Differential impact of lymphatic outflow pathways on cerebrospinal fluid homeostasis

Abstract

Dysfunctional lymphatic drainage from the central nervous system (CNS) has been linked to neuroinflammatory and neurodegenerative disorders, but our understanding of the lymphatic contribution to CNS fluid autoregulation remains limited. Here, we studied forces that drive the outflow of the cerebrospinal fluid (CSF) into the deep and superficial cervical lymph nodes (dcLN and scLN) and tested how the blockade of lymphatic networks affects CNS fluid homeostasis. Outflow to the dcLN occurred spontaneously in the absence of lymphatic pumping and was coupled to intracranial pressure (ICP), whereas scLN drainage was driven by pumping. Impaired dcLN drainage led to elevated CSF outflow resistance and delayed CSF-to-blood efflux despite the recruitment of the nasal-to-scLN pathway. Fluid regulation was better compensated after scLN obstruction. The dcLN pathway exhibited steady, consistent drainage across conditions, while the nasal-to-scLN pathway was dynamically activated to mitigate perturbances. These findings highlight the complex physiology of CSF homeostasis and lay the groundwork for future studies aimed at assessing and modulating CNS lymphatic function.

Figure 1.

A combination of extrinsic forces and contractile pumping drives lymphatic outflow of CSF in scLN and dcLN pathways. (A) Schematic of cranial lymphatic anatomy showing: (a) initial lymphatics in the nasal epithelium, (b) collecting lymphatics leading to the scLN, (c) initial intracranial meningeal lymphatics leading to the (f) dcLN via the (e) nasopharyngeal lymphatic plexus and (d) along the jugular vein. Boxes showing the location of (g) scLN- and (h) dcLN-afferent cLV imaging. (i) Inset of a single lymphangion showing contractile center (arrowhead) and flanking lymphatic valves (asterisks). (B) Representative frame captures of collecting vessel contractility afferent to the scLN and dcLN indicating the contractile center of the lymphangion (solid arrows) and flanking lymphatic valves (asterisks). Scale = 100 µm. (C) Representative traces from individual animals of cLV diameter afferent to the scLN and dcLN. (D) Peak amplitude in lymphatic pumping in the typical range of 2–20 min⁻¹ (percent mean diameter, Lomb-Scargle periodogram, n = 10, 8, Student’s t test, ***, P = 0.00035). (E) Amplitude of mean max-min vessel diameter in lymphatic contractility of scLN and dcLN afferents (percent mean diameter, n = 10, 8, Student’s t test, ***, P = 0.00025). (F) Frequency of spectral amplitude peaks in lymphatic contractility in the typical range of 2–20 min⁻¹ (Lomb-Scargle periodogram, n = 10, 8, Student’s t test, ***, P = 0.00012). (G) Representative frame of fluorescent microsphere drainage into the dcLN used for particle velocimetry with the dcLN (dashed line) and dcLN-afferent cLV (solid lines) indicated. Scale = 200 µm. (H) Representative traces from individual animals of mean particle velocity over time in dcLN afferents showing rapid oscillatory flow. (I) Representative frames of collecting lymphatic valve opening and closing in dcLN and scLN afferents (red: raw image signal, green: filtered image signal). Scale = 50 µm. (J) Representative traces of relative lymphatic valve movement in dcLN and scLN afferents with 0 indicating the mean inter-leaflet distance. (K) Frequency of peak amplitude in lymphatic valve opening in the range of 2–200 min⁻¹ (Lomb-Scargle periodogram, n = 4, 3, Student’s t test, *, P = 0.011). (L) Representative frame captures of GCaMP6f calcium indicator fluorescence in LMCs in scLN- and dcLN-afferent lymphatic vessels. Scale = 100 µm. (M) Kymographs of GCaMP6f calcium indicator fluorescence in LMCs along scLN- and dcLN-afferent lymphatic vessels. (N) Representative traces from individual animals of vessel diameter and GCaMP6f ∆F/F₀ in scLN- and dcLN-afferent lymphatic vessels (n = 3, 3).

Figure S1.

Supplemental measurements of contractile pumping and extrinsic forces acting on lymphatic outflow to the scLN and dcLN . (A) Peak amplitude in lymphatic contractility in the typical range of 2–20 min⁻¹ (µm deviation from mean diameter, Lomb-Scargle periodogram, n = 10, 8, Student’s t test, ****, P = 0.0001). (B) Amplitude of mean max-min vessel diameter in lymphatic contractility of scLN and dcLN afferents (µm deviation from mean diameter, n = 10, 8, Student’s t test, ***, P = 0.0001). (C) Mean diameter of measured lymphatic vessels (n = 10, 8, Student’s t test, P = 0.79). (D) Representative Lomb-Scargle periodograms of dcLN-afferent cLV particle flow oscillation showing a peak in respiratory frequency range (mean 137.6 ± 3.3 min⁻¹). (E) Typical mouse respiratory rates (RR, 172 ± 26 min⁻¹) and heart rates (HR, 341 ± 31 min⁻¹) measured under ketamine/xylazine anesthesia (n = 5). (F) Representative ICP trace showing cardiorespiratory oscillation in the waveform (n = 3). (G) Lomb-Scargle periodogram of F showing respiratory (RP, 174 ± 4.6) and cardiac (CP, 348 ± 9.2) frequency peaks. (H) Lomb-Scargle periodograms showing valve-indicated flow oscillation in the frequency range of lymphatic pumping for scLN afferents and dcLN afferents with false alarm probability <0.05 highlighted by −log(p) (n = 4, 3). (I) Lomb-Scargle periodograms showing valve-indicated flow oscillation in the frequency range of the respiratory rate for scLN afferents and dcLN afferents with false alarm probability <0.05 highlighted by −log(p) (n = 4, 3). (J) Frequency of spectral amplitude peaks in LMC GCaMP6f fluorescent signal in the range of 0-20 min⁻¹ (Lomb-Scargle periodogram, n = 3, 3, Student’s t test, *, P = 0.024). (K) Peak amplitude of LMC GCaMP6f fluorescent signal of scLN and dcLN afferents (percent mean diameter, n = 3, 3, Student’s t test, *, P = 0.048). (L) Peak amplitude of LMC GCaMP6f vessel diameter in scLN and dcLN afferents (percent mean diameter, n = 3, 3, Student’s t test, *, P = 0.017).

