Patent Application
20170284975
The disclosure provides methods for rapid fractionation of circulating microRNAs, viral RNA and long-non-coding RNA (lncRNA) based on their associated carrier molecules. The disclosure further provides that the methods of the disclosure can be used for diagnosing a disorder in a subject by identifying specific microRNA, lncRNA and viral RNA makers associated with that disorder and specific carriers.
Circulating microRNAs have been thought to be good biomarkers for disease diagnosis, because they could be specifically secreted by diseased cells, such as cancer cells. Considering the key roles of microRNAs in regulating gene expression, the active secretion of microRNAs could be highly relevant to disease development.
Conventionally, total microRNAs are extracted from patient's serum, and the expression profiles are analyzed to see whether the patterns can be used to indicate disease stage. But such patterns have not revealed very convincing miRNA markers, although a lot of screenings have been done.
The disclosure provides a fractionation method for determining the distribution of circulating RNAs in a sample, comprising fractionating a biological fluid sample obtained from a subject into fractions comprising at least an exosome fraction, protein fraction and lipoprotein fraction, wherein each fraction comprises RNA carriers; and determining or quantitating the RNAs in each of the fractions to generate a distribution profile for the RNAs to RNA carriers in the sample. In one embodiment, the fractionating is by performing asymmetrical flow field-flow fractionation (AF4) and collecting a plurality of eluents. In another embodiment, the fractionating is by a chip-based microfluidics system. In still another embodiment of any of the foregoing the biological fluid sample is a serum sample. In a further embodiment, a serum sample is fractionated using a trapezoidal separation channel about 0.350 mm in thickness and a tip-to-tip length of about 275 mm, with an inlet triangle width of about 20 mm and outlet width of about 5 mm. In yet a further embodiment, the surface area of the accumulation wall is about 3160 mm2 with a molecular weight cutoff value of 10 kDA. In still a further embodiment, the plurality of eluents are collected as 1 minute eluents over a period of 20 to 25 minutes. In another embodiment, at least six fractions of the biological fluid sample are generated from the plurality of eluents. In a further embodiment, the six fractions result from combining 1 minute eluents collected over six separate and non-overlapping time periods. In a further embodiment, each of the six factions are enriched with an RNA carrier protein of a specific hydrodynamic diameter. In another embodiment, the fractions are enriched with proteins, high density lipoprotein (HDL), low density lipoprotein (LDL) and exosome. In still another embodiment, the RNAs are determined or quantified by deep sequencing or RT-qPCR. In another embodiment of any of the foregoing embodiments, the RNAs include microRNAs or lncRNAs, or viral RNAs. In a further embodiment, the microRNAs or lncRNAs or viral RNAs are biomarkers associated with a disease or disorder. In still a further embodiment, the disorder is cancer. In yet a further embodiment, the cancer is breast cancer. In one embodiment, of any of the foregoing, the RNAs are microRNAs comprising the sequence of SEQ ID NO:1, 2, 3, 4, 5, 6, 7, 8, and/or 9. In another embodiment, the RNAs are selected from the group consisting of let-7a, let-7b, let-7c, let-7d, let-7e, let-7f, let-7g, let-7i, miR-1, miR-100, miR-101, miR-103, miR-105, miR-106a, miR-106b, miR-107, miR-10a, miR-10b, miR-122a, miR-124a, miR-125a, miR-125b, miR-126, miR-126*, miR-127, miR-128a, miR-128b, miR-129, miR-130a, miR-130b, miR-132, miR-133a, miR-133b, miR-134, miR-135a, miR-135b, miR-136, miR-137, miR-138, miR-139, miR-140, miR-141, miR-142-3p, miR-142-5p, miR-143, miR-144, miR-145, miR-146a, miR-146b, miR-147, miR-148a, miR-148b, miR-149, miR-150, miR-151, miR-152, miR-153, miR-154, miR-154*, miR-155, miR-15a, miR-15b, miR-16, miR-17-3p, miR-17-5p, miR-181a, miR-181b, miR-181c, miR-181d, miR-182, miR-182*, miR-183, miR-184, miR-185, miR-186, miR-187, miR-188, miR-189, miR-18a, miR-18a*, miR-18b, miR-190, miR-191, miR-191*, miR-192, miR-193a, miR-193b, miR-194, miR-195, miR-196a, miR-196b, miR-197, miR-198, miR-199a, miR-199a*, miR-199b, miR-19a, miR-19b, miR-200a, miR-200a*, miR-200b, miR-200c, miR-202, miR-202*, miR-203, miR-204, miR-205, miR-206, miR-208, miR-20a, miR-20b, miR-21, miR-210, miR-211, miR-212, miR-213, miR-214, miR-215, miR-216, miR-217, miR-218, miR-219, miR-22, miR-220, miR-221, miR-222, miR-223, miR-224, miR-23a, miR-23b, miR-24, miR-25, miR-26a, miR-26b, miR-27a, miR-27b, miR-28, miR-296, miR-299-3p, miR-299-5p, miR-29a, miR-29b, miR-29c, miR-301, miR-302a, miR-302a*, miR-302b, miR-302b*, miR-302c, miR-302c*, miR-302d, miR-30a-3p, miR-30a-5p, miR-30b, miR-30c, miR-30d, miR-30e-3p, miR-30e-5p, miR-31, miR-32, miR-320, miR-323, miR-324-3p, miR-324-5p, miR-325, miR-326, miR-328, miR-329, miR-33, miR-330, miR-331, miR-335, miR-337, miR-338, miR-339, miR-33b, miR-340, miR-342, miR-345, miR-346, miR-34a, miR-34b, miR-34c, miR-361, miR-362, miR-363, miR-363*, miR-365, miR-367, miR-368, miR-369-3p, miR-369-5p, miR-370, miR-371, miR-372, miR-373, miR-373*, miR-374, miR-375, miR-376a, miR-376a*, miR-376b, miR-377, miR-378, miR-379, miR-380-3p, miR-380-5p, miR-381, miR-382, miR-383, miR-384, miR-409-3p, miR-409-5p, miR-410, miR-411, miR-412, miR-421, miR-422a, miR-422b, miR-423, miR-424, miR-425, miR-425-5p, miR-429, miR-431, miR-432, miR-432*, miR-433, miR-448, miR-449, miR-450, miR-451, miR-452, miR-452*, miR-453, miR-455, miR-483, miR-484, miR-485-3p, miR-485-5p, miR-486, miR-487a, miR-487b, miR-488, miR-489, miR-490, miR-491, miR-492, miR-493, miR-493-3p, miR-494, miR-495, miR-496, miR-497, miR-498, miR-499, miR-500, miR-501, miR-502, miR-503, miR-504, miR-505, miR-506, miR-507, miR-508, miR-509, miR-510, miR-511, miR-512-3p, miR-512-5p, miR-513, miR-514, miR-515-3p, miR-515-5p, miR-516-3p, miR-516-5p, miR-517*, miR-517a, miR-517b, miR-517c, miR-518a, miR-518a-2*, miR-518b, miR-518c, miR-518c*, miR-518d, miR-518e, miR-518f, miR-518f*, miR-519a, miR-519b, miR-519c, miR-519d, miR-519e, miR-519e*, miR-520a, miR-520a*, miR-520b, miR-520c, miR-520d, miR-520d*, miR-520e, miR-520f, miR-520g, miR-520h, miR-521, miR-522, miR-523, miR-524, miR-524*, miR-525, miR-525*, miR-526a, miR-526b, miR-526b*, miR-526c, miR-527, miR-532, miR-542-3p, miR-542-5p, miR-544, miR-545, miR-548a, miR-548b, miR-548c, miR-548d, miR-549, miR-550, miR-551a, miR-552, miR-553, miR-554, miR-555, miR-556, miR-557, miR-558, miR-559, miR-560, miR-561, miR-562, miR-563, miR-564, miR-565, miR-566, miR-567, miR-568, miR-569, miR-570, miR-571, miR-572, miR-573, miR-574, miR-575, miR-576, miR-577, miR-578, miR-579, miR-580, miR-581, miR-582, miR-583, miR-584, miR-585, miR-586, miR-587, miR-588, miR-589, miR-590, miR-591, miR-592, miR-593, miR-594, miR-595, miR-596, miR-597, miR-598, miR-599, miR-600, miR-601, miR-602, miR-603, miR-604, miR-605, miR-606, miR-607, miR-608, miR-609, miR-610, miR-611, miR-612, miR-613, miR-614, miR-615, miR-616, miR-617, miR-618, miR-619, miR-620, miR-621, miR-622, miR-623, miR-624, miR-625, miR-626, miR-627, miR-628, miR-629, miR-630, miR-631, miR-632, miR-633, miR-634, miR-635, miR-636, miR-637, miR-638, miR-639, miR-640, miR-641, miR-642, miR-643, miR-644, miR-645, miR-646, miR-647, miR-648, miR-649, miR-650, miR-651, miR-652, miR-653, miR-654, miR-655, miR-656, miR-657, miR-658, miR-659, miR-660, miR-661, miR-662, miR-663, miR-7, miR-9, miR-9*, miR-92, miR-93, miR-95, miR-96, miR-98, miR-99a, miR-99b and any combination thereof. In another embodiment, the chip-based microfluidic system comprises a microfluidic chip comprising at least 3 channels; at least 3 reservoirs; and a sample reservoir, wherein the channels fluidly connect the at least 3 reservoirs and sample reservoir; a