Patent Application
20160034637
The instant application contains a Sequence Listing which has been submitted electronically in ASCII format and is hereby incorporated by reference in its entirety. Said ASCII copy, created on Feb. 5, 2014, is named 15892-0005_SL.txt and is 93,776 bytes in size.
The invention relates to methods for identifying T-cell receptor antibody in a population of cells and methods for using that information to measure immune status of a patient and predict the likelihood of which disease the patient might have.
Scientists have known for a number of years that certain discos associated with particular genes or genetic mutations. Genetic causation, however, accounts for only a portion the diseases diagnosed in humans. Many diseases appear to be linked in some way to the immune system's response to infectious and environmental agents, but bow the immune system plays a role in diseases such as cancer, Alzheimer's, costochondritis, fibromyalgia, lupus, and other diseases is still being determined.
The human genome comprises a total number of 567-588 IG (immunoglobulin) and TR (T cell receptor) genes (339-354 IG and 228-234 TR) per haploid genome, localized in the 7 major loci. They comprise 405-418 V, 32 D, 105-109 J and 25-29 C genes. The number of functional IG and TR genes is 321-353 per haploid genome. They comprise 187-215 V, 28 D, 86-88 J and 20-21 C genes (http://imgt.cines.fr). Through rearrangement of these genes, an estimated 2.5×102 possible antibodies or T cell receptors can be generated.
A few diseases to date have been associated with the body's reaction to a common antigen (Prinz, J. et al., Eur. J. Immunol. (1999) 29(10): 3360-3368, “Selection of Conserved TCR VDJ Rearrangements in Chronic Psoriatic Plaques Indicates a Common Antigen in Psoriasis Vulgaris) and/or to specific VDJ rearrangements (Tamaru, J. et al., Blood (1994) 84(3): 708-715. “Hodgkin's Disease with a B-cell Phenotype Often Shows a VDJ Rearrangement and Somatic Mutations in the VH Genes). What is needed is a better method for evaluating changes in human immune response cells and associating those changes with specific diseases.
The invention relates to a method for evaluating changes in immune response populations and associating those changes with a specific disease. In one aspect of the invention, the method composes the steps of (a) isolating a subpopulation of white blood cells from at least one human or animal subject, (b) isolating RNA from the subpopulation of cells, (c) amplifying the RNA using RT-PCR in a first amplification reaction to produce amplicons using nested primers, at least a portion of the nested primers comprising additional nucleotides to incorporate into a resulting amplicon a binding site for a communal primer, (d) separating the amplicons from the first amplification reaction from one or more unused primers from the first amplification reaction, (e) amplifying, by the addition of communal primers in a second amplification reaction, the amplicons of the first amplification reaction having at least one binding site for a communal primer, and (f) sequencing the amplicons of the second amplification reaction to identify antibody and,or receptor rearrangements in the subpopulation of cells. In one embodiment, the subpopulation may comprise a whole blood population or another mixed population sample.
In one embodiment, the step of isolating a subpopulation of white blood cells may be performed by flow cytometry to separate naïve B cells, mature B cells, memory B cells, naïve T cells, mature T cells, and memory T cells. In various embodiments of the method, the recombinations in the subpopulation of cells are rearrangements of B-cell immunoglobulin heavy chain (IgH), kappa and/or lambda light chains (IgK, IgL) T-cell receptor Alpha Beta, Gamma, Delta. In an additional embodiment.
In another aspect of the invention, the method may optionally comprise an additional step comprising (g) comparing the rearrangements identified for a population of individuals to whom a vaccine has been administered with the rearrangements identified for a population of individuals to whom the vaccine was not administered to evaluate the efficacy of the vaccine in producing an immune response.
The method may also optionally comprise the additional step of (g) comparing the rearrangements identified for a population of normal individuals with the rearrangements identified for a population of individuals who have been diagnosed with a disease to determine if there is a correlation between a specific rearrangement or set of rearrangements and the disease.
In various aspects, the method can produce semi-quantitative amplification of polynucleotides comprising complementarity determining region 3 (CDR3s), which result from genetic rearrangements within T or B cells and are responsible for the affinity and specificity of antibodies and/or T cell receptors for specific antigens. Semi-quantitative amplification provides a method to not only detect the presence of specific CDR3 sequences, but also determine the relative abundance of cells which have produced the necessary recombination events to produce those CDR3 sequences.
