INDIVIDUAL PROTEIN VARIANTS

RELATED APPLICATION

This application claims priority to U.S. Provisional Patent Application No.

61/617,476, filed on March 29, 2012, the entirety of which is incorporated herein by reference.

FEDERAL GRANT SUPPORT

The invention was made with Government support under Grants No.

2RO1EB004761-06 and R21EB010191-01 Al, awarded by the National Institutes of Health. The government has certain rights in the invention.

SEQUENCE LISTING

The instant application contains a Sequence Listing which has been submitted in ASCII format via EFS-Web and is hereby incorporated by reference in its entirety. Said ASCII copy, created on February 21, 2013, is named 17555.007WOl_SL.txt and is 34,904 bytes in size.

BACKGROUND OF THE INVENTION

Protein variants including alternatively processed and misfolded proteins have been associated with many different human diseases. For example, misfolded protein aggregates play a critical role in many devastating human diseases including Alzheimer's (AD) and Parkinson's diseases, diabetes and cancer. However, assessing the variant's role in the onset and progression of different diseases is hampered by a lack of reagents that can distinguish between protein isoforms in vivo. Development of such reagents has been hindered by two main factors: protein variants often occur at only trace levels in vivo and the variants may differ only subtly from the parent form, making them difficult to separate and purify. Novel separation technologies that can tease apart subtle protein variants along with novel molecular recognition protocols are needed to create reagents with sufficient specificity to distinguish between these protein variants. Therefore, there is a continuing need for technologies capable of generating highly selective reagents to specific protein isoforms isolated from mammalian tissue. SUMMARY OF THE INVENTION

The present invention provides a DC-iGDEP device for separating target protein species in a biological sample based on a variety of chemical and physical parameters comprising an open sawtooth microfluidic channel having an inlet port and an outlet port, and gates between each tooth of the channel, wherein spacing of the gates starts at 50 microns and decreases over two centimeters to 1 micron, and wherein the teeth insulate adjoining gates. As used herein the gap distance of a gate is the distance between teeth across from each other on opposite sides of the device.

The present invention provides a method of separating from a biological sample a target species based on various chemical and physical parameters including charge, size, permittivity, deformation, shape and other factors comprising

(a) providing the device for separating target protein species in a biological sample based on a variety of chemical and physical parameters comprising an open sawtooth microfluidic channel having an inlet port and an outlet port, and gates between each tooth of the channel, wherein spacing of the gates starts at 50 microns and decreases over two centimeters to 1 micron, and wherein the teeth insulate adjoining gates,

(b) loading a loading volume of the sample into the inlet port,

(c) applying a field to the device to separate particles or molecules in the sample, and

(d) recovering the target species.

In certain embodiments, the target species is an Αβ aggregate. In certain

embodiments, the recovered Αβ aggregates with an antibody to confirm the size of the Αβ aggregate. In certain embodiments, the antibody is specific for oligomeric Αβ aggregates. As used herein, the term "oligomer" refers to a dimer, trimer, or tetramer or larger aggregate. In certain embodiments, the antibody is a nanobody. As used herein, the term "antibody" includes scFv (also called a "nanobody"), humanized, fully human or chimeric antibodies, single-chain antibodies, diabodies, and antigen-binding fragments of antibodies (e.g., Fab fragments). In certain embodiments, the nanobody is a C6, A4, El , D5, 10H, 6E, D10 or BSEC1 nanobody.

In certain embodiments, the biological sample has a volume of less than 100 microliters. In certain embodiments, has a volume of about 50 microliters. In certain embodiments, the biological sample is brain tissue, serum, cerebrospinal fluid (CSF), urine or saliva. In certain embodiments, the force is applied for a period of time that is less than 20 minutes, such as between 5-15 minutes. In certain embodiments, target Αβ aggregate is concentrated by several orders of magnitude. In certain embodiments, the target Αβ aggregate is concentrated by 10 ⁶ as compared to the loading volume. In certain

embodiments, the protein is p53, islet amyloid polypeptide, beta-amyloid, tau), alpha- synuclein, huntingtin, or superoxide dismutase.

BRIEF DESCRIPTION OF THE FIGURES

Fig. 1. Schematic showing current DC-iGDEP design (top) used to capture Αβ fibrils

(middle, right) but not monomers (middle, left). The proposed modified device (bottom) will 1) increase maximum DEP force -700 times to capture monomelic AJ3, 2) refine the difference in local DEP force between adjoining "gates," and 3) enable separation and concentration of various Αβ aggregates (dimers, trimers,...up to fibrils) in separate chambers.

Fig. 2 provides nucleic acid and amino acid sequences for several nanobodies.

Underlining indicates CDR regions (or nucleic acids that encode CDR regions).

Fig. 3 depicts an embodiment of the present device. Diagram of the direct current insulator gradient dielectrophoresis (DC-iGDEP) device. An insulated sawtooth pattern is integrated with a tapered microfluidic channel to generate localized non-uniform electric field gradients of increasing strength from left to right.

Figs. 4A-4D. Fluorescence images of the narrowest portion of the DC-iGDEP for Αβ monomer samples (30 μΜ) with (A) 400 V, (B) 600 V, (C) 800 V, and (D) 1000 V applied. The diffuse light areas show the fluorescently tagged Αβ monomer distributed throughout the channel. Streaming of monomer is more visually apparent in (C) and (D).

Fig. 5A-5D. Fluorescence images of the narrowest portion of the DCiGDEP (same as in Fig. 4) showing capture of Αβ fibrils (30 μΜ) at (A) 400 V, (B) 600 V, (C) 800 V, and (D) 1000 V applied. The areas of high fluorescence localized between the tips of the sawtooth patterned insulator indicate small zones where the fibrils were captured and concentrated.

Fig. 6. TEM of the mature Αβ fibrils used in Fig. 5. The scale bar is 100 nm, and the fibril sample was in PBS before buffer exchange to Tris.

Fig. 7. Regions of interest (ROIs) indicated as described in the text. Three different areas were selected to determine the enrichment of fibril concentration in the capture zone using the fluorescence intensities in each region. The areas depicted here are larger than the actual ROIs used for ease of viewing.

Fig. 8. Plot showing computed magnitudes of V |E| ² along the microchannel centerline versus gate width. Vertical axis is plotted using a logarithmic scale. The trend line shows that the relationship between data points can be approximated using a power function.

DETAILED DESCRIPTION OF THE INVENTION

The inventors have developed methods to separate and concentrate protein variants at a microliter scale and to generate reagents to those variants with exquisite selectivity for specific protein isoforms using only picograms of target material. This capability is broadly applicable to protein variants associated with many human diseases. In certain embodiments, this method is used to generate reagents against isoforms of the protein amyloid-beta (Αβ) that have been implicated in Alzheimer's disease (AD). Misfolding and assembly of Αβ into an array of different aggregate species has been linked to the onset and progression of AD. A variety of different length and aggregate forms of Αβ have been identified in human brain tissue. Unique Αβ forms present in AD brain tissue are separated and isolated using a novel electric field-based separation process. Antibody-based reagents are utilized that can selectively bind the different Αβ species and identify which reagents can best distinguish healthy and AD brain tissue.

The present work benefits from the current availability of well characterized postmortem human AD and healthy brain tissue. The present work utilizes novel electric field- based capabilities, which can quickly isolate and concentrate different protein isoforms using only minimal amounts of material. This enables the collection in separate nanoliter volumes several distinct Αβ isoforms isolated from human AD tissue. Novel protocols have been developed that allow the isolation of single chain variable domain antibody fragments (scFvs, also called "nanobodies") against specific protein morphologies by utilizing Atomic Force Microscopy (AFM) in conjunction with surface display antibody libraries. These protocols enable the isolation of nanobodies to specific aggregate morphologies using only picograms of material without the need for any protein modification. Panning techniques are used to generate nanobodies against the isolated Αβ isoforms from AD tissue. Protocols have been developed to characterize nanobody binding specificity to different target morphologies without the need to purify or modify the target antigen and again requiring only microliters of protein target solution. Techniques have been developed both to isolate subtle protein variants involved in human disease, in this case Αβ isoforms isolated from AD brain tissue, and also to generate and characterize nanobody reagents to the key isoforms that distinguish AD from healthy brain tissue.

