Patent Application
20210222254
The present invention relates to methods for determining an adenocarcinoma subtype of a lung sample and for predicting the response to a treatment for a patient inflicted with specific subtypes of lung cancer.
The Sequence Listing associated with this application is provided in text format in lieu of a paper copy, and is hereby incorporated by reference into the specification. The name of the text file containing the Sequence Listing is GNCN_009_02US_SeqList_ST25.txt. The text file is 194 KB, and was created on Jan. 22, 2021, and is being submitted electronically via EFS-Web.
Lung cancer is the leading cause of cancer deaths in both the United States and worldwide. Approximately 172,000 tumors of the lung were diagnosed in 2005 with an estimated 163,000 deaths, more than colon, breast, and prostate combined. At least 75% of patients present with locally advanced disease. Although there has been much effort to improve screening using technology such as high-resolution CT, these methods often produce false positive results and usually do not change outcome. Thus, even small tumors detected early present a significant threat to patients with postoperative 5-year survival rates for stage I lung cancer estimated between 47 to 63 percent. For patients with advanced disease the prognosis is worse with median survivals well under a year. In general, palliative therapy is effective but not sustainable and the average impact on overall survival is approximately 3 months.
At the population level, the underlying cause of lung cancer is clearly tobacco use, with 90% of all lung cancers attributed directly to smoking. Smoking is so tightly correlated with lung cancer that it confounds definitive association with most other risk factors; although asbestos, radon, and a number of lung irritants are generally accepted as lung cancer risk factors. A genetic association is strongly suspected, however, the exact mechanism remains to be determined outside of a select group of rare Mendelian cancer syndromes. Despite many classification schemes and ongoing clinical trials, there has been overall disappointing progress in the field of clinical diagnostics and therapeutics.
Most lung cancers are classified as non-small cell lung carcinoma (NSCLC) (>85%), which is a diverse group with subtypes occurring throughout the respiratory tract. Adenocarcinoma (AD) and squamous cell carcinomas (SCC or SQ), the two main subtypes of NSCLC, are diagnosed at near equal frequency but are often found at different locations with SCC occurring more centrally. The 6th edition of the consensus classification of lung cancers developed by the World Health Organization (WHO) describes no fewer than 90 malignant morphologic classes and variants. There can often be heterogeneity, especially in larger tumors >1.5 cm, making morphological classification more difficult and leading to designations such as adeno-squamous carcinoma. Further, studies of histologic diagnosis reproducibility have shown limited intra-pathologist agreement and inter-pathologist agreement. Variability in morphology, limited tissue samples, and the need for assessment of a growing list of therapeutically targeted markers pose challenges to the current diagnostic standard. This is further highlighted by the idea that differentiation among various morphologic subtypes of lung cancer can be essential in guiding patient management and additional molecular testing can be used to identify specific therapeutic target markers.
Currently, gene expression based lung adenocarcinoma (AD) subtyping has been primarily restricted to a research protocol involving the extraction of RNA from fresh frozen lung tumors, followed by application of a nearest centroid predictor using quantitative gene expression of over 500 genes. Gene expression based adenocarcinoma subtyping has been shown to classify adenocarcinoma tumors into 3 biologically distinct subtypes (Terminal Respiratory Unit (TRU; formerly referred to as Bronchioid), Proximal Inflammatory (PI; formerly referred to as Squamoid), and Proximal Proliferative (PP; formerly referred to as Magnoid)) which can vary in their genomic profiles including gene expression, mutational spectrum, and copy number alterations. Further, these three subtypes can vary in their prognosis, in their distribution of smokers vs. nonsmokers, in their prevalence of EGFR alterations, ALK rearrangements, TP53 mutations, in their angiogenic features, and in their immunogenic response features. Despite evidence of prognostic and predictive benefits from AD subtyping, the requirement for gene expression of >500 genes in combination with complex bioinformatics analyses, has hindered the application of AD subtyping in drug development and/or in the clinic.
Cancer immunosurveillance is the principle that the immune system can identify precancerous and cancerous cells and kill these cells before they become clinically relevant, which has been demonstrated in immunodeficient mouse models. Innate and adaptive immune responses can work together to either promote or inhibit cancer growth, and evasion of immune destruction is an emerging hallmark of cancer. Historically, methods of immune stimulation were not effective for lung cancer patients in the clinic. Deficiencies in tumor antigen expression and presentation on antigen presenting cells (APCs), infiltration of immunosuppressive cells and cytokines, and ineffective T-cell activation can lead to immunosuppression at the tumor site. Advances in the understanding of cancer and the immune system have led to effective therapies that activate antitumor responses, even in tumors that have highly developed methods of immune evasion, such as lung cancer. However the high immunosuppressive effects caused by lung tumors limit the beneficial effects of these advances due to a delicate balance between immunoactivation and immunosuppression in a patient. For example, in NSCLC, the role of immunosuppressive cells hampering immune activation is high, which is suggested to be related to the type of tumor, advanced stage of the disease, and the tumor load.
Therefore, developing a method to effectively distinguish intrinsic lung adenocarcinoma subtypes is critical for clinical diagnosis and disease management. Accordingly, new methods are needed to further define populations that might be likely to respond to immunotherapy. The present invention addresses these and other needs in the field for determining a prognosis or disease outcome for adenocarcinoma patient populations based in part on the adenocarcinoma subtype (Terminal Respiratory Unit (TRU), Proximal Inflammatory (PI), Proximal Proliferative (PP)) of the patient. The methods of the invention provide a means for determining the cellular and molecular origins of lung cancer (e.g., subtyping AD) and can provide for more accurate diagnosis and applicable treatments as compared to diagnostic methods known in the art.
In one aspect, provided herein is a method for determining an adenocarcinoma (AD) subtype of a lung tissue sample obtained from a patient, the method comprising detecting an expression level of at least one classifier biomarker of Table 1, wherein the detection of the expression level of the classifier biomarker specifically identifies a Terminal Respiratory Unit (TRU), Proximal Proliferative (PP), or Proximal Inflammatory (PI) AD subtype. In some cases, the method further comprises comparing the detected levels of expression of the at least one classifier biomarkers of Table 1 to the expression of the at least one classifier biomarkers of Table 1 in at least one sample training set(s), wherein the at least one sample training set comprises expression data of the at least one classifier biomarkers of Table 1 from a reference AD TRU sample, expression data of the at least one classifier biomarkers of Table 1 from a reference AD PP sample, expression data of the at least one classifier biomarkers of Table 1 from a reference AD PI sample or a combination thereof; and classifying the sample as TRU, PP or PI subtype based on the results of the comparing step. In some cases, the comparing step comprises applying a statistical algorithm which comprises determining a correlation between the expression data obtained from the sample and the expression data from the at least one training set(s); and classifying the sample as a TRU, PP or PI subtype based on the results of the statistical algorithm. In some cases, the expression level of the classifier biomarker is detected at the nucleic acid level. In some cases, the nucleic acid level is RNA or cDNA. In some cases, the detecting an expression level comprises performing quantitative real time reverse transcriptase polymerase chain reaction (qRT-PCR), RNAseq, microarrays, gene chips, nCounter Gene Expression Assay, Serial Analysis of Gene Expression (SAGE), Rapid Analysis of Gene Expression (RAGE), nuclease protection assays, Northern blotting, or any other equivalent gene expression detection techniques. In some cases, the expression level is detected by performing qRT-PCR. In some cases, the detection of the expression level comprises using at least one pair of oligonucleotide primers specific for at least one classifier biomarker of Table 1. In some cases, the sample is a formalin-fixed, paraffin-embedded (FFPE) lung tissue sample, fresh or a frozen tissue sample, an exosome, wash fluids, cell pellets, or a bodily fluid obtained from the patient. In some cases, the bodily fluid is blood or fractions thereof, urine, saliva, or sputum. In some cases, the at least one classifier biomarker comprises a plurality of classifier biomarkers. In some cases, the plurality of classifier biomarkers comprises at least two classifier biomarkers, at least 8 classifier biomarkers, at least 16 classifier biomarkers, at least 24 classifier biomarkers, at least 32 classifier biomarkers, at least 40 classifier biomarkers, or at least 48 classifier biomarkers of Table 1. In some cases, the at least one classifier biomarker comprises all the classifier biomarkers of Table 1.
In another aspect, provided herein is a method for determining an adenocarcinoma (AD) subtype of a lung tissue sample obtained from a patient comprising detecting an expression level of at least one nucleic acid molecule that encodes a classifier biomarker having a specific expression pattern in lung cancer cells, wherein the classifier biomarker is selected from the group consisting of the classifier genes set forth in Table 1, the method comprising: (a) isolating nucleic acid material from a lung tissue sample from a patient; (b) mixing the nucleic acid material with oligonucleotides that are substantially complementary to portions of nucleic acid molecule of the classifier biomarker; and (c) detecting expression of the classifier biomarker. In some cases, the method further comprises comparing the detected levels of expression of the at least one classifier biomarkers of Table 1 to the expression of the at least one classifier biomarkers of Table 1 in at least one sample training set(s), wherein the at least one sample training set comprises expression data of the at least one classifier biomarkers of Table 1 from a reference AD TRU sample, expression data of the at least one classifier biomarkers of Table 1 from a reference AD PP sample, expression data of the at least one classifier biomarkers of Table 1 from a reference AD PI sample or a combination thereof; and classifying the sample as TRU, PP or PI subtype based on the results of the comparing step. In some cases, the comparing step comprises applying a statistical algorithm which comprises determining a correlation between the expression data obtained from the sample and the expression data from the at least one training set(s); and classifying the sample as a TRU, PP or PI subtype based on the results of the statistical algorithm. In some cases, the detecting the expression level comprises performing qRT-PCR or any hybridization-based gene assays. In some cases, the expression level is detected by performing qRT-PCR. In some cases, the detection of the expression level comprises using at least one pair of oligonucleotide primers specific for at least one classifier biomarker of Table 1. In some cases, the method further comprises predicting the response to a therapy for treating a subtype of lung adenocarcinoma (AD) based on the detected expression level of the classifier biomarker. In some cases, the therapy is chemotherapy, angiogenesis inhibitors and/or immunotherapy. In some cases, the subtype of lung AD is TRU and the therapy is chemotherapy or angiogenesis inhibitor. In some cases, the subtype of lung AD is PP and the therapy is chemotherapy. In some cases, the subtype of lung AD is PI and the therapy is an immunotherapy. In some cases, the sample is a formalin-fixed, paraffin-embedded (FFPE) lung tissue sample, fresh or a frozen tissue sample, an exosome, wash fluids, cell pellets or a bodily fluid obtained from the patient. In some cases, the bodily fluid is blood or fractions thereof, urine, saliva, or sputum. In some cases, the at least one nucleic acid molecule that encodes a classifier biomarker comprises a plurality of nucleic acid molecules that encode a plurality of classifier biomarkers. In some cases, the plurality of classifier biomarkers comprises at least two classifier biomarkers, at least 5 classifier biomarkers, at least 10 classifier biomarkers, at least 20 classifier biomarkers or at least 30 classifier biomarkers of Table 1. In some cases, the at least one classifier biomarker comprises all the classifier biomarkers of Table 1.
In yet another aspect, provided herein is a method of detecting a biomarker in a lung tissue sample obtained from a patient, the method comprising measuring the expression level of a plurality of biomarker nucleic acids selected from Table 1 using an amplification, hybridization and/or sequencing assay. In some cases, the lung tissue sample was previously diagnosed as being adenocarcinoma. In some cases, the previous diagnosis was by histological examination. In some cases, the amplification, hybridization and/or sequencing assay comprises performing quantitative real time reverse transcriptase polymerase chain reaction (qRT-PCR), RNAseq, microarrays, gene chips, nCounter Gene Expression Assay, Serial Analysis of Gene Expression (SAGE), Rapid Analysis of Gene Expression (RAGE), nuclease protection assays, Northern blotting, or any other equivalent gene expression detection techniques. In some cases, the expression level is detected by performing qRT-PCR. In some cases, the detection of the expression level comprises using at least one pair of oligonucleotide primers per each of the plurality of biomarker nucleic acids selected from Table 1. In some cases, the sample is a formalin-fixed, paraffin-embedded (FFPE) lung tissue sample, fresh or a frozen tissue sample, an exosome, wash fluids, cell pellets, or a bodily fluid obtained from the patient. In some cases, the bodily fluid is blood or fractions thereof, urine, saliva, or sputum. In some cases, the plurality of biomarker nucleic acids comprises, consists essentially of or consists of at least two biomarker nucleic acids, at least 10 biomarker nucleic acids, at least 20 biomarker nucleic acids, at least 30 biomarker nucleic acids, at least 40 biomarker nucleic acids, at least 50 biomarker nucleic acids, at least 60 biomarker nucleic acids, or at least 70 biomarker nucleic acids of Table 1. In some cases, the plurality of biomarker nucleic acids comprises, consists essentially of or consists of all the classifier biomarker nucleic acids of Table 1.
In one aspect, provided herein is a method of detecting a biomarker in a lung tissue sample obtained from a patient, the method consisting essentially of measuring the expression level of a plurality of biomarker nucleic acids selected from Table 1 using an amplification, hybridization and/or sequencing assay. In some cases, the lung tissue sample was previously diagnosed as being adenocarcinoma. In some cases, the previous diagnosis was by histological examination. In some cases, the amplification, hybridization and/or sequencing assay comprises performing quantitative real time reverse transcriptase polymerase chain reaction (qRT-PCR), RNAseq, microarrays, gene chips, nCounter Gene Expression Assay, Serial Analysis of Gene Expression (SAGE), Rapid Analysis of Gene Expression (RAGE), nuclease protection assays, Northern blotting, or any other equivalent gene expression detection techniques. In some cases, the expression level is detected by performing qRT-PCR. In some cases, the detection of the expression level comprises using at least one pair of oligonucleotide primers per each of the plurality of biomarker nucleic acids selected from Table 1. In some cases, the sample is a formalin-fixed, paraffin-embedded (FFPE) lung tissue sample, fresh or a frozen tissue sample, an exosome, wash fluids, cell pellets, or a bodily fluid obtained from the patient. In some cases, the bodily fluid is blood or fractions thereof, urine, saliva, or sputum. In some cases, the plurality of biomarker nucleic acids comprises, consists essentially of or consists of at least two biomarker nucleic acids, at least 10 biomarker nucleic acids, at least 20 biomarker nucleic acids, at least 30 biomarker nucleic acids, at least 40 biomarker nucleic acids, at least 50 biomarker nucleic acids, at least 60 biomarker nucleic acids, or at least 70 biomarker nucleic acids of Table 1. In some cases, the plurality of biomarker nucleic acids comprises, consists essentially of or consists of all the classifier biomarker nucleic acids of Table 1.
In another aspect, provided herein is a method of detecting a biomarker in a lung tissue sample obtained from a patient, the method consisting of measuring the expression level of a plurality of biomarker nucleic acids selected from Table 1 using an amplification, hybridization and/or sequencing assay. In some cases, the lung tissue sample was previously diagnosed as being adenocarcinoma. In some cases, the previous diagnosis was by histological examination. In some cases, the amplification, hybridization and/or sequencing assay comprises performing quantitative real time reverse transcriptase polymerase chain reaction (qRT-PCR), RNAseq, microarrays, gene chips, nCounter Gene Expression Assay, Serial Analysis of Gene Expression (SAGE), Rapid Analysis of Gene Expression (RAGE), nuclease protection assays, Northern blotting, or any other equivalent gene expression detection techniques. In some cases, the expression level is detected by performing qRT-PCR. In some cases, the detection of the expression level comprises using at least one pair of oligonucleotide primers per each of the plurality of biomarker nucleic acids selected from Table 1. In some cases, the sample is a formalin-fixed, paraffin-embedded (FFPE) lung tissue sample, fresh or a frozen tissue sample, an exosome, wash fluids, cell pellets, or a bodily fluid obtained from the patient. In some cases, the bodily fluid is blood or fractions thereof, urine, saliva, or sputum. In some cases, the plurality of biomarker nucleic acids comprises, consists essentially of or consists of at least two biomarker nucleic acids, at least 10 biomarker nucleic acids, at least 20 biomarker nucleic acids, at least 30 biomarker nucleic acids, at least 40 biomarker nucleic acids, at least 50 biomarker nucleic acids, at least 60 biomarker nucleic acids, or at least 70 biomarker nucleic acids of Table 1. In some cases, the plurality of biomarker nucleic acids comprises, consists essentially of or consists of all the classifier biomarker nucleic acids of Table 1.
In another aspect, provided herein is a method of determining whether an adenocarcinoma patient is likely to respond to immunotherapy, the method comprising, determining the adenocarcinoma subtype of a lung tissue sample from the patient, wherein the adenocarcinoma subtype is selected from the group consisting of squamoid (proximal inflammatory), bronchoid (terminal respiratory unit) and magnoid (proximal proliferative); and based on the subtype, assessing whether the patient is likely to respond to immunotherapy. In some cases, the immunotherapy comprises checkpoint inhibitor therapy. In some cases, the checkpoint inhibitor targets PD-1 or PD-L1. In some cases, the checkpoint inhibitor targets CTLA-4. In some cases, the checkpoint inhibitor is Pembrolizumab, Nivolumab or an antigen fragment binding fragment thereof. In some cases, the checkpoint inhibitor is Ipilimumab or an antigen binding fragment thereof. In some cases, the patient is initially determined to have adenocarcinoma via a histological analysis of a sample. In some cases, the patient's adenocarcinoma molecular subtype is selected from squamoid (proximal inflammatory), bronchoid (terminal respiratory unit) or magnoid (proximal proliferative), and is determined via a histological analysis of a sample obtained from the patient. In some cases, the sample is a formalin-fixed, paraffin-embedded (FFPE) lung tissue sample, fresh or a frozen tissue sample, an exosome, or a bodily fluid obtained from the patient. In some cases, the bodily fluid is blood or fractions thereof, urine, saliva, or sputum. In some cases, the determining the adenocarcinoma subtype comprises determining expression levels of a plurality of classifier biomarkers. In some cases, the determining the expression levels of the plurality of classifier biomarkers is at a nucleic acid level by performing RNA sequencing, reverse transcriptase polymerase chain reaction (RT-PCR) or hybridization based analyses. In some cases, the plurality of classifier biomarkers for determining the adenocarcinoma subtype is selected from a publically available lung adenocarcinoma dataset. In some cases, the publically available lung adenocarcinoma dataset is TCGA Lung AD RNAseq dataset. In some cases, the plurality of classifier biomarkers for determining the adenocarcinoma subtype is selected from Table 1. In some cases, the RT-PCR is quantitative real time reverse transcriptase polymerase chain reaction (qRT-PCR). In some cases, the RT-PCR is performed with primers specific to the plurality of classifier biomarkers of Table 1. In some cases, the method further comprises comparing the detected levels of expression of the plurality of classifier biomarkers of Table 1 to the expression of the plurality of classifier biomarkers of Table 1 in at least one sample training set(s), wherein the at least one sample training set comprises expression data of the plurality of classifier biomarkers of Table 1 from a reference adenocarcinoma TRU sample, expression data of the plurality of classifier biomarkers of Table 1 from a reference adenocarcinoma PP sample, expression data of the plurality of classifier biomarkers of Table 1 from a reference adenocarcinoma PI sample, or a combination thereof; and classifying the first sample as TRU, PP, or PI based on the results of the comparing step. In some cases, the comparing step comprises applying a statistical algorithm which comprises determining a correlation between the expression data obtained from the sample and the expression data from the at least one training set(s); and classifying the sample as a TRU, PP, or PI subtype based on the results of the statistical algorithm. In some cases, the plurality of the classifier biomarkers comprise each of the classifier biomarkers set forth in Table 1.
