WO2002042779A2

WO2002042779A2 - Axonal response to chemotropic gradients

Info

Publication number: WO2002042779A2
Application number: PCT/US2001/041935
Authority: WO
Inventors: Jeffrey S. Urbach; Geoffrey J. Goodhill; Mark J. Esrick
Original assignee: Georgetown University
Priority date: 2000-08-30
Filing date: 2001-08-30
Publication date: 2002-05-30
Also published as: WO2002042779A3; AU2002241466A1

Abstract

In one aspect, the present invention provides an assay for the characterization of axon-growth guidance properties of a molecule comprising: e) preparing a three-dimensional collagen substrate; f) forming a concentration gradient of the molecule into the collage substrate; g) exposing the axon to the concentration gradient formed in the collagen substrate; and h) measuring a change in axon size, texture and/or shape; wherein the collagen substrate has a thickness to length ratio adjusted to stabilize the concentration gradient for a period of time sufficient to induce a change in the axon size, texture and/or shape.

Description

ASSAY FOR THE CHARACTERIZATION AND THE QUATIFICATION OF AXONAL RESPONSE TO CHEMOTROPIC GRADIENTS

BACKGROUND OF THE INVENTION

1. Related Applications

[0001] This application is based on US Provisional Application Serial No. 60/228,996, filed August 30, 2000, the contents of which are hereby incorporated by reference in their entirety.

2. Field of the Invention

[0002] The present invention relates to assays for the investigation of axon growth mechanisms. In particular, the present invention provides an assay based on stable concentration gradients for the identification of chemotropic factors and the elucidation of their role in axon growth.

3. Summary of the Related Art

[0003] An early step in the development of the brain is the growth of neuronal axons from their cell bodies of origin to their appropriate targets, to form a precise pattern of neuronal connections. The growth of axons is highly directed, as axons are guided to their targets by specific molecular guidance cues arrayed in the extracellular environment. These cues can be attractive, steering axons towards particular sources of the cues, or repulsive steering axons away from inappropriate regions. Several families of attractants and repellents, including the netrin, semphorin and slit protein families, as well receptors involved in mediating the actions of these factors are being investigated.

[0004] In the developing nervous system, axons project considerable distances along stereotyped pathways to reach their targets. Axon growth and guidance depends partly on the recognition of cell-surface and extracellular matrix cues along these pathways. The identification of such nerve cell growth and guidance cues is the holy grail of neurobiology. These are the compounds that tell neurons when to grow, where to grow, and when to stop growing. The medical applications of such compounds are enormous and include modulating neuronal growth regenerative capacity, treating neurodegenerative disease, and mapping (e.g. diagnosing) genetic neurological defects. [0005] Current research efforts are aimed at identifying the function of cues in guiding axons in vivo, and in identifying the signal transduction mechanisms through which receptor activation leads to axonal steering. In particular, axon outgrowth and path-finding is currently one of the most exciting areas of developmental neuroscience. The idea that axons could be guided by gradients was first proposed over 100 years ago, however, direct evidence for this has only recently appeared. Within the past few years, several new factors implicated in axon guidance have been identified and cloned.

[0006] A key mechanism of guidance is chemotropism. In chemotropism, axons are directed to a final or intermediate target by growing up (or down) concentration gradients of guidance factors. Axon guidance gradients may be classified as substrate-bound or diffusible. For the substrate-bound case, the concentration gradient exists in the nervous system either because the molecule is differentially expressed on the surface of cells, or because it was once free but was then bound to the cell surface or extracellular matrix in such a way as to create a gradient. The best-known examples of substrate-bound gradients are located in the retinotectal. By contrast, diffusible gradients axe set up by molecules that move more or less freely, and form a gradient by diffusing away from the region (often the target) where they are released. Target- derived diffusible factors have now been identified in a large number of systems. The evidence for this comes mostly from experiments, where a piece of target tissue is embedded in a three dimensional collagen gel near to a piece of tissue containing the appropriate population of neurons. Axon growth is then observed directed towards the target, presumably caused by the gradient of a target-derived signal.

[0007] Over decades of concentrated research, various hypotheses involving chemo- attractants and repellents, labeled pathways, cell adhesion molecules, etc. have been invoked to explain guidance. Molecules such as N-CAM and N-cadherin have been reported to provide favorable substrates for axon growth and certain sensory axons may be responsive to NGF and NGF-like factors. Recent reports suggest the existence of diffusible chemotropic molecule(s) which influence the pattern and orientation of com issural axon growth.

[0008] A better characterized axon guidance molecule that has been shown to exert a long range chemotropic effect during normal development is the 75 kDa protein netrin- 1. Netrin binds to the receptor DCC with an affinity of order 1. Netrin- 1 was originally identified from the study of axon guidance to and from the midline in both the developing brain and spinal cord. It is released from the floor plate and attracts axons of dorsally generated ventrally projecting commissural neurons towards the floor plate, cerebellofugal axons in the rostral hindbrain towards the floor plate, and alar plate axons in the myelencephalon towards the mesencephalon. Netrin- 1 also guides axons in the forebrain. Different populations of axons in the embryonic cerebral cortex project medially to form the corpus callosum or laterally to subcortical targets. The lateral projection is guided in part by a diffusible attractive cue released from the nascent internal capsule, i.e. the mantle zone of the ganglionic eminence. The fact that netrin-1 is expressed in the internal capsule, netrin-1 transfected cell lines can attract cortical axons, and a function-blocking anti-netrin-1 antiserum prevents cortical axon guidance all implicate netrin-1 as an important guidance molecule in this system. However, exactly how a growth cone converts a netrin-1 gradient into a signal for directed movement is unknown.

