Patent Application
20140005058
The present disclosure relates to methods and compositions for diagnosing and defining the staging or progress of disorders such as prostate cancer.
The use of prostate specific antigen (PSA) as a diagnostic biomarker for prostate cancer was approved by the US Federal Drug Agency in 1994. In the nearly two decades since this approval, the PSA test has remained the primary tool for use in prostate cancer diagnosis, in monitoring for recurrence of prostate cancer, and in following the efficacy of treatments. However the PSA test has multiple shortcomings and, despite its widespread use, has resulted in only small changes in the death rate from advanced prostate cancers. To reduce the death rate and the negative impacts on quality of life caused by prostate cancer, new tools are required not only for more accurate primary diagnosis, but also for assessing the risk of spread of primary prostate cancers, and for monitoring responses to therapeutic interventions.
Today, a blood serum level of around 4 ng per ml of PSA is considered indicative of prostate cancer, while a PSA level of 10 ng per ml or higher is considered highly suggestive of prostate cancer. The PSA blood test is not used in isolation when checking for prostate cancer; a digital rectal examination (DRE) is usually also performed. If the results of the PSA test or the DRE are abnormal, a biopsy is generally performed in which small samples of tissue are removed from the prostate and examined. If the results are positive for prostate cancer, further tests may be needed to determine the stage of progression of the cancer, such as a bone scan, a computed tomography (CT) scan or a pelvic lymph node dissection.
While the PSA test has a good sensitivity (80%), it suffers from a false positive rate that approaches 75%. For example, it has been estimated that for PSA values of 4-10 ng/ml, only one true diagnosis of prostate cancer was found in approximately four biopsies performed (Catalona et al. J. Urol. 151(5):1283-90, 1994). Tests that measure the ratio of free to total (i.e., free plus bound) PSA do not have significantly greater specificity or sensitivity than the standard PSA test.
Higher PSA levels often lead to biopsies to determine the presence or absence of cancer cells in the prostate, and may lead to the surgical removal of the localized prostate gland. While surgery removes the localized cancer and often improves prostate cancer-specific mortality, it also masks the fact that many patients with prostate cancer, even in the absence of surgery, do not experience disease progression to metastasis or death.
The high false positive rate associated with the PSA test leads to many unnecessary biopsies. In addition to the physical discomfort and psychological distress associated with biopsies, it has been suggested that performing a biopsy may promote inflammation of cancerous tissue and increase the risk of cancer metastasis.
Currently, the established prognostic factors of histological grade and cancer stage from biopsy results, and prostate-specific antigen level in blood at diagnosis are insufficient to separate prostate cancer patients who are at high risk for cancer progression from those who are likely to die of another cause.
Once high risk or virulent forms of prostate cancer have been diagnosed, control strategies may involve surgery to remove the prostate gland if identified before metastasis, radiation to destroy cancer cells within the prostate and drug-based testosterone repression, generally referred to as androgen depletion therapy. These various treatments may bring about cures in some instances, or slow the time to death. However, for those with the most virulent forms of prostate cancer, the cancer will usually recur after surgery or radiation therapy and progress to resistance to androgen depletion therapy, with death a frequent outcome.
Early detection of virulent forms of prostate cancer is critical but the conclusion of specialist physicians is that the PSA test alone is inadequate for distinguishing patients whose cancers will become virulent and progress to threaten life expectancy from those with indolent cancers.
The following are some key reasons why the PSA test does not meet the needs of men's health:
There are at least two basic cell types involved in prostate cancer. Adenocarcinoma is a cancer of epithelial cells in the prostate gland and accounts for approximately 95% of prostate cancers. Neuroendocrine cancers may arise from cells of the endocrine (hormonal) and nervous systems of the prostate gland and account for approximately 5% of prostate cancers. Neuroendocrine cells have common features such as special secretory granules, produce biogenic amines and polypeptide hormones, and are most common in the intestine, lung, salivary gland, pituitary gland, pancreas, liver, breast and prostate. Neuroendocrine cells co-proliferate with malignant adenocarcinomas and secrete factors which appear to stimulate adenocarcinoma cell growth. Neuroendocrine cancers are rarer, and are considered non-PSA secreting and androgen-independent for their growth.
Some 15 to 17% of men with prostate cancer have cancers that grow but do not produce increasing or high blood levels of PSA. In these patients, who are termed asymptomatic, the PSA test often returns false negative test results as the cancer grows.
iii) BPH, Prostatitis and PIN
Benign prostate hypertrophy (BPH), a non-malignant growth of epithelial cells, and prostatitis are diseases of the prostate that are usually caused by an infection of the prostate gland. Both BPH and prostatitis are common in men over 50 and can result in increased PSA levels. Incidence rates increase from 3 cases per 1000 man-years at age 45-49 years, to 38 cases per 1000 man-years by the age of 75-79 years. Whereas the prevalence rate is 2.7% for men aged 45-49, it increases to at least 24% by the age of 80 years. While prostate cancer results from the deregulated proliferation of epithelial cells, BPH commonly results from proliferation of normal epithelial cells and frequently does not lead to malignancy (Ziada et al. (1999) Urology 53(3 Suppl 3D):1-6). Bacterial infection of the prostate can be demonstrated in only about 10% of men with symptoms of chronic prostatitis/chronic pelvic pain syndrome. Bacteria able to be cultured from patients suffering chronic bacterial prostatitis are mainly Gram-negative uropathogens. The role of Gram-positives, such as staphylococci and enterococci, and atypicals, such as chlamydia, ureaplasmas, mycoplasmas, are still debatable.
Another condition, known as prostate intraepithelial neoplasia (PIN), may precede prostate cancer by five to ten years. Currently there are no specific diagnostic tests for PIN, although the ability to detect and monitor this potentially pre-cancerous condition would contribute to early detection and enhanced survival rates for prostate cancer.
The phenotype of prostate cancer varies from one patient to another. More specifically, in different individuals prostate cancers display heterogeneous cellular morphologies, growth rates, responsiveness to androgens and pharmacological blocking agents for androgens, and varying metastatic potential. Each prostate cancer has its own unique progression involving multiple steps, including progression from localized carcinoma to invasive carcinoma to metastasis. The progression of prostate cancer likely proceeds, as seen for other cancers, via events that include the loss of function of cell regulators such as cancer suppressors, cell cycle and apoptosis regulators, proteins involved in metabolism and stress response, and metastasis related molecules (Abate-Shen et al. Polypeptides Dev. 14(19):2410-34, 2000; Ciocca et al. Cell Stress Chaperones 10(2):86-103, 2005).
At present health authorities do not universally recommend widespread screening for prostate cancer with the PSA test. There are concerns that many men may be diagnosed and treated unnecessarily as a result of being screened, at high cost to health systems as well as risking the patient's quality of life, such as through incontinence or impotence. Despite these concerns, prostate cancer is the most prevalent form of cancer and the second most common cause of cancer death in New Zealand, Australian and North American males (Jemal et al. CA Cancer J. Clin., 57(1):43-66, 2007). In reality, at least some of the men incubating life threatening forms of prostate cancer are being missed until their cancer is too advanced, due to the economic costs of national screening, the need to avoid unnecessary over-treatment, and/or the presence of progressive cancers producing only low or background levels of PSA. The need for a better diagnostic test could not be clearer.
The lack of a diagnostic test that distinguishes a non-life threatening from a potentially life-threatening cancer raises the important clinical question as to how aggressively to treat patients with localized prostate cancer. Treatment options for more aggressive cancers are invasive and include radical prostatectomy and/or radiation therapy.
Androgen-depletion therapy, for example using gonadotropin-releasing hormone agonists (e.g., leuprolide, goserelin, etc.), is designed to reduce the amount of testosterone that enters the prostate gland and is used in patients with metastatic disease, some patients who have a rising PSA and choose not to have surgery or radiation, and some patients with a rising PSA after surgery or radiation. Treatment options usually depend on the stage of the prostate cancer. Men with a 10-year life expectancy or less, who have a low Gleason score from a biopsy and whose cancer has not spread beyond the prostate are often not treated. Younger men with a low Gleason score and a prostate-restricted cancer may enter a phase of “watchful waiting” in which treatment is withheld until signs of progression are identified. However, these prognostic indicators do not accurately predict clinical outcome for individual patients.