Understanding the molecular basis of disease progression is often a very challenging and complex problem because the specific biochemical species involved may be extremely difficult to identify. Individual biomolecules may be unstable and differ from each other in only very subtle ways, therefore identification and isolation of different forms is a

technologically demanding problem. One critically important and challenging class of closely related protein variants are different folded conformations of the same protein. Misfolded and aggregated protein variants are often indistinguishable by many analytical strategies. Over 30 human health diseases have already been connected to misfolding or misprocessing of proteins including cancer (p53), diabetes (islet amyloid polypeptide), Alzheimer's (beta- amyloid and tau), Parkinson's (alpha-synuclein), and Huntington's diseases (huntingtin), Amyotrophic lateral sclerosis (superoxide dismutase) and prion based diseases. In many of these diseases, specific misfolded protein variants such as small soluble oligomeric forms of the amyloid-beta (Αβ) protein or a misfolded form of the prion protein have been associated with cell dysfunction and disease progression. While the role of protein expression in disease can be very effectively studied using proteomic and genomic analyses or inhibitory RNA techniques, these methods are generally not capable of distinguishing between misfolded or alternatively processed protein variants. The different protein variants are often metastable and contain only subtle differences. Robust tools that can selectively identify and manipulate aberrant protein forms would be extremely useful in studying and controlling the many diseases associated with protein variants.

Currently, studies to probe the mechanisms underlying these diseases and to develop appropriate therapeutic strategies have been greatly hampered because reagents that can selectively recognize specific protein variants are scarce. In order to address this critically important need, unique electric field-based methods have been developed to efficiently isolate and concentrate subtly different protein variants, and simultaneously developing novel biopanning methods to generate very selective molecular recognition reagents that can be used to identify specific disease related species and mechanisms.

To demonstrate the current technology, reagents were isolated that selectively bind a diverse array of different aggregate species of the protein amyloid-beta (Αβ). While Αβ was first implicated in Alzheimer's Disease (AD) over 20 years ago the role of Αβ in AD is still unclear and has proven to be much more elusive than originally hoped. Much of the confusion around the role of Αβ in AD is due to the variety of different Αβ species that can occur in vivo. Aggregation of Αβ is a critically important though poorly understood factor in the progression of AD. While the amyloid plaques of AD contain fibrillar Αβ aggregates, a variety of other smaller aggregate species of Αβ can also be formed and increasing evidence implicates various small soluble oligomeric Αβ species in neurotoxicity and loss of synaptic function. Cortical levels of soluble Αβ correlate well with cognitive impairment and loss of synaptic function. Small soluble aggregates of Αβ, termed Αβ-derived diffusible ligands (ADDLs), and spherical or annular aggregates termed protofibrils were shown to be neurotoxic. Oligomeric forms of Αβ, created in vitro or derived from cell cultures, were shown to inhibit long term potentiation. The concentration of oligomeric forms of Αβ is also elevated in transgenic mouse models of AD and in AD brain. Disruption of neural connections was shown to occur near Αβ plaques and fibrils, suggesting a toxic role for the fibrillar form, however the disruption also occurred in regions without fibrillar Αβ deposition suggesting that the toxicity may be due to small amounts of oligomeric Αβ, some in equilibrium with the fibrillar form, some existing on their own. A halo of oligomeric Αβ was shown to surround Αβ plaques and correlate with synapse loss. Oligomeric Αβ was also shown to disrupt cognitive function in transgenic animal models of AD.

A major barrier impeding studies of the connection between Αβ and AD is the lack of suitable reagents to identify and localize the different Αβ species. Antibodies to selected Αβ forms have been previously generated by immunization methods using for example soluble oligomeric Αβ (Lambert, M.P., et al., Vaccination with soluble Abeta oligomers generates toxicity-neutralizing antibodies. J Neurochem, 2001. 79: p. 595-605) or molecular mimics of oligomeric Αβ (Kayed, R., et al., Common structure of soluble amyloid oligomers implies common mechanism of pathogenesis. Science, 2003. 300: p. 486-9). These reagents have been useful to demonstrate the importance of oligomeric Αβ in AD for example by confirming the presence of oligomeric Αβ in AD brains, showing they target synaptic ligands and bind strongly to neuronal dendrites, and that oligomeric Αβ correlates with synapse loss near amyloid plaques. However, while the current pool of reagents has been useful in pointing out the importance of soluble Αβ aggregates, their binding specificities are generally not well characterized, the affinities and specificities of the reagents are not sufficient to identify the presence of specific oligomeric species or of low concentration species and are not generally suitable for precise histochemical studies or to localize species to specific cellular locations in situ or in vivo. In addition, they are not suitable for intracellular studies.

Since soluble aggregates of Αβ, α-synuclein (α-syn) and tau have all been associated with neurodegenerative diseases, reagents that can recognize specific morphologies of specific proteins are needed to elucidate the roles of the different forms of these proteins in disease. Therefore there is a critical need for well characterized reagents that recognize specific Αβ forms to facilitate studies on the role of Αβ aggregation in AD, to identify the relevant toxic Αβ species, and to clarify where and when aggregation of Αβ begins and how it progresses in AD. This information is vitally important to understand the mechanism of AD, to facilitate diagnoses, and to develop appropriate markers for monitoring therapeutic interventions. Similar reagents are needed to probe the role of protein variants in other significant human diseases such as Parkinson's, diabetes, and cancer.

To address this need, novel techniques have been developed to separate and purify subtle protein variants and to generate and characterize reagents that recognize specific protein forms. To demonstrate the capabilities of the present technology to generate reagents to specific protein variants, protocols are modified so that the various in vitro generated Αβ species are separated. These conditions are used to separate the much more complex Αβ pool present in AD brain.

Subtle protein variants including alternatively processed and folded forms play critical roles in numerous human diseases, many of which cannot be distinguished by genomic and proteomic techniques. Despite the growing importance of protein variants in human disease, few tools are available to study and treat these phenomena because of the difficulty in identifying individual protein isoforms. Since there are often only subtle differences between protein variants, and many may be present only at trace levels or may not be particularly stable, isolating individual protein variants and generating reagents that selectively recognize each form are extremely challenging problems. To overcome these difficulties, protocols have been developed that enable the isolation and concentration of specific protein variant forms, and the generation of reagents that selectively bind specific protein variants using only trace amounts of target.

DC-iGDEP Separation Devices and Methods

One aspect of the present work utilizes the unique abilities of dielectrophoretic (DEP) and electrokinetic (EK) forces to separate protein isoforms. DEP forces typically are used to bifurcate particle systems or uniquely capture specific targets in areas of high gradient (or low gradient, negative versus positive DEP). The approach described herein uniquely harnesses these forces to form a separation scheme not unlike isoelectric focusing (IEF), in that an overarching gradient is formed and specific targets localize according to their unique properties: pi in the case of IEF and DEP/EK force balances in the case of DC-iGDEP. A local electric field gradient can be formed with insulating structures using a static electric field applied from electrodes located in remote inlet and outlet reservoirs (Cummings, E.B. and A.K. Singh, Dielectrophoresis in Microchips Containing Arrays of Insulating Posts: Theoretical and Experimental Results. Anal. Chem., 2003. 75: p. 4724-4731; Cummings, E.B., Streaming Dielectrophoresis for Continuous-Flow Microfluidic Devices. IEEE

Engineering in Medicine and Biology Magazine, 2003. Nov/Dec: p. 75-84; Lapizco-Encinas, B.H., et al., Insulator-based dielectrophoresis for the selective concentration and separation of live bacteria in water. Electrophoresis, 2004. 25: p. 1695-1704; Lapizco-Encinas, B.H., et al., Dielectrophoretic Concentration and Separation of Live and Dead Bacteria in an Array of Insulators. Anal. Chem., 2004. 76: p. 1571-1579; Barrett, L.M., et al., Dielectrophoretic manipulation of particles and cells using insulating ridges in faceted prism microchannels. Analytical Chemistry, 2005. 77: p. 6798-6804; Lapizco-Encinas, B.H., et al., An insulator- based (electrodeless) dielectrophoretic concentrator for microbes in water. Journal of

Microbiological Methods, 2005. 62: p. 317-326; Simmons, B.A., et al., The development of polymeric devices as dielectrophoretic separators and concentrators. MRS Bulletin, 2006. 31: p. 120-124; Davalos, R.V., et al., Performance impact of dynamic surface coatings on polymeric insulator-based dielectrophoretic particle separators. Analytical and Bioanalytical Chemistry, 2008. 390: p. 847-855).