[0009] Netrin-1 (and the floor plate) can also act as repellents. For instance, ventrally generated trochlear motor neurons are repelled by netrin-1 and floor plate, as are dorsally projecting cranial motor axons. Attraction to a source of netrin-1 can be converted to repulsion by lowering levels of cAMP in the growth cone.

[0010] A further influence is the substrate on which the axons are growing. Xenopus retinal axons cultured on glass, poly-D-lysine or fibronectin coated coverslips are attracted to a source of netrin-1, whereas they are repelled if the substrate is laminin. Growth cones on laminin have reduced levels of cAMP, and this may play a role in guidance of retinal axons at the optic nerve head. The fact that a chemical switch can apparently convert an attractive response to a repulsive one might provide biomedical researchers with a powerful mechanism for controlling the trajectories of developing axons.

[0011] Previous methods for establishing gradients include techniques for establishing controlled gradients of tectal membranes on which retinal axons can be grown. This method is restricted to molecules that bind to cell membranes, and sets up gradients only of these membranes rather than of the molecule directly. As a consequence, it does not represent guidance by a diffusible factor. In addition, although this method can reliably produce linear gradients of variable slope, its ability to produce gradients of arbitrary shape is limited.

[0012] A more quantitative approach for diffusible guidance molecules is to glue together two glass slides so that they are closed in on three sides with a gap between them. One end is filled with a solution of nerve growth factor (NGF) in agar which then hardens, followed by a suspension of dissociated chick dorsal root ganglion (DRG) neurons in agar. NGF diffuses from the end block through the block containing the cells so as to create a gradient. However, detailed quantification of the concentration or gradient steepness at the growth cone is not obtained, and the shape of the gradient is generally not well controlled in this type of experiment. Furthermore, this technique is not easily adapted to the production of a standard assays, and is relatively cumbersome to prepare repeatedly. [0013] In another demonstration of the chemotropic effect of NGF on DRGs, chick DRG neurons are cultured on a two-dimensional surface of collagen and poly-L-lysine. After 24 hours, a micropipette filled with a solution of NGF is placed approximately 25 microns from a DRG growth cone. Gravity produces a flow of about 1-2 μl of NGF solution per hour, and a perfusion system directed this flow past the DRG growth cone. Growth cone turning towards the source of NGF is observed over a timescale of minutes. The concentration was found to have fallen to about 10% of the source concentration at a distance of 25 microns. In a similar approach, axons are grown on a two dimensional substrate in a bath of fluid medium, and a fine pipette containing the molecule of interest introduced a short distance from the growth cone. The molecules are slowly ejected, setting up a gradient by diffusion. The shape of this gradient has been measured, and matches what would be expected theoretically. Although this is a more quantitative approach than the conventional collagen gel assay, it has the limitation that only one gradient shape can be set up which is of very short range, and this can only be held constant for a few hours. Thus, only the short-term response of growth cones can be examined. Also the operator has very limited knowledge of and control over the actual concentration profile.

[0014] Despite its limitations, the pipette technique has been extremely successful at harnessing recent advances in molecular biology to uncover new information about axon guidance. These include the nature of axon turning in a glutamate gradient, the ability to switch the response to netrin-1 of some types of neurons from attractive to repulsive by manipulating levels of cAMP or the substrate material, the receptors involved in the chemotactic response to netrin-1, clarifying the chemotropic effects of neurotrophins, and demonstrating the effects of manipulating intracellular Ca on netrin-1 signaling. This success is due in large part to the fact that the technique is compatible with standard two-dimensional axon growth assays and many of the diffusible proteins of interest. Equally important, it is easy to implement repeatedly, so it is possible to build up meaningful statistical measures for a range of concentrations and for different biochemical perturbations.

[0015] In contrast to the field of axon guidance, the study of chemotaxis bacteria is now well grounded in both quantitative results regarding gradient detection and detailed theoretical models that draw on this data. Bacteria move very fast compared to growth cones (about 20 cell diameters per second for bacteria versus about 1 growth cone diameter per several minutes), and use a temporal rather than spatial sensing mechanism (i.e. they compare concentrations at two different points in time while moving, rather than two different points in space while stationary. One of the remarkable features of bacterial chemotaxis is the property of adaptation: bacteria maintain their sensitivity to chemical gradients over a wide range of attractant or repellant concentrations. This fact, established from carefully controlled quantitative experiments, significantly constrains the possible biochemical networks involved in the signal transduction, and was recently used to motivate a model of the chemotactic network that produced quantitative predictions about the effects of perturbations of that network. The predictions were recently verified in an experiment that manipulated the concentrations of the proteins involved in the network. While the signal transduction mechanisms of axonal growth cones are much more complex than that of bacteria, these results demonstrate the power of quantitative experimental constraints to advance our understanding of the mechanisms of signal transduction. Therefore, there remains a need for assays that will allow the study of axon - growth guidance over time scales that are physiologically relevant and allow quantitative analysis of the axon growth mechanisims.

SUMMARY OF THE INVENTION

[0016] In one aspect, the present invention provides an assay for the characterization of axon-growth guidance properties of a molecule comprising: forming a concentration gradient of the molecule into the collagen substrate; exposing the axon to the concentration gradient formed in the collagen substrate; and measuring a change in axon size, texture and/or shape; wherein the collagen substrate has a thickness to length ratio adjusted to stabilize the concentration gradient for a period of time sufficient to induce a change in the axon size, texture and/or shape..