In yet another aspect, provided herein is a method for selecting an adenocarcinoma patient for immunotherapy, the method comprising, determining an adenocarcinoma subtype of a lung tissue sample from the patient, based on the subtype; and selecting the patient for immunotherapy. In some cases, the immunotherapy comprises checkpoint inhibitor therapy. In some cases, the checkpoint inhibitor targets PD-1 or PD-L1. In some cases, the checkpoint inhibitor targets CTLA-4. In some cases, the checkpoint inhibitor is Pembrolizumab, Nivolumab or an antigen fragment binding fragment thereof. In some cases, the checkpoint inhibitor is Ipilimumab or an antigen binding fragment thereof. In some cases, the patient is initially determined to have adenocarcinoma via a histological analysis of a sample. In some cases, the patient's adenocarcinoma molecular subtype is selected from squamoid (proximal inflammatory), bronchoid (terminal respiratory unit) or magnoid (proximal proliferative), and is determined via a histological analysis of a sample obtained from the patient. In some cases, the sample is a formalin-fixed, paraffin-embedded (FFPE) lung tissue sample, fresh or a frozen tissue sample, an exosome, or a bodily fluid obtained from the patient. In some cases, the bodily fluid is blood or fractions thereof, urine, saliva, or sputum. In some cases, the determining the adenocarcinoma subtype comprises determining expression levels of a plurality of classifier biomarkers. In some cases, the determining the expression levels of the plurality of classifier biomarkers is at a nucleic acid level by performing RNA sequencing, reverse transcriptase polymerase chain reaction (RT-PCR) or hybridization based analyses. In some cases, the plurality of classifier biomarkers for determining the adenocarcinoma subtype is selected from a publically available lung adenocarcinoma dataset. In some cases, the publically available lung adenocarcinoma dataset is TCGA Lung AD RNAseq dataset. In some cases, the plurality of classifier biomarkers for determining the adenocarcinoma subtype is selected from Table 1. In some cases, the RT-PCR is quantitative real time reverse transcriptase polymerase chain reaction (qRT-PCR). In some cases, the RT-PCR is performed with primers specific to the plurality of classifier biomarkers of Table 1. In some cases, the method further comprises comparing the detected levels of expression of the plurality of classifier biomarkers of Table 1 to the expression of the plurality of classifier biomarkers of Table 1 in at least one sample training set(s), wherein the at least one sample training set comprises expression data of the plurality of classifier biomarkers of Table 1 from a reference adenocarcinoma TRU sample, expression data of the plurality of classifier biomarkers of Table 1 from a reference adenocarcinoma PP sample, expression data of the plurality of classifier biomarkers of Table 1 from a reference adenocarcinoma PI sample, or a combination thereof; and classifying the first sample as TRU, PP, or PI based on the results of the comparing step. In some cases, the comparing step comprises applying a statistical algorithm which comprises determining a correlation between the expression data obtained from the sample and the expression data from the at least one training set(s); and classifying the sample as a TRU, PP, or PI subtype based on the results of the statistical algorithm. In some cases, the plurality of the classifier biomarkers comprise each of the classifier biomarkers set forth in Table 1.
In one aspect, provided herein is a method of treating lung cancer in a subject, the method comprising: measuring the expression level of at least one biomarker nucleic acid in a lung cancer sample obtained from the subject, wherein the at least one biomarker nucleic acid is selected from a set of biomarkers listed in Table 1, wherein the presence, absence and/or level of the at least one biomarker indicates a subtype of the lung cancer; and administering an immunotherapeutic agent based on the subtype of the lung cancer. In some cases, the lung cancer sample is an adenocarcinoma lung cancer sample, In some cases, the at least one biomarker nucleic acid selected from the set of biomarkers comprises, consists essentially of or consists of at least two biomarker nucleic acids, at least 8 biomarker nucleic acids, at least 16 biomarker nucleic acids, at least 32 biomarker nucleic acids, or all 48 biomarker nucleic acids of Table 1. In some cases, the lung tissue sample was previously diagnosed as being adenocarcinoma. In some cases, the previous diagnosis was by histological examination. In some cases, the method further comprises measuring the expression of at least one biomarker from an additional set of biomarkers. In some cases, the additional set of biomarkers comprise gene expression signatures of Innate immune Cells (IIC), Adaptive immune Cells (AIC), one or more individual immune biomarkers, one or more interferon (IFN) genes, one or more major histocompatibility complex, class II (MHCII) genes or a combination thereof. In some cases, the additional set of biomarkers comprises genes selected from Tables 4A, 4B, 5, 6, 7, or a combination thereof. In some cases, the gene expression signatures of AICs are selected from Table 4A. In some cases, the gene expression signature of IICs are selected from Table 4B. In some cases, the one or more individual immune biomarkers are selected from Table 5. In some cases, the one or more IFN genes are selected from Table 6. In some cases, the one or more MHCII genes are selected from Table 7. In some cases, the measuring the expression level is conducted using an amplification, hybridization and/or sequencing assay. In some cases, the amplification, hybridization and/or sequencing assay comprises performing quantitative real time reverse transcriptase polymerase chain reaction (qRT-PCR), RNAseq, microarrays, gene chips, nCounter Gene Expression Assay, Serial Analysis of Gene Expression (SAGE), Rapid Analysis of Gene Expression (RAGE), nuclease protection assays, Northern blotting, or any other equivalent gene expression detection techniques. In some cases, the expression level is detected by performing qRT-PCR. In some cases, the sample is a formalin-fixed, paraffin-embedded (FFPE) lung tissue sample, fresh or a frozen tissue sample, an exosome, wash fluids, cell pellets, or a bodily fluid obtained from the patient. In some cases, the bodily fluid is blood or fractions thereof, urine, saliva, or sputum. In some cases, the subject's adenocarcinoma subtype is selected from squamoid (proximal inflammatory), bronchoid (terminal respiratory unit) or magnoid (proximal proliferative). In some cases, the lung cancer subtype is proximal inflammatory and wherein the immunotherapeutic agent comprises a checkpoint inhibitor. In some cases, the checkpoint inhibitor targets PD-1 or PD-L1. In some cases, the checkpoint inhibitor targets CTLA-4. In some cases, the checkpoint inhibitor is Pembrolizumab, Nivolumab or an antigen fragment binding fragment thereof. In some cases, the checkpoint inhibitor is Ipilimumab or an antigen binding fragment thereof. In some cases, the at least one biomarker nucleic acid is a plurality of biomarker nucleic acids, wherein the plurality of biomarker nucleic acids comprises at least one biomarker nucleic acid listed in Table 1 in combination with one or more biomarker nucleic acids from a publically available lung adenocarcinoma dataset, wherein the presence, absence and/or level of the plurality of biomarker nucleic acids indicates a subtype of the lung cancer. In some cases, the at least one biomarker nucleic acid is a plurality of biomarker nucleic acids, wherein the plurality of biomarker nucleic acids comprises all of the biomarker nucleic acids listed in Table 1 in combination with one or more biomarker nucleic acids from a publically available lung adenocarcinoma dataset, wherein the presence, absence and/or level of the plurality of biomarker nucleic acids indicates a subtype of the lung cancer. In some cases, the publically available lung adenocarcinoma dataset is TCGA Lung AD RNAseq dataset.
The present invention provides kits, compositions and methods for identifying or diagnosing lung cancer. That is, the methods can be useful for molecularly defining subsets of lung cancer, specifically lung adenocarcinoma (AD). The methods provide a classification of lung cancer that can be prognostic and predictive for therapeutic response. While a useful term for epidemiologic purposes, “lung cancer” may not refer to a specific disease, but rather can represent a heterogeneous collection of tumors of the lung, bronchus, and pleura. For practical purposes, lung cancer can generally be divided into two histological subtypes-small cell lung cancer (SCLC) and non-small cell lung cancer (NSCLC). These main tumor types can present at different frequencies, can have different anatomic locations, can have different predilections for metastasis, may respond differently to therapy, and may likely be derived from different cell progenitors.
“Determining an adenocarcinoma subtype” can include, for example, diagnosing or detecting the presence and type of lung adenocarcinoma, monitoring the progression of the disease, and identifying or detecting cells or samples that are indicative of subtypes.
In one embodiment, lung cancer status is assessed through the evaluation of expression patterns, or profiles, of a plurality of classifier genes or biomarkers in one or more subject samples. For the purpose of discussion, the term subject, or subject sample, refers to an individual regardless of health and/or disease status. A subject can be a subject, a study participant, a control subject, a screening subject, or any other class of individual from whom a sample is obtained and assessed in the context of the invention. Accordingly, a subject can be diagnosed with lung adenocarcinoma (including subtypes, or grades thereof), can present with one or more symptoms of lung AD cancer, or a predisposing factor, such as a family (genetic) or medical history (medical) factor, for lung cancer, can be undergoing treatment or therapy for lung cancer, or the like. Alternatively, a subject can be healthy with respect to any of the aforementioned factors or criteria. It will be appreciated that the term “healthy” as used herein, is relative to lung cancer status, as the term “healthy” cannot be defined to correspond to any absolute evaluation or status. Thus, an individual defined as healthy with reference to any specified disease or disease criterion, can in fact be diagnosed with any other one or more diseases, or exhibit any other one or more disease criterion, including one or more other cancers.
As used herein, an “expression profile” or a “biomarker profile” or “gene signature” comprises one or more values corresponding to a measurement of the relative abundance, level, presence, or absence of expression of a discriminative or classifier gene or biomarker. An expression profile can be derived from a subject prior to or subsequent to a diagnosis of lung cancer, can be derived from a biological sample collected from a subject at one or more time points prior to or following treatment or therapy, can be derived from a biological sample collected from a subject at one or more time points during which there is no treatment or therapy (e.g., to monitor progression of disease or to assess development of disease in a subject diagnosed with or at risk for lung cancer), or can be collected from a healthy subject. The term subject can be used interchangeably with patient. The patient can be a human patient. The one or more biomarkers of the biomarker profiles provided herein are selected from one or more biomarkers of Table 1.
As used herein, the term “determining an expression level” or “determining an expression profile” or “detecting an expression level” or “detecting an expression profile” as used in reference to a biomarker or classifier means the application of a biomarker specific reagent such as a probe, primer or antibody and/or a method to a sample, for example a sample of the subject or patient and/or a control sample, for ascertaining or measuring quantitatively, semi-quantitatively or qualitatively the amount of a biomarker or biomarkers, for example the amount of biomarker polypeptide or mRNA (or cDNA derived therefrom). For example, a level of a biomarker can be determined by a number of methods including for example immunoassays including for example immunohistochemistry, ELISA, Western blot, immunoprecipitation and the like, where a biomarker detection agent such as an antibody for example, a labeled antibody, specifically binds the biomarker and permits for example relative or absolute ascertaining of the amount of polypeptide biomarker, hybridization and PCR protocols where a probe or primer or primer set are used to ascertain the amount of nucleic acid biomarker, including for example probe based and amplification based methods including for example microarray analysis, RT-PCR such as quantitative RT-PCR (qRT-PCR), serial analysis of gene expression (SAGE), Northern Blot, digital molecular barcoding technology, for example Nanostring Counter Analysis, and TaqMan quantitative PCR assays. Other methods of mRNA detection and quantification can be applied, such as mRNA in situ hybridization in formalin-fixed, paraffin-embedded (FFPE) tissue samples or cells. This technology is currently offered by the QuantiGene ViewRNA (Affymetrix), which uses probe sets for each mRNA that bind specifically to an amplification system to amplify the hybridization signals; these amplified signals can be visualized using a standard fluorescence microscope or imaging system. This system for example can detect and measure transcript levels in heterogeneous samples; for example, if a sample has normal and tumor cells present in the same tissue section. As mentioned, TaqMan probe-based gene expression analysis (PCR-based) can also be used for measuring gene expression levels in tissue samples, and this technology has been shown to be useful for measuring mRNA levels in FFPE samples. In brief, TaqMan probe-based assays utilize a probe that hybridizes specifically to the mRNA target. This probe contains a quencher dye and a reporter dye (fluorescent molecule) attached to each end, and fluorescence is emitted only when specific hybridization to the mRNA target occurs. During the amplification step, the exonuclease activity of the polymerase enzyme causes the quencher and the reporter dyes to be detached from the probe, and fluorescence emission can occur. This fluorescence emission is recorded and signals are measured by a detection system; these signal intensities are used to calculate the abundance of a given transcript (gene expression) in a sample.
In one embodiment, the “expression profile” or a “biomarker profile” or “gene signature” associated with the gene cassettes or classifier genes described herein (e.g., Tables 1 and 2) can be useful for distinguishing between normal and tumor samples. In another embodiment, the tumor samples are lung adenocarcinoma (AD). In another embodiment, AD can be further classified as bronchioid, squamoid, and magnoid based upon an expression profile determined using the methods provided herein. The characterization of bronchioid, squamoid, and magnoid adenocarcinomas using tumor biopsy tissue has been described in Hayes et al. (2006) J. Clin Oncol. 24(31):5079-90Expression profiles using the classifier genes disclosed herein (e.g., Table 1) can provide valuable molecular tools for specifically identifying lung adenocarcinoma subtypes, and for evaluating therapeutic efficacy in treating lung adenocarcinoma. Accordingly, the invention provides methods for screening and classifying a subject for molecular AD subtypes and methods for monitoring efficacy of certain therapeutic treatments for lung AD.
In some instances, a single classifier gene provided herein is capable of identifying subtypes of lung adenocarcinoma with a predictive success of at least about 70%, at least about 71%, at least about 72%, about 73%, about 74%, about 75%, about 76%, about 77%, about 78%, about 79%, about 80%, about 81%, about 82%, about 83%, about 84%, about 85%, about 86%, about 87%, about 88%, about 89%, about 90%, about 91%, about 92%, about 93%, about 94%, about 95%, about 96%, about 97%, about 98%, about 99%, up to 100%.
In some instances, a single classifier gene as provided herein is capable of determining lung adenocarcinoma subtypes with a sensitivity or specificity of at least about 70%, at least about 71%, at least about 72%, about 73%, about 74%, about 75%, about 76%, about 77%, about 78%, about 79%, about 80%, about 81%, about 82%, about 83%, about 84%, about 85%, about 86%, about 87%, about 88%, about 89%, about 90%, about 91%, about 92%, about 93%, about 94%, about 95%, about 96%, about 97%, about 98%, about 99%, up to 100%.
The present invention also encompasses a system capable of distinguishing various subtypes of lung adenocarcinoma not detectable using current methods. This system c a n b e capable of processing a large number of subjects and subject variables such as expression profiles and other diagnostic criteria. The methods described herein can also be used for “pharmacometabonomics,” in analogy to pharmacogenomics, e.g., predictive of response to therapy. In this embodiment, subjects could be divided into “responders” and “nonresponders” using the expression profile as evidence of “response,” and features of the expression profile could then be used to target future subjects who would likely respond to a particular therapeutic course.
The expression profile can be used in combination with other diagnostic methods including histochemical, immunohistochemical, cytologic, immunocytologic, and visual diagnostic methods including histologic or morphometric evaluation of lung tissue.
In various embodiments of the present invention, the expression profile derived from a subject is compared to a reference expression profile. A “reference expression profile” can be a profile derived from the subject prior to treatment or therapy; can be a profile produced from the subject sample at a particular time point (usually prior to or following treatment or therapy, but can also include a particular time point prior to or following diagnosis of lung cancer); or can be derived from a healthy individual or a pooled reference from healthy individuals. A reference expression profile can be generic for lung cancer, or can be specific to different subtypes of lung adenocarcinoma.
The reference expression profile can be compared to a test expression profile. A “test expression profile” can be derived from the same subject as the reference expression profile except at a subsequent time point (e.g., one or more days, weeks or months following collection of the reference expression profile) or can be derived from a different subject. In summary, any test expression profile of a subject can be compared to a previously collected profile from a subject that has TRU, PP, or PI subtype.
The classifier biomarkers of the invention can include nucleic acids (RNA, cDNA, and DNA) and proteins, and variants and fragments thereof. Such biomarkers can include DNA comprising the entire or partial sequence of the nucleic acid sequence encoding the biomarker, or the complement of such a sequence. The biomarkers described herein can include RNA comprising the entire or partial sequence of any of the nucleic acid sequences of interest, or their non-natural cDNA products, obtained synthetically in vitro in a reverse transcription reaction. The biomarker nucleic acids can also include any expression product or portion thereof of the nucleic acid sequences of interest. A biomarker protein can be a protein encoded by or corresponding to a DNA biomarker of the invention. A biomarker protein can comprise the entire or partial amino acid sequence of any of the biomarker proteins or polypeptides. The biomarker nucleic acid can be extracted from a cell or can be cell free or extracted from an extracellular vesicular entity such as an exosome.
A “classifier biomarker” or “biomarker” or “classifier gene” can be any gene or protein whose level of expression in a tissue or cell is altered compared to that of a normal or healthy cell or tissue. For example, a “classifier biomarker” or “biomarker” or “classifier gene” can be any gene or protein whose level of expression in a tissue or cell is altered in a specific lung adenocarcinoma subtype. The detection of the biomarkers of the invention can permit the determination of the specific subtype. The “classifier biomarker” or “biomarker” or “classifier gene” may be one that is up-regulated (e.g. expression is increased) or down-regulated (e.g. expression is decreased) relative to a reference or control as provided herein. The reference or control can be any reference or control as provided herein. In some embodiments, the expression values of genes that are up-regulated or down-regulated in a particular subtype of lung adenocarcinoma can be pooled into one gene cassette. The overall expression level in each gene cassette is referred to herein as the “‘expression profile” and is used to classify a test sample according to the subtype of lung adenocarcinoma. However, it is understood that independent evaluation of expression for each of the genes disclosed herein can be used to classify tumor subtypes without the need to group up-regulated and down-regulated genes into one or more gene cassettes. In some cases, as shown in Table 2, a total of 48 biomarkers can be used for AD subtype determination. For each AD subtype, 8 of the 16 biomarkers can be negatively correlated genes while 8 can be positively correlated genes which can be selected as the gene signature of a specific AD subtype.