Unlike many cancer types, specific patterns of gene expression have not been consistently identified in prostate cancer progression, although a number of candidate genes and pathways likely to be important in individual cases have been identified (Tomlins et al., Annu. Rev. Pathol. 1:243-71, 2006). Several groups have attempted to examine prostate cancer progression by comparing gene expression of primary carcinomas to normal prostate tissue. Because of differences in technique, the integrity of the tissue samples used as well as the biological heterogeneity of prostate cancers, these studies have reported thousands of candidate genes that share only moderate consensus. Also sample type differences could contribute to the lack of consensus seen from these studies. For example formalin fixed paraffin embedded (FFPE) tissues allow a convenient comparison of tumor and adjacent tissues but many of the cDNA microarray studies have used snap frozen tissues (Bibikova et al., Genomics 89:666-72, 2007; van't Veer et al., Nature 415:530-6, 2002). In addition, some studies have included accident victim donors as controls to overcome potential field effects (Aryee et al. Sci Trans' Med 5, 169ra10 2013; Chandran et al. BMC Cancer, 5:45 doi:10.1186/1471-2407-5-45, 2005). However, a few genes have emerged including hepsin (HPN; Rhodes et al., Cancer Res. 62:4427-33, 2002), alpha-methylacyl-CoA racemase (AMACR; Rubin et al., JAMA 287:1662-70, 2002, Lin et al. Biosensors 2:377-387, 2012), enhancer of Zeste homolog 2 (EZH2; Varambally et al. Nature, 419:624-9, 2002), L-dopa decarboxylase (DDC; Koutalellis et al. BJU International, 110:E267-E273, 2012) and anterior-gradient 2 (AGR2; Hu et al. Carcinogenesis 33:1178-1186, 2012) which have been shown experimentally to have probable roles in prostate carcinogenesis.
More recently, bioinformatic approaches employing data from gene expression profiling using both microarray and RNA-seq have generated lists of dysregulated genes in prostate cancer. RNA-seq is a technique based on enumeration of RNA transcripts using next-generation sequencing methodologies. However, because of their different experimental approaches, these studies have also shown few consensus genes, (Aryee et al. Sci Trans' Med 5, 169ra10, 2013; Chandran et al. BMC Cancer, 5:1471-2407 2005; Pflueger et al. Genome Res. 21:56-67, 2011; Prensner et al. Nature Biotechnology 29:742-749, 2011; Shancheng Ren et al. Cell Research 22:806-821, 2012).
A number of studies have also shown distinct classes of prostate cancers separable by their gene expression profiles (Glinsky et al., J. Clin. Invest. 113:913-23, 2004; Hsieh et al., Nature doi:10.1038/nature.10912, 2012; Lapointe et al., Proc. Natl. Acad. Sci. USA 101:811-6, 2004; LaTulippe et al., Cancer Res. 62:4499-506, 2002; Markert et al., Proc. Natl. Acad. Sci. doi:10.1073/pnas.1117029108, 2012; Rhodes et al., Cancer Res. 62:4427-33, 2002; Singh et al., Cancer Cell 1:203-9, 2002; Yu et al., J. Clin. Oncol. 22:2790-9, 2004; Varambally et al., Nature 419:624-9, 2002). Additionally, these approaches have been used to identify the genomic fusion of androgen-regulated genes including transmembrane protease, serine 2 (TMPRSS2) with members of the erythroblast transformation specific (ETS) DNA transcription factor family (Tomlins et al., Science 310:644-8, 2005, Tomlins, Nature 448: 595-599, 2007). These fusions appear commonly in prostate cancers and have been shown to be prevalent in more aggressive cancers (Attard et al., Oncogene 27:253-63, 2008; Barwick et al. Br. J. Cancer 102:570-576, 2010; Demichelis et al., Oncogene 26:4596-9, 2007; Nam et al., Br. J. Cancer 97:1690-5, 2007). Transcriptional modulation of TMPRSS2-ERG fusions has been shown to be associated with prostate cancer biomarkers and TGF-beta signalling (Brase et al., BMC Cancer 11:507 doi: 10.1186/1471, 2011). In addition to specific gene fusions, a vast array of mutational changes, including copy number variants, have been associated with prostate cancer tumours (Berger et al., Nature 470:214-220, 2011; Demichellis et al., Proc. Natl. Acad. Sci. doi:10.1073/pnas.117405109, 2012; Kumar et al., Proc. Natl. Acad. Sci. 108:17087-17092, 2011). Intratumor heterogeneity has also been found which has been suggested to result in underestimation of the degree of tumor heterogeneity (Gerlinger et al., New Eng, J. Med. 66:883-892, 2012). In particular mutations involving the substrate binding cleft of SPOP, which was found in 6-15% of prostate tumors, lacked ETS family gene rearrangements suggesting that tumors with SPOP mutations define a new class of prostate tumors. Also tumors with SPOP mutations lacked PTEN deletions in primary tumors but not in metastatic tumors (Barbieri et al., Nature Gen. 44:685-689, 2012).
Gene expression is the transcription of DNA into messenger RNA by RNA polymerase. Up-regulation describes a gene which has been observed to have higher expression (higher RNA levels) in one sample (for example, from cancer tissue) compared to another (usually healthy tissue from a control sample). Down-regulation describes a gene which has been observed to have lower expression (lower RNA levels) in one sample (for example, from cancer tissue) compared to another (usually healthy tissue from a control sample).
A common technology used for measuring RNA abundance is RT-qPCR where reverse transcription (RT) is followed by real-time quantitative PCR (qPCR). Reverse transcription first generates a DNA template from the RNA. This single-stranded template is called cDNA. The cDNA template is then amplified in the quantitative step, during which the fluorescence emitted by labeled hybridization probes or intercalating dyes changes as the DNA amplification process progresses. Quantitative PCR produces a measurement of an increase or decrease in copies of the original RNA and has been used to attempt to define changes of gene expression in cancer tissue as compared to comparable healthy tissues (Nolan T, et al. Nat Protoc 1:1559-1582, 2006; Paik S. The Oncologist 12:631-635, 2007; Costa C, et al. Trans' Lung Cancer Research 2:87-91, 2013). Massive parallel sequencing made possible by next generation sequencing (NGS) technologies is another way to approach the enumeration of RNA transcripts in a tissue sample and RNA-seq is a method that utilizes this. It is currently the most powerful analytical tool used for transcriptome analyses, including gene expression level difference between different physiological conditions, or changes that occur during development or over the course of disease progression. Specifically, RNA-seq can be used to study phenomena such as gene expression changes, alternative splicing events, allele-specific gene expression, and chimeric transcripts, including gene fusion events, novel transcripts and RNA editing. However, there are currently no methods that allow the use of RNA-seq for the accurate and reproducible quantification of multiple specific RNAs for reliable applications in the field of diagnostics.
Using multiple biomarkers in a diagnostic or prognostic test is preferable to using a single biomarker because of the following:
Each individual tumor is heterogeneous with respect to all of the different aspects of their genome, transcriptome and proteome;
Multiple tumor foci are commonly found in tissues;
A single biomarker does not allow tumors of different lethality, aggressiveness or specificity to be differentiated;
A single biomarker may be affected by a treatment regime or other environmental influence;
A single biomarker may be affected by a field effect either as part of the progression of the disease or due to the tumor itself; and
A single biomarker may be less effective in particular ethnic groups.
Why does RT-qPCR not Allow the Accurate Detection of Multiple Biomarkers?
RT-qPCR is a time consuming technique as expression differences are determined for a single gene at a time, which does not allow multiple biomarkers to be compared/assessed at one time.
Comparing expression levels for genes across different experiments is often difficult, and can require complicated normalization methods that may not be suitable for integration into a diagnostic.
RT-qPCR does not allow the accurate detection of down-regulated genes because it is limited in its fluorescence detection range, compared to NGS based methods. This causes genes that are at a low and/or high abundance to be problematic. Very often these transcripts, for which differential expression is difficult to measure, are the ones with the most diagnostic and/or progonostic value. RT-PCR does not allow multiplexing which causes a rise in cost per RNA biomarker, and hence the overall cost of the diagnostic test.
There thus remains a need in the art for an accurate test for prostate cancer.