[0017] Preferably, the concentration gradient is formed by depositing a plurality of drops of the molecule onto a surface of the collagen substrate, the drops being distributed along the length of the substrate and the concentration and/or volume of the drops is varied such that a the concentration gradient is formed along the length of the substrate. The drops deposited onto the surface of the substrate diffuse through the thickness of the substrate such that a uniform concentration of the molecule is obtained through the thickness of the substrate for a given position along the length of the substrate. A preferred thickness to length ratio is in the range of 1:2 to 1 :1000. More preferably, the substrate has a thickness to length ratio of 1 : 10 to 1 : 100.

[0018] Preferably ,the molecule is a chemotropic factor selected from the group consisting of nerve growth factor (NGF); brain-derived neurotrophic factor (BDNF); neurotrophin-3 (NT-3); neurotrophin-4 (NT-4); neurotrophin-5 (NT-5); acidic fibroblast factor (aFGF); basic fibroblast groeth factor (bFGF); epidermal growth factor (EGF); transforming growth factor alpha (TGF-alpha); schwanoma-derived growth factor (SDGF); insulin; insulinlike growth factor-I (IGF-I); insulin-like growth factor-II (IGF -II); interleukin-6 (IL-6); cholinergic differentiation factor/leukemia inhibitory factor (CDF/LIF); transforming growth factor-beta-1 (TGF-beta); interleukin-1 (IL-1); interleukin-2 (IL-2); interleukin-3 (IL-3); activin; heparin-binding neurotrophic factor (HBNF), protease nexin I and II, gamma-interferon, choline acetyltransferase development factor; and platelet derived growth factor (PDGF). A more preferred molecule is a protein from a family selected from the group consisting of netrin, semphorin and slit protein families.

[0019] The invention also provides a method for the modulation of neural growth regenerative capacity comprising identifying a molecule having axon growth guidance properties through the assay of the invention and administering to a subject said molecule in an amount sufficient to regenerate neural growth. Also provided is a method for the treatment of a neurodegenerative disease comprising identifying through the assay of the invention a molecule having axon growth guidance properties and administering to a subject said molecule in an amount sufficient to induce neural growth.

[0020] The invention also provides a method of diagnosis of a neurological defect comprising extracting a sample of body fluid from a subject, isolating a molecule from the sample and assaying the molecule through the assay of the invention to characterize the axon growth guidance of the molecule.

BRIEF DESCIPTION OF THE DRAWINGS

[0021] Another method of diagnosis of a neurological defect provided by the invention comprises extracting a sample of body fluid from a subject, and screening the sample for the presence of a molecule whose axon growth guidance properties are determined through the assay of the invention.

[0022] Figure 1 - Shows ingredients of the gradient assay

[0023] Figure 2 - Simulation results showing that after initial transience have died away

[0024] Figure 3 - Distribution of fluorescence intensity for an initial pattern of equally spaced lines of fluorescence casein

[0025] Figure 4 - Response of rat cortical explants to Netrin-1

[0026] Figure 5 - Method for establishing chemical gradient with micro-machining technology DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0027] The present invention is based on the appreciation that elucidating and controlling axon guidance by chemical gradients require a quantitative understanding of the

_ chemotactic response. In addition to the model discussed above, several detailed theoretical models of gradient sensing have been developed, mostly based on data from leukocyte and bacterial chemotaxis. These models start from the hypothesis that gradient detection is limited by inherent statistical fluctuations in receptor binding. The basic premis is that the presence of a ligand gradient will produce a variation in the average occupancy of receptors across the growth cone. However, at any instant in time, the actual occupancy of the receptors will differ from the average. If these fluctuations are large compared to the difference in the average 'that arises from the concentration gradient, the growth cone will not be able to obtain a clean guidance signal from an instantaneous measurement. The random noise can be overcome by averaging statistically independent measurements over some averaging time. Putting in reasonable parameters for growth cones, the inventors have determined that the minimum detectable gradient is about 1%, and the resulting maximum guidance distance by chemotaxis is about 1 mm. This result is in rough agreement with experimental results, but the data necessary to test and refine these theoretical models did not exist prior to the subject application, because the technology for generating stable gradients of chemotropic molecules in the appropriate^' environment did not exist. While the statistical noise inherent in the receptor binding may be the ultimate limit of sensitivity, limitations in the signal transduction networks probably reduce the range of concentrations over which small gradients can be detected. The present invention provides a tool for precise determination of the minimum detectable gradient as a function of concentration, thereby providing models for gradient detection allowing for more reliable determination of the maximum guidance distance.

