A method of identifying variants in gene products from gene constructs used for cell therapy applications

Abstract

A method of ensuring that a gene product used for cell therapy is free of the risk of reduced efficacy or toxicity due to production of an undesirable variant. The method includes performing a computer simulation analysis on a gene construct to identify and alter sequences that can cause variants. In addition, the method includes performing an in vivo analysis consisting of RNA sequences based on the product of the construct. Variant detection can then be performed based on gap reads from the RNA sequences to determine variant expression levels, variant significance. If the identified variants are not acceptable, the method can include repeating the computer simulation analysis.

Description

[0001] Cross-reference to related applications

[0002] This application claims priority to U.S. Provisional Patent Application No. 63 / 217,933, filed July 2, 2021, the entire contents of which are incorporated herein by reference. Technical Field

[0003] This disclosure relates to methods for detecting and replacing sequences that can cause undesirable variants in gene constructs. Background Technology

[0004] In recent years, advancements in medical technology have led to the emerging use of immunotherapy in treating various types of ailments and diseases, including all forms of cancer. Typically, immunotherapy treats disease by stimulating or suppressing the immune response. Often, modified forms of the patient's own biological materials, such as immune cells, are reintroduced into the patient to trigger and / or supplement the immune response.

[0005] For example, engineered immune cells have demonstrated promising properties in therapeutic applications, particularly in oncology. Two main types of engineered immune cells are those containing chimeric antigen receptors (called “CAR” or “CAR-T”) and T-cell receptors (“TCR”). These engineered cells are engineered to confer antigen specificity while retaining or enhancing their ability to identify and kill target cells. For example, a chimeric antigen receptor may contain (i) an antigen-specific component (“antigen-binding molecule”), (ii) an extracellular domain, (iii) one or more co-stimulatory domains, and (iv) one or more activation domains. Each domain can be heterogeneous, i.e., composed of sequences derived from (or corresponding to) different protein chains.

[0006] Introducing genetic elements into cells using gene constructs (such as viral vectors) is a method for generating cells for applications such as cell therapy. The generation of gene constructs and the transduction of cells require multiple biological steps, which can potentially introduce heterogeneity into the product. Due to the specific requirements of viral vectors, defects in viral packaging, transduction, or transgenic transcription can introduce undesirable contaminants that differ from the intended sequence. These "variants" can express unexpected protein sequences and, depending on their frequency and nature, can cause side effects such as reduced efficacy, impaired production, or even increased toxicity. To prevent potential procedural failures, it is necessary to reduce the likelihood of variant generation during the development of transgenic technologies and cell therapies.

[0007] There is a need for a systematic method to detect and identify variants and potential variant-causing sequences in gene products for cell therapy applications. Furthermore, there is a need for improved methods for risk reduction strategies, such as, in the case of variant identification, employing variant characterization and / or DNA sequence modification during iterative processes to remove detected variants. Summary of the Invention

[0008] In summary, this disclosure generally relates to a system and method for preparing gene products for cell therapy. In one embodiment, the method includes performing a computer simulation analysis on a gene construct to identify and alter sequences that cause variants. Furthermore, the method includes performing an in vivo analysis comprising a first RNA sequencing step to identify a percentage frequency of variants, and repeating the computer simulation analysis if the frequency of variants identified in the first RNA sequencing step is greater than 5%. The in vivo analysis may include a second RNA sequencing step to identify a percentage frequency of variants, and repeating the computer simulation analysis if the variant is unacceptable. The method may also include performing variant detection based on gap readings from at least one of the first and second RNA sequencing to determine variant expression levels and variant significance. If the variant is determined to be unacceptable, the computer simulation analysis may be repeated to produce a new gene construct.

[0009] In one embodiment of the disclosed method, the computer simulation analysis may include modifying the variant in the gene construct with synonymous codon substitutions. Furthermore, the computer simulation analysis may include identifying and removing homologous sequences from the gene construct. In one embodiment, the computer simulation analysis includes identifying identical sequences in the gene construct. Additionally, the computer simulation analysis may include computing a matrix of segmented combinations from the input sequence and obtaining the Hamming distance for each combination. One embodiment of the method includes replacing random synonymous codons, provided that the replacement increases the sum of the matrix.

[0010] In one embodiment, the in vivo analysis includes RNA sequencing of the product prepared from the gene construct. In another embodiment, the RNA sequencing is performed in multiple stages to identify high-frequency variants and then low-frequency variants. Furthermore, the in vivo analysis may include performing an analysis to determine whether the lower-frequency variants should be replaced.

[0011] In one embodiment of the method, variant detection includes extracting RNA from a donor sample. Furthermore, when performing gap-sensing alignment, three separate aligners may be used in one embodiment. The p-value for variant significance is calculated using the Wilcox rank-sum test.

[0012] This disclosure also relates to a method for ensuring that gene products used in cell therapy do not pose a risk of reduced efficacy or toxicity due to the generation of undesirable variants. The method includes performing a computer simulation analysis on the gene construct to identify and alter sequences that may cause variants. Furthermore, the method includes performing an in vivo analysis consisting of an RNA sequence based on the product of the construct. Variant detection can then be performed based on gap readings from this RNA sequence to determine variant expression levels and variant significance. If the identified variant is unacceptable, the method may include repeating the computer simulation analysis.

[0013] One embodiment of this disclosure relates to a method for detecting and replacing sequences capable of causing undesirable variants in a gene construct. Such a method includes: performing a computer simulation analysis on the gene construct to detect the presence of the sequence capable of causing the undesirable variant; replacing the detected sequence capable of causing the undesirable variant with a substitution sequence, wherein the substitution sequence is a derived sequence including synonymous codon substitutions; measuring the frequency percentage of the undesirable variant expressed by the gene construct, the measurement including performing an in vivo analysis on one or more genes expressed by the gene construct, the in vivo analysis including performing RNA sequencing analysis on RNA products transcribed from the gene construct, wherein the frequency percentage of the undesirable variant is determined at least in part by using a splice-sensing aligner from the RNA sequencing analysis; and repeating the computer simulation analysis and replacement steps if the frequency percentage of the undesirable variant in the gene product from the in vivo analysis is greater than a predetermined value for an acceptable frequency percentage of the undesirable variant.

[0014] One embodiment of this disclosure relates to a method for preparing a gene product for cell therapy. Such a method includes the steps of: performing computer simulation analysis on a gene construct encoding the gene product to identify and alter sequences capable of causing undesirable variants; replacing the detected sequences capable of causing the undesirable variants with a substitution sequence, wherein the substitution sequence is a derived sequence including synonymous codon substitutions; measuring the percentage frequency of the undesirable variant expressed by the gene construct, the measurement including performing an in vivo analysis on one or more genes expressed by the gene construct, the in vivo analysis including performing RNA sequencing analysis on an RNA product transcribed from the gene construct, wherein the undesirable variant is determined at least in part by using a splice-sensing aligner from the RNA sequencing analysis. The percentage of the frequency of the variant; if the percentage of the frequency of the undesired variant in the gene product from the in vivo analysis is greater than a predetermined value of the acceptable percentage of the frequency of the undesired variant, the computer simulation and replacement steps are repeated to generate a new gene construct; and the percentage of the frequency of the undesired variant expressed by the new gene construct is measured, the measurement comprising performing an in vivo analysis on one or more genes expressed by the new gene construct, the in vivo analysis comprising performing RNA sequencing analysis on the RNA product transcribed from the new gene construct, wherein the percentage of the frequency of the undesired variant is determined at least in part by using a splice-sensing aligner from the RNA sequencing analysis.

[0015] Other aspects and advantages of this disclosure will become apparent from the following detailed description taken in conjunction with the accompanying drawings, which illustrate the principles of this disclosure by way of example only. Attached Figure Description

[0016] The teachings protected by and / or described herein are further described according to exemplary embodiments. These exemplary embodiments are described in detail with reference to the accompanying drawings. These embodiments are non-limiting exemplary embodiments, wherein the same reference numerals denote the same structures in various views of the drawings, and wherein:

[0017] Figure 1 illustrates an overview of one possibility for generating variants during cell therapy manufacturing.

[0018] Figure 2 An implementation scheme for the variant prediction, detection, and elimination process is described.

[0019] Figure 3 An example of a sequence divergence matrix that can be used in the disclosed process is depicted.

[0020] Figure 4 An exemplary process for a duplicate remover tool that can be used in the disclosed process is described.

[0021] Figure 5 An example screenshot showing a variant report template that can be used in the published process is depicted.

[0022] Figure 6 An example screenshot is shown illustrating the output from the duplicate finder and observer tools that identify identical sequences.

[0023] Figure 7A This is a flowchart depicting an exemplary process according to one embodiment of the present disclosure for detecting and replacing sequences that can cause undesirable variants in a gene construct.

[0024] Figure 7B This is a flowchart depicting an exemplary process for generating a gene therapy product for cell therapy according to one embodiment of the present disclosure.

[0025] Figures 8A to 8C Example code snippets are depicted that can be used to practice the disclosed methods. Detailed Implementation

[0026] This disclosure addresses the need for improved systems and methods for identifying variants in gene constructs and then selecting gene constructs for cell therapy. The following disclosure describes a systematic method for detecting and identifying variants and potential variant-causing sequences in gene products for cell therapy applications. In the case of variant identification, risk reduction strategies, such as variant characterization and / or DNA sequence modification, can be employed during the iterative process to remove detected variants.

[0027] It should be understood that the description herein is exemplary and illustrative only, and not a limitation on the technology protected by the claims. In this application, unless otherwise specified, the singular is used to include the plural meaning.

[0028] All documents or portions thereof (including, but not limited to, patents, patent applications, articles, books, and papers) cited in this application are expressly incorporated herein by reference in their entirety for any purpose. Unless otherwise specified, the following terms shall be understood to have the following meanings as used in accordance with this disclosure:

[0029] As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” include plural references unless the context clearly specifies otherwise.

[0030] Unless otherwise specified or obvious from the context, as used herein, the term “or” is understood to be inclusive and encompasses both “or” and “and”.

[0031] As used herein, the term “and / or” should be considered as a specific disclosure of each of two specified features or components, together or apart from the other. Thus, the term “and / or” as used herein in phrases such as “A and / or B” is intended to include A and B; A or B; A (alone); and B (alone). Similarly, the term “and / or” as used in phrases such as “A, B, and / or C” is intended to cover each of the following: A, B, and C; A, B, or C; A or C; A or B; B or C; A and C; A and B; B and C; A (alone); B (alone); and C (alone).

[0032] As used herein, the terms “for example” and “that is” are used only by way of example and are not intended to be restrictive, and should not be construed as referring only to those items explicitly listed in the specification.

[0033] The terms "or more," "at least," "more than," etc., such as "at least one," are understood to include, but are not limited to, at least 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40. 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 12 9, 130, 131, 132, 133, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149 or 150, 200, 300, 400, 500, 600, 700, 800, 900, 1000, 2000, 3000, 4000, 5000 or more of the stated values. This also includes any larger numbers or fractions in between.

[0034] Conversely, the term "not exceeding" includes every value less than the stated value. For example, "not exceeding 100 nucleotides" includes 100, 99, 98, 97, 96, 95, 94, 93, 92, 91, 90, 89, 88, 87, 86, 85, 84, 83, 82, 81, 80, 79, 78, 77, 76, 75, 74, 73, 72, 71, 70, 69, 68, 67, 66, 65, 64, 63, 62, 61, 60, 59, 58, 57, 56, 55, 5 4, 53, 52, 51, 50, 49, 48, 47, 46, 45, 44, 43, 42, 41, 40, 39, 38, 37, 36, 35, 34, 33, 32, 31, 30, 29, 28, 27, 26, 25, 24, 23, 22, 21, 20, 19, 18, 17, 16, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, and 0 nucleotides. Also includes any smaller numbers or fractions in between.

[0035] The terms "multiple," "at least two," "two or more," "at least second," etc., are understood to include, but are not limited to, at least 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, etc. 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 8 9, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149 or 150, 200, 300, 400, 500, 600, 700, 800, 900, 1000, 2000, 3000, 4000, 5000 or more. Also includes any larger numbers or fractions in between.