The present invention provides methods for determining the presence and progression of a disorder in a subject. Such methods employ modified RNA-seq techniques to determine the relative frequency of one or more RNA biomarkers (also referred to as gene transcript biomarkers) specific for the disorder in the subject compared to that in healthy controls.
Determination of the relative frequency of expression levels of specific combinations of RNA biomarkers using the methods disclosed herein can also be used to determine the type and/or stage of a disorder, and to monitor the progression of a disorder and/or the effectiveness of treatment. Disorders that can be diagnosed and monitored using the methods disclosed herein include, but are not limited to, cancers, such as prostate and breast cancers.
The methods disclosed herein allow the determination of the frequency of multiple RNA biomarkers simultaneously using a process known as multiplexing. Multiplexing is a process wherein oligonucleotides specific for multiple biomarkers are amplified together to produce a pool of amplicons. The advantages of multiplexing are that it allows simultaneous testing of multiple RNA biomarkers in one or a small number of tubes, which in turn:
Reduces cost;
Reduces the amount of tissue required;
Increases the level of reproducibility due to less hands-on manipulation;
Reduces time involved in set-up; and
Increases throughput.
More specifically, the disclosed methods employ oligonucleotides specific for RNA biomarkers known to be associated with the presence and/or progression of a disorder, such as prostate cancer, at specific steps of a RNA-seq protocol to selectively identify cDNAs for the RNA biomarkers, and compare their relative frequency of expression between prostate cancer donors and healthy donors, as well as defining differences in expression between different stages of the disorder.
In conventional RNA-seq methodologies, the actual frequency of expression of each transcript is determined for the whole genome. These frequencies can be biased by differences in the efficiency of the cDNA production and subsequent PCR amplification steps for each transcript. The inventors believe that the methods disclosed herein avoid these biases by determining the relative, rather than actual, frequency of expression of RNA biomarkers. The biases are not relevant as long as they are neutral with respect to the comparisons made. The relative changes in frequency of expression of RNA biomarkers specific for prostate cancer allows detection of prostate cancers, distinguishing prostate cancers from benign prostate hypertrophy (BPH) and prostatitis, and detection of prostate cancers in asymptomatic men whose prostate cancer may produce low levels of PSA with high sensitivity and specificity. In certain embodiments, the disclosed methods determine changes in frequency of expression of RNA biomarkers in order to distinguish between indolent cancers, which have a low likelihood of progressing to a lethal disease, and more aggressive forms of prostate cancer which are life threatening and require treatment.
In one aspect, the present disclosure provides methods for detecting the presence of a disorder in a subject, comprising: (a) determining the relative frequency of expression of at least one RNA biomarker in a biological sample obtained from the subject using RNA sequencing; and (b) comparing the relative frequency of expression of the at least one RNA biomarker in the biological sample with a predetermined threshold value, wherein increased or decreased relative frequency of expression of the at least one RNA biomarker in the biological sample indicates the presence of the disorder in the subject. In related aspects, the disclosed methods comprise: (a) determining the relative frequency of expression of a plurality of RNA biomarkers in the biological sample; and (b) comparing the relative frequency of expression of the plurality of RNA biomarkers in the biological sample with predetermined threshold values, wherein increased or decreased relative frequency of expression of at least two or more of the RNA biomarkers in the biological sample indicates the presence of the disorder in the subject.
In one embodiment, the relative frequency of expression of at least one RNA biomarker is determined by: (a) isolating total RNA from the biological sample; (b) generating first strand cDNA from the total RNA using a first oligonucleotide primer specific for the at least one RNA biomarker; (c) synthesizing second strand cDNA to provide double-stranded cDNA (dsDNA); (d) adding at least one sequencing adapter to the double-stranded cDNA; (e) amplifying the double-stranded cDNA to provide a cDNA library from the double-stranded cDNA; and (f) sequencing the cDNA library and determining the relative frequency of expression of the at least one RNA biomarker. Optionally, such methods also comprise: (i) removing rRNA from the total RNA prior to step (b); (ii) end repairing the double stranded cDNA and adding an overhanging adenine (A) base to the 3′ end of the double stranded cDNA after step (c) and prior to step (d); and/or (iii) purifying and, optionally, size selecting the cDNA in the cDNA library after step (e) and prior to step (f).
In a related embodiment, such methods further comprise the option of synthesizing cDNA by polymerase chain reaction (PCR) using an oligonucleotide primer pair specific for the at least RNA biomarker after step (b) and prior to step (d) or by the standard methods. In certain embodiments, one of the oligonucleotides in the primer pair will be the same as the oligonucleotide primer used in the generation of the first strand cDNA.
In a further embodiment, the relative frequency of expression of the at least one RNA biomarker is determined by: (a) isolating total RNA from a biological sample; (b) generating first strand cDNA from the total RNA; (c) amplifying cDNA by polymerase chain reaction using an oligonucleotide primer pair specific for the at least one RNA biomarker to provide amplified double-stranded cDNA; (d) adding at least one sequencing adapter to the amplified double-stranded cDNA; (e) further amplifying the amplified double-stranded cDNA using primers specific for the at least one sequencing adapter to provide a cDNA library; and (f) sequencing the cDNA library and determining the relative frequency of expression of the at least one RNA biomarker. Optionally, such methods also comprise: (i) removing rRNA from the total RNA prior to step (b); (ii) end repairing the double stranded cDNA and adding an overhanging adenine (A) base to the 3′ end of the double stranded cDNA after step (c) and prior to step (d); and/or (iii) purifying and, optionally, size selecting the cDNA in the cDNA library after step (e) and prior to step (f).
In certain embodiments, the disclosed methods comprise determining the expression level of multiple RNA biomarkers corresponding to polynucleotide biomarkers selected from the group consisting of those listed in Tables 1, 2 and 3. Oligonucleotide primers that can be employed in the methods disclosed herein include, but are not limited to, those provided in SEQ ID NO: 76-232 and 293-326. In certain embodiments, the methods disclosed herein include detecting the relative frequency of expression of a RNA biomarker comprising an RNA sequence that corresponds to a DNA sequence of SEQ ID NO: 1-75 and 235-287 or a variant thereof, as defined herein. Those of skill in the art will appreciate that the RNA sequences for the disclosed RNA biomarkers are identical to the cDNA sequences disclosed herein except for the substitution of thymine (T) residues with uracil (U) residues.
In a further aspect, the present disclosure provides an oligonucleotide primer comprising, or consisting of, a sequence selected from the group consisting of SEQ ID NO: 76-232 and 293-326, and variants thereof. In certain embodiments, such oligonucleotide primers have a length equal to or less than 30 nucleotides. The disclosed oligonucleotide primers can be effectively employed in methods for diagnosing the presence of, and/or monitoring the progression of, prostate cancer using methods well known to those of skill in the art, including quantitative real time PCR or small scale oligonucleotide microarrays.
Biological samples that can be effectively employed in the disclosed methods include, but are not limited to, urine, blood, serum, cell lines, peripheral blood mononuclear cells (PBMCs), biopsy tissue and prostatectomy tissue.
As used herein, the term “biomarker” refers to a molecule that is associated either quantitatively or qualitatively with a biological change. Examples of biomarkers include polypeptides, proteins, fragments of a polypeptide or protein; polynucleotides, such as a gene product, RNA or RNA fragment; and other body metabolites.
As used herein, the term “RNA biomarker” or “gene transcript biomarker” refers to an RNA molecule produced by transcription of a gene that is associated either quantitatively or qualitatively with a biological change.
As used herein the term “RNA sequence corresponding to a DNA sequence” refers to a sequence that is identical to the DNA sequence except for the substitution of all thymine (T) residues with uracil (U) residues.
As used herein, the term “oligonucleotide specific for a biomarker” refers to an oligonucleotide that specifically hybridizes to a polynucleotide biomarker or a polynucleotide encoding a polypeptide biomarker, and that does not significantly hybridize to unrelated polynucleotides. In certain embodiments, the oligonucleotide hybridizes to a gene, a gene fragment or a gene transcript. In specific embodiments, the oligonucleotide hybridizes to the polynucleotide of interest under stringent conditions, such as, but not limited to, prewashing in a solution of 6×SSC, 0.2% SDS; hybridizing at 65° C., 6×SSC, 0.2% SDS overnight; followed by two washes of 30 minutes each in lx SSC, 0.1% SDS at 65° C. and two washes of 30 minutes each in 0.2×SSC, 0.1% SDS at 65° C.