[0028] The present invention is based on a three-dimensional collagen gel assay, which provides an environment that is much more faithful to that found where guidance by diffusible factors has been identified, and provides the longer time-scales appropriate for most axon guidance contexts. In addition, the assay of the present invention allows quantitative control over the chemical gradients imposed, allowing much more stringent tests of the chemotactic sensitivity of developing axons. [0029] The assay of the invention is designed to allow axonal response to gradients to be measured with unprecedented accuracy. In accordance with the present invention, gradients are established by "printing" chemotropic molecules onto the surface of a collagen gel and allowing them to diffuse into the collagen. Diffusion of chemotropic molecules in collagen is quite slow. Our initial order of magnitude estimate for the rate of diffusion of netrin-1 in a collagen gel is 10^"7cm ² second. In a block of collagen of length L = 1 cm the time scale for significant diffusion along this length, L2 D, is approximately 4 months. It is thus impractical to set up a gradient by simply placing opposite sides of the collagen block in contact with reservoirs of different concentrations. However, the present invention utilizes this long-term stability to produce a stable, yet easy to prepare assay. In accordance with the present invention, a block of collagen 1 cm long but only 1 mm thick is associated with a time scale for diffusion through this thickness of less than one day. In a ccordance with the present invention, a distribution of a chemotropic factor applied to the surface of the collagen block produces a uniform concentration in the vertical direction relatively rapidly, reflecting the profile on the surface. However, along the length of the block the profile will dissipate and flatten out only very slowly. To obtain optimum results with the assay of the invention the collagen should be ' "dry", i.e. is not in contact with any fluid such as culture medium; since the fluid would have a much higher diffusion constant for the chemotropic factor than collagen, the factor would move rapidly through the fluid from high to low concentration regions (producing a "'short circuit") and the gradient in the collagen would rapidly dissipate.

[0030] In a preferred embodiment of the present invention, a gradient is established by printing a pattern of a chemothropic factor of variable concentration onto the surface of a thin collagen block. For example, the pattern is produced in the form of thin, equally spaced parallel lines of chemothropic factor with a regular variation in the concentration from one end of the block to the other, as shown in figure 1A. An instantaneous image of a stream of droplets ejected from a pump onto the surface of the collagen block can be captured with strobe illumination. Preferred drops are of volume 1 nl, and the ruler spacing is 1mm.

[0031] The time-dependent concentration profile resulting from the initial pattern on chemical factor is calculated using finite element modeling of the diffusion equation with the appropriate boundary conditions (or analytic calculation for some initial conditions). Figure 2 shows the result of a two-dimensional simulation of a ² cm by 1 mm block, according to the invention as illustrated in Figure 1A, in the limit of an infinitely wide block (i.e. infinite in the x direction). The pattern consists of 21 equally spaced lines, and the quantity of factor in the initial lines varies linearly from one end of the block to the other in the y direction. Figure 2 A shows the profile at various times along the length of the block. The initial oscillations quickly ' die away, followed by a long period of a stable gradient. The time scale for stabilization and eventual decay of the gradient is inversely proportional to the diffusion coefficient D=10^"7 cm ²/sec for the simulation results shown in Figure 2.

- [0032] Although an axon embedded in the gel will initially experience a time- dependent concentration of the chemotropic factor, the concentration in between the top and the bottom of the gel very quickly reaches a nearly stable level Figure 2B. Theprecision of the technique of the invention allows for accurate control of the concentration over several orders of magnitude. The technique of the invention can also be used to provide a source of factor while axons are growing, since no direct contact is made with the gel.

[0033] Thus the way axons respond to gradients that are changing shape over time in a precisely controlled way can be quantitatively analyzed by the assay of the present invention. Finite element modeling is employed to calculate the time dependent concentration profile that results from an initial distribution of chemotropic factor, and to optimize the initial distribution in accordance with the desired properties of the assay.

[0034] Simulation result showing that, after the initial transients have died away, the gradient (y axis) along the length of the block (x axis) remains highly stable. Simulation result showing that the gradient (y axis) rapidly becomes uniform through the vertical thickness of the collagen block (x axis).

[0035] In accordance with a preferred embodiment of the present invention, precisely controlled volumes of chemotropic factors are deposited onto precisely controlled locations using a method similar to that employed in inkjet printers. A commercially available micropump that can be advantageously used in conjunction with this invention consists of a small reservoir between a glass plate and a membrane in contact with a piezo-electric actuator. When a short (100) voltage pulse is applied to the pump, the membrane rapidly contracts the reservoir, forcing a small spherical droplet out of the micro-machined nozzle of the micropump (see Figure IB). The droplet volume depends in a controllable way on the parameters of the voltage pulse, ranging from 0.5 to 1 nl. The amount of factor delivered can be precisely controlled by varying the pulse rate, up to a maximum of 1000 drops per second. The micropump sits above the dish containing the gel, which can be mounted on a commercial high precision x-y translation stage. By varying the rate of factor delivery from the pump, the length of time the factor is delivered to a particular spot, and the pattern of spots chosen, the present invention provides virtually unlimited control over the spatial and temporal characteristics of the factor concentration profile, subject to the limits of molecular diffusion. Actual concentration profiles can be directly measured by substituting a fluorescent molecule of similar molecular weights to the factor (e.g., netrin-1), and also directly using fiuorescently labeled chemotropic factor. Quantitative fluorescent imaging requires some care, but with the low concentrations of molecules appropriate for axon guidance, it is usually possible to work in a regime where the fluorescent intensity can be reliably converted to the concentration of the dye.