first bead reagent comprising magnetic beads and an antibody that interacts with an antigen on exosomes; and a second bead reagent comprising cationically charged beads. In a further embodiment, the antibody is an anti-CD63 antibody. In a further embodiment, the method comprises (i) adding serum to the sample reservoir; (a) adding the first bead reagent to the sample reservoir; applying a magnetic field to the sample reservoir and moving the first bead reagent with the magnetic field through a first channel of the at least 3 channels to a first reservoir of the at least 3 reservoirs; disrupting the exosomes in the first reservoir; removing the first bead reagent; adding a second bead reagent to the first reservoir; (b) adding GuHCl, KCl, and a detergent to the sample reservoir to dissociate RNA from proteins; add the second bead reagent to the sample reservoir to bind RNA; moving the second bead reagent through a second channel of the at least 3 channels to a second reservoir of the at least 3 reservoirs; and (c) adding guanidine thiocyanate, a detergent, and ethanol to the sample reservoir to dissociate RNA from lipoproteins; add the second bead reagent to the sample reservoir to bind RNA; moving the second bead reagent through a third channel of the at least 3 channels to a third reservoir of the at least 3 reservoirs, (ii) extracting RNA from each of the first, second and third reservoir. In a further embodiment, the method further comprises reagents that can destroy the protein-RNA interaction, or the lipoprotein complexes. In a further embodiment, the reagents are a mixture of surfactant, organic solvent, chaotropic salts.
The disclosure also provides a method for diagnosing whether a subject has a disorder, comprising comparing the distribution of circulating RNAs obtained by using the method of any of the foregoing embodiments between a healthy subject(s) and subject(s) with the disorder, wherein a difference identifies a risk of the disease or disorder.
The disclosure also provides a kit for carrying out any of the methods described herein, wherein the kit is compartmentalized to contain reagents and devices for performing the methods. In a further embodiment, the kit comprises a microfluidic device, a first bead reagent, a second bead reagent, and reagents that can destroy the protein-RNA interaction, or can destroy the lipoprotein complexes.
The disclosure provides methods for the rapid fractionation of circulating microRNAs (miRNAs) based on the type of associated carrier. The fractionated miRNAs are collected, identified, and quantified by RT-qPCR. A distribution profile of each of the targeted miRNAs is then obtained. The methods disclosed herein feature rapid fractionation, high recovery, and have a low possibility of disrupting the binding between miRNAs and their carriers. Further, the methods of the disclosure enable comprehensive profiling of the location of miRNAs in various carriers, which can reveal the more sensitive and specific microRNA markers for disorder diagnosis. The distribution profile contains much richer information for interpreting the secretion and transportation pathway of the microRNAs, and their roles in disease development. Comparison of the distribution profiles of circulating miRNAs collected from healthy subject(s) and from patient(s) with a disorder(s) can not only reveal which miRNAs are associated with the disorder but can also indicate the stage of the disorder based upon which carrier is associated with the miRNA.
In a particular embodiment, the disclosure provides a rapid fractionation method for determining the distribution of circulating miRNAs in a sample, comprising: fractionating a serum sample obtained from a subject, by performing asymmetrical flow field-flow fractionation (AF4) on the sample and collecting a plurality of eluents; combining the plurality of eluents into fractions, wherein each fraction is enriched with a different miRNA carrier; quantitating the level of a set of miRNAs in each of the collected fractions to generate distribution profiles for the miRNAs to carriers in the sample; and determining the distribution of circulating miRNAs in the sample. In a further embodiment, the serum sample is fractionated using a trapezoidal separation channel about 0.350 mm in thickness and a tip-to-tip length of about 275 mm, with an inlet triangle width of about 20 mm and outlet width of about 5 mm. In yet a further embodiment, the surface area of the AF4 accumulation wall is about 3160 mm2 with a molecular weight cutoff value of 10 kDA. In another embodiment, the plurality of eluents are collected as 1 minute eluents over a period of 20 to 25 minutes. In yet another embodiment, at least six fractions of the serum sample is generated from the plurality of eluents. In a further embodiment, the six fractions result from combining 1 minute eluents collected over six separate and non-overlapping time periods. In yet a further embodiment, each of the six factions is enriched with a miRNA carrier protein of a specific hydrodynamic diameter.
In a certain embodiment, a method of the disclosure comprises fractions that are enriched with a miRNA carrier protein selected from high density lipoprotein (HDL), low density lipoprotein (LDL), and exosome. In another embodiment, a method disclosed herein comprises quantifying miRNAs by using RT-qPCR.
In a particular embodiment, a method of the disclosure comprises a set of miRNAs that are biomarkers associated with a disorder, such as a cancer, diabetes, obesity, epilepsy, liver disease (e.g., NASH or NAFLD), coronary artery disease, Alzheimer Disease, polycystic ovary syndrome, endometriosis, kidney disease (e.g., minimal change disease, focal segmental glomerulosclerosis). In a further embodiment, a method of the disclosure comprises a set of microRNAs that are biomarkers associated with breast cancer, such as those microRNAs comprising the sequence of SEQ ID NO:1, 2, 3, 4, 5, 6, 7, 8, and/or 9.
In a certain embodiment, a method disclosed herein can be used to diagnose whether a subject has a disorder, comprising: comparing the distribution of circulating microRNAs obtained by using a method of the disclosure with the distribution of circulating microRNAs from a healthy subject(s) and/or subject(s) with the disorder obtained by using that same method.
As used herein and in the appended claims, the singular forms “a,” “and,” and “the” include plural referents unless the context clearly dictates otherwise. Thus, for example, reference to “a fraction” includes a plurality of such fractions and reference to “the miRNA” includes reference to one or more miRNAs and equivalents thereof known to those skilled in the art, and so forth.
Also, the use of “or” means “and/or” unless stated otherwise. Similarly, “comprise,” “comprises,” “comprising” “include,” “includes,” and “including” are interchangeable and not intended to be limiting.
It is to be further understood that where descriptions of various embodiments use the term “comprising,” those skilled in the art would understand that in some specific instances, an embodiment can be alternatively described using language “consisting essentially of” or “consisting of.”
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood to one of ordinary skill in the art to which this disclosure belongs. Although many methods and reagents similar or equivalent to those described herein can be used in the practice of the disclosed methods and compositions, the exemplary methods and materials are now described.
All publications mentioned herein are incorporated herein by reference in full for the purpose of describing and disclosing the methodologies that might be used in connection with the description herein. With respect to any term that is presented in the one or more publications that is similar to, or identical with, a term that has been expressly defined in this disclosure, the definition of the term as expressly provided in this disclosure will control in all respects.
Cells communicate with their surrounding environment via many different pathways, including cell-cell interactions, cell-matrix interactions, hormones, growth factors, cytokines, hormones and the like. Long range effects between cells can be performed through a process of secreting factors that travel through the blood stream to act upon a distant cells. More recently, evidence shows the vesicles such as exosomes are capable of mediating such communications.
Early detection of cancer can enhance the survival rate of patients but the success strongly relies on the availability of specific and sensitive biomarkers. One class of promising biomarkers for cancer diagnosis are the microRNAs (miRNAs) and long non-coding RNAs (lncRNAs). miRNAs bind to target mRNAs and inhibit translation or induce degradation of target transcripts. Overexpression of miRNAs that inhibit the tumor suppressor genes can interfere with the anti-oncogenic pathway; while deletion or epigenetic silencing of miRNAs that target oncogenes can increase oncogenic potency. It is also recognized that miRNA profiles more accurately reflect the developmental lineage and tissue origin of human cancers than mRNA profiles. Compared to proteins, miRNAs have simpler structures and less complex post-synthesis processing; and can be detected by the highly sensitive PCR methods. More appealing, miRNAs can be released into the circulation system and be stably present at levels detectible by sensitive techniques like RT-PCR. Accumulating evidence shows that circulating miRNAs exhibit varied patterns between cancer patients and healthy controls, with the patterns of some secretory miRNAs altered in the early stage of cancer initiation. Since sampling from circulating body fluids, like blood, urine, saliva, etc. is considered to be convenient and non-invasive compared to other biopsy methods, more and more research efforts have been devoted to obtaining the comprehensive profiles of circulating miRNAs, and validate their utility as biomarkers.
The microRNAs are bound to certain carriers, such as proteins, lipoprotein particles (like HDL), and exosomes (membranous vesicles with diameter around 30-100 nm, released by cells). The carriers are highly relevant to how the microRNAs are secreted and transported in the circulation system. Therefore, RNAs in particular carriers, but not the sum quantity, are directly related to disease development. Moreover, current methods for fractionating circulating RNAs bound to carriers in serum or plasma are exclusively based upon size exclusion chromatography or ultracentrifugation.
RNA interference (RNAi) is a biological process for the control of gene expression and activity. Recently, RNAi molecules (e.g., miRNA) have been reported to be present in exosomes, high- and low-density lipoproteins (Vickers et al, 2011) (HDL/LDL), large extracellular vesicles, termed microvesicles, and are associated with Argonaut 2 (AGO2) (Arroyo et al., 201 1; Li et al., 2012; Turchinovich et al., 2011).
miRNAs are small non-coding RNAs of 18-24 nucleotides (nt) in length that control gene expression post-transcriptionally. They are synthesized via sequential actions of Drosha and Dicer endonucleases and loaded into the RISC (RNA induced silencing complex) to target mRNAs (Bartel, 2009; Maniataki and Mourelatos, 2005).
miRNAs operate via sequence-specific interaction and pairing of the miRNA-associated RISC (composed of Dicer, TRBP and AG02 proteins) with the target mRNAs (Bartel, 2009). This action inhibits translation and/or causes mRNA destabilization (Filipowicz, 2005). The degree of complementarity of the miRNA and its mRNA target dictates the process of mRNA silencing, either via mRNA destabilization/degradation or by inhibition of translation (Ambros, 2004; Bartel, 2009). If complete complementation is encountered between the miRNA and target mRNA sequence, the RISC complex acts to cleave the bound mRNA for degradation (Ambros, 2004; Bartel, 2009). If absolute complementation is not encountered, as in most cases of miRNAs in animal cells, translation is prevented to achieve gene silencing (Ambros, 2004; Bartel, 2009).
To achieve efficient miRNA-mediated gene silencing, the miRNA must be complexed with the RLC (RISC-loading complex) proteins Dicer, TRBP and AGO2. Within the RLC, Dicer and TRBP are required to process precursor miRNAs (pre-miRNAs), after they emerge from the nucleus via exportin-5, to generate miRNAs and associate with AG02. AG02 bound to the mature miRNA constitutes the minimal RISC and may subsequently dissociate from Dicer and TRBP. Single-stranded miRNAs by themselves incorporate into RISC very poorly and therefore cannot be efficiently directed to its target mRNA for post-transcriptional regulation.
Exosomes are released by cells in vivo and in vitro. By the term “exosome” is meant a lipid-based microparticle or nanoparticle present in a sample (e.g., a biological fluid) obtained from a subject. The term exosome is also referred to in the art as a microvesicle, nanovesicle or extracellular vesicles. In some embodiments, an exosome is between about 20 nm to about 90 nm in diameter. Exosomes are secreted or shed from a variety of different mammalian cell types. Exosomes are small membrane-bound vesicles that carry biological macromolecules from the site of production to target sites either in the microenvironment or at distant sites away from the origin. The content of exosomal content varies with the cell type that produces them as well as environmental factors that alter the normal state of the cell such as viral infection. Exosomes have been shown to contain viral RNA, viral proteins, viral miRNA, cellular miRNA and the like (Singh et al., Viruses, 7(6):3204-25, 2015; Hubert et al., Future Virol., 10(4):357-370, 2015).
Long noncoding RNAs (lncRNAs) include RNA molecules greater than 200 nucleotides in length that have low protein-coding potential. Traditionally viewed as transcriptional noise, they are now emerging as important regulators of cellular functions such as protein synthesis, RNA maturation/transport, chromatin remodeling, and transcriptional activation and/or repression programs. They have been shown to influence biological processes such as stem cell pluripotency, cell cycle, and DNA damage response. Indicative of their important regulatory functions, aberrant expression and function of some lncRNAs have been observed in several types of cancers (see, e.g., U.S. Pat. Publ. No. 2013/0178428, the disclosure of which is incorporated herein by reference).
Circulating microRNAs (miRNAs) are potential biomarkers useful in cancer, diabetes, obesity, epilepsy, liver disease (e.g., NASH or NAFLD), coronary artery disease, Alzheimer Disease, polycystic ovary syndrome, endometriosis, and kidney disease (e.g., minimal change disease, focal segmental glomerulosclerosis) diagnosis. Similarly, long-non-coding RNA (lncRNA) molecules have been associated with various disease as having an effect on gene expression. These RNA molecules have been found to be bound to various carriers such as proteins, lipoprotein particles, and exosomes. It is likely that the miRNAs and lncRNA associated with particular carriers, but not the overall quantity, are related to the disease states (e.g., cancer, cardiovascular, kidney, endometriosis etc.).
One obstacle to using circulating miRNAs as a diagnostic is that not all circulating miRNAs are related to cancer development or disease. The cancer/disease-irrelevant miRNAs can be secreted by blood cells; or be shed after cells die. They could then contribute to large variances in miRNA abundances between individuals and subsidize signals from the cancer-related miRNAs during quantification. It has been known that, the cell-free miRNAs are protected from nucleases in extracellular environments and in body fluids by various types of carriers. The carriers can be proteins like Argonaute (AGO) 2 and GW182 that belong to the RNA-induced silencing complex (RISC); lipoprotein (high-density lipoprotein (HDL) and low density lipoprotein (LDL)) particles that could mediate intracellular communication; or vesicles like the exosomes, which are believed to be one of the exportation routes for miRNAs from malignant cells. While active miRNA secretion by malignant cells could be the consequence of dysregulation of cellular pathways, for-purpose exportation and uptake could be related to tumor progression and metastasis. Therefore, to better eliminate the cancer-irrelevant miRNAs and reveal the more specific miRNA markers, isolation of miRNAs from carriers that are specifically secreted by cancer cells could provide a solution. Thus, HDL and exosomes have recently been focused in study of circulating miRNAs.
Furthermore, viral RNA (vRNA) associated with various carriers can be used to determine the presence of an infection, viral load, or the state of the infection (e.g., active or latent).
As used herein an “RNA carrier” refers to a macromolecule present in a fluid or tissue of a subject and to which RNA is bound or carried in the subject. In one embodiment, the RNA carrier is not a cell (“non-cellular”) (e.g., not a stem cell, parenchymal or other cell). By “bound” means covalently or non-covalently associated with the RNA carrier (e.g., encapsulated in an exosome, linked by H-bonds or other charge association and the like). Examples of RNA carriers include proteins (e.g., Argonaute (AGO)2 and GW182 that belong to the RISC complex), lipids, lipoproteins (e.g., high-density lipoproteins (HDLs) and/or low density lipoproteins (LDLs)), extracellular vesicles (e.g., exosomes), and the like. The term “RNAs” as used herein refers to one or more of miRNA, lncRNA, and viral RNA.
By the term “sample” or “biological sample” is meant any biological fluid obtained from a mammalian subject (e.g., composition containing blood, plasma, urine, saliva, breast milk, tears, vaginal discharge, or amniotic fluid).
While miRNAs enclosed in exosomes, may provide disease state information, the methods disclosed herein have found RNAs bound to other carriers are also highly relevant to disease development, as different carriers are secreted by different pathways and transported to different locations. The actual distribution pattern of RNAs among various carriers is therefore indicative to the stage of a disease and disease diagnosis. By using the methods of the disclosure, RNA quantities in separate carriers can be analyzed, allowing for the identification of specific microRNA, lncRNA and vRNA disease states.
Pure HDL or exosomes are often obtained by ultracentrifugation and immunoaffinity capture. Ultracentrifugation can provide good size/density resolution; but it requires large sample volumes, is very tedious and time-consuming, and typically provides low recovery. Immunoaffinity capture is easy to perform and provides high specificity, but can only target one type of carrier at a time. In one study of miRNA carriers, serum was fractionated with size exclusion chromatography (SEC) to reveal the existence of exosomal and exosome-free circulating miRNAs. In another study, SEC was used to further characterize the HDL isolated by ultracentrifugation. However, in SEC, good separation resolution can only be achieved within a small size range; interaction of biomolecules with the column materials is problematic; and integrity of biocomplexes or vesicle structures after passing through the packed column is questionable.
While recovering RNAs from either pure HDL or exosomes could possibly remove the cancer/disease-irrelevant RNAs shed by normal cells, it is actually not conclusive about which carriers are more important in cancer and disease diagnosis. Thus, study of RNA distribution among all types of carriers is important in answering this question.
The disclosure provides a method for rapid separation of different RNA carriers in a fluid (e.g., serum) from a subject using (i) asymmetrical flow field flow fractionation (AF4) (or an improvement thereof, see, e.g., U.S. Pat. Publ. No. 2009/0301942, which is incorporated herein by reference) or (b) a bead-based microfluidic/chip-based methods. Compared to SEC and ultracentrifugation, asymmetrical flow field-flow fractionation (AF4) and the bead-based microfluidic method are gentler and thus provide for better preservation of the binding between RNAs and their carriers. Due to its non-interactive separation ability, AF4 and the bead-based microfluidic method can be used to isolate intact macromolecular complexes of protein-RNA, lipoprotein-RNA and exosomes containing RNA.
The A4F apparatus and variants thereof are described in various publication including Giddings et al. (Science, 260:1456-1465, 1993) and Carl-Gustav Wahlund et al. (“Properties of an asymmetrical flow field-flow fractionation channel having one permeable wall,” Analytical Chemistry 59, 1332-39, 1987).
Generally an A4F unit includes (1) a bottom assembly structure holding a liquid-permeable frit, usually made from sintered stainless steel particles, (2) a permeable membrane that lies over the frit, (3) a spacer of thickness from about 75 to 800 μm containing a cavity, and (4) a top assembly structure generally holding a transparent plate of material such as glass. The resulting sandwich is held together with screws, bolts, glue or some other means. A generally rectangular or coffin-shaped cavity in the spacer serves as the channel in which separation will occur. The top assembly structure typically contains three holes that pass through the top plate, referred to as ports, that are centered above the channel and permit the attachments of fittings thereto. These ports are: (a) a mobile phase inlet port located near the beginning of the channel and through which is pumped the carrier liquid (the “mobile phase”), (b) a sample port, very close to and downstream of the inlet port, into which an aliquot of the sample to be separated is introduced to the channel, and (c) an exit port through which the fractionated aliquot leaves the channel, downstream from the inlet port and sample port.
A4F channels are used to separate particles including serum proteins, lipids and the like and spanning a size range from a few nanometers to tens of micrometers. The separation of a sample aliquot comprised of such particles depends in turn on the length, breadth, and thickness of the rectangular or coffin-shaped cavity. In addition, it depends on the channel flow rate, the ratio of the cross flow to channel flow, temperature, liquid viscosity, pH, ionicity, the physical composition of the particles themselves, and the type of permeable membrane lying over the frit. By suitably programming the time variation of the channel-to-cross flow ratio, separations of different particle classes may be improved significantly and often a great range of particle sizes present in the injected sample aliquot may be separated in the same run. Indeed, for each class of particles to be separated an optimal separation may be developed by empirically varying the foregoing factors. The only variable that cannot be changed for a specific AF4 device is the channel length.
Historically, the channel length for A4F has been on the order of 25 to 30 cm with a greatest breadth of the order on 1 to 3 cm that tapers along its length and ends at a breadth comparable to the breadth of the exit port. Recent studies have suggested that a channel of shorter length would provide certain benefits and, on this basis.
AF4 has been used for analysis of exosomes in serum. Therefore, it is a useful method for rapid separation of different miRNA carriers based on their hydrodynamic diameters, enabling the screening of RNA distribution among various carriers. Comparing the distribution profiles obtained from healthy individuals and cancer patients may help to reveal which types of carriers are more relevant to cancer development, and thus enhance the sensitivity and specificity in diagnosis when using the miRNAs enclosed in those carriers as the markers.
Accordingly, AF4 can be used for separation of different carriers in human serum. In one embodiment, AF4 is used to separate RNA carriers. RNA on (or in) such carriers can then be isolated. For example, the eluted RNAs are collected and quantified to obtain their distribution profiles among the various molecular carriers.
The disclosure also provides a device to carry out quick fractionation of RNAs based upon the main carriers. In comparison to the existing separation techniques used for miRNA fractionation, the methods and compositions of the disclosure are much faster and easier to perform; require smaller sample volumes and can be done with higher degree of automation to avoid variations introduced by human operators; and are more suitable for processing a high number of patient samples. The disclosure provides methods and devices for comprehensive screening of the distribution of circulating RNAs among various carriers. Such methods and devices facilitate the discovery of specific RNA biomarkers for disease diagnosis, and help to understand the biogenesis and functions of circulating RNAs, contributing to better diagnosis, therapy and prognosis.
Although the AF4-based method provides comprehensive distribution profiling by separating the carriers into various fractions, recovering RNAs from the large elution volumes is labor intensive and time consuming. Additionally, improved resolution between different carriers would provide better quantification. To further improve sample recovery, work efficiency, carrier resolution, and analysis throughput, while reducing sample consumption, a microchip-based distribution profiling technique was developed. This technique combines immuno-capture of the exosomes with detergent-based disruption of the protein-RNA binding to separately isolate the RNAs bound to proteins, associated with lipoprotein complexes, and enclosed in exosomes in three microchannels on a microchip. The total isolation process in the preliminary devices and methods took about 1.5 hrs with minimum manual sample handling; and only 25 μL or less serum is required. Improvements are being made in both the volume and time for processing. The eluted RNAs are of good quality and can be quantified by RT-real-time PCR or other RNA detection techniques.
The disclosure thus further describes a microfluidic/chip system is used to separate RNA carriers. In one embodiment a fluidic device is used that comprises at least one channel (e.g., 2, 3, 4, 5 or more channels), a sample reservoir for receiving a biological sample (e.g., serum) and at least one bead reservoir that comprises beads and/or can be used to remove and store beads that can bind to RNA carriers or RNA in the sample.
The microchip-based RNA distribution profiling method quantifies the circulating RNAs bound to three well-recognized carriers in a quick, high-throughput, and semi-automatic manner. The three channels on the chip separately yield the protein-bound, lipoprotein-associated, and exosomal RNAs, taking advantage of immuno-affinity and chemical reagents. As described more fully below, the on-chip method indeed yields the intended distribution profiling, and the obtained profiles can be used to distinguish between the serum samples collected from cancer patients and from healthy individuals.
The channels 40 and reservoirs 30, 50, 60, 70, 80 contain different fluids/buffers. For example, channels 40 can comprise an oil (e.g., silicone oil, mineral oil etc.), while reservoirs 30, 50, 60, 70, 80 can comprise droplets formed from an aqueous-based buffer in the oil. In this way, different reaction components can be separated in the different reservoirs 30, 50, 60, 70, 80, while remaining fluidly connected by the channels 40.
The device 10 can be made using common microfluidic fabrication technology.
During operation a sample (e.g., serum) is provided into sample reservoir 30. In order to prevent unwanted and non-specific adsorption of either miRNA or other factors (e.g., serum components) on the surfaces of the channels or wells, the surfaces can be modified with an octamethyl siloxane species to block the surface and render the channels and wells hydrophobic and inert. Referring again to
For example, for serum protein bound RNA extraction, approximately 400 μg of 1 μm bare magnetic silica beads in about 0.6 MKCl, about 0.01% Tween 20, and about 4.5M Guanidine HCl are used. For serum lipoprotein bound RNA extraction approximately 400 μg of 1 μm bare magnetic silica beads, in about 1M KCl, about 0.11% Tween 20, about 3M Guanidine HCl, about 2.5M Guanidine Thiocyanate, and about 10% EtOH are used. For serum exosome, the captured exosomes are incubated in a solution comprising about 50% EtOH/3 M Guanidine Thiocyanate, remove capture beads and add about 400 μg magnetic silica beads, about 0.1% tween 20, and about 0.6M KCl.
Disclosed herein are methods to rapidly fractionate the microRNAs based on where they locate. The methods of the disclosure employ asymmetrical flow field flow fractionation to separate the microRNA carriers in serum. The eluted fractions can then be collected. For example, if a total of 6 fractions are collected, each fraction will comprise an enriched population of a particular carrier (e.g., faction #3 is enriched with high density lipoprotein (HDL) particle, and fraction #6 is enriched with exosomes). From the eluted fractions, the microRNAs from each fraction can be extracted and quantitated. In the experiments presented herein, it was further found that quantitated microRNAs from the fractions showed significant differences between healthy individuals and those with a disorder. But if the fractions were combined, the quantitated miRNAs in sum did not demonstrate as significant a difference between healthy subjects and those with a disorder.
In a particular embodiment, the methods disclosed herein utilize AF4 or a microfluidic process to fractionate the whole serum. By utilizing the methods of the disclosure, discrete elution fractions were collected; total RNAs were extracted from each fraction; and the amounts of 8 selected miRNAs in each fraction were quantified by RT-qPCR. Alternatively, the extracted RNA can undergo deep sequencing. Proteins eluted in each fraction were also extracted and identified to reveal the identities of carriers enriched in each fraction. Accurate quantification of the miRNA in each fraction yielded the distribution profile. The distribution profiles acquired from the sera of healthy individuals were compared with those from patients with breast cancer.
The term “deep sequencing,” as used herein, refers to nucleic acid sequencing to a depth that allows each base to be read hundreds of times, typically at least about 500 times, more typically at least about 1000 times, and even more typically at least about 1500 times. Deep sequencing methods provide for greater coverage (depth) in targeted sequencing approaches. “Deep sequencing,” “deep coverage,” or “depth” refers to having a high amount of coverage for every nucleotide being sequenced. The high coverage allows not only the detection of nucleotide changes, but also the degree of heterogeneity at every single base in a genetic sample. Moreover, deep sequencing is able to simultaneously detect small indels and large deletions, map exact breakpoints, calculate deletion heterogeneity, and monitor copy number changes. In some aspects, deep sequencing strategies, as provided by Myllykangas and Ji, Biotechnol Genet Eng Rev. 27:135-58 (2010), may be employed with the methods of the present disclosure.
It was found that by using the methods of the disclosure that the quantity of some miRNAs in particular fractions exhibited more distinct difference between healthy individuals and breast cancer patients, than the overall quantity, indicating that such miRNAs, when present in some type of carriers, could be more specific and sensitive biomarkers for cancer diagnosis. The knowledge of the carrier could help to improve the understanding on the fundamentals behind differential secretion of the miRNA markers by cancer cells and their transportation pathways in the circulation system. Such information can help to interpret their functions, and help with discovery of more effective therapeutic approaches. Accordingly, compared to current SEC based fractionation methods for collecting and quantifying miRNAs bound to carriers, the methods of the disclosure allows for a comprehensive screening of the miRNAs distributed in serum and the simultaneous evaluation of the quantity of different carriers. The methods of the disclosure therefore provide rich information that is not only useful for discovering biomarkers associated with disorders, such as indicating the particular cancer stage, but also for understanding the dynamics of the section and transportation of the circulating microRNAs.
To exemplify one embodiment of the disclosure a study of two groups of human samples, one from healthy individuals (control) and the other from cancer patients (case) have revealed that, different types of miRNA carriers, such as the lipoprotein particles and exosomes, could be enriched in individual eluted fractions after AF4 separation. The quantities of eight selected miRNAs in some of the fractions also showed larger changes between the “control” and the “case”, compared to the sum values. Moreover, statistical analysis on the distribution profiles revealed more potential miRNA markers than analysis on the overall miRNA quantity.
Sera from two healthy individuals (control) or from two cancer patients (case) were fractionated. Six fractions enriching different types of miRNA carriers, such as the lipoprotein particles and exosomes, were collected. The quantities of eight selected miRNAs in each fraction were obtained by RT-qPCR to yield their distribution profiles among the carriers. Larger changes in miRNA quantity between the control and the case were detected in the fractionated results compared to the sum values. Statistical analysis on the distribution profiles also proved that, the quantities of 4 miRNAs within particular fractions showed significant difference between the controls and the cases. On contrary, if the overall quantity of the miRNA was subject to the same statistical analysis, only 2 miRNAs exhibited significant difference. Moreover, principle component analysis revealed good separation between the controls and the cases with the fractionated miRNA amounts. Accordingly, the methods disclosed herein allow for the comprehensive screening of the distribution of circulating miRNAs in carriers. The obtained distribution profiles enlarge the miRNA expression difference between healthy individuals and cancer patients, facilitating the discovery of specific miRNA biomarkers for cancer diagnosis.
The following examples are intended to illustrate but not limit the disclosure. While they are typical of those that might be used, other procedures known to those skilled in the art may alternatively be used.
Chemicals and Biomaterials.
HDL and low-density lipoprotein (LDL) were purchased from CalBioChem (EMD Millipore, Billerica, Mass.). Trizol LS reagent, 3,3′-dioctadecyloxacarbocyanine perchlorate (DiO) and Total Exosome Isolation kit were purchased from Invitrogen (Life Technologies). MicroRNA standards were purchased from Integrated DNA Technologies (Coralville, Iowa). TaqMan MicroRNA Assays specific to each miRNA strand were purchased from Applied Biosystems (Life Technologies). All chemicals used to prepare the AF4 running buffer of 1×PBS (10 mM phosphate at pH 7.4, 137 mM NaCl, 2.7 mM KCl, and 1.0 mM MgCl2), ethylene glycol, dimethyl sulfoxide, guanidine hydrochloride, RNA-grade glycogen, 2-propanol, and chloroform were purchased from Thermo Fisher (Pittsburgh, Pa.). All single proteins used as AF4 standards were purchased from Sigma-Aldrich (St. Luis, Mo.). Taq 5× master mix was purchased from New England Biolabs.
Serum Samples.
The serum sample used for exosome extraction and separation method optimization was the pooled healthy male serum from Sigma-Aldrich. The serum samples used in the distribution profile study were from voluntarily consenting patients (females) under institutional review board-approved protocols. Both breast cancer patients had infiltrating ductal carcinoma and were ER/PR/HER2-positive (ER-estrogen receptor; PR-progesterone receptor; HER2-human epidermal growth factor receptor).