As used herein the term “oligonucleotide primer pair” refers to a pair of oligonucleotide primers that span an intron in the cognate RNA biomarker.
As used, herein the term “polynucleotide(s),” refers to a single or double-stranded polymer of deoxyribonucleotide or ribonucleotide bases and includes DNA and corresponding RNA molecules, including hnRNA, mRNA, and non-coding RNA, molecules, both sense and anti-sense strands, and includes cDNA, genomic DNA and recombinant DNA, as well as wholly or partially synthesized polynucleotides. An hnRNA molecule contains introns and corresponds to a DNA molecule in a generally one-to-one manner. An mRNA molecule corresponds to an hnRNA and DNA molecule from which the introns have been excised. A non-coding RNA is a functional RNA molecule that is not translated into a protein, although in some circumstances non-coding RNA can be coding and vice a versa.
As used herein, the term “subject” refers to a mammal, preferably a human, who may or may not have a disorder, such as prostate cancer. Typically, the terms “subject” and “patient” are used interchangeably herein in reference to a human subject.
As used herein, the term “healthy subject” refers to a subject who is not inflicted with a disorder of interest.
As used herein in connection with prostate cancer, the term “healthy male” refers to a male who has an undetectable PSA level in serum or non-rising PSA levels up to 1 ng/ml, no evidence of prostate gland abnormality following a DRE and no clinical symptoms of prostatic disorders.
As used herein in connection with prostate cancer, the term “asymptomatic male” refers to a male who has a PSA level in serum of greater than 4 ng/ml, which is considered indicative of prostate cancer, but whose DRE is inconclusive and who has no symptoms of clinical disease.
The term “benign prostate hypertrophy” (BPH) refers to a prostatic disease with a non-malignant growth of epithelial cells in the prostate gland and the term “prostatitis” refers to another prostatic disease of the prostate, usually due to a microbial infection of the prostate gland. Both BPH and prostatitis can result in increased PSA levels.
As used herein, the term “metastatic prostate cancer” refers to prostate cancer which has spread beyond the prostate gland to a distant site, such as lymph nodes or bone. As used herein, the term “biopsy tissue” refers to a sample of tissue (e.g., prostate tissue) that is removed from a subject for the purpose of determining if the sample contains cancerous tissue. The biopsy tissue is then examined (e.g., by microscopy) for the presence or absence of cancer.
As used herein, the term “prostatectomy” refers to the surgical removal of the prostate gland.
As used herein, the term “sample” is used herein in its broadest sense to include a sample, specimen or culture obtained from any source. Biological samples include blood products (such as plasma, serum and whole blood), urine, saliva and the like. Biological samples also include tissue samples, such as biopsy tissues or pathological tissues, that have previously been fixed (e.g., formalin, snap frozen, cytological processing, etc.).
As used herein, the term “predetermined threshold value of expression” of a RNA biomarker refers to the level of expression of the same RNA biomarker in a corresponding control/normal sample or group of control/normal samples obtained from normal, or healthy, subjects, e.g. from males who do not have prostate cancer.
As used herein, the term “altered frequency of expression” of a RNA biomarker in a test biological sample refers to a frequency that is either below or above the predetermined threshold value of expression for the same RNA biomarker in a control sample and thus encompasses either high (increased) or low (decreased) expression levels.
As used herein, the term “relative frequency of expression” refers to the frequency of expression of a RNA biomarker in a test biological sample relative to the frequency of expression of the same RNA biomarker in a corresponding control/normal sample or group of control/normal samples obtained from normal, or healthy, subjects, (e.g., from males who do not have prostate cancer). In preferred embodiments, the frequency of expression of the RNA biomarker is also normalized to the frequency of an internal reference transcript.
As used herein, the term “prognosis” or “providing a prognosis” for a disorder, such as prostate cancer, refers to providing information regarding the likely impact of the presence of prostate cancer (e.g., as determined by the diagnostic methods) on a subject's future health (e.g., the risk of metastasis).
As outlined above, the present disclosure provides methods for detecting the presence or absence of a disorder, such as prostate cancer, in a subject, determining the stage of the disorder and/or the phenotype of the disorder, monitoring progression of the disorder, and/or monitoring treatment of the disorder by determining the frequency of expression of specific RNA biomarkers in a biological sample obtained from the subject. The methods disclosed herein employ one or more modifications of standard RNA-seq protocols. RNA-seq is a relatively new technology that has been employed for mass sequencing of whole transcriptomes, and that offers significant advantages over other methods employed for transcriptome sequencing, such as microarrays, including low levels of background noise, the ability to detect low levels of expression, the ability to detect novel mutations and transcripts, and the ability to use relatively small amounts of RNA (for a review of RNA-seq, see Wang et al., Nat. Rev. Genet. (2009) 10:57-63).
The disclosed methods employ oligonucleotides specific for one or more RNA biomarker in combination with RNA-seq technology to perform directed sequencing and thereby determine the relative frequency of expression of the RNA biomarker(s). Such methods have significant advantages over other technologies typically employed to determine expression levels of polynucleotide biomarkers, including improved accuracy, reproducibility and speed, the ability to easily determine the frequency of expression of a multitude of RNA biomarkers in a large number of samples at a relatively low cost, and the ability to identify novel mutations and transcripts.
In specific embodiments, such methods use oligonucleotides specific for one or more biomarkers selected from those shown in Tables 1, 2 and 3.
In one embodiment, the disclosed methods comprise determining the relative frequency of expression levels of at least two, three, four, five, six, seven, eight, nine, ten or more RNA biomarkers selected from the group consisting of: SEQ ID NO: 76-223 and 293-326 in a biological sample taken from a subject, and comparing the relative frequency of expression levels with predetermined threshold values.
The disclosed methods can be employed to diagnose the presence of prostate cancer in subjects with early stage prostate cancer; subjects who have had surgery to remove the prostate (radical prostatectomy); subjects who have had radiation treatment for prostate cancer; subjects who are undergoing, or have completed, androgen ablation therapy; subjects who have become resistant to hormone ablation therapy; and/or subjects who are undergoing, or have had, chemotherapy.
In certain embodiments, the RNA biomarkers disclosed herein appear in subjects with prostate cancer at levels that are at least two-fold higher or lower than, or at least two standard deviations above or below, the mean level in normal, healthy individuals, or are at least two-fold higher or lower than, or at least two standard deviations above or below, a predetermined threshold of expression.
All of the biomarkers and oligonucleotides disclosed herein are isolated and purified, as those terms are commonly used in the art. Preferably, the biomarkers and oligonucleotides are at least about 80% pure, more preferably at least about 90% pure, and most preferably at least about 99% pure.
In certain embodiments, the oligonucleotides employed in the disclosed methods specifically hybridize to a variant of a polynucleotide biomarker disclosed herein. As used herein, the term “variant” comprehends nucleotide or amino acid sequences different from the specifically identified sequences, wherein one or more nucleotides or amino acid residues is deleted, substituted, or added. Variants may be naturally occurring allelic variants, or non-naturally occurring variants. Variant sequences (polynucleotide or polypeptide) preferably exhibit at least 80%, 85%, 90%, 95%, 96%, 97%, 98% or 99% identity to a sequence disclosed herein. The percentage identity is determined by aligning the two sequences to be compared as described below, determining the number of identical residues in the aligned portion, dividing that number by the total number of residues in the inventive (queried) sequence, and multiplying the result by 100.
In addition to exhibiting the recited level of sequence identity, variants of the disclosed biomarkers are preferably themselves expressed in subjects with prostate cancer at a frequency that are higher or lower than the levels of expression in normal, healthy individuals.
Polypeptide and polynucleotide sequences may be aligned, and percentages of identical amino acids or nucleotides in a specified region may be determined against another polypeptide or polynucleotide sequence, using computer algorithms that are publicly available. The percentage identity of a polynucleotide or polypeptide sequence is determined by aligning polynucleotide and polypeptide sequences using appropriate algorithms, such as BLASTN or BLASTP, respectively, set to default parameters; identifying the number of identical nucleic or amino acids over the aligned portions; dividing the number of identical nucleic or amino acids by the total number of nucleic or amino acids of the polynucleotide or polypeptide of the present invention; and then multiplying by 100 to determine the percentage identity.
