The modulation of developmental biochemical pathways by mechanical cues is an emerging feature of animal development, but its evolutionary origins have not been explored. Here we show that a common mechanosensitive pathway involving β-catenin specifies early mesodermal identity at gastrulation in zebrafish and Drosophila. Mechanical strains developed by zebrafish epiboly and Drosophila mesoderm invagination trigger the phosphorylation of β-catenin–tyrosine-667. This leads to the release of β-catenin into the cytoplasm and nucleus, where it triggers and maintains, respectively, the expression of zebrafish brachyury orthologue notail and of Drosophila Twist, both crucial transcription factors for early mesoderm identity. The role of the β-catenin mechanosensitive pathway in mesoderm identity has been conserved over the large evolutionary distance separating zebrafish and Drosophila. This suggests mesoderm mechanical induction dating back to at least the last bilaterian common ancestor more than 570 million years ago, the period during which mesoderm is thought to have emerged.

Heparan sulphate (glucosamine) 3-O-sulphotransferase 2 (HS3ST2, also known as 3OST2) is an enzyme predominantly expressed in neurons wherein it generates rare 3-O-sulphated domains of unknown functions in heparan sulphates. In Alzheimer's disease, heparan sulphates accumulate at the intracellular level in disease neurons where they co-localize with the neurofibrillary pathology, while they persist at the neuronal cell membrane in normal brain. However, it is unknown whether HS3ST2 and its 3-O-sulphated heparan sulphate products are involved in the mechanisms leading to the abnormal phosphorylation of tau in Alzheimer's disease and related tauopathies. Here, we first measured the transcript levels of all human heparan sulphate sulphotransferases in hippocampus of Alzheimer's disease (n = 8; 76.8 ± 3.5 years old) and found increased expression of HS3ST2 (P < 0.001) compared with control brain (n = 8; 67.8 ± 2.9 years old). Then, to investigate whether the membrane-associated 3-O-sulphated heparan sulphates translocate to the intracellular level under pathological conditions, we used two cell models of tauopathy in neuro-differentiated SH-SY5Y cells: a tau mutation-dependent model in cells expressing human tau carrying the P301L mutation hTau(P301L), and a tau mutation-independent model in where tau hyperphosphorylation is induced by oxidative stress. Confocal microscopy, fluorescence resonance energy transfer, and western blot analyses showed that 3-O-sulphated heparan sulphates can be internalized into cells where they interact with tau, promoting its abnormal phosphorylation, but not that of p38 or NF-κB p65. We showed, in vitro, that the 3-O-sulphated heparan sulphates bind to tau, but not to GSK3B, protein kinase A or protein phosphatase 2, inducing its abnormal phosphorylation. Finally, we demonstrated in a zebrafish model of tauopathy expressing the hTau(P301L), that inhibiting hs3st2 (also known as 3ost2) expression results in a strong inhibition of the abnormally phosphorylated tau epitopes in brain and in spinal cord, leading to a complete recovery of motor neuronal axons length (n = 25; P < 0.005) and of the animal motor response to touching stimuli (n = 150; P < 0.005). Our findings indicate that HS3ST2 centrally participates to the molecular mechanisms leading the abnormal phosphorylation of tau. By interacting with tau at the intracellular level, the 3-O-sulphated heparan sulphates produced by HS3ST2 might act as molecular chaperones allowing the abnormal phosphorylation of tau. We propose HS3ST2 as a novel therapeutic target for Alzheimer's disease.