The present method adds a global gradient to the system, allowing the longitudinal separation of mixtures as opposed to a simple bifurcation of two components (Pysher, M.D. and M.A. Hayes, Electrophoretic and dielectrophoretic field gradient technique for separating bioparticles. Analytical Chemistry, 2007. 79: p. 4552-4557; Chen, K.P., et al., Insulator- based Dielectrophoretic Separations of Small Particles in Sawtooth Channel. Electrophoresis, 2009. 30: p. 1441-1448; Jones, P.V., S.J.R. Staton, and M.A. Hayes, Blood Cell Capture in a Gradient Dielectrophoretic MicroChannel. Anal. Bioanal. 2011. 401: p. 2103-2111; Staton, S.J.R., et al., Characterization of particle capture in a sawtooth patterned insulating electrokinetic microfluidic device. Electrophoresis, 2010. 31: p. 3634-3641; Weiss, N.G., et al., Dielectrophoretic mobility determination in DC insulator-based dielectrophoresis. Electrophoresis, 2011. 32, 2292-2297; Staton, S.R., et al., Gradient Insulator Based

Dielectrophoresis Isolation and Concentration of A-beta Amyloid Fibrils. Analyst, 2012. 137, in press (RSC ID: AN-COM-01-2012-035138)).

The insulating structures are fabricated such that an increasing local gradient is induced along the length of a microfluidic channel (Fig. 1). This sawtooth structure enables separations to be based on high multipole moments. Generally, simple electrophoresis accesses the monopole electric properties of a molecule or particle. For complex molecules and bioparticles, several intricate electrical geometries exist which can be used for separation (see for ex. Jones, T.B. and M. Washizu, Generalized multipolar dielectrophoretic force and electrorotational torque calculation. Journal of Electrostatics, 1996. 38: p. 199-211).

Protocols have been developed to gently (cells remain viable) separate several targets along a gradient based on specific features of each target. These features can be as subtle as deformability or as simple as size. In addition to earlier work focused on model particles, several species of bacteria and cells from whole human blood have been isolated (along with a model in support of the present work (Chen, K.P., et al., Insulator-based Dielectrophoretic Separations of Small Particles in Sawtooth Channel. Electrophoresis, 2009. 30: p. 1441- 1448)), and it has been shown that different Αβ species can be uniquely captured. All devices used to generate these results were created with photolithographic templates fabricated with AZ P4620 photoresist (Pysher, M.D. and M.A. Hayes, Electrophoretic and dielectrophoretic field gradient technique for separating bioparticles. Analytical Chemistry, 2007. 79: p. 4552- 4557).

The basic approach is demonstrated with polystyrene particles varying in size from 20 nm to 1 micron in diameter being isolated along the increasing gradient (Staton, S.J.R., et al., Characterization of particle capture in a sawtooth patterned insulating electrokinetic microfluidic device. Electrophoresis, 2010. 31: p. 3634-3641). The small space between the tips of the insulating structures (gates) on the narrowest portion of the device do not clog, consistent with the inventors' modeling studies showing that the particles never actually penetrate the narrowest zone (Chen, K.P., et al., Insulator-based Dielectrophoretic

Separations of Small Particles in Sawtooth Channel. Electrophoresis, 2009. 30: p. 1441- 1448). Differential behaviors were observed for two populations of the same sized particles, indicating that dielectrophoretic forces act on factors beyond just size— erhaps

heterogeneous permittivity (Staton, S.J.R., et al., Characterization of particle capture in a sawtooth patterned insulating electrokinetic microfluidic device. Electrophoresis, 2010. 31: p. 3634-3641). Live and dead bacteria were easily separated using this setup without clogging, verifying the modeled behavior of the micro and nano particles (Pysher, M.D. and M.A. Hayes, Electrophoretic and dielectrophoretic field gradient technique for separating bioparticles. Analytical Chemistry, 2007. 79: p. 4552-4557). It was also possible to separate red blood cells (RBCs) from other types of cell debris or proteins (Blood Cell Capture in a Gradient Dielectrophoretic MicroChannel. Anal. Bioanal. 2011. 401: p. 2103-2111). A component of this system has also been modeled, which was a series of seven tooth pairs that do not converge and no particle-particle interactions are included (Chen, K.P., et al.,

Insulator-based Dielectrophoretic Separations of Small Particles in Sawtooth Channel.

Electrophoresis, 2009. 30: p. 1441-1448). Using just these limiting constraints the ratio of electrokinetic forces (electrophoretic and electroosmotic, με _Ρ +μ _ε οί mobilities) and

dielectrophoretic forces was shown to be unique at each gate. The shape of the projected electric field and resulting field gradient is strongly influenced by the shape of the insulators.

Development of DC-iGDEP device to isolate a full range of in vitro generated Αβ species. The unique capabilities of DC-iGDEP are used to extend the production of nanobodies to create a suite of precise probes of Αβ aggregates specific to AD patients. To accomplish this, a DC-iGDEP device is fabricated that maximizes resolution of the Αβ aggregate species ranging from dimers to full fibrils, including metastable (several minutes time-scale) intermediates. The location within the DC-iGDEP device of selected isolated and concentrated Αβ40 and 42 aggregate species is confirmed, exploiting existing nanobodies with precise affinity for various protein regions and morphologies.

Separation results have been obtained using the open channel sawtooth design on a variety of systems including different size polystyrene particles (Staton, S.J.R., et al., Characterization of particle capture in a sawtooth patterned insulating electrokinetic microfluidic device. Electrophoresis, 2010. 31: p. 3634-3641), red blood cells (Blood Cell Capture in a Gradient Dielectrophoretic MicroChannel. Anal. Bioanal. 2011. 401: p. 2103- 2111), and bacteria (Pysher, M.D. and M.A. Hayes, Electrophoretic and dielectrophoretic field gradient technique for separating bioparticles. Analytical Chemistry, 2007. 79: p. 4552- 4557), and have developed a theoretical model to predict separation properties (Chen, K.P., et al., Insulator-based Dielectrophoretic Separations of Small Particles in Sawtooth Channel. Electrophoresis, 2009. 30: p. 1441-1448).

The inventors have also shown that the design can isolate and concentrate fully formed Αβ fibrils while allowing monomers to pass freely (Fig. 1). In this design, the Αβ fibrils were captured at a "gate" (closest approach of the insulating teeth) with a 27 micron spacing, resulting in a squared field strength gradient calculated (COMSOL multiphysics) to be approximately 10 ¹⁸ V ² /m ³ . These results were obtained within 5-15 minutes after loading the device with approximately 40 microliters of the monomer or fibril solution/colloid. The data demonstrate that this unique separation scheme is well controlled and is ideally suited for the selective isolation and concentration of aggregation intermediates.

According to the basic accepted theories underlying DEP and EK, the net velocity of a particle/molecule is proportional to particle radius ( ) squared [ v _DEP = f-i _DEP

μ _ΟΕΡ ∞ a ² : VDEP s the velocity of a particle due to DEP forces and E is the local electric field (Chen, K.P., et al., Insulator-based Dielectrophoretic Separations of Small Particles in Sawtooth Channel. Electrophoresis, 2009. 30: p. 1441-1448; Weiss, N.G., et al.,

Dielectrophoretic mobility determination in DC insulator-based dielectrophoresis.

Electrophoresis, 2011. 32, 2292-2297)]. Using this relationship and noting that monomelic Αβ nominally are about 1.5 nanometers across and the characteristic of length of fibrils is in the 10s to 100s of nanometers (Roychaudhuri, R., et al., Amyloid beta-Protein Assembly and Alzheimer Disease. Journal Of Biological Chemistry, 2009. 284: p. 4749-4753), the range of expected DEP forces needed is effectively bracketed [estimate E, AE ² from COMSOL, use Αβ (1-40) monomer μ _ΕΡ (10.0 mM TRIS buffer, pH 7.8) of 1.20 x 10 ^"4 cm ² /V s and Αβ (1-42) monomer μ _ΕΡ (10.0 mM TRIS buffer, pH 7.8) is 1.072 x 10 ^"4 cm ² /V s (about four peak widths difference between Αβ 40 and 42) (Picou, R., et al., Analysis of monomelic A beta (1-40) peptide by capillary electrophoresis. Analyst, 2010. 135: p. 1631-1635; Picou, R.A., et al., Analysis of A-beta (1-40) and A-beta (1-42) Monomer and Fibrils by Capillary

Electrophoresis. Journal of Chromatography B, 2011DOI: 10.1016/j.jchromb.2011.01.030)]. A device with the "gate" spacing starting at 50 microns and decreasing over two centimeters to 1 micron effectively captures, at varying points, Αβ species ranging from monomers to fibrils (Fig 1, bottom graphic). COMSOL multiphysics was used to design these devices, and the resulting underlying structures are transferred to AutoCAD for creation of

photolithographic plates.