Two exemplary algorithms for aligning and identifying the identity of polynucleotide sequences are the BLASTN and FASTA algorithms. The alignment and identity of polypeptide sequences may be examined using the BLASTP algorithm. BLASTX and FASTX algorithms compare nucleotide query sequences translated in all reading frames against polypeptide sequences. The FASTA and FASTX algorithms are described in Pearson and Lipman, Proc. Natl. Acad. Sci. USA 85:2444-2448, 1988; and in Pearson, Methods in Enzymol. 183:63-98, 1990. The FASTA software package is available from the University of Virginia, Charlottesville, Va. 22906-9025. The FASTA algorithm, set to the default parameters described in the documentation and distributed with the algorithm, may be used in the determination of polynucleotide variants. The readme files for FASTA and FASTX Version 2.0× that are distributed with the algorithms describe the use of the algorithms and describe the default parameters.
The BLASTN software is available on the NCBI anonymous FTP server and is available from the National Center for Biotechnology Information (NCBI), National Library of Medicine, Building 38A, Room 8N805, Bethesda, Md. 20894. The BLASTN algorithm Version 2.0.6 [Sep.-10-1998] and Version 2.0.11 [Jan.-20-2000] set to the default parameters described in the documentation and distributed with the algorithm, is preferred for use in the determination of variants according to the present invention. The use of the BLAST family of algorithms, including BLASTN, is described at NCBI's website and in the publication of Altschul, et al., “Gapped BLAST and PSI-BLAST: a new generation of protein database search programs,” Nucleic Acids Res. 25:3389-3402, 1997.
Variant sequences generally differ from the specifically identified sequence only by conservative substitutions, deletions or modifications. As used herein with regards to amino acid sequences, a “conservative substitution” is one in which an amino acid is substituted for another amino acid that has similar properties, such that one skilled in the art of peptide chemistry would expect the secondary structure and hydropathic nature of the polypeptide to be substantially unchanged. In general, the following groups of amino acids represent conservative changes: (1) ala, pro, gly, glu, asp, gln, asn, ser, thr; (2) cys, ser, tyr, thr; (3) val, ile, leu, met, ala, phe; (4) lys, arg, his; and (5) phe, tyr, trp, his. Variants may also, or alternatively, contain other modifications, including the deletion or addition of amino acids that have minimal influence on the antigenic properties, secondary structure and hydropathic nature of the polypeptide. For example, a polypeptide may be conjugated to a signal (or leader) sequence at the N-terminal end of the protein which co-translationally or post-translationally directs transfer of the protein. The polypeptide may also be conjugated to a linker or other sequence for ease of synthesis, purification or identification of the polypeptide (e.g., poly-His), or to enhance binding of the polypeptide to a solid support. For example, a polypeptide may be conjugated to an immunoglobulin Fc region.
In another embodiment, variant polypeptides are encoded by polynucleotide sequences that hybridize to a disclosed polynucleotide under stringent conditions. Stringent hybridization conditions for determining complementarity include salt conditions of less than about 1 M, more usually less than about 500 mM, and preferably less than about 200 mM. Hybridization temperatures can be as low as 5° C., but are generally greater than about 22° C., more preferably greater than about 30° C., and most preferably greater than about 37° C. Longer DNA fragments may require higher hybridization temperatures for specific hybridization. Since the stringency of hybridization may be affected by other factors such as probe composition, presence of organic solvents and extent of base mismatching, the combination of parameters is more important than the absolute measure of any one alone. An example of “stringent conditions” is prewashing in a solution of 6×SSC, 0.2% SDS; hybridizing at 65° C., 6×SSC, 0.2% SDS overnight; followed by two washes of 30 minutes each in 1×SSC, 0.1% SDS at 65° C. and two washes of 30 minutes each in 0.2×SSC, 0.1% SDS at 65° C.
The expression levels of one or more RNA biomarkers in a biological sample can be determined, for example, using one or more oligonucleotides that are specific for the RNA biomarker. In one method, the expression level of one or more RNA biomarkers disclosed herein is determined by first collecting urine from a subject following DRE or prostate massage via a bicycle or exocycle. RNA is isolated from the urine sample, and the frequency of expression of the RNA biomarker is determined as described below using modified RNA-seq technology in combination with oligonucleotides specific for the RNA biomarker of interest.
In other embodiments, the levels of mRNA corresponding to a prostate cancer biomarker disclosed herein can be detected using oligonucleotides in Southern hybridizations, in situ hybridizations, or quantitative real-time PCR amplification (qRT-PCR). Solid phase substrates, or carriers, that can be effectively employed in such assays are well known to those of skill in the art and include, but are not limited to, microporous membranes constructed, for example, of nitrocellulose, nylon, polyvinylidene difluoride, polyester, cellulose acetate, mixed cellulose esters and polycarbonate. Suitable microporous membranes include, for example, those described in US Patent Application Publication no. US2010/0093557A1. Methods for performing such assays are well known to those of skill in the art.
The oligonucleotides employed in the disclosed methods are generally single-stranded molecules, such as synthetic antisense molecules or cDNA fragments, and are, for example, 6-60 nt, 15-30 or 20-25 nt in length.
Oligonucleotides specific for a polynucleotide, or RNA, biomarker disclosed herein are prepared using techniques well known to those of skill in the art. For example, oligonucleotides can be designed using known computer algorithms to identify oligonucleotides of a defined length that are unique to the polynucleotide, have a GC content within a range suitable for hybridization, and lack predicted secondary structure that may interfere with hybridization. Oligonucleotides can be synthesized using methods well known to those in the art. In specific embodiments, the oligonucleotides employed in the disclosed methods and compositions are selected from the group consisting of: SEQ ID NO: 76-223 and 293-326.
For tests involving alterations in RNA expression levels, it is important to ensure adequate standardization. Accordingly, in tests such as the adapted RNA-seq technology disclosed herein, quantitative real time PCR or small scale oligonucleotide microarrays, at least one expression standard is employed. Expression standards that can be employed in such methods include, but are not limited to, those listed in Table 3 below.
The present disclosure further provides methods employing a plurality of oligonucleotides that are specific for a plurality of the prostate cancer RNA biomarkers disclosed herein.
The following examples are intended to illustrate, but not limit, this disclosure.
RNA was isolated from LNCaP and A549 cell lines that had been harvested from cell culture and stored in Trizol using a ZYMO Direct-zol™ kit (Ngaio Diagnostics Ltd.) following the manufacturer's instructions. RNA quality was assessed using the Agilent BioAnalyser and the Agilent RNA 6000 nano assay protocol. The LNCaP and A549 RNA had a RIN value of 9.5 and 9.8 respectively. The RNA was also checked on the NanoDrop 2000 spectrophotometer, (Thermo Scientific), and its concentration ascertained by the Qubit® 2.0 Fluorometer (Life Technologies).
Histological blocks from subjects were reviewed by a clinical histopathologist, and tumor and histologically adjacent regions deemed “normal” were identified. These sections were then excised and reset in paraffin. Approximately fifteen freshly cut sections at a thickness of ten microns were then processed using a Qiagen RNeasy FFPE kit (Cat No: 74404, 73504). The method used in all extractions for deparaffinization step was the original method from the Cat no: 74404 kit, and the remainder of the protocol was performed following the manufacturer's instructions. The RNA was checked on the NanoDrop, and its concentration ascertained by the Qubit® 2.0 Fluorometer (Life Technologies).
