The agent responsible for transmissible spongiform encephalopathies (TSEs) is thought to be a malfolded, protease-resistant version (PrPres) of the normal cellular prion protein (PrP). The interspecies transmission of bovine spongiform encephalopathy (BSE) to mice was studied. Although all of the mice injected with homogenate from BSE-infected cattle brain exhibited neurological symptoms and neuronal death, more than 55 percent had no detectable PrPres. During serial passage, PrPres appeared after the agent became adapted to the new host. Thus, PrPres may be involved in species adaptation, but a further unidentified agent may actually transmit BSE.

Various experimental models are used to study brain development and degeneration. They range from whole animal models, which preserve anatomical structures but strongly limit investigations at the cellular level, to dissociated cell culture systems that allow detailed observation of cell phenotypes but lack the highly ordered physiological neuron connection architecture. We describe here a platform comprising independent cell culture chambers separated by an array of "axonal diodes". This array involves asymmetric micro-channels, imposing unidirectional axon connectivity with 97% selectivity. It allows the construction of complex, oriented neuronal networks not feasible with earlier platforms. Different neuronal subtypes could be co-cultivated for weeks, and sequential seeding of different cell populations reproduced physiological network development. To illustrate possible applications, we created and characterized a cortico-striatal oriented network. Functional synaptic connections were established. The activation of striatal differentiation by cortical axons, and the synchronization of neural activity were demonstrated. Each neuronal population and subcompartment could be chemically addressed individually. The directionality of neural pathways being a key feature of the nervous system organization, the axon diode concept brings in a paradigmatic change in neuronal culture platforms, with potential applications for studying neuronal development, synaptic transmission and neurodegenerative disorder such as Alzheimer and Parkinson diseases at the sub-cellular, cellular and network levels.

IntroductionWe recently identi®ed the 37-kDa laminin receptor precursor (LRP) as an interactor for the prion protein (PrP) (Rieger et al., 1997; for reviews see Rieger et al., 1999;Gauczynski et al., 2001a). Employing a series of neuronal and non-neuronal cells, we proved that the 37-kDa LRP/67-kDa high-af®nity laminin receptor (LR) acts as the receptor for the cellular PrP (Gauczynski et al., 2001b). In the present manuscript we used the yeast twohybrid system and cell-binding studies on neuronal as well as non-neuronal cells involving the Semliki Forest virus (SFV) system (for reviews see Liljestrom and Garoff, 1991;Tubulekas et al., 1997) to identify domains on the PrP and the LRP involved in the PrP±LRP interaction on the cell surface. We identi®ed two binding domains for LRP on PrP termed PrPLRPbd1 and PrPLRPbd2. The ®rst one binds directly to LRP, whereas the second one depends on the presence of heparan sulfate proteoglycans (HSPGs) on the cell surface. The yeast two-hybrid system and cell-binding assays on wild-type and mutant HSPGde®cient Chinese hamster ovary (CHO) cells also identi®ed two binding domains for PrP on LRP.The relationship between 37-kDa LRP and 67-kDa LR is not yet fully understood and has been explained with homodimerization of 37-kDa LRP (Landowski et al., 1995) or an additional factor, such as a polypeptide (Castronovo et al., 1991), which might bind to 37-kDa LRP to form the 67-kDa form of the receptor. The 67-kDa heterodimer might be stabilized by hydrophobic interactions mediated by fatty acids such as palmitate, oleate and stearate bound to 37-kDa LRP and to a galectin-3 (gal-3) cross reacting polypeptide (Landowski et al., 1995;Buto et al., 1998). However, we recently proved that the b-galactoside lectin gal-3 is not present on the surface of neuronal or non-neuronal cells used for PrP-binding/ internalization studies (Gauczynski et al., 2001b) and antigal-3 antibodies failed to compete for the 37-kDa LRP/67-kDa LR-mediated binding and internalization of the cellular PrP (Gauczynski et al., 2001b), suggesting that gal-3 is not a partner of the 37-kDa LRP in this context. In this study we investigated by a yeast two-hybrid system analysis whether gal-3 interacts with 37-kDa LRP and/or the cellular PrP. In addition, we investigated whether 37-kDa LRP interacts with itself in the yeast two-hybrid and analysed the monomer/dimer status of the receptor by sizeexclusion chromatography. Both PrP (Gabizon et al., 1993;Caughey et al., 1994;Chen,S.G. et al., 1995;Brimacombe et al., 1999) and the 37-kDa/67-kDa LR (Guo et al., 1992;Kazmin et al., 2000) bind to heparan sulfates. HSPGs are required for the binding of the ®broblast growth factor (FGF) to its FGFR receptor (Yayon et al., 1991;Spivak et al., 1994;Venkataraman et al., 1999) and act as initial attachment receptors for bacteria (Chen,T. et al., 1995) and viruses including alphaviruses (Byrnes and Grif®n, 1998), human immunode®ciency virus (HIV) type 1 (Mondor et al., 1998) and vaccinia virus (Chung et al., 1998). Heparan sulfates are...