With this design, various Αβ aggregates are separated, although not homogeneously (evenly spaced or uniquely isolated), ranging from monomers to fully formed fibrils. This takes place in a matter of minutes as a typical velocity (EP only, open portions of the device) is -lO^m/s. Further, compared to the loading volume of approximately 50 microliters, specific targets can be concentrated by several orders of magnitude (up 10 ⁶ ). Device:

Microchannel geometry consisted of sets of successively larger, equilateral triangular features lining both sides of the channel (Fig. 3). The tip of each triangle corresponded with another opposing triangle on the other side of the channel, forming sequentially narrower gaps along a converging sawtooth pattern. The smallest triangles (located near the entrance to the channel) possessed 6-μηι sides and a 5.2-μηι height. The side-length of the equilateral triangles increased by 40 μπι after every sixth repeated unit. This created a channel with an initial gap pitch of 945 μτη and a final gap pitch of 27 μπι. The whole channel length was approximately 4.1 cm with an average depth of 14 ±1 μπι.

Materials and Methods:

Microfluidic devices constructed from both glass and polydimethylsiloxane (PDMS) were used to perform the experiments. Platinum electrodes were inserted through small access ports into reservoirs at each end of the microchannel and used to apply potential across the device.

PDMS (Sylgard 184, Dow/Corning, Midland, MI) was poured over a photoresist template patterned on a silicon wafer and allowed to cure. Two-millimeter diameter holes were punched through the PDMS at each end of the channel to access the reservoirs. The PDMS, along with a glass slide, were then oxidized via oxygen-plasma treatment and then irreversibly bonded.

Cell samples were typically collected in phosphate buffer and fluorescently stained to aid visualization during experiments (Vybrant DiO, Invitrogen, Inc., Carlsbad, CA).

Buffer was pipetted into the inlet reservoir, causing the channel to fill passively via capillary action. After inspecting the device and ensuring uniform fluid distribution, sample was then pipetted into the inlet reservoir. After analyte bioparticles had entered the device via pressure-driven flow and attained uniform distribution within the channel, buffer was added to the opposite reservoir in order to balance the hydrodynamic pressure. Once the device was prepared in this manner, platinum electrodes (0.404 mm external diameter 99.9% purity, Alfa Aesar, Ward Hill, MA) were inserted through the access ports into the reservoirs and connected to a Series 225 DC power supply (Bertan High Voltage Corp., Hicksville, NY). Experiments were observed with an Olympus 1X70 microscope. Samples were illuminated using a broad-spectrum mercury lamp (H30 102 w/2, OSRAM) and an Olympus DAPI, FITC, Texas Red triple band-pass cube (Olympus, Center Valley, PA). Generation ofNanobodies. The inventors have developed novel technology enabling the generation of reagents that recognize specific protein conformations by combining the powerful imaging capabilities of Atomic Force Microscopy (AFM) with the molecular recognition diversity of phage display libraries (Barkhordarian, H., et al., Isolating recombinant antibodies against specific protein morphologies using atomic force microscopy and phage display technologies. Protein Eng Des Sel, 2006. 19: p. 497-502; Shlyakhtenko, L.S., et al., Single-molecule selection and recovery of structure-specific antibodies using atomic force microscopy. Nanomedicine, 2007. 3: p. 192-7). The inventors have also developed innovative technologies that allow the characterization of binding specificity of these reagents using only nanograms of material again utilizing AFM (Wang, M.S., et al., Characterizing Antibody Specificity to Different Protein Morphologies by AFM. Langmuir, 2008). In order to isolate single chain antibody fragments (or nanobodies) to individual aggregate forms, an AFM biopanning technology was developed that allows the visualization of the target protein morphology (Barkhordarian, H., et al., Isolating recombinant antibodies against specific protein morphologies using atomic force microscopy and phage display technologies. Protein Eng Des Sel, 2006. 19: p. 497-502; PCT US11/57887;

PCT US11/57925; PCT/US11/57904).

The protocol is particularly well suited to isolate nanobodies against different protein morphologies since it minimizes protein handling, as the target protein is added to the mica surface without modification, it uses small amounts of protein, low nanogram quantities are more than sufficient, and the target protein does not have to be purified. This basic technology ahs been utilized to isolate nanobodies that recognize different areas of monomelic Αβ and a-syn (Emadi, S., et al., Inhibiting Aggregation of alpha-Synuclein with Human Single Chain Antibody Fragments. Biochemistry, 2004. 43: p. 2871-2878; Zhou, C, et al., A human single-chain Fv intrabody blocks aberrant cellular effects of overexpressed alpha-synuclein. Mol Ther, 2004. 10: p. 1023-31 ; Liu, R., et al., Single chain variable fragments against beta-amyloid (Abeta) can inhibit Abeta aggregation and prevent abeta- induced neurotoxicity. Biochemistry, 2004. 43: p. 6959-67; Zameer, A., et al., Single Chain Fv Antibodies against the 25-35 Abeta Fragment Inhibit Aggregation and Toxicity of Abeta42. Biochemistry, 2006. 45: p. 1 1532-9), fibrillar Αβ and a-syn (Barkhordarian, H., et al., Isolating recombinant antibodies against specific protein morphologies using atomic force microscopy and phage display technologies. Protein Eng Des Sel, 2006. 19: p. 497-502; Marcus, W.D., et al., Characterization of an antibody scFv that recognizes fibrillar insulin and beta-amyloid using atomic force microscopy. Nanomedicine, 2008. 4: p. 1-7), two different oligomeric a-syn species (Emadi, S., et al., Isolation of a human single chain antibody fragment against oligomeric alpha-synuclein that inhibits aggregation and prevents alpha-synuclein-induced toxicity. J Mol Biol, 2007. 368: p. 1132-44; Emadi, S., et al., Detecting morphologically distinct oligomeric forms of alpha-synuclein. J Biol Chem, 2009. 284: p. 11048-58), and three different oligomeric Αβ species (Zameer, A., et al., Anti- oligomeric Abeta single-chain variable domain antibody blocks Abeta-induced toxicity against human neuroblastoma cells. J Mol Biol, 2008. 384: p. 917-28; Kasturirangan, S., et al., Nanobody specific for oligomeric beta-amyloid stabilizes non-toxic form. Neurobiol Aging, 2010. In press; Kasturirangan, S., et al., Isolation and Characterization of a Nanobody that Selectively Binds Brain Derived Oligomeric Beta-Amyloid. (Submitted)).

The different oligomer specific nanobodies do not show cross-reactivity, so the nanobodies binding oligomeric Αβ do not bind oligomeric a-syn and vice versa. The nanobodies work well in standard ELISA and immunohistochemistry assays as it has been shown that each of the different aggregate species recognized by the different nanobodies naturally occur in human AD or PD tissue, and that the nanobodies can be used to distinguish between AD, PD and healthy brain tissue, and block toxicity of different aggregate species (Emadi, S., et al., Isolation of a human single chain antibody fragment against oligomeric alpha-synuclein that inhibits aggregation and prevents alpha-synuclein-induced toxicity. J Mol Biol, 2007. 368: p. 1132-44; Emadi, S., et al., Detecting morphologically distinct oligomeric forms of alpha-synuclein. J Biol Chem, 2009. 284: p. 11048-58; Zameer, A., et al., Anti-oligomeric Abeta single-chain variable domain antibody blocks Abeta-induced toxicity against human neuroblastoma cells. J Mol Biol, 2008. 384: p. 917-28; Kasturirangan, S., et al., Nanobody specific for oligomeric beta-amyloid stabilizes non-toxic form. Neurobiol Aging, 2010. In press). The nanobodies currently developed have several significant advantages over conventional antibodies including: 1) they target specific morphologies of a single protein target for example recognizing a selected Αβ oligomer form, but not any a-syn oligomer forms, 2) they can be affinity matured to femtomolar levels; 3) they can be genetically modified with targeting or tag sequences; 4) they can be expressed intracellularly as intrabodies to identify intracellular Αβ species; and 5) their specificities can be carefully characterized.

