Early detection of hepatocellular carcinoma

Abstract

The present invention relates to early detection of hepatocellular carcinoma. Specifically, the present invention provides methods, assays, and kits for detecting hepatocellular carcinoma, as well as methods for stratifying subjects into higher risk and lower risk categories for developing hepatocellular carcinoma, and methods for treating and managing subjects suspected of having or at risk for developing hepatocellular carcinoma. While previous work has attempted to address the need for highly sensitive early predictive indicators for hepatocellular carcinoma by evaluating one or more biological factors, none has approached the sensitivity required for a clinically relevant determination of whether a subject, particularly an asymptomatic subject, has the disease. The inventors have discovered that certain combinations of factors meet this need by conferring previously unattainable levels of high accuracy.

Description

[0001] This application is a divisional application of application PCT / US2017 / 018040, with an international application date of February 16, 2017, which entered the Chinese national phase on August 24, 2018, with application number 201780013280.1 and invention name “Early Detection of Hepatocellular Carcinoma”.

[0002] CROSS-REFERENCE TO RELATED APPLICATIONS

[0003] This application claims the benefit of priority to U.S. Provisional Application No. 62 / 300,142, filed February 26, 2016, the entire contents of which are incorporated herein by reference. Technical Field

[0004] The present disclosure relates to assays, kits, and methods for detecting liver disease, patient stratification, and therapeutic intervention. Background Art

[0005] Infection with hepatitis B virus (HBV) and / or hepatitis C virus (HCV) is the main cause of hepatocellular carcinoma (HCC). Both HBV and HCV cause acute and chronic liver infections, and most chronically infected individuals remain asymptomatic for many years, as clinical disease (HCC and cirrhosis) typically takes decades to develop. 10% to 40% of all chronic HBV carriers eventually develop liver cancer, and it is estimated that more than one million people die from HBV / HCV-related liver cancer worldwide. In fact, HBV and HCV infection are associated with more than 80% of all HCC cases worldwide, and can be as high as 96% in areas where HBV is prevalent.

[0006] The progression of liver disease to liver cancer is primarily monitored by measuring serum levels of the oncofetal glycoprotein, alpha-fetoprotein (AFP), or the core fucosylated glycoform of AFP (AFP-L3). However, AFP can be produced in many conditions, including other liver diseases such as human hepatocellular carcinoma, hepatoblastoma, and hepatitis B infection, and is not present in all HCC patients. Therefore, the use of AFP as an initial screen for HCC has been questioned, and there is a need for more sensitive serum biomarkers for HCC.

[0007] Glycosylation of proteins can be cell-specific, and the N-linked glycans carried by a protein are modifications that occur in its cell of origin. The sugar (glycan) structures on proteins secreted from malignant or diseased tissues can and often do differ from the glycans present on the same protein from normal cells. In fact, many studies have observed that N-linked glycosylation changes with the development of cirrhosis and hepatocellular carcinoma (HCC) (see, for example, Naitoh, A. et al., Highly enhanced fucosylation of serum glycoproteins in patients with hepatocellular carcinoma. J Gastroenterol Hepatol, 14: 436-445, 1999; Block, TM et al., Use of targeted glycoproteomics to identify serum glycoproteins that correlate with liver cancer in woodchucks and humans. Proc Natl Acad Sci USA, 102: 779-784, 2005). For example, in individuals chronically infected with HCV and diagnosed with HCC, the amount of fucosylated N-linked glycans derived from total protein preparations isolated from serum is consistently greater than in healthy subjects or those with HCV and "inactive" disease (Comunale, MA et al., Proteomic analysis of serum associated fucosylated glycoproteins in the development of primary hepatocellular carcinoma. Journal of Proteome Research., 6:308-315, 2006).

[0008] However, it has not yet been possible to identify a single glycoprotein, glycosylation pattern, biomarker panel, or other indicator of HCC that confers high sensitivity and capability for early detection.There remains an unmet need for early detection agents for HCC. Summary of the Invention

[0009] Provided herein are methods for detecting hepatocellular carcinoma in a subject, the methods comprising removing IgG and IgM proteins from a biological fluid from the subject; measuring the amount of one or more biomarkers in the biological fluid; determining the age and sex of the subject; and determining the absence or presence of hepatocellular carcinoma in the subject using an optimized output of a function of the determined age and sex and the biomarkers measured in the biological fluid.

[0010] The identity of each biomarker measured according to the present disclosure is both specific and critical to the accuracy of hepatocellular carcinoma detection, but for the sake of brevity, the biomarker combinations of the present invention are omitted from this disclosure.

[0011] Also disclosed are assays for detecting hepatocellular carcinoma in a subject, the assay comprising measuring the amount of one or more biomarkers in the biological fluid, wherein the assay utilizes a reagent to remove IgG and IgM proteins from the biological fluid of the subject prior to the measuring step; determining the age and sex of the subject; and determining the absence or presence of hepatocellular carcinoma in the subject using an optimized output of a function of the determined age and sex and the biomarkers measured in the biological fluid.

[0012] The present disclosure also relates to a kit for detecting hepatocellular carcinoma in a subject, the kit comprising a reagent for removing IgG protein from a biological fluid from the subject and a reagent for removing IgM protein from the biological fluid; reagents for respectively measuring one or more biomarkers in the biological fluid; and instructions for determining the absence or presence of hepatocellular carcinoma in the subject using an optimized output of a function of the determined age and sex and the biomarkers measured in the biological fluid.

[0013] Also provided herein are methods of assigning a subject to a group having a higher or lower probability of hepatocellular carcinoma, the method comprising removing IgG and IgM proteins from a biological fluid from the subject; measuring the amount of one or more biomarkers in the biological fluid; determining the age and sex of the subject; and assigning the subject to a group having a higher or lower probability of hepatocellular carcinoma based on an optimized output of a function of the determined age and sex and the biomarkers measured in the biological fluid.

[0014] The present disclosure also provides a method for the therapeutic management of a subject suspected of having hepatocellular carcinoma, the method comprising removing IgG and IgM proteins from a biological fluid from the subject; measuring the amount of one or more biomarkers in the biological fluid; determining the age and sex of the subject; and treating the subject based on an optimized output of a function of the determined age and sex and the biomarkers measured in the biological fluid, wherein when the subject is determined to have hepatocellular carcinoma based on the output of the function, the subject is treated for the disease. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figures 1A-1C Depicted are the results of the evaluation of a panel consisting of fucosylated low molecular weight kininogen, alpha-fetoprotein, aspartate aminotransferase (AST), alkaline phosphatase (ALK), age, and sex. Figure 1A The results of the clinical trials for differentiating all HCC patients (N=115) from all cirrhosis patients (N=93), differentiating early HCC (N=69) from cirrhosis (N=93), and differentiating AFP-negative (AFP-negative) patients were summarized. - , AFP levels were within the range observed in normal subjects, e.g., subjects without HCC) HCC (N=39) versus AFP-negative cirrhosis (N=84), as well as AUROC data distinguishing early-stage and AFP-negative HCC (N=29) from AFP-negative cirrhosis (N=84). Figure 1B Summarizes the sensitivity values ​​for differentiating the same corresponding patient groups in the same study at a 90% specificity cutoff, and Figure 1C Sensitivity values ​​for differentiating the same corresponding patient groups in the same study at a 95% specificity cutoff are summarized. DETAILED DESCRIPTION

[0016] The present invention may be more readily understood by reference to the following detailed description in conjunction with the accompanying drawings and examples which form a part of this disclosure. It is to be understood that these inventions are not limited to the specific methods, assays, kits, conditions or parameters described and / or shown herein, and that the terminology used herein is for illustrative purposes only and is not intended to limit the claimed invention.

[0017] The entire disclosure of each patent, patent application, and publication cited or described in this document is hereby incorporated by reference.

[0018] As used above and throughout the disclosure, the following terms and abbreviations shall be understood to have the following meanings unless otherwise indicated.

[0019] In this disclosure, references without specific quantitative indications include plural forms, and references to specific values ​​include at least that specific value unless the context clearly indicates otherwise. Thus, for example, reference to an "agent" refers to one or more of such agents and equivalents thereof known to those skilled in the art, and so forth. Furthermore, when it is indicated that an element "can be" X, Y, or Z, such usage is not intended to exclude other options for that element in all cases.

[0020] When a value is expressed as an approximation by using the antecedent "about," it should be understood that the particular value forms another embodiment. As used herein, "about X" (where X is a numerical value) preferably refers to ±10% of the recited value, inclusive. For example, the phrase "about 8" can refer to values ​​from 7.2 to 8.8, inclusive; as another example, the phrase "about 8%" can refer to values ​​from 7.2% to 8.8%, inclusive. All ranges, if any, are inclusive and combinable. For example, when reciting a range of "1 to 5," the recited range should be interpreted as optionally including the ranges "1 to 4," "1 to 3," "1-2," "1-2 and 4-5," "1-3 and 5," etc. In addition, when a list of optional options is affirmatively provided, such a list may also include embodiments in which any optional option may be excluded. For example, when describing a range of "1 to 5", such description can support the situation where any of 1, 2, 3, 4, or 5 are excluded; thus, the statement "1 to 5" can support "1 and 3-5, but not 2", or simply "wherein 2 is not included."