[0036] Unless otherwise specified or apparent from the context, as used herein, the term “about” refers to a value or composition within an acceptable margin of error for a particular value or composition, as determined by one of ordinary skill in the art, which will depend in part on how the value or composition is measured or determined, i.e., the limitations of the measurement system. For example, according to practice in the art, “about” or “approximately” may mean within one or more standard deviations. “About” or “approximately” may mean a range of up to 10% (i.e., ±10%). Thus, “about” can be understood as being within 10%, 9%, 8%, 7%, 6%, 5%, 4%, 3%, 2%, 1%, 0.5%, 0.1%, 0.05%, 0.01%, or 0.001% greater or less than the stated value. For example, about 5 mg can include any amount between 4.5 mg and 5.5 mg. Furthermore, particularly for biological systems or processes, these terms may mean up to one order of magnitude or up to five times a certain value. When a particular value or composition is provided in this disclosure, unless otherwise stated, the meaning of “about” or “approximately” shall be assumed to be within the acceptable margin of error for that particular value or composition.

[0037] As stated herein, any concentration range, percentage range, ratio range, or integer range should be understood to include any integer value within the range, and, where appropriate, its fraction (such as one-tenth and one-hundredth of an integer), unless otherwise stated.

[0038] The units, prefixes, and symbols used in this article are provided in their SI-accepted form. Numerical ranges include the numbers that define the range.

[0039] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art relating to this disclosure. For example, Juo, “The Concise Dictionary of Biomedicine and Molecular Biology,” 2nd edition, 2001, CRC Press; “The Dictionary of Cell & Molecular Biology,” 5th edition, 2013, Academic Press; and “The Oxford Dictionary of Biochemistry and Molecular Biology,” edited by Cammack et al., 2nd edition, 2006, Oxford University Press, provide a general dictionary for many of the terms used in this disclosure.

[0040] "Administration" means the physical introduction of a pharmaceutical agent into a subject using any of the various methods and delivery systems known to those skilled in the art. Exemplary routes of administration for the formulations disclosed herein include intravenous, intramuscular, subcutaneous, intraperitoneal, spinal, or other parenteral administration routes (e.g., by injection or infusion). Exemplary routes of administration for the compositions disclosed herein include intravenous, intramuscular, subcutaneous, intraperitoneal, spinal, or other parenteral administration routes (e.g., by injection or infusion). As used herein, the phrase "parenteral administration" refers to a mode of administration other than enteral and local administration (typically by injection) and includes, but is not limited to, intravenous, intramuscular, intraarterial, intrathecal, intralymphatic, intralesional, intracapsular, intraorbital, intracardiac, intradermal, intraperitoneal, tracheal, subcutaneous, subepidermal, intra-articular, subcapsular, subarachnoid, intraspinal, epidural, and intrasternal injections and infusions, as well as in vivo electroporation. In some embodiments, the formulation is administered via a non-parenteral route (e.g., orally). Other non-parenteral routes include local, epidermal, or mucosal administration, such as intranasal, vaginal, rectal, sublingual, or topical. Administration may also be performed, for example, once, multiple times, and / or over one or more extended time periods. In one embodiment, CAR T-cell therapy is administered via an "infusion product" containing CAR T-cells.

[0041] The term "antibody" (Ab) includes, but is not limited to, glycoprotein immunoglobulins that specifically bind to antigens. Generally, an antibody may comprise at least two heavy (H) chains and two light (L) chains linked together by disulfide bonds, or its antigen-binding molecule. Each H chain contains a heavy chain variable region (abbreviated VH herein) and a heavy chain constant region. The heavy chain constant region contains three constant domains CH1, CH2, and CH3. Each light chain contains a light chain variable region (abbreviated VL herein) and a light chain constant region. The light chain constant region contains one constant domain CL. The VH and VL regions can be further subdivided into hypervariable regions called complementarity-determining regions (CDRs), interspersed with more conserved regions called framework regions (FRs). Each VH and VL contains three CDRs and four FRs, arranged from the amino terminus to the carboxyl terminus as follows: FR1, CDR1, FR2, CDR2, FR3, CDR3, and FR4. The variable regions of the heavy and light chains contain binding domains that interact with the antigen. The constant region of Ab can mediate the binding of immunoglobulins to host tissues or factors, including various cells of the immune system (e.g., effector cells) and the first component (C1q) of the classical complement system.

[0042] Antibodies can include, for example, monoclonal antibodies, recombinant antibodies, monospecific antibodies, multispecific antibodies (including bispecific antibodies), human antibodies, engineered antibodies, humanized antibodies, chimeric antibodies, immunoglobulins, synthetic antibodies, tetrameric antibodies comprising two heavy chain and two light chain molecules, antibody light chain monomers, antibody heavy chain monomers, antibody light chain dimers, antibody heavy chain dimers, antibody light chain-antibody heavy chain pairs, endosomes, antibody fusions (sometimes referred to herein as “antibody conjugates”), heteroconjugate antibodies, single-domain antibodies, monovalent antibodies, single-chain antibodies or single-chain Fv (scFv), camelified antibodies, affinity molecules, Fab fragments, F(ab')2 fragments, disulfide-linked Fv (sdFv), anti-idiotypic (anti-Id) antibodies (including, for example, anti-Id antibodies), microantibodies, domain antibodies, synthetic antibodies (sometimes referred to herein as “antibody mimics”), and antigen-binding fragments of any of the above antibodies. In some embodiments, the antibodies described herein refer to a population of polyclonal antibodies.

[0043] "Antigen-binding molecule," "antigen-binding moiety," or "antibody fragment" refers to any molecule containing an antigen-binding moiety (e.g., CDR) of an antibody derived from that molecule. An antigen-binding molecule may contain an antigen complementarity-determining region (CDR). Examples of antibody fragments include, but are not limited to, Fab, Fab', F(ab')2, and Fv fragments, dAb, linear antibodies, scFv antibodies, and multispecific antibodies formed from antigen-binding molecules. Peptide bodies (i.e., Fc fusion molecules containing peptide-binding domains) are another example of suitable antigen-binding molecules. In some embodiments, the antigen-binding molecule binds to antigens on tumor cells. In some embodiments, the antigen-binding molecule binds to antigens on cells involved in hyperproliferative diseases or viral or bacterial antigens. In some embodiments, the antigen-binding molecule binds to CD19. In other embodiments, the antigen-binding molecule is an antibody fragment that specifically binds to an antigen, including one or more of its complementarity-determining regions (CDRs). In other embodiments, the antigen-binding molecule is a single-chain variable region fragment (scFv). In some embodiments, the antigen-binding molecule contains or is composed of a high-affinity multimer.

[0044] An "antigen" is any molecule that elicits an immune response or can be bound by an antibody or antigen-binding molecule. An immune response may involve antibody production, or activation of specific immune-active cells, or both. Those skilled in the art will readily understand that any macromolecule (including virtually all proteins or peptides) can be used as an antigen. Antigens may be endogenously expressed, i.e., expressed from genomic DNA, or may be recombinantly expressed. Antigens may be specific to a particular tissue (such as cancer cells), or they may be widely expressed. Additionally, fragments of larger molecules can act as antigens. In some embodiments, the antigen is a tumor antigen.

[0045] The term "neutralization" refers to an antigen-binding molecule, scFv, antibody, or fragment thereof that binds to a ligand and prevents or reduces the biological function of that ligand. In some embodiments, the antigen-binding molecule, scFv, antibody, or fragment thereof directly blocks the binding site on the ligand, or indirectly alters the binding ability of the ligand (e.g., structural or energy changes within the ligand). In some embodiments, the antigen-binding molecule, scFv, antibody, or fragment thereof prevents the protein it binds to from performing its biological function.

[0046] The term "autologous" refers to any material derived from the same individual and subsequently reintroduced into that individual. For example, the engineered autologous cell therapy (eACT) described herein... TM The method involves collecting lymphocytes from a patient, then engineering them to express, for example, a CAR construct, and subsequently administering them back to the same patient.

[0047] The term "allogeneic" refers to any material derived from one individual and subsequently introduced into another individual of the same species, such as allogeneic T-cell transplantation.

[0048] The terms “transduction” and “transduced” refer to the process of introducing exogenous DNA into cells via a viral vector (see Jones et al., “Genetics: principles and analysis,” Boston: Jones & Bartlett Publ. (1998)). In some embodiments, the vector is a retroviral vector, a DNA vector, an RNA vector, an adenovirus vector, a baculovirus vector, an Epstein-Barr virus vector, a papillomavirus vector, a vaccinia virus vector, a herpes simplex virus vector, an adenovirus-associated vector, a lentiviral vector, or any combination thereof.

[0049] “Cancer” refers to a broad group of various diseases characterized by the uncontrolled growth of abnormal cells in the body. Unregulated cell division and growth lead to the formation of malignant tumors, which invade adjacent tissues and can also metastasize to distant parts of the body via the lymphatic system or bloodstream. “Cancer” or “cancer tissue” can include tumors. In this application, the term cancer is synonymous with malignant tumors. Examples of cancers that can be treated by the methods disclosed herein include, but are not limited to, cancers of the immune system, including lymphoma, leukemia, myeloma, and other white blood cell malignancies. In some implementations, the methods disclosed herein can be used to reduce the size of tumors derived from, for example, the following: bone cancer, pancreatic cancer, skin cancer, head and neck cancer, malignant melanoma of the skin or eye, uterine cancer, ovarian cancer, rectal cancer, anal cancer, gastric cancer, testicular cancer, uterine cancer, fallopian tube cancer, endometrial cancer, cervical cancer, vaginal cancer, vulvar cancer, [add other solid tumors] multiple myeloma, Hodgkin's disease, non-Hodgkin's lymphoma (NHL), primary mediastinal large B-cell lymphoma (PMBC), diffuse large B-cell lymphoma (DLBCL), follicular lymphoma (FL), transformed follicular lymphoma, splenic marginal zone lymphoma (SMZL), esophageal cancer, small bowel cancer, endocrine system cancers, thyroid cancer. Cancer, parathyroid carcinoma, adrenal carcinoma, soft tissue sarcoma, urethral cancer, penile cancer, chronic or acute leukemia, acute myeloid leukemia, chronic myeloid leukemia, acute lymphoblastic leukemia (ALL) (including non-T-cell ALL), chronic lymphocytic leukemia (CLL), childhood solid tumors, lymphocytic lymphoma, bladder cancer, kidney or ureter cancer, renal pelvis cancer, central nervous system (CNS) tumors, primary CNS lymphoma, tumor angiogenesis, spinal cord axis tumors, brainstem glioma, pituitary adenoma, Kaposi's sarcoma, epidermoid carcinoma, squamous cell carcinoma, T-cell lymphoma, environmentally induced cancers (including those caused by asbestos), other B-cell malignancies, and combinations of the aforementioned cancers. In some embodiments, the cancer is multiple myeloma. In some embodiments, the cancer is NHL. Certain cancers may respond to chemotherapy or radiation therapy, or the cancer may be refractory. Refractory cancers are those that are not suitable for surgical intervention and that initially do not respond to chemotherapy or radiation therapy, or that become unresponsive over time.

[0050] As used in this article, "antitumor effect" refers to biological effects that can manifest as a reduction in tumor volume, a reduction in the number of tumor cells, a reduction in tumor cell proliferation, a reduction in the number of metastases, an increase in overall or progression-free survival, an increase in life expectancy, or an improvement in various physiological symptoms associated with tumors. Antitumor effect can also refer to the prevention of tumor development, such as through vaccines.

[0051] As used herein, “cytokine” refers to a non-antibody protein released by a cell in response to contact with a specific antigen, wherein the cytokine interacts with a second cell to mediate the response in the second cell. As used herein, “cytokine” refers to a protein released by a population of cells that acts as an intercellular mediator to another cell. Cytokines can be expressed endogenously by cells or administered to a subject. Cytokines can be released by immune cells, including macrophages, B cells, T cells, and mast cells, to propagate an immune response. Cytokines can induce various responses in recipient cells. Cytokines can include homeostatic cytokines, chemokines, pro-inflammatory cytokines, effectors, and acute-phase proteins. For example, homeostatic cytokines, including interleukin (IL)7 and IL-15, promote immune cell survival and proliferation, and pro-inflammatory cytokines can promote inflammatory responses. Examples of homeostatic cytokines include, but are not limited to, IL-2, IL-4, IL-5, IL-7, IL-10, IL-12p40, IL-12p70, IL-15, and interferon (IFN)γ. Examples of pro-inflammatory cytokines include, but are not limited to, IL-1a, IL-1b, IL-6, IL-13, IL-17a, tumor necrosis factor (TNF)-α, TNF-β, fibroblast growth factor (FGF)2, granulocyte-macrophage colony-stimulating factor (GM-CSF), soluble intercellular adhesion molecule-1 (sICAM-1), soluble vascular cell adhesion molecule-1 (sVCAM-1), vascular endothelial growth factor (VEGF), VEGF-C, VEGF-D, and placental growth factor (PLGF). Examples of effectors include, but are not limited to, granzyme A, granzyme B, soluble Fas ligand (sFasL), and perforin. Examples of acute-phase proteins include, but are not limited to, C-reactive protein (CRP) and serum amyloid A (SAA).