Combination of the novel separation and molecular recognition technologies enables the ability to identify and concentrate specific protein isoforms connected with diseased human tissue and to generate nanobody reagents that selectively recognize the different protein isoforms characteristic of the disease. The resulting panel of nanobodies provide extremely powerful tools in this case for the AD community to better define the roles of the different Αβ species in the progression of AD. This effort provides a clear and immediately useful example of the overall strategy of highly refined reagent development.

In certain embodiments, the antibody fragments that can be used in the present invention are those listed in Table 1 below (See also Fig. 2):

Table 1

Antibody Library Specificity Assays to Applications SEQ ID

Fragment source validate demonstrated NO

A4 Tomlinson Oligomeric Dot blot, Human AD brain amino acid

(MRC) Abeta 3- time tissue, Human CSF, SEQ ID day course, Mouse AD brain NO:2; aggregates ELISA, tissue, (Dot blot nucleic acid

AFM assays). SEQ ID

(Note: 3- Immunohistochemistry NO: 8 day

oligomeric

target is

not stable

for

westerns)

El Tomlinson Oligomeric Dot blot, Human AD brain amino acid

(MRC) Abeta 1- time tissue, Human CSF, SEQ ID day course, Mouse AD brain NO:3; aggregates ELISA, tissue, (Dot blot nucleic acid

AFM assays) SEQ ID

(Note: 1- NO: 9 day

oligomeric

target is

not stable

for

westerns)

C6 Sheets Oligomeric AFM Human AD brain amino acid

(UCSF) Abeta: tissue, Human CSF, SEQ ID brain Mouse AD brain NO:l ; derived tissue, (Dot blot nucleic acid assays) SEQ ID

NO: 10 and amino acid SEQ

ID NO: 15

D5 Tomlinson Oligomeric Dot blot, Human PD brain amino acid Antibody Library Specificity Assays to Applications SEQ ID

Fragment source validate demonstrated NO

(MRC) a- time tissue, Human CSF, SEQ ID synuclein course, Mouse PD brain NO:4;

3 -day ELISA, tissue, (Dot blot nucleic acid aggregates AFM, assays, western blot, SEQ ID western Immunohistochemistry NO: 1 1 blot with tissue and cells

analysis

10H Tomlinson Oligomeric Dot blot, Human PD brain amino acid

(MRC) a- time tissue, Mouse PD SEQ ID synuclein course, brain tissue, (Dot blot NO:5;

7-day ELISA, assays, western blot, nucleic acid aggregates AFM, Immunohistochemistry SEQ ID western with tissue and cells NO: 12 blot

analysis

6E Tomlinson Fibrillar Dot blot, Human PD brain amino acid

(MRC) aggregates time tissue, Human CSF, SEQ ID

(likely not course, Mouse PD brain NO:6; protein ELISA, tissue, (Dot blot nucleic acid specific) AFM assays, western blot, SEQ ID

Immunohistochemistry NO: 13 with tissue and cells

D10 Tomlinson All forms Dot blot, Human PD brain amino acid

(MRC) of a- time tissue, Human CSF, SEQ ID synuclein course, Mouse PD brain NO:7;

ELISA, tissue, (Dot blot nucleic acid

AFM, assays, western blot, SEQ ID western Immunohistochemistry NO: 14 with tissue and cells

BSEC1 PNRL BACE1 BACE-1 Cell toxicity, cell

(Pacific cleavage catalytic culture assays

National site on assay

Research APP (Does

Labs?) not bind

Yeast soluble

library Abeta

In certain embodiments, the C6 nanobody has a sequence of SEQ ID NO: 1 :

EXPIAYGSRWIVITRGPAGHGPGTAAGVGGGLVQPGGSLRLSCAASGF TFSSYAMSWVRQAPGKGLEWVSAISGSGGSTYYADSVKGRFTISRDNS

KNTLYLQMNSLRAEDTAVYYCAKSYGSVKISCFDYWGQSTLVTVSSG

GGGSGGGGSGGGGSEIVLTQSPDSLAVSLGERATINCKSSQSVLYNSN

NKNYLAWYQQKPGQSPELLIYWASTRESGVPDRFSGSGSGTEFTLTISS LQAEDVAVYYCQQFYSTPPTFGQGTKLEIKRAAAHHHHHHGAAEQKL ISEED

In certain embodiments, the antibody fragment comprises amino acid residues 16-292 of SEQ ID NO: 1. In certain embodiments, the antibody fragment comprises or consists of amino acid sequence SEQ ID NO:l. In certain embodiments, the antibody fragment is less than 500 amino acids in length, such as between 200-450 amino acids in length, or less than 300 amino acids in length. Identification of isolated Αβ40/42 fractions that react with existing morphology- specific nanobodies.

Several different nanobodies to different Αβ morphologies including ones that selectively bind only fibrillar Αβ (Marcus, W.D., et al., Characterization of an antibody scFv that recognizes fibrillar insulin and beta-amyloid using atomic force microscopy.

Nanomedicine, 2008. 4: p. 1-7) or three different oligomeric species of Αβ (Zameer, A., et al., Anti-oligomeric Abeta single-chain variable domain antibody blocks Abeta-induced toxicity against human neuroblastoma cells. J Mol Biol, 2008. 384: p. 917-28; Kasturirangan, S., et al., Nanobody specific for oligomeric beta-amyloid stabilizes non-toxic form. Neurobiol Aging, 2010. In press; Kasturirangan, S., et al., Isolation and Characterization of a Nanobody that Selectively Binds Brain Derived Oligomeric Beta-Amyloid. (Submitted)) have been developed by the inventors. This existing pool of morphology specific nanobodies is used to verify that distinct Αβ species are separated in the DC-iGEP protocols, and to identify novel Αβ species for which there are not selective reagents.

Separation and collection of Αβ fractions. Αβ fractions are isolated and concentrated along the sawtooth gates. Each Αβ fraction is captured within the so called "cell" (the area between the gates) (Staton, S.J.R., et al., Characterization of particle capture in a sawtooth patterned insulating electrokinetic microfluidic device. Electrophoresis, 2010. 31: p. 3634- 3641). An additional design feature is needed to keep the collected targets within a confined volume, yet accessible to various reagents. Lateral channels are used that are "electrically silent" when the electric field is placed across the long axis of the device. By placing electrodes in reservoirs placed above and below the device and corresponding channels, the captured Αβ species for subsequent analysis is recovered. Reactivity of Αβ fractions with nanobody. There are numerous techniques that are used to determine binding specificity of each of the nanobodies isolated against the different target Αβ morphologies depending on the availability and stability of the target antigen. In characterizing binding specificity to the different Αβ species isolated by DC-iGDEP, most Αβ species are present in only very small amounts and low volume. For these cases, reactivity of the existing species is analyzed using AFM, which utilizes minimal material without the need for any modification. For those Αβ morphologies that are obtained in reasonable quantity, binding specificity is determined by ELIS A, western or dot blot, depending on how easy it is to purify the target aggregate morphology. The protocols for each of these assays are routinely used (Emadi, S., et al., Inhibiting Aggregation of alpha-Synuclein with Human

Single Chain Antibody Fragments. Biochemistry, 2004. 43: p. 2871-2878; Zhou, C, et al., A human single-chain Fv intrabody blocks aberrant cellular effects of overexpressed alpha- synuclein. Mol Ther, 2004. 10: p. 1023-31; Liu, R., et al., Single chain variable fragments against beta-amyloid (Abeta) can inhibit Abeta aggregation and prevent abeta-induced neurotoxicity. Biochemistry, 2004. 43: p. 6959-67; Zameer, A., et al., Single Chain Fv

Antibodies against the 25-35 Abeta Fragment Inhibit Aggregation and Toxicity of Abeta42. Biochemistry, 2006. 45: p. 11532-9; Emadi, S., et al., Isolation of a human single chain antibody fragment against oligomeric alpha-synuclein that inhibits aggregation and prevents alpha-synuclein-induced toxicity. J Mol Biol, 2007. 368: p. 1132-44; Liu, R., et al., Residues 17-20 and 30-35 of beta-amyloid play critical roles in aggregation. JNeurosci Res, 2004. 75: p. 162-71; Liu, R., et al., Proteolytic antibody light chains alter beta-amyloid aggregation and prevent cytotoxicity. Biochemistry, 2004. 43: p. 9999-10007). In certain circumstances, the nanobody specificity for some of the oligomeric Αβ samples may not be able to determined by conventional means such as western blot as described above. Several different AFM based methods are used to determine antibody specificity for antigen targets that are not suitable for analysis as described above, or that are available in only limited amounts. For example, nanobody specificity by height distribution analysis (Wang, M.S., et al.,

Characterizing Antibody Specificity to Different Protein Morphologies by AFM. Langmuir, 2008) or by recognition imaging (Marcus, W.D., et al., Isolation of an scFv targeting BRG1 using phage display with characterization by AFM. Biochem Biophys Res Commun, 2006. 342: p. 1123-9).