One aspect of the invention therefore relates to a method for analyzing semi-quantitative sequence information to provide one or more immune status reports for a human or animal. The method for producing an immune status report comprising the steps of (a) identifying one or more distinct CDR3 sequences that are shared between a subject's immunoprofile and a cumulative immunoprofile from a disease library stored in a database, summing a total number of a subjects detected sequences corresponding to those shared distinct CDR3 sequences, and computing the percentage of the total number of detected sequences in the subject's immunoprofile that are representative of those distinct CDR3s shared between the subject's immunoprofile and the disease library to create one or more original sharing indices, (b) randomly selecting sequences from a public library stored in a database to form a sub-library, the sub-library comprising a number of sequences that is approximately equal to the number of distinct CDR3 sequences in the disease library, identifying one or more distinct CDR3 sequences that are shared between the subject's immunoprofile and the sub-library, summing a total number of detected sequences corresponding to those shared CDR3 sequences, and calculating a percentage of the total number of detected sequences in the subject's immunoprofile that are shared between the subject's immunoprofile and the sub-library to create a sampling sharing index (c) repeating step (b) at least 1000 or more times and (d) estimating the P-value as the fraction of times the sampling sharing indices are greater then or equal to the original sharing index between a patient's immunoprofile and a disease library.
The disclosure can be better understood with reference to the following drawings. The elements of the drawings are not necessarily to scale relative to each other, emphasis instead being placed upon dearly illustrating the principles of the disclosure. Furthermore, like reference numerals designate corresponding parts throughout the several views.
a and
a and
The inventors have developed methods for evaluating antibody and T cell receptor rearrangements from a large number of cells, the methods being useful for comparing rearrangements identified in populations of individuals to determine whether there is a correlation between a specific rearrangement or set of rearrangements and a disease, or certain symptom of a disease. The method is also useful for establishing a history of the immune response of an individual or individuals in response to infectious and/or environmental agents as well as for evaluating the efficacy of vaccines.
The invention relates to a method for evaluating changes in immune response cell populations and associating those changes with a specific disease. In one aspect of the invention, the method comprises the amps of (a) isolating a subpopulation of white blood cells from at least one human or animal subject, (b) isolating RNA from the subpopulation of cells, (c) amplifying the RNA using RT-PCR in a first amplification reaction to produce amplicons using nested primers at least a portion of the nested primers comprising additional nucleotides to incorporate into a resulting amplicon a binding site for a communal primer, (d) separating the amplicons from the first amplification reaction from one or more unused primers from the first amplification reaction, (e) amplifying, by the addition of communal primers in a second amplification reaction, the amplicons of the first amplification reaction having at least one binding site for a communal primer, and (f) sequencing the amplicons of the second amplification reaction to identify antibody and/or receptor rearrangements in the subpopulation of cells. In one embodiment, the subpopulation may comprise a whole blood population or another mixed population sample.
In one embodiment, a peripheral blood sample is taken from a patient and the step of isolating a subpopulation of white blood cells may be performed by flow cytometry to separate naïve B cells, mature B cells, memory B cells, naïve T cells, mature T cells, and memory T cells. In various embodiments of the method, the recombinations in the subpopulation of cells are rearrangements of B-cell immunoglobulin heavy chain (IgH), kappa and/or lamba light chains (IgK, IgL), T-cell receptor Beta, Gamma, or Delta.
In a second aspect of the invention, the method may comprise an additional step (g) comparing the rearrangements identified for a population of normal individuals with the rearrangements identified for a population of individuals who have been diagnosed with a disease to determine if there is a correlation between a specific rearrangement or set of rearrangements and the disease.
In another aspect of the invention, the method may comprise an additional step comprising (g) comparing the rearrangements identified for a population of individuals to whom a vaccine has been administered with the rearrangements identified for a population of individuals to whom the vaccine was not administered to evaluate the efficacy of the vaccine in producing an immune response.
In some embodiments, the step of separating the amplicons from the first amplification reaction from one or more unused primers from the first amplification reaction may be omitted and the two amplification reactions may be performed in the same reaction tube.
The inventor previously developed a PCR method known as tem-PCR, which has been described in publication number WO20051038039, the disclosure of which is herein incorporated by reference in its entirety. More recently, the inventor has developed a method called arm-PCR, which was described in U.S. provisional patent application No. 61/042,259, the disclosure of which is herein incorporated by reference in its entirety. Also described is an apparatus for detecting target polynucleotides in a sample, the apparatus comprising a first amplification chamber for thermocycling to amplify one or more target polynucleotides to produce amplicons using nested primers, at least a portion of the nested primers composing additional nucleotides to incorporate into a resulting amplicon a binding site for a communal primer; a means for separating the amplicons from the first amplification reaction from one or more unused primers from the first amplification reaction and a second amplification chamber for thermocycling to amplify one or more amplicons produced during the first amplification reaction by the addition of communal primers in a second amplification reaction, the amplicons of the first amplification reaction having at least one binding site for at least one communal primer.