[0052] Chemokines are cytokines that mediate cellular chemotaxis or directed movement. Examples of chemokines include, but are not limited to, IL-8, IL-16, eosinophil activation chemokine, eosinophil activation chemokine-3, macrophage-derived chemokines (MDC or CCL22), monocyte chemoattractant protein 1 (MCP-1 or CCL2), MCP-4, macrophage inflammatory protein 1α (MIP-1α, MIP-1a), MIP-1β (MIP-1b), γ-inducible protein 10 (IP-10), and thymus activation-regulated chemokine (TARC or CCL17).

[0053] As used herein, a "chimeric receptor" refers to an engineered surface-expressed molecule capable of recognizing a specific molecule. Chimeric antigen receptors (CARs) and engineered T-cell receptors (TCRs), which include binding domains capable of interacting with specific tumor antigens, allow T cells to target and kill cancer cells expressing specific tumor antigens. In one embodiment, T-cell therapy is based on T cells engineered to express a chimeric antigen receptor (CAR) or a T-cell receptor (TCR) comprising (i) an antigen-binding molecule, (ii) a co-stimulatory domain, and (iii) an activation domain. The co-stimulatory domain may comprise an extracellular domain, a transmembrane domain, and an intracellular domain, wherein the extracellular domain comprises a truncated hinge domain.

[0054] The “therapeutic effective amount,” “effective dose,” “effective amount,” or “therapeutic effective dose” of a therapeutic agent (e.g., engineered CAR T cells) is any amount that, when used alone or in combination with another therapeutic agent, protects a subject from disease onset or promotes disease remission (demonstrated by a reduction in the severity of disease symptoms, an increase in the frequency and duration of asymptomatic disease periods, or prevention of impairment or disability caused by the disease). These terms are used interchangeably. The ability of a therapeutic agent to promote disease remission can be assessed using a variety of methods known to those skilled in the art, such as in human subjects during clinical trials, in animal model systems predicting efficacy in humans, or by measuring the activity of the agent in in vitro assays.

[0055] As used in this article, the term "lymphocyte" includes natural killer (NK) cells, T cells, or B cells. NK cells are cytotoxic (toxic to cells) lymphocytes that represent a major component of the innate immune system. NK cells repel tumors and virus-infected cells. They function through apoptosis, or programmed cell death. They are called "natural killers" because they can kill cells without activation. T cells play a major role in cell-mediated immunity (without antibody involvement). Their T cell receptors (TCRs) distinguish them from other lymphocyte types. The thymus is a specialized organ of the immune system and is primarily responsible for the maturation of T cells. There are six types of T cells: helper T cells (e.g., CD4+ cells), cytotoxic T cells (also known as TC, cytotoxic T lymphocytes, CTL, T killer cells, cytolytic T cells, CD8+ T cells, or cytotoxic T cells), memory T cells (i.e., stem memory TCM cells, such as immature cells), are CD45RO-, CCR7+, CD45RA+, CD62L+ (L-selectin), CD27+, CD28+, and IL-7Rα+, but they also express large amounts of CD95, IL-2Rβ, and C. XCR3 and LFA-1, and exhibit many functional properties specific to memory cells; (ii) central memory TCM cells express L-selectin and CCR7, and they secrete IL-2 but not IFNγ or IL-4; and (iii) however, effector memory TEM cells do not express L-selectin or CCR7, but produce effector cytokines (such as IFNγ and IL-4), regulatory T cells (Treg, suppressor T cells, or CD4+CD25+ regulatory T cells), natural killer T cells (NKT), and γδ T cells. On the other hand, B cells play a major role in humoral immunity (with antibody involvement). They produce antibodies and antigens and act as antigen-presenting cells (APCs), and are activated through antigen-antigen interactions to become memory B cells. In mammals, immature B cells are formed in the bone marrow, from which their name originates.

[0056] The terms "genetically engineered" or "engineered" refer to methods of modifying the genome of a cell, including but not limited to deleting coding or non-coding regions or portions thereof, or inserting coding regions or portions thereof. In some embodiments, the modified cells are lymphocytes, such as T cells, which may be obtained from a patient or donor. Cells may be modified to express exogenous constructs incorporated into the cell's genome, such as chimeric antigen receptors (CARs) or T-cell receptors (TCRs).

[0057] "Immune response" refers to the action of cells of the immune system (e.g., T lymphocytes, B lymphocytes, natural killer (NK) cells, macrophages, eosinophils, mast cells, dendritic cells, and neutrophils) and any of these cells or soluble macromolecules (including antibodies, cytokines, and complement) produced by the liver, resulting in the selective targeting, binding, damage, destruction, and / or clearance of invading pathogens from the vertebrate body, pathogen-infected cells or tissues, cancer cells or other abnormal cells, or normal human cells or tissues in cases of autoimmunity or pathological inflammation.

[0058] The term "immunotherapy" refers to the treatment of a subject who has a disease or is at risk of developing or relapsing into a disease by means of methods including inducing, enhancing, suppressing, or otherwise altering an immune response. Examples of immunotherapy include, but are not limited to, T-cell therapy. T-cell therapy may include adoptive T-cell therapy, tumor-infiltrating lymphocyte (TIL) immunotherapy, autologous cell therapy, and engineered autologous cell therapy (eACT). TM This includes allogeneic T-cell transplantation. However, those skilled in the art will recognize that the conditioning methods disclosed herein will enhance the efficacy of any transplanted T-cell therapy. Examples of T-cell therapies are described in U.S. Patent Publications 2014 / 0154228 and 2002 / 0006409, U.S. Patent 7,741,465, U.S. Patent 6,319,494, U.S. Patent 5,728,388, and International Publication WO 2008 / 081035. In some embodiments, immunotherapy includes CAR-T cell therapy. In some embodiments, CAR T-cell therapy products are administered via infusion.

[0059] The T cells for immunotherapy can be derived from any source known in the art. For example, T cells can be differentiated in vitro from hematopoietic stem cell populations, or T cells can be obtained from a subject. T cells can be obtained from, for example, peripheral blood mononuclear cells (PBMCs), bone marrow, lymph node tissue, umbilical cord blood, thymus tissue, tissue from the site of infection, ascites, pleural effusion, spleen tissue, and tumors. Additionally, T cells can be derived from one or more T cell lines available in the art. Various techniques known to those skilled in the art (such as FICOLL) can also be used. TM T cells are obtained from blood units collected from a subject using apheresis (isolation and / or apheresis). Additional methods for isolating T cells for T cell therapy are disclosed in U.S. Patent Publication 2013 / 0287748, the entire contents of which are incorporated herein by reference.

[0060] The term "engineered autologous cell therapy" or "eACT" TM(Also known as adoptive cell transfer) is the process of collecting a patient's own T cells and subsequently genetically modifying these cells to recognize and target one or more antigens expressed on the surface of one or more specific tumor cells or malignant tumor cells. T cells can be engineered to express, for example, chimeric antigen receptors (CARs). CAR-positive (+) T cells are engineered to express extracellular single-chain variable region fragments (scFvs) specific to a particular tumor antigen, linked to an intracellular signaling motif containing at least one co-stimulatory domain and at least one activation domain. CAR scFvs can be engineered to target, for example, CD19, a transmembrane protein expressed by cells in the B-cell lineage (including all normal B cells and B-cell malignancies, including but not limited to diffuse large B-cell lymphoma (DLBCL) nonspecific type, primary mediastinal large B-cell lymphoma, high-grade B-cell lymphoma, and DLBCL, NHL, CLL, and non-T-cell ALL caused by follicular lymphoma). Exemplary CARs T-cell therapies and constructs are described in U.S. Patent Publications 2013 / 0287748, 2014 / 0227237, 2014 / 0099309, and 2014 / 0050708, the full text of which is incorporated herein by reference.

[0061] As used herein, "patient" includes anyone with cancer (e.g., lymphoma or leukemia). The terms "subject" and "patient" are used interchangeably in this document.

[0062] As used herein, the term "in vitro cells" refers to any cells cultured in vitro. Specifically, in vitro cells may include T cells. The term "in vivo" refers to cells within a patient's body.

[0063] The terms “peptide,” “polypeptide,” and “protein” are used interchangeably and refer to compounds composed of amino acid residues covalently linked by peptide bonds. A protein or peptide contains at least two amino acids, and there is no limit to the maximum number of amino acids that can constitute a protein or peptide sequence. A polypeptide includes any peptide or protein containing two or more amino acids linked together by peptide bonds. As used herein, the term refers to both short chains (which are also commonly referred to in the art, for example, peptides, oligopeptides, and oligomers) and long chains (which are commonly referred to in the art, for example, proteins, of which there are many types). “Polypeptide” includes, for example, biologically active fragments, substantially homologous polypeptides, oligopeptides, homodimers, heterodimers, variants of polypeptides, modified polypeptides, derivatives, analogs, fusion proteins, and so on. Polypeptides include native peptides, recombinant peptides, synthetic peptides, or combinations thereof.

[0064] As used herein, “stimulus” refers to a primary response induced by the binding of a stimulating molecule to its homologous ligand, wherein such binding mediates a signal transduction event. A “stimulating molecule” is a molecule on a T cell (e.g., the T cell receptor (TCR) / CD3 complex) that specifically binds to a homologous stimulating ligand presented on an antigen-presenting cell. A “stimulating ligand” is a ligand that, when presented on an antigen-presenting cell (e.g., APCs, dendritic cells, B cells, etc.), specifically binds to a stimulating molecule on a T cell, thereby mediating primary responses of the T cell (including, but not limited to, activation, initiation of an immune response, proliferation, etc.). Stimulating ligands include, but are not limited to, anti-CD3 antibodies, peptide-loaded MHC class I molecules, superagonist anti-CD2 antibodies, and superagonist anti-CD28 antibodies.

[0065] As used in this article, “co-stimulatory signal” refers to a signal that, in conjunction with primary signals such as TCR / CD3 linkage, elicits a T cell response (such as, but not limited to, proliferation and / or upregulation or downregulation of key molecules).

[0066] As used herein, “co-stimulatory ligands” include molecules on antigen-presenting cells that specifically bind to homologous co-stimulatory molecules on T cells. The binding of co-stimulatory ligands provides signals that mediate T cell responses, including but not limited to proliferation, activation, and differentiation. Co-stimulatory ligands induce signals in addition to the primary signals provided by stimulatory molecules, for example, through the binding of the T cell receptor (TCR) / CD3 complex to peptide-loaded major histocompatibility complex (MHC) molecules. Costimulatory ligands may include, but are not limited to, 3 / TR6, 4-1BB ligands, agonists or antibodies that bind to Toll ligand receptors, B7-1 (CD80), B7-2 (CD86), CD30 ligands, CD40, CD7, CD70, CD83, herpesvirus invasion mediators (HVEM), human leukocyte antigen G (HLA-G), ILT4, immunoglobulin-like transcript (ILT)3, inducible costimulatory ligands (ICOS-L), intercellular adhesion molecule (ICAM), ligands that specifically bind to B7-H3, lymphotoxin β receptors, MHC class I chain-associated protein A (MICA), MHC class I chain-associated protein B (MICB), OX40 ligands, PD-L2, or programmed death (PD)L1. In some implementations, the costimulatory ligands include, but are not limited to, antibodies that specifically bind to costimulatory molecules present on T cells, such as, but not limited to, 4-1BB, B7-H3, CD2, CD27, CD28, CD30, CD40, CD7, ICOS, ligands that specifically bind to CD83, lymphocyte function-associated antigen-1 (LFA-1), natural killer cell receptor C (NKG2C), OX40, PD-1, or tumor necrosis factor superfamily member 14 (TNFSF14 or LIGHT).

