamoxifen. (A) Diagram of Pump-Test preparation (see also Online Video X), (B) adverse pressure gradient limit from Pump-Tests (pressure level at which lymphatic vessels failed to open the downstream/output valve), and representative traces (top section: valve opening, middle section: input (blue line), output (red line), and middle lymphangion intraluminal pressures (black line), and bottom section: tracked luminal diameter) for (C) Cx45^fx/fx[TMX] and (D) Smmhc-CreER^T2;Cx45^fx/fx[TMX] respectively.

Figure 9.. Lack of coordinated spontaneous contractions in popliteal lymphatic vessels in vivo from mice with SMC-specific Cx45 deficiency.

Assessment of the contractile activity of popliteal lymphatic vessels (via fluorescence imaging following FITC-solution injection at the dorsal aspect of the foot) from control and Cx45-deficient (Nestin-Cre;Cx45^fx/fx and Smmhc-CreER^T2;Cx45^fx/fx) mice in vivo. STMs of the in vivo contractions of popliteal collecting lymphatics from (A) WT, (B) Nestin-Cre;Cx45^fx/fx, and (C) Smmhc-CreER^T2;Cx45^fx/fx mice respectively. In contrast to the rapidly propagating, highly entrained contractions observed in WT mice, contractions of popliteal lymphatics from mice deficient in Cx45 were uncoordinated and slowly propagating. (D-G) Conduction speed, normalized number of different initiation sites for contractions, conduction length of propagating contraction waves, and mean length of in vivo imaged popliteal lymphatic segments.

Figure 10.. In vivo lymph transport in popliteal lymphatics of SMC-specific Cx45 deficient mice is inhibited when a gravitational hydrostatic load is imposed.

(A) NIRF imaging of (mouse hindlimb) popliteal afferent lymphatics following infrared dye (IRDye® 800CW PEG Contrast Agent) injection in the dorsal aspect of the foot. (B) Assessment of lymph transport. Transport speed was determined by tracking the position of the wavefront (at 1 fps) after dye uptake had initiated. Time-lapse images of three vessels (labeled 1–3) are shown in a Cx45-deficient mouse in the horizontal position: in vessels 1 and 3, the wavefront of the dye moving downstream is marked by a dashed-circle or square respectively; vessel 2 filled with dye up to the popliteal node within 2 seconds, at an estimated speed >2 mm/s, indicating it was driven by the pressure head generated by the injection (such vessels were excluded from analysis). (C) Mean lymph transport speed and (D) normalized distance traveled (per minute) by lymph in popliteal lymphatics from control (Cx45^fx/fx) and Cx45-deficient (Nestin-Cre;Cx45^fx/fx) mice in horizontal and near-vertical (~78^o) positions. No significant differences were found in lymph transport between control and Cx45-deficient mice in the horizontal position or in control mice in the near-vertical position. However, lymph transport was completely inhibited in popliteal lymphatics from SMC-Cx45 deficient mice in the near-vertical position. Significant differences (*) were determined using a 1-way ANOVA followed by Tukey’s multiple comparison test (comparing means of all groups) with P<0.05 (n represents number of imaged limbs following a single injection with infrared dye). (*) indicates that the Nestin-Cre;Cx45^fx/fx (in the near-vertical position) group was significantly different from all other groups.