Neck Lymph Node Removal Disrupts Brain Waste Clearance and Triggers Cognitive Decline

Meningeal lymphatic vessels carry cerebrospinal fluid and brain interstitial fluid, along with the waste they hold, to the cervical lymph nodes. This drainage route has been tied to brain aging, stroke recovery, and neurodegeneration. Yet surgeons routinely remove these nodes in head and neck squamous cell carcinoma, and the effect on the brain had never been systematically examined. This Neuron study combines two patient cohorts with three rodent models to ask whether abrupt loss of lymphatic drainage harms cognition and brain protein handling.

In the Einstein-Montefiore cohort (2013-2023), 1,035 patients with head and neck squamous cell carcinoma underwent neck node resection. About 17% had lymphadenectomy without chemotherapy or radiation, and 10.2% of those had at least two years of follow-up. Among patients under 65, 18% developed mild-to-severe cognitive impairment, compared with a published age-matched risk of 4-6%. Among those 65 or older, 25% did, compared with a reported 10-22%. Patients with vascular events were excluded.

A second cohort from the Technical University of Munich included 619 patients. Of these, 190 had lymphadenectomy without chemoradiation, and 59 had pre- and multiple post-operative MRI scans over up to six years. A neuroradiologist measured temporal horn widening, a proxy for medial temporal and hippocampal tissue loss. After adjustment for age, sex, and comorbidities, bilateral dissection produced faster widening than unilateral dissection (p<0.001). The rate was 17% ± 2.7 SEM versus 4.2% ± 2.9 SEM, and the unilateral rate resembled age-matched norms.

To probe mechanism, the authors ligated the afferent lymphatics to the deep cervical node in mice and cauterized them in rats, with sham and naïve controls, and sampled tissue at one month. Cerebrospinal fluid from ligated mice showed upregulated de novo NAD+ synthesis, an activated pentose phosphate pathway, and higher short-chain acylcarnitines. It also showed more glutamate, cysteine, glycine, glutathione, and glutathione disulfide, and more extracellular ATP, ADP, AMP, calreticulin, and S100. Proteomics of 729 proteins found 153 altered (111 up, 42 down; FDR<0.05), and ELISA showed higher IL-1b, GM-CSF, CXCL2, and CCL3. Rats showed a comparable signature.

Hippocampal proteomics identified 72 differentially expressed proteins pointing to mitochondrial dysfunction, impaired synaptogenesis, and neuroinflammation. Protein carbonylation rose, though Congo red and oligomer staining showed no macro-aggregates. Field flow fractionation isolated microaggregates that made up 8 ± 2% of the injected proteome in ligated mice versus under 1% in sham. These were round and amorphous, unlike Alzheimer's fibrils. Of 880 proteins identified, 78% were 2- to 33-fold more abundant after ligation. Over 70% mapped to presynaptic mitochondria, including all five oxidative phosphorylation complexes, along with synapsins I and II. Tau, α-synuclein, and amyloid-beta were present at low counts. Most proteins scored low on the amyloid predictor ANuPP (71.4% below 200) but high on Zagg, TANGO, and GAP, and showed carbonylation, glycation, 4-hydroxynonenal adducts, and ubiquitination near aggregation hotspots.

Critically, mice with gradual meningeal lymphatic atrophy from AAV-sR3 (VEGF-C/D trap) showed no prominent hippocampal proteome changes across more than 5,000 proteins. This suggests that abruptness matters, or that compensatory drainage routes develop over time. The results point to presynaptic mitochondria as a vulnerable site for waste-driven oxidative damage. Caveats include the retrospective, observational patient data compared with published population rates rather than matched controls, small rodent groups, and the contrast between surgical ligation and the clinical setting.