RNA was isolated from one or more separate fresh urine samples from donors by sedimentation of the cellular material using centrifgation at 1000 g for five minutes at 4° C. The urine was decanted and the cell pellet resuspended in 1.8 ml of ice cold 1×PBS containing 2.5% Fetal Bovine Serum (Invitrogen). The cell suspension was transferred to a 2 ml Eppendorf tube and the cellular material collected by centrifugation at 400 g for 5 minutes at 4° C. The supernatant was removed (leaving around 50 μl) and the cell pellet resuspended in 1.8 ml of ice cold 1×PBS containing 2.5% Fetal Bovine Serum (Invitrogen). The cells were again collected by centrifugation at 400 g for 5 minutes at 4° C. The supernatant was removed (leaving around 50 μl) and the cell pellet resuspended in 1.8 ml of ice cold 1×PBS containing 2.5% Fetal Bovine Serum (Invitrogen). The cells were collected by centrifugation at 400 g for 5 minutes at 4° C. and all but 100 μL1 of the supernatant removed. The cells were resuspended in the remaining 100 μA of supernatant, and 8 μl was taken for microscopic analysis. A total of 300 μA of Trizol LS (Invitrogen) and 5 μg of E. coli 5S rRNA was added and the cell suspension was stored at −80° C. RNA was extracted as described by ZYMO using the Direct-zol™ kit, or as described by Invitrogen and further purified using Qiagen RNeasy™ spin columns. RNA was stored at −80° C. prior to use.
cDNA Preparation
cDNA was produced from approximately 1-1.5 ug of total RNA from either cell lines, biopsy tissue or urine extracts using random primers for the production of the first strand cDNA using the SuperScript® VILO™ cDNA Synthesis Kit (Life Technologies) or RNA biomarker-specific primers. The cDNA preparations were stored at −80° C. prior to use and then diluted 1/5 in sterile water prior to qRT-PCR.
qRT-PCR Methods
RNA biomarker specific primers were used to perform real time SYBR green PCR quantification from cell line-, biopsy- or urine-derived cDNA using the Roche Lightcycler 480 using standard protocols for determining the specificity and efficiency of the amplification. The relative amount of the marker gene in each of the samples tested was determined by comparing the cycle threshold (Ct value: number of PCR cycles required for the SYBR green fluorescent signal to cross the threshold exceeding background level within the exponential growth phase of the amplification curve). Following 30 cycle RT-PCR reactions, the amplicons were electrophoresed on a 2% agarose gel and sequenced with standard Sanger chemistry using an Applied Biosystems 3130×1 DNA sequencer.
The relative frequency of expression of specific RNA biomarkers was determined using the isolated RNA in one or more of the four methods described below. Each of these methods includes at least one modification of conventional RNA-seq technologies. Conventional RNA-seq technologies are well known to those of skill in the art and are described, for example, in Wang et al. (Nat. Rev. Genet. (2009) 10:57-63), and Marguerat and Bahler (Cell. Mol. Life. Sci. (2010) 67:569-579).
In a first method, sequence specific priming is employed during the generation of first strand cDNA. An optional first step in this method is to deplete the total RNA of rRNA using an industry-provided kit, if necessary. An industry-provided first strand cDNA kit is used to combine total RNA or rRNA-depleted total RNA with at least one strand specific oligonucleotide primer (i.e. an oligonucleotide primer specific for the RNA biomarker of interest) and generate first strand cDNA according to the manufacturer's protocol. Second strand cDNA is then synthesized in an unbiased manner using standard techniques. The resulting double-stranded cDNA is fragmented if necessary using standard methods, and the cDNA ends are repaired using standard methods in which any overhangs at the cDNA ends are converted into blunt ends using T4 DNA polymerase. An overhanging adenine (A) base is added to the 3′ end of the blunt DNA fragments by the use of Klenow fragment to assist with ligation of adapters required for the sequencing process. The adapters are ligated to the ends of the cDNA fragments using standard procedures, and then the cDNA fragments are run on a gel for purification and removal of excess adapters. The cDNA is amplified using adapter primers, purified, denatured and further diluted for cluster generation and sequencing, for example on a HiSeq2000 according to Illumina Corporation's standard protocols (208 cycles sequencing program, paired-end with indexing). The cDNA library is sequenced, and the relative frequency of expression of the specific RNA biomarkers in cancer patients and healthy controls is determined.
As in method 1, sequence specific priming is employed during the generation of first strand cDNA. This is achieved using an industry provided first strand cDNA kit and at least one strand specific oligonucleotide primer to generate first strand cDNA from total RNA (or rRNA depleted total RNA if necessary) according to the manufacturer's protocol. The second strand cDNA can either be prepared in an unbiased manner using standard techniques, or it can be directly amplified using a set of specific oligonucleotide primers (i.e. oligonucleotide primers specific for the RNA biomarkers of interest) to amplify a specific set of PCR amplicons by either primer limited or cycle limited PCR. In preferred embodiments, the oligonucleotide primer employed to generate the first strand cDNA can be the same as one of the pair of oligonucleotide primers used to amplify the double-stranded cDNA. The cDNA is then purified via a cleanup procedure to remove excess PCR reagents. The cDNA is fragmented if necessary using standard methods, and the cDNA ends are repaired using standard methods in which any overhangs at the cDNA ends are converted into blunt ends using T4 DNA polymerase. An overhanging adenine (A) base is added to the 3′ end of the blunt DNA fragments by the use of Klenow fragment to assist with ligation of adapters required for the sequencing process. The adapters are ligated to the ends of the cDNA fragments using standard procedures, and the cDNA fragments are then purified to remove excess adapters. The cDNA is amplified using adapter primers, purified, denatured and further diluted for cluster generation and sequencing, for example on a HiSeq2000 according to Illumina Corporation's standard protocols (208 cycles sequencing program, paired-end with indexing). The cDNA library is sequenced and the relative frequency of expression of the specific RNA biomarkers in cancer patients and healthy controls is determined.
This method employs total RNA or rRNA-depleted RNA if necessary. The first strand cDNA is synthesized using standard methods. The first strand cDNA is then directly amplified using a set of specific oligonucleotide primers (i.e. oligonucleotide primers specific for the RNA biomarkers of interest) to amplify a specific set of PCR amplicons using either primer limited or cycle limited PCR. The cDNA is purified via a cleanup procedure to remove excess PCR reagents. The cDNA is fragmented if necessary using standard methods, and the cDNA ends are repaired using standard methods, in which any overhangs at the cDNA ends are converted into blunt ends using T4 DNA polymerase. An overhanging adenine (A) base is added to the 3′ end of the blunt DNA fragments by the use of Klenow fragment to assist with ligation of adapters required for the sequencing process. Adapters are ligated to the ends of the cDNA fragments using standard procedures, and the cDNA is purified to remove excess adapters. The cDNA is then amplified using adapter primers and purified. The cDNA can be size selected via gel electrophoresis using standard methods if necessary. The cDNA library is sequenced, and the relative frequency of expression of the specific RNA biomarkers in cancer patients and healthy controls is determined.
Method 4 differs from Method 3 in that all sequences necessary for next generation sequencing are incorporated via either a one or two step PCR amplification.
An optional first step in this method is to deplete the total RNA of rRNA using an industry-provided kit, if necessary. The first strand cDNA is then synthesized using standard methods. The first strand cDNA is directly amplified using a set of specific oligonucleotide primers (i.e. oligonucleotide primers specific for the RNA biomarkers of interest) also containing Next Generation Sequencing (NGS) primer sites, using either primer limited or cycle limited PCR. The cDNA is then purified via a cleanup procedure to remove excess PCR reagents, and re-amplified with another set of primers, if necessary, in order to add further sites required for NGS using either primer limited or cycle limited PCR. The cDNA is then purified to remove excess PCR reagents and, if necessary, is again amplified using adaptor primers and purified. The cDNA is amplified using adapter primers, purified, denatured and further diluted for cluster generation and sequencing, for example on a HiSeq2000 according to Illumina Corporation's standard protocols (208 cycles sequencing program, paired-end with indexing). The cDNA library is sequenced, and the relative frequency of expression of the specific RNA biomarkers in cancer patients and healthy controls is determined.
RNA biomarkers were selected using annotation and analysis of publicly available RNA expression profile data in the NCBI databases GSE6919 and GSE38241 as these data-sets include data from cancer free donors. The biomarkers shown in Table 1 below is a unique set identified as being over-expressed in subjects with prostate cancer. Similarly, the biomarkers shown in Table 2 is a second unique combination of RNA biomarkers identified as being under-expressed in subjects with prostate cancer.
The NCBI database GSE6919, which was developed at the University of Pittsburgh, contains data from three Affymetrix chips (U95A, U95B and U95C), representing more than 36,000 gene reporters. The database, which has been analyzed by Chandran et al. (BMC Cancer 2005, 5:45; BMC Cancer 2007, 9:64), and Yu et al. (J Clin Oncol 2004, 22:2790-2799), contains RNA profiles from more than 200 individual prostate tumor samples, combined with adjacent “normal” or “healthy” tissues, or prostate tissues from individuals believed to be free of prostate cancer.
Homo sapiens
Homo sapiens laminin,
sapiens leucine-rich
Homo sapiens SET
Homo sapiens
Homo sapiens
Homo sapiens
Homo sapiens SMAD
Homo sapiens v-src
Homo sapiens tudor
Homo sapiens
Homo sapiens isolate
Homo sapiens mitogen-
For tests measuring the changes in frequency of RNA expression levels, it is essential to ensure adequate standardization. For this reason we have analyzed the NCBI database to identify reporters with the least variation between gene expression profiles, as shown in Table 3 below, in prostate cancer and healthy donor tissues. These reporters form a robust set of RNA expression standards that can be used where appropriate in tests involving quantification of RNA expression, such as in the modified RNA-seq technology described herein.
Primers for the production of an RNA biomarker specific amplicon were created using a multistep primer design strategy. Specific intron-spanning primers were created to amplify an amplicon of a specific size (60-300 bp) that can be used for Next Generation Sequencing (NGS).
The primers were designed using Primer3 (v. 0.4.0) software and the primers were checked to ensure that certain criteria were met:
In order to use these RNA specific amplicon primer sets for the RNA Biomarker Amplicon Sequencing (RBAS), nucleotides incorporating sequencing primers were added to the 5′ end of the primers in the first round PCR as described in Table 4 below, and a second set of primers used for a second round of PCR were used to add further sequences containing an index and adaptor sequence.
All primers used in the studies described herein were designed by the inventors and supplied by Invitrogen or IDT, except for a set of primers for PSA (KLK3) which are taught by Hessels et al. (European Urology 44: 8-16, 2003.
The ability of RNA Biomarker Amplicon Sequencing (RBAS) to be used for the accurate detection and relative quantification of multiple RNA biomarkers was demonstrated by:
An amplicon is defined as the specific amplification product obtained by PCR using a pair of oligonucleotide primers targeted to a specific RNA biomarker. The template used for the amplicon production was the single strand DNA complementary to the RNA extracted from LNCaP and A549 cells (see method section above). The cDNA was produced using random primers in this example but biomarker specific primers can also be used to initiate the reverse transcription from the extracted RNA.
DNA amplicons compatible with Illumina Corporation's Next Generation Sequencing technology were produced in this example. Amplicons compatible for sequencing using other NGS technology can also be prepared using the same rationale. The 25 specific primer pairs were targeted to 21 prostate cancer RNA biomarkers and 4 reference RNA biomarkers and contained added sequences for adaptor introduction to the 5′ and 3′ ends of the amplicons according to Illumina's specification (the RNA biomarker selection and primer design strategies are presented in the method section above).
Technical triplicates for each individual RNA biomarker were produced during a first round of PCR. The same cDNAs produced from RNA of LNCaP or A549 cells were used as a template for each of the three separate first round PCR amplifications. Six amplicon pools were then prepared by combining equal volumes of each of the 25 biomarker specific amplicons produced individually during the first round PCRs. These six amplicon pools, technical triplicates for each of the two cell types, were purified to remove residual primers and dNTPs using Agencourt AMPureXP system (Beckman Coulter, Inc.), and then analyzed with the 2100 Bioanalyser (Agilent Technologies Inc.) and Qubit® 2.0 Fluorometer (Life Technologies) to ascertain quality, average size distribution and the concentration of amplicons in each pool.
After dilution, the six cleaned amplicon pools were used as individual templates for the second round PCR performed with sequencing primers specific for the adaptor added during the first round PCR. The sequencing primers also contained a barcode sequence for indexing and a tag sequence for clustering. The amplicon libraries produced during the second round PCR were analyzed and the concentration determined using the 2100 Bioanalyser (Agilent Technologies, Inc.) and Qubit® (Life Technologies—Invitrogen). Residual primers and dNTPs were removed using Agencourt AMPureXP system (Beckman Coulter, Inc.) and then pooled together at equimolar concentration to produce a single amplicon library sequencing pool. The sequencing pool was denatured and further diluted for cluster generation and sequenced on a HiSeq2000 according to Illumina Corporation's standard protocols (208 cycles sequencing program, paired-end with indexing).
Illumina bcl2fastq conversion software (version 1.8.3) was used for the de-multiplexing of the sequence reads acquired during the sequencing program and base call conversion to fastq paired end read data. Quality statistics for percentage of bases>Q30 and mean QScore for all reads showed that all amplicon libraries sequenced and de-multiplexed very well. This data set was used to generate the read counts per amplicon (Read counts (Rc) Tables 5 and 6). This is the number of sequencing reads of at least 50 bp in length that map to the corresponding amplicon. This number is directly proportional to the amount of the amplicon in the library, and is also proportional to the specific RNA biomarker abundance from which the amplicon was derived.
By using the read count obtained for each amplicon it is thus possible to establish a precise assessment of the relative abundance of the corresponding RNA biomarkers in each sample studied.
Different methods can be used for the normalization of the read count to minimize biases generated by the acquisition of wide count distribution by massive parallel sequencing. The average of the read counts obtained from the four reference amplicons were used to normalize the raw read counts of the amplicons produced from the LNCaP and A549 RNA using the 21 primer pairs specific for the prostate cancer RNA biomarkers. The reference amplicons were made with specific primers targeted to four different RNA biomarkers selected due to their low level of expression variation between different prostate cancer and healthy donor control tissues. The raw counts obtained for the four reference amplicons derived from A549 and LNCaP RNA were consistent between replicates and between the two cell types compared (Table 5). The data confirms the low level of differential expression of these reference RNAs and validates the selection of these RNA biomarkers as reference amplicons.
In Table 6, the normalization of the read count for each of the non-reference RNA biomarker specific amplicons derived from LNCaP and A549 RNA (termed target amplicons) was calculated by dividing each target read count by the average read count calculated from the mean of the four reference amplicons either from LNCaP or A549 RNA. This normalization was performed for each replicate (Table 6: target amplicon read counts/average references read counts).
The assessment of the RNA biomarker differential expression fold change (FC) between the LNCaP and A549 cells was performed by comparing the normalized read counts per amplicon converted to a log2 number. The log2 FC was calculated for the read counts before (raw read counts) and after normalization (Normalised read counts) and was compared in order to assess the effect of the amplicon library count distributions on the evaluation of the differential expression (Table 6). The data in Table 6 compares the expression of 21 target RNA biomarkers in LNCaP and A549 cells. A negative log2 number indicates a decrease, or down regulation of RNA biomarkers while a positive log2 number indicates an increase, or up regulation of RNA biomarkers.
The data shows that the difference between FC values calculated either using the log2 value for raw counts or the log2 value for the normalized counts is not large. However, the normalization process allows a more accurate detection of the relative difference in expression of RNA biomarkers in A549 and LNCaP cells.
For the data in Table 7 we have accepted Log2 FC values greater than 2 are significant and grouped the expression levels of the 21 prostate cancer specific RNA biomarkers tested using LNCaP and A549 RNA in two groups: Log2FC>2; and Log2FC<2.
The data reveals an even split of RNA biomarkers with Log2 FC>2 between the two RNAs.
The data contained in Table 8 are basic statistical analyses of the Log2 FC differences between the 21 RNA biomarkers expressed in LNCaP and A549 RNA calculated by dividing the normalized Log2 FC of each RNA biomarker from LNCaP RNA by the corresponding Log2 FC from A549 RNA. The level of differential expression calculated by the limma-based linear model fit analysis (T=limma moderated t−statistic) highlights some significant levels of differential expression of the RNA biomarker between the LNCaP and A549 cell types (T value) with correlating P value.
These two cell lines, LNCAP and A549, were chosen for this example to demonstrate a proof of concept by comparing RNA biomarker expression in two cell lines; one (LNCaP cells) of prostate origin and the other (A549 cells) of lung origin. As might be expected, there is significant differential expression between these two cell lines of the RNA biomarkers chosen on the basis of their possible involvement in prostate cancer.
The data provided in the above example shows that it is possible to detect the change in expression of specific RNA biomarkers through quantitative amplicon synthesis followed by enumeration using a Next Generation DNA sequencing methodology.