At least five Αβ fractions are separated, including monomelic, several oligomeric and a fibrillar form. In certain situation if the lateral channels are not sufficiently electrically silent or they are accessible under the diffusive conditions during processing, a simple pneumatic valve concept is used (Unger, M.A., et al., Monolithic microfabricated valves and pumps by multilayer soft lithography. Science, 2000. 288: p. 113-116), where the roof of the lateral channels are closed by applying pressure because the device is made of elastic polymers. Pressure is nominal, and can be applied mechanically.

The invention will now be illustrated by the following non-limiting Examples.

Example 1

Here we report a novel method for the manipulation and concentration of Αβ amyloid fibrils, implicated in Alzheimer's disease, using DC insulating gradient dielectrophoresis

(DC-iGDEP). Fibril enrichment was found to be -400%. Simulations suggest that capture of the full range of amyloid protein aggregates is possible with optimized device design.

Protein aggregates have been associated with more than 25 human diseases, including Alzheimer's disease, Parkinson's disease and type 2 diabetes. As a result, researchers in medical, biochemical and analytical chemistry fields are striving to understand the mechanisms of formation of amyloid aggregates and to identify toxic species. Protofibrils, transient metastable oligomeric aggregates that develop prior to the formation of amyloid fibrils, have been implicated in disease pathogenesis and are considered to be cytotoxic. Amyloid aggregation has been shown to form a heterogeneous mixture of oligomeric species in vitro, but the specific structure or structures of the cytotoxic species, their relationship to amyloid fibril formation and the mechanism of toxicity are unknown. Techniques capable of manipulating and concentrating various aggregate structures are of paramount importance because specific knowledge of all constituent aggregate species is needed to effectively develop therapeutics based on disrupting or altering the aggregation process. Each of the different aggregate species from monomer to mature fibrils can have unique chemical and physical properties providing a basis for differential toxicity.

Traditional means of isolating various oligomers and fibrils of Αβ amyloid and other misfolded proteins have relied heavily on techniques such as sedimentation, size-exclusion chromatography (SEC) and capillary electrophoresis (CE). Sedimentation typically uses large volume preparations and cannot effectively separate oligomers into subpopulations, given their structural similarities and the typical small volumes of aggregate samples. While CE uses small volume samples and provides greater resolution compared to other separation techniques, target populations generally are diluted rather than concentrated during separation. Similarly, SEC has proven effective for isolating relatively stable protofibril species, but as with CE, samples are generally diluted during separation, and it has been shown to disrupt some protein aggregates. Several other methods used to isolate and characterize Αβ aggregates include HPLC, gel electrophoresis, and transmission electron microscopy (TEM); however, these methods do not provide a comprehensive account of the coexisting oligomeric populations during amyloid fibril formation. TEM measurements are made after extensive sample preparation and generally only capture stable end-point species. Considering amyloid fibril formation as a series of reversible reactions suggests that all separation schemes can alter the populations of aggregate species present before separation. For instance, protofibrils can convert back to monomers. To provide the most realistic view of the population of aggregates in a sample, a separation-based technique should be as rapid and gentle as possible.

The technique described here, direct current based insulator gradient dielectrophoresis (DC-iGDEP), rapidly resolves particulates by balancing electrophoresis and dielectrophoresis within a single separation channel. (S. J. R. Staton, K. P. Chen, T. J. Taylor, J. R. Pacheco and M. A. Hayes, Electrophoresis, 2010, 31, 3634-3641.) Relative to a capillary

electrophoretic separation, the addition of dielectrophoretic forces increases the analyte- specific separation vectors to include the permeability and conductivity of the particle as well as the same parameters for the surrounding medium, which can be tuned. Combining these forces in opposition translates into discrete collection points that also concentrate the target species. This achieves both the separation and concentration of Αβ amyloid structures in a short time frame for direct quantification or use in downstream research on the properties of those structures. The separation is not based on the interaction of the aggregates with a stationary phase, reducing the impact of the separation process on the aggregate structures.

Previously, it has been demonstrated that DC-iGDEP can separate and concentrate spherical synthetic polymer nano- and microparticles from 20 nm to 1 μηι. (S. J. R. Staton, K. P. Chen, T. J. Taylor, J. R. Pacheco and M. A. Hayes, Electrophoresis, 2010, 31, 3634- 3641; M. D. Pysher and M. A. Hayes, Anal. Chem. 2007, 79, 4552-4557.) It was not known if DCiGDEP could be used successfully to manipulate species with the size, shape and chemical properties of amyloid aggregates. Here we present initial studies demonstrating that DC-iGDEP is able to rapidly and selectively concentrate Αβ amyloid aggregates, and has the potential to analyze populations of smaller protofibrillar aggregates with further tailoring of the channel design. The design and fabrication of the DC-iGDEP device is detailed in Staton et al. (S. J. Pv. Staton, K. P. Chen, T. J. Taylor, J. R. Pacheco and M. A. Hayes, Electrophoresis, 2010, 31, 3634-3641) Briefly, the DC-iGDEP microfluidic channel was fabricated in

polydimethylsiloxane (PDMS) with a glass cover slide (Figure 3), and 20 xL of Αβ sample was introduced into the reservoir at the broader end of the channel. The Αβ monomer and fibrils in PBS were prepared as described in detail previously. (B. O'Nuallain, A. K. Thakur, A. D. Williams, A. M. Bhattacharyya, S. Chen, T. Geetha and R. Wetzel, Methods Enzymol, 2006, 413, 34-74; R. Picou, J. P. Moses, A. D. Wellman, I. Kheterpal and S. D. Gilman, Analyst, 2010, 135, 1631-1635; I. Kheterpal, K. D. Cook and R. Wetzel, Methods Enzymol. 2006, 39, 584-593; R. A. Picou, I. Kheterpal, A. D. Wellman, M. Minnamreddy, G. Ku and S. D. Gilman, J. Chromatogr. B, 2011, 879, 627-632.) Additional information about the fabrication of the DC-iGDEP and Αβ sample preparation can be found in Example 2. The monomer and fibril samples were tested separately by DC-iGDEP. The results reported here were consistent and representative of multiple experiments using three independent preparations of the monomer and fibrils and multiple devices. Experiments were performed by applying voltages between 400 to 1000 V (over the entire channel length) for 1 to 15 min for each experiment.

The DC-iGDEP experiments were monitored by fluorescence microscopy. When introduced into the DC-iGDEP channel, the monomer consistently created streaming patterns at all of the applied voltages (400 to 1000 V, Figure 4), which were more apparent at higher voltages. The streamlines appearing along the centerline of the channel indicate significant influence of electrophoretic forces; however, none of the conditions tested resulted in the monomers being captured.

In contrast, when the Αβ amyloid fibril samples were tested using the DC-iGDEP channel under similar conditions, separation and concentration of the fibrils were observed (Figure 5). A TEM of the fibril sample used in this experiment is presented in Fig. 6 and clearly shows the presence of mature amyloid fibrils. Thioflavin T (10 μΜ) fluorescence measured that aggregation was complete after 7 days. Images of experiments below 400 V applied over the length of the channel indicated movement of fibrils without capture. At 400- 1000 V applied, the Αβ fibrils were captured and concentrated at the narrow points in the channel with similar local fluorescence intensity. Previous studies with synthetic nanospheres have demonstrated that particle capture typically does not obstruct the channels, and nanoparticles below the capture threshold size will pass through an intersection containing larger particles that have been concentrated and captured. (S. J. R. Staton, K. P. Chen, T. J. Taylor, J. R. Pacheco and M. A. Hayes, Electrophoresis, 2010, 31, 3634-3641.)