The classifier biomarkers of the invention include any gene or protein that is selectively expressed in lung adenocarcinoma, as defined herein above. Sample biomarker genes are listed in Table 1 or 2, below. In Table 2, the first column of the table represents the biomarker list selected for distinguishing Terminal Respiratory Unit (TRU). The middle column of the table represents the biomarker list selected for distinguishing Proximal Proliferative (PP). The last column of the table represents the biomarker list selected for distinguishing Proximal Inflammatory (PI).
The relative gene expression levels as represented by the tsat as described herein of the classifier biomarkers for lung AD subtyping are shown in Table 1. In one embodiment, the gene expression levels of the classifier biomarkers for lung adenocarcinoma subtyping are shown in Table 1. In one embodiment, all 48 genes can be used to classify the subtypes of AD. In one embodiment, the first 16 genes are the selected gene signature biomarkers for Terminal Respiratory Unit, with gene numbers 1-8 up-regulated and gene numbers 9-16 down-regulated compared to a non-TRU sample. In another embodiment, gene numbers 17-32 are the selected gene signature biomarkers specific for Proximal Proliferative (PP), with gene numbers 17-24 up-regulated and gene numbers 25-32 down-regulated compared to a non-PP sample. In yet another embodiment, gene numbers 33-48 are the selected gene signature biomarkers specific for Proximal Inflammatory (PI), with gene numbers 33-40 up-regulated and gene numbers 41-48 down-regulated compared to a non-PI sample.
In one embodiment, the methods and compositions provided herein allow for the differentiation of the three subtypes of adenocarcinoma: (1) Terminal Respiratory Unit (TRU), formerly referred to as bronchioid; (2) Proximal Proliferative (PP), formerly referred to as magnoid; and (3) Proximal Inflammatory (PI), formerly referred to as squamoid, with fewer genes needed that the molecular AD subtyping methods known in the art.
In general, the methods provided herein are used to classify a lung cancer sample as a particular lung cancer subtype (e.g. subtype of adenocarcinoma). In one embodiment, the method comprises measuring, detecting or determining an expression level of at least one of the classifier biomarkers of any publically available Lung AD expression dataset. In one embodiment, the method comprises detecting or determining an expression level of at least one of the classifier biomarkers of Table 1 in a lung cancer sample obtained from a patient or a subject. The lung cancer sample for the detection or differentiation methods described herein can be a sample previously determined or diagnosed as an adenocarcinoma sample. The previous diagnosis can be based on a histological analysis. The histological analysis can be performed by one or more pathologists.
In one embodiment, the measuring or detecting step is at the nucleic acid level by performing RNA-seq, a reverse transcriptase polymerase chain reaction (RT-PCR) or a hybridization assay with oligonucleotides that are substantially complementary to portions of cDNA molecules of the at least one classifier biomarker (such as the classifier biomarkers of Table 1) under conditions suitable for RNA-seq, RT-PCR or hybridization and obtaining expression levels of the at least one classifier biomarkers based on the detecting step. The expression levels of the at least one of the classifier biomarkers are then compared to reference expression levels of the at least one of the classifier biomarker (such as the classifier biomarkers of Table 1) from at least one sample training set. The at least one sample training set can comprise, (i) expression levels of the at least one biomarker from a sample that overexpresses the at least one biomarker, (ii) expression levels from a reference squamoid (proximal inflammatory), bronchioid (terminal respiratory unit) or magnoid (proximal proliferative) sample, or (iii) expression levels from an adenocarcinoma free lung sample, and classifying the lung tissue sample as a squamoid (proximal inflammatory), bronchioid (terminal respiratory unit) or a magnoid (proximal proliferative) subtype. The lung cancer sample can then be classified as a bronchioid, squamoid, or magnoid subtype of adenocarcinoma based on the results of the comparing step. In one embodiment, the comparing step can comprise applying a statistical algorithm which comprises determining a correlation between the expression data obtained from the lung tissue or cancer sample and the expression data from the at least one training set(s); and classifying the lung tissue or cancer sample as a squamoid (proximal inflammatory), bronchioid (terminal respiratory unit) or a magnoid (proximal proliferative) subtype based on the results of the statistical algorithm.
In one embodiment, the method comprises probing the levels of at least one of the classifier biomarkers provided herein, such as the classifier biomarkers of Table 1 at the nucleic acid level, in a lung cancer sample obtained from the patient. The lung cancer sample can be a sample previously determined or diagnosed as an adenocarcinoma sample. The previous diagnosis can be based on a histological analysis. The histological analysis can be performed by one or more pathologists. The probing step, in one embodiment, comprises mixing the sample with one or more oligonucleotides that are substantially complementary to portions of cDNA molecules of the at least one classifier biomarkers provided herein, such as the classifier biomarkers of Table 1 under conditions suitable for hybridization of the one or more oligonucleotides to their complements or substantial complements; detecting whether hybridization occurs between the one or more oligonucleotides to their complements or substantial complements; and obtaining hybridization values of the at least one classifier biomarkers based on the detecting step. The hybridization values of the at least one classifier biomarkers are then compared to reference hybridization value(s) from at least one sample training set. For example, the at least one sample training set comprises hybridization values from a reference TRU adenocarcinoma, PP adenocarcinoma, and/or PI adenocarcinoma sample. The lung cancer sample is classified, for example, as TRU, PP, or PI based on the results of the comparing step.
The lung tissue sample can be any sample isolated from a human subject or patient. For example, in one embodiment, the analysis is performed on lung biopsies that are embedded in paraffin wax. In one embodiment, the sample can be a fresh frozen lung tissue sample. In another embodiment, the sample can be a bodily fluid obtained from the patient. The bodily fluid can be blood or fractions thereof (i.e., serum, plasma), urine, saliva, sputum or cerebrospinal fluid (CSF). The sample can contain cellular as well as extracellular sources of nucleic acid for use in the methods provided herein. The extracellular sources can be cell-free DNA and/or exosomes. In one embodiment, the sample can be a cell pellet or a wash. This aspect of the invention provides a means to improve current diagnostics by accurately identifying the major histological types, even from small biopsies. The methods of the invention, including the RT-PCR methods, are sensitive, precise and have multi-analyte capability for use with paraffin embedded samples. See, for example, Cronin et al. (2004) Am. J Pathol. 164(1):35-42, herein incorporated by reference.
Formalin fixation and tissue embedding in paraffin wax is a universal approach for tissue processing prior to light microscopic evaluation. A major advantage afforded by formalin-fixed paraffin-embedded (FFPE) specimens is the preservation of cellular and architectural morphologic detail in tissue sections. (Fox et al. (1985) J Histochem Cytochem 33:845-853). The standard buffered formalin fixative in which biopsy specimens are processed is typically an aqueous solution containing 37% formaldehyde and 10-15% methyl alcohol. Formaldehyde is a highly reactive dipolar compound that results in the formation of protein-nucleic acid and protein-protein crosslinks in vitro (Clark et al. (1986) J Histochem Cytochem 34:1509-1512; McGhee and von Hippel (1975) Biochemistry 14:1281-1296, each incorporated by reference herein).
In one embodiment, the sample used herein is obtained from an individual, and comprises formalin-fixed paraffin-embedded (FFPE) tissue. However, other tissue and sample types are amenable for use herein. In one embodiment, the other tissue and sample types can be fresh frozen tissue, wash fluids, or cell pellets, or the like. In one embodiment, the sample can be a bodily fluid obtained from the individual. The bodily fluid can be blood or fractions thereof (e.g., serum, plasma), urine, sputum, saliva or cerebrospinal fluid (CSF). A biomarker nucleic acid as provided herein can be extracted from a cell or can be cell free or extracted from an extracellular vesicular entity such as an exosome.
Methods are known in the art for the isolation of RNA from FFPE tissue. In one embodiment, total RNA can be isolated from FFPE tissues as described by Bibikova et al. (2004) American Journal of Pathology 165:1799-1807, herein incorporated by reference. Likewise, the High Pure RNA Paraffin Kit (Roche) can be used. Paraffin is removed by xylene extraction followed by ethanol wash. RNA can be isolated from sectioned tissue blocks using the MasterPure Purification kit (Epicenter, Madison, Wis.); a DNase I treatment step is included. RNA can be extracted from frozen samples using Trizol reagent according to the supplier's instructions (Invitrogen Life Technologies, Carlsbad, Calif.). Samples with measurable residual genomic DNA can be resubjected to DNaseI treatment and assayed for DNA contamination. All purification, DNase treatment, and other steps can be performed according to the manufacturer's protocol. After total RNA isolation, samples can be stored at −80° C. until use.
General methods for mRNA extraction are well known in the art and are disclosed in standard textbooks of molecular biology, including Ausubel et al., ed., Current Protocols in Molecular Biology, John Wiley & Sons, New York 1987-1999. Methods for RNA extraction from paraffin embedded tissues are disclosed, for example, in Rupp and Locker (Lab Invest. 56:A67, 1987) and De Andres et al. (Biotechniques 18:42-44, 1995). In particular, RNA isolation can be performed using a purification kit, a buffer set and protease from commercial manufacturers, such as Qiagen (Valencia, Calif.), according to the manufacturer's instructions. For example, total RNA from cells in culture can be isolated using Qiagen RNeasy mini-columns. Other commercially available RNA isolation kits include MasterPure™. Complete DNA and RNA Purification Kit (Epicentre, Madison, Wis.) and Paraffin Block RNA Isolation Kit (Ambion, Austin, Tex.). Total RNA from tissue samples can be isolated, for example, using RNA Stat-60 (Tel-Test, Friendswood, Tex.). RNA prepared from a tumor can be isolated, for example, by cesium chloride density gradient centrifugation. Additionally, large numbers of tissue samples can readily be processed using techniques well known to those of skill in the art, such as, for example, the single-step RNA isolation process of Chomczynski (U.S. Pat. No. 4,843,155, incorporated by reference in its entirety for all purposes).
In one embodiment, a sample comprises cells harvested from a lung tissue sample, for example, an adenocarcinoma sample. Cells can be harvested from a biological sample using standard techniques known in the art. For example, in one embodiment, cells are harvested by centrifuging a cell sample and resuspending the pelleted cells. The cells can be resuspended in a buffered solution such as phosphate-buffered saline (PBS). After centrifuging the cell suspension to obtain a cell pellet, the cells can be lysed to extract nucleic acid, e.g, messenger RNA. All samples obtained from a subject, including those subjected to any sort of further processing, are considered to be obtained from the subject.
The sample, in one embodiment, is further processed before the detection of the biomarker levels of the combination of biomarkers set forth herein. For example, mRNA in a cell or tissue sample can be separated from other components of the sample. The sample can be concentrated and/or purified to isolate mRNA in its non-natural state, as the mRNA is not in its natural environment. For example, studies have indicated that the higher order structure of mRNA in vivo differs from the in vitro structure of the same sequence (see, e.g., Rouskin et al. (2014). Nature 505, pp. 701-705, incorporated herein in its entirety for all purposes).
mRNA from the sample in one embodiment, is hybridized to a synthetic DNA probe, which in some embodiments, includes a detection moiety (e.g., detectable label, capture sequence, barcode reporting sequence). Accordingly, in these embodiments, a non-natural mRNA-cDNA complex is ultimately made and used for detection of the biomarker. In another embodiment, mRNA from the sample is directly labeled with a detectable label, e.g., a fluorophore. In a further embodiment, the non-natural labeled-mRNA molecule is hybridized to a cDNA probe and the complex is detected.
In one embodiment, once the mRNA is obtained from a sample, it is converted to complementary DNA (cDNA) in a hybridization reaction or is used in a hybridization reaction together with one or more cDNA probes. cDNA does not exist in vivo and therefore is a non-natural molecule. Furthermore, cDNA-mRNA hybrids are synthetic and do not exist in vivo. Besides cDNA not existing in vivo, cDNA is necessarily different than mRNA, as it includes deoxyribonucleic acid and not ribonucleic acid. The cDNA is then amplified, for example, by the polymerase chain reaction (PCR) or other amplification method known to those of ordinary skill in the art. For example, other amplification methods that may be employed include the ligase chain reaction (LCR) (Wu and Wallace, Genomics, 4:560 (1989), Landegren et al., Science, 241:1077 (1988), incorporated by reference in its entirety for all purposes, transcription amplification (Kwoh et al., Proc. Natl. Acad. Sci. USA, 86:1173 (1989), incorporated by reference in its entirety for all purposes), self-sustained sequence replication (Guatelli et al., Proc. Nat. Acad. Sci. USA, 87:1874 (1990), incorporated by reference in its entirety for all purposes), incorporated by reference in its entirety for all purposes, and nucleic acid based sequence amplification (NASBA). Guidelines for selecting primers for PCR amplification are known to those of ordinary skill in the art. See, e.g., McPherson et al., PCR Basics: From Background to Bench, Springer-Verlag, 2000, incorporated by reference in its entirety for all purposes. The product of this amplification reaction, i.e., amplified cDNA is also necessarily a non-natural product. First, as mentioned above, cDNA is a non-natural molecule. Second, in the case of PCR, the amplification process serves to create hundreds of millions of cDNA copies for every individual cDNA molecule of starting material. The numbers of copies generated are far removed from the number of copies of mRNA that are present in vivo.
In one embodiment, cDNA is amplified with primers that introduce an additional DNA sequence (e.g., adapter, reporter, capture sequence or moiety, barcode) onto the fragments (e.g., with the use of adapter-specific primers), or mRNA or cDNA biomarker sequences are hybridized directly to a cDNA probe comprising the additional sequence (e.g., adapter, reporter, capture sequence or moiety, barcode). Amplification and/or hybridization of mRNA to a cDNA probe therefore serves to create non-natural double stranded molecules from the non-natural single stranded cDNA, or the mRNA, by introducing additional sequences and forming non-natural hybrids. Further, as known to those of ordinary skill in the art, amplification procedures have error rates associated with them. Therefore, amplification introduces further modifications into the cDNA molecules. In one embodiment, during amplification with the adapter-specific primers, a detectable label, e.g., a fluorophore, is added to single strand cDNA molecules. Amplification therefore also serves to create DNA complexes that do not occur in nature, at least because (i) cDNA does not exist in vivo, (i) adapter sequences are added to the ends of cDNA molecules to make DNA sequences that do not exist in vivo, (ii) the error rate associated with amplification further creates DNA sequences that do not exist in vivo, (iii) the disparate structure of the cDNA molecules as compared to what exists in nature, and (iv) the chemical addition of a detectable label to the cDNA molecules.
In some embodiments, the expression of a biomarker of interest is detected at the nucleic acid level via detection of non-natural cDNA molecules.
In some embodiments, the method for lung cancer AD subtyping includes detecting expression levels of a classifier biomarker set. In some embodiments, the detecting includes all of the classifier biomarkers of Table 1 at the nucleic acid level or protein level. In another embodiment, a single or a subset of the classifier biomarkers of Table 1 are detected, for example, from about 8 to about 16. For example, in one embodiment, from about 5 to about 10, from about 5 to about 15, from about 5 to about 20, from about 5 to about 25, from about 5 to about 30, from about 5 to about 35, from about 5 to about 40, from about 5 to about 45, from about 5 to about 48 of the biomarkers in Table 1 are detected in a method to determine the lung cancer AD subtype. In another embodiment, each of the biomarkers from Table 1 is detected in a method to determine the lung cancer subtype. In another embodiment, 16 of the biomarkers from Table 1 are selected as the gene signatures for a specific lung cancer AD subtype.
The detecting can be performed by any suitable technique including, but not limited to, RNA-seq, a reverse transcriptase polymerase chain reaction (RT-PCR), a microarray hybridization assay, or another hybridization assay, e.g., a NanoString assay for example, with primers and/or probes specific to the classifier biomarkers, and/or the like. In some cases, the primers useful for the amplification methods (e.g., RT-PCR or qRT-PCR) are any forward and reverse primers suitable for binding to a classifier gene provided herein, such as the classifier biomarkers listed in Table 1.
The biomarkers described herein include RNA comprising the entire or partial sequence of any of the nucleic acid sequences of interest, or their non-natural cDNA product, obtained synthetically in vitro in a reverse transcription reaction. The term “fragment” is intended to refer to a portion of the polynucleotide that generally comprise at least 10, 15, 20, 50, 75, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 800, 900, 1,000, 1,200, or 1,500 contiguous nucleotides, or up to the number of nucleotides present in a full-length biomarker polynucleotide disclosed herein. A fragment of a biomarker polynucleotide will generally encode at least 15, 25, 30, 50, 100, 150, 200, or 250 contiguous amino acids, or up to the total number of amino acids present in a full-length biomarker protein of the invention.
In some embodiments, overexpression, such as of an RNA transcript or its expression product, is determined by normalization to the level of reference RNA transcripts or their expression products, which can be all measured transcripts (or their products) in the sample or a particular reference set of RNA transcripts (or their non-natural cDNA products). Normalization is performed to correct for or normalize away both differences in the amount of RNA or cDNA assayed and variability in the quality of the RNA or cDNA used. Therefore, an assay typically measures and incorporates the expression of certain normalizing genes, including well known housekeeping genes, such as, for example, GAPDH and/or β-Actin. Alternatively, normalization can be based on the mean or median signal of all of the assayed biomarkers or a large subset thereof (global normalization approach).
Isolated mRNA can be used in hybridization or amplification assays that include, but are not limited to, Southern or Northern analyses, PCR analyses and probe arrays, NanoString Assays. One method for the detection of mRNA levels involves contacting the isolated mRNA or synthesized cDNA with a nucleic acid molecule (probe) that can hybridize to the mRNA encoded by the gene being detected. The nucleic acid probe can be, for example, a cDNA, or a portion thereof, such as an oligonucleotide of at least 7, 15, 30, 50, 100, 250, or 500 nucleotides in length and sufficient to specifically hybridize under stringent conditions to the non-natural cDNA or mRNA biomarker of the present invention.
As explained above, in one embodiment, once the mRNA is obtained from a sample, it is converted to complementary DNA (cDNA) in a hybridization reaction. Conversion of the mRNA to cDNA can be performed with oligonucleotides or primers comprising sequence that is complementary to a portion of a specific mRNA. Conversion of the mRNA to cDNA can be performed with oligonucleotides or primers comprising random sequence. Conversion of the mRNA to cDNA can be performed with oligonucleotides or primers comprising sequence that is complementary to the poly(A) tail of an mRNA. cDNA does not exist in vivo and therefore is a non-natural molecule. In a further embodiment, the cDNA is then amplified, for example, by the polymerase chain reaction (PCR) or other amplification method known to those of ordinary skill in the art. PCR can be performed with the forward and/or reverse primers comprising sequence complementary to at least a portion of a classifier gene provided herein, such as the classifier biomarkers in Table 1. The product of this amplification reaction, i.e., amplified cDNA is necessarily a non-natural product. As mentioned above, cDNA is a non-natural molecule. Second, in the case of PCR, the amplification process serves to create hundreds of millions of cDNA copies for every individual cDNA molecule of starting material. The number of copies generated is far removed from the number of copies of mRNA that are present in vivo.
In one embodiment, cDNA is amplified with primers that introduce an additional DNA sequence (adapter sequence) onto the fragments (with the use of adapter-specific primers). The adaptor sequence can be a tail, wherein the tail sequence is not complementary to the cDNA. For example, the forward and/or reverse primers comprising sequence complementary to at least a portion of a classifier gene provided herein, such as the classifier biomarkers from Table 1 can comprise tail sequence. Amplification therefore serves to create non-natural double stranded molecules from the non-natural single stranded cDNA, by introducing barcode, adapter and/or reporter sequences onto the already non-natural cDNA. In one embodiment, during amplification with the adapter-specific primers, a detectable label, e.g., a fluorophore, is added to single strand cDNA molecules. Amplification therefore also serves to create DNA complexes that do not occur in nature, at least because (i) cDNA does not exist in vivo, (ii) adapter sequences are added to the ends of cDNA molecules to make DNA sequences that do not exist in vivo, (iii) the error rate associated with amplification further creates DNA sequences that do not exist in vivo, (iv) the disparate structure of the cDNA molecules as compared to what exists in nature, and (v) the chemical addition of a detectable label to the cDNA molecules.
In one embodiment, the synthesized cDNA (for example, amplified cDNA) is immobilized on a solid surface via hybridization with a probe, e.g., via a microarray. In another embodiment, cDNA products are detected via real-time polymerase chain reaction (PCR) via the introduction of fluorescent probes that hybridize with the cDNA products. For example, in one embodiment, biomarker detection is assessed by quantitative fluorogenic RT-PCR (e.g., with TaqMan® probes). For PCR analysis, well known methods are available in the art for the determination of primer sequences for use in the analysis.
Biomarkers provided herein in one embodiment, are detected via a hybridization reaction that employs a capture probe and/or a reporter probe. For example, the hybridization probe is a probe derivatized to a solid surface such as a bead, glass or silicon substrate. In another embodiment, the capture probe is present in solution and mixed with the patient's sample, followed by attachment of the hybridization product to a surface, e.g., via a biotin-avidin interaction (e.g., where biotin is a part of the capture probe and avidin is on the surface). The hybridization assay, in one embodiment, employs both a capture probe and a reporter probe. The reporter probe can hybridize to either the capture probe or the biomarker nucleic acid. Reporter probes e.g., are then counted and detected to determine the level of biomarker(s) in the sample. The capture and/or reporter probe, in one embodiment contain a detectable label, and/or a group that allows functionalization to a surface.
For example, the nCounter gene analysis system (see, e.g., Geiss et al. (2008) Nat. Biotechnol. 26, pp. 317-325, incorporated by reference in its entirety for all purposes, is amenable for use with the methods provided herein.
Hybridization assays described in U.S. Pat. Nos. 7,473,767 and 8,492,094, the disclosures of which are incorporated by reference in their entireties for all purposes, are amenable for use with the methods provided herein, i.e., to detect the biomarkers and biomarker combinations described herein.
Biomarker levels may be monitored using a membrane blot (such as used in hybridization analysis such as Northern, Southern, dot, and the like), or microwells, sample tubes, gels, beads, or fibers (or any solid support comprising bound nucleic acids). See, for example, U.S. Pat. Nos. 5,770,722, 5,874,219, 5,744,305, 5,677,195 and 5,445,934, each incorporated by reference in their entireties.
In one embodiment, microarrays are used to detect biomarker levels. Microarrays are particularly well suited for this purpose because of the reproducibility between different experiments. DNA microarrays provide one method for the simultaneous measurement of the expression levels of large numbers of genes. Each array consists of a reproducible pattern of capture probes attached to a solid support. Labeled RNA or DNA is hybridized to complementary probes on the array and then detected by laser scanning hybridization intensities for each probe on the array are determined and converted to a quantitative value representing relative gene expression levels. See, for example, U.S. Pat. Nos. 6,040,138, 5,800,992 and 6,020,135, 6,033,860, and 6,344,316, each incorporated by reference in their entireties. High-density oligonucleotide arrays are particularly useful for determining the gene expression profile for a large number of RNAs in a sample.
Techniques for the synthesis of these arrays using mechanical synthesis methods are described in, for example, U.S. Pat. No. 5,384,261. Although a planar array surface is generally used, the array can be fabricated on a surface of virtually any shape or even a multiplicity of surfaces. Arrays can be nucleic acids (or peptides) on beads, gels, polymeric surfaces, fibers (such as fiber optics), glass, or any other appropriate substrate. See, for example, U.S. Pat. Nos. 5,770,358, 5,789,162, 5,708,153, 6,040,193 and 5,800,992, each incorporated by reference in their entireties. Arrays can be packaged in such a manner as to allow for diagnostics or other manipulation of an all-inclusive device. See, for example, U.S. Pat. Nos. 5,856,174 and 5,922,591, each incorporated by reference in their entireties.
Serial analysis of gene expression (SAGE) in one embodiment is employed in the methods described herein. SAGE is a method that allows the simultaneous and quantitative analysis of a large number of gene transcripts, without the need of providing an individual hybridization probe for each transcript. First, a short sequence tag (about 10-14 bp) is generated that contains sufficient information to uniquely identify a transcript, provided that the tag is obtained from a unique position within each transcript. Then, many transcripts are linked together to form long serial molecules, that can be sequenced, revealing the identity of the multiple tags simultaneously. The expression pattern of any population of transcripts can be quantitatively evaluated by determining the abundance of individual tags, and identifying the gene corresponding to each tag. See, Velculescu et al. Science 270:484-87, 1995; Cell 88:243-51, 1997, incorporated by reference in its entirety.
An additional method of biomarker level analysis at the nucleic acid level is the use of a sequencing method, for example, RNAseq, next generation sequencing, and massively parallel signature sequencing (MPSS), as described by Brenner et al. (Nat. Biotech. 18:630-34, 2000, incorporated by reference in its entirety). This is a sequencing approach that combines non-gel-based signature sequencing with in vitro cloning of millions of templates on separate 5 μm diameter microbeads. First, a microbead library of DNA templates is constructed by in vitro cloning. This is followed by the assembly of a planar array of the template-containing microbeads in a flow cell at a high density (typically greater than 3.0×106 microbeads/cm2). The free ends of the cloned templates on each microbead are analyzed simultaneously, using a fluorescence-based signature sequencing method that does not require DNA fragment separation. This method has been shown to simultaneously and accurately provide, in a single operation, hundreds of thousands of gene signature sequences from a yeast cDNA library.
Another method of biomarker level analysis at the nucleic acid level is the use of an amplification method such as, for example, RT-PCR or quantitative RT-PCR (qRT-PCR). Methods for determining the level of biomarker mRNA in a sample may involve the process of nucleic acid amplification, e.g., by RT-PCR (the experimental embodiment set forth in Mullis, 1987, U.S. Pat. No. 4,683,202), ligase chain reaction (Barany (1991) Proc. Natl. Acad. Sci. USA 88:189-193), self-sustained sequence replication (Guatelli et al. (1990) Proc. Natl. Acad. Sci. USA 87:1874-1878), transcriptional amplification system (Kwoh et al. (1989) Proc. Natl. Acad. Sci. USA 86:1173-1177), Q-Beta Replicase (Lizardi et al. (1988) Bio/Technology 6:1197), rolling circle replication (Lizardi et al., U.S. Pat. No. 5,854,033) or any other nucleic acid amplification method, followed by the detection of the amplified molecules using techniques well known to those of skill in the art. Numerous different PCR or qRT-PCR protocols are known in the art and can be directly applied or adapted for use using the presently described compositions for the detection and/or quantification of expression of discriminative genes in a sample. See, for example, Fan et al. (2004) Genome Res. 14:878-885, herein incorporated by reference. Generally, in PCR, a target polynucleotide sequence is amplified by reaction with at least one oligonucleotide primer or pair of oligonucleotide primers. The primer(s) hybridize to a complementary region of the target nucleic acid and a DNA polymerase extends the primer(s) to amplify the target sequence. Under conditions sufficient to provide polymerase-based nucleic acid amplification products, a nucleic acid fragment of one size dominates the reaction products (the target polynucleotide sequence which is the amplification product). The amplification cycle is repeated to increase the concentration of the single target polynucleotide sequence. The reaction can be performed in any thermocycler commonly used for PCR.
Quantitative RT-PCR (qRT-PCR) (also referred as real-time RT-PCR) is preferred under some circumstances because it provides not only a quantitative measurement, but also reduced time and contamination. As used herein, “quantitative PCR” (or “real time qRT-PCR”) refers to the direct monitoring of the progress of a PCR amplification as it is occurring without the need for repeated sampling of the reaction products. In quantitative PCR, the reaction products may be monitored via a signaling mechanism (e.g., fluorescence) as they are generated and are tracked after the signal rises above a background level but before the reaction reaches a plateau. The number of cycles required to achieve a detectable or “threshold” level of fluorescence varies directly with the concentration of amplifiable targets at the beginning of the PCR process, enabling a measure of signal intensity to provide a measure of the amount of target nucleic acid in a sample in real time. A DNA binding dye (e.g., SYBR green) or a labeled probe can be used to detect the extension product generated by PCR amplification. Any probe format utilizing a labeled probe comprising the sequences of the invention may be used.
Immunohistochemistry methods are also suitable for detecting the levels of the biomarkers of the present invention. Samples can be frozen for later preparation or immediately placed in a fixative solution. Tissue samples can be fixed by treatment with a reagent, such as formalin, gluteraldehyde, methanol, or the like and embedded in paraffin. Methods for preparing slides for immunohistochemical analysis from formalin-fixed, paraffin-embedded tissue samples are well known in the art.
In one embodiment, the levels of the biomarkers provided herein, such as the classifier biomarkers of Table 1 (or subsets thereof, for example 8 to 16, 16 to 32, or 32 to 48 biomarkers), are normalized against the expression levels of all RNA transcripts or their non-natural cDNA expression products, or protein products in the sample, or of a reference set of RNA transcripts or a reference set of their non-natural cDNA expression products, or a reference set of their protein products in the sample.
In one embodiment, lung adenocarcinoma subtypes can be evaluated using levels of protein expression of one or more of the classifier genes provided herein, such as the classifier biomarkers listed in Table 1. The level of protein expression can be measured using an immunological detection method. Immunological detection methods which can be used herein include, but are not limited to, competitive and non-competitive assay systems using techniques such as Western blots, radioimmunoassays, ELISA (enzyme linked immunosorbent assay), “sandwich” immunoassays, immunoprecipitation assays, precipitin reactions, gel diffusion precipitin reactions, immunodiffusion assays, agglutination assays, complement-fixation assays, immunoradiometric assays, fluorescent immunoassays, protein A immunoassays, and the like. Such assays are routine and well known in the art (see, e.g., Ausubel et al, eds, 1994, Current Protocols in Molecular Biology, Vol. I, John Wiley & Sons, Inc., New York, which is incorporated by reference herein in its entirety).
In one embodiment, antibodies specific for biomarker proteins are utilized to detect the expression of a biomarker protein in a body sample. The method comprises obtaining a body sample from a patient or a subject, contacting the body sample with at least one antibody directed to a biomarker that is selectively expressed in lung cancer cells, and detecting antibody binding to determine if the biomarker is expressed in the patient sample. A preferred aspect of the present invention provides an immunocytochemistry technique for diagnosing lung cancer subtypes. One of skill in the art will recognize that the immunocytochemistry method described herein below may be performed manually or in an automated fashion.
As provided throughout, the methods set forth herein provide a method for determining the lung cancer AD subtype of a patient. Once the biomarker levels are determined, for example by measuring non-natural cDNA biomarker levels or non-natural mRNA-cDNA biomarker complexes, the biomarker levels are compared to reference values or a reference sample, for example with the use of statistical methods or direct comparison of detected levels, to make a determination of the lung cancer molecular AD subtype. Based on the comparison, the patient's lung cancer sample is AD classified, e.g., as TRU, PP, or PI.
In one embodiment, expression level values of the at least one classifier biomarkers provided herein, such as the classifier biomarkers of Table 1 are compared to reference expression level value(s) from at least one sample training set, wherein the at least one sample training set comprises expression level values from a reference sample(s). In a further embodiment, the at least one sample training set comprises expression level values of the at least one classifier biomarkers provided herein, such as the classifier biomarkers of Table 1 from a proximal inflammatory (squamoid), proximal proliferative (magnoid), a terminal respiratory unit (bronchioid) sample, or a combination thereof.
In a separate embodiment, hybridization values of the at least one classifier biomarkers provided herein, such as the classifier biomarkers of Table 1 are compared to reference hybridization value(s) from at least one sample training set, wherein the at least one sample training set comprises hybridization values from a reference sample(s). In a further embodiment, the at least one sample training set comprises hybridization values of the at least one classifier biomarkers provided herein, such as the classifier biomarkers of Table 1 from a proximal inflammatory (squamoid), proximal proliferative (magnoid), a terminal respiratory unit (bronchioid) sample, or a combination thereof. Methods for comparing detected levels of biomarkers to reference values and/or reference samples are provided herein. Based on this comparison, in one embodiment a correlation between the biomarker levels obtained from the subject's sample and the reference values is obtained. An assessment of the lung cancer AD subtype is then made.
Various statistical methods can be used to aid in the comparison of the biomarker levels obtained from the patient and reference biomarker levels, for example, from at least one sample training set.
In one embodiment, a supervised pattern recognition method is employed. Examples of supervised pattern recognition methods can include, but are not limited to, the nearest centroid methods (Dabney (2005) Bioinformatics 21(22):4148-4154 and Tibshirani et al. (2002) Proc. Natl. Acad. Sci. USA 99(10):6576-6572); soft independent modeling of class analysis (SIMCA) (see, for example, Wold, 1976); partial least squares analysis (PLS) (see, for example, Wold, 1966; Joreskog, 1982; Frank, 1984; Bro, R., 1997); linear descriminant analysis (LDA) (see, for example, Nillson, 1965); K-nearest neighbour analysis (KNN) (sec, for example, Brown et al., 1996); artificial neural networks (ANN) (see, for example, Wasserman, 1989; Anker et al., 1992; Hare, 1994); probabilistic neural networks (PNNs) (see, for example, Parzen, 1962; Bishop, 1995; Speckt, 1990; Broomhead et al., 1988; Patterson, 1996); rule induction (RI) (see, for example, Quinlan, 1986); and, Bayesian methods (see, for example, Bretthorst, 1990a, 1990b, 1988). In one embodiment, the classifier for identifying tumor subtypes based on gene expression data is the centroid based method described in Mullins et al. (2007) Clin Chem. 53(7):1273-9, each of which is herein incorporated by reference in its entirety.
In other embodiments, an unsupervised training approach is employed, and therefore, no training set is used.
Referring to sample training sets for supervised learning approaches again, in some embodiments, a sample training set(s) can include expression data of a plurality or all of the classifier biomarkers (e.g., all the classifier biomarkers of Table 1) from an adenocarcinoma sample. The plurality of classifier biomarkers can comprise at least two classifier biomarkers, at least 8 classifier biomarkers, at least 16 classifier biomarkers, at least 24 classifier biomarkers, at least 32 classifier biomarkers, at least 40 classifier biomarkers, or at least 48 classifier biomarkers of Table 1. In some embodiments, the sample training set(s) are normalized to remove sample-to-sample variation.
In some embodiments, comparing can include applying a statistical algorithm, such as, for example, any suitable multivariate statistical analysis model, which can be parametric or non-parametric. In some embodiments, applying the statistical algorithm can include determining a correlation between the expression data obtained from the human lung tissue sample and the expression data from the adenocarcinoma training set(s). In some embodiments, cross-validation is performed, such as (for example), leave-one-out cross-validation (LOOCV). In some embodiments, integrative correlation is performed. In some embodiments, a Spearman correlation is performed. In some embodiments, a centroid based method is employed for the statistical algorithm as described in Mullins et al. (2007) Clin Chem. 53(7):1273-9, and based on gene expression data, which is herein incorporated by reference in its entirety.
Results of the gene expression performed on a sample from a subject (test sample) may be compared to a biological sample(s) or data derived from a biological sample(s) that is known or suspected to be normal (“reference sample” or “normal sample”, e.g., non-adenocarcinoma sample). In some embodiments, a reference sample or reference gene expression data is obtained or derived from an individual known to have a particular molecular subtype of adenocarcimona, i.e., squamoid (proximal inflammatory), bronchioid (terminal respiratory unit) or magnoid (proximal proliferative).
The reference sample may be assayed at the same time, or at a different time from the test sample. Alternatively, the biomarker level information from a reference sample may be stored in a database or other means for access at a later date.
The biomarker level results of an assay on the test sample may be compared to the results of the same assay on a reference sample. In some cases, the results of the assay on the reference sample are from a database, or a reference value(s). In some cases, the results of the assay on the reference sample are a known or generally accepted value or range of values by those skilled in the art. In some cases the comparison is qualitative. In other cases the comparison is quantitative. In some cases, qualitative or quantitative comparisons may involve but are not limited to one or more of the following: comparing fluorescence values, spot intensities, absorbance values, chemiluminescent signals, histograms, critical threshold values, statistical significance values, expression levels of the genes described herein, mRNA copy numbers.
In one embodiment, an odds ratio (OR) is calculated for each biomarker level panel measurement. Here, the OR is a measure of association between the measured biomarker values for the patient and an outcome, e.g., lung adenocarcinoma subtype. For example, see, J. Can. Acad. Child Adolesc. Psychiatry 2010; 19(3): 227-229, which is incorporated by reference in its entirety for all purposes.