This example demonstrates that the RNA amplicon biomarker sequencing (RBAS) method is diagnostically and prognostically relevant by quantifying the relative expression of 79 RNA biomarkers using amplicon production and NGS to establish their RNA expression profile in prostate cancer tissues.
Stored formalin-fixed paraffin embedded (FFPE) prostatectomy tissue blocks were reviewed by a clinical histopathologist to select tissues for analysis. Prostatectomy tissue from two subjects was selected.
Subject 1 is a 63 year old male who underwent a prostate biopsy in 2007 and was diagnosed with prostate cancer with a Gleason score of 4+5. The subject underwent a radical prostatectomy at the age of 58. A stored FFPE block containing the original prostatectomy tissue was re-examined and a tumor region was identified with a Gleason score of 4+5. The region identified was reset in paraffin and then sectioned. Three tissue samples were selected from Subject 1 for RNA extraction: Tumor tissue 4+5 (T); adjacent glandular tissue (Adj.G); and adjacent muscle tissue (Adj.M) deemed histologically normal.
Subject 2 is a 67 year old male who underwent a prostate biopsy in 2012 and was diagnosed with prostate cancer with a Gleason score of 3+4. The subject underwent a radical prostatectomy at the age of 66. A stored FFPE block containing the prostatectomy tissue was re-examined. Three tumors were identified with different Gleason scores, 4+5 (T1), 3+4 (T2) and 3+3 (T3) respectively. The different regions from the blocks were reset, and then sectioned. Tissue samples were selected from each of the three tumor regions as well as an adjacent glandular tissue (Adj.G) deemed histologically normal. No Adj.M region was identified in Subject 2 tissue samples.
Total RNA was extracted separately from the seven selected tissue samples from Subject 1 and 2 using a Qiagen FFPE RNeasy extraction kit (Cat No: 74404, 73504). The RNA was then used to generate cDNA for each tissue sample as described above in the methods section. This cDNA was used for amplicon production in triplicate, using a total of 79 RNA biomarker primer pairs that included five reference amplicons from four RNA biomarkers. The second round PCR sequencing of the 79 RNA biomarker specific amplicons produced in the first round PCR was done in two separate runs. During the second round PCR, the barcode sequence for indexing and a tag sequence were added and the amplicon libraries were pooled together for clustering and sequencing on the Illumina Hiseq2500 instrument as described in Example 1.
As described in Example 1, Illumina bcl2fastq conversion software (version 1.8.3) was used to obtain the number of sequence reads per amplicon (read counts).
The raw counts of the five reference amplicons from each of the sequencing runs (Run1, Run2) is presented in Table 9. The sequence counts for all the reference amplicons were lower in run 1 than the run 2. However, the ratio of the individual reference RNA biomarkers to each other was very similar in the two runs.
The raw counts obtained for the reference amplicons presented in Table 9 were generally consistent between replicates across the prostatectomy-derived RNA samples and the data supports the selection of these RNA biomarkers as reference amplicons.
The average of the read counts from the five reference amplicons was used to normalize the raw read counts of the amplicons produced from the appropriate tumor and adjacent glandular and muscular tissue pairings.
For the analysis of Subject 1, the data compared the relative expression of the RNA biomarkers between tumor tissue and both adjacent glandular and adjacent muscular tissue. The raw counts of triplicate samples from tumor tissue and both adjacent glandular and adjacent muscular tissue is given followed by the log2 normalized counts. The log2 FC expression of each RNA biomarker from the tumor region of the prostatectomy tissue RNA samples is given relative to the adjacent glandular and muscular adjacent muscular tissue RNA. Finally the log2 FC of the adjacent glandular relative to the muscular adjacent muscular tissue RNA is presented (Table 10).
Those RNA biomarkers with a differential amplicon count (Loge FC>2) from Subject 1 were selected from the tumor, adjacent glandular and adjacent muscular samples with the data being presented in Table 11.
A number of biomarkers are found to be differentially expressed in either the tumor samples or the adjacent glandular or muscular tissues and these have been grouped in Table 12 below.
It is common practice in this area of cancer research, particularly when using archival FFPE blocks as the source of tumor tissue, to use tissue adjacent to the tumor as control healthy tissue when studying differential expression. However, studies that have compared gene expression profiles or the chromatin status of prostate tumor tissue with adjacent tissue and benign prostate tissue from brain dead organ donors with no evidence of prostate cancer have suggested that the adjacent tissue has a genome and transcriptome that is more similar to the tumor than to the donor control tissues, suggesting that field effects exist (Chandran et al. 2005, Aryee et al. 2013).
The RBAS analysis using Subject 1 tissue shows that the glandular adjacent tissue has an RNA expression profile more similar to the tumor which is very likely due to field effects as described for prostate cancer tissues by Chandran et al (2005), Rizzi et al. (PLoS One 3(10):e3617, 2008) and reviewed in Trujillo et al. (Prostate Cancer, 2012).
The analysis of Subject 2 used prostatectomy tissue and the data compares the relative expression of the RNA biomarkers between three tumor tissues with different Gleason scores (termed T1, T2, and T3) to the adjacent glandular tissue only. The raw counts of triplicate samples from T1, T2 and T3 tumor tissues and adjacent glandular tissue is given followed by the log2 normalised counts. Finally the log2 FC expression of each RNA biomarker from the tumor region of the prostatectomy tissue RNA samples is given relative to the adjacent glandular tissue RNA.
The raw counts acquired for each amplicon from Subject 2 samples is presented in Table 13 with the calculation of the normalized count and FC.
In Table 14, the data represents those RNA biomarkers with a Loge FC>2 in the differential expression in the tumour compare to the adjacent gland. Most of these RNA biomarkers are up regulated in the tumor compared with the adjacent glandular tissue. Only two biomarkers were detected in a higher amount in the adjacent glandular tissue compared with all tumors. Some distinctions between the different grades of tumors can be made, for example with the OPRK1 and PSMA RNA biomarkers.
Before proceeding with the amplicon production for RBAS analysis, the efficiency of all the RNA specific primers was tested by real time PCR or by visualization of the produced amplicon of the expected size. Therefore, the lower sequence counts observed for certain amplicons produced from prostatectomy tissues RNA cannot be attributed to the inefficiency of the amplicon production. As seen in Example 1, raw sequence counts of 900 and 13,000 were obtained from the MUC1 amplicon produced from LNCaP and A549 cell RNA respectively (Table 6).
The process used to select RNA biomarkers disclosed herein is by selecting those that are up-regulated or down-regulated in a small number of prostate tumors, rather than in all prostate tumors. For this reason it is not expected that differential expression of all the RNA biomarkers would be seen in all prostate tumors or their adjacent tissues. The data indicate that tumors examined from Subjects 1 and 2 are likely not to have some of the RNA dysregulated within their tissue. The analysis of tumors from a range of subjects will will likely reveal differences in the expression of these and other RNA biomarkers. That is the major reason why, for diagnostic and prognostic use, RNA biomarker panels are selected from a large RNA biomarker pool. RBAS methodology has been developed to allow rapid screening of tumor samples for a large number of RNA biomarkers simultaneously.
In conclusion, these observations highlight the issue with staging prostate cancers and illustrate reasons for developing multi-RNA biomarker diagnostics, as it is unlikely that a single RNA biomarker can diagnose and stage prostate cancers, or distinguish prostate cancer from benign prostate hyperplasia or prostatitis.
While the present invention has been described with reference to the specific embodiments thereof, it should be understood by those skilled in the art that various changes may be made and equivalents may be substituted without departing from the true spirit and scope of the invention. In addition, many modifications may be made to adapt a particular situation, material, composition of matter, method, method step or steps, for use in practicing the present invention. All such modifications are intended to be within the scope of the claims appended hereto.
All of the publications, patent applications and patents cited in this application are herein incorporated by reference in their entirety to the same extent as if each individual publication, patent application or patent was specifically and individually indicated to be incorporated by reference in its entirety.
SEQ ID NO: 1-326 are set out in the attached Sequence Listing. The codes for nucleotide sequences used in the attached Sequence Listing, including the symbol “n,” conform to WIPO Standard ST.25 (1998), Appendix 2, Table 1.
| Number | Date | Country | |
|---|---|---|---|
| 61665849 | Jun 2012 | US | |
| 61691743 | Aug 2012 | US | |
| 61709517 | Oct 2012 | US |