[0021] As used herein, the terms "treatment" or "therapy" (and their variant forms) include preventative (e.g., prophylactic), curative, palliative, or antiproliferative treatments. Such preventative, curative, palliative, or antiproliferative treatments may be complete or partial. For example, complete elimination of an unwanted symptom or partial elimination of one or more unwanted symptoms would represent "treatment" as contemplated herein.

[0022] The present disclosure relates particularly to methods, assays, and kits for detecting hepatocellular carcinoma, as well as methods for stratifying subjects into higher and lower risk categories for developing hepatocellular carcinoma, and methods for treating and managing subjects suspected of having or at risk for developing hepatocellular carcinoma. Unless otherwise indicated, steps, reagents, and factors (e.g., biomarkers) disclosed for one aspect of the disclosure (e.g., in conjunction with any of the methods, kits, and assays of the invention) can be used in conjunction with any other aspect of the disclosure.

[0023] Although previous work has attempted to address the need for highly sensitive early predictive indicators for hepatocellular carcinoma by evaluating one or more biological factors, none has approached the sensitivity required for clinically relevant determination of whether a subject, particularly an asymptomatic subject, has the disease. The inventors have discovered that certain combinations of factors meet this need by conferring previously unattainable levels of high accuracy. Importantly, the identified combinations can utilize factors that, on an individual basis, do not represent accurate indicators for the detection of hepatocellular carcinoma, can omit factors that are traditionally considered indicators of the disease, or both. These and other features of the subject matter of the present disclosure are described more fully herein.

[0024] Provided herein are methods for detecting hepatocellular carcinoma in a subject or determining a likelihood that a subject has hepatocellular carcinoma, the methods comprising removing IgG and IgM proteins from a biological fluid from the subject; measuring the amount of one or more biomarkers in the biological fluid; determining the age and sex of the subject; and determining the absence or presence of hepatocellular carcinoma in the subject using an optimized output of a function of the determined age and sex and the biomarkers measured in the biological fluid.

[0025] During the discovery process of the subject matter of the present disclosure, the inventors observed that it was impossible to identify certain glycoforms in the biological fluid from the subject, including glycoforms containing fucose. The inventors determined that some substances present in the subject sample produced nonspecific signals in the assay for identifying glycoforms, and ultimately found that the main contaminant was IgM. It was further discovered that removing both IgM and IgG removed the contaminating signals and allowed accurate analysis of specific protein glycoforms in a manner that was previously impossible. It has been unexpectedly determined that the steps of removing IgM and IgG and the reagents used to complete these steps represent clinically essential aspects of the method for early detection of hepatocellular carcinoma according to the methods, assays and kits of the present invention. In some embodiments, IgG is removed by incubating the biological fluid with protein A / G, which was found to be more effective than other tested strategies, and IgM is removed by passing the biological fluid through a filter. The filter can be, for example, a 1000 kD filter, a 900 kD filter, an 800 kD filter, a 500 kD filter, a 450 kD filter, a 400 kD filter, a 350 kD filter, a 300 kD filter, a 250 kD filter, a 200 kD filter, a 175 kD filter, a 150 kD filter, a 125 kD filter, a 100 kD filter, an 80 kD filter, a 75 kD filter, a 70 kD filter, a 60 kD filter, a 50 kD filter, a 40 kD filter, a 30 kD filter, a 25 kD filter, a 20 kD filter, a 15 kD filter, or a 10 kD filter. Filtration can be accomplished, for example, by gravity or centrifugation. Alternatively, IgG and IgM can be removed by precipitation with polyethylene glycol. According to such an embodiment, serum can be incubated with polyethylene glycol and then centrifuged, and the resulting supernatant represents the material in which the corresponding amount of target biomarker is measured.

[0026] Any biological fluid can be used according to the present invention. For example, the biological fluid can be whole blood, serum, plasma, urine, saliva, tears, mucus, or a combination of two or more of these fluids.

[0027] After removing IgG and IgM from the biological fluid, the amount of one or more biomarkers in the biological fluid is measured. The "amount" of a biomarker in the biological fluid can refer to its concentration in the biological fluid, for example, expressed in ng / mL, or, in the case of a fucosylated glycoprotein, can refer to the amount of fucosylated biomarker in the sample relative to the amount of non-fucosylated biomarker in the biological sample. In some embodiments, the amount of fucosylated biomarker can be expressed as a percentage relative to the total amount of biomarker in the sample.

[0028] Previous work has identified over 50 glycoproteins that exhibit increased fucosylation in subjects known to have hepatocellular carcinoma, cirrhosis, or both (Comunale, MA et al., J Proteome Res. 2006 Feb;5(2):308-15). Various published studies have evaluated the ability of individual fucosylated glycoproteins to serve as markers for hepatocellular carcinoma, including fetuin-A (also known as α-2-HS-glycoprotein) (Comunale, MA et al., J. Proteome Res. 2009;8;595-602), kininogen (Wang, M et al., Cancer Epidemiol Biomarkers Prev. 2009;18;1914-1921), α-1 antitrypsin (supra), hemopexin (Comunale, MA et al., 2009), α-1-antichymotrypsin (Comunale, MA et al., Proteomics Clin. Appl. 2013;7;690-700), GP73 (Drake, RR et al., Mol Cell Proteomics. 2006;5;1957-67), and LRAGG (lectin-reactive anti-α-galactose IgG) (Mehta, AS et al., J Cancer Res. 2009;18;1914-1921). Virol. 2008; 82; 1259-1270). Among these glycoproteins, the best performers include, for example, hemopexin (AUROC = 0.87) and fetuin-A (AUROC = 0.90). The present inventors have determined that certain combinations of factors (which may be referred to as panels), some of which are fucosylated glycoproteins, can be used to identify the presence of hepatocellular carcinoma with an accuracy significantly greater than that obtained by evaluating any single glycoprotein or other single factor. Unexpectedly, certain glycoproteins that serve as the best indicators of hepatocellular carcinoma when evaluated alone have only a moderate contribution or even a negative contribution to the panel of factors evaluated that include such proteins, etc. (e.g., one or more other fucosylated glycoproteins). Thus, it was found that previous work identifying individual indicators of hepatocellular carcinoma cannot be used to predict the contribution of these indicators to a panel of factors used to identify subjects with hepatocellular carcinoma.

[0029] According to the method of the present invention, the biomarkers whose corresponding amounts are measured can be one or more of the following: alkaline phosphatase, GP-73, hemopexin, HBsAg, hepatitis B virus particles, α-acid-glycoprotein, α-1-antichymotrypsin, α-1-antichymotrypsin His-Pro-less, α-1-antitrypsin, serum transferrin, plasma ceruloplasmin, α-2-macroglobulin, fetuin-A / α-2-HS-glycoprotein, alpha-fetoprotein, haptoglobin, fibrinogen gamma chain precursor, immunoglobulins (including, for example, IgA, IgD, IgE), APO-D, kininogen, histidine-rich glycoprotein, complement factor 1 precursor, complement factor I heavy chain, complement factor I light chain, complement C1, complement factor B precursor, complement factor B Ba fragment, complement factor B Bb fragment, complement C3 precursor, complement C3 beta chain, complement C3 alpha chain, C3a anaphylatoxin, complement, C3b alpha' chain, complement C3c fragment, complement C3dg fragment, complement C3g fragment, complement C3d fragment, complement C3f fragment, complement C5, complement C5 beta chain, complement C5 alpha chain, C5a anaphylatoxin, complement C5 alpha' chain, complement C7, beta-2-glycoprotein, vitamin D binding protein, inter-alpha-trypsin inhibitor heavy chain H2, alpha-1B-glycoprotein, angiotensinogen precursor, angiotensin-1, angiotensin-2, angiotensin-3, GARP protein, beta-2-glycoprotein, clusterin (Apo J), integrin alpha-8 precursor glycoprotein, integrin alpha-8 heavy chain, integrin alpha-8 light chain, hepatitis C virus particle, elf-5, kininogen, HSP33 homolog, lysyl endopeptidase, and leucine-rich repeat-containing protein 32 precursor. In some embodiments, the glycoforms of these glycoproteins are measured according to the methods of the present invention. For example, the N-linked glycosylation forms of these glycoproteins can be measured, such as the fucosylated N-linked glycoforms of the biomarkers, and in particular, the core-fucosylated N-linked glycoforms can be measured. Additional biomarkers can be one or more of alanine aminotransferase, aspartate aminotransferase, bilirubin, albumin, platelet count, or white blood cell count. In addition, according to the present disclosure, the change in the amount of any of the above-mentioned factors over time can represent a "biomarker."