In order to differentiate between nonspecific and DC-iGDEP controlled capture, the applied voltage was removed after the capture and concentration of Αβ fibrils, and bright areas were monitored for dispersion away from the capture area. The bright fluorescence regions shown in Figure 5 rapidly dissipated due to diffusion and slight convection of the Αβ fibrils away from the capture regions. Following a short period of time, the voltage was reapplied to allow selective capture to occur again. This process was repeated several times to verify that the capture was the result of the DC-iGDEP as well as to demonstrate the ability of the technique to reproducibly manipulate the amyloid aggregates within a given experimental session. The fibrils controllably collected near tips of the sawtooth-patterned insulating channel. The position of capture indicates that, under these experimental conditions, the fibrils exhibited properties consistent with positive dielectrophoresis, meaning that the collection points were where the electric field intensity is greatest. (S. J. R. Staton, K. P. Chen, T. J. Taylor, J. R. Pacheco and M. A. Hayes, Electrophoresis, 2010, 31, 3634-3641 ; H. A. Pohl, Dielectrophoresis, Cambridge University Press, 5 Cambridge, 1978; K. P. Chen, J. R. Pacheco, M. A. Hayes and S. J. R. Staton, Electrophoresis, 2009, 30, 1441-1448.) According to classic theory, particles that undergo positive dielectrophoretic capture are less permeable than the surrounding medium. Some evidence of nonspecific adsorption of the amyloid aggregates was observed along the channel surface. Without surface treatments or coatings nonspecific protein adsorption is common for PDMS channels.

The ability of the DC-iGDEP method to concentrate fibrils was examined semi- quantitatively by noting the fluorescence intensity at the collection points compared to background levels in the channel prior to collection and areas where no detectable capture had occurred. The resulting enrichment of the fibril material ranged from about 350% to over 500% depending on the applied voltage (see Example 2). However, under more ideal conditions with lower fibril loads, the potential capture efficiency could be as high as 600%, simply by reducing the amount of background fluorescence. The maximum observed enrichment was 520% at 600 V.

The overall goal of this work is to separate, capture and concentrate the full range of aggregate structures generated during mature amyloid fibril formation. The experimental results presented here demonstrate that mature Αβ fibrils, but not Αβ monomer, are captured and concentrated using the current device design and the described experimental conditions. COMSOL (finite element multiphysics modeling software) calculations allow for directed device design development by exploring alternative device design parameters not yet tested experimentally. Modeling with COSMOL confirmed that for the current DC-iGDEP device design, Αβ monomer should not be captured for any reasonable experimental conditions. Details of the models used to evaluate the particle capture potential of the device can be found in Example 2. Modeling also predicts that by reducing the smallest sawtooth gap distance from 27 μπι in the current design to -10 nm, the increased field gradient strength required to allow monomer capture along with all intermediate species could be generated. By modeling the amyloid aggregates, oligomers, and monomers based on their unique electrophoretic and dielectrophoretic properties, the speed of the device design evolution can be accelerated while also being tailored for capturing particular bioanalytes and positioning them along the length of the channel.

In conclusion, this study demonstrates that DC-iGDEP can be used to manipulate and selectively capture Αβ amyloid fibrils, while influencing but not capturing Αβ monomer. DC- iGDEP successfully combines high enrichment of fibrils (up to 520%) with short analysis times (1-15 min) on a cost effective platform. Production and operation of the DC-iGDEP microdevices using PDMS are very simple. This initial study and related simulations indicate that this technique has the potential to rapidly isolate and concentrate various Αβ aggregate structures intermediate between monomer and mature fibrils. Development of new rapid and gentle separation techniques for isolating and characterizing amyloid aggregates is essential for understanding the role of protein aggregation in amyloid-linked diseases.

Example 2

A) Details of DC-iGDEP Device Fabrication

The DC-iGDEP devices were fabricated utilizing standard photolithography, fabrication, and bonding techniques. (C. Mack, Fundamental Principles of Optical

Lithography: The Science of Microfabrication, Wiley, Hoboken, 2008.) Photomasks were designed in AutoCAD (Autodesk; San Rafael, CA, USA), and photolithographic positive stamps were made using AZ P4620 photoresist (AZ Electronic Materials; Branchburg, NJ, USA) and contrast enhancement material CEM388SS (Shin-Etsu MicroSi; Phoenix, AZ, USA). Device channels were fabricated from polydimethylsiloxane (PDMS) with a microscope slide coverplate. The PDMS channels were cast using Sylgard 184 silicone elastomer kit PDMS (Dow/Corning; Midland, MI, USA). Shortly after the PDMS portion of the device was fabricated, access holes were made using a hole punch (3 mm diameter through 0.5-1 cm of PDMS), and then the PDMS portion of the device was sealed to the glass cover plate by plasma oxidation followed by contact sealing. (K. Haubert, T. Drier and D. Beebe, Lab Chip, 2006, 6, 1548-1549.) The geometry of the separatory portion of the DC- iGDEP channel consisted of successive triangular units that extended into the open volume to induce local electric field gradients. The insulating PDMS 60° triangles began with a base length of 6 μηι and a height of 5.2 μηι. Their side length and width increased by 40 μιη after every six repeats (Fig. 3), resulting in an initial gap distance of 945 μπι and a final gap distance of 27 μπι. The separatory portion of the DC-iGDEP channel connected the two reservoirs created by the hole punch, where sample and buffer were introduced into the channel. The channel depth ranged from 13 to 16 μηι.

Sample was introduced via the reservoir at the end of the channel with the larger gap distance. After sample introduction, platinum wire electrodes (0.404 mm diameter, 99.9% purity; Alfa Aesar; Ward Hill, MA, USA) were placed in each of the reservoirs in contact with the solution and attached to a power supply (Series 225, Bertram). The voltage was applied at a potential between 0 and 1000 V, for 1-15 min depending upon the experiment. Visualization was achieved using an Olympus inverted 1X70 microscope with a mercury short arc H30 103 w/2 light source from OSRAM and an Olympus DAPI, FITC, Texas Red triple band pass cube (Olympus; Center Valley, PA, USA).

Videos and still images were collected with a monochrome QICAM cooled CCD camera (Qlmaging, Inc.; Surrey, BC, Canada) and Streampix III image capture software (Norpix, Inc.; Montreal, QC, Canada). The fluorescence intensity was then analyzed with ImageJ (NIH; Bethesda, MD). Three different regions of interest (ROI) were selected (Fig. 7). The ROIs represent the capture zone (gate), a background area within the channel (recess), and a background outside of the channel (PDMS). Table 2 uses the same ROIs described in Fig. 7. The data presented in Table 3 represent multiple replicates (n>3) with isolation events occurring in the same position. Ratios of the various ROIs were utilized to determine enrichment of fluorescent fibrils in the capture zone when compared to the rest of the microfiuidic channel filled with sample as well as a background of the device outside of the channel. Each ROI represents equally sized areas.