Also described is a PCR chip comprising a first PCR chamber fluidly connected to both a waste reservoir and a second PCR chamber, the waste reservoir and second PCR chamber each additionally comprising at least one electrode, the electrodes comprising, a means for separating amplicons produced from the first PCR chamber. The second PCR chamber is fluidly connected to a hybridization and detection chamber, the hybridization and detection chamber comprising microspheres, or beads, arranged so that the physical position of the beads is an indication of a specific target polynucleotide's presence in the sampled analyzed by means of the chip.
The tem-PCR, and especially the arm-PCR, methods provide semi-quantitative amplification of multiple polynucleotides in one reaction. Additionally, arm-PCR provides added sensitivity. Both provide the ability to amplify multiple polynucleotides in one reaction, which is beneficial in the present method because the repertoire of various T and B cells, for example, is so large. The addition of a communal primer binding site in the amplification reaction, and the subsequent amplification of target molecules using communal primers, gives a quantitative, or semi-quantitative result—making it possible to determine the relative amounts of the cells comprising various rearrangements within a patient blood sample. Clonal expansion due to recognition of antigen results in a larger population of cells which recognize that antigen, and evaluating cells by their relative numbers provides, a method for determining whether an antigen exposure has influenced expansion of antibody-producing B cells or receptor-bearing T cells. This is helpful for evaluating whether there may be a particular population of cells that is prevalent in individuals who have been diagnosed with a particular disease, for example, and may be especially helpful in evaluating whether or not a vaccine has achieved the desired immune response in individuals to whom the vaccine has been given.
There are several commercially available high throughput sequencing technologies, such as Roche Life Sciences's 454 sequencing. In the 454 sequencing method, 454A and 454B primers are linked onto PCR products either during PCR or ligated on after the PCR reaction. When done in conjunction with tem-PCR or arm-PCR, 454A and 454B primers may be used as communal primers in the amplification reactions. PCR products, usually a mixture of different sequences, are diluted to about 200 copies per μl. In an “emulsion PCR” reaction, (a semisolid gel like environment) the diluted PCR products are amplified by primers (454A or 454B) on the surface of the microbeads. Because the PCR templates are so dilute, usually only one bead is adjacent to one template, and confined in the semisolid environment, amplification only occurs on and around the beads. The beads are then eluted and put onto a plate with specially designed wells. Each well can only hold one bead. Reagents are then added into the wells to came out pyrosequencing. A fiber-optic detector may be used to read the sequencing reaction from each well and the data is collected in parallel by a computer. One such high throughput reaction could generate up to 60 million reads (60 million beads) and each read can generate about 300 bp sequences.
One aspect of the invention involves the development of a database of “personal immunorepertoires,” or immunoprofiles, so that each individual may establish a baseline and follow the development of immune responses to antigens, both known and unknown, over a period of years. This information may, if information is gathered from a large number of individuals, provide an epidemiological database that will produce valuable information, particularly in regard to the development of those diseases, such as cancer and heart disease, which are thought to often arise from exposure to viral or other infectious agents or transformed cells, many of which have as yet been unidentified. One particularly important use for the method of the invention involves the evaluation of children to determine whether infectious disease, environmental agents, or vaccines may be the cause of autism. For example, many have postulated that vaccine administration may trigger the development of autism. However, many also attribute that potential correlation to the use of agents such as thimerosol in the vaccine, and studies have demonstrated that thimerosol does not appear to be a causative agent of the disease. There is still speculation that the development of cocktail vaccines has correlated with the rise in the number of cases of autism, however, gathering data to evaluate a potential causal connection for multiple antigens is extremely difficult. The method of the present invention simplifies that process and may provide key information for a better understanding of autism and other diseases in which the immune response of different individuals may provide an explanation for the differential development of disease in some individuals exposed to an agent or a group of agents, while others similarly exposed do not develop the disease.
Imbalances of the immunoprofile, triggered by infection, may lead to many diseases, including cancers, leukemia, neuronal diseases (Alzheimer's, Multiple Sclerosis, Parkinson's, autism etc.), autoimmune diseases, and metabolic diseases. These diseases may be celled immunoprofile diseases. There may be two immunoprofile disease forms. (1) a “loss of function” form, and (2) a “gain of function” form, in the “loss of function” form, a person is susceptible to a disease because his/her restricted and/or limited immunoprofile lacks the cells that produce the most efficient and necessary IGs and TRs. In the “gain of function” form, a person is susceptible to a disease because his/her immunoprofile gained cells that produce IGs and TRs that normally should not be there. In the “loss of a function” (LOF) immunoprofile diseases, an individual does not have the appropriate functional B or T cells to fight a disease. His/her HLA typing has determined that those cells are eliminated during the early stages of the immune cell maturation process, the cells generally being eliminated because they react to strongly to his/her own proteins.