[0067] "Co-stimulatory molecules" are homologous binding partners on T cells. They specifically bind to co-stimulatory ligands, thereby mediating co-stimulatory responses in T cells, such as, but not limited to, proliferation. Co-stimulatory molecules include, but are not limited to, 4-1BB / CD137, B7-H3, BAFFR, BLAME (SLAMF8), BTLA, CD33, CD45, CD100 (SEMA4D), CD103, CD134, CD137, CD154, CD16, CD160 (BY55), CD18, CD19, CD19a, CD2, CD22, CD247, CD27, CD276 (B7-H3), and CD2... 8. CD29, CD3 (α; β; δ; ε; γ; ζ), CD30, CD37, CD4, CD4, CD40, CD49a, CD49D, CD49f, CD5, CD64, CD69, CD7, CD8 0. CD83 ligand, CD84, CD86, CD8α, CD8β, CD9, CD96 (Tactile), CD11a, CD11b, CD11c, CD11d, CDS, CEACAM1, CRT AM, DAP-10, DNAM1 (CD226), Fcγ receptor, GADS, GITR, HVEM (LIGHTR), IA4, ICAM-1, ICOS, Igα (CD79a), IL2Rβ, IL2Rγ, IL7Rα, integrin, ITGA4, ITGA6, ITGAD, ITGAE, ITGAL, ITGAM, ITGAX, ITGB2, ITGB7, ITG1, KIRDS2, LAT, LFA-1, LIGHT (tumor necrosis factor superfamily member 14; TNFSF14), LTBR, ​​Ly9 (CD229), lymphocyte function-associated antigen-1 (LFA-1 (CD11a / CD18), MHC) Class I molecules, NKG2C, NKG2D, NKp30, NKp44, NKp46, NKp80 (KLRF1), OX40, PAG / Cbp, PD-1, PSGL1, SELPLG (CD162), signaling lymphocyte activating molecules, SLAM (SLAMF1; CD150; IPO-3), SLAMF4 (CD244; 2B4), SLAMF6 (NTB-A; Lyl08), SLAMF7, SLP-76, TNF, TNFr, TNFR2, Toll ligand receptor, TRANCE / RANKL, VLA1 or VLA-6, or fragments, truncations, or combinations thereof.

[0068] The terms “reduction” and “lowering” are used interchangeably herein and indicate any change less than the original value. “Reduction” and “lowering” are relative terms requiring a comparison between before and after measurement. “Reduction” and “lowering” include complete exhaustion. Similarly, the term “increase” indicates any change greater than the original value. “Increase,” “higher,” and “lower” are relative terms requiring a comparison between before and after measurement and / or between reference standards. In some embodiments, the reference value is obtained from values ​​of a general population, which may be a general patient population. In some embodiments, the reference value is derived from an quartile analysis of the general patient population.

[0069] "Treatment" in a subject refers to any type of intervention or procedure performed on a subject, or the administration of an active agent to a subject, with the aim of reversing, alleviating, improving, suppressing, slowing, or preventing the onset, progression, development, severity, or recurrence of disease-related symptoms, complications, or conditions, or biochemical indicators. In some implementations, "treatment" includes partial remission. In another implementation, "treatment" or "treatment" includes complete remission.

[0070] As used herein, the term "multifunctional T cell" refers to a cell that co-secres at least two proteins from a pre-specified group of cells, coupled with the amount of each protein produced (i.e., the number of proteins secreted and the combination of their strengths). In some embodiments, a single-cell functional profile is determined for each evaluable population of engineered T cells. The profile can be categorized into effector groups (granulase B, IFN-γ, MIP-1α, perforin, TNF-α, TNF-β), stimulatory groups (GM-CSF, IL-2, IL-5, IL-7, IL-8, IL-9, IL-12, IL-15, IL-21), regulatory groups (IL-4, IL-10, IL-13, IL-22, TGF-β1, sCD137, sCD40L), chemoattractive groups (CCL-11, IP-10, MIP-1β, RANTES), and inflammatory groups (IL-1b, IL-6, IL-17A, IL-17F, MCP-1, MCP-4). In some implementations, the functional profile of each cell enables the calculation of other metrics, including the breakdown of each sample based on cell versatility (i.e., what percentage of cells are secreting multiple cytokines compared to non-secreting or monofunctional cells), and the breakdown of the sample by functional groups (i.e., which monofunctional and polyfunctional groups are being secreted by the cells in the sample, and their frequency).

[0071] As used in this article, the terms "quartiles" or "quadrants" are statistical terms that describe how observations are divided into four defined intervals based on data values ​​and how they are compared to the whole set of observations.

[0072] As used in this article, the term "Study Day 0" is defined as the day a subject receives their first CAR T-cell infusion. Days preceding Study Day 0 will be Study Day -1. Any days after enrollment and before Study Day -1 will be consecutive and negative integers.

[0073] As used herein, the term “objective response” refers to a complete response (CR), a partial response (PR), or no response. The criteria are based on the revised IWG response criteria for malignant lymphomas.

[0074] As used in this article, the term "complete response" refers to the complete regression of disease, which becomes undetectable by radiographic imaging and clinical laboratory assessment. There is no evidence of cancer at a given time.

[0075] As used herein, the term "partial response" refers to a reduction in tumor size of more than 30% but not complete regression. The criteria are based on the revised IWG response criteria for malignant lymphomas, where a partial response (PR) is defined as "a reduction of at least 50% in the sum of the diameter products (SPD) of up to six largest dominant nodes or node masses." These nodes or masses should be selected based on all of the following: they should be clearly measurable in at least two vertical dimensions; they should originate from different regions of the body, if possible; and whenever these sites are involved, they should include mediastinal and retroperitoneal disease areas.

[0076] As used in this article, the term "unresponsive" refers to subjects who have never experienced CR or PR after CAR T-cell infusion.

[0077] As used in this article, the term "durable response" refers to subjects who are in an ongoing response during at least one year of follow-up following CAR T-cell infusion. The 6-month f / u was only used for Z1 and C3 because no more f / u was available for that cohort. However, the conclusion remains the same.

[0078] As used in this article, the term “relapse” refers to a subject who achieved a complete response (CR) or a partial response (PR) and subsequently experienced disease progression.

[0079] As used in this article, the expansion and persistence of CAR T cells in peripheral blood can be monitored by qPCR analysis, for example, using CAR-specific primers targeting the scFv portion of the CAR (e.g., the heavy chain of the CD19 binding domain) and its hinge / CD28 transmembrane domain. Alternatively, it can be measured by calculating the number of CAR cells per unit blood volume.

[0080] As used in this article, scheduled blood draws for CAR T cells can be taken before CAR T cell infusion, on day 7, week 2 (day 14), week 4 (day 28), month 3 (day 90), month 6 (day 180), month 12 (day 360), and month 24 (day 720).

[0081] As used in this article, “peak CAR T cell count” is defined as the maximum absolute number of CAR+PBMCs / μL reached in serum after day 0.

[0082] As used in this article, “time to reach peak CAR T cells” is defined as the number of days from day 0 to the day when peak CAR T cells are reached.

[0083] As used in this article, the area under the curve (AUC) of CAR T cell levels from day 0 to day 28 is defined as the area under the curve in a graph relative to the planned CAR T cell levels from day 0 to day 28. This AUC measures the total level of CAR T cells over time.

[0084] As used in this article, scheduled blood draws for cytokines are taken before or on the same day as conditioning chemotherapy (day -5), day 0, day 1, day 3, day 5, day 7, every other day until hospitalization (if any), week 2 (day 14), and week 4 (day 28).

[0085] As used in this article, the “baseline” for cytokines is defined as the last value measured before conditioning chemotherapy.

[0086] As used in this article, the multiple change from baseline on day X is defined as...

[0087]

[0088] As used in this article, “post-baseline peak cytokine” is defined as the maximum level of serum cytokines reached from baseline (day -5) up to day 28.

[0089] As used in this article, the “time to reach peak cytokine levels” after CAR T cell infusion is defined as the number of days from day 0 to the day when peak cytokine levels are reached.

[0090] As used in this paper, the area under the curve (AUC) of cytokine levels from day -5 to day 28 is defined as the area under the curve of the cytokine level relative to the planned visit from day -5 to day 28. This AUC measures the total cytokine level over time. Given that cytokines and CAR+T cells are measured at certain discrete time points, the trapezoidal rule can be used to estimate the AUC.

[0091] As used herein, those skilled in the art will readily understand the term "negligible effect" and its boundaries and limits. As a non-limiting example, those skilled in the art will understand that a negligible effect can refer to one or more of the following: a statistically insignificant effect on the expression of the chimeric antigen receptor; a statistically insignificant effect on the therapeutic efficacy of the chimeric antigen receptor; a statistically insignificant effect on the toxicity of the chimeric antigen receptor to patients; or an effect on the expression and / or efficacy and / or toxicity of the chimeric antigen receptor, wherein such effect does not exceed a predetermined threshold for the expression and / or efficacy and / or toxicity of the chimeric antigen receptor.

[0092] It should be understood that chimeric antigen receptors (CARs or CAR-T) and T-cell receptors (TCRs) can be genetically engineered receptors. According to techniques known in the art, these engineered receptors can be readily inserted into and expressed by immune cells, including T cells. Using CARs, a single receptor can be programmed to both recognize a specific antigen and, upon binding to that antigen, activate immune cells to attack and destroy cells carrying that antigen. When these antigens are present on tumor cells, CAR-expressing immune cells can target and kill the tumor cells.

[0093] CARs can be engineered to bind to antigens (such as cell surface antigens) by incorporating antigen-binding molecules that interact with the target antigen. As used herein, “antigen-binding molecule” refers to any protein that binds to a specific target molecule. Antigen-binding molecules include, but are not limited to, antibodies and their binding moieties, such as immune functional fragments. Peptide bodies (i.e., Fc fusion molecules containing peptide-binding domains) are another example of suitable antigen-binding molecules.

[0094] Preferably, the target molecules may include, but are not limited to, hematologic cancer-associated antigens. Non-limiting examples of hematologic cancer-associated antigens include antigens associated with one or more cancers selected from the group consisting of: acute myeloid leukemia (AML), chronic myeloid leukemia (CML), chronic myelomonocytic leukemia (CMML), juvenile myelomonocytic leukemia, atypical chronic myeloid leukemia, acute promyelocytic leukemia (APL), acute monocytic leukemia, acute erythroblastic leukemia, acute megakaryocytic leukemia, lymphocytic leukemia, B-cell acute lymphoblastic leukemia, B-cell chronic lymphocytic leukemia, B-cell non-Hodgkin's lymphoma, myelodysplastic syndromes (MDS), myeloproliferative disorders, myeloid neoplasms, myeloid sarcomas, and blastic plasmacytoid dendritic cell tumors (BPDCN).

[0095] In some implementations, the antigen is selected from tumor-associated surface antigens, such as 5T4, alpha-fetoprotein (AFP), B7-1 (CD80), B7-2 (CD86), BCMA, beta-human chorionic gonadotropin, CA-125, carcinoembryonic antigen (CEA), CD123, CD133, CD138, CD19, CD20, CD22, CD23, CD24, CD25, CD30, CD33, CD34, CD4, CD40, CD44, CD56, CD8, CLL-1, c-Met, CMV-specific antigen, CSPG4, CTLA-4, disialotetrahexosylganglioside GD2, ductal-epithelial mucin, EBV-specific antigen, and EGFR variant III (EGFRvIII). ELF2M, endothelial glycoprotein, liver glycoprotein B2, epidermal growth factor receptor (EGFR), epithelial cell adhesion molecule (EpCAM), epithelial tumor antigen, ErbB2 (HER2 / neu), fibroblast-associated protein (fap), FLT3, folate-binding protein, GD2, GD3, glioma-associated antigen, glycosphingolipid, gp36, HBV-specific antigen, HCV-specific antigen, HER1-HER2, HER2-HER3 combination, HERV-K, high molecular weight melanoma-associated antigen (HMW-MAA), HIV-1 envelope glycoprotein gp41, HPV-specific antigen, human telomerase reverse transcriptase, IGFI receptor, IGF-II, IL-11Rα, IL-13R-a2, influenza virus-specific antigen;CD38, Insulin-Growth Factor (IGF1)-1, Intestinal Carboxyesterase, κ chain, LAGA-1a, λ chain, Lassa virus-specific antigen, Lectin-Reactive AFP, Lineage-Specific or Tissue-Specific Antigens such as CD3, MAGE, MAGE-A1, Major Histocompatibility Complex (MHC) molecule, Major Histocompatibility Complex (MHC) molecule presenting tumor-specific peptide epitopes, M-CSF, Melanoma-Associated Antigen, Mesothelin, MN-CAIX, MUC-1, mut hsp72, mutant p53, mutant p53, mutant ras, neutrophil elastase, NKG2D, Nkp30, NY-ESO-1, p53, PAP, prostate enzymes, prostate-specific antigen (PSA), prostate cancer tumor antigen-1 (PCTA-1), prostate-specific antigen, proteins, PSMA, RAGE-1, ROR1, RU1, RU2 (AS), surface adhesion molecules, survivin and telomerase, TAG-72, extra domain A (EDA) and extra domain B (EDB) of fibronectin and the A1 domain (TnC A1) of tendinin-C, thyroglobulin, tumor matrix antigen, vascular endothelial growth factor receptor-2 (VEGFR2), virus-specific surface antigens such as HIV-specific antigens (such as HIV gpl20), and any derivatives or variants of these surface markers.

[0096] In some embodiments, the target molecule may include viral infection-associated antigens. The viral infection described herein can be caused by any virus, including, for example, HIV. This list of possible target molecules is not intended to be exclusive.

[0097] The TCR disclosed herein can bind to, for example, tumor-associated antigens. As used herein, "tumor-associated antigen" refers to any antigen associated with one or more cancers selected from a group of components: adrenocortical tumor, anal cancer, bladder cancer, bone cancer, brain cancer, breast cancer, carcinoid tumor, carcinoma, cervical cancer, colon cancer, endometrial cancer, esophageal cancer, extrahepatic bile duct cancer, extracranial germ cell cancer, eye cancer, gallbladder cancer, gastric cancer, germ cell tumor, gestational trophoblastoma, head and neck cancer, hypopharyngeal cancer, pancreatic islet cell cancer, kidney cancer, colorectal cancer, laryngeal cancer, leukemia, lip and oral cancer, liver cancer, lung cancer, lymphoma, malignant mesothelioma, Merkel cell carcinoma, mushroom cancer. Mycosis fungal infection, myelodysplastic syndrome, myeloproliferative disorders, nasopharyngeal carcinoma, neuroblastoma, oral cancer, oropharyngeal cancer, osteosarcoma, ovarian epithelial cancer, ovarian germ cell cancer, pancreatic cancer, sinus and nasal cavity cancer, parathyroid cancer, penile cancer, pituitary cancer, plasmacytoma, prostate cancer, rhabdomyosarcoma, rectal cancer, renal cell carcinoma, transitional cell carcinoma of the renal pelvis and ureter, salivary gland cancer, Sezary syndrome, skin cancer, small intestine cancer, soft tissue sarcoma, gastric cancer, testicular cancer, thymoma, thyroid cancer, urethral cancer, uterine cancer, vaginal cancer, vulvar cancer, and Wilms' tumor.

[0098] In some embodiments, this disclosure is applicable to target molecules of blood cancers. In some embodiments, the cancer is a cancer of white blood cells. In other embodiments, the cancer is a cancer of plasma cells. In some embodiments, the cancer is leukemia, lymphoma, or myeloma. In some embodiments, the cancer is acute lymphoblastic leukemia (ALL) (including non-T-cell ALL), acute lymphoblastic leukemia (ALL) and hemophagocytic lymphohistiocytosis (HLH)), B-cell prolymphoblastic leukemia, B-cell acute lymphoblastic leukemia (“BALL”), blastic plasmacytoid dendritic cell tumor, Burkitt lymphoma, chronic lymphocytic leukemia (CLL), chronic myeloid leukemia (CML), or chronic myeloid leukemia (CML). Chronic or acute granulomatous diseases, chronic or acute leukemia, diffuse large B-cell lymphoma, diffuse large B-cell lymphoma (DLBCL), follicular lymphoma, follicular lymphoma (FL), hairy cell leukemia, hemophagocytic syndrome (macrophage activation syndrome (MAS), Hodgkin's disease, large cell granuloma, leukocyte adhesion defect, malignant lymphoproliferative disorders, MALT lymphoma, mantle cell lymphoma, marginal zone lymphoma, monoclonal globulinemia of undetermined significance (MGU) S), multiple myeloma, myelodysplastic syndromes (MDS), myeloid diseases (including but not limited to acute myeloid leukemia (AML)), non-Hodgkin lymphoma (NHL), plasmacytic proliferative disorders (e.g., asymptomatic myeloma (depressive multiple myeloma or indolent myeloma), plasmablastic lymphoma, plasmacytoid dendritic cell tumor, plasmacytoma (e.g., plasmacytic cachexia; solitary myeloma; solitary plasmacytoma; extramedullary plasmacytoma and multiple plasmacytoma), POEMS syndrome (Crow-Fukase syndrome; Takatsuki disease; PEP syndrome), primary mediastinal large B-cell lymphoma (PMBCL), small cell or large cell follicular lymphoma, splenic marginal zone lymphoma (SMZL), systemic amyloid light chain amyloidosis, T-cell acute lymphoblastic leukemia (“TALL”), T-cell lymphoma, transformed follicular lymphoma, Waldenstrom macroglobulinemia, or combinations thereof.

[0099] The TCR disclosed herein can also bind to viral infection-associated antigens. Virus infection-associated antigens include antigens associated with any viral infection, including, for example, viral infection caused by HIV.

[0100] The various implementation schemes are described in further detail in the following sections.

[0101] Variants generated

[0102] Cell therapy products (including autologous, allogeneic, neoantigen, and other types of products) have the potential to express RNA and protein sequences other than the desired or transfected genes. The expression of these non-standard sequences (called variants) can occur due to at least two different mechanisms.

[0103] The first mechanism is RNA splicing, which can lead to variant expression of the product even when it has been transfected with the intended gene sequence, because the product is transcribed into an RNA product, which is then spliced. If the product manufacturing process involves reverse transcription, RNA splicing can also lead to transfection of the product with a variant sequence (at the DNA level), because the RNA can be spliced ​​before it is reverse transcribed into DNA, which is then transfected into the cell.

[0104] A second mechanism occurs when reverse transcription is part of the manufacturing process: homologous recombination. In this phenomenon, the reverse transcriptase that actively transcribes "jumps" between two highly similar (or identical) sequences on the RNA template, skipping the insertion sequence and producing a non-standard DNA transcript that can then be transfected into the product. Because this mechanism relies on highly similar sequences in the template, it is particularly dangerous for bicistronic CAR products, where the two CARs may share some of the same domains (e.g., co-stimulatory domains).

[0105] Considering both mechanisms and focusing on standard CAR-T cell manufacturing procedures using lentiviral vectors, several points where variants can be generated are identified. First, as illustrated in the example of Figure 1, HEK293 cells can be transfected with the product and a lentiviral encoding plasmid to generate lentiviral vectors. The RNA produced by HEK293 cells can be spliced ​​before being packaged into these vectors. Second, the lentiviral vectors are used to transfect T cells. This involves reverse transcription of the product sequence before it is integrated into the T cell genome, during which homologous recombination can occur. Third, the transfected T cells (now CAR-T cells) express CAR, and the expressed CAR mRNA can be spliced ​​before it is translated into a protein.

[0106] Variants, regardless of their origin, are undesirable. Variants in cell therapy products may have little to no efficacy. Moreover, variants can cause cell therapy products to identify different antigens, thus leading to off-target toxicity. Below is a description of different implementations of a general pipeline or process for predicting, detecting, and eliminating variants to prevent these outcomes.

[0107] One embodiment of this disclosure relates to a method for detecting and replacing sequences capable of causing undesirable variants in a gene construct. Such a method includes: performing a computer simulation analysis on the gene construct to detect the presence of the sequence capable of causing the undesirable variant; replacing the detected sequence capable of causing the undesirable variant with a substitution sequence, wherein the substitution sequence is a derived sequence including synonymous codon substitutions; measuring the frequency percentage of the undesirable variant expressed by the gene construct, the measurement including performing an in vivo analysis on one or more genes expressed by the gene construct, the in vivo analysis including performing RNA sequencing analysis on RNA products transcribed from the gene construct, wherein the frequency percentage of the undesirable variant is determined at least in part by using a splice-sensing aligner from the RNA sequencing analysis; and repeating the computer simulation analysis and replacement steps if the frequency percentage of the undesirable variant in the gene product from the in vivo analysis is greater than a predetermined value for an acceptable frequency percentage of the undesirable variant.

[0108] One embodiment of this disclosure relates to the above method, wherein gap-sensing alignment includes using at least two separate aligners.

[0109] One embodiment of this disclosure relates to any of the above methods, wherein the computer simulation analysis further includes: detecting at least one of a plurality of homologous sequences and a plurality of identical sequences in the gene construct, wherein the plurality of homologous sequences and the plurality of identical sequences are capable of causing an undesirable variant in the gene construct; and replacing any such detected plurality of homologous sequences and plurality of identical sequences by performing a synonymous codon substitution step.

[0110] One embodiment of this disclosure relates to any of the above methods, wherein the computer simulation analysis further includes: calculating a matrix of segment combinations from the gene construct and obtaining the Hamming distance of each segment combination in the segment combinations.

[0111] One embodiment of this disclosure relates to any of the above methods, wherein the computer simulation analysis further includes: replacing multiple random synonymous codons in the gene construct with multiple substitution sequences, such that the multiple substitution sequences increase the sum of the array.

[0112] One embodiment of this disclosure relates to any of the methods described above, wherein the gene construct contains a sequence encoding a chimeric antigen receptor.

[0113] One embodiment of this disclosure relates to any of the methods described above, wherein a predetermined value of an acceptable frequency percentage of the undesired variant is determined based on whether the undesired variant is associated with at least one of the following: whether the undesired variant adversely affects the output of the chimeric antigen receptor to the cell surface; whether the undesired variant is associated with changes in the binding domain of the chimeric antigen receptor; and whether the undesired variant has previously been characterized as having a negligible effect on the expression or function of the chimeric antigen receptor.

[0114] One embodiment of this disclosure relates to any of the methods described above, wherein if the undesired variant adversely affects the output of the chimeric antigen receptor to the cell surface, the predetermined value of the acceptable frequency percentage of the undesired variant is 0.1%; and wherein if the undesired variant is associated with a change in the binding domain of the chimeric antigen receptor, the predetermined value of the acceptable frequency percentage of the undesired variant is 0.01%.

[0115] One embodiment of this disclosure relates to any of the methods described above, wherein the computer simulation analysis and replacement steps are not repeated if the undesired variant has previously been characterized as having a negligible effect on the expression or function of the chimeric antigen receptor.

[0116] One embodiment of this disclosure relates to any of the methods described above, the method further comprising the steps of: identifying and removing a subpopulation of high-frequency variants; and identifying a subpopulation of low-frequency variants; and wherein the in vivo analysis further comprises: performing an analysis to determine whether the subpopulation of the low-frequency variant should be replaced.

[0117] One embodiment of this disclosure relates to any of the methods described above, wherein if an unwanted variant is detected and none of the conditions or requirements described above are met, the computer simulation analysis and sequence removal steps are repeated to attempt to eliminate the unwanted variant, and / or the unwanted variant is further characterized in an additional study to assess the risk to potential patients.

[0118] One embodiment of this disclosure relates to a method for preparing a gene product for cell therapy. Such a method includes the steps of: performing computer simulation analysis on a gene construct encoding the gene product to identify and alter sequences capable of causing undesirable variants; replacing the detected sequences capable of causing the undesirable variants with a substitution sequence, wherein the substitution sequence is a derived sequence including synonymous codon substitutions; measuring the percentage frequency of the undesirable variant expressed by the gene construct, the measurement including performing an in vivo analysis on one or more genes expressed by the gene construct, the in vivo analysis including performing RNA sequencing analysis on an RNA product transcribed from the gene construct, wherein the undesirable variant is determined at least in part by using a splice-sensing aligner from the RNA sequencing analysis. The percentage of the frequency of the variant; if the percentage of the frequency of the undesired variant in the gene product from the in vivo analysis is greater than a predetermined value of the acceptable percentage of the frequency of the undesired variant, the computer simulation and replacement steps are repeated to generate a new gene construct; and the percentage of the frequency of the undesired variant expressed by the new gene construct is measured, the measurement comprising performing an in vivo analysis on one or more genes expressed by the new gene construct, the in vivo analysis comprising performing RNA sequencing analysis on the RNA product transcribed from the new gene construct, wherein the percentage of the frequency of the undesired variant is determined at least in part by using a splice-sensing aligner from the RNA sequencing analysis.