[0030] In certain embodiments, the biomarkers whose corresponding amounts are measured are one or more of the following: alpha-fetoprotein, fucosylated fetuin-A, fucosylated hemopexin, fucosylated kininogen, fucosylated alpha-1-antitrypsin, aspartate aminotransferase (AST), alanine aminotransferase (ALT), and alkaline phosphatase (ALK). In one aspect, the method of the present invention comprises measuring alpha-fetoprotein, fucosylated kininogen, ALK, and AST. In another aspect, the method of the present invention comprises measuring alpha-fetoprotein, fucosylated kininogen, ALK, AST, and fucosylated alpha-1-antitrypsin. In yet another aspect, the method of the present invention comprises measuring alpha-fetoprotein, fucosylated kininogen, ALK, AST, and fucosylated fetuin-A. Alternatively, the method of the present invention may comprise measuring alpha-fetoprotein, fucosylated kininogen, ALK, AST, fucosylated fetuin-A, and fucosylated alpha-1-antitrypsin. Other combinations of biomarkers measured according to the methods of the present invention include: alpha-fetoprotein and fucosylated fetuin-A; alpha-fetoprotein and fucosylated hemopexin; alpha-fetoprotein and fucosylated kininogen; alpha-fetoprotein, fucosylated fetuin-A, and fucosylated hemopexin; alpha-fetoprotein, fucosylated fetuin-A, and fucosylated kininogen; alpha-fetoprotein, fucosylated alpha-1-antitrypsin, and fucosylated kininogen; and alpha-fetoprotein, fucosylated alpha-1-antitrypsin, fucosylated fetuin-A, and fucosylated kininogen.

[0031] According to the methods of the present invention, any acceptable method for determining the corresponding amount of the biomarker can be used. Total protein biomarkers, such as total alpha-fetoprotein, can be detected using conventional techniques, such as immunoassays. When the biomarker is glycosylated, the detection reagent can directly label the glycosyl moiety, for example, via a carbohydrate-specific chemical or dye, or via a labeled lectin, a labeled carbohydrate-binding protein, or a labeled antibody. The detection reagent can be a secondary reagent, for example, by first capturing the target analyte and then contacting the capture reagent-target complex with a labeled secondary reagent. In some embodiments, detection and quantification can be performed by isolating the glycosyl moiety from the protein, and then quantifying the glycosylation. In other embodiments, the glycoprotein can be isolated from a biological fluid, and then quantifying the glycosylation. These and other approaches are described in U.S. Publication No. 2012 / 0196277, filed April 10, 2012, the entire contents of which are incorporated herein by reference. When a lectin is used to detect glycosylated proteins, the lectin can be wild-type or a recombinant lectin with enhanced binding affinity for a specific target glycan. For example, U.S. Publication No. 2015 / 0198610, filed February 3, 2012, which is incorporated herein by reference, discloses various recombinant Aleuria aurantia lectins characterized by enhanced binding to core fucosyl moieties, which are known to be present to a greater extent on glycoproteins in subjects with hepatocellular carcinoma compared to glycoproteins in subjects without the disease.

[0032] The corresponding amounts of ALT and AST can be determined using kinetic assays based on the corresponding enzymatic activities of AST and ALT, using colorimetric, spectrophotometric, chemiluminescent, chromatographic, fluorescent or UV absorbance, radiochemical and electrochemical techniques for detection. Those of ordinary skill in the art are familiar with these assays and techniques.

[0033] According to the method of the present invention, the age and sex of the subject are determined. The age of the subject can be expressed in years, months, or days as needed. Age and sex have previously been shown to be relevant factors in subjects with hepatocellular carcinoma (see Wang, M. et al., Proceedings. IEEE International Conference on Bioinformatics and Biomedicine 2012; Wang, M. et al., BMC Medical Genomics, 2013, 6 Suppl 3: S9).

[0034] After measuring the corresponding amount of a specific biomarker and determining the age and sex of the subject, the presence or absence of hepatocellular carcinoma is determined using the optimized output of the function of the determined age and sex and the corresponding measured amount of the biomarker. For example, the function can include corresponding optimized weighting coefficients for the determined age and sex and for the measured amount of the biomarker. Appropriate statistical methods can be used to identify the optimized function of age, sex, and biomarker amount. For example, logistic regression can be used for each desired factor group (i.e., biomarker, age, and sex) to generate a logistic regression algorithm. In order to judge the goodness of fit of each regression, one or more of the following can be derived: AIC, R 2 , Dxy, likelihood ratio test, Pearson goodness of fit, log likelihood, deviance statistic, Tau-a, NRI, and area under the ROC curve (AUC) of apparent validation (see Steyerberg, E. (2009), "Clinical Prediction Models", Springer-Verlag New York). Based on these criteria and tests, the logistic regression model with the highest goodness of fit can be selected for further evaluation. Optionally, in order to avoid overfitting, one or more of leave-one-out cross-validation, bootstrap-validation, and 3-fold cross-validation can be applied to validate the candidate model.

[0035] The optimized function of age, sex, and biomarker amount produces a value representing the probability that the subject has hepatocellular carcinoma. According to the method of the present invention, the value can reveal that the subject has hepatocellular carcinoma at a clinically relevant level of certainty. For example, the level of certainty can be about 50% or higher, 60% or higher, 70% or higher, 75% or higher, 80% or higher, 85% or higher, 90% or higher, about 91% or higher, about 92% or higher, about 93% or higher, about 94% or higher, about 95% or higher, about 96% or higher, about 97% or higher, or about 98% or higher. Therefore, the method of the present invention can be used to determine whether a subject has hepatocellular carcinoma with a high degree of certainty.

[0036] As used herein, "detecting hepatocellular carcinoma" or "determining the absence or presence of hepatocellular carcinoma" can include detecting a type or stage of hepatocellular carcinoma in a subject that is not normally detectable by methods that exist at the time of this disclosure. For example, the methods, assays, and kits of the present disclosure can be used to determine whether a subject has early-stage hepatocellular carcinoma; a negative alpha-fetoprotein (AFP) -) hepatocellular carcinoma; or early hepatocellular carcinoma and negative alpha-fetoprotein (AFP - ) Hepatocellular carcinoma (ie, early stage and AFP - of HCC).

[0037] "AFP negative - ) hepatocellular carcinoma refers to a disease that is not characterized by AFP levels that are significantly different from the AFP levels found in subjects without liver disease. For example, a subject may be said to have AFP when the amount of AFP in a biological sample from the subject is less than about 20 ng / mL. - .

[0038] Significantly, the data provided in the present disclosure, such as the data in Tables 3A and 3B, demonstrate the ability of the methods, assays, and kits of the present disclosure to provide a system that can detect early-stage and / or alpha-fetoprotein negative hepatocellular carcinoma in a subject. Appropriate treatment can then be administered. The high clinical significance of this aspect of the present subject matter will be apparent to one of ordinary skill in the art, particularly given that early-stage hepatocellular carcinoma is curable (whereas advanced disease is typically only amenable to palliative treatment). Unlike the presently disclosed subject matter, previous systems for detecting hepatocellular carcinoma are unable to detect early-stage hepatocellular carcinoma, alpha-fetoprotein negative (AFP negative) hepatocellular carcinoma, and / or alpha-fetoprotein negative (AFP negative) hepatocellular carcinoma. - ) hepatocellular carcinoma, or early hepatocellular carcinoma and negative alpha-fetoprotein (AFP - )Hepatocellular carcinoma.

[0039] Thus, the present disclosure also provides methods for treating or recommending treatment for a subject specifically for early-stage hepatocellular carcinoma, alpha-fetoprotein-negative hepatocellular carcinoma, or early-stage hepatocellular carcinoma that is also alpha-fetoprotein-negative hepatocellular carcinoma. Such methods can be adjunct to the disclosed methods for detecting hepatocellular carcinoma, or can be the result of using the assays and kits disclosed herein. Appropriate treatment specifically tailored for early-stage hepatocellular carcinoma, alpha-fetoprotein-negative hepatocellular carcinoma, or early-stage hepatocellular carcinoma that is also alpha-fetoprotein-negative hepatocellular carcinoma can be identified by a practicing physician.