[0036] Figure 3 shows the spatial distribution of fluorescent intensity produced by a pattern of fluorescent casein, a protein of a molecular weight similar to axon guidance molecules (26D). Figure 3 A was taken approximately 30 minutes after 10 1cm long lines of casein were deposited onto the surface of a collagen block. The lines were 1 mm apart, with 100 drops in the first line, and an increase of 100 drops for each subsequent line. Distribution of fluorescent intensity for an initial pattern of 11 equally spaced lines of fluorescent casein, allowed the measurement of gradient of fluorescent dye in a section of a collagen block 30 minutes after generation of the initial pattern, and the same gradient after 4 days. Note that gradient shape remains highly stable, x, y axis are uncalibrated, but the total extent of the pattern in the x and y directions is 1 cm. The z axis units are arbitrary units of fluorescence intensity. The remnant of the initial pattern is the series of steps evident in Figure 3 A. The smearing of the initial pattern is due to diffusion, and the fact that, although the drops are very small (less than 100 μm diameter), they spread out some when they hit the collagen. The ridges along the side of the gradient are the result of a synchronization error between the micropump and the translation stage, which actually helps to stabilize the concentration in the central region. Figure 3B shows the gradient four days later. The steepness has reduced somewhat, but the overall shape is remarkably stable. While it appears in the image that the roughness of the initial pattern is still evident, this is likely an artifact of chemical changes in the collagen surface caused by the liquid in the droplets. Alternative imaging approaches can minimize this effect. For example, since the collagen is in a transparent dish, imaging can be performed from below the dish, instead of from above.

[0037] As discussed above, an important aspect of the chemotropic factor gradient assay of the present invention is that, preferably, no fluid medium (or as little as possible) is present in the collagen, blovk since this would form a "short circuit" allowing the gradient to flatten out too quickly. This is feasible based on the recognition that it is possible to grow cells and explants in a "dry" collagen gel such that they remain healthy and continue to extend neurites for periods of up to a week. It is important to note that the present inventors have established that this modification to the standard collagen assay does not impede axon growth as shown by the results on cortical explants provided in Figure 4 and DRG neurons clearly demonstrate that the absence of the fluid medium does not significantly alter the behavior of the neurons. It is anticipated that there will be no difficulty in extending this technique to other systems, including dissociated cells.

Examples

[0038] Explants approximately 300 microns in diameter are taken from an E15 rat embryo, and cultured in Type I collagen for 10 hrs. They are co-cultured with filter paper (outlined with dotted line) soaked in either (A) PBS (as a control), or (B) 1 nM recombinant netrin-1 in PBS. Response as a function of concentration of netrin-1 in the solution. Approximately 20 explants for each concentration are scored for asymmetry by 2 observers in a blind test. The point on the abscissa is the control. Neurons are labeled with TuJl, a neuron specific anti-tubulin antibody, and visualized with a goat anti-mouse antibody conjugated to Oregon Green (Molecular Probes).

[0039] We have tested the dry collagen assay and the netrin-1 by culturing rat cortical explants nearby small squares of filter paper that were soaked in a solution containing purified netrin-1 prior to being placed in the collagen. The filter paper acts as a source of netrin-1, producing a gradient of decreasing concentration with increasing distance from the paper. Figure 4A shows a control case, where the filter paper is soaked in the solution without netrin-1 , and shows normal, essentially symmetric axonal outgrowth. When the filter paper is soaked in the same solution but with netrin-1 added, a strongly asymmetric response is observed. We have used this assay to determine the concentration range over which the axons are sensitive to netrin- 1. Figure 4 shows a plot of the percentage of explants showing an asymmetric response as a function of the initial netrin-1 concentration. The response is strong over a surprisingly limited range of concentrations, suggesting that axons do not display adaptation to netrin-1, and can only be guided over a narrow range of concentrations. This result has profound implications for guidance, since it severely constrains the chemical profiles necessary for guidance and the maximum guidance distance. It also suggests that the biochemical networks responsible for signal transduction do not contain the feedback necessary for adaptation. While these results are strongly suggestive, this technique has severe limitations that are overcome with the micropump technique of the invention. It is difficult to carefully control the spacing between the filter paper and the explants, so the concentration at the explants varies from one dish to the next. Also, the concentration profile that can be generated is not under control.

[0040] While the preliminary results shown above demonstrate the feasibility of the method of gradient generation according to the invention. It should be understood that improvements may be obtained without departing from the scope of the invention. These include more careful measurements of the gradients produced and validation of the diffusion modeling for different chemotropic factors, including techniques for gradient stabilization when the diffusion coefficient is not sufficiently large, and accurate measurements of the diffusion coefficients of the relevant factors in various culture media.

[0041] Further refinements that are contemplated by the inventors include testing the predictions of gradient sensitivity described above by growing explants and dissociated cells of cortical axons in gradients of netrin-1. In particular, the minimum gradient necessary for a turning response can be determined as a function of the overall concentration of netrin-1. The present invention also affords tests of the prediction that developing axons can turn faster (more sharply) in steeper gradients. In addition, conditions necessary for a repulsive response to netrin-1 can show if the limits of repulsion are the same as attraction.

[0042] The ability afforded by the present invention to predict the time-dependent concentration profile produced by an initial pattern of a neurotrophic factor in a collagen gel is dependent upon accurate knowledge of the diffusion coefficient. Existing measurements of the relevant coefficients, however, are very limited. No measurements exist for many important axon guidance molecules, and very little attention has been paid to the precise composition of the host media. One collateral benefit of the micropump and calibrated fluorescent imaging capability of the invention provides a simple, efficient method for the determination of diffusion coefficients: measuring the spreading of a spot of fluorescently labeled protein "printed' at the center of a dish containing collagen gel. While at short times the time dependent profile depends on the details of the initial dye distribution, the longer time decay is indistinguishable from that of an ideal point source, because the concentration profile relaxes to a Gaussian. The diffusion coefficient can then be determined from the slope of a plot of the width of the Gaussian as a function of time. (The distribution will deviate from a Gaussian once a significant concentration diffuses to the boundaries of the 3.5 cm diameter circular dish, but that takes weeks for the molecules of interest.)