Table 2: Fluorescence Intensity Values in ROIs for Fibril Samples. 400 V

Photo label Intensity/Area

Sate 1 212.8 Ratio

Recess 2 52.2 Gate/Recess 4.1

PDMS 3 35.9 Gate/PDMS 5.9

600 V

Photo label Intensity/Area

Gate 1 207.0 Ratio

Recess 2 40.0 Gate/Recess 5.2

PDMS 3 36.1 Gate/PDMS 5.7 ioo v

Photo label Intensity/Area

Gate 1 214.5 Ratio

Recess 2 60.0 Gate/Recess 3.6

PDMS 3 36.1 Gate/PDMS 5.9

1000 V

Photo label Intensity/Area

Gate 1 211.7 Ratio

Recess 2 61.2 Gate/Recess 3.5

PDMS 3 36.1 Gate/PDMS 5.9

B) Preparation of Αβ (1-40) Monomers and Fibrils

Both Αβ (1-40) monomer and fibril samples were prepared as described in detail previously. (R. Picou, J. P. Moses, A. D. Wellman, I. Kheterpal and S. D. Gilman, Analyst, 2010, 135, 1631-1635; R. A. Picou, I. Kheterpal, A. D. Wellman, M. Minnamreddy, G. Ku and S. D. Gilman, J Chromatogr. B, 2011, 879, 627-632; B. CWuallain, A. K. Thakur, A. D. Williams, A. M. Bhattacharyya, S. Chen, G. Thiagarajan and R. Wetzel, in Methods Enzymol. 2006, 413, 34-74.) Briefly, Αβ (1-40) peptide (W.M Keck Foundation Biotechnology Research Laboratory, Yale University; New Haven, CT) and carboxyfluorescein (FAM) labeled Αβ (1-40) peptide (Anaspec Inc.; Fremont, CA) were first treated with trifluoroacetic acid (TFA) and hexafluoroisopropanol (HFIP) to remove any preexisting aggregates. For Αβ monomer samples, the solvent was evaporated off, and the peptides were dissolved in 10.00 mM Tris at pH 7.79. FAM-Αβ (1-40) monomer was mixed with Αβ (1-40) monomer at a mass ratio of 1 :4. The total Αβ concentration of the monomer solution was determined to be 30 μΜ using a Shimadzu HPLC-UV instrument with detection at 215 nm as described previously. (R. Picou, J. P. Moses, A. D. Wellman, I. Kheterpal and S. D. Gilman, Analyst, 2010, 135, 1631-1635; R. A. Picou, I. Kheterpal, A. D. Wellman, M. Minnamreddy, G. Ku and S. D. Gilman, J Chromatogr. B, 201 1, 879, 627-632; B. O'Nuallain, A. K. Thakur, A. D. Williams, A. M. Bhattacharyya, S. Chen, G. Thiagarajan and R. Wetzel, in Methods Enzymol. 2006, 413, 34-74.)

For the Αβ (1-40) fibril samples, TFA was evaporated off, and the peptides were dissolved in HFIP. The concentration of each peptide was determined using HPLC-UV.

FAM-Αβ (1-40) monomer was mixed at a mass ratio of 1 :4 with Αβ (1-40) monomer. HFIP was evaporated off, and the peptide mixture was dissolved stepwise in equal volumes of 2.0 mM NaOH and 2X phosphate buffered saline (PBS) containing 22.8 mM phosphate, 274 mM NaCl, 5.4 mM KC1 and 0.1% NaN3 at pH 7.4. The samples were centrifuged at 50,000 g for a minimum of 10 h at 4 °C. Fibril formation was initiated by addition of a small quantity (0.1% by weight of total Αβ monomer) of fibrillar aggregates to the supernatant from a previous fibril synthesis. The mixture was then incubated at 37 °C for 7 d. Depletion in monomer during fibril formation was monitored using HPLC-UV with detection at 215 nm as described previously. (R. Picou, J. P. Moses, A. D. Wellman, I. Kheterpal and S. D. Gilman, Analyst, 2010, 135, 1631-1635; R. A. Picou, I. Kheterpal, A. D. Wellman, M. Minnamreddy, G. Ku and S. D. Gilman, J Chromatogr. B, 201 1, 879, 627-632; B. O'Nuallain, A. K.

Thakur, A. D. Williams, A. M. Bhattacharyya, S. Chen, G. Thiagarajan and R. Wetzel, in Methods Enzymol. 2006, 413, 34-74.) Fibril growth was monitored using ThT fluorescence until complete (5-7 days), and the quality of fibrils was assessed by electron microscopy. For direct comparison with Αβ (1-40) monomer samples in DC-iGDEP analysis, fibril samples were buffer exchanged from PBS to 10.00 mM Tris electrophoresis buffer at pH 7.79. (I. Kheterpal, S. Zhou, K. D. Cook and R. Wetzel, Proc. Natl. Acad. Sci. USA, 2000, 97, 13597- 13601.) The monomer-equivalent concentrations of all samples were determined to be 30 μΜ by HPLC-UV.

C) COMSOL Mathematical Modeling

In order to aid in understanding and interpreting particle behavior within the sawtooth-patterned microchannel, the electric field distribution within the microdevice was determined numerically and plotted. This was accomplished using finite element analysis software COMSOL Multiphysics 4.1 (COMSOL, Inc., Burlington, MA) in the "conductive media DC" mode. Models created in this mode are time-independent calculations of field strength within an aqueous medium. COMSOL computes the electric field by solving the Laplace equation, V (φ) = 0, for potential distribution at various points throughout the channel, along a predefined mesh. The boundary conditions are defined as distinct potentials at the channel inlet and outlet and electrical insulators at the channel walls.

A properly scaled model of the main channel geometry was produced and imported into the COMSOL environment. In order to simplify the model, a 2D approximation of the channel was utilized. Due to the placement of electrodes in distant reservoirs and the high width to depth ratio of the channel, the effects of channel depth on electrical potential were neglected. The effects of particles on the electric field distribution were also neglected.

Reservoir-channel junctions were set to predetermined potentials similar to those used during experiments. Vj _n i _et and V _ou tiet were 0 V and -550 V, respectively. All other boundaries (representing PD S and glass walls) were set to be perfect insulators; an approximation justified by the large difference in conductivity between the fluid medium and the channel walls. The conductivity and relative permittivity of the medium were set at 1.2 S/m and 78, respectively. This conductivity is consistent with values that can be obtained using standard phosphate-buffered saline.

A triangular mesh was applied to the entire channel area. The mesh contained approximately 97,000 triangular elements and 680 vertex elements. Through finite element analysis, the software approximates the electric potential at each mesh point. From these numerical values other useful parameters relevant to electrokinesis and dielectrophoresis may be determined, such as electric field strength, |E|, and the gradient of square of the electric field, V |E| . Built-in tools enable graphic representation of the resulting data.

Since dielectrophoretic force is proportional to V |E| , the magnitude of this term along the channel centerline was extracted from the COMSOL model using a 2D cut line. Localized maxima from this dataset correspond to channel constrictions or gates along the sawtooth pattern. When these values of V |E| ² are plotted as a function of gate width, the relationship can be reasonably approximated with a power function (Fig. 8). Visualizing the data in this manner illustrates the rapid scaling of dielectrophoretic force with increasing geometric constriction of the channel.

Amongst other things, dielectrophoretic force is proportional to particle size. Thus, capture of smaller particles such as protein monomers will require larger values of V |E| ² . Fig. 8 illustrates that drastically larger values of V |E| are attainable by increasing the constriction ratio at gates. Using the following assumptions and the V |E| values determined in Fig. 8, it was found that a gate width of -10 nm would be sufficient to capture Αβ monomers. The assumptions used in the calculation were that there was negligible electroosmotic flow, the Αβ monomer is spherical with a diameter of -2.5 nm, the monomer μΕΡ is 1.2 xlO ^"4 cm /(Vs), the viscosity and permittivity of the medium are similar to that of water, and the conductivity of the Αβ monomer is significantly less than that of the fluid medium. Although the foregoing specification fully discloses and enables the present invention, it is not intended to limit the scope of the invention, which is defined by the claims appended hereto.

All publications, patents and patent applications are incorporated herein by reference. While in the foregoing specification this invention has been described in relation to certain embodiments thereof, and many details have been set forth for purposes of illustration, it will be apparent to those skilled in the art that the invention is susceptible to additional embodiments and that certain of the details described herein may be varied considerably without departing from the basic principles of the invention.

The use of the terms "a" and "an" and "the" and similar referents in the context of describing the invention are to be construed to cover both the singular and the plural, unless otherwise indicated herein or clearly contradicted by context. The terms "comprising," "having," "including," and "containing" are to be construed as open-ended terms (i.e., meaning "including, but not limited to") unless otherwise noted. Recitation of ranges of values herein are merely intended to serve as a shorthand method of referring individually to each separate value falling within the range, unless otherwise indicated herein, and each separate value is incorporated into the specification as if it were individually recited herein. All methods described herein can be performed in any suitable order unless otherwise indicated herein or otherwise clearly contradicted by context. The use of any and all examples, or exemplary language (e.g., "such as") provided herein, is intended merely to better illuminate the invention and does not pose a limitation on the scope of the invention unless otherwise claimed. No language in the specification should be construed as indicating any non-claimed element as essential to the practice of the invention.

Embodiments of this invention are described herein, including the best mode known to the inventors for carrying out the invention. Variations of those embodiments may become apparent to those of ordinary skill in the art upon reading the foregoing description. The inventors expect skilled artisans to employ such variations as appropriate, and the inventors intend for the invention to be practiced otherwise than as specifically described herein.

Accordingly, this invention includes all modifications and equivalents of the subject matter recited in the claims appended hereto as permitted by applicable law. Moreover, any combination of the above-described elements in all possible variations thereof is encompassed by the invention unless otherwise indicated herein or otherwise clearly contradicted by context.