One aspect of the invention also provides a method comprising (a) amplifying and sequencing one or more RNAs from the T cells and/or B cells from one or more individuals, (b) inputting the sequences into a database to provide data which may be stored on a computer, server, or other electronic storage device, (c) inputting identifying information and characteristics for an individual corresponding to the sequences of the one or more RNAs as data which may also be stored on a computer, server, or other electronic storage device, and (d) evaluating the data of step (b) end step (e) for one or more individuals to determine whether a conviction exists between the one or more RNA sequences and one or more characteristics of the individual corresponding to the sequence(s). Identifying information may include, for example, a patient identification number, a code comprising the patient's HLA type, a disease code comprising one or more clinical diagnoses that may have been made, a “staging code” comprising the date of the sample, a cell type code comprising the type of cell subpopulation from which the RNA was amplified and sequenced, and one or more sequence codes comprising the sequences identified for the sample.
The described method includes a novel primer design that riot only allows amplification of the entire immunorepertoire, but also allows amplification in a highly multiplex fashion and semiquantitatively. Multiplex amplification requires that only a few PCR or RT-PCR reactions will be needed. For example, all IGs may be amplified in one reaction, or it could be divided into two or three reactions for IgH, IgL or IgK. Similarly, the T-cell receptors (TRs) may be amplified in just one reaction, or may be amplified in a few reactions including TRA, TRB, TRD, and TRG. Semi-quantitative amplification means that all the targets in the multiplex reaction will be amplified independently, so that the end point analysis of the amplified products will reflect the original internal ratio among the targets.
In various aspects, the method can produce semi-quantitative amplification of polynucleotides comprising complementarity determining regions (CDRs), which result from genetic rearrangements within T or B cells and are responsible for the affinity and specificity of antibodies and/or T cell receptors for specific antigens. Semi-quantitative amplification provides a method to not only detect the presence of specific CDR3 sequences, but also determine the relative numbers of cells have produced the necessary recombination events to produce those CDR3 sequences.
One aspect of the invention therefore relates to a method for analyzing semi-quantitative sequence information to provide one or more immune status reports for a human or animal. The method for producing an immune status report comprising the steps of (a) identifying one or more distinct CDR3 sequences that are shared between a subject's immunoprofile and a disease library stored in a database summing the total of those shared CDR3 sequences and computing the percentage of the total number of sequences in the subject's immunoprofile that are shared between the subject's immunoprofile and the disease library to create one or more original sharing indices; (b) randomly selecting sequences from a public library stored in a database to form a sub-library, the sub-library comprising a number of sequences that is approximately equal to the number of distinct sequences in the disease library, identifying one or more distinct CDR3 sequences that are shared between the subject's immunoprofile and the sub-library, summing the total of those shared CDR3 sequences and calculating the percentage of the total number of sequences in the subject's immunoprofile that are shared between the subject's immunoprofile and the sub-library to create a sampling sharing index; (c) repeating step (b) at least 1000 or more times; and (d) estimating the P-value as the fraction of times the sampling sharing indices are greater than or equal to the original sharing index between a patient's immunoprofile and a disease library.
The inventors have discovered that the immunoprofile of individuals who have certain diseases, such as, for example, cancer, autoimmune disease, etc., may be characterized by a lack of diversity in one or more immune cell population(s).
The list of each distinct CDR3-expressing cell, and the numbers of such cells represented within a blood or tissue sample from a human or animal, can constitute an immunoprofile for that human or animal. Compiling the immunoprofiles from a group of humans, for example, the group comprising both healthy individuals and individuals with various different diseases may provide a “public library” that is representative of the type of diversity found in a normal population (
An immunoprofile, comprising a listing of distinct CDR3-expressing cells (“distinct CDR3s”, those cells sharing a unique CDR3 sequence) and the numbers of each distinct CDR3 present in a blood or tissue sample from an individual may be produced for an individual patient. The patient's immunoprofile is compared to the combined immunoprofiles of a group of patients who have been diagnosed with a particular disease (a disease library, stored in a database). This can be done for a series of disease libraries, and shown in
Millions of possible combinations are possible for the public library, the immune systems of most of those individuals generally exhibiting increased diversity over that of a group of individuals who have been diagnosed with a specific disease. Therefore, the inventors determined that an accurate assessment and comparison for the method of the invention would be facilitated by the step of preparing sub-libraries by randomly sampling/selecting from the lists of distinct CDR3s and their numbers in the public library. The number of distinct CDR3 s, represented by unique peptide sequence of CDR3 fragments, should be approximately equal to the number of distinct CDR3s identified in the disease library, or an average calculated from more than one disease library. Producing a significant number of sub-frames, such as, for example, 1000 or more sub-frames, produced by randomly sampling from the public library, increases the presence of a variety of distinct CDR3s and produces a result that is statistically significant effective for identifying and characterizing an individual patient's immunoprofile as normal (“healthy”) or characterized by the presence of a type and number of cells that have been associated with a particular disease.