[0043] Our preliminary measurements suggest that the diffusion coefficient is quite sensitive to the composition of the culture medium, in particular to the concentration of collagen in the gel. An accurate measurement of this dependence for molecules implicated in axon guidance is important for understanding how gradients are set up, and for interpreting axon guidance assays, where different labs sometimes use significantly different densities of homemade collagen. In addition, it is an important application of the interesting and poorly understood problem of the diffusion of complex polymers in tortuous media.

[0044] One embodiment of the present invention incorporates finite element modeling to calculate the time dependent concentration that will result from an initial chemical concentration, which allows for the comparison of those calculations with the measured profiles of fluorescent molecules according to the technique of the subject invention. Actual concentration profiles can be directly measured by substituting a fluorescent molecule of similar molecular weights to netrin-1, and also directly using fluorescently labeled netrin-1 (Alexa 546 protein labeling kit, Molecular Probes). Similarly, our initial work described herein has been with casein, an inexpensive molecule of similar molecular weight to NGF, the assay of the invention can be practiced with fluorescent NGF, which can be obtained commercially (Molecular Probes Inc., Eugene, OR).

[0045] In order to increase the stability of the gradients for situations where the diffusion coefficient is too large, or gradients that decay more quickly than linear gradients, the assay of the present invention allows experimentation with different initial distributions of chemical factors. By providing excess molecules at the high concentration end of the gradient, for example, the decay in the middle region of the gradient can be slowed somewhat. This possibility is first explored through computer simulations, and then implemented to measure the idealized distributions with fluorescent molecules and measure the resulting gradients.

[0046] Accurate validation of this technology will require more precise quantitative fluorescent imaging. While the current testing of the assay of the invention is based on imaging systems suitable for concentration variation of a factor of 100, the assay can be easily implemented to extended the variations by an order of magnitude with a more sensitive CCD. It is also contemplated that improvements can be obtained by measuring the absorption length for both the excitation and emission wavelengths in the collagen gel, in order to more accurately interpret the fluorescent intensity. Another improvement is based on imaging from the bottom of the gel, both to get increased information about the vertical concentration distribution, and to eliminate the artifacts generated by the interaction of the gel with the drops from the micropump.

[0047] Netrin-1 purification Serum free media are conditioned by a human embryonic kidney 293 cell line, and transfected with chick netrin-1 (a gift from Marc Tessier-Lavigne). Netrin-1 is then eluted from a heparin column using a gradient of NaCl. The fractions analyzed for netrin-1 content using silver stained SDS-PAGE gels and the selected fractions are combined and concentrated. The final concentration of netrin-1 obtained using this protocol is in the range 0.5 - 1.0 mg/ml (Marc Tessier-Lavigne, personal communication). We calculate that only roughly 30 ng of netrin are required per dish to establish appropriate gradients.

[0048] Tissue preparation: Rat embryos (Sprague-Dawley, Taconic) is removed and their brains dissected under sterile conditions on embryonic days 15-17. The meninges are removed and the optic stalk, brainstem and sub-cortical regions dissected away from the cortex. If explants are to be used in the culture, they are processed at this point. To prepare single cell suspensions, explants are washed and triturated gently to dissociate the cells. The suspension is allowed to settle and the supernatant is removed, (the dissociated cells are predominantly in the supernatant). The explants are triturated and allowed to settle twice more, each time removing and saving the supernatant. The supernatant is then centrifuged, resuspended and filtered to generate single cells.

[0049] One ml of a 0.3% collagen gel per 35 mm dish is prepared under sterile conditions by mixing on ice; type I rat tail collagen diluted with water, sodium bicarbonate solution, 10X OptiMEM (Gibco), 100X antibiotic/antimytotic (Biofluids). The solution is quickly placed in a 35 mm dish, spread uniformly over the bottom of the dish and allowed to set, forming a bottom layer of collagen. Several explants, or a drop of the cell single cell suspension (see below), are then placed in the middle of the dish. A fresh layer of collagen is added immediately in the case of the explants, or after 1 hour in the case of single cells to allow them to settle. The dishes are returned to the incubator for 15 minutes for the top layer of collagen to set, and are then ready for the addition of the gradient. It is important that the explants and cells are sandwiched between these two layers of collagen as then each growth cone is completely surrounded by a uniform gradient: if the axons were growing on the bottom of the dish only the top of the growth cone would be exposed to the gradient. No fluid culture medium is added on top of the collagen since this would provide a short-circuit for the gradient. Providing a tray of water is kept in the incubator we have not experienced problems with the collagen drying out. Dishes are returned for 3 days to an incubator dedicated to this experiment. At the end of this time, the tissue is fixed by covering the collagen with 2-3 mis of a 4% solution of paraformaldehyde. Axons extending from the explant are then imaged using an inverted microscope equipped with phase contrast optics (Nikon TE300 with fluorescence, available in GICCS).

[0050] To compare theoretical predictions with experimental results, it is necessary to compare growth cone responses at different positions in the gradient (for details of setting up the gradient see below). The single cell suspension is diluted so that a drop contains 40,000 cells: the drop is then plated so that it covers approximately the middle 1 cm² of the bottom layer of collagen (the region over which the gradient will be established), so that the average spacing between cells is about 50 μm. All cells are in (roughly) the same plane, making them easier to visualize than if they were distributed throughout the collagen. There are now many cells at each concentration, i.e. along each narrow strip perpendicular to the gradient, each strip corresponding to a different concentration.