[0040] An exemplary method may include removing IgG and IgM proteins from a biological fluid from a subject suspected of having early-stage hepatocellular carcinoma, AFP-negative hepatocellular carcinoma, or both; measuring the amount of one or more biomarkers in the biological fluid; determining the age and sex of the subject; determining the absence or presence of early-stage hepatocellular carcinoma, AFP-negative hepatocellular carcinoma, or both in the subject using an optimized output of a function of the determined age and sex and the biomarkers measured in the biological fluid; and, if the presence of early-stage hepatocellular carcinoma, AFP-negative hepatocellular carcinoma, or both is determined in the subject, treating the subject or providing a treatment recommendation for the subject, wherein the treatment is specifically selected for treating early-stage hepatocellular carcinoma, AFP-negative hepatocellular carcinoma, or both. Aspects of such methods involving removing IgG and IgM proteins from a biological fluid, measuring the amount of one or more biomarkers in the biological fluid, determining the age and sex of the subject, and determining the absence or presence of disease using an optimized output of a function of the determined age and sex and the biomarkers measured in the biological fluid can be performed according to corresponding aspects of the previously disclosed methods for detecting the absence or presence of hepatocellular carcinoma.

[0041] Also disclosed are assays for detecting hepatocellular carcinoma in a subject, the assay comprising measuring the amount of one or more biomarkers in the biological fluid, wherein the assay utilizes a reagent to remove IgG and IgM proteins from the biological fluid of the subject prior to the measuring step; determining the age and sex of the subject; and determining the absence or presence of hepatocellular carcinoma in the subject using an optimized output of a function of the determined age and sex and the biomarkers measured in the biological fluid.

[0042] The reagent for removing IgG from biological fluids can be protein A / G. In other embodiments, the reagent can be protein L. When used in the assays, methods and kits disclosed herein, the "reagent" for removing IgM can be an item of laboratory equipment or a mechanical system. For example, the reagent for removing IgM can be a filter or filtration system based on molecular weight. For example, filtration can be performed by gravity-based or centrifugal filtration. In other embodiments, the reagent for removing both IgG and IgM can be polyethylene glycol, non-limiting examples of which include PEG-400, PEG-3350, PEG-4000, PEG-6000 and PEG-8000.

[0043] In order to quantitatively detect the one or more biomarkers, any of a variety of different assay formats can be used according to the assay method of the present invention. Immunoassay is a preferred assay method and includes but is not limited to ELISA, radioimmunoassay, competitive assay, Western blot, bead aggregation assay, lateral flow immunoassay, immunochromatographic test strips, test strips, migration form immunoassay, etc. Other suitable immunoassays are known to those skilled in the art. Microscopy can also be used. In some embodiments, chromatography represents the selected assay format. High performance liquid chromatography (HPLC) is particularly preferred. In some embodiments, mass spectrometry is a preferred assay method. In some embodiments, gel electrophoresis combined with densitometry is used as an assay format.

[0044] The general format of the assay can include contacting a suitable reagent with a test sample containing a target analyte, i.e., a biomarker, that can be distinguished from other components found in the sample. The sample can be a biological fluid from the subject, or it can be a separate sample derived from the biological fluid. After the analyte interacts with the reagent, the system can be washed and then detected directly or by a secondary reagent.

[0045] In some embodiments, the reagents for detecting biomarkers are immobilized on a solid support. In other preferred embodiments, the test sample or molecules separated or purified from the test sample, such as post-translationally modified polypeptides, are immobilized on a solid support. Techniques for purifying biomolecules from samples such as cells, tissues, or biological fluids are well known in the art. The selected technique may vary with the tissue or sample being examined, but matching the appropriate purification procedure to the test sample source is well within the scope of the art.

[0046] Examples of suitable solid supports include, but are not limited to, glass, plastic, metal, latex, rubber, ceramic, polymers such as polypropylene, polyvinylidene fluoride, polyethylene, polystyrene and polyacrylamide, dextran, cellulose, nitrocellulose, PVDF, nylon, amylase, etc. The solid support can be planar, concave or convex, spherical, cylindrical, etc., and can be a particle, bead, membrane, strip, precipitate, gel, sheet, container, well, capillary, film, plate, slide, etc. The solid support can be magnetic, or a column.

[0047] Other aspects of the assays of the invention comprising the steps of measuring the amount of one or more biomarkers in a biological fluid, determining the age and sex of the subject, and determining the absence or presence of hepatocellular carcinoma using the determined age and sex and an optimized output of a function of the biomarkers measured in the biological fluid can be performed according to any of the aspects described above for the methods of the invention for detecting hepatocellular carcinoma in a subject.

[0048] The present disclosure also relates to a kit for detecting hepatocellular carcinoma in a subject, the kit comprising a reagent for removing IgG protein from a biological fluid of the subject, and a reagent for removing IgM protein from the biological fluid; reagents for respectively measuring one or more biomarkers in the biological fluid; and instructions for determining the absence or presence of hepatocellular carcinoma in the subject using an optimized output of a function of the determined age and sex and the biomarkers measured in the biological fluid.

[0049] In some embodiments, the reagent that is used to remove IgG from biological fluid can be protein A / G. In other embodiments, the reagent can be protein L. In some embodiments, the reagent that is used to remove IgM from biological fluid can be a filter based on molecular weight, for example a 1000kD filter, a 900kD filter, an 800kD filter, a 500kD filter, a 450kD filter, a 400kD filter, a 350kD filter, a 300kD filter, a 250kD filter, a 200kD filter, a 175kD filter, a 150kD filter, a 125kD filter, a 100kD filter, an 80kD filter, a 75kD filter, a 70kD filter, a 60kD filter, a 50kD filter, a 40kD filter, a 30kD filter, a 25kD filter, a 20kD filter, a 15kD filter, or a 10kD filter. Filtration can for example be accomplished by gravity or centrifugal.

[0050] In certain embodiments, the reagent used to remove IgG and the reagent used to remove IgM can be the same. An exemplary embodiment of this class is polyethylene glycol.

[0051] When the biomarker is a polypeptide, such as a glycoprotein, the reagent for measuring the one or more biomarkers in the biological fluid can take advantage of the fact that the polypeptide contains one or more post-translational modifications (e.g., glycosylation). Detection of post-translational modifications can be accomplished by detecting a certain polypeptide-part complex. In another preferred embodiment, detection of post-translational modifications can be performed by separating the polypeptide and the part and detecting the part. Detection of the polypeptide-part complex can be performed using a reagent that specifically recognizes the part, a particular class of parts, or as part of a complex with the polypeptide. Suitable detection reagents are apparent to those skilled in the art, and non-limiting examples are described below. The reagent may comprise a plurality of molecules, each molecule having specificity for a different target part, resulting in multiple reagent-target interactions.

[0052] The reagent used to detect the biomarker can be an antibody. For example, any antibody that specifically binds to a target moiety of interest can be used in accordance with the kit of the present invention. Monoclonal and / or polyclonal antibodies produced from any source can be used, as can recombinant antibodies such as single-chain antibodies and phage-displayed antibodies, as well as chimeric and humanized antibodies. Antigen-binding fragments of antibodies, such as Fab or Fv, can also be used.

[0053] In some embodiments, the reagent used to detect the biomarker is an antibody that specifically recognizes a glycoprotein. Preferably, the antibody specifically recognizes a carbohydrate moiety, including monosaccharides and polysaccharides. In some cases, the antibody specifically recognizes a fucose moiety, such as a core fucosyl moiety. Antibodies that can specifically recognize fucose have been described. See, for example, Roy SS et al., (2002) Ann. Bot. 89: 293-9; and, Srikrishna G et al., (1998) Glycobiology 8: 799-811. Alternatively, antibodies against various moieties, including fucose, can also be generated and used in the present invention. Methods for generating and purifying antibodies are well known in the art. In addition, monoclonal antibodies can be prepared by many techniques known in the art, including the technique originally developed by Kohler and Milstein (1975) Nature 256: 495-497.

[0054] Other proteins with carbohydrate recognition domains can also be used as reagents for detecting biomarkers. Proteins with carbohydrate recognition domains have been described, see for example Bouyain S et al., (2002) J. Biol. Chem. 277: 22566-72 (Drosophila melanogaster protein CG2958 that recognizes fucose).

[0055] In a particularly preferred embodiment, lectins are used as reagents for detecting biomarkers. Lectins can be obtained from any organism, including plants, animals, yeast, bacteria, protozoa, and the like. Purified lectins are commercially available, see, for example, the Sigma-Aldrich catalog (St. Louis, MO). Lectins can also be isolated from their naturally occurring sources, or recombinantly expressed and purified, by means known to those skilled in the art. Lectins can, but need not, be specific for a particular carbohydrate moiety. Fucose-specific lectins have been described. See, for example, Mansour MH et al. (2005) Immunobiology. 210: 335-48; Amano K et al. (2003) Biosci. Biotechnol. Biochem. 67: 2277-9; Loris R et al. (2003) J. Mol. Biol. 331: 861-70; and Ishida H et al. (2002) Biosci. Biotechnol. Biochem. 66: 1002-8. U.S. Publication No. 2015 / 0198610, filed February 3, 2012, which is incorporated herein by reference, discloses various recombinant Aleuria aurantia lectins characterized by enhanced binding to a core fucosyl moiety. It is contemplated that lectins identified or developed in the future are suitable for use as reagents for detecting one or more biomarkers according to the present disclosure.

