Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
NON-INVASIVE DETECTION OF COLORECTAL NEOPLASIA USING CIRCULAR CONSENSUS SEQUENCING
Document Type and Number:
WIPO Patent Application WO/2014/093186
Kind Code:
A1
Abstract:
A non-invasive test for colorectal neoplasia is described that relies on Single Molecule Real Time Circular Consensus (SMRT-CCS) Sequencing. Genomic DNA from the patient is obtained from either a blood or a stool sample and regions of genes known to mutated in colorectal neoplasia are amplified for SMRT-CCS sequencing. For the APC gene, it is necessary to scan at least 500 bp in order to detect mutations at an acceptable frequency. When a mutation is present, the mutated DNA is often between 0.5-2%, hence a sensitive method is required. SMRT-CCS allows for thousands of individual molecules of DNA to be sequenced at high accuracy. The combination of sequencing depth and accuracy allows detection of low level variants. Other genes such as Beta Catenin, KRAS, BRAF, and P53 can also be scanned in parallel for increased sensitivity.

Inventors:
GARVIN ALEX M (FR)
HOEHN FREDERIC (FR)
Application Number:
PCT/US2013/073772
Publication Date:
June 19, 2014
Filing Date:
December 09, 2013
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
GARVIN ALEX M (FR)
HOEHN FREDERIC (FR)
International Classes:
C12Q1/68; G01N33/48; G01N33/50
Domestic Patent References:
WO2012159754A22012-11-29
Foreign References:
US20110207134A12011-08-25
Other References:
KOGA. ET AL.: "Detection of the DNA point mutation of colorectal cancer cells isolated from feces stored under different conditions.", JPN J CLIN ONCOL., vol. 39, no. 1, 2009, pages 62 - 9
DIEHL ET AL.: "Detection and quantification of mutations in the plasma of patients with colorectal tumors.", PROC NATL ACAD SCI USA, vol. 102, no. 45, 2005, pages 16368 - 73
"PacBio RS Sequencing.", May 2012 (2012-05-01), pages 1, 4, 5, 11, Retrieved from the Internet [retrieved on 20140225]
SMITH ET AL.: "Single Molecule Real Time (SMRT) Sequencing Sensitively Detects Polyclonal and Compound BCR-ABL in Patients Who Relapse on Kinase Inhibitor Therapy.", 3 February 2012 (2012-02-03), Retrieved from the Internet [retrieved on 20140224]
TRAVERS ET AL.: "A flexible and efficient template format for circular consensus sequencing and SNP detection.", NUCLEIC ACIDS RES., vol. 38, no. 15, 2010, pages E159
ONOUCHI ET AL.: "New method for colorectal cancer diagnosis based on SSCP analysis of DNA from exfoliated colonocytes in naturally evacuated feces.", ANTICANCER RES., vol. 28, no. 1A, 2008, pages 145 - 50
LECOMTE ET AL.: "Circulating free tumor DNA and colorectal cancer.", GASTROENTOROLOGIE CLINIQUE ET BIOLOGIQUE, vol. 34, no. 12, 2010, pages 662 - 681
RUSSO ET AL.: "Early and high sensitive detection of colorectal cancer mutations using third generation sequencing.", April 2013 (2013-04-01), pages 64, Retrieved from the Internet [retrieved on 20140225]
RUSSO ET AL.: "High sensitivity detection of colorectal cancer mutations using third generation single molecule sequencing.", 9 November 2013 (2013-11-09), Retrieved from the Internet [retrieved on 20140225]
Attorney, Agent or Firm:
GARVIN, Alex, M (Mulhouse, FR)
Download PDF:
Claims:
CLAIMS

1 ) A non-invasive method to detect colorectal cancer in humans using stool DNA as analyte that requires:

a) Isolation of stool DNA,

b) Amplification of at least 500 bases between codons 1209 and 1581 of the human APC gene to generate amplification products, and

c) Sequencing individual molecules of said amplification products by Single Molecule Real Time Circular Consensus Sequencing.

2) A non-invasive method to detect colorectal cancer in humans using cell-free DNA from blood plasma as analyte that requires:

a) Isolation of cell-free DNA from blood plasma,

b) Amplification of at least 500 bases between codons 1209 and 1581 of the human APC gene to generate amplification products, and

c) Sequencing individual molecules of said amplification products by Single Molecule Real Time Circular Consensus Sequencing.

Description:
NON-INVASIVE DETECTION OF COLORECTAL NEOPLASIA USING CIRCULAR CONSENSUS SEQUENCING

CROSS-REFERENCE TO RELATED APPLICATIONS

The present application claims benefit of US Provisional Patent Application No 61 /736, filed December 12th, 2012, the full disclosures of which are incorporated herein by reference in their entirety for all purposes. STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH

Not Applicable.