Serum Fractionation by AF4.
An AF2000 system manufactured by Postnova Analytics (Salt Lake City, Utah) was used in this study. The trapezoidal separation channel was 0.350 mm thick (thickness of the spacer), and its tip-to-tip length was 275 mm, with an inlet triangle width of 20 mm and an outlet width of 5 mm. The injection loop volume was 20 μL. The surface area of the accumulation wall was 3160 mm2, which was made out of the regenerated cellulose ultrafiltration membrane (Postnova Analytics) with the molecular weight cutoff (MWCO) value of 10 kDa. The eluate exiting AF4 passed through a SPD-20A absorbance detector (Shimadzu) followed by a fraction collector (Bio-Rad). The running buffer for all samples was the 1×PBS mentioned above.
During serum fractionation, an initial focusing step of eight minutes was used, with the cross flow (the flow exiting the channel through the membrane wall) at 3.00 mL/min, tip flow (the flow entering the channel from the inlet) at 0.30 mL/min, and focus flow (a flow entering at a position further down from the inlet to focus the analyte into a narrow sample zone) at 3.00 mL/min. After focusing, there was a 1 minute transition period where the tip flow increased to 3.30 mL/min and the focus flow was reduced to zero. Afterwards, the tip flow was kept at 3.30 mL/min for five minutes, and was then reduced to 0.30 mL/min over the course of 15 minutes. In each case, the cross flow was reduced to keep the detector flow (the flow exiting the channel from the outlet) at 0.30 mL/min. A fraction collector (Bio-rad) was used to perform step-wise collection at every minute interval. These 1-min collections for each sample were then combined into 6 fractions, with fraction #1 (F1) containing the eluents collected from 6 to 9 min, F2 from 9 to 13 min, F3 from 13 to 16 min, F4 from 16 to 19 min, F5 from 19 to 23 min, and F6 from 23 to 28 min.
Protein and Particle Standards Used in Method Optimization.
Protein standards, as well as the pure HDL and LDL from Sigma, were prepared in solutions of 0.1 mg/mL for cytochrome C, albumin, transferrin, IgG, or thyroglobin. The 50-nm polystyrene beads were suspended at a concentration of 0.1 μM. Exosomes were prepared using an exosome precipitation kit (Invitrogen). In brief, whole serum was incubated with an exosome isolation reagent at a 5:1 v/v ratio for 20 minutes. The sample was then centrifuged at 4° C. to precipitate the exosomes. The supernatant was removed, and the exosomes were re-suspended in 1×PBS to give a 2× concentrated solution. The exosomes were either run in the system as-is or pre-incubated with DiO (final concentration of 5 μM) for 20 minutes at room temperature. All standards were analyzed using the same flow program but without the 5-min constant flow window.
LC-MS/MS Identification of Proteins.
Protein samples were subjected to tryptic digestion prior to LC-MS/MS analysis. Ammonium bicarbonate was added to reach a final concentration of ˜50 mM. Samples were reduced and alkylated using the standard DTT/IAA reduction alkylation protocols. Trypsin was added to the samples, and the digestion proceeded overnight at 37° C. After digestion, samples were purified using a C18 ZipTip (Millipore), and eluted in 50% acetonitrile/0.1% trifluoroacetic acid. After elution, samples were dried and resuspended in 0.1% TFA. These samples were then subjected to nano-LC-MS/MS analysis using a Waters 2695 Separations Module interfaced with a Finnegan LTQ (Thermo).
The raw data was uploaded to the Protein Prospector search engine (provided online by the University of California, San Francisco) for peptide and protein identification. Spectral counting was conducted for relative protein quantitation using the number of identified peptides for each protein (keeping replicates). In addition, specific searches were conducted for lower-abundance proteins of interest.
RNA and Protein Extraction from Collected Fractions.
Each fraction was spiked with 0.31 fmol C. elegans miRNA, cel-miR-67, and subjected to phenol-chloroform extraction using the Trizol® LS reagent (Invitrogen). Each fraction was split into several ˜450 μL aliquots, each aliquot homogenized with 1 mL Trizol LS reagent followed with the addition of 300 μL chloroform. After phase separation, the RNA-containing aqueous phase was mixed with RNA grade glycogen and the RNAs were precipitated by isopropanol (IPA). The RNA pellet was washed once by 80% ethanol, dried, and then all pellets for the same fraction were combined before going through another round of IPA precipitation and ethanol wash. The fractions were then dried and stored at −20° C. until RT-qPCR analysis. The protein-containing organic faction was precipitated using IPA and washed with 0.3 M guanidine hydrochloride in ethanol. After drying, the protein pellets were reconstituted in water.
MicroRNA Analysis.
To acquire sufficient miRNA amounts, two collections were carried out for each serum in each repeat. One collection was used to quantify hsa-miR-16, miR-191, let-7a, miR-17, miR-155, and miR-375, in which the miRNA pellets were reconstituted in 31 μL TE buffer. The other collection was for quantification of hsa-miR-21 and miR-122; and reconstitution of the miRNA pellets was done in 16 μL. The cel-miR-67 spiked into each fraction before RNA extraction was used as an internal standard to correct for sample loss during extraction, and the absolute miRNA quantity in each sample was obtained using an external standard calibration curve prepared from reactions with standard miRNAs.
The six high-abundance strands (hsa-miR-16, miR-191, let-7a, miR-17, miR-155, and miR-375) were all analyzed from a single collection. The remaining two strands (hsa-miR-21, and miR-122) were analyzed from another single collection. Prior to reverse transcription, lyophilized miRNA pellets were reconstituted in either 31 μL (for the high abundance strands) or 16 uL (for the low abundance strands). In each RT reaction, 5 μL of sample was mixed with 3 μL of a reverse transcription master mix and 2 μL of a corresponding RT primer for reach miRNA strand (TaqMan reverse transcription probe). The master mix consisted of 1.1 μL nuclease-free water, 1 μL of a 10× buffer mix, 0.13 μL of RNAse inhibitor, 0.1 μL of a dNTP mix, and 0.67 μL reverse transcriptase (all components were provided in a TaqMan reverse transcription kit). After mixing, 5 μL of silicone oil was layered on top of the RT mixture, and reverse transcription conducted on a Perkin-Elmer 2400 GeneAmp PCR system. The RT reaction consisted of a 30-minute annealing step at 16° C., a 32-minute transcription step at 42° C., and a 5-minute denaturing step at 85° C.
After RT, the samples underwent quantitative PCR (qPCR). On the qPCR plate, 1 μL of the RT product was mixed with 9 μL of qPCR master mix for a final volume of 10 μL. As an overlay, 5 μL of silicone oil was added to the top of each sample to limit evaporative loss. The master mix consisted of 4.9 μL of nuclease-free water, 1 μL of ethylene glycol, 0.1 μL of DMSO, 0.5 μL of 25 mM magnesium chloride, 2 μL of Taq 5× master mix, and 0.5 μL of TaqMan microRNA Assay 20×qPCR reagent (containing miRNA RT product specific forward and reverse PCR primers, and also a RT product specific TaqMan fluorescent probe). Each sample was plated in triplicate, as were any standards corresponding to the samples analyzed (high-versus low-abundance). The qPCR analysis was conducted on a Bio-Rad CFX real-time instrument, with an initial activation step at 95° C. for 90 seconds followed by a initial annealing step at 59° C. for 50s, then followed by a 40-cycle PCR with 30 second denaturation at 95° C. and 70 second annealing/extension at 53° C. for each cycle. Cel-miR-67 was used as an exogenous standard to account for sample loss during extraction, and miRNA levels were normalized and quantified using a standard calibration curve.
ELISA for Exosome Detection.