[0056] Proteins having a lectin-like domain are also suitable for use as reagents for detecting the one or more biomarkers. Proteins having a lectin-like domain are known in the art. See, for example, Drickamer K (1999) Curr. Opin. Struct. Biol. 9: 585-90.

[0057] Nucleic acid-based lectin alternatives can also be used. Such reagents are called aptamers, which utilize the huge conformational flexibility of single-stranded nucleic acids. From a large library of random short nucleic acids, individual molecules with high affinity for many non-nucleic acid ligands have been isolated by iterative selection. The advantage of polysaccharide binding "lectamer" reagents is that the leached DNA is unlikely to mess up or interfere with downstream analysis. The lectamer can work under uniform binding conditions (pH, ionic strength). Synthetic nucleic acids can be prepared into various derivatized forms (e.g., terminal biotinylation). Target polysaccharides are not limited by the specificity of existing lectins, significantly expanding existing classification and analytical capabilities.

[0058] The reagent used to detect the biomarker can be directly labeled with a detectable moiety. Alternatively, a secondary reagent that specifically recognizes the primary reagent and is labeled with a detectable moiety can be used. The secondary reagent can be any molecule, such as an antibody. The secondary reagent is labeled with a detectable moiety. The detectable moieties contemplated for the present invention include, but are not limited to, radioisotopes, fluorescent dyes such as fluorescein, phycoerythrin, Cy-3, Cy5, allophycocyanin, DAPI, Texas Red, rhodamine, Oregon Green, fluorescein yellow, green fluorescent protein, red fluorescent protein, cyan fluorescent protein, yellow fluorescent protein, cerianthus orange fluorescent protein, alkaline phosphatase, beta-lactamase, chloramphenicol acetyltransferase, adenosine deaminase, aminoglycoside phosphotransferase (neor, G418r) dihydrofolate reductase, hygromycin-B-phosphotransferase, thymidine kinase, lacZ (encoding α-galactosidase) and xanthine guanine phosphoribosyltransferase, beta-glucuronidase, placental alkaline phosphatase, secretory embryonic alkaline phosphatase, or firefly or bacterial luciferase. Enzyme tags are used together with their associated substrates. As with other standard procedures relevant to the practice of the present invention, the skilled person will appreciate operable other labels. In some embodiments, the reagent or secondary reagent is conjugated to biotin and contacted with avidin or streptavidin tagged with a detectable moiety.

[0059] In some embodiments, the portion that is linked to the polypeptide biomarker through a post-translational modification can be directly labeled and detected, thereby avoiding the need for a labeled reagent that specifically recognizes a particular portion and any need for a labeled secondary reagent. For the purposes of this disclosure, detection of a biomarker includes detecting the portion that is linked to the biomarker through a post-translational modification. In some embodiments, the portion that is linked to the biomarker through a post-translational modification can be separated from the biomarker and directly labeled and detected. For example, but not limited to, carbohydrates and carbohydrate moieties can be directly labeled using various methods known in the art. Kits for labeling carbohydrates and carbohydrate moieties are commercially available. Carbohydrates and carbohydrate moieties can also be biotinylated and labeled with a detectable moiety that is conjugated to avidin or streptavidin, such as those described herein. Non-limiting examples of reagents that can directly label oligosaccharides include 2-aminobenzamide and 2-aminobenzoic acid.

[0060] Post-translationally linked moieties can be separated from polypeptide biomarkers by any means suitable in the art, including chemical means, such as treatment with hydrazine or acids such as hydrofluoric acid or trifluoromethanesulfonic acid; enzymatic means, such as treatment with N-glycosidases such as PNGase F, O-glycosidases, endoglycosidases, or exoglycosidases; or physical means. Commercially available kits are available for removing post-translational modifications, including deglycosylation. Chemical bases such as hydrazine or chemical reagents that induce β-elimination reactions can also be used for deglycosylation reactions. Other techniques and reagents are within the skill of the art and are contemplated within the scope of this disclosure.

[0061] In some embodiments, the isolated, previously post-translationally linked fraction is purified prior to labeling or detection. Solid or liquid phase extraction techniques known in the art can be used to purify the isolated fraction for further analysis.

[0062] As described above, total protein biomarkers such as total alpha-fetoprotein can be detected using conventional techniques such as immunoassays, and the reagents required for performing such techniques can be included in the kits of the present invention. The same is true for kinetic assays based on the corresponding enzyme activities of AST and ALT using colorimetric, spectrophotometric, chemiluminescent, chromatographic, fluorescent or UV absorbance, radiochemical, and electrochemical techniques for detection.

[0063] The kit of the present invention includes instructions for determining the presence or absence of hepatocellular carcinoma in a subject using an optimized output of a function of the determined age and sex and the measured amount of a biomarker in a biological fluid. The function can be provided according to the description provided above for the method of the present invention for detecting hepatocellular carcinoma in a subject. The present disclosure also provides specific examples of how appropriate functions of the determined age and sex and the measured amount of a biomarker can be generated.

[0064] Also provided herein are methods of assigning a subject to a group having a higher or lower probability of hepatocellular carcinoma, the method comprising removing IgG and IgM proteins from a biological fluid from the subject; measuring the amount of one or more biomarkers in the biological fluid; determining the age and sex of the subject; and assigning the subject to a group having a higher or lower probability of hepatocellular carcinoma based on an optimized output of a function of the determined age and sex and the biomarkers measured in the biological fluid.

[0065] As provided above in accordance with the disclosed methods for detecting hepatocellular carcinoma in a subject, the optimized function of age, sex, and biomarker amount generates a value representing the probability that the subject has hepatocellular carcinoma. This value can indicate with a certainty of about 90% or greater, about 91% or greater, about 92% or greater, about 93% or greater, about 94% or greater, about 95% or greater, about 96% or greater, about 97% or greater, or about 98% or greater that the subject has hepatocellular carcinoma. According to the disclosed methods for assigning subjects to groups with a higher or lower probability of hepatocellular carcinoma, a clinically appropriate determination can be made as to whether the value generated by the optimized function should assign the subject to a group with a higher or lower probability of hepatocellular carcinoma. For example, when the value generated by the optimized function of age, sex, and biomarker amount indicates that the subject has hepatocellular carcinoma and the level of certainty associated with the function is clinically significant, the subject can be assigned to the "higher probability" group. For example, the level of certainty can be at least 50% higher, at least 60% higher, at least 65% higher, at least 70% higher, at least 75% higher, at least 80% or higher, at least 82% or higher, at least 84% or higher, at least 85% or higher, at least 86% or higher, at least 88% or higher, or at least 90% or higher. An experienced pathologist, in collaboration with a biostatistician, bioinformatician, or the like, can apply appropriate statistical methods to determine whether the values ​​generated by a particular optimized function of age, sex, and biomarker amount assign the subject to a group with a higher or lower probability of liver cancer.

[0066] The present disclosure also provides a method for the therapeutic management of a subject suspected of having hepatocellular carcinoma, the method comprising removing IgG and IgM proteins from a biological fluid from the subject; measuring the amount of one or more biomarkers in the biological fluid; determining the age and sex of the subject; and treating the subject based on an optimized output of a function of the determined age and sex and the biomarkers measured in the biological fluid, wherein when the subject is determined to have hepatocellular carcinoma based on the output of the function, the subject is treated for the disease.

[0067] After applying appropriate statistical evaluation, when the output of the optimization function of the determined age and sex of the object and the biomarker measured in the biological fluid indicates that there is hepatocellular carcinoma in the object, the object can be treated according to the method of the present invention. When the object is asymptomatic, the method of the present invention advantageously allows to start a treatment plan that takes into account the early stage hepatocellular carcinoma in the object. The treatment of early stage hepatocellular carcinoma can be carried out according to generally accepted medical practice. When it is determined that the object does not suffer from hepatocellular carcinoma according to the method of the present invention, the object can be treated for any optional or precursor disease that may be applicable. For example, when the object is suspected of suffering from hepatocellular carcinoma, it may be known that the object has shown or is subsequently found to have (for example, as a result of a test prompted by a negative diagnosis of hepatocellular carcinoma) any common risk factor, such as alcoholism, hepatitis B or hepatitis C, aflatoxin, cirrhosis, non-alcoholic steatohepatitis, hemochromatosis, Wilson's disease (Wilson'sdisease), type 2 diabetes, NASH (non-alcoholic steatohepatitis) or hemophilia. Treatment of any one or more conditions representing risk factors for hepatocellular carcinoma can be performed according to appropriate measures determined by one of ordinary skill in the art. In addition, if a subject is determined not to have hepatocellular carcinoma, the subject may be required to submit to regular monitoring, which may include testing the subject for hepatocellular carcinoma on one or more future occasions. For the purposes of the methods of the present invention, "treatment" may include regular monitoring of the subject to confirm whether the subject develops hepatocellular carcinoma after the initial determination that the subject does not have hepatocellular carcinoma. For example, according to the methods of the present invention, if a subject is determined not to have hepatocellular carcinoma based on the output of an optimization function of determined age and sex and biomarkers measured in a biological fluid, one or more additional series of tests may be performed on the subject to determine whether the subject subsequently develops hepatocellular carcinoma. Additional testing may be performed, for example, two months, three months, six months, eight months, one year, 18 months, two years, 30 months, or three years after an initial negative diagnosis of hepatocellular carcinoma, and, for example, every three months, every six months, every year, every 18 months, every two years, every 30 months, every three years, every four years, or every five years until the subject is found to have hepatocellular carcinoma. The additional testing can be performed according to the methods of the present invention for detecting hepatocellular carcinoma in a subject.