[0051] This case is more complicated than single cells, since the area of the dish covered by processes from each explant only covers a small proportion of the total extent of the gradient. We therefore plate 9 explants per dish in a 3 by 3 grid pattern, allowing sufficient distance between explants such that their processes will not reach each other in the 3-day time scale of the experiments. This arrangement, with slight perturbations across several dishes, provides multiple growth cones at each position in the gradient, and at least one growth cone at multiple positions in the gradient.

[0052] To study the properties of the repulsive response of cortical axons to netrin-1 we use conditions identical to those used in studying attraction, except for the addition to the collagen gel at the time of preparation of cis-N-(2-Phenylcyclopentyl)-azacyclotridec-l-en-2- aminemonohydrochloride (Calbiochem). This is an adenylate cyclase inhibitor, and will reduce the concentration of cAMP in the growth cone. The netrin-1 gradient experiments described above will then be repeated: we now expect to see repulsion rather than attraction. By varying the concentration of adenylate cyclase inhibitor between different dishes with other conditions held constant, we will be able to measure a detailed dose-response curve, and thus determine the specific concentration threshold at which attraction is converted to repulsion. Growth cones are assayed for cAMP levels using an anti-cAMP antibody. cAMP immunofluorescence intensity is be measured from captured images and normalized by the area of the growth cone (calculated using the growth cone boundary from a superimposed phase contrast image). This yields a measurement of the dosejresponse curve of attraction/repulsion for relative levels of c MP. Knowledge of the quantitative details of the conversion from attraction to repulsion significantly facilitates the understanding of the biochemical networks responsible for transducing the chemotactic signal.

[0053] Neurites are identified based on morphology and by staining with TUJ1 (Chemicon) (Figure 4). Axons are distinguished from dendrites by staining with a MAP2 antibody, which specifically labels dendrites. The images are digitized, the neurites traced, and their final orientation measured. Two principle methods are used to quantify the degree of turning response. The simplest measure, applicable to both explants and single cells, is the average angle of the neurites relative to an axis perpendicular to the gradient. Because the concentration varies substantially along the growth region, averages are calculated across narrow strips where the concentration is roughly constant. Additionally, for the explant case we locate the center of the explant from the images, and measure the orientation of the neurites relative to the line connecting them to the center of the explant as a function of their position in the gradient. In the absence of any systematic turning response, the angles are symmetrically distributed about zero. Both of these methods are based on the final segment of each neurite: due to the very high density of neurites initially emerging from an explant it is often difficult to trace the complete trajectory of a neurite. Note that this focus on turning responses rather than comparing amounts of growth for different regions of the explant avoids the ambiguity that can arise in conventional collagen gel assays due to trophic effects. Time-lapse imaging of axon extension in the gradient is performed for instance to examine the detailed behavior of filopodia in different types of gradient.

[0054] The assays of the present invention can be advantageously utilized in investigating why some classes of neurons fail to regenerate after injury, while others regenerate quite well without experimental manipulation. The present invention allows for testing the role of particular molecular factors play in guiding regeneration. The present invention is advantageous in that it allows for the control of the profile bf guidance molecules to test hypotheses about the role of diffusible chemical factors for example, in the normal development of the corpus callosum.

[0055] The present invention also encompasses other approaches to gradient generation for situations where the micropump approach is not adequate (e.g. for rapidly diffusing molecules). A chemical gradient can be setup and maintained indefinitely by putting the collagen in contact with two or more reservoirs of differing concentration. One way to do this is by embedding semi-porous tubing into the gel, but this technique may be cumbersome when hundreds of repetitions are necessary. Another more powerful approach is to etch channels into a silicon or glass substrate. Anisotropic etching using TMAH on silicon 100 wafers can be used to fabricate precise channels and wells. An example of the sort of structure that is adaptable for use in conjuction with the present invention, fabricated for a different purpose, is shown in Figure 5 A. Figure 5B shows a schematic of the design suitable for use in conjunction with the present invention. A reservoir for the solution containing the guidance factors is formed, for example, by anodic bonding of Pyrex glass to the bottom of the silicon wafer. The collagen gel for neuronal culturing is set in a well on top of the wafer, separated from the reservoir by a thin window. In one possible implementation, a spatially varying pattern of holes etched into the window results in a spatially varying exchange from the reservoir to the collagen, leading to the formation of a concentration gradient. The collagen gel, a polymer matrix, will not be affected by the small holes. (If necessary, the holes can be blocked while the collagen sets.) Channels etched into the Pyrex provide a mechanism for filling the reservoirs and emptying them once the desired concentration profile is reached. It is also possible to vary the concentration in the reservoir or even change to a different chemical factor while the axons are growing. Multiple reservoirs with different concentrations of guidance molecules or different molecules could also be used, mimicking the effect of the semi-porous tubing described above. This technique provides a flexible and powerful method of gradient generation. We begin by fabricating a simple porous membrane, and measure the rate of diffusion of fluorescently labeled proteins into the collagen through the membrane. With this information, it is possible to determine the pattern of holes necessary to generate a desired gradient. As with the micropump patterning, we use finite element modeling to determine the expected time-dependent concentration profile, and use fluorescent imaging to measure the actual concentration profiles.

