The ephrinA1 ligand exerts antioncogenic effects in tumor cells through activation and downregulation of the EphA2 receptor and has been described as a membrane-anchored protein requiring clustering for function. However, while investigating the ephrinA1/EphA2 system in the pathobiology of glioblastoma multiforme (GBM), we uncovered that ephrinA1 is released from GBM and breast adenocarcinoma cells as a soluble, monomeric protein and is a functional form of the ligand in this state. Conditioned media containing a soluble monomer of ephrinA1 caused EphA2 internalization and downregulation, dramatic alteration of cell morphology and suppression of the Ras–MAPK pathway. Moreover, soluble monomeric ephrinA1 was functional in a physiological context, eliciting collapse of embryonic neuronal growth cones. We also found that ephrinA1 is cleaved from the plasma membrane of GBM cells, an event which involves the action of a metalloprotease. Thus, the ephrinA1 ligand can, indeed, function as a soluble monomer and may act in a paracrine manner on the EphA2 receptor without the need for juxtacrine interactions. These findings have important implications for further deciphering the function of these proteins in pathology and physiology, as well as for the design of ephrinA1-based EphA2-targeted antitumor therapeutics.

An abundance of evidence exists that shows calcium channel blockade promotes injury in cultured neurons. However, few studies have addressed the in vivo toxicity of such agents. We now show that the L-type calcium channel antagonist nimodipine promotes widespread and robust injury throughout the neonatal rat brain, in an age-dependent manner. Using both isolated neuronal as well as brain slice approaches, we address mechanisms behind such injury. These expanded studies show a consistent pattern of injury using a variety of agents that lower intracellular calcium. Collectively, these observations indicate that postnatal brain development represents a transitional period for still developing neurons, from being highly sensitive to reductions in intracellular calcium to being less vulnerable to such changes. These observations directly relate to current therapeutic strategies targeting neonatal brain injury.

Loss of neuronal calcium is associated with later apoptotic injury but observing reduced calcium and increased apoptosis in the same cell would provide more definitive proof of this apparent correlation. Thus, following exposure to vehicle or the calcium chelator BAPTA (1-20 μM), primary cortical neurons were labeled with Calcium Green-1 which was then cross-linked with EDAC, prior to immunostaining for various proteins. We found that BAPTA-induced changes in calcium were highly correlated with changes in expression of activated caspase-3 as well as the calcium binding proteins calbindin, calretinin, and parvalbumin. Additionally, in brain slices from P7 neonatal rats, BAPTA induced significant loss of calcium in a brain region we have previously shown to express only moderate levels of calcium binding proteins as well as display robust apoptosis following calcium entry blockade. In contrast, BAPTA had little influence on calcium levels in a brain region we have previously shown to express robust calcium binding proteins as well as display far less apoptosis following calcium entry blockade. These data suggest that the ability of developing neurons to buffer changes in calcium may be critical to their long-term survival.

It has become increasingly clear that agents that disrupt calcium homeostasis may also be toxic to developing neurons. Using isolated primary neurons, we sought to understand the neurotoxicity of agents such as MK801 (which blocks ligand-gated calcium entry), BAPTA (which chelates intracellular calcium), and thapsigargin (TG; which inhibits the endoplasmic reticulum Ca 2+ -ATPase pump). Thus, E18 at cotical neuons wee grown for 1 day in vitro (DIV) and then exposed to vehicle (0.1% DMSO), MK801 (0.01-20 μM), BAPTA (0.1-20 μM), or TG (0.001-1 μM) for 24 h. We found that all three agents could profoundly influence early neuronal maturation (growth cone expansion, neurite length, neurite complexity), with the order of potency being MK801 < BAPTA < TG. We next asked if cultures exposed to these agents were able to re-establish their developmental program once the agent was removed. When we examined network maturity at 4 and 7 DIV, the order of recovery was MK801 > BAPTA > TG. Thus, mechanistically distinct ways of disrupting calcium homeostasis differentially influenced both short-term and long-term neuronal maturation. These observations suggest that agents that act by altering intracellular calcium and are used in obstetrics or neonatology may be quite harmful to the still-developing human brain.