[0068] Example

[0069] The following examples are set forth in order to provide those of ordinary skill in the art with a complete disclosure and description of how the methods, kits, and assays claimed herein are developed and evaluated, and are intended to be purely exemplary of the invention and are not intended to limit the scope of what the inventors regard as their invention. Efforts have been made to ensure accuracy with respect to numbers (e.g., amounts), but some errors and deviations should be accounted for.

[0070] Example 1 - Clinical data analysis and identification of HCC predictive markers

[0071] The development of the present predictive index was completed using data from two clinical cohorts: one clinical cohort consisting of 115 HCC patients and 93 cirrhosis patients (liver function test data were available for this cohort) and the other clinical cohort consisting of 66 HCC patients and 38 cirrhosis patients (liver function test data were not available for this cohort).

[0072] Twenty-four feature selection algorithms were used to explore each potential panel constituent (serum biomarkers; patient demographic factors) and combination. The feature selection algorithms used included: correlation feature selection filtering (cfs filtering), chi-square filtering, consistency-based filtering, linear correlation filtering, rank correlation filtering, information gain filtering, information ratio filtering, systematic uncertainty, exhaustive search algorithm, greedy forward search, greedy backward search, hill climbing search, oneR algorithm, random forest filtering, relief filtering with 3 nearest neighbors, filter filtering with 5 nearest neighbors, stepwise logistic regression, stepwise penalized logistic regression, penalized svm, bestglm (bic), bestglm (aic), robust logistic regression, Bayesian fractional polynomials of GLM (glmBfp), and Bayesian model averaging (BMA). Subsequently, logistic regression was applied to the subset of predictors provided by the 24 feature selection algorithms. Twenty-one subset features were selected from the feature selection algorithms and market basket analysis. To this end, all predictors and AFP were added to the discussion subset, and 23 logistic regression algorithms were constructed. In order to judge the goodness of fit of each regression, the inventors derived AIC, R 2 , Dxy, likelihood ratio test, Pearson goodness of fit, log likelihood, deviation statistic, tau-a, NRI, and area under the ROC curve (AUROC) of apparent validation. Based on these criteria and tests, the four logistic regression models with the highest goodness of fit were selected for further evaluation. To avoid overfitting, leave-one-out cross validation, bootstrap validation, and three-fold cross validation were applied to validate the four candidate models.

[0073] Fourteen biomarker, clinical, and demographic variables were assessed: alanine aminotransferase (ALT), aspartate aminotransferase (AST), bilirubin (BIL), albumin (ALB), platelets (PLT), alkaline phosphatase (ALK), white blood cells (WBC), alpha-fetoprotein (AFP), fucosylated A1AT, fucosylated low-molecular-weight (LMW) kininogen, fucosylated fetuin-A, fucosylated hemopexin, age, and sex. AFP values ​​were distributed over a wide range (1 to 347,000 ng / ml) and were log-transformed for further analysis. The raw data were used for the remaining variables, as their distribution did not affect statistical analysis and algorithm exploration.

[0074] Preliminary analysis showed that AFP, ALK, AST, age, and male sex were positively correlated with the probability of HCC. The male ratio was higher in the HCC group, with an odds ratio of 1.75 and a χ 2 =5.36, df=1, p=0.02056. The inclusion of AFP yielded a significantly higher AUROC value for distinguishing HCC. When AFP was used without fucosylation biomarkers, the AUROC value varied from 0.8016 for AFP alone to 0.9066 when four additional factors (age, sex, ALK, and ALT) were included. Increasing the number of variables did not alter the AUROC beyond this value, suggesting the need to include new biomarkers to drive detection.

[0075] To this end, an evaluation of a panel containing various biomarkers was performed. The combination of AFP, fucosylated low molecular weight (LMW) kininogen, fucosylated A1AT, age, and sex produced an AUROC of 0.9405. Adding AST and ALK increased the AUROC to 0.9835. Eliminating fucosylated A1AT slightly decreased the AUROC to 0.9826. This latter minimal panel, consisting of fucosylated LMW kininogen, AFP, AST, ALK, age, and sex, yielded the following algorithm A and was used to generate the following: Figures 1A-1C Data for patient subsets shown (divided by early-stage and AFP-negative disease).

[0076] Algorithm A

[0077]

[0078] Figures 1A-1C The results of the evaluation of a panel consisting of fucosylated low molecular weight kininogen, alpha-fetoprotein, aspartate aminotransferase (AST), alkaline phosphatase (ALK), age, and sex are depicted. The panel was evaluated in one of three different cohorts of HCC and cirrhosis patients. The three patient cohorts each had the characteristics described in Table 1 below:

[0079] Table 1

[0080]

[0081] In the population used to generate the data for Figure 1, the causes of HCC were 61% HCV, 6% HBV, and 33% other, and the causes of cirrhosis were 48% HCV, 10% HBV, and 42% other. Figure 1A Data were summarized for the area under the receiver operating characteristic curve (AUROC) for distinguishing all HCC patients (N=115) from all cirrhotic patients (N=93), distinguishing early-stage HCC (N=69) from cirrhosis (N=93), distinguishing AFP-negative HCC (N=39) from AFP-negative cirrhosis (N=84), and distinguishing early-stage and AFP-negative HCC (N=29) from AFP-negative cirrhosis (N=84). The 95% confidence intervals (CIs) were non-overlapping in all cases. Figure 1B Summarizes the sensitivity values ​​for differentiating the same corresponding patient groups in the same study at a 90% specificity cutoff, and Figure 1C Sensitivity values ​​for differentiating the same corresponding patient groups in the same study at a 95% specificity cutoff are summarized.

[0082] As described above, incorporating AST and ALK into the developed model increased the AUC from 0.9405 to 0.9835. Inclusion of these two clinical variables unexpectedly offset the need to include one of the fucosylation biomarkers, in this case fucosylated A1AT: When AST and ALK were added to the equation and fucosylated A1AT was discarded from the algorithm, the AUC remained at 0.9826, suggesting that a highly predictive base model incorporating age, sex, kininogen, ALK, and AST was sufficient.

[0083] Previous work identified fucosylated fetuin-A and fucosylated hemopexin as very good independent indicators of hepatocellular carcinoma (Comunale, MA et al., J. Proteome Res. 2009; 8; 595-602). The contribution of these independent fucosylation biomarkers to a panel that also included age, sex, and AFP (log) was evaluated. Algorithm B shown below was used to test the panel including fucosylated fetuin-A and fucosylated hemopexin:

[0084] Algorithm B

[0085]

[0086] Separate testing of the same panel members (panel 4, shown in Table 2 below) using another algorithm yielded an AUC of 0.8461. When kininogen was substituted for hemopexin (panel 5b in Table 2), the AUC increased to 0.8661.

[0087] A summary of the evaluation results for the panel including age, alpha-fetoprotein, and other specified biomarkers is provided in Table 2 below.

[0088] Table 2

[0089]

[0090]

[0091] Tables 3A and 3B below provide the AUROC (95% confidence interval), sensitivity (95% confidence interval) at a specified fixed specificity, and positive (LR+) and negative (LR-) likelihood ratios for discriminating the following patients by Group 1 (as shown in Table 2 above): (1) HCC (N=115) versus cirrhosis (N=93); (2) early HCC (UNOS T1 / 2; N=69) versus cirrhosis (N=93); (3) AFP-negative HCC (<20 ng / mL; N=39) versus AFP-negative cirrhosis (<20 ng / mL; N=84); and (4) early (UNOS T1 / 2) and AFP-negative HCC (<20 ng / mL; N=29) versus AFP-negative cirrhosis (<20 ng / mL; N=84). [HCC etiology (%) = HCV (61), HBV (6), other (33); HCC gender (M / F%) = 73 / 27; Cirrhosis etiology (%) = HCV (48), HBV (10), other (42); Cirrhosis gender (M / F%) = 51 / 49]. These data were obtained using the HCC group score (P value) obtained directly from algorithm A.