BACKGROUND OF THE INVENTION

Colorectal cancer (CRC) is the fourth most deadly malignancy worldwide, resulting in 608,000 deaths in 2008. However, this disease can be prevented by early detection and surgical removal of the neoplasia as demonstrated by the 80% reduction in mortality seen in patients that are screened by colonoscopy (Winawer et al., 1993). CRC is caused by mutations in specific genes, and trace amounts of mutated DNA from cancerous tissue obtained from patient samples can be used as the basis for a diagnostic test for CRC. For example, venous blood contains cell free DNA (cf-DNA) present in the plasma fraction of blood at between 1 -100 ng of DNA per ml of plasma. When a patient has a colorectal tumor, mutated DNA from the tumor can be detected in cf-DNA(Grutzmann et al., 2008). Furthermore, mutated DNA from CRC tumors and even from pre-cancerous polyps can be found in DNA isolated from stool samples (Traverso et al., 2002). The difficulty with these approaches of screening for tumors using DNA from blood or stool is that the mutations present in the tumors can be found in a number of genes, at different positions in these genes, and are present at low frequency in the DNA sample, as most of the DNA from stool or plasma will be wildtype even if the patient has a tumor. Therefore any method for detecting the mutated DNA in a screening assay must be highly sensitive, preferably capable of detecting mutations present at 1 part in 1000 (0.1 %) and must be able to detect a wide range of mutations in a number of genes because neither the presence of a mutation nor the precise sequence of the mutation, when present, is known ahead of time. Single Molecule Real Time (SMRT) sequencing performed on circular single stranded DNA templates derived from PCR products can be used to determine a highly accurate circular consensus sequence (CCS) from a single DNA molecule. The ability of SMRT-CCS sequencing to detect mutations anywhere in a test sequence at high accuracy enables the detection of rare tumor derived mutations in DNA isolated from patient samples.

The present invention is directed towards a sensitive and specific method for detecting mutations in genes known to be altered in colorectal neoplasia using DNA isolated from either blood or stool from a patient. The mutations are detected by sequencing thousands of individual DNA molecules derived from the patient sample using SMRT-CCS sequencing.

BRIEF SUMMARY OF THE INVENTION

A non-invasive test for colorectal neoplasia is performed by isolating DNA from either cell free DNA derived from blood, or from a stool sample and this DNA is purified to allow PCR amplification. Regions of a number of genes known to be mutated in colorectal neoplasia are then amplified, the most important being at least 500 bases present between codons 1209 and 1581 of the APC gene. Thousands of individual DNA molecules from these amplicons are then sequenced using Single Molecule Real Time Circular Consensus Sequencing. When mutations known to cause colorectal neoplasia are found in these sequences, the patient is likely to have a pre- cancerous or a cancerous growth in the lower gastrointestinal tract.

BRIEF DESCRIPTION OF THE DRAWINGS

Drawing 1 lists the 5 genes that are analysed in example 1 . The codons that are covered, the number of segments required, and the amplicon sizes are listed.

Drawing 2 shows the 30 oligonucleotide primers used to amplify the test sequences from drawing 1 .

Drawing 3 presents gels of the PCR products obtained using the primers in drawing 2 when genomic DNA from a cell line(3a), stool DNA (3b) and cell free DNA from blood plasma(3c) are used as template.

Drawing 4 shows the sequence variants detected when amplicons derived from a DNA template that is 1 % mutant (from the DLD1 cell line) are sequenced using SMRT-CCS. Roughly 3,000 high accuracy circular consensus reads were obtained for each amplicon.

DETAILED DESCRIPTION OF THE INVENTION DEFINITIONS

To help in understanding the current invention, a number of definitions are provided.

"Neoplasia" as used herein refers to an abnormal and new growth of cells in a patient.

The term "adenoma" means a neoplastic growth of epithelial cell origin that has not yet invaded the surrounding tissue. The term "PCR amplification" refers to the Polymerase Chain Reaction process of in vitro amplification of DNA using sequence specific oligonucleotides directed against the target sequence and a thermostable polymerase. Multiple thermal cycles consisting of denaturation, hybridization, and extension generate multiple copies of the target sequence.

The term "amplicon" refers to the double stranded DNA product of PCR from a single pair of oligonucleotide primers.

The term "primer" refers to a synthetic oligonucleotide, designed to amplify a given target sequence.

The term "sensitivity" measures the proportion of actual positives which are correctly identified as positive by the test.

The term "specificity" is defined as the percentage of negatives as called by the test that are actually negative.

One method to detect any mutation in a sequence of interest is to PCR amplify the sequence from genomic DNA taken from a patient sample and to sequence the PCR product, provided that the mutated sequence is present at a high enough frequency to be detected by the sequencing method. Standard Sanger sequencing can detect a mutant present at roughly 25%, which is far too insensitive for the application under consideration because most patients with colorectal neoplasia have a mutant allele frequency much less than 25% in stool or blood DNA samples. Second generation sequencing (Ion Torrent, lllumina, and 454) can obtain higher sensitivity by sequencing a large number of clones derived from single molecules, although the sensitivity for all of these methods is limited by the fact that these sequence-by-synthesis methods all require a cluster, or clone of template molecules derived from a single DNA molecule. As each step of synthesis is less than 100% efficient, the template becomes increasingly heterogeneous with each round of extension, and this heterogeneity creates noise that degrades the quality of the raw data.

One way to minimize this problem of signal degradation with increased read length is to only use short reads, and in a recent application of the lllumina HiSeq platform for detecting mutations in tumor DNA samples (Frampton et al., 2013), which are heterogeneous due to the presence of non- mutant stromal cells and leukocytes, reads of only 49 bases were used, even though this platform can read up to 150 bases per run. In the cited study the sensitivity of detecting base substitutions was enhanced by using a Bayesian method, which allows false positives at positions not previously known to be mutated in tumors to be discounted. The sensitivity in this study dropped off rapidly when the mutant allele frequency fell below 5%. Commercially available products for detecting mutