A deeper dive into amyloid clearance by meningeal lymphatic vessels

Acute depletion of meningeal lymphatic vessels impairs the clearance of cerebrospinal fluid and brain macromolecules. A new study by Antila et al. shows that amyloid pathology in Alzheimer’s disease is neither improved nor aggravated by genetic expansion or depletion of meningeal lymphatic vessels.

Amyloid-β (Aβ)-clearing therapies have emerged as a promising strategy to slow the progression of Alzheimer’s disease (AD), the leading cause of cognitive impairment in older people and a huge unmet medical need¹. Aβ immunotherapy has shown promise in preserving cognition in individuals with mild Alzheimer’s disease, but the role of the recently rediscovered meningeal lymphatic vessels that draining cerebrospinal fluid (CSF) into the deep cervical lymph nodes (dCLN) has gained increasing attention following initial evidence that CSF clearance is impaired in Alzheimer’s disease and that Aβ accumulates in dCLN^2,3. These observations in humans have been complemented by an emerging body of work in mouse models of amyloid accumulation demonstrating that meningeal lymphatics contribute greatly to clearing Aβ from the brain (Fig. 1 and Table 1). In this issue of Nature Cardiovascular Research, Antila et al.⁴ counter these previous observations by providing extensive evidence that meningeal lymphatics have no role in Aβ clearance from the brain.

Fig. 1: Potential mechanisms of brain Aβ disposal.

Aβ deposition in amyloid plaques depends on the balance between phagocytosis, enzymatic degradation and clearance from the brain. Aβ is thought to reach the perivascular space that surrounds cerebral arteries and veins and be transported across the cerebrovascular wall into the blood. Aβ also reaches the CSF in the subarachnoid space through perivascular or paravascular (glymphatic) interstitial fluid flow¹¹. Once in the CSF, Aβ may (1) enter the brain’s venous outflow via arachnoid protrusions in the superior sagittal sinus, (2) be removed by lymphatics (dural, nasopharyngeal or skull base), or (3) leave the brain by traveling along cranial nerves that exit the skull, especially the olfactory nerve¹¹.

Table 1 Effects of manipulating meningeal lymphatic drainage on brain Aβ pathology

Antila et al.⁴ took two approaches to make their case. First, they examined the effect of lymphatic depletion on amyloid pathology. To this end, they crossed K14-sVEGFR3-Ig (K14-sR3) mice, which express a VEGF-C/D trap that inhibits lymphatic development, with APdE9 mice, which develop amyloid plaques owing to enhanced amyloid precursor protein (APP) processing. Surprisingly, they found no increase in brain Aβ plaque load despite marked suppression of meningeal lymphatic drainage. To rule out compensatory mechanisms arising from curtailed lymphatic development in K14-sR3 mice, the authors also induced subacute lymphatic regression using adenoviral vector (AAV) delivery of a VEGF-C trap encoding the ligand-binding domains of VEGFR3 fused to the immunoglobulin G Fc domain (mVEGFR3_1–4-Ig (sR3)) in two-month-old APdE9 mice. sR3 treatment did not increase Aβ load at 6 or 16 months of age or impair cognition at 13 months of age, despite a documented loss of meningeal lymphatic function. The same result was obtained in 5xFAD mice, a model with more aggressive amyloid pathology. The second approach was to assess whether enhancing meningeal lymphatic drainage resulted in more efficient Aβ removal from the brain. Antila et al.⁴ induced pronounced meningeal lymphangiogenesis using AAV-VEGFC in young (2-month-old) and old (12-month-old) APdE9 mice and in young 5xFAD mice, but found no reduction in Aβ load despite enhanced meningeal drainage.

These findings are in stark contrast to previous work from several independent laboratories demonstrating that modulation of meningeal lymphatics can have marked effects on brain amyloid pathology and cognition (Table 1). Reduced lymphatic function as a result of photoablation of lymphatic vessels^5,6,7 or blockage of lymph drainage⁸ into dCLN in APP/PS1, 5xFAD or J20 mice, or in a mouse model with reduced meningeal lymphatic drainage⁹ (Twist1^+/− crossed with 5xFAD), was found to enhance Aβ accumulation and/or increase cognitive deficits.

A detailed analysis of the available literature in light of the findings of Antila et al.⁴ provides clues on the potential bases for the conflicting observations and suggests a more nuanced understanding of meningeal lymphatic clearance mechanisms. The first notable difference between the studies is in the methodology, including the approaches used to modulate lymphatic drainage (for example, a developmental defect in K14-sR3 mice or subacute depletion with sR3 versus acute photoablation or ligation of lymphatic vessels draining into dCLN), the mouse models with different dynamics of amyloid production, degradation and accumulation (APdE9, APP/PS1, 5xFAD and J20), and the ages at which mice were studied.

The speed of the removal of lymphatic vessels may influence the engagement of alternative mechanisms for Aβ clearance and degradation. The disposal of brain Aβ involves a wide variety of mechanisms including phagocytosis, enzymatic degradation, trans-vascular transport into the blood, as well as CSF drainage systems by sagittal sinus arachnoid granulations, meningeal, nasopharyngeal¹⁰ and skull base lymphatics, and exiting cranial and spinal nerves¹¹ (Fig. 1). Given the multiplicity of the pathways regulating fluid balance in the brain, it is imperative to consider these mechanisms as acting in unison rather than independently of each other. Thus, the speed of the interruption of lymphatic flow by photoablation or ligation may have prevented compensation by alternative pathways, more clearly revealing the lymphatic contribution to Aβ clearance. This view is consistent with the observation by Antila et al.⁴ that subacute meningeal lymphatic depletion with sR3 in APdE9 mice increases macromolecular transfer from the CSF to the peripheral blood via pathways that are seemingly not operative prior to lymphatic deletion (figure 1o of Antila et al.⁴). The effect of chronic lymphatic depletion on cerebrovascular function also remains to be assessed. CSF flow through perivascular spaces requires intact cerebrovascular vasoactivity¹¹, which was not tested in the chronic depletion models.

Confounders related to the mouse models also deserve consideration. The role of Aβ lymphatic clearance relative to other Aβ disposal mechanisms may differ among models, which could influence the outcome of lymphatic expansion. For example, meningeal expansion in young APdE9, 5xFAD and J20 mice does not alter Aβ load or improve cognition, possibly because Aβ meningeal lymphatic clearance was already maximal relative to Aβ production and degradation rates^4,5. In 5xFAD mice, AAV-mVEGFC has no effect when delivered at two months of age⁴, but begins to have effects on Aβ plaque load at four to five months of age^6,12, when lymphatic vessels start to deteriorate¹². By contrast, meningeal lymphatics do not seem to deteriorate with age in APdE9 mice⁴, which may explain the lack of effectiveness of meningeal expansion in this model. In 9-month-old AAP/PS1 mice, meningeal expansion reduced brain levels of soluble Aβ and improved cognition, despite having no effect on Aβ plaques¹³. Thus, the balance between Aβ production, deposition, parenchymal degradation, vascular and lymphatic clearance, and the resulting cognitive dysfunction may vary in different models of cerebral amyloidosis, complicating comparisons between models.

Another question concerns whether meningeal lymphatics clear some macromolecules more efficiently than others. Meningeal depletion or expansion was able to reduce or enhance, respectively, the CSF clearance of macromolecular markers, but did not influence brain Aβ accumulation or removal. In addition, K14-sR3 mice lacking meningeal lymphatics do not efficiently clear tau injected into the brain¹⁴, but do not retain more Aβ when crossed with APdE9 mice⁴. In addition, in an independent model of neurodegeneration, lymphatic drain ligation enhances α-synuclein accumulation¹⁵. Notwithstanding the differences in animal models, these findings raise the possibility of preferential clearance of some macromolecules by the meningeal lymphatic system.

The discordant results of Antila et al.⁴ highlight the complexity of Aβ-clearing mechanisms. Animal models, age, baseline lymphatic function and methods of lymphatic ablation have a profound influence on experimental outcomes. Although this conclusion applies to any research area, the complexity, multiplicity, potential molecular selectivity and interactive nature of fluid-clearance mechanisms in the brain offer a particular challenge to investigators exploring the therapeutic potential of clearing proteins associated with neurodegeneration.

Nevertheless, promoting brain Aβ clearance has proven to be beneficial in Alzheimer’s disease¹ and approaches to enhance the elimination of toxic molecular species from the brain via meningeal lymphatics would be a valuable addition to the limited therapeutic options for Alzheimer’s disease and other neurodegenerative diseases. The observation that increasing meningeal lymphatic clearance may enhance the effectiveness of Aβ immunotherapy in mice⁶ provides an impetus to redouble efforts in this research area.

However, the way forward is obscured by the conflicting observations. The available evidence suggests that our understanding of Aβ-clearance mechanisms remains incomplete, and key questions remains unanswered (Fig. 1). For example, the relative contributions of Aβ disposal mechanisms (cellular, enzymatic, vascular and lymphatic) in different animal models, the effects of aging on the relative roles of these disposal pathways, and more generally, whether vascular or lymphatic mechanisms clear different neurodegeneration-associated molecules, such as Aβ, tau and synuclein at different rates, are unknown.

In parallel with tackling these fundamental questions, the exploration of the therapeutic potential of modulating meningeal lymphatic clearance demands a more concerted approach by the scientific community. Different laboratories with relevant expertise could use the same methods and experimental approaches to investigate meningeal clearance of neurotoxic molecules by adopting an experimental design and analysis framework similar to those of clinical trials. Such preclinical trials are already ongoing for other neurological diseases (for example, the Stroke Preclinical Assessment Network) and may shed light on the potential translational value of this highly promising and novel therapeutic approach for one of our most devastating public health challenges.