In one embodiment, a specified statistical confidence level may be determined in order to provide a confidence level regarding the lung cancer subtype. For example, it may be determined that a confidence level of greater than 90% may be a useful predictor of the lung cancer subtype. In other embodiments, more or less stringent confidence levels may be chosen. For example, a confidence level of about or at least about 50%, 60%, 70%, 75%, 80%, 85%, 90%, 95%, 97.5%, 99%, 99.5%, or 99.9% may be chosen. The confidence level provided may in some cases be related to the quality of the sample, the quality of the data, the quality of the analysis, the specific methods used, and/or the number of gene expression values (i.e., the number of genes) analyzed. The specified confidence level for providing the likelihood of response may be chosen on the basis of the expected number of false positives or false negatives. Methods for choosing parameters for achieving a specified confidence level or for identifying markers with diagnostic power include but are not limited to Receiver Operating Characteristic (ROC) curve analysis, binormal ROC, principal component analysis, odds ratio analysis, partial least squares analysis, singular value decomposition, least absolute shrinkage and selection operator analysis, least angle regression, and the threshold gradient directed regularization method.
Determining the lung adenocarcinoma subtype in some cases can be improved through the application of algorithms designed to normalize and or improve the reliability of the gene expression data. In some embodiments of the present invention, the data analysis utilizes a computer or other device, machine or apparatus for application of the various algorithms described herein due to the large number of individual data points that are processed. A “machine learning algorithm” refers to a computational-based prediction methodology, also known to persons skilled in the art as a “classifier,” employed for characterizing a gene expression profile or profiles, e.g., to determine the lung adenocarcinoma subtype. The biomarker levels, determined by, e.g., microarray-based hybridization assays, sequencing assays, NanoString assays, etc., are in one embodiment subjected to the algorithm in order to classify the profile. Supervised learning generally involves “training” a classifier to recognize the distinctions among subtypes such as squamoid (proximal inflammatory) positive, bronchioid (terminal respiratory unit) positive or magnoid (proximal proliferative) positive, and then “testing” the accuracy of the classifier on an independent test set. Therefore, for new, unknown samples the classifier can be used to predict, for example, the class (e.g., squamoid vs bronchioid vs magnoid) in which the samples belong.
In some embodiments, a robust multi-array average (RMA) method may be used to normalize raw data. The RMA method begins by computing background-corrected intensities for each matched cell on a number of microarrays. In one embodiment, the background corrected values are restricted to positive values as described by Irizarry et al. (2003). Biostatistics April 4 (2): 249-64, incorporated by reference in its entirety for all purposes. After background correction, the base-2 logarithm of each background corrected matched-cell intensity is then obtained. The background corrected, log-transformed, matched intensity on each microarray is then normalized using the quantile normalization method in which for each input array and each probe value, the array percentile probe value is replaced with the average of all array percentile points, this method is more completely described by Bolstad et al. Bioinformatics 2003, incorporated by reference in its entirety. Following quantile normalization, the normalized data may then be fit to a linear model to obtain an intensity measure for each probe on each microarray. Tukey's median polish algorithm (Tukey, J. W., Exploratory Data Analysis. 1977, incorporated by reference in its entirety for all purposes) may then be used to determine the log-scale intensity level for the normalized probe set data.
Various other software programs may be implemented. In certain methods, feature selection and model estimation may be performed by logistic regression with lasso penalty using glmnet (Friedman et al. (2010). Journal of statistical software 33(1): 1-22, incorporated by reference in its entirety). Raw reads may be aligned using TopHat (Trapnell et al. (2009). Bioinformatics 25(9): 1105-11, incorporated by reference in its entirety). In methods, top features (N ranging from 10 to 200) are used to train a linear support vector machine (SVM) (Suykens J A K, Vandewalle J. Least Squares Support Vector Machine Classifiers. Neural Processing Letters 1999; 9(3): 293-300, incorporated by reference in its entirety) using the e1071 library (Meyer D. Support vector machines: the interface to libsvm in package e1071. 2014, incorporated by reference in its entirety). Confidence intervals, in one embodiment, are computed using the pROC package (Robin X, Turck N, Hainard A, et al. pROC: an open-source package for R and S+ to analyze and compare ROC curves. BMC bioinformatics 2011; 12: 77, incorporated by reference in its entirety).
In addition, data may be filtered to remove data that may be considered suspect. In one embodiment, data derived from microarray probes that have fewer than about 4, 5, 6, 7 or 8 guanosine+cytosine nucleotides may be considered to be unreliable due to their aberrant hybridization propensity or secondary structure issues. Similarly, data deriving from microarray probes that have more than about 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, or 22 guanosine+cytosine nucleotides may in one embodiment be considered unreliable due to their aberrant hybridization propensity or secondary structure issues.
In some embodiments of the present invention, data from probe-sets may be excluded from analysis if they are not identified at a detectable level (above background).
In some embodiments of the present disclosure, probe-sets that exhibit no, or low variance may be excluded from further analysis. Low-variance probe-sets are excluded from the analysis via a Chi-Square test. In one embodiment, a probe-set is considered to be low-variance if its transformed variance is to the left of the 99 percent confidence interval of the Chi-Squared distribution with (N−1) degrees of freedom. (N−1)*Probe-set Variance/(Gene Probe-set Variance). Chi-Sq(N−1) where N is the number of input CEL files, (N−1) is the degrees of freedom for the Chi-Squared distribution, and the “probe-set variance for the gene” is the average of probe-set variances across the gene. In some embodiments of the present invention, probe-sets for a given mRNA or group of mRNAs may be excluded from further analysis if they contain less than a minimum number of probes that pass through the previously described filter steps for GC content, reliability, variance and the like. For example in some embodiments, probe-sets for a given gene or transcript cluster may be excluded from further analysis if they contain less than about 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or less than about 20 probes.
Methods of biomarker level data analysis in one embodiment, further include the use of a feature selection algorithm as provided herein. In some embodiments of the present invention, feature selection is provided by use of the LIMMA software package (Smyth, G. K. (2005). Limma: linear models for microarray data. In: Bioinformatics and Computational Biology Solutions using R and Bioconductor, R. Gentleman, V. Carey, S. Dudoit, R. Irizarry, W. Huber (eds.), Springer, New York, pages 397-420, incorporated by reference in its entirety for all purposes).
Methods of biomarker level data analysis, in one embodiment, include the use of a pre-classifier algorithm. For example, an algorithm may use a specific molecular fingerprint to pre-classify the samples according to their composition and then apply a correction/normalization factor. This data/information may then be fed in to a final classification algorithm which would incorporate that information to aid in the final diagnosis.
Methods of biomarker level data analysis, in one embodiment, further include the use of a classifier algorithm as provided herein. In one embodiment of the present invention, a diagonal linear discriminant analysis, k-nearest neighbor algorithm, support vector machine (SVM) algorithm, linear support vector machine, random forest algorithm, or a probabilistic model-based method or a combination thereof is provided for classification of microarray data. In some embodiments, identified markers that distinguish samples (e.g., of varying biomarker level profiles, and/or varying molecular subtypes of adenocarcinoma (e.g., squamoid, bronchioid, magnoid)) are selected based on statistical significance of the difference in biomarker levels between classes of interest. In some cases, the statistical significance is adjusted by applying a Benjamin Hochberg or another correction for false discovery rate (FDR).
In some cases, the classifier algorithm may be supplemented with a meta-analysis approach such as that described by Fishel and Kaufman et al. 2007 Bioinformatics 23(13): 1599-606, incorporated by reference in its entirety for all purposes. In some cases, the classifier algorithm may be supplemented with a meta-analysis approach such as a repeatability analysis.
Methods for deriving and applying posterior probabilities to the analysis of biomarker level data are known in the art and have been described for example in Smyth, G. K. 2004 Stat. Appl. Genet. Mol. Biol. 3: Article 3, incorporated by reference in its entirety for all purposes. In some cases, the posterior probabilities may be used in the methods of the present invention to rank the markers provided by the classifier algorithm.
A statistical evaluation of the results of the biomarker level profiling may provide a quantitative value or values indicative of one or more of the following: molecular subtype of adenocarcinoma (squamoid, bronchioid or magnoid); the likelihood of the success of a particular therapeutic intervention, e.g., angiogenesis inhibitor therapy, chemotherapy, or immunotherapy. In one embodiment, the data is presented directly to the physician in its most useful form to guide patient care, or is used to define patient populations in clinical trials or a patient population for a given medication. The results of the molecular profiling can be statistically evaluated using a number of methods known to the art including, but not limited to: the students T test, the two sided T test, Pearson rank sum analysis, hidden Markov model analysis, analysis of q-q plots, principal component analysis, one way ANOVA, two way ANOVA, LIMMA and the like.
In some cases, accuracy may be determined by tracking the subject over time to determine the accuracy of the original diagnosis. In other cases, accuracy may be established in a deterministic manner or using statistical methods. For example, receiver operator characteristic (ROC) analysis may be used to determine the optimal assay parameters to achieve a specific level of accuracy, specificity, positive predictive value, negative predictive value, and/or false discovery rate.
In some cases, the results of the biomarker level profiling assays, are entered into a database for access by representatives or agents of a molecular profiling business, the individual, a medical provider, or insurance provider. In some cases, assay results include sample classification, identification, or diagnosis by a representative, agent or consultant of the business, such as a medical professional. In other cases, a computer or algorithmic analysis of the data is provided automatically. In some cases the molecular profiling business may bill the individual, insurance provider, medical provider, researcher, or government entity for one or more of the following: molecular profiling assays performed, consulting services, data analysis, reporting of results, or database access.
In some embodiments of the present invention, the results of the biomarker level profiling assays are presented as a report on a computer screen or as a paper record. In some embodiments, the report may include, but is not limited to, such information as one or more of the following: the levels of biomarkers (e.g., as reported by copy number or fluorescence intensity, etc.) as compared to the reference sample or reference value(s); the likelihood the subject will respond to a particular therapy, based on the biomarker level values and the lung adenocarcinoma subtype and proposed therapies.
In one embodiment, the results of the gene expression profiling may be classified into one or more of the following: squamoid (proximal inflammatory) positive, bronchioid (terminal respiratory unit) positive, magnoid (proximal proliferative) positive, squamoid (proximal inflammatory) negative, bronchioid (terminal respiratory unit) negative, magnoid (proximal proliferative) negative; likely to respond to angiogenesis inhibitor, immunotherapy or chemotherapy; unlikely to respond to angiogenesis inhibitor, immunotherapy or chemotherapy; or a combination thereof.
In some embodiments of the present invention, results are classified using a trained algorithm. Trained algorithms of the present invention include algorithms that have been developed using a reference set of known gene expression values and/or normal samples, for example, samples from individuals diagnosed with a particular molecular subtype of adenocarcinoma. In some cases, a reference set of known gene expression values are obtained from individuals who have been diagnosed with a particular molecular subtype of adenocarcinoma, and are also known to respond (or not respond) to angiogenesis inhibitor therapy. In some cases, a reference set of known gene expression values are obtained from individuals who have been diagnosed with a particular molecular subtype of adenocarcinoma, and are also known to respond (or not respond) to immunotherapy. In some cases, a reference set of known gene expression values are obtained from individuals who have been diagnosed with a particular molecular subtype of adenocarcinoma, and are also known to respond (or not respond) to chemotherapy.
Algorithms suitable for categorization of samples include but are not limited to k-nearest neighbor algorithms, support vector machines, linear discriminant analysis, diagonal linear discriminant analysis, updown, naive Bayesian algorithms, neural network algorithms, hidden Markov model algorithms, genetic algorithms, or any combination thereof.
When a binary classifier is compared with actual true values (e.g., values from a biological sample), there are typically four possible outcomes. If the outcome from a prediction is p (where “p” is a positive classifier output, such as the presence of a deletion or duplication syndrome) and the actual value is also p, then it is called a true positive (TP); however if the actual value is n then it is said to be a false positive (FP). Conversely, a true negative has occurred when both the prediction outcome and the actual value are n (where “n” is a negative classifier output, such as no deletion or duplication syndrome), and false negative is when the prediction outcome is n while the actual value is p. In one embodiment, consider a test that seeks to determine whether a person is likely or unlikely to respond to angiogenesis inhibitor therapy. A false positive in this case occurs when the person tests positive, but actually does respond. A false negative, on the other hand, occurs when the person tests negative, suggesting they are unlikely to respond, when they actually are likely to respond. The same holds true for classifying a lung cancer subtype.
The positive predictive value (PPV), or precision rate, or post-test probability of disease, is the proportion of subjects with positive test results who are correctly diagnosed as likely or unlikely to respond, or diagnosed with the correct lung cancer subtype, or a combination thereof. It reflects the probability that a positive test reflects the underlying condition being tested for. Its value does however depend on the prevalence of the disease, which may vary. In one example the following characteristics are provided: FP (false positive); TN (true negative); TP (true positive); FN (false negative). False positive rate (α)=FP/(FP+TN)-specificity; False negative rate (β)=FN/(TP+FN)-sensitivity; Power=sensitivity=1−β; Likelihood-ratio positive=sensitivity/(1−specificity); Likelihood-ratio negative=(1−sensitivity)/specificity. The negative predictive value (NPV) is the proportion of subjects with negative test results who are correctly diagnosed.
In some embodiments, the results of the biomarker level analysis of the subject methods provide a statistical confidence level that a given diagnosis is correct. In some embodiments, such statistical confidence level is at least about, or more than about 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% 99.5%, or more.
In some embodiments, the method further includes classifying the lung tissue sample as a particular lung cancer subtype based on the comparison of biomarker levels in the sample and reference biomarker levels, for example present in at least one training set. In some embodiments, the lung tissue sample is classified as a particular subtype if the results of the comparison meet one or more criterion such as, for example, a minimum percent agreement, a value of a statistic calculated based on the percentage agreement such as (for example) a kappa statistic, a minimum correlation (e.g., Pearson's correlation) and/or the like.
It is intended that the methods described herein can be performed by software (stored in memory and/or executed on hardware), hardware, or a combination thereof. Hardware modules may include, for example, a general-purpose processor, a field programmablegate array (FPGA), and/or an application specific integrated circuit (ASIC). Software modules (executed on hardware) can be expressed in a variety of software languages (e.g., computer code), including Unix utilities, C, C++, Java™, Ruby, SQL, SAS®, the R programming language/software environment, Visual Basic™, and other object-oriented, procedural, or other programming language and development tools. Examples of computer code include, but are not limited to, micro-code or micro-instructions, machine instructions, such as produced by a compiler, code used to produce a web service, and files containing higher-level instructions that are executed by a computer using an interpreter. Additional examples of computer code include, but are not limited to, control signals, encrypted code, and compressed code.
Some embodiments described herein relate to devices with a non-transitory computer-readable medium (also can be referred to as a non-transitory processor-readable medium or memory) having instructions or computer code thereon for performing various computer-implemented operations and/or methods disclosed herein. The computer-readable medium (or processor-readable medium) is non-transitory in the sense that it does not include transitory propagating signals per se (e.g., a propagating electromagnetic wave carrying information on a transmission medium such as space or a cable). The media and computer code (also can be referred to as code) may be those designed and constructed for the specific purpose or purposes. Examples of non-transitory computer-readable media include, but are not limited to: magnetic storage media such as hard disks, floppy disks, and magnetic tape; optical storage media such as Compact Disc/Digital Video Discs (CD/DVDs), Compact Disc-Read Only Memories (CD-ROMs), and holographic devices; magneto-optical storage media such as optical disks; carrier wave signal processing modules; and hardware devices that are specially configured to store and execute program code, such as Application-Specific Integrated Circuits (ASICs), Programmable Logic Devices (PLDs), Read-Only Memory (ROM) and Random-Access Memory (RAM) devices. Other embodiments described herein relate to a computer program product, which can include, for example, the instructions and/or computer code discussed herein.
In some embodiments, a single biomarker, or from about 5 to about 10, from about 8 to about 16, from about 5 to about 15, from about 5 to about 20, from about 5 to about 25, from about 5 to about 30, from about 5 to about 35, from about 5 to about 40, from about 5 to about 45, from about 5 to about 48 biomarkers (e.g., as disclosed in Table 1) is capable of classifying subtypes of lung adenocarcinoma with a predictive success of at least about 70%, at least about 71%, at least about 72%, about 73%, about 74%, about 75%, about 76%, about 77%, about 78%, about 79%, about 80%, about 81%, about 82%, about 83%, about 84%, about 85%, about 86%, about 87%, about 88%, about 89%, about 90%, about 91%, about 92%, about 93%, about 94%, about 95%, about 96%, about 97%, about 98%, about 99%, up to 100%, and all values in between. In some embodiments, any combination of biomarkers disclosed herein (e.g., in Table 1) can be used to obtain a predictive success of at least about 70%, at least about 71%, at least about 72%, about 73%, about 74%, about 75%, about 76%, about 77%, about 78%, about 79%, about 80%, about 81%, about 82%, about 83%, about 84%, about 85%, about 86%, about 87%, about 88%, about 89%, about 90%, about 91%, about 92%, about 93%, about 94%, about 95%, about 96%, about 97%, about 98%, about 99%, up to 100%, and all values in between.
In some embodiments, a single biomarker, or from about 5 to about 10, from about 8 to about 16, from about 5 to about 15, from about 5 to about 20, from about 5 to about 25, from about 5 to about 30, from about 5 to about 35, from about 5 to about 40, from about 5 to about 45, from about 5 to about 48 biomarkers (e.g., as disclosed in Table 1) is capable of classifying lung adenocarcinoma subtypes with a sensitivity or specificity of at least about 70%, at least about 71%, at least about 72%, about 73%, about 74%, about 75%, about 76%, about 77%, about 78%, about 79%, about 80%, about 81%, about 82%, about 83%, about 84%, about 85%, about 86%, about 87%, about 88%, about 89%, about 90%, about 91%, about 92%, about 93%, about 94%, about 95%, about 96%, about 97%, about 98%, about 99%, up to 100%, and all values in between. In some embodiments, any combination of biomarkers disclosed herein can be used to obtain a sensitivity or specificity of at least about 70%, at least about 71%, at least about 72%, about 73%, about 74%, about 75%, about 76%, about 77%, about 78%, about 79%, about 80%, about 81%, about 82%, about 83%, about 84%, about 85%, about 86%, about 87%, about 88%, about 89%, about 90%, about 91%, about 92%, about 93%, about 94%, about 95%, about 96%, about 97%, about 98%, about 99%, up to 100%, and all values in between.