[0092] Table 3A

[0093]

[0094] Table 3B

[0095]

[0096] Tables 4A and 4B below provide the AUROC (95% confidence interval) and sensitivity (95% confidence interval) at a specified fixed specificity for distinguishing the following patients by panel 11 (as shown in Table 2 above): (1) HCC (N=115) from cirrhosis (N=93); (2) early HCC (UNOS T1 / 2; N=69) from cirrhosis (N=93); (3) AFP-negative HCC (<20 ng / mL; N=39) from AFP-negative cirrhosis (<20 ng / mL; N=84); and (4) early (UNOS T1 / 2) and AFP-negative HCC (<20 ng / mL; N=29) from AFP-negative cirrhosis (<20 ng / mL; N=84). [HCC etiology (%) = HCV (61), HBV (6), other (33); HCC gender (M / F%) = 73 / 27; Cirrhosis etiology (%) = HCV (48), HBV (10), other (42); Cirrhosis gender (M / F%) = 51 / 49]. These data were obtained using algorithm A (but using ALT values ​​instead of AST values) and three-fold cross validation.

[0097] Table 4A

[0098]

[0099] Table 4B

[0100]

[0101] When reviewing the overall data, it was surprising that the variables that contributed to the highest performing groups tended to be those that were least effective as indicators of hepatocellular carcinoma on a standalone basis. For example, hemopexin produced the highest AUC value when used alone (0.8695 in patient set b), and A1AT (patient set a), AFP (patient set a), and kininogen (patient set a) produced the lowest AUC values ​​when used alone (0.7395, 0.8346, and 0.8192, respectively), but the latter contributed to groups that performed higher than those containing hemopexin. From these data, it is learned that the contribution of a particular biomarker to a group cannot be predicted based on its independent ability to determine the presence of hepatocellular carcinoma.

[0102] The data also revealed the ability of the tested panel to allow detection of early stage hepatocellular carcinoma, AFP negative hepatocellular carcinoma and early stage hepatocellular carcinoma that is also AFP negative hepatocellular carcinoma.

[0103] Example 2 - Identification of IgG and IgM as assay contaminants

[0104] Plate-based assays for detecting hepatocellular carcinoma are often hampered by the presence of heterophilic antibodies and potential other lectin-binding contaminants. Such contaminants are known to lead to false-positive results. A study was conducted to identify contaminating lectin-reactive factors present in the serum of patients with cirrhosis and liver cancer. As described below, the inventors identified IgM as a contaminating lectin-reactive factor, and when IgM was removed from the serum prior to lectin-ELISA, a lectin-reactive signal associated with the specific protein could be detected. This approach was used in two independent sample sets to validate the method and also to verify the performance of fucosylated glycoforms as biomarkers for hepatocellular carcinoma.

[0105] method

[0106] Serum samples were obtained from the blood of each test subject, and demographic and clinical information of the patient population was also obtained. Consecutive HCC patients were enrolled in the study, as well as cirrhotic patients whose age, sex and race / ethnicity matched those of the HCC patients. The diagnosis of HCC was made by histopathology including all T1 lesions, and if histopathology was unavailable, two imaging modalities (ultrasound, magnetic resonance imaging or computed tomography) were used to show a vascular enhancement mass > 2 cm. The diagnosis of cirrhosis was based on liver histology or clinical, laboratory and imaging evidence of liver decompensation or portal hypertension. Each cirrhotic patient had a normal ultrasound, and if serum AFP was elevated, no liver mass was shown by liver MRI within 3 months before enrollment and another MRI 6 months after enrollment. After enrollment, the median follow-up of cirrhotic controls was 12 months (range: 7-18 months), and no one developed HCC. The TNM staging system revised by the United Network for Organ Sharing was used to determine tumor stage for HCC. Early-stage HCC was defined as T1 (single lesion <2 cm in diameter) and T2 (single lesion between 2 and 5 cm in diameter; or <3 lesions, each <3 cm in diameter) lesions that met US liver transplant criteria. A 20 ml blood sample was drawn from each subject, centrifuged, aliquoted, and the resulting serum stored at -80°C until testing. Blood samples were drawn before starting HCC treatment. A commercially available immunoassay was used to test for AFP using enhanced chemiluminescence.

[0107] Lectin FLISA

[0108] Briefly, to remove fucosylation from the capture antibody (mouse anti-human AAT or rabbit anti-fetoprotein, AbD SeloteC, Raleigh, NC), the antibody was incubated with 10 mM sodium periodate at 4°C for 1 hour. An equal volume of ethylene glycol was added, and the oxidized antibody was brought to a concentration of 10 μg / ml with sodium carbonate buffer, pH 9.5. Antibodies (5 μg / well) were added to the plate and, after incubation, washed with 0.1% Tween 20 / PBS 7.4 and blocked overnight with 3% BSA / PBS. For analysis, 5 μL of serum was diluted in 95 μL of heterophil blocking tubes. TM (Heterophilic Blocking Tubes TM (Scantibodies Laboratory, Inc., Santee, CA) and incubated at room temperature for 1 hour. Subsequently, the samples were added to the plate for 2 hours and washed five times in lectin incubation buffer (10 mM Tris pH 8.0, 0.15 M NaCl, 0.1% Tween 20). Fucosylated proteins were then detected using biotin-conjugated Auricularia aurantifolia (AAL) lectin (Vector Laboratories, Burlingame, CA). TM 800 μg of streptavidin to detect bound lectin and use Signal intensity was measured using an infrared imaging system (LI-COR Biotechnology, Lincoln, NE). In all cases, signal intensity was compared to that detected using commercially available human serum (Sigma-Aldrich, St. Louis, MO). It should be noted that the lectin-FLISA detects the amount of fucosylation present on equal amounts of capture molecules from each patient sample and does so in a manner independent of the total amount of protein in any given patient.

[0109] Proteomic identification of pollutants

[0110] Lectin-Western

[0111] Serum was depleted of IgG using protein A / G-coated agarose beads and the serum was purified using a magnetic column coated with monoclonal anti-A1AT (AbDSerotec, Raleigh, NC). A1AT was immunoprecipitated using ELISA (Thermo Fisher Scientific Inc., Waltham, MA). A1AT was then eluted and resolved by SDS-PAGE. Fucosylated A1AT was detected using biotin-conjugated AAL agglutinin (AAL). Bound AAL was detected using streptavidin conjugated to IRDye™ 800 and the ligands were stained using ELISA. Signal intensity was measured using an infrared imaging system (LI-COR Biotechnology, Lincoln, NE). Subsequently, A1AT was detected using polyclonal anti-A1AT (Sigma-Aldrich, St. Louis, MO) and the cells were visualized using an IRDye-conjugated TM Bound antibodies were detected with 700 anti-rabbit antibody.

[0112] In a previous analysis of a small sample set of 40 patients (20 with cirrhosis and 20 with cirrhosis and HCC), the lectin-ELISA method was unable to specifically detect changes in fucosylation of a given protein. For example, a lectin ELISA was performed for fucosylated alpha-1-antitrypsin (A1AT). In this method, an A1AT antibody that had been modified to remove its inherent fucosylation was coated on the bottom of a 96-well plate. Serum that had been depleted of IgG using protein A / G was added, and the fucosylation level of the captured A1AT was detected using recombinant Dictyophora aurantiacus agglutinin (AAL). When this assay was performed using HCC serum, a lectin-reactive signal was observed. However, it was not possible to compete this signal with non-fucosylated A1AT, even when non-fucosylated A1AT was bound to the capture antibody. Furthermore, the use of nonspecific antibodies, such as AFP in AFP-negative patients, resulted in the same nonspecific signal. Trypsinization of the samples prior to analysis confirmed that the signal was protein-based.

[0113] Identification of IgM as a potential contaminant

[0114] To identify nonspecific lectin-reactive substances found in serum, a lectin ELISA was performed for A1AT in a 96-well plate. The plates were incubated with serum using the same conditions as for a normal lectin ELISA, but before adding the lectin, the samples were incubated with SDS lysis buffer and examined by SDS-PAGE and lectin blotting with a fucose-binding lectin using colloidal Coomassie Brilliant Blue staining, or by blotting for the presence of human A1AT. Strong lectin staining was observed at a band of ~80 kD, and a weaker band was observed at 50 kD. Although this 50 kD band was also observed after staining with colloidal Coomassie Brilliant Blue and by staining with an A1AT antibody, no ~80 kD band was observed by either colloidal Coomassie Brilliant Blue staining or A1AT lectin blotting. This suggests that the nonspecific lectin-reactive substance in these samples is a highly fucosylated 80 kD glycoprotein. Subsequently, identically treated wells were collected after ELISA capture and subjected to proteomic analysis after trypsin digestion. A list of glycoproteins identified in these samples is shown in Table 5 below, with two major proteins observed of similar size to those observed in the lectin blot: complement B and IgM heavy chain.

[0115] Table 5

[0116]

[0117]

[0118] Subsequently, the same lectin ELISA experiment was repeated using anti-complement B and anti-IgM antibodies. Although no staining was observed in the material captured with the anti-complement B antibody, immunoblotting revealed that the 80 kD lectin-reactive material found in these samples was an IgM heavy chain. The 80 kD band represents the IgM heavy chain, which indicates the entire IgM molecule. In total, there are 10 heavy chains per IgM molecule.

[0119] To remove native IgM from serum samples, a method was developed to remove this substance from serum prior to lectin-ELISA analysis. Initially, protein L was used, but the results were unsatisfactory. A subsequent method involved incubating the serum with 20 μl of Pierce™ Protein A / G Plus (Thermo Fisher Scientific Inc., Waltham, MA) for 1 hour, followed by filtering the mixture through a 100 kD centrifugal filter to remove both IgG and IgM from the serum prior to lectin ELISA. When this method was used, both IgG and IgM were effectively removed from the serum, as determined by both immunoblotting and lectin blotting. Subsequently, after IgM and IgG were removed using this method, non-fucosylated A1AT could be used to block the lectin-reactive A1AT signal, confirming the finding that the presence of IgG and IgM was responsible for the contaminating signal observed in these samples.

[0120] The ability of the lectin-ELISA method to analyze fucosylated glycoforms of A1AT was examined using a small sample set in which the method had previously failed. This set consisted of 20 patients with cirrhosis and 20 patients with HCC in the setting of cirrhosis. Importantly, in this set, lectin-western data showing lectin-reactive A1AT were determined, and the data served as the gold standard for lectin-ELISA (Comunale, MA et al., Proteomics Clin. Appl. 2013; 7; 690–700) and allowed comparison between the performance of lectin-ELISA with and without filtration. Using the unfiltered method, the mean values ​​were 3.2 (± 2.0) for cirrhotic samples and 2.9 (± 1.8) for HCC samples. There was no statistical difference between the lectin-reactive signals in these samples (p = 0.74). The AUROC of this assay was 0.578. In contrast, using the filtration method of the present invention, the lectin ELISA resulted in a mean value of 1.4 (±0.75) in cirrhotic samples and 2.4 (±1.1) in HCC samples. This difference was statistically significant (P=0.0016). When comparing pre- and post-filtration samples, there was a statistical difference between pre- and post-filtration cirrhotic samples (P=0.0005), but no statistical difference was found between pre- and post-filtration HCC samples (P=0.5249).

[0121] The AUROC for the comparison between HCC and cirrhosis samples after filtration was 0.788, which is almost identical to that observed by lectin-western (Comunale et al., 2013). Importantly, there was a high correlation between lectin-western and lectin-ELISA after filtration. Importantly, every lectin-ELISA-positive sample after filtration was also positive by lectin-western, and every sample that was negative by lectin-ELISA after filtration was also negative by lectin-western.

[0122] The assay was then used to assess the presence of fucosylated α-1-antitrypsin and fucosylated kininogen in 80 independent samples. The samples were examined using both filtered and unfiltered lectin-ELISA. While the unfiltered analysis resulted in no differences between HCC and control samples, the filtered analysis did produce statistically significant differences.

[0123] Optional procedure for removal of IgG and IgM contaminants.

[0124] An alternative method for the removal of IgG and IgM was developed, whereby the serum was incubated with PEG-8000, centrifuged, and the supernatant was assayed in a lectin ELISA. For example, to analyze fucosylated A1AT or fucosylated fetuin-A, 2 μL of serum was added to 8 μL of PBS and then 6 μL of a 40% aqueous solution of PEG-8000 was added to give a final PEG-8000 concentration of 15%. The sample was mixed by drawing in and out of the pipette 10-20 times, followed by a brief vortex. The sample was then incubated on a shaker at 1000-1500 rpm for 30 minutes and then continued to incubate at 4°C overnight with slow shaking. The next morning, the sample was centrifuged at 14,000 rpm at 4°C, and the supernatant was quickly transferred to a fresh tube and then subjected to a lectin ELISA. IgG and IgM removal efficacy was confirmed using methods similar to the original procedure for immunoglobulin removal, such as by immunoblotting or by lectin blotting.

Claims

1. Use of a reagent for removing IgG from a biological fluid from a subject, a reagent for removing IgM from a biological fluid from a subject, and a reagent for measuring a biomarker in the biological fluid in the preparation of a kit for detecting hepatocellular carcinoma in a subject by an in vitro method comprising: measuring the amount of the biomarker in a biological fluid from the subject, wherein IgG and IgM proteins are removed from the biological fluid from the subject prior to measuring the amount of the biomarker that is a fucose-containing glycoform; determining the age and sex of the subject; and determining the absence or presence of hepatocellular carcinoma in the subject using the optimized output of the function of the determined age and sex and the biomarkers measured in the biological fluid, wherein the biological fluid is whole blood or plasma, Wherein the reagent for removing IgM includes an item of laboratory equipment or mechanical system, and The biomarkers are the following combinations: i) total alpha-fetoprotein, fucosylated kininogen, aspartate aminotransferase, and alkaline phosphatase, or ii) Total alpha-fetoprotein, fucosylated kininogen, aspartate aminotransferase, alkaline phosphatase, and fucosylated alpha-1-antitrypsin.

2. A kit for detecting hepatocellular carcinoma in a subject, the kit comprising: a reagent for removing IgG from a biological fluid from the subject; a reagent for removing IgM from a biological fluid from the subject; Reagents for measuring biomarkers in the biological fluid; and instructions for determining the absence or presence of hepatocellular carcinoma in the subject using the optimized output of a function of the determined age and sex and the biomarkers measured in the biological fluid, wherein the biological fluid is whole blood or plasma, Wherein the reagent for removing IgM includes an item of laboratory equipment or mechanical system, and The biomarkers are the following combinations: i) total alpha-fetoprotein, fucosylated kininogen, aspartate aminotransferase, and alkaline phosphatase, or ii) Total alpha-fetoprotein, fucosylated kininogen, aspartate aminotransferase, alkaline phosphatase, and fucosylated alpha-1-antitrypsin.

3. Use of a reagent for removing IgG from a biological fluid from a subject, a reagent for removing IgM from a biological fluid from a subject, and a reagent for measuring a biomarker in the biological fluid in the preparation of a kit for assigning a subject to a group having a higher or lower probability of hepatocellular carcinoma by an in vitro method comprising: measuring the amount of the biomarker in a biological fluid from the subject, wherein IgG and IgM proteins are removed from the biological fluid from the subject prior to measuring the amount of the biomarker that is a fucose-containing glycoform; determining the age and sex of the subject; and assigning the subject to a group having a higher or lower probability of hepatocellular carcinoma based on the optimized output of the function of the determined age and sex and the biomarkers measured in the biological fluid, wherein the biological fluid is whole blood or plasma, Wherein the reagent for removing IgM includes an item of laboratory equipment or mechanical system, and The biomarkers are the following combinations: i) total alpha-fetoprotein, fucosylated kininogen, aspartate aminotransferase, and alkaline phosphatase, or ii) Total alpha-fetoprotein, fucosylated kininogen, aspartate aminotransferase, alkaline phosphatase, and fucosylated alpha-1-antitrypsin.

4. Use according to claim 1 or 3, wherein the function comprises corresponding optimized weighting coefficients for the determined age and gender and for the measured amount of the biomarker.

5. The use according to claim 1 or 3, wherein IgG is removed by incubating the biological fluid with protein A / G and IgM is removed by passing the biological fluid through a filter based on molecular weight.

6. Use according to claim 1 or 3, wherein IgG and IgM are removed by incubating the biological fluid with polyethylene glycol.

7. The kit according to claim 2, wherein the function comprises corresponding optimized weighting coefficients for the determined age and gender and for the measured amount of the biomarker.

8. The kit according to claim 2 or 7, wherein the reagent for removing IgG is protein A / G, and the reagent for removing IgM is a filter based on molecular weight.

9. The use according to claim 1 or 3, wherein the hepatocellular carcinoma is early-stage hepatocellular carcinoma.

10. The use according to claim 1 or 3, wherein the hepatocellular carcinoma is alpha-fetoprotein-negative hepatocellular carcinoma.

11. The use according to claim 1 or 3, wherein the hepatocellular carcinoma is an early stage hepatocellular carcinoma which is also alpha-fetoprotein-negative hepatocellular carcinoma. The kit according to claim 2 , wherein the hepatocellular carcinoma is early-stage hepatocellular carcinoma.

13. The kit according to claim 2, wherein the hepatocellular carcinoma is alpha-fetoprotein-negative hepatocellular carcinoma.

14. The kit according to claim 2, wherein the hepatocellular carcinoma is an early stage hepatocellular carcinoma that is also alpha-fetoprotein-negative hepatocellular carcinoma.

Images