Figure 2.

Postural differences control nasal-to-scLN CSF outflow through ICP changes. (A) Schematic of experimental conditions. (B) Mean ICP in head-up, level, and down positions (n = 3, one-way repeated-measures ANOVA: P = 0.0004, pairwise, paired t tests (Holm’s method): *P < 0.05, **P < 0.01). (C) Representative images of dcLN and scLN sections showing tracer drainage across posture conditions (n = 6, 6, 12, 11; scale = 200 µm). (D) Quantification of dcLN tracer drainage (ANOVA: P = 0.057). (E) Quantification of scLN tracer drainage (ANOVA: P = 0.00007, Tukey HSD: **P < 0.01, ***P < 0.005). (F) Representative horizontal and sagittal projections of light-sheet imaging of tracer distribution from cleared skulls (n = 5, 5; scale = 1 mm). (G) Quantification of F (Student’s t test, *, P = 0.046). (H) Representative horizontal view of tracer distribution in skulls after dissection (n = 6, 5, scale = 500 µm). (I) Quantification of H (Student’s t test, *, P = 0.037).

Figure S2.

Additional data on intracranial pressure and lymphatic outflow across postural changes . (A) Representative traces of ICP across tilt conditions (n = 3). (B) Representative images of lumbar lymph node sections showing tracer drainage across posture conditions (n = 6, 6, 6; scale = 200 µm). (C) Quantification of lumbar LN tracer drainage.

Figure 3.

Impairment of dcLN CSF outflow recruits the nasal-to-scLN pathway. (A) Schematic of experimental conditions. (B) Representative images of dcLN and scLN sections showing tracer drainage across ligation conditions (scale = 200 µm, n = 6, 5, 6, 6; scale = 100 µm). (C) Quantification of dcLN tracer drainage (ANOVA: P = 0.00073, Tukey HSD: *P < 0.05, **P < 0.01). (D) Quantification of scLN tracer drainage (ANOVA: P = 0.000083, Tukey HSD: ***P < 0.001). (E) Representative sagittal projections of light-sheet imaging of tracer distribution from cleared skulls (n = 6, 5, 6, 6; scale = 1 mm). (F) Quantification of E (ANOVA: P = 0.0034, Tukey HSD: *P < 0.05, **P < 0.01). (G) Representative horizontal view of tracer distribution in skulls after dissection (n = 6, 5, 6, 6; scale = 500 µm). (H) Quantification of G (ANOVA: P = 0.0000073, Tukey HSD: *P < 0.05, **P < 0.01, ****P < 0.0001).

Figure 4.

dcLN and dual-lymphatic impairment impedes CSF-to-blood outflow and decreases CSF throughput. (A) Representative final frames (t = 60 min) of femoral vein showing blood content of i.c.m.-injected DB53 tracer (n = 5, 5, 6, 6; scale = 500 µm). (B) Mean traces of DB53 fluorescent intensity in blood at the femoral vein over 60 min after i.c.m. injection (mixed ANOVA: P_group = 0.0019, P_t = 1.4*10⁻¹⁰, P_group:t = 0.0093). (C) Comparison of area-under-curve integration of B (ANOVA: P = 0.0019, Tukey HSD: *P < 0.05, **P < 0.01). (D) Mean ICP across ligation groups (ANOVA: P = 0.70, n = 7, 7, 7, 6). (E) Mean ICP traces during constant-rate infusion (1–7 min: 0.5 μl/min, 7–13 min: 1 μl/min, 13–19 min: 2 μl/min, 19–25 min: 4 μl/min, 25–31 min: 8 μl/min). (F) Mean CSF outflow resistance across ligation conditions (ANOVA: P = 0.00091, Tukey HSD: *P < 0.05, **P < 0.01). (G) Mean CSF outflow resistance at each infusion rate (See Fig. S3 B).

Figure S3.

Additional data on the impact of cervical lymphatic ligation on CSF-to-blood transit and intracranial outflow resistance . (A) Comparison of final (t = 75 min) DB53 fluorescence intensity of Fig. 4 B (ANOVA: P = 0.0012, Tukey HSD: *P < 0.05, **P < 0.01). (B) Table of significant pairwise t tests from Fig. 4 G corrected for multiple comparisons with Holm’s method. (C) Plot of each stable ICP against infusion rate from Fig. 4 E, linear fit of which gives the overall outflow resistance in Fig. 4 F.