In the method of the invention, a patient supplies a clinical sample comprising, for example, blood or tissue, from which distinct CDR3s are semi-quantitatively amplified and sequenced. This provides the identity and the relative abundance of each CDR3 for all distinct CDR3s. This information may be entered into a program which accesses a database containing at least one public library and one or more disease libraries. Software used for data entry and/or analysis may be accessed via internet access to the database, or may be located on an individual personal computer, with internet access to the sequence information in the database. Comparisons are obtained between the individual immunoprofile and the various libraries and sub-libraries, and results are generated as generally illustrated in
As sequence data is compiled and stored in one or more databases for multiple populations of individuals, it may additionally be possible to associate certain sharing indexes with libraries representing populations with pre-conditions predispositions to certain diseases. The immune system is both proactive and reactive, and changes in the immune system, reflected in the immunoprofile, may provide the first—and sometimes the only—signal that a predisposition, a precondition, or even an established disease is present. The inventors have utilized the method to demonstrate that certain types of cancers, inflammatory bowel disease, and certain viral infections may be detected by determining the sharing index between a patient and an established disease library, obtained by sequencing CDR3s using the ARM-PCR method to produce a subset of the immunorepertoire representing the CDR3s present.
The results are even more reliable when a filter is applied to the sequence data. For example, the inventors have developed a “SMART” filter for the sequence data that aids in the generation of significantly more reliable results. This is described further in the Examples.
By way of further explanation, the following example may be illustrative of the methods of the invention. Blood samples may be taken from children prior to administration of any vaccines, those blood samples for each child establishing a “baseline” from which future samples may be evaluated. For each child, the future samples may be utilized to determine whether there has been an exposure to an agent which has expanded a population of cells known to be correlated with a disease, and this may serve as a “marker” for the risk of development of the disease in the future. Individuals so identified may then be more closely monitored so that early detection is possible, and any available treatment options may be provided at an earlier stage in the disease process.
By means of providing another example, blood samples may be taken from children prior to administration of any vaccines, those blood samples from each child establishing a “baseline” from which future samples may be evaluated. For each child and for the entire population of children in the study, those baselines may be compared to the results of RNA sequencing of T and B cells using target-specific primers to amplify antibody and T-cell receptor, after vaccine administration. The comparison may further involve the evaluation of data regarding symptoms, diagnosed diseases, and other information associated for each individual with the corresponding antibody, and T-cell receptor sequences. If a relationship exists between the administration of a vaccine and the development of a particular disease, individuals who exhibit symptoms of that disease may also share a corresponding antibody or T-cell receptor, for example, or a set of corresponding antibodies or T-cell receptors.
The method of the invention may be especially useful for identifying commonalities between individuals with autoimmune diseases, for example, and may provide epidemiological data that will better describe the correlation between infectious and environmental factors and diseases such as heart disease, atherosclerosis, diabetes, and cancer—providing “biomarkers” that signal either the presence of a disease, or the tendency to develop disease.
The method may also be useful for development passive immunity therapies. For example, following exposure to an infectious agent, certain antibody-producing B cells anchor T cells are expanded. The method of the invention enables the identification of protective antibodies, for example, and those antibodies may be utilized to provide passive immunity therapies in situations where such therapy is needed.
The method of the invention may also provide the ability to accomplish targeted removal of cells with undesirable rearrangements, the method providing a means by which such cells rearrangements may be identified.
The inventor has identified and developed target-specific primers for use in the method of the invention. T-cell-specific primers are shown in Table 1, and antibody-specific primers are shown in Table 2. An additional embodiment of the invention is a method of using any one or a combination of primers of Table 1 or Table 2, to amplify RNA from a blood sample, and more particularly to identify antibodies, T-cell receptors, and HLA molecules within a population of cells.
Arm-PCR or tem-PCR may be used to amplify genes coding for the immunoglobulin superfamily molecules in am amplification method described previously by the inventor (Han et al, 2006, Simultaneous Amplification and Identification of 25 Human Papillomavirus Types with Templex Technology, J. Clin. Micro. 44(11), 4157-4162). In a tem-PCR reaction, nested gene-specific primers are designed to enrich the targets during initial PCR cycling. Later universal “Super” primers are used to amplify all targets. Primers are designated as Fo (forward out), Fi (forward in), Ri (reverse in), Ro (reverse out), FS (forward super primer) and RS (reverse super primer), with super primers being common to a variety of the molecules due to the addition of a binding site for those primers at the end of a target-specific primer. The gene-specific primers (Fo, Fi, Ri and Ro) are used at extremely low concentrations. Different primers are involved in the tem-PCR process at each of the three major stages. First, at the “enrichment” stage, low-concentration gene-specific primers are given enough time to find the templates. For each intended target, depending on which primers are used, four possible products may be generated Fo/Ro, F/Ro, F/Ri, and Fo/Ri. The enrichment stage is typically carried out for 10 cycles. In the second, or “tagging” stage, the annealing temperature is raised to 72° C., and only the long 40-nucleotide inside primers (Fi and Ri) will work. After 10 cycles of this tagging stage, all PCR products are “tagged” with the universal super primer sequences. Then, at the third “amplification” stage, high-concentration super primers work efficiently to amplify all targets and label the PCR products with biotin during the process. Specific probes may be covalently linked with Luminex color-mated beads.
To amplify the genes coding for immunoglobulin superfamily molecules, the inventor designed nested primers based on sequence information in the public domain. For studying B and T cell VDJ rearrangement, the inventor designed primers to amplify rearranged and expressed RNAs. Generally, a pair of nested forward primers is designed from the V genes and a set of reverse nested primers are designed from the J or C genes. The average amplicon size is 250-350 bp. For the igHV genes, for example, there are 123 genes that can be classified into 7 different families, and the present primers are designed to be family specific. However, if sequencing the amplified cDNA sequences, there are enough sequence diversities to allow further differentiation among the gene within the same family. For the MHC gene locus, the intent is to amplify genomic DNA.
Assuming that S is a subject's immunoprofile (IP), which is represented by N unique CDR3 sequences CDR31, CDR32, . . . CDR3n, each CDR3 has its own frequency s1, s2, . . . sn.
D is a disease library, which is the sum of a certain number of patients' immunoprofile with M unique CDR3s. All patients in the disease library were diagnosed to have the same disease.
P is a public library, which is the sum of a large number of control's immunoprofile.
The Sharing Index is defined as the sum of sx, sy, . . . xz, where CDR3x, CDR3y, . . . CDR3z are shared in the subject's immunoprofile and a library. Note that sx, sy, . . . sz is the frequency of CDR3s in the subject's immunoprofile, not in the library.
Assuming that there are always more unique CDR3s in a public library (P) than in a disease library (D), M unique CDR3s in the public library are randomly selected and used to create a sub-library P1 and the sharing index (SIp1) between the subject and the sub-library computed according to above formula. The sampling procedure is repeated 1000 or more times and 1000 or more SIpx are computed.
The sharing index SId between the subject and the disease library are computed in the same manner. The P-value is defined as the fraction of all SIs (SIp1, SIp2, . . . SIpx, SId. (Note that SId is included), which is equal to or greater than SId. Note that when sampling CDR3s in the public library, CDR3s found in x control's immunoprofiles are given x times of chances to be sampled.
All oligos were resuspended using 1× TE. All oligos except 454A and 454B were resuspended to a concentration of 100 pmol/μL. 454A and 454B were resuspended to a concentration of 1000 pmol/μL. 454A and 454B are functionally the same as the communal primers described previously, the different sequences were used for follow up high throughput sequencing procedures.
Three different primer mixes were made. An Alpha Delta primer mix included 82 primers (all of TRAV-C+TRDV-C), a Beta Gamma primer mix included 79 primers (all of TRBVC and TRGV-C) and a B cell primer mix that included a total of 70 primers. Fo, Fi, and Ri primers were at a concentration of 1 pmol/μL. Ro primers were at a concentration of 5 pmol/μL. 454A and 454B were at a concentration of 30 pmol/μL.
Three different RNA samples were ordered from ALLCELLS (www.allcells.com). All samples were diluted down to a final concentration of 4 ng/uL. The samples ordered were:
RT-PCR was performed using a Qiagen One-Step RT-PCR kit. Each sample contained the following:
10 μL of Qiagen Buffer
2 μL of DNTP's
2 μL of Enzyme
23.5 μL of dH2O
10 μL of the appropriate primer mix
2.5 μL of the appropriate template (10 ng of RNA total)
The samples were run using the following cycling conditions:
15 cycles of
6 cycles of
30 cycles of
The order of samples placed in the gel shown in
The initial experiment showed that a smear is generated from PCR reactions where templates were included. The smears indicate different sizes of PCR products were generated that represented a mixture of different VDJ rearrangements. There is some background amplification from the B cell reaction. Further improvement on that primer mix was required to clean up the reaction.
To determine whether the PCR products indeed include different VDJ rearrangements, it was necessary to isolate and sequence the single clones. Instead of using the routine cloning procedures, the inventor used a different strategy. PCR products generated from the Alpha Delta mix and the Beta Gamma mix (lanes 2 and 3 in
30 cycles of
A Qiagen PCR kit was used to amplify the products. The Master Mix used for the PCR contained the following:
The photograph of the gel in
As illustrated by
To investigate the impact of artifacts on the overall repertoire analysis of the TCRβ transcriptome, the inventors conducted control experiments using chemically synthesized TCRβ CDR3 templates. For this, the inventors chemically synthesized four distinct clones, clonally purified each clone, and prepared different mixes of the four constucts as templates for amplicon rescue multiplex (ARM)-PCR. Two different reaction mixtures were subjected to two independent ARM-PCR reactions, and the pooled PCR products were sequenced at a length of 100 bp from both ends using the Illimuna HiSeq2000®. The inventors first joined together paired-end reads through overlapping alignment with a modified Needleman-Wunsch algorithm, and then mapped the merged sequences to germline V, D and J reference sequences.
Without cleaning, the inventors obtained a total of 5,729,613 sequences from template mix I that could be mapped to TCRβ V, D and J segments. Surprisingly, the sequence reads purportedly represented a total of 36,439 unique CDR3 variants. Therefore, given that only four distinct CDR3 variants were present in the template mixtures, virtually all of the identified CDR3 variants must be non-authentic. Similar results were obtained for the second template mix, in which a total of 9,131,681 VDJ-mapped sequences were identified that mimicked the existence of 50,354 unique TCRβ CDR3 variants. The inventors' independent sequencing experiments show that only a few distinct CDR3 template variants can create artifactual repertoire diversities that far outweigh the real template diversity, and thus the inventors set out to eliminate these artifacts.
The quality of 3′ end Illumina sequencing reads is generally considered to be low. In the context of repertoire sequencing, this is troublesome because PCR primers need to be positioned distal enough from the hypervariable V(D)J junctions to avoid negative effects due to primer-template mismatching. As a consequence, the CDR3 segments of interest are generally “shifted” closer to the 3′ end of the sequencing reads, the region with increased sequencing error rates. Another technical issue that deserves attention is the observation that sequencing errors are context-specific end consequently strand-specific. Therefore, it is realistic to assume that the probability that a sequencing error a forward read coincides with that in the corresponding reverse read is rare.
Considering this, the inventors devised a paired-end strategy that affords double-strand sequencing of complete TCR CDR3 segments on the basis of the Illumina® technology. In this approach, forward and reverse sequencing primers are positioned at the framework region 3 and at the TCR J region or the 5′ end of the C region, respectively. Taking into account the average length of Illumina sequence reads (currently 100-150 bp) this design enables the complete sequencing of both strands that define a CDR3 segment. In a second step, the forward and reverse reads are then analyzed for sequence mismatches and CDR3 sequences that exhibit non-identity of both strands are eliminated using a newly developed paired-end filtering algorithm.
Applying this sequencing error filter to the 5,729,613 CDR3 sequences obtained for template mix I, the inventors identified a total of 2,751,131 (48%) CDR3 sequences that contained conflicting sequence information on their opposite strands. Discarding of these sequences resulted in the elimination of 35,455 (97.2%) distinct artifactual CDR3 variants. Consistent with this, the paired-end filter removed 4,308,020 (47%) CDR3 sequences from template mix II, leading to the elimination of 49,063 (97.4%) artifectual CDR3 variants. A total of 973 and 1271 unique CDR3 variants, respectively, passed through the filter. These results indicate that paired-end sequencing and filtering reduces the total number of non-authentic unique CDR3 sequences by almost two orders of magnitude.
Detailed analysis of the frequency distribution of the non-authentic CDR3 variants after the sequencing error filter revealed that in both mixtures approximately 50% of all artifacts were single-copy sequences. About 10% of these artifactual CDR3s displayed >100 copy numbers and accounted for >80% of all artifactual CDR3 variants. Given that variable TCR genes do not undergo somatic hypermutation, the inventors developed a reference algorithm that identifies and removes CDR3 sequence reads that display nucleotide mismatches relative to the mapped germline V, D and J reference sequences, as these must be artifacts generated at the level of PCR amplification or sequencing.
Applying this filtering algorithm to the “paired-end filtered” sequences of template mix I, a total of 29,804 sequences, which corresponded to 609 unique CDR3 variants, were removed. For template mix II, 54,516 artifactual sequences (831 unique CDR3 variants) were identified. Thus, the use of the reference sequence filter leads to a 60% reduction of non-authentic distinct CDR3 sequences. The reference filter is ineffective at the V-J and D-J junctions because the randomly added nucleotides in these regions during somatic recombination cannot be mapped. Therefore, the inventors implemented a PCR filter after computational simulation experiments to better understand four variables: the impact of the initial template number, the replication efficiency of each cycle, the cycle number (n), and the DNA polymerase error rate (μ) on the total end-point error rate. In contrast, the inventors noted that the PCR polymerase error rate has a pronounced effect on the number of accumulated errors
In the inventors' control sequencing experiments, PCR amplification was performed with 15 cycles and 45 cycles in the first and second reaction, using Taq polymerase. To simulate error accumulation during the ARM-PCR reactions more realistically, the PCR efficiency was set to decreased 5% per cycle for the first 25 cycles and 10% per cycle for the remaining cycles. The PCR efficiency was reset to 1.0 for each fresh PCR reaction. Furthermore, the inventors allowed mutation at the second position. Published substitution error rates for Taq enzyme, expressed as errors per bp per cycle, range from 0.023×10−4 to 2.1×10−4. In the simulation experiments, the substitution error rate was set at 2.7×10−5, and the insertion-deletion (indel) error rate was set as 1.0×10−6. Taq polymerase is known to have a much higher insertion-and-deletion (indel) mutation rate in homopolymeric region of templates. For a homopolymeric region, indel mutation in any position of this region generates identical pattern. Therefore, the indel error rate in a homopolymeric region was set n×μ, where n is the length of the homopolymeric region and μ is 1.0×10−6.
Because the impact of the initial template number and the PCR efficiency on the endpoint error rate is small, it should be safe to apply the same end-point error rate estimated from the simulation experiments to molecules with different initial number and different replication efficiencies in a multiplex PCR reaction. The cutoff error rates (μ) were empirically set as error rates at the 9999th 10000-quantiles point for each category. For two similar CDR3 sequences, A and B, of frequency NA and NB (NA>>NB) that differ in less than three positions, if NA*μ≧NB, where μ is the corresponding cutoff error rate, CDR3 sequence B will be excluded. Applying this filtering algorithm to the “reference filtered” sequences of template mix I, a total of 22,369 sequences, which corresponded to 281 unique CDR3 variants, were removed. For template mix II, 39,920 artifactual sequences (348 unique CDR3 variants) were identified (Table 1). Thus, the use of the PCR amplification error filter leads to a further reduction of non-authentic distinct CDR3 sequences by around 80%.
In the pool of sequences that had passed through the above filters, the inventors identified several high-abundance CDR3 variants, which differed from their most similar input template sequences at multiple positions. Because the occurrence of PCR substitution and/or indel mutation at multiple positions of CDR3 fragments is extremely rare according to simulation experiments, those CDR3 variants must arise from other source of artifacts. Intriguingly, the inventors noted that some of these sequences were composed of the fragments of two distinct input templates and exhibited clear breakpoints, which identified them as chimeras. Chimeric sequences are PCR artifacts that arise from incomplete primer extension or template switching during PCR and form mosaic-like structures. In light of this unexpected PCR artifact, the inventors developed a computational “mosaic filter.” Using this filtering algorithm, the inventors identified a total of 17 and 15 chimeric sequences in template mixtures I and II respectively. Of note, some of these CDR3 chimeras displayed sequence copy numbers >1000, indicating that the inventors algorithm for the filter is capable of identifying high-abundance chimeric CDR3 sequences.
Application of the filtering algorithms resulted in the elimination of 99.8% of the non-authentic unique CDR3 sequences generated by high-throughput sequencing of only four defined TCR CDR3 templates. Only 62 and 73 artifactual CDR3 sequences, respectively, passed through all filters. Among these, the two most abundant CDR3 sequences were identical in both mixing experiments. Most likely they represent chimeric artifacts which escaped filtering because of a single nucleotide substitution located exactly at the breakpoint. Among the remaining erroneous CDR3, 85% (n=53) and 75% (n=55) were single reads, respectively. To eliminate this minor fraction of artifacts, the inventors propose that high-stringency data analysis of TCR immune repertoires should include an additional filter that removes single copy CDR3 reads (frequency threshold filter).
This application is a continuation of and claims priority to U.S. Provisional Application No. 61/763,341, entitled “Method for Evaluating an Immunorepertoire” and filed on Feb. 11, 2013, which is incorporated herein by reference.
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/US14/15841 | 2/11/2014 | WO | 00 |
| Number | Date | Country | |
|---|---|---|---|
| 61763451 | Feb 2013 | US |