The total amount of proteins in each of the 6 collected fractions added into the well of the microtiter plate (Thermo, Microfluor 2 coated, flat bottom) were around 22 ng and diluted up to 50 μL with 1×PBS. The ELISA plate was incubated overnight at 4° C. to let the proteins be adsorbed onto the bottom of the well. Then, the protein solution was discarded, and the plate was washed with 200 μL 1×PBS for two times (all washing buffers used in the assay were 1×PBS), before 200 μL of the blocking buffer containing 5% non-fat milk in 1×PBS was added for each well. After 2-hr incubation at room temperature with gentle shaking, the blocking buffer was dumped and the wells were washed twice. Next, 100 μL of the primary antibody (mouse anti-human CD63, Catalog #ab8219, Abcam, Cambridge, Mass.) in 1:5000 dilution with 1×PBS was added to the wells, followed with another 2-hr incubation at room temperature. Following 4 washes, 100 μL of the secondary antibody, HRP conjugated rabbit anti mouse IgG (Catalog # ab97046, Abcam) in 1:25000 dilution was added and incubated for 1 hour at room temperature with gentle shaking. The plate was washed 4 times before 30 μL of the Perice ECL substrate (Thermo Fisher) was added, and incubated for 5 minutes. The resulted chemiluminescence was detected. Two repeats were done on the same plate. For the standard curve, two repeats of human CD63 (Sino Biology) with gradient concentrations were added in the same plate. The blank contained only 1×PBS in the adsorption step.
AF4 Separation of miRDA Carriers.
Due to the large differences in the hydrodynamic diameter (dh) between proteins and exosomes, the AF4 separation condition needs to be optimized to elute all carriers in a reasonable period of time while maintaining modest resolution between different species. In particular, elution of large particles like exosomes could take a very long time, since their diffusion rate is slow. Under a constant channel/cross flow condition, the exosomes prepared by the Total Exosome Isolation kit was injected but not eluted within 30 minutes, unless the cross flow was turned off gradually (See
The whole human serum purchased from Sigma was fractionated by the optimized AF4 method. The serum was spiked with pure HDL and LDL to determine their exact elution windows (see
Fractionation of Patient Serum and Confirmation of Carriers Eluted in Each Fraction.
Once the approximate windows for elution of the known miRNA carriers were known, sera samples collected from 2 healthy females (control, referred as Control #1 and #2) and 2 breast cancer (BC) patients (case, Case #1 and #2) (see
To confirm the identities of carriers enriched in each fraction, proteins eluted in F1-F6 were collected, digested by trypsin, and analyzed by LC-MS/MS. The relative abundance of the eluted proteins were evaluated by spectral counting, which counts the number of mass spectra collected for a specific protein. The percentage of the spectra number for a particular protein among all spectra identified in one sample should be semi-quantitatively proportional to its relative abundance in the mixture.
Apolipoproteins belonging to various lipoprotein complexes, such as apolipoprotein A-I (ApoA-I), A-II (ApoA-II) and B-100 (ApoB), were found in multiple fractions (see
LC-MS/MS did not identify marker proteins for exosomes, probably due to the signal suppression resulting from the highly abundant serum proteins like IgG and albumin. Instead, the marker protein for exosomes, CD-63, was detected in each fraction by ELISA (see
Overall, the above results point out that, F1 contained mainly albumin or proteins with MW<100 kDa. HDL and LDL should be enriched in F3 and F5, respectively; and exosomes mainly in F6, but could also be in F5. VLDL was co-eluted with exosomes in F6. Although co-elution of multiple carriers was seen using the current separation method, such as the overlap of HDL and LDL in F4, and the co-elution of exosomes and VLDL in F6, enriching specific carriers in particular fractions should already allow the look at the general distribution of miRNAs among the carriers. Higher resolution will indeed enhance the accuracy in distribution profiling, and can be achieved by injecting lower amounts serum in each round of the separation, but multiple collections are needed, increasing the overall labor in the analysis, which is not a favorable choice. Increasing the separation force by using a higher crossflow may also be beneficial to separation resolution, but the risk of losing more miRNAs due to membrane adsorption is increased. Thus, the current fractionation conditions were used in the subsequent experiments. The results demonstrate that the coarse distribution profiles were adequate in differentiating the cancer patients from healthy controls, as well as in revealing strands and particular carriers that were important to the differentiation.
Distribution of miRNAs in Serum.
The total RNAs were precipitated and reconstituted in water for quantification by RT-PCR. As stated above, sera from two groups of donors (all females) were tested. The sera from healthy individuals (Control #1 and #2); and those from breast cancer patients (Case #1 and #2) were analyzed, each with two repeated measurements. Eight miRNAs were quantified by RT-qPCR. Their sequences are listed in TABLE 1, together with the rationale of their inclusion in the study.
miRCancer; an exosomal
miRNAdola as an
miRNAdola as a
Recovery of miRNAs in the method was evaluated by quantification of miR-16 in the serum from Sigma. The total content of miR-16 was directly extracted from the whole 20-μL serum by the TRIzol reagent was compared with the sum miRNA quantity recovered from all AF4 fractions obtained with the injection of the same serum volume. A recovery as high as 98% was achieved (see
The high reproducibility in the separation step and careful processing in miRNA extraction and quantification ensured high analytical reproducibility: the RSD for the Log value of the total miRNA content in the two repeated measurements was <5% for most of the strands, except for miR-375, -21, and -122, which could vary by up to 15%. The results agreed with previous reports, large variations in the miRNA amounts were observed among individuals, even between the two samples within the same health group: the controls or the BC cases. Evaluation of the RSD of the total miRNA amount in all serum samples points out that, miR-16 and -17 had relatively more stable expression among individuals than other miRNA species. Their RSD was below 15%. However, this RSD already corresponds to about 10-fold alteration in the miRNA copy numbers if the base value is around 106. For miR-122, RSD values close to 120% were observed between the two samples within the same group.
Since each fraction enriched a particular type of miRNA carrier, the copy number found in each fraction corresponded to the miRNA level in that particular carrier. Different miRNAs showed distinct distribution patterns among the carriers, as demonstrated by the distribution profile of Case #1 (see
The miRNA copy number found in each fraction was then compared between the control and BC samples.
Statistical Analysis of the miRNA Distribution Profiles.
To see whether the distribution profile could tell the difference between healthy donors and BC patients, and whether more reliable miRNA biomarkers can be found, for the 8 miRNAs listed in TABLE 1, their quantities in each fraction were fitted in the linear mixed effects model of EQ. 1, using R 3.0.2.
Y
ijk
=μ+b
i
+b
j(i)+εijk (EQ. 1)
where i=1,2(# of patient group),
j=1,2(sample # in each group),
k=1,2 (replication),
bi: effect of ith group (fixed, 1 for the control group, 2 for the BC case group),
bj(i): effect of jth sample in group i (random, 1(1) for Control #1,2(1) for Control #2,1(2) for Case#1, 2(2) for Case #2) bj(i)˜N(0,σb2), εijkl˜N(0,σ2), bj(i) and εijk are independent.
For a miRNA in a given fraction, Y is the log value of the observed miRNA copy number. For example, for miR-16 in F1, Y111 is the Log value of the miRNA copy number from one of the two repeats of Control #1. This linear mixed effects model accounted for sample to sample variation σb2, as well as within sample variation σ2, when comparing healthy donors to BC patients, i.e., testing the hypothesis H0:b1=b2=0. This hypothesis was tested for each fraction of each one of the eight miRNAs using likelihood ratio test. To compare with standard approach, the same test was also performed on the sum of all fractions for each miRNAs. More miRNA strands (miR-16, -17, -375, and -122) in particular fractions (miR-16 in F5 and F6, -17 in F4, -375 in F4, and -122 in F4) yielded significant difference between healthy donors and BC patients at the level of 0.05, as marked by the “*” sign in see
It is interesting to see that miRNA quantity in F4 or F6 seems to matter the most in differentiating cases from controls. While F6 mainly contained exosomes, F4 enriched HDL and LDL. Then it is possible that, while all four markers may be valuable in diagnosis of breast cancer, they may be released by cancer cells via different pathways. miR-16 could be secreted in exosomes; but miR-17, -375, and -122 in the lipoprotein complexes could be more relevant to the development breast cancer than the exosomal fraction. This highlights the necessity of testing the miRNA quantities in multiple carriers, instead of in only one.
To visualize the effectiveness of the quantity of miR-16 in F5 and F6; miR-17 in F4; miR-375 in F4, and miR-122 in F4, in discriminating healthy donors and BC patients, principal component analysis (PCA) was performed using XLSTAT 2014 (Addinsoft™). The contents of each miRNA in individual fractions were considered as the variables. For example, the miR-16 content in F6 is one variable and named as miR-16-F6. A total of 8 observations were made in the study, two repeats for each sample were counted as two independent observations. PCA suggests that the first principle component with loadings −0.436, −0.598, −0.167, −0.258, 0.599 on miR-16-F5, miR-16-F6, 17-F4, 375-F4, and 122-F4, respectively, can potentially separate healthy donors from BC patients, as shown in the scores plot in
Microchip Fabrication.
In brief, the microchip was fabricated as generally depicted in
Preparation of Microbeads.
The polystyrene magnetic microbeads with an average diameter of 350 nm were conjugated to goat anti-Mouse IgG using carbodiimide crosslinking (see, e.g.,
Extraction of miRNAs.
The chip layout is shown in
The extraction started by adding 25 μL serum and 100 μg immune-beads conjugated with the anti-human CD63 IgG to the sample reservoir. The sample was pipetted up and down for 3-5 times to mix well and incubated for 30 minutes at room temperature. The beads were then moved towards Channel 3, through a wash reservoir, and then into a disruption reservoir, using a permanent magnet underneath the microfluidic chip. The wash reservoir contained 1×PBS and the disruption reservoir held 30 μL solution consisting of 75% EtOH and 2 M guanidine thiocyanate and 1% tween-20. After 15-minute incubation in the disruption reservoir the beads were removed into the connected bead reservoir, and then 20 μL of 9 M GuHCl and 4 μL of 6M KCl were added to the well, followed by 200 μg of the 1 μm magnetic silica beads. After mixing and another round of 15-minute incubation, the silica beads travelled to the elution reservoir that contained RNase-free ultrapure water, mixed, and incubated for 15 minutes to unload the miRNAs, before the silica beads were removed into the corresponding bead reservoir.
Extraction of the protein and lipoprotein bound miRNAs was started while the exosomal miRNAs were being isolated. After removal of the exosomes from the serum, 30 μL of 9M GuHCl, 6 μL of 6 M KCl, and 0.1% Tween 20 were added to the sample reservoir. Next 200 μg of the magnetic silica beads in 2 μL water were added, mixed well, and incubated for 15 minutes. Subsequently, the silica beads were magnetically dragged into Channel 1. Once the silica beads carrying the protein-bound miRNAs left the sample reservoir, 60 μL of 6 M guanidine thiocyanate, 1 μL of 10% tween 20, 15 μL of 100% ethanol and 200 μg of silica beads were added, mixed well, and incubated for 15 minutes. This time the beads would extract the lipoprotein-bound miRNAs and be moved to Channel 2. In both Channel 1 and 2, the silica beads were moved through the wash and elution reservoirs, and eventually collected in the corresponding bead reservoir.
The three eluents corresponding to the exosome, protein, and lipoprotein-bound miRNA fractions were removed from the chip, and quantified by RT-qPCR with the commercial Taqman miRNA primer assay kits specific to each target miRNA.
Confirmation of Exosome Isolation and Disruption to Release the Exosomal miRNAs.
Extraction of exosomes was confirmed by analyzing the extracted samples with asymmetrical flow field flow fractionation (AF4) and comparison of the CD63 amounts in the microfluidic-chip extraction and in the exosomes prepared by the Invitrogen Exosome Isolation Kit. AF4 separates analytes based on their hydration sizes. All samples were examined by UV-Vis absorption; and also stained with the DiO dye and detected by fluorescence for illustration of the lipid-enriched portions. As shown in
Once the exosomes were isolated, they were treated with a solution containing 75% EtOH. The high organic content destroyed the membrane structure of the exosomes, and released the enclosed miRNAs.
Confirmation of Disruption of the miRNA-Protein Complexes.
Once the exosomes were depleted, the remaining miRNAs in the sample were bound to lipoprotein complexes or to proteins such as AGO2. Protein-RNA interaction relies on H-bonding and electrostatic interaction between the negatively charged phosphate groups on RNA and the positively charged primary amines on proteins. The presence of denaturants for both RNAs and proteins would definitely affect the stability of the protein-RNA complexes, releasing the protein-bound miRNAs. To disrupt the more compact lipoprotein complexes, higher concentrations or stronger denaturants should be employed. The denaturants chosen were the combination of two Chaotropic salts, Guanidine HCl (GuHCl) and Guanidine Thiocyanate (GuSCN), the surfactant Tween 20, and the organic solvent EtOH. To realize consecutive extraction of the protein- and lipoprotein-bound miRNAs from the same serum sample, the serum depleted of exosomes was treated using the mild solution that contained about 0.5 M KCl, 0.0015% Tween 20, and 4 M Guanidine HCl to release the miRNAs bound to proteins. Once these miRNAs were removed by silica beads, more Tween 20 and the stronger denaturants of Guanidine Thiocyanate (GuSCN) and EtOH were supplied to break up the compact structures of the lipoprotein complexes and free the associated miRNAs. The final mixture contained roughly 0.25 M KCl, 1.8 M GuHCl, 2.5 M GuSCN, and 10% EtOH. Again, DiO staining and AF4 analysis was used to visualize the integrity of the HDL and LDL complexes. Before any treatment, serum stained with the DiO dye gave two large peaks when injected into the AF4 system (
Extraction of Free miRNAs by Silica Beads.
Disruption of exosomes, lipoprotein complexes, and simple protein-RNA complexes release the miRNAs from their carriers. The freed miRNAs then bind to the silica beads. Silica-based DNA or RNA extraction has been widely employed. Typically, chaotropic salts like GuHCl would be used to weaken the hydration effect of nucleic acids in aqueous solutions and promote hydrophobic interaction with the silica surface. KCl could also be included to enhance binding by forming salt bridges between the ionized silanol groups on silica surface and the phosphate backbone of RNA. Both were either present in the mild denaturing solution for protein-RNA disruption; or added with the silica beads after exosome disruption. Any extraction procedure could experience sample loss due to binding equilibrium and diffusion. To evaluate the recovery of in the microfluidic-chip method, total miRNA extraction using the Trizol reagent and two other commercial kits, the GeneJet Kit (Thermo Scientific Catalog# K0731) and PureLink Kit (Ambion, Catalog#12183020) were used. These represent the common methods employed to extract RNAs from biological samples. Following the manufacturer's protocols, only a very small fraction of the cel-miR-67 spiked in the serum was recovered using the commercial kits (
Fractionation Effect Evaluation.
The above results prove that, the fractionation method does not induce more sample loss and takes much less time than the conventional RNA extraction methods, while separately obtain the miRNAs bound to three different carriers. To further evaluate the fractionation effect, the miRNA amounts recovered from on-chip extraction were compared with those obtained from the AF4 method. Fraction 1 and Fraction 6 obtained by the AF4 method represented the protein-bound (grey bars in
As shows in
Analysis of Human Sera.
7 sera samples from human subject were obtained. These samples were collected from breast cancer patients. In addition, healthy sera was purchased from Innovative Research Inc., matching the age and race of each of the cancer subjects, as the controls. The miRNA distribution profiles of one case and one control are provided in
A number of embodiments have been described herein. Nevertheless, it will be understood that various modifications may be made without departing from the spirit and scope of this disclosure. Accordingly, other embodiments are within the scope of the following claims.
This application claims priority under 35 U.S.C. §119 from Provisional Application Ser. No. 62/045,503, filed Sep. 3, 2014, the disclosure of which is incorporated herein by reference.
This invention was made with Government support under Grant Nos. CHE-1057113 and DGE-0813967, awarded by National Science Foundation. The Government has certain rights in the invention.
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/US2015/048341 | 9/3/2015 | WO | 00 |
| Number | Date | Country | |
|---|---|---|---|
| 62045503 | Sep 2014 | US |