In one embodiment, the methods and compositions provided herein are useful for determining the AD subtype of a sample (e.g., lung tissue sample) from a patient by analyzing the expression of a set of biomarkers, whereby the set of biomarkers comprise a fewer number of biomarkers that methods known in the art for molecularly classifying lung AD subtype. In some cases, the set of biomarkers is less than 250, 240, 230, 220, 210, 200, 150, 100, 95 or 90 biomarkers. In some cases, the set of biomarkers is less than 50 biomarkers. In some cases, the set of biomarkers is the set of 48 biomarkers listed in Table 1. In some cases, the set of biomarkers is a sub-set of biomarkers listed Table 1. The biomarkers or classifier genes useful in the methods and compositions provided herein can be selected from one or more lung adenocarcinoma datasets from one or more databases. The databases can be public databases. In one embodiment, classifier genes (e.g., one or more genes listed in Table 1 and Table 2) useful in the methods and compositions provided herein for detecting or diagnosing lung adenocarcinoma subtypes were selected from a lung adenocarcinoma RNAseq dataset from The Cancer Genome Atlas (TCGA). In one embodiment, classifier genes useful for the methods and compositions provided herein such as those in Table 1 are selected by subjecting a large set of classifier genes to an in silico based process in order to determine the minimum number of genes whose expression profile can be used to determine an AD subtype of sample obtained from a subject. In some cases, the large set of classifier genes can be a lung AD RNAseq dataset such as, for example, from TCGA. In some cases, the large set of classifier genes can be 506-gene classifier described herein, whereby the 506-gene classifier can serve to define gold standard subtype. The in silico process for selecting a gene cassette as provided herein for determining lung AD subtype of a sample from a patient can comprise, applying or using a Classifying arrays to Nearest Centroid (CLaNC) algorithm with modification on the standard 506 classifier genes to choose an equal number of negatively and positively correlated genes for each subtype. For determination of the optimal number of genes (e.g, 16 per subtype as shown in Table 1) to include in the signature, the process can further comprise performing a 5-fold cross validation using TCGA lung adenocarcinoma dataset as provided herein to produce cross-validation curves as shown in
In one embodiment, the method further comprises validating the gene classifiers. Validation can comprise testing the expression of the classifiers in several fresh frozen publicly available array and RNAseq datasets and calling the subtype based on said expression levels and subsequently comparing the expression with the gold standard subtype calls as defined by the previously published 506-gene signature. Final validation of the gene signature (e.g., Table 1) can then be performed in a newly collected RNAseq dataset of archived formalin-fixed paraffin-embedded (FFPE) adenocarcinoma samples to assure comparable performance in the FFPE samples. In one embodiment, the classifier biomarkers of Table 1 were selected based on the in silico CLaNC process described herein. The gene symbols and official gene names are listed in column 2 and column 3, respectively.
In one embodiment, the methods of the invention require the detection of at least 1, at least 2, at least 3, at least 4, at least 5, at least 6, at least 7, at least 8, at least 9, at least 10, at least 11, at least 12, at least 13, at least 14, at least 15, or up to 16 classifier biomarkers in a lung cancer cell sample obtained from a patient which expression is altered in order to identify a TRU, a PP, or a PI lung adenocarcinoma subtype. The same applies for other classifier gene expression datasets as provided herein.
In another embodiment, the methods of the invention require the detection of a total of at least 1, at least 2, at least 5, at least 8, at least 10, at least 16, at least 20, at least 30, at least 32, or up to 48 classifier biomarkers out of the 48 gene biomarkers of Table 1 in a lung cancer cell sample (e.g., lung AD sample) obtained from a patient in order to identify a TRU, a PP, or a PI lung adenocarcinoma subtype. The same applies for other classifier gene expression datasets as provided herein.
In one embodiment, at least 1, at least 2, at least 3, at least 4, at least 5, at least 6, at least 7, or up to 8 biomarkers of Table 1 are “up-regulated” in a specific subtype of lung adenocarcinoma. In another embodiment, at least 1, at least 2, at least 3, at least 4, at least 5, at least 6, at least 7, or up to 8 biomarkers of Table 1 are “down-regulated” in a specific subtype of lung adenocarcinoma. The same applies for other classifier gene expression datasets as provided herein.
In one embodiment, the expression level of an “up-regulated” biomarker as provided herein is increased by about 0.5-fold, about 1-fold, about 1.5-fold, about 2-fold, about 2.5-fold, about 3-fold, about 3.5-fold, about 4-fold, about 4.5-fold, about 5-fold, and any values in between. In another embodiment, the expression level of a “down-regulated” biomarker as provided herein is decreased by about 0.8-fold, about 1.4-fold, about 2-fold, about 2.6-fold, about 3.2-fold, about 3.6-fold, about 4-fold, and any values in between.
It is recognized that additional genes or proteins can be used in the practice of the invention. For example, vimentin, a member of the intermediate filament family of proteins can be used to identify the adenocarcinoma subtype Proximal Proliferative (magnoid), and SMA can be used to identify Proximal Inflammatory (squamoid) subtype. In general, genes useful in classifying the subtypes of lung adenocarcinoma, include those that are independently capable of distinguishing between normal versus tumor, or between different classes or grades of lung cancer. A gene is considered to be capable of reliably distinguishing between subtypes if the area under the receiver operator characteristic (ROC) curve is approximately 1.
In one embodiment, a method is provided herein for determining a disease outcome or prognosis for a patient suffering from cancer. In some cases, the cancer is lung cancer. The disease outcome or prognosis can be measured by examining the overall survival for a period of time or intervals (e.g., 0 to 36 months or 0 to 60 months). In one embodiment, survival is analyzed as a function of subtype (e.g., for lung cancer, adenocarcinoma (TRU, PI, and PP)). Relapse-free and overall survival can be assessed using standard Kaplan-Meier plots as well as Cox proportional hazards modeling.
In one embodiment, upon determining a patient's lung cancer subtype, the patient is selected for suitable therapy, for example chemotherapy or drug therapy with an angiogenesis inhibitor or immunotherapy. In one embodiment, upon determining a patient's lung cancer subtype, the patient is administered a suitable therapeutic agent, for example chemotherapeutic agent(s) or an angiogenesis inhibitor or immunotherapeutic agent(s). In one embodiment, the therapy is immunotherapy, and the immunotherapeutic agent is a checkpoint inhibitor, monoclonal antibody, biological response modifier, therapeutic vaccine or cellular immunotherapy.
The methods of present invention are also useful for evaluating clinical response to therapy, as well as for endpoints in clinical trials for efficacy of new therapies. The extent to which sequential diagnostic expression profiles move towards normal can be used as one measure of the efficacy of the candidate therapy.
In one embodiment, the methods of the invention also find use in predicting response to different lines of therapies based on the subtype of lung adenocarcinoma (AD). For example, chemotherapeutic response can be improved by more accurately assigning tumor subtypes. Likewise, treatment regimens can be formulated based on the tumor subtype. For example, clinical trials have shown convincing evidence that the VEGF inhibitor, bevacizumab, can be effective in the treatment of NSCLC.
In one embodiment, the Terminal Respiratory Unit (TRU) subtype may have enhanced response to EGFR inhibitors and Pemetrexed. In another embodiment, Proximal Proliferative (PP) can have enhanced response to chemotherapy. In another embodiment, Proximal Inflammatory (PI) can have enhanced response to immunotherapy. In another embodiment, all subtypes can have enhanced response to chemotherapies, angiogenesis inhibitor treatments, and immunotherapies.
Angiogenesis Inhibitors
In one embodiment, upon determining a patient's lung adenocarcinoma subtype, the patient is selected for drug therapy with an angiogenesis inhibitor.
In one embodiment, the angiogenesis inhibitor is a vascular endothelial growth factor (VEGF) inhibitor, a VEGF receptor inhibitor, a platelet derived growth factor (PDGF) inhibitor or a PDGF receptor inhibitor.
Each biomarker panel can include one, two, three, four, five, six, seven, eight or more biomarkers usable by a classifier (also referred to as a “classifier biomarker”) to assess whether an adenocarcinoma patient is likely to respond to angiogenesis inhibitor therapy; to select an adenocarcinoma patient for angiogenesis inhibitor therapy; to determine a “hypoxia score” and/or to subtype an adenocarcinoma sample as squamoid (also referred to as proximal inflammatory), bronchioid (also referred to as terminal respiratory unit) or magnoid (also referred to as proximal proliferative) molecular subtype. As used herein, the term “classifier” can refer to any algorithm for statistical classification, and can be implemented in hardware, in software, or a combination thereof. The classifier can be capable of 2-level, 3-level, 4-level, or higher, classification, and can depend on the nature of the entity being classified. One or more classifiers can be employed to achieve the aspects disclosed herein.
In general, methods of determining whether an adenocarcinoma patient is likely to respond to angiogenesis inhibitor therapy, or methods of selecting an adenocarcinoma patient for angiogenesis inhibitor therapy are provided herein. In one embodiment, the method comprises assessing whether the patient's adenocarcinoma subtype is squamoid (proximal inflammatory), bronchioid (terminal respiratory unit) or magnoid (proximal proliferative) using the methods described herein (e.g., assessing the expression of one or more classifier biomarkers of Table 1) and probing an adenocarcinoma sample from the patient for the levels of at least five biomarkers selected from the group consisting of RRAGD, FABP5, UCHL1, GAL, PLOD, DDIT4, VEGF, ADM, ANGPTL4, NDRG1, NP, SLC16A3, and C140RF58 (see Table 3) at the nucleic acid level. In a further embodiment, the probing step comprises mixing the sample with five or more oligonucleotides that are substantially complementary to portions of nucleic acid molecules of the at least five biomarkers under conditions suitable for hybridization of the five or more oligonucleotides to their complements or substantial complements, detecting whether hybridization occurs between the five or more oligonucleotides to their complements or substantial complements; and obtaining hybridization values of the sample based on the detecting steps. The hybridization values of the sample are then compared to reference hybridization value(s) from at least one sample training set, wherein the at least one sample training set comprises (i) hybridization value(s) of the at least five biomarkers from a sample that overexpresses the at least five biomarkers, or overexpresses a subset of the at least five biomarkers, (ii) hybridization values of the at least five biomarkers from a reference squamoid (proximal inflammatory), bronchioid (terminal respiratory unit) or magnoid (proximal proliferative) sample, or (iii) hybridization values of the at least five biomarkers from an adenocarcinoma free lung sample. A determination of whether the patient is likely to respond to angiogenesis inhibitor therapy, or a selection of the patient for angiogenesis inhibitor is then made based upon (i) the patient's adenocarcinoma subtype and (ii) the results of comparison.
The aforementioned set of thirteen biomarkers, or a subset thereof, is also referred to herein as a “hypoxia profile”.
In one embodiment, the method provided herein includes determining the levels of at least five biomarkers, at least six biomarkers, at least seven biomarkers, at least eight biomarkers, at least nine biomarkers, or at least ten biomarkers, or five to thirteen, six to thirteen, seven to thirteen, eight to thirteen, nine to thirteen or ten to thirteen biomarkers selected from RRAGD, FABP5, UCHL1, GAL, PLOD, DDIT4, VEGF, ADM, ANGPTL4, NDRG1, NP, SLC16A3, and C14ORF58 in an adenocarcinoma sample obtained from a subject. Biomarker expression in some instances may be normalized against the expression levels of all RNA transcripts or their expression products in the sample, or against a reference set of RNA transcripts or their expression products. The reference set as explained throughout, may be an actual sample that is tested in parallel with the adenocarcinoma sample, or may be a reference set of values from a database or stored dataset. Levels of expression, in one embodiment, are reported in number of copies, relative fluorescence value or detected fluorescence value. The level of expression of the biomarkers of the hypoxia profile together with adenocarcinoma subtype as determined using the methods provided herein can be used in the methods described herein to determine whether a patient is likely to respond to angiogenesis inhibitor therapy.
In one embodiment, the levels of expression of the thirteen biomarkers (or subsets thereof, as described above, e.g., five or more, from about five to about 13), are normalized against the expression levels of all RNA transcripts or their non-natural cDNA expression products, or protein products in the sample, or of a reference set of RNA transcripts or a reference set of their non-natural cDNA expression products, or a reference set of their protein products in the sample.
In one embodiment, angiogenesis inhibitor treatments include, but are not limited to an integrin antagonist, a selectin antagonist, an adhesion molecule antagonist, an antagonist of intercellular adhesion molecule (ICAM)-1, ICAM-2, ICAM-3, platelet endothelial adhesion molecule (PCAM), vascular cell adhesion molecule (VCAM)), lymphocyte function-associated antigen 1 (LFA-1), a basic fibroblast growth factor antagonist, a vascular endothelial growth factor (VEGF) modulator, a platelet derived growth factor (PDGF) modulator (e.g., a PDGF antagonist).
In one embodiment of determining whether a subject is likely to respond to an integrin antagonist, the integrin antagonist is a small molecule integrin antagonist, for example, an antagonist described by Paolillo et al. (Mini Rev Med Chem, 2009, volume 12, pp. 1439-1446, incorporated by reference in its entirety), or a leukocyte adhesion-inducing cytokine or growth factor antagonist (e.g., tumor necrosis factor-α (TNF-α), interleukin-1β (IL-1β), monocyte chemotactic protein-1 (MCP-1) and a vascular endothelial growth factor (VEGF)), as described in U.S. Pat. No. 6,524,581, incorporated by reference in its entirety herein.
The methods provided herein are also useful for determining whether a subject is likely to respond to one or more of the following angiogenesis inhibitors: interferon gamma 1β, interferon gamma 1β (Actimmune®) with pirfenidone, ACUHTR028, αVβ5, aminobenzoate potassium, amyloid P, ANG1122, ANG1170, ANG3062, ANG3281, ANG3298, ANG4011, anti-CTGF RNAi, Aplidin, Astragalus membranaceus extract with salvia and Schisandra chinensis, atherosclerotic plaque blocker, Azol, AZX100, BB3, connective tissue growth factor antibody, CT140, danazol, Esbriet, EXC001, EXC002, EXC003, EXC004, EXC005, F647, FG3019, Fibrocorin, Follistatin, FT011, a galectin-3 inhibitor, GKT137831, GMCT01, GMCT02, GRMD01, GRMD02, GRN510, Heberon Alfa R, interferon α-2β, ITMN520, JKB119, JKB121, JKB122, KRX168, LPA1 receptor antagonist, MGN4220, MIA2, microRNA 29a oligonucleotide, MMI0100, noscapine, PBI4050, PBI4419, PDGFR inhibitor, PF-06473871, PGN0052, Pirespa, Pirfenex, pirfenidone, plitidepsin, PRM151, Px102, PYN17, PYN22 with PYN17, Relivergen, rhPTX2 fusion protein, RXI109, secretin, STX100, TGF-β Inhibitor, transforming growth factor, β-receptor 2 oligonucleotide, VA999260, XV615 or a combination thereof.
In another embodiment, a method is provided for determining whether a subject is likely to respond to one or more endogenous angiogenesis inhibitors. In a further embodiment, the endogenous angiogenesis inhibitor is endostatin, a 20 kDa C-terminal fragment derived from type XVIII collagen, angiostatin (a 38 kDa fragment of plasmin), a member of the thrombospondin (TSP) family of proteins. In a further embodiment, the angiogenesis inhibitor is a TSP-1, TSP-2, TSP-3, TSP-4 and TSP-5. Methods for determining the likelihood of response to one or more of the following angiogenesis inhibitors are also provided a soluble VEGF receptor, e.g., soluble VEGFR-1 and neuropilin 1 (NPR1), angiopoietin-1, angiopoietin-2, vasostatin, calreticulin, platelet factor-4, a tissue inhibitor of metalloproteinase (TIMP) (e.g., TIMP1, TIMP2, TIMP3, TIMP4), cartilage-derived angiogenesis inhibitor (e.g., peptide troponin I and chrondomodulin I), a disintegrin and metalloproteinase with thrombospondin motif 1, an interferon (IFN), (e.g., IFN-α, IFN-β, IFN-γ), a chemokine, e.g., a chemokine having the C-X-C motif (e.g., CXCL10, also known as interferon gamma-induced protein 10 or small inducible cytokine B10), an interleukin cytokine (e.g., IL-4, IL-12, IL-18), prothrombin, antithrombin III fragment, prolactin, the protein encoded by the TNFSF15 gene, osteopontin, maspin, canstatin, proliferin-related protein.
In one embodiment, a method for determining the likelihood of response to one or more of the following angiogenesis inhibitors is provided is angiopoietin-1, angiopoietin-2, angiostatin, endostatin, vasostatin, thrombospondin, calreticulin, platelet factor-4, TIMP, CDAI, interferon α, interferon β, vascular endothelial growth factor inhibitor (VEGI) meth-1, meth-2, prolactin, VEGI, SPARC, osteopontin, maspin, canstatin, proliferin-related protein (PRP), restin, TSP-1, TSP-2, interferon gamma 1β, ACUHTR028, αVβ5, aminobenzoate potassium, amyloid P, ANG1122, ANG1170, ANG3062, ANG3281, ANG3298, ANG4011, anti-CTGF RNAi, Aplidin, Astragalus membranaceus extract with salvia and Schisandra chinensis, atherosclerotic plaque blocker, Azol, AZX100, BB3, connective tissue growth factor antibody, CT140, danazol, Esbriet, EXC001, EXC002, EXC003, EXC004, EXC005, F647, FG3019, Fibrocorin, Follistatin, FT011, a galectin-3 inhibitor, GKT137831, GMCT01, GMCT02, GRMD01, GRMD02, GRN510, Heberon Alfa R, interferon α-2β, ITMN520, JKB119, JKB121, JKB122, KRX168, LPA1 receptor antagonist, MGN4220, MIA2, microRNA 29a oligonucleotide, MMI0100, noscapine, PBI4050, PBI4419, PDGFR inhibitor, PF-06473871, PGN0052, Pirespa, Pirfenex, pirfenidone, plitidepsin, PRM151, Px102, PYN17, PYN22 with PYN17, Relivergen, rhPTX2 fusion protein, RXI109, secretin, STX100, TGF-β Inhibitor, transforming growth factor, β-receptor 2 oligonucleotide, VA999260, XV615 or a combination thereof.
In yet another embodiment, the angiogenesis inhibitor can include pazopanib (Votrient), sunitinib (Sutent), sorafenib (Nexavar), axitinib (Inlyta), ponatinib (Iclusig), vandetanib (Caprelsa), cabozantinib (Cometrig), ramucirumab (Cyramza), regorafenib (Stivarga), ziv-aflibercept (Zaltrap), motesanib, or a combination thereof. In another embodiment, the angiogenesis inhibitor is a VEGF inhibitor. In a further embodiment, the VEGF inhibitor is axitinib, cabozantinib, aflibercept, brivanib, tivozanib, ramucirumab or motesanib. In yet a further embodiment, the angiogenesis inhibitor is motesanib.
In one embodiment, the methods provided herein relate to determining a subject's likelihood of response to an antagonist of a member of the platelet derived growth factor (PDGF) family, for example, a drug that inhibits, reduces or modulates the signaling and/or activity of PDGF-receptors (PDGFR). For example, the PDGF antagonist, in one embodiment, is an anti-PDGF aptamer, an anti-PDGF antibody or fragment thereof, an anti-PDGFR antibody or fragment thereof, or a small molecule antagonist. In one embodiment, the PDGF antagonist is an antagonist of the PDGFR-α or PDGFR-β. In one embodiment, the PDGF antagonist is the anti-PDGF-β aptamer E10030, sunitinib, axitinib, sorefenib, imatinib, imatinib mesylate, nintedanib, pazopanib HCl, ponatinib, MK-2461, dovitinib, pazopanib, crenolanib, PP-121, telatinib, imatinib, KRN 633, CP 673451, TSU-68, Ki8751, amuvatinib, tivozanib, masitinib, motesanib diphosphate, dovitinib dilactic acid, linifanib (ABT-869).