[0056] Commercially available micropums provide the mechanism for factor delivery, but integrating the pump into a reliable, flexible gradient generating device requires a combination of numerical analysis of the factor diffusion in the gel and direct measurement of the chemical concentration profiles produced. The present invention incorporates finite element modeling to calculate the time dependent concentration profile that will result from an initial distribution of growth factor, and to optimize our initial distribution for the needs of each application. Actual concentration profiles are directly measured by substituting a fluorescent molecule of similar molecular weights to the chemotropic factors, and also directly using fluorescently labeled factors. The diffusion coefficients of the factors in collagen are directly measured. The present invention allows the design and fabrication of structures using micromachining technology to allow for the controlled release of axon guidance molecules into collagen gels. The gradients generated by this process are quantitatively measured.

[0057] While the invention has been described in terms of preferred embodiments, the skilled artisan will appreciate that various modifications, substitutions, omissions and changes may be made without departing from the spirit thereof. Accordingly, it is intended that the scope of the present invention be limited solely by the scope of the following claims, including equivalents thereof.

Claims

WHAT IS CLAIMED IS:

1. An assay for the characterization of axon-growth guidance properties of a molecule comprising:

(a) preparing a three-dimensional collagen substrate;

(b) forming a concentration gradient of the molecule into the collagen substrate;

(c) exposing the axon to the concentration gradient formed in the collagen substrate; and

(d) measuring a change in axon size, texture and/or shape; wherein the collagen substrate has a thickness to length ratio adjusted to stabilize the concentration gradient for a period of time sufficient to induce a change in the axon size, texture and/or shape.

2. The assay of Claim 1 , wherein the concentration gradient is formed by depositing a plurality of drops of the molecule onto a surface of the collagen substrate,_' the drops being distributed along the length of the substrate and the concentration and/or volume of the drops is varied such that a the concentration gradient is formed along the length of the substrate.

3. The assay of Claim 2, wherein the drops deposited onto the surface of the substrate diffuse through the thickness of the substrate such that a uniform concentration of the molecule is obtained through the thickness of the substrate for a given position along the length of the substrate.

4. The assay of Claim 2, wherein the substrate has a thickness to length ratio of 1 :2 to 1:1000.

5. The assay of Claim 2, wherein the substrate has a thickness to length ratio of 1 : 10 to 1:100.

6. The assay of Claim 1, wherein the molecule is a chemotropic factor selected from the group consisting of nerve growth factor (NGF); brain-derived neurotrophic factor (BDNF); neurotrophin-3 (NT_r3); neurotrophin-4 (NT-4); neurotrophin-5 (NT-5); acidic fibroblast factor (aFGF); basic fibroblast groeth factor (bFGF); epidermal growth factor (EGF); transforming growth factor alpha (TGF-alpha); schwanoma-derived growth factor (SDGF); insulin; insulinlike growth factor-I (IGF-I); insulin-like growth factor-II (IGF-lϊ); interleukin-6 (IL-6); cholinergic differentiation factor/leukemia inhibitory factor (CDF/LIF); transforming growth factor-beta-1 (TGF-beta); interleukin-1 (IL-1); interleukin-2 (IL-2); interleukin-3 (IL-3); activin; heparin-binding neurotrophic factor (HBNF), protease nexin I and II, gamma-interferon, choline acetyltransferase development factor;, and platelet derived growth factor (PDGF).

7. The assay of Claim 1 , wherein the molecule is a protein from a family selected from the group consisting of netrin, semphorin and slit protein families.

8. A method for the modulation of neural growth regenerative capacity comprising identifying a molecule having axon growth guidance properties through the assay of Claim 1 and administering to a subject said molecule in an amount sufficient to regenerate neural growth.

9. A method for the treatment of a neurodegenerative disease comprising identifying through the assay of Claim 1 a molecule having axon growth guidance properties and administering to a subject said molecule in an amount sufficient to induce neural growth.

10. A method of diagnosis of a neurological defect comprising extracting a sample of body fluid from a subject, isolating a molecule from the sample and assaying the molecule through the assay of Claim 1 to characterize the axon growth guidance of the molecule.

11. A method of diagnosis of a neurological defect comprising extracting a sample of body fluid from a subject, and screening the sample for the presence of a molecule whose axon growth guidance properties are determined by through the assay of Claim 1.

12. The assay of Claim 1, wherein the characterization of the axon growth guidance properties includes the determination whether the molecule has an attractive vs repulsive growth guidance property.

13. The assay of Claim 1, wherein the concentration gradient is generated by depositing varying quantities of the molecule along the length of the substrate through an ink jet type machine containing a micropump.

14. The assay of Claim 1, wherein the collagen substrate is substantially fluid-free.

15. The assay of Claim 1 , wherein the substrate thickness to length ratio is adjusted according to the diffusion coefficient of the molecule through the substrate.

16. A method of forming a substrate suitable for the characterization of axon growth guidance properties of a molecule, wherein the method comprises preparing a three-dimensional collagen substrate having a thickness to length ratio adjusted according to the diffuison coefficient of the molecule through the substrate such that the molecule diffuses through the thickness of the substrate at a rate that is substantially higher than the rate of diffusion of the molecule along the length of the substrate.

17. The method of Claim 16, further comprising depositing drops of the molecule having different concentration and/or volume such that a concentration gradient of the molecule is formed along the length of the substrate, wherein the substrate dimensions are adjusted such that the concentration gradient remains substantially unchanged for a period of time sufficient to determine the axon growth guidance properties of the molecule by exposure of an axon to the concentration gradient.

18. The method of Claim 17, wherein the concentration gradient remains substantially unchanged for at least four days.

19. The method of Claim 17, wherein the concentration gradient remains substantially unchanged for at least seven days.