[0119] One embodiment of this disclosure relates to the above method, wherein gap-sensing alignment includes using at least two separate aligners.

[0120] One embodiment of this disclosure relates to any of the methods described above, wherein the computer simulation analysis further includes the steps of: detecting at least one of a plurality of homologous sequences and a plurality of identical sequences in the gene construct, wherein the plurality of homologous sequences and the plurality of identical sequences are capable of causing an undesirable variant in the gene construct; and replacing any such detected plurality of homologous sequences and plurality of identical sequences, the replacement including a synonymous codon substitution step.

[0121] One embodiment of this disclosure relates to any of the above methods, wherein the computer simulation analysis further includes: calculating a matrix of segment combinations from the gene construct and obtaining the Hamming distance of each segment combination in the segment combinations.

[0122] One embodiment of this disclosure relates to any of the above methods, wherein the computer simulation analysis further includes: replacing multiple random synonymous codons in the gene construct with multiple substitution sequences, such that the multiple substitution sequences increase the sum of the array.

[0123] One embodiment of this disclosure relates to any of the methods described above, wherein the gene construct contains a sequence encoding a chimeric antigen receptor.

[0124] One embodiment of this disclosure relates to any of the methods described above, wherein a predetermined value of an acceptable frequency percentage of the undesired variant is determined based on whether the undesired variant is associated with at least one of the following: whether the undesired variant adversely affects the output of the chimeric antigen receptor to the cell surface; whether the undesired variant is associated with changes in the binding domain of the chimeric antigen receptor; and whether the undesired variant has previously been characterized as having a negligible effect on the expression or function of the chimeric antigen receptor.

[0125] One embodiment of this disclosure relates to any of the methods described above, wherein if the undesired variant adversely affects the output of the chimeric antigen receptor to the cell surface, the predetermined value of the acceptable frequency percentage of the undesired variant is 0.1%; and wherein if the undesired variant is associated with a change in the binding domain of the chimeric antigen receptor, the predetermined value of the acceptable frequency percentage of the undesired variant is 0.01%.

[0126] One embodiment of this disclosure relates to any of the methods described above, wherein the computer simulation analysis and replacement steps are not repeated if the undesired variant has previously been characterized as having a negligible effect on the expression or function of the chimeric antigen receptor.

[0127] One embodiment of this disclosure relates to any of the methods described above, the method further comprising the steps of: identifying and removing a subpopulation of high-frequency variants; and identifying a subpopulation of low-frequency variants; and wherein the in vivo analysis further comprises: performing an analysis to determine whether the subpopulation of the low-frequency variant should be replaced.

[0128] One embodiment of this disclosure relates to any of the methods described above, wherein if an unwanted variant is detected and none of the conditions or requirements described above are met, the computer simulation analysis and sequence removal steps are repeated to attempt to eliminate the unwanted variant, and / or the unwanted variant is further characterized in an additional study to assess the risk to potential patients.

[0129] One embodiment of this disclosure relates to a method for mitigating the risk that a gene product for cell therapy may have reduced efficacy or toxicity due to the generation of undesirable variants. Such a method includes the steps of: performing a computer simulation analysis on a gene construct encoding the gene product to detect the presence of a sequence capable of causing the undesirable variant; replacing the detected sequence capable of causing the undesirable variant with a substitution sequence, wherein the substitution sequence is a derived sequence including synonymous codon substitutions; measuring the frequency percentage of the undesirable variant expressed by the gene construct, the measurement including performing an in vivo analysis on one or more genes expressed by the gene construct, the in vivo analysis including performing RNA sequencing analysis on an RNA product transcribed from the gene construct, wherein the frequency percentage of the undesirable variant is determined at least in part by using a splice-sensing aligner from the RNA sequencing analysis; and repeating the computer simulation analysis and substitution steps if the frequency percentage of the undesirable variant in the gene product from the in vivo analysis is greater than a predetermined value for an acceptable frequency percentage of the undesirable variant.

[0130] Variant prediction, detection and elimination

[0131] One implementation discloses a general pipeline for processing variants. This pipeline relies on algorithms for identifying and removing sequences caused by potential variants, as well as sequencing and alignment methods for detecting the presence of variants. The algorithm-based and sequencing-based portions of this pipeline are described in further detail in subsequent sections.

[0132] Figure 2 The document provides an overview of one implementation of this pipeline. This pipeline uses recurrent logic in which efforts are made to identify and remove variant-generating sequences through computer simulation, followed by physical testing and further computer simulation work if variants are found. Variant detection is performed in multiple sequencing steps, thus eliminating the need to invest significant effort in construct sequences that can be easily identified as variant-generating.

[0133] In this implementation, the production line is designed to be flexible. Figure 2 The frequency cutoff values ​​shown are examples, and alternative methods for determining whether variant addressing is necessary can be used (discussed in the sequencing section). Similarly, the number of donors in the sequencing steps can be increased / decreased, or additional sequencing steps can be added / removed (e.g., sequencing of lentiviral vectors) to maximize efficiency or to search for variants more deeply.

[0134] refer to Figure 2In one implementation, prior to any physical work, algorithms designed to identify splice sites and highly homologous sequences (which may lead to homologous recombination) are run on the planned construct sequence. If any of these splice sites and highly homologous sequences are identified, they are removed using alternative algorithms via synonymous codon substitution (the construct protein sequence remains unchanged). Once this initial computer simulation screening and modification is complete, a small batch of product using a single donor is produced, and the RNA sequence from this batch is used to identify any high-frequency (e.g., greater than about 5%) variants. If high-frequency variants are found, the computer simulation screening and modification process is repeated, guided by knowledge of the identified variants, before retrying small-batch production and RNA sequencing. If no high-frequency variants are found, a larger batch of product using 5 to 10 donors is produced, and the RNA sequence from this batch is used to identify any low-frequency (<5%) variants. If any of these variants are found, the sequence and expression levels can be analyzed via computer simulation to determine if the variant poses a safety / efficacy risk. As an example of this evaluation, variants of CARs with splice co-stimulatory domains occurring only in 1% of the product are unlikely to be a problem, as this would likely result in only a very small reduction in efficacy. Conversely, splicing in the scFv domain of the CAR, even in low percentages, can lead to off-target binding and thus toxicity. If safety / efficacy risks cannot be ruled out, products expressing these specific variants can be generated and physically tested. If these tests indicate safety / efficacy issues or cannot be performed, the construct sequence can be redesigned using computer simulations, and the entire process can be restarted. Otherwise, if no low-frequency variants are found, or if no safety / efficacy risks are confirmed, the construct sequence can be eliminated for further development.

[0135] Splice site prediction and removal algorithm

[0136] In one implementation, during the computer simulation step of the pipeline, the SpliceAI-Jaganathan et al. Cell 2019 (https: / / github.com / Illumina / SpliceAI) and MaxEntScan-Yeo and Burge, Journal of Computational Biology 2003 (http: / / hollywood.mit.edu / burgelab / maxent / Xmaxentscan_scoreseq.html) algorithms can be used to identify potential donor and acceptor splicing sites. In other implementations, other algorithms can be used.

[0137] In one implementation, if sites in the construct are identified using any of these algorithms, synonymous codon substitutions can be used to modify these sites until variants are no longer detected or their predicted splicing probability is significantly reduced. This step leaves the construct protein sequence unchanged. Similarly, if sequencing has already been performed and variants have been detected, these algorithms can be used to determine whether variants are likely caused by splicing, and synonymous codon substitutions can be used to reduce the likelihood of variant occurrence.

[0138] Homology identification and removal algorithms

[0139] Since homologous recombination is driven by highly similar sequences within the construct, this disclosure describes algorithms and tools to help locate and remove these sequences. Similar to splice site removal algorithms, these algorithms are run during the initial computer simulation step and can be re-run after the sequencing step if sequencing indicates homologous recombination.

[0140] To identify sequences that may cause homologous recombination, one implementation includes a repeat finder tool. This tool analyzes and displays the construct sequence and plots an arch connecting all identical sequence pairs longer than a given (user-defined) length. Alternatively, the tool can connect sequence pairs of a given size with a given similarity level determined by a user-defined Lewinstein distance. Figure 6 The image shows an example screenshot displaying the output from the Duplicate Finder tool. The thickness of the arch depends on the length of the identical sequences it connects to. By searching for thick arches or multiple clusters of arches, users can easily identify highly similar sequences, even if these sequences are not identical.

[0141] To reduce the similarity between all segments of a given construct, one implementation includes a sequence divergence tool. This tool takes a construct sequence of length n (in base pairs) and a user-selected segment size k (also in base pairs). The tool then creates a matrix of length n-k+1 in both dimensions. Each position (x, y) in this matrix corresponds to a pair of segments starting at bases x and y in the construct sequence. The value at this position is the Hamming distance between these two segments. An example of this matrix is ​​shown below. Figure 3As shown. Using this design, the sum of the matrix effectively describes the similarity of the construct sequence to itself. A construct sequence with many highly similar segments will have a smaller sum than a construct sequence where most of its segments are distinct from each other. This means that reducing the similarity between all segments of the construct, and thus reducing the probability of homologous recombination events, is a problem of increasing the sum of the matrix. The sequence diverger achieves this by performing random synonymous codon substitutions throughout the construct sequence and retaining only those substitutions that increase the sum of the matrix. A user-specified number of substitutions are performed before the algorithm terminates, and the modified sequence is returned. The user can track the sum of the matrix relative to the number of substitutions attempted and thus understand how many steps are required.

[0142] like Figure 3 As shown in the example, for a given sequence (AACGAACG) and a given segment size (4), the Hamming distance (HD) is calculated for all possible sub-segment pairs. The sum of this matrix describes how the different segments of the sequence are similar to each other, so maximizing this sum should reduce the probability of homologous recombination.

[0143] In another implementation, a duplication remover tool can be used to help prevent the construct sequence from containing any identical segments that could cause homologous recombination. This tool takes the construct sequence and a user-selected segment size k (in base pairs). In one implementation, the duplication remover tool then goes through a process where each segment of size k, starting from position 1 of the construct, is compared with all other segments in the construct. If an identical segment is found, a random synonymous codon substitution within that segment is used to eliminate the similarity. This process is repeated cyclically until the entire construct is scanned without finding any identical segments or reaching the user-defined number of iterations. Figure 4 An example of this situation is shown. Multiple cyclic repetitions may be necessary because synonymous codon substitutions may be performed to eliminate one pair of identical sequences in order to introduce another pair of identical sequences.

[0144] like Figure 4 As shown, the duplicate remover tool compares all segments of a user-selected size (e.g., 6 to 10 base pairs) to see if they are identical. If an identical pair is found, it is removed using synonymous codon substitution. In this implementation, this process continues until all identical pairs have been removed, or the user-defined number of cycles is reached.

[0145] Although they share the same goal of removing highly similar sequences that may cause homologous recombination, sequence divergent and duplication remover tools employ different methods and can produce potentially different results. Duplication removers focus on removing identical sequences, and at the cost of doing so, they may produce highly similar (but not identical) sequences. Conversely, sequence divergents focus on globally reducing sequence similarity, and at the cost of doing so, they may leave several identical sequences in place. The computer simulation steps of an implementation can be performed using both tools in either order or with only one tool, depending on what works best for the particular construct being developed. In other implementations, no tools may be used in this process; instead, manual removal of highly similar sequences identified via a duplication remover (or duplication watcher) can be employed.

[0146] These implementations are flexible because the tools described above can be used individually or in combination, and additional tools or features can be added. For example, all three tools can be set to ignore highly similar / identical sequences within a given distance. This is relevant because there are indications that homologous recombination events require a minimum distance between highly similar sequences to occur. Furthermore, the sequence diverger and duplication remover tools can be configured so that certain regions of the construct remain unchanged, even when these regions are known to be highly sensitive to codon usage. Finally, the sequence diverger tool can be modified based on simulated annealing logic to accept or reject synonymous codon substitutions, rather than only accepting those that increase the matrix sum. This helps the tool better find global maxima.

[0147] In some implementations, the duplicate finder, sequence diverger, and duplicate remover tools can be separate computer modules. In other implementations, these three tools can be programmed into a single module or computer.

[0148] Sequencing-based variant detection

[0149] There is some flexibility in how RNA sequencing is performed to identify variants. It is desirable that the sequencing has sufficient quality and depth to allow for the observation of variants. For example, in one implementation, CAR-T products can be sequenced with HiSeq 2500 sequencing channels, a depth of approximately 300 million reads per sample, and paired end reads of 150 bp in length.

[0150] After sequencing, the reads should be aligned with the construct using one or more splice-sensing aligners. Flexibility exists in this step depending on the precise alignment parameters and the aligner used. This implementation has shown good results using three different aligners simultaneously (STAR, HISAT2, and TopHat2) because it helps identify and ignore errors in any one of the aligners (discussed later).

[0151] In one implementation, after alignment, all reads with gap alignment (since these reads are the ones that will indicate variants) are extracted and translated to produce protein sequences. The percentage of each variant (the protein sequence that differs from the expected sequence) in the sample is calculated using the following formula:

[0152]

[0153] Where x is the percentage, R is the number of readings supporting the variant, n is the number of bases in the Open Reading Frame (ORF), and d is the reading depth at each position in the ORF (all readings).

[0154] After specifying readings and sample percentages for each variant, the final step in this implementation is to determine which variants require further investigation. This step takes flexibility into account and should consider whether it is an initial small screening for identifying highly expressed variants or a large screening for identifying all variants (see [link to implementation details]). Figure 2 ).

[0155] One option is to establish a cutoff value based on the number of readings. For example, if a sample has five or more readings supporting the variant, the sample can be considered positive for the variant. Another option is to consider the percentage of samples supporting the variant (e.g., ...). Figure 2 The threshold (shown) is whether the sample is above a certain cutoff value. In both cases, if samples from multiple donors are used, a secondary cutoff value based on multiple donors can be established. For example, only positive variants from samples from a given number or percentage of donors would be considered. In another implementation, a statistical test, such as the Wilcox rank-sum test, can be used, where each donor sample is considered an independent experiment to determine whether the donor positivity rate for a given variant is significantly different from 0. Finally, if multiple aligners are used, a cutoff value based on them can be used to address inaccuracies in any one of the pipelines. For example, if a sample has five or more readings supporting a variant in at least two of the three aligners, the sample can be considered positive only for the variant. Alternatively, a variant can only be considered if a Wilcox rank-sum test returns positive results for multiple donors using at least two pipelines.

[0156] After statistical analysis of the variants, variant-supporting and normal readings can be visually examined in a program such as the Integrated Genomic Observer (IGV). Some noticeable variations may be the result of alignment errors that can be detected visually. In some implementations of this process, in addition to performing the above-described tests, the process may also include separating all readings from the sample supporting the specific variant and placing them in a single BAM file that can be easily examined.

[0157] RNA sequence analysis and bioinformatics analysis

[0158] In another implementation, RNA sequence analysis and bioinformatics analysis include quality control checks using tools such as FastQC (version 0.11.7). Using FastQC's default parameters, the gene construct should pass quality control. The RNA sequence analysis and bioinformatics analysis may also include an alignment step. To maximize splice event detection, the readings are aligned with the construct sequence using three different splice-sensing aligners: STAR (version 2.7.3a) (PMID: 23104886), HISAT2 (version 2.1.0) (PMID: 31375807), and TopHat2 (version 2.1.0) (PMID: 23618408). Custom alignment indexes corresponding to the construct sequence can be generated for each tool. In one implementation, alignment can be performed on the Seven Bridges platform, but it can also be performed on other computing platforms.

[0159] In one implementation, the RNA sequence analysis and bioinformatics analysis may include STAR alignment. The reference index for STAR alignment can be created using the `genomeGenerate` command, where the `genomeSAindexNbases` parameter is set to 5, and all other parameters are set to their default values. In one implementation, alignment in STAR can be performed using all default parameters. In one implementation, the following command can be used to create the index, and the alignment can be performed using the STAR tool:

[0160] STAR indexing command:

[0161] STAR--runMode genomeGenerate--genomeDir. / genomeDir--runThreadN 32--genomeSAindexNbases 5-genomeFastaFiles construct_name.fa--limitGenomeGenerateRAM 60000000000

[0162] STAR alignment command:

[0163] STAR--runThreadN 32--readFilesCommand zcat--genomeDir. / genomeDir

[0164] --limitBAMsortRAM 0--outSAMtype BAM Unsorted--readFilesInR1.fastq.gzR2.fastq.gz

[0165] It should be understood that other commands can be used to create indexes and perform alignments.

[0166] Furthermore, the RNA sequence analysis and bioinformatics analysis may include HISAT2 alignment. In one implementation, a HISAT2-aligned reference index is created using the Hhisat2-build command with HISAT2 version 2.0.1. HISAT2 alignment can be performed with the --no-softclip --no-unal options enabled and the --pen-cansplice and --pen-noncansplice parameters set to 0. All other parameters can be set to their default values, and then the readings are sorted using Sambamba (version 0.6.6) (PMID: 25697820). In one implementation, the index can be created using the following command, and alignment can be performed using the HISAT2 tools:

[0167] HISAT2 indexing commands:

[0168] hisat2-build-p1 construct_name.fa index / construct_name_HISAT2-2.0.1

[0169] HISAT2 alignment command:

[0170] hisat2--met-file metrics.txt--no-softclip--no-unal-p 20--pen-cansplice 0--pen-

[0171] noncansplice 0-x. / index_files_path-1R1_001.fastq.gz-2R2_001.fastq.gz-S

[0172] / dev / stdout

[0173] It should be understood that other commands can be used to create indexes, and alignment can be performed using the HISAT2 tool.

[0174] In one implementation, the RNA sequence analysis and bioinformatics analysis steps may include TopHat2 alignment. The reference index for TopHat2 alignment can be created using the BowTie2-build command (version 2.2.6) (PMID: 21154709), etc. In one implementation, TopHat2 alignment can be performed using all default parameters. The following commands can be used to create the index and perform alignment using the TopHat2 tools:

[0175] BowTie2-build command:

[0176] bowtie2-build-f construct.fa. / construct_name

[0177] TopHat2 alignment:

[0178] tophat2--num-threads 1--output-dir. / tophat_out--no-coverage-

[0179] search. / construct_name

[0180] R1_001.fastq.gz R2_001.fastq.gz

[0181] In one implementation, the analysis step includes processing the readings. In this implementation, the aligned readings from each alignment method can be further processed on the Seven Bridges platform. First, all gapless readings can be removed using SAMtools (version 1.6) (PMID: 19505943), viewing functions, etc. Then, the remaining gap readings in BAM file format can be converted to SAM format using SAMtools (version 1.9). Next, R (version 3.6.2) can be used... https: / / www.R- project.org / The script (translate_and_group.R) translates the nucleotide sequence from each gap readout into its corresponding amino acid sequence. Furthermore, the script can be used to calculate the number of reads supporting each unique gap event. In one embodiment, gap reads with overhangs of fewer than 10 base pairs can be removed. In one embodiment, the script utilizes Seqinr (version 3.6.1) (ISBN: 978-3-540-35305-8, https: / / cran.r- project.org / web / packages / seqinr / index.html The library package translates interstitial DNA sequences into amino acid sequences.

[0182] SAMtools commands are used to remove non-interval readings:

[0183] samtools view-h / path / to / input_bam.ext|awk'{if($0~ / ^@ / ||$6~ / N / ){print$0}}'|samtools view-Sb->input_bam_gapped.bam

[0184] The SAMtools command is used to convert BAM files to SAM file format:

[0185] samtools view--output-fmt SAM-o hits_gapped.sam hits_gapped.bam

[0186] In one implementation, the final step of this bioinformatics analysis can be performed on an Amazon Web Services (AWS) Virtual Private Cloud (VPC) using an EC2 instance running an R (version 3.6.2) script (app.R). This script imports the output file (translate_and_group.R) generated by the s-script, along with the BAM output file from the alignment, to calculate the prevalence of each gap event. The formula used for this calculation is:

[0187]

[0188] Where x is the coverage percentage, R is the number of readings supporting the gap, n is the number of bases in the open read frame (ORF), and d is the reading depth at each position in the ORF.

[0189] In one implementation, the app.R script relies on SAMtools (version 1.10) depth functions to calculate the coverage of the construct and generates a BAM file (with calls to SAMtools view functions) containing readings from each unique gap event for visualization purposes. Following this analysis, a gap event is considered a variant if it meets the following screening criteria: a p-value (nonparametric single-sample Wilcoxon rank-sum test) < 0.01 in two of the three methods, where cross-validation is used to minimize method-specific artifacts.

[0190] The null hypothesis and the alternative hypothesis are:

[0191] H0: μ = 0

[0192] H a μ≠0

[0193] In one implementation, a conservative threshold with a p-value <0.01 may be selected to minimize the false positive rate, which is caused by a large number of false putative variants supported by <5 reads in a single donor and may be the result of sequencing artifacts.

[0194] Report generation

[0195] In one implementation, the variant detection method can generate and display a report for each variant, which includes information such as... Figure 5 The following should be shown: 1) the sequence of the expected protein product; 2) the variant ID or number; 3) a graph or diagram annotating the variation of the characteristics of the protein product (i.e., CAR); 4) the alignment frequency of each aligner used (e.g., Tophat, HISAT, STAR); 5) the p-value of the detected statistical significance (if available); and 6) visualization of the readings aligned with the construct.

[0196] Those skilled in the art will recognize that this disclosure may be implemented in other specific forms without departing from the spirit or essential characteristics thereof. Therefore, the foregoing embodiments are to be considered in all respects as illustrative rather than limiting of the subject matter described herein.

[0197] Example

[0198] The following embodiments disclose an exemplary table of potential software packages that can be used for RNA sequencing analysis and bioinformatics analysis, as well as exemplary software code portions that can be used for the disclosed processes.

[0199] Example 1

[0200] This embodiment provides an example of a method 101 for detecting and replacing sequences that can cause undesirable variants in a gene construct. For example... Figure 7AAs shown, method 101 includes at least the following steps. Method 101 includes step 103: performing a computer simulation analysis on the gene construct to detect the presence of the sequence capable of causing the undesirable variant. Method 101 further includes step 105: replacing the detected sequence capable of causing the undesirable variant with a substitution sequence. In step 105, the substitution sequence is obtained by using synonymous codon substitution. Method 101 includes step 107: measuring the frequency percentage of the undesirable variant expressed by the gene construct. Step 107 includes performing an in vivo analysis on one or more genes expressed by the gene construct by performing RNA sequencing analysis on the RNA product transcribed from the gene construct, wherein the frequency percentage of the undesirable variant is determined at least in part by using a splice-sensing aligner from the RNA sequencing analysis. Method 101 further includes repeating the computer simulation analysis and substitution steps 103 and 105 if the frequency percentage 107 of the undesirable variant in the gene product from the in vivo analysis is greater than a predetermined value of an acceptable frequency percentage of the undesirable variant.

[0201] In method 101, step 107 requires the use of at least two separate aligners. Furthermore, method 101 is preferably used for gene constructs encoding chimeric antigen receptors.

[0202] In method 101, if the undesirable variant adversely affects the output of the chimeric antigen receptor to the cell surface, the predetermined value of the acceptable frequency percentage of the undesirable variant is 0.1%; and if the undesirable variant is associated with a change in the binding domain of the chimeric antigen receptor, the predetermined value of the acceptable frequency percentage of the undesirable variant is 0.01%.

[0203] In method 101, if the undesirable variant has previously been characterized as having a negligible effect on the expression or function of the chimeric antigen receptor, the computer simulation analysis and replacement steps 103 and 105 are not repeated.

[0204] Example 2

[0205] This embodiment provides an example of a method 201 for generating a gene product for cell therapy. For example... Figure 7BAs shown, method 201 includes at least the following steps. Method 201 includes step 203: performing computer simulation analysis on a gene construct encoding a gene product to identify and alter sequences capable of causing undesirable variants. Method 201 further includes step 205: replacing the detected sequences capable of causing the undesirable variant with alternative sequences. In step 205, the alternative sequences are derived using synonymous codon substitution. Method 201 includes step 207: measuring the percentage frequency of the undesirable variant expressed by the gene construct. Step 207 includes performing in vivo analysis on one or more genes expressed by the gene construct by performing RNA sequencing analysis on the RNA product transcribed from the gene construct. In step 207, the percentage frequency of the undesirable variant is determined at least in part by using a splice-sensing aligner from the RNA sequencing analysis. Method 201 includes repeating the computer simulation and replacement steps 203 and 205 to generate a new gene construct if the frequency percentage of the undesired variant in the gene product from the in vivo analysis is greater than a predetermined value of an acceptable frequency percentage of the undesired variant. Method 201 also includes step 209: measuring the frequency percentage of the undesired variant expressed by the new gene construct. Step 209 includes performing an in vivo analysis on one or more genes expressed by the new gene construct by performing RNA sequencing analysis on the RNA product transcribed from the new gene construct, wherein the frequency percentage of the undesired variant is determined at least in part by using a splice-sensing aligner from the RNA sequencing analysis.