Upon making a determination of whether a patient is likely to respond to angiogenesis inhibitor therapy, or selecting a patient for angiogenesis inhibitor therapy, in one embodiment, the patient is administered the angiogenesis inhibitor. The angiogenesis in inhibitor can be any of the angiogenesis inhibitors described herein.
Immunotherapy
In one embodiment, provided herein is a method for determining whether an adenocarcinoma (AD) lung cancer patient is likely to respond to immunotherapy by determining the subtype of AD of a sample obtained from the patient and, based on the AD lung cancer subtype, assessing whether the patient is likely to respond to immunotherapy. In another embodiment, provided herein is a method of selecting a patient suffering from AD for immunotherapy by determining an AD subtype of a sample from the patient and, based on the AD subtype, selecting the patient for immunotherapy. The determination of the AD subtype of the sample obtained from the patient can be performed using any method for subtyping AD known in the art. In one embodiment, the sample obtained from the patient has been previously diagnosed as being AD, and the methods provided herein are used to determine the AD subtype of the sample. The previous diagnosis can be based on a histological analysis. The histological analysis can be performed by one or more pathologists. In one embodiment, the AD subtyping is performed via gene expression analysis of a set or panel of biomarkers or subsets thereof in order to generate an expression profile. The gene expression analysis can be performed on a lung cancer sample (e.g., lung cancer AD sample) obtained from a patient in order to determine the presence, absence or level of expression of one or more biomarkers selected from a publically available lung cancer database described herein and/or Table 1 provided herein. The AD subtype can be selected from the group consisting of squamoid (proximal inflammatory), bronchioid (terminal respiratory unit) and magnoid (proximal proliferative). The immunotherapy can be any immunotherapy provided herein. In one embodiment, the immunotherapy comprises administering one or more checkpoint inhibitors. The checkpoint inhibitors can be any checkpoint inhibitor provided herein such as, for example, a checkpoint inhibitor that targets PD-1, PD-LI or CTLA4.
As disclosed herein, the biomarkers panels, or subsets thereof, can be those disclosed in any publically available AD gene expression dataset or datasets. In one embodiment, the lung cancer is AD and the biomarker panel or subset thereof is, for example, the cancer genome atlas (TCGA) lung AD RNAseq gene expression dataset (n=515). In one embodiment, the lung cancer is AD and the biomarker panel or subset thereof is, for example, the AD gene expression dataset (n=442) disclosed in Shedden et al. (Nat Med 2008; 14(8): 822-827), the contents of which are herein incorporated by reference in its entirety. In one embodiment, the lung cancer is AD and the biomarker panel or subset thereof is, for example, the AD gene expression dataset (n=117) disclosed in Tomida et al. (J Clin Oncol 2009; 27(17):2793-2799), the contents of which are herein incorporated by reference in its entirety. In one embodiment, the lung cancer is AD and the biomarker panel or subset thereof is, for example, the AD gene expression dataset (n=116) disclosed in Wilkerson et al. (PLoS One 2012; 7(5):e36530), the contents of which are herein incorporated by reference in its entirety. In one embodiment, the lung cancer is AD and the biomarker panel or subset thereof is, for example, the AD gene expression dataset disclosed in Table 1. In one embodiment, the lung cancer is AD and the biomarker panel or subset thereof is, for example, the AD gene expression dataset disclosed in Table 1 in combination with one or more biomarkers from a publically available AD expression dataset. In Table 2, the first column of the table represents the biomarker list for distinguishing Terminal Respiratory Unit (TRU). The middle column of the table represents the biomarker list for distinguishing Proximal Proliferative (PP). The last column of the table represents the biomarker list for distinguishing Proximal Inflammatory (PI). In some cases, as shown in Table 2, a total of 48 biomarkers can be used for AD subtype determination. For each AD subtype, 8 of the 16 biomarkers can be negatively correlated genes, while 8 can be positively correlated genes which can be selected as the gene signature of a specific AD subtype.
In some embodiments, the method for lung cancer subtyping (e.g., AD subtyping) includes detecting expression levels of a classifier biomarker set. The classifier biomarker set can be a set of biomarkers from a publically available database such as, for example, TCGA lung AD RNASeq gene expression dataset(s) or any other dataset provided herein. In some embodiments, the detecting includes all of the classifier biomarkers of Table 1 or any other dataset provided herein at the nucleic acid level or protein level. In another embodiment, a single classifier biomarker of Table 1 or a subset of the classifier biomarkers of Table 1 or any other dataset provided herein are detected, for example, from about five to about twenty. In another embodiment, a single classifier biomarker of Table 1 or a subset of the classifier biomarkers of Table 1 and/or any other dataset provided herein are detected, for example, from about 16 to about 48. In another embodiment, all of the classifier biomarkers of Table 1 or any other dataset provided herein are detected. In another embodiment, at least one or all of the classifier biomarkers of Table 1 in combination with one or more classifier biomarkers of any other AD dataset provided herein are detected. The detecting can be performed by any suitable technique including, but not limited to, RNA-seq, a reverse transcriptase polymerase chain reaction (RT-PCR), a microarray hybridization assay, or another hybridization assay, e.g., a NanoString assay for example, with primers and/or probes specific to the classifier biomarkers, and/or the like. In some cases, the primers useful for the amplification methods (e.g., RT-PCR or qRT-PCR) are any forward and reverse primers suitable for binding to a classifier gene from a dataset provided herein alone or in combination.
In one embodiment, from about 1 to about 5, about 5 to about 10, from about 5 to about 15, from about 5 to about 20, from about 5 to about 25, from about 5 to about 30, from about 5 to about 35, from about 5 to about 40, from about 5 to about 45, from about 5 to about 50, from about 5 to about 55, from about 5 to about 60, from about 5 to about 65, from about 5 to about 70, from about 5 to about 75, or from about 5 to about 80 of the biomarkers in any of the AD gene expression datasets provided herein, including, for example, Table 1 for an AD lung sample are detected in a method to determine the lung cancer subtype as provided herein. In another embodiment, each of the biomarkers from any one of the AD gene expression datasets provided herein, including, for example, Table 1 for an AD lung sample are detected in a method to determine the lung cancer subtype as provided herein.
In one embodiment, the methods provided herein further comprise determining the presence, absence or level of immune activation in an AD subtype. The presence or level of immune cell activation can be determined by creating an expression profile or detecting the expression of one or more biomarkers associated with innate immune cells and/or adaptive immune cells associated with each AD subtype in a sample (e.g., lung cancer sample) obtained from a patient. In one embodiment, immune cell activation associated with an AD subtype is determined by monitoring the immune cell signatures of Bindea et al (Immunity 2013; 39(4); 782-795), the contents of which are herein incorporated by reference in its entirety. In one embodiment, the method further comprises measuring single gene immune biomarkers, such as, for example, CTLA4, PDCD1 and CD274 (PD-LI), PDCDLG2 (PD-L2) and/or IFN gene signatures. The presence or a detectable level of immune activation (Innate and/or Adaptive) associated with an AD subtype can indicate or predict that a patient with said AD subtype may be amendable to immunotherapy. The immunotherapy can be treatment with a checkpoint inhibitor as provided herein. In one embodiment, the PI subtype of AD has immune expression. In one embodiment, a method is provided herein for detecting the expression of at least one classifier biomarker provided herein in a sample (e.g., lung cancer AD sample) obtained from a patient further comprises administering an immunotherapeutic agent following detection of immune activation as provided herein in said sample.
In one embodiment, the method comprises determining a subtype of a lung cancer AD sample and subsequently determining a level of immune cell activation of said sub-type. In one embodiment, the subtype is determined by determining the expression levels of one or more classifier biomarkers using sequencing (e.g., RNASeq), amplification (e.g., qRT-PCR) or hybridization assays (e.g., microarray analysis) as described herein. The one or more biomarkers can be selected from a publically available database (e.g., TCGA lung AD RNASeq gene expression datasets or any other publically available AD gene expression datasets provided herein). In some embodiments, the biomarkers of Table 1 can be used to specifically determine the subtype of an AD lung sample obtained from a patient. In one embodiment, the level of immune cell activation is determined by measuring gene expression signatures of immunomarkers. The immunomarkers can be measured in the same and/or different sample used to subtype the lung cancer sample as described herein. The immunomarkers that can be measured can comprise, consist of, or consistently essentially of innate immune cell (IIC) and/or adaptive immune cell (AIC) gene signatures, interferon (IFN) gene signatures, individual immunomarkers, major histocapability complex class II (MEW class II) genes or a combination thereof. The gene expression signatures for both IICs and AICs can be any known gene signatures for said cell types known in the art. For example, the immune gene signatures can be those from Bindea et al. (Immunity 2013; 39(4); 782-795). In one embodiment, the immunomarkers for use in the methods provided herein are selected from Table 4A and/or Table 4B. The individual immunomarkers can be CTLA4, PDCD1 and CD274 (PD-L1). In one embodiment, the individual immunomarkers for use in the methods provided herein are selected from Table 5. The immunomarkers can be one or more interferon (INF) genes. In one embodiment, the immunomarkers for use in the methods provided herein are selected from Table 6. The immunomarkers can be one or more WWII genes. In one embodiment, the immunomarkers for use in the methods provided herein are selected from Table 7. In yet another embodiment, the immunomarkers for use in the methods provided herein are selected from Tables 4A, 4B, 5, 6, 7, or a combination thereof.
Homo sapiens
Homo sapiens CD74 molecule (CD74)
Homo sapiens major histocompatibility
Homo sapiens major histocompatibility
Homo sapiens major histocompatibility
Homo sapiens major histocompatibility
Homo sapiens major histocompatibility
Homo sapiens major histocompatibility
Homo sapiens nuclear
In one embodiment, upon determining a patient's AD lung cancer subtype using any of the methods and classifier biomarkers panels or subsets thereof as provided herein alone or in combination with determining expression of one or more immune cell markers as provided herein, the patient is selected for treatment with or administered an immunotherapeutic agent. The immunotherapeutic agent can be a checkpoint inhibitor, monoclonal antibody, biological response modifiers, therapeutic vaccine or cellular immunotherapy.
In another embodiment, the immunotherapeutic agent is a checkpoint inhibitor. In some cases, a method for determining the likelihood of response to one or more checkpoint inhibitors is provided. In one embodiment, the checkpoint inhibitor is a PD-1/PD-LI checkpoint inhibitor. The PD-1/PD-LI checkpoint inhibitor can be nivolumab, pembrolizumab, atezolizumab, durvalumab, lambrolizumab, or avelumab. In one embodiment, the checkpoint inhibitor is a CTLA-4 checkpoint inhibitor. The CTLA-4 checkpoint inhibitor can be ipilimumab or tremelimumab. In one embodiment, the checkpoint inhibitor is a combination of checkpoint inhibitors such as, for example, a combination of one or more PD-1/PD-LI checkpoint inhibitors used in combination with one or more CTLA-4 checkpoint inhibitors.
In one embodiment, the immunotherapeutic agent is a monoclonal antibody. In some cases, a method for determining the likelihood of response to one or more monoclonal antibodies is provided. The monoclonal antibody can be directed against tumor cells or directed against tumor products. The monoclonal antibody can be panitumumab, matuzumab, necitumunab, trastuzumab, amatuximab, bevacizumab, ramucirumab, bavituximab, patritumab, rilotumumab, cetuximab, immu-132, or demcizumab.
In yet another embodiment, the immunotherapeutic agent is a therapeutic vaccine. In some cases, a method for determining the likelihood of response to one or more therapeutic vaccines is provided. The therapeutic vaccine can be a peptide or tumor cell vaccine. The vaccine can target MAGE-3 antigens, NY-ESO-1 antigens, p53 antigens, survivin antigens, or MUC1 antigens. The therapeutic cancer vaccine can be GVAX (GM-CSF gene-transfected tumor cell vaccine), belagenpumatucel-L (allogeneic tumor cell vaccine made with four irradiated NSCLC cell lines modified with TGF-beta2 antisense plasmid), MAGE-A3 vaccine (composed of MAGE-A3 protein and adjuvant AS15), (1)-BLP-25 anti-MUC-1 (targets MUC-1 expressed on tumor cells), CimaVax EGF (vaccine composed of human recombinant Epidermal Growth Factor (EGF) conjugated to a carrier protein), WT1 peptide vaccine (composed of four Wilms' tumor suppressor gene analogue peptides), CRS-207 (live-attenuated Listeria monocytogenes vector encoding human mesothelin), Bec2/BCG (induces anti-GD3 antibodies), GV1001 (targets the human telomerase reverse transcriptase), tergenpumatucel-L (consists of human lung cancer cells genetically modified to include a mouse gene to which the immune system responds strongly), TG4010 (targets the MUC1 antigen), racotumomab (anti-idiotypic antibody which mimicks the NGcGM3 ganglioside that is expressed on multiple human cancers), tecemotide (liposomal BLP25; liposome-based vaccine made from tandem repeat region of MUC1) or DRibbles (a vaccine made from nine cancer antigens plus TLR adjuvants).
In one embodiment, the immunotherapeutic agent is a biological response modifier. In some cases, a method for determining the likelihood of response to one or more biological response modifiers is provided. The biological response modifier can trigger inflammation such as, for example, PF-3512676 (CpG 7909) (a toll-like receptor 9 agonist), CpG-ODN 2006 (downregulates Tregs), Bacillus Calmette-Guerin (BCG), Mycobacterium vaccae (SRL172) (nonspecific immune stimulants now often tested as adjuvants). The biological response modifier can be cytokine therapy such as, for example, IL-2+ tumor necrosis factor alpha (TNF-alpha) or interferon alpha (induces T-cell proliferation), interferon gamma (induces tumor cell apoptosis), or Mda-7 (IL-24) (Mda-7/IL-24 induces tumor cell apoptosis and inhibits tumor angiogenesis). The biological response modifier can be a colony-stimulating factor such as, for example granulocyte colony-stimulating factor. The biological response modifier can be a multi-modal effector such as, for example, multi-target VEGFR: thalidomide and analogues such as lenalidomide and pomalidomide, cyclophosphamide, cyclosporine, denileukin diftitox, talactoferrin, trabecetedin or all-trans-retinmoic acid.
In one embodiment, the immunotherapy is cellular immunotherapy. In some cases, a method for determining the likelihood of response to one or more cellular therapeutic agents. The cellular immunotherapeutic agent can be dendritic cells (DCs) (ex vivo generated DC-vaccines loaded with tumor antigens), T-cells (ex vivo generated lymphokine-activated killer cells; cytokine-induce killer cells; activated T-cells; gamma delta T-cells), or natural killer cells.
In some cases, specific subtypes of AD have different levels of immune activation (e.g., innate immunity and/or adaptive immunity) such that subtypes with elevated or detectable immune activation (e.g., innate immunity and/or adaptive immunity) are selected for treatment with one or more immunotherapeutic agents described herein. In one embodiment, the PP subtype of AD has low immune activation (e.g., innate immunity and/or adaptive immunity) as compared to other AD subtypes or lung cancer subtypes. In some cases, specific subtypes of AD have high or elevated levels of immune activation. In some cases, the PI subtype of AD has elevated levels of immune activation (e.g., innate immunity and/or adaptive immunity) as compared to other AD subtypes or lung cancer subtypes. In one embodiment, AD subtypes with low levels of or no immune activation (e.g., innate immunity and/or adaptive immunity) are not selected for treatment with one or more immunotherapeutic agents described herein.
In one embodiment, the methods and compositions provided herein allow for the detection of at least one nucleic acid in a lung cancer sample (e.g. adenocarcinoma lung cancer sample) obtained from a subject. The at least one nucleic acid can be a classifier biomarker provided herein. In one embodiment, the at least one nucleic acid detected using the methods and compositions provided herein are selected from Table 1. In one embodiment, the methods of detecting the nucleic acid(s) (e.g., classifier biomarkers) in the lung cancer sample obtained from the subject comprises, consists essentially of, or consists of measuring the expression level of at least one or a plurality of biomarkers using any of the methods provided herein. The biomarkers can be selected from Table 1. In some cases, the plurality of biomarker nucleic acids comprises, consists essentially of or consists of at least two biomarker nucleic acids, at least 8 biomarker nucleic acids, at least 16 biomarker nucleic acids, at least 24 biomarker nucleic acids, at least 32 biomarker nucleic acids, or all 48 biomarkers nucleic acids of Table 1. The detection can be at the nucleic acid level. The detection can be by using any amplification, hybridization and/or sequencing assay disclosed herein.
In another embodiment, the methods and compositions provided herein allow for the detection of at least one nucleic acid or a plurality of nucleic acids in a lung cancer sample (e.g. adenocarcinoma lung cancer sample) obtained from a subject such that the at least one nucleic acid is or the plurality of nucleic acids are selected from the biomarkers listed in Table 1 and the detection of at least one biomarker from a set of biomarkers whose presence, absence and/or level of expression is indicative of immune activation. The set of biomarkers for indicating immune activation can be gene expression signatures of and/or Adaptive Immune Cells (AIC) (e.g., Table 4A) and/or Innate Immune Cells (IIC) (e.g., Table 4B), individual immune biomarkers (e.g., Table 5), interferon genes (e.g., Table 6), major histocompatibility complex, class II (MHC II) genes (e.g., Table 7) or a combination thereof. The gene expression signatures of both IIC and AIC can be any gene signatures known in the art such as, for example, the gene signature listed in Bindea et al. (Immunity 2013; 39(4); 782-795). The detection can be at the nucleic acid level. The detection can be by using any amplification, hybridization and/or sequencing assay disclosed herein.
Kits for practicing the methods of the invention can be further provided. By “kit” can encompass any manufacture (e.g., a package or a container) comprising at least one reagent, e.g., an antibody, a nucleic acid probe or primer, etc., for specifically detecting the expression of a biomarker of the invention. The kit may be promoted, distributed, or sold as a unit for performing the methods of the present invention. Additionally, the kits may contain a package insert describing the kit and methods for its use.
In one embodiment, kits for practicing the methods of the invention are provided. Such kits are compatible with both manual and automated immunocytochemistry techniques (e.g., cell staining). These kits comprise at least one antibody directed to a biomarker of interest, chemicals for the detection of antibody binding to the biomarker, a counterstain, and, optionally, a bluing agent to facilitate identification of positive staining cells. Any chemicals that detect antigen-antibody binding may be used in the practice of the invention. The kits may comprise at least 2, at least 3, at least 4, at least 5, at least 6, at least 7, at least 8, at least 9, at least 10, or more antibodies for use in the methods of the invention.
The present invention is further illustrated by reference to the following Examples. However, it should be noted that these Examples, like the embodiments described above, are illustrative and are not to be construed as restricting the scope of the invention in any way.
Gene expression based subtyping in Lung Adenocarcinoma (AD) classifies AD tumors into distinct subtypes with variable biologic and clinical features. Gene expression based subtyping has consistently identified 3 distinct biologic types in Lung AD, Terminal Respiratory Unit (TRU), formerly Bronchioid, Proximal Proliferative (PP), formerly Magnoid, and Proximal Inflammatory (PI), formerly Squamoid (1,2)) (see
Using previously published Bindea et al. (3) immune cell gene signatures (24 in total) and AD subtyping gene expression signatures (1-2), several publically available lung AD datasets (2, 4 and 5) and 1 recently collected gene expression dataset (see
The AD datasets included several publically available lung cancer gene expression data sets as described above and a newly collected adenocarcinoma dataset of Formalin Fixed Paraffin Embedded (FFPE) lung tumor samples (n=88). The newly collected AD dataset of 88 formalin fixed paraffin embedded (FFPE) samples were archived residual lung tumor samples collected under an approved IRB protocol at the University of North Carolina at Chapel Hill (UNC-CH). FFPE sample sections (3 10 um sections) were macrodissected prior to RNA extraction. Transcriptome-enriched RNAseq was performed using Illumina's RNA-Access kits (San Diego, Calif.) with input of 100 ng/sample. Sequence data was aligned using hg19 as reference and the transcriptome was built using cufflinks (Trapnell 2010). Cuff compare was used to annotate the transcriptome and gene expression counts were calculated.
For AD, 4 published and 1 recently collected gene expression data sets (i.e., GeneCentric expression data set) of lung adenocarcinoma samples with a total of 1278 patient samples were used. The published data sets included TCGA [2], Shedden et al [4], Tomida et al [5], and Wilkerson et al [1], derived from fresh frozen specimens. The GeneCentric expression data set was derived from Formalin Fixed Paraffin Embedded (FFPE) specimens. For TCGA, upper quantile normalized RSEM data was downloaded from Firehose and log 2 transformed. Affymetrix Cel files from Shedden et al [4] were downloaded from the caIntegrator website and robust multi-array average expression measures were generated using the Affy package in R. Normalized Agilent array data was downloaded from the Gene Expression Omnibus (GEO) website for Tomida et al [5] (GSE13213) and Wilkerson et al [1] (GSE26939).
To determine adenocarcinoma subtype (TRU, PP, and PI), the published 506-gene nearest centroid classifier as described previously in Wilkerson et al [1] was used. After median centering of genes in the signature, each sample was assigned the subtype corresponding to the centroid with which it was maximally correlated. (Pearson)
Using the TCGA data for adenocarcinoma, correlations were assessed among the 30 markers by plotting matrices of pairwise Spearman rank correlation coefficients where markers were ordered by hierarchical clustering (see
Immune cell signature associations with tumor subtype and with CD274 expression were evaluated using linear regression. More specifically, to assess the prediction strength of subtype as a predictor of immune markers relative to that of PD-L1, a linear regression model of each signature with subtype the sole predictor, and again with PD-L1 the sole predictor, was fitted in the TCGA dataset. PD-L1 expression was treated as a low/medium/high categorical variable with equal proportions in each group. Scatter plots of adjusted R-squared when subtype was the predictor against adjusted R-squared when PD-L1 was the predictor were inspected for overall trends (see
Using non-silent mutation burden per Mb data, available in the supplementary information from TCGA adenocarcinoma (Lawrence 2013), mutation burden-Tcell expression associations was investigated using the Kruskal Wallis test and the Spearman correlation coefficients, respectively. For TCGA adenocarcinoma, STK11 CNV and mutation status were downloaded from Firehose, and STK11 inactivation-subtype association was evaluated using Fisher's exact test. Here, a sample was called inactive when it was reported as deleted and/or mutated. To test whether STK11 in AD showed evidence of association after adjusting for subtype, a linear model for Tcell expression was fit with inactive STK11 in AD as sole predictors and again following adjustment for subtype.
Subtype and immune signature associations with a 13-gene MHC class II signature [Forero [6]; Table 7], calculated as an average of all genes in the list (Table 7), were investigated using the Kruskal-Wallis test. For immune signature-MHC class II associations, Spearman correlation coefficients were calculated.
Hierarchical clustering of immune signatures and pairwise signature correlations were also analyzed. Survival signature associations of Stages 1-III samples were evaluated with stratified cox proportional hazard models allowing for different baseline hazards in each dataset. More specifically, immune marker-survival associations in the TCGA data sets were tested, overall and separately within each subtype, using Cox proportional hazards models. Immune markers were centered and scaled to have mean 0 and variance 1, and stage IV patients were excluded. Evaluations within a specific subtype adjusted for stage, and overall evaluations adjusted for both stage and subtype. Forest plots showing hazard ratios and confidence intervals for each signature were made (see
Heatmap analysis and unsupervised hierarchical clustering of immune cell gene signatures provided separation of intrinsic subtypes of AD (see
Overall, immune activation was most prominent in the PI subtype of AD demonstrating activation of both innate as well as adaptive immune cells. In contrast, the PP subtype of AD demonstrated lower immune activation.
Using hierarchical clustering, correlation matrices revealed clustering of adaptive immune cells and innate immune cells (see
Strength of association of CD274 (PD-L1) expression with adaptive immune cell signatures, as compared to AD subtype was conducted. As shown in
Immune cell signatures were primarily evaluated in the TCGA datasets, however AD subtype immune differences, as measured by the immune cell signatures, were found to be very reproducible across multiple datasets (see
Non-silent mutation burden in the TCGA AD data differed by subtype with PI showing the highest burden and TRU the lowest burden (
Several other genomic features such as loss of STK11 in AD (Cao [7], Shabath [8], Koyama [9]) have been suggested as possible contributors to reduced immune response in NSCLC. STK11 inactivation was enriched in the low immune response adenocarcinoma PP subtype. STK11 inactivation in ADwere associated with lower immune cell expression, however after adjustment for subtype using linear regression, STK11 was not a significant predictor (STK11 in AD p=0.0007 to p=0.43 following adjustment for subtype).
The association of immune cell expression in AD lung cancer with MHC class II genes was investigated using a published 13 gene MHC class II signature (Forero [6]). MHC class II gene expression was strongly correlated with several immune cells in AD including T-cell expression (Spearman correlation:=0.66 in AD), B-cell expression (Spearman correlation=0.5 in AD) and DC expression (Spearman correlation=0.69 in AD). MHC class II gene expression was significantly higher in tumor adjacent normal lung tissue as compared with tumor and was differentially expressed across tumor subtypes (
Using cox proportional hazard models, subtype specific hazard ratios (HRs) for one unit of increased expression were calculated. Subtype specific HRs were adjusted for pathologic stage and confidence intervals (CI) were calculated. Hazard ratios and confidence intervals for markers that were significant (nominal p-value<0.05) for at least one subtype are shown in
Lung AD gene expression subtypes vary in their immune landscape. Intrinsic biologic subtypes of AD reveal key differences in immune cell activation, which were not always correlated with CD274 expression and demonstrated variable association with survival. AD PP subtype showed minimal immune infiltration (depressed immune cell expression) suggesting reduced response to immunoRX. The AD PI subtype showed immune feature expression associated with improved survival. Further, non-silent mutation burden was not correlated with immune cell expression across subtypes, however, MI-IC class II gene expression was highly correlated. Increased immune and MHC II gene expression was associated with improved survival in the TRU and PI subtype of AD.
The following references are incorporated by reference in their entireties for all purposes.
Several genomic studies have demonstrated three distinct intrinsic lung adenocarcinoma subtypes that can vary in their genomic profiles including gene expression, mutational spectrum, and copy number alterations [1-3]. The three biologic AD subtypes TRU, PP, and PI differ not only in their genomic features, but also demonstrate potentially important differences in clinical features [1-4]. The gene expression subtypes of AD can demonstrate significant differences in tumor differentiation, likelihood of distant recurrence, stage specific survival, underlying tumor drivers and inflammatory response [1-4] and may not be readily distinguishable by standard morphology-based techniques (microscopy & immunohistochemistry). Potential response differences to chemotherapy [2], Pemetrexed [5], and/or EGFR inhibitor therapies have also been suggested [2]. Enrichment for EGFR over-expression was demonstrated in the terminal respiratory unit (TRU) subtype [2, 3]. Greater frequency of KRAS mutations, in combination with LKB1/STK11 deletions, are more likely in the proximal proliferative (PP) subtype [2, 3]. TP53 mutations and immune gene activation are hallmarks of the proximal inflammatory (PI) subtype [2-4]. Preliminary data may demonstrate potential for enhanced response to EGFR inhibitors in the TRU subtype, enhanced response to chemotherapy in the PP subtype, enhanced Pemetrexed response in the TRU subtype, and potential response to immunotherapy in the PI subtype [2-6]. The emerging data suggests that AD classification by gene expression subtype may provide valuable information complementing drug target mutation testing and informing lung cancer patient management.
Lung Adenocarcinoma (AD) subtyping has been primarily restricted to a research protocol involving the extraction of RNA from Fresh Frozen lung tumors, followed by application of a nearest centroid predictor using quantitative gene expression of over 500 genes. Despite evidence of prognostic and predictive benefits from adenocarcinoma subtyping, the need for Fresh Frozen tissue, the requirement for gene expression of >500 genes in combination with complex bioinformatic analyses, has hindered the application of AD subtyping in drug development and/or the clinic. The goal of this study was to develop a robust and efficient gene signature (with fewer genes needed) for differentiating the three subtypes of adenocarcinoma (Terminal Respiratory Unit (TRU); formerly referred to as Bronchioid, Proximal Proliferative (PP); formerly referred to as Magnoid, and Proximal Inflammatory (PI); formerly referred to as Squamoid). The new efficient gene signature may serve to reliably subtype AD from fresh frozen or FFPE tumor samples, making it amenable for diagnostic applications and/or drug development using any of the available quantitative RNA platforms (qRT-PCR, RNAseq, Affymetrix or Agilent Arrays). Development of the 48 gene signature for differentiating the subtypes of adenocarcinoma was performed as described in the methods herein.
Using the 515 lung adenocarcinoma The Cancer Genome Atlas (TCGA) RNAseq dataset for training and the 506-gene classifier to define gold standard subtype, a 48-gene signature was developed that maintains low misclassification rates when applied to several independent test sets. Starting with the standard 506 classifier genes, the Classifying arrays to Nearest Centroid (CLaNC) [7] algorithm was used with modification to select an equal number of negatively and positively correlated genes for each subtype. The optimal number of genes (16 per subtype) to include in the signature was chosen based on 5-fold cross validation curves was performed using the TCGA lung adenocarcinoma dataset (see
In order to validate the consistent performance of the selected 48 gene signature, the newly collected FFPE samples were lung adenocarcinoma (AD) residual archived samples (primarily surgical samples) that had been collected under an IRB approved protocol at the University of North Carolina in Chapel Hill, N.C. The samples were reviewed by a pathologist for tumor cells and three 10 μm tissue sections were macrodissected prior to extraction to enrich for tumor cells. RNA was quantitated and 100 ng was input per sample. Sequencing libraries were constructed using Illumina RNA-Access kits that enrich for the transcriptome. Sequencing libraries were under quality control by using a BA analyzer and quantified using qPCR. Sequence data was generated on an Illumina HiSeq platform (50 bp PE, 20-30 million reads) and was under quality control by using fastQC. Sequence results were aligned against hg19 reference sequence using STAR aligner and the transcriptome was built using Cufflinks [10]. Cuffcompare was used to annotate the transcriptome and counts of various expressed genes were calculated. RSEM expression count estimates were upper quartile normalized and log 2 transformed following the approach used in the Cancer Genome Atlas lung adenocarcinoma analysis [3, 11].
The 48 gene signature gene list developed in this study is shown in Table 2, while the T statistics for the 48 gene signature gene list for each AD subtype can be found in Table 1. The median gene expression of the 16 genes selected for each AD subtype (bronchioid, magnoid, squamoid) is shown in
Development and validation of an efficient 48 gene signature for AD subtyping was described. The resulting 48 gene signature maintains low misclassification rates when applied to several independent test sets. Thus, the new signature reliably subtypes AD from fresh frozen or FFPE tumor samples and can perform reliably using gene expression data generated from a variety of platforms including RNAseq and Arrays.
The following references are incorporated by reference in their entireties for all purposes.
Using previously published Bindea et al. (3) immune cell gene signatures (24 in total) and the Lung AD subtyping gene signature described in Example 2 for subtyping AD, several publically available lung AD datasets (1-2, 4-5; see
Using the TCGA AD dataset and the 48 gene AD subtyping signature of Example 2, heatmap analysis and unsupervised hierarchical clustering of immune cell gene signatures provided separation of intrinsic subtypes of AD in a similar fashion as to what was observed in Example 1 (see
Using cox proportional hazard models, subtype specific hazard ratios for one unit of increased expression were calculated. Subtype specific HR's were adjusted for pathologic stage and confidence intervals were calculated. The HR and CI for cell signatures or genes showing significant survival associations for one or more of the subtypes are shown in
The 48 gene signature for AD subtyping described in Example 2 shows similar results to the AD subtyping gene signature(s) used in Example 1 in terms of showing how Lung AD subtypes vary in their immune landscape. In agreement with the AD subtyping gene signatures of Example 1, the AD subtyping gene signature used in this example shows that Lung AD gene expression subtypes vary in their immune landscape. Intrinsic biologic subtypes of AD revealed key differences in immune cell activation, which were not always correlated with CD274 expression and demonstrated variable association with survival. AD PP subtype showed minimal immune infiltration suggesting reduced response to immunoRX. AD PI subtype showed immune feature expression associated with improved survival.
The following references are incorporated by reference in their entireties for all purposes.
Introduction: Gene expression based subtyping in Lung Adenocarcinoma (AD) classifies AD tumors into distinct subtypes with variable outcomes and potential response to therapy. Gene expression based subtyping has consistently identified 3 distinct biologic types in Lung AD, Terminal Respiratory Unit (TRU), formerly Bronchioid, Proximal Proliferative (PP), formerly Magnoid, and Proximal Inflammatory (PI), formerly Squamoid (1,2) (see
Methods: As a follow up to the experiments conducted in Example 1, differential drug target gene expression was evaluated in the lung AD subtypes from Example 1 that were determined using the TCGA lung cancer gene expression dataset (AD n=515)2 shown in
Results: As shown in
Conclusion: Molecular subtypes of lung AD vary in expression of the majority of key drug target genes included in a clinical solid tumor sequencing panel. Molecular subtypes of lung AD revealed differential expression of host immune response and immune targets. Evaluation of subtypes as potential biomarkers for drug sensitivity should be investigated alone, and in combination with immune cell features and key mutation targets.
The following references are incorporated by reference in their entireties for all purposes.
Introduction: Just like in Example 4, the purpose of this Example was to assess the differential expression of clinically important genes across previously defined gene expression subtypes of Adenocarcinoma (AD). In contrast to Example 4 where the AD and gene expression based subtyping was performed using the TCGA lung cancer gene expression dataset (AD n=515)2 as described in Example 1, gene expression based AD subtyping in this Example was performed using the 48 gene sets described in Example 2. Further, the clinically important genes were 322 genes (see Table 8) that constituted a clinical solid tumor mutation sequencing panel used in the management of oncology patients to identify genomic alterations impacting therapeutic management and/or to determine eligibility for targeted drug clinical trials. Just like in Example 4, differences in tumor proliferation were also assessed across the AD subtypes using an 11 gene proliferation signature (see Table 9).
Methods: Using the TCGA lung cancer gene expression dataset (Adenocarcinoma (AD) n=515),1 differential drug target gene expression was evaluated in lung AD subtypes. Subtype was defined in AD using the Clanc48 AD subtyper (see Example 2 and described herein) as previously described (nearest centroid prediction).2 AD subtypes Terminal Respiratory Unit (TRU), Proximal Proliferative (PP), and Proximal Inflammatory (PI) were examined. Variable expression of genes from a clinical oncology solid tumor mutation panel (322 genes),3 was examined in relation to AD. Differential gene expression was assessed using the Kruskal-Wallis (KW) test with Bonferroni correction. Further, a proliferation score was calculated as the average expression (log 2(RSEM+1)) of available genes in the 11-gene PAM50 proliferation signature4. Subtype-proliferation association was tested using the Kruskal-Wallis test.
Results: Similar to
Conclusion: Just like in Example 4, molecular subtypes of lung AD vary in expression of the majority of key drug target genes included in a clinical solid tumor sequencing panel. Molecular subtypes of lung AD revealed differential expression of host immune response and immune targets.
The following references are incorporated by reference in their entireties for all purposes.
The various embodiments described above can be combined to provide further embodiments. All of the U.S. patents, U.S. patent application publications, U.S. patent application, foreign patents, foreign patent application and non-patent publications referred to in this specification and/or listed in the Application Data Sheet are incorporated herein by reference, in their entirety. Aspects of the embodiments can be modified, if necessary to employ concepts of the various patents, application and publications to provide yet further embodiments.
These and other changes can be made to the embodiments in light of the above-detailed description. In general, in the following claims, the terms used should not be construed to limit the claims to the specific embodiments disclosed in the specification and the claims, but should be construed to include all possible embodiments along with the full scope of equivalents to which such claims are entitled. Accordingly, the claims are not limited by the disclosure.
This application is a continuation of U.S. application Ser. No. 16/302,167, filed on Nov. 16, 2018, which is a national phase of International Application No. PCT/US2017/033110, filed May 17, 2017, which claims priority from U.S. Provisional Application No. 62/337,591 filed May 17, 2016, U.S. Provisional Application No. 62/337,645 filed May 17, 2016, U.S. Provisional Application No. 62/396,587 filed Sep. 19, 2016, U.S. Provisional Application No. 62/420,836 filed Nov. 11, 2016, and U.S. Provisional Application No. 62/425,717 filed Nov. 23, 2016, each of which is incorporated by reference herein in its entirety for all purposes.
| Number | Date | Country | |
|---|---|---|---|
| 62337591 | May 2016 | US | |
| 62337645 | May 2016 | US | |
| 62396587 | Sep 2016 | US | |
| 62420836 | Nov 2016 | US | |
| 62425717 | Nov 2016 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 16302167 | Nov 2018 | US |
| Child | 17156024 | US |