[0206] In method 201, step 207 requires the use of at least two separate aligners. Furthermore, method 201 is preferably used for gene constructs encoding chimeric antigen receptors.

[0207] In method 201, if the undesirable variant adversely affects the export of the chimeric antigen receptor to the cell surface, the predetermined value of the acceptable frequency percentage of the undesirable variant is 0.1%; and if the undesirable variant is associated with a change in the binding domain of the chimeric antigen receptor, the predetermined value of the acceptable frequency percentage of the undesirable variant is 0.01%.

[0208] In method 201, if the undesirable variant has previously been characterized as having a negligible effect on the expression or function of the chimeric antigen receptor, the computer simulation analysis and replacement steps 203 and 205 are not repeated.

[0209] Example 3

[0210] Table 1: Exemplary software packages for RNA sequence analysis and bioinformatics analysis.

[0211]

[0212] Example 4

[0213] This embodiment provides sample code portions that can be used to practice the disclosed methods. Figure 8A In the example code section depicted, x is the coverage percentage, R is the number of readings supporting the gap, n is the number of bases in the open read frame (ORF), and d is the reading depth at each position in the ORF (see lines 25 to 45 of the app.R script).

[0214] Example 5

[0215] This embodiment provides sample code portions that can be used to practice the disclosed methods. Figure 8B In the example code section depicted, the app.R script relies on the SAMtools (version 1.10) depth function to calculate the coverage of the construct and produce a BAM file (with calls to the SAMtools view function) that has readings from each unique gap event for visualization purposes (see app.R 48 to 82).

[0216] Example 6

[0217] This embodiment provides sample code portions that can be used to practice the disclosed methods. Figure 8C In the example code section depicted, a conservative threshold OA with a p-value < 0.01 was chosen to minimize the false positive rate, which is due to a large number of false putative variants supported by < 5 reads in a single donor, as well as possible sequencing artifacts. (See app.R script lines 131-136)

[0218] All publications, patents, patent applications and other documents cited in this application are incorporated herein by reference in their entirety for all purposes, to the extent that each individual publication, patent, patent application or other document is individually indicated as being incorporated by reference for all purposes.

[0219] Although various specific implementation schemes / aspects have been described and illustrated, it should be understood that various changes may be made without departing from the spirit and scope of this disclosure.

Claims

1. A method for detecting and replacing sequences capable of causing undesirable variants in a gene construct, the method comprising: Computer simulation analysis was performed on the gene construct to detect the presence of the sequence that could cause the undesirable variant; Replace the detected sequence that could cause the undesirable variant with an alternative sequence, wherein the alternative sequence is a derived sequence that includes synonymous codon substitutions; The frequency percentage of the unwanted variant expressed by the gene construct is measured, the measurement comprising performing an in vivo analysis on one or more genes expressed by the gene construct, the in vivo analysis comprising performing RNA sequencing analysis on RNA products transcribed from the gene construct, wherein the frequency percentage of the unwanted variant is determined at least in part by using a splice-sensing aligner from the RNA sequencing analysis. as well as If the frequency percentage of the unwanted variant in the gene product from the in vivo analysis is greater than a predetermined value for the acceptable frequency percentage of the unwanted variant, then the computer simulation analysis and replacement steps are repeated. The gene construct described herein contains a sequence encoding a chimeric antigen receptor. A predetermined value for the acceptable percentage of frequency of the undesired variant is determined based on whether the undesired variant is associated with at least one of the following: whether the undesired variant adversely affects the export of the chimeric antigen receptor to the cell surface; whether the undesired variant is associated with changes in the binding domain of the chimeric antigen receptor; and whether the undesired variant has previously been characterized as having a negligible effect on the expression or function of the chimeric antigen receptor. If the undesired variant adversely affects the output of the chimeric antigen receptor to the cell surface, the predetermined value of the acceptable frequency percentage of the undesired variant is 0.1%; and if the undesired variant is associated with a change in the binding domain of the chimeric antigen receptor, the predetermined value of the acceptable frequency percentage of the undesired variant is 0.01%.

2. The method of claim 1, wherein using the splice-aware aligner includes using at least two separate aligners.

3. The method according to any one of claims 1 or 2, wherein the computer simulation analysis further comprises: Detecting at least one of a plurality of homologous sequences and a plurality of identical sequences in the gene construct, wherein the plurality of homologous sequences and the plurality of identical sequences are capable of causing undesirable variants in the gene construct; and Replace any such detected multiple homologous sequences and multiple identical sequences, the replacement including a synonymous codon substitution step.

4. The method according to claim 1, wherein the computer simulation analysis further comprises: Calculate the matrix of segment combinations from the gene construct and obtain the Hamming distance of each segment combination in the segment combination.

5. The method according to claim 4, wherein the computer simulation analysis further comprises: Multiple random synonymous codons in the gene construct are replaced with multiple substitution sequences, such that the multiple substitution sequences increase the sum of the matrix.

6. The method of claim 1, wherein the computer simulation analysis and replacement steps are not repeated if the undesired variant has previously been characterized as having a negligible effect on the expression or function of the chimeric antigen receptor.

7. The method according to claim 1, further comprising: Identify and remove subgroups of high-frequency variants; And to identify subgroups of low-frequency variants; and The in vivo analysis further includes performing an analysis to determine whether the subpopulation of the low-frequency variant should be replaced.

8. A method for preparing a gene product for cell therapy, the method comprising: Computer simulation analysis was performed on the gene construct encoding the gene product to identify and modify sequences that could cause undesirable variants; Replace the detected sequence that could cause the undesirable variant with an alternative sequence, wherein the alternative sequence is a derived sequence that includes synonymous codon substitutions; The frequency percentage of the unwanted variant expressed by the gene construct is measured, the measurement comprising performing an in vivo analysis on one or more genes expressed by the gene construct, the in vivo analysis comprising performing RNA sequencing analysis on RNA products transcribed from the gene construct, wherein the frequency percentage of the unwanted variant is determined at least in part by using a splice-sensing aligner from the RNA sequencing analysis. If the percentage of the frequency of the undesired variant in the gene product from the in vivo analysis is greater than a predetermined value for the acceptable percentage of the frequency of the undesired variant, then the computer simulation and replacement steps are repeated to generate a new gene construct. as well as The measurement involves measuring the percentage frequency of the unwanted variant expressed by the novel gene construct, the measurement comprising performing in vivo analysis on one or more genes expressed by the novel gene construct, the in vivo analysis comprising performing RNA sequencing analysis on RNA products transcribed from the novel gene construct, wherein the percentage frequency of the unwanted variant is determined at least in part by using a splice-sensing aligner from the RNA sequencing analysis. The gene construct described herein contains a sequence encoding a chimeric antigen receptor. A predetermined value for the acceptable percentage of frequency of the undesired variant is determined based on whether the undesired variant is associated with at least one of the following: whether the undesired variant adversely affects the export of the chimeric antigen receptor to the cell surface; whether the undesired variant is associated with changes in the binding domain of the chimeric antigen receptor; and whether the undesired variant has previously been characterized as having a negligible effect on the expression or function of the chimeric antigen receptor. If the undesired variant adversely affects the output of the chimeric antigen receptor to the cell surface, the predetermined value of the acceptable frequency percentage of the undesired variant is 0.1%; and if the undesired variant is associated with a change in the binding domain of the chimeric antigen receptor, the predetermined value of the acceptable frequency percentage of the undesired variant is 0.01%.

9. The method of claim 8, wherein using the splice-aware aligner includes using at least two separate aligners.

10. The method according to any one of claims 8 or 9, wherein the computer simulation analysis further comprises: Detecting at least one of a plurality of homologous sequences and a plurality of identical sequences in the gene construct, wherein the plurality of homologous sequences and the plurality of identical sequences are capable of causing undesirable variants in the gene construct; and Replace any such detected multiple homologous sequences and multiple identical sequences, the replacement including a synonymous codon substitution step.

11. The method of claim 8, wherein the computer simulation analysis further comprises: Calculate the matrix of segment combinations from the gene construct and obtain the Hamming distance of each segment combination in the segment combination.

12. The method of claim 11, wherein the computer simulation analysis further comprises: Multiple random synonymous codons in the gene construct are replaced with multiple substitution sequences, such that the multiple substitution sequences increase the sum of the matrix.

13. The method of claim 8, wherein the computer simulation analysis and replacement steps are not repeated if the undesired variant has previously been characterized as having a negligible effect on the expression or function of the chimeric antigen receptor.

14. The method according to claim 8, further comprising: Identify and remove subgroups of high-frequency variants; And to identify subgroups of low-frequency variants; and The in vivo analysis further includes performing an analysis to determine whether the subpopulation of the low-frequency variant should be replaced.

15. A system for detecting and replacing sequences capable of causing undesirable variants in a gene construct, said system being configured to perform methods including: Computer simulation analysis was performed on the gene construct to detect the presence of the sequence that could cause the undesirable variant; Replace the detected sequence that could cause the undesirable variant with an alternative sequence, wherein the alternative sequence is a derived sequence that includes synonymous codon substitutions; The frequency percentage of the unwanted variant expressed by the gene construct is measured, the measurement comprising performing an in vivo analysis on one or more genes expressed by the gene construct, the in vivo analysis comprising performing RNA sequencing analysis on RNA products transcribed from the gene construct, wherein the frequency percentage of the unwanted variant is determined at least in part by using a splice-sensing aligner from the RNA sequencing analysis. as well as If the percentage of the frequency of the unwanted variant in the gene product from the in vivo analysis is greater than a predetermined value for the acceptable percentage of the frequency of the unwanted variant, then the computer simulation analysis and replacement steps are repeated. The gene construct described herein contains a sequence encoding a chimeric antigen receptor. A predetermined value for the acceptable percentage of frequency of the undesired variant is determined based on whether the undesired variant is associated with at least one of the following: whether the undesired variant adversely affects the export of the chimeric antigen receptor to the cell surface; whether the undesired variant is associated with changes in the binding domain of the chimeric antigen receptor; and whether the undesired variant has previously been characterized as having a negligible effect on the expression or function of the chimeric antigen receptor. If the undesired variant adversely affects the output of the chimeric antigen receptor to the cell surface, the predetermined value of the acceptable frequency percentage of the undesired variant is 0.1%; and if the undesired variant is associated with a change in the binding domain of the chimeric antigen receptor, the predetermined value of the acceptable frequency percentage of the undesired variant is 0.01%.

16. The system of claim 15, wherein using the splice-aware aligner includes using at least two separate aligners.

17. The system according to any one of claims 15 or 16, wherein the computer simulation analysis further comprises: Detecting at least one of a plurality of homologous sequences and a plurality of identical sequences in the gene construct, wherein the plurality of homologous sequences and the plurality of identical sequences are capable of causing undesirable variants in the gene construct; and Replace any such detected multiple homologous sequences and multiple identical sequences, the replacement including a synonymous codon substitution step.

18. The system of claim 15, wherein the computer simulation analysis further comprises: Calculate the matrix of segment combinations from the gene construct and obtain the Hamming distance of each segment combination in the segment combination.

19. The system of claim 18, wherein the computer simulation analysis further comprises: Multiple random synonymous codons in the gene construct are replaced with multiple substitution sequences, such that the multiple substitution sequences increase the sum of the matrix.

20. The system of claim 15, wherein the computer simulation analysis and replacement steps are not repeated if the undesired variant has previously been characterized as having a negligible effect on the expression or function of the chimeric antigen receptor.

21. The system of claim 15, wherein the method further comprises: Identify and remove subgroups of high-frequency variants; And to identify subgroups of low-frequency variants; and The in vivo analysis further includes performing an analysis to determine whether the subpopulation of the low-frequency variant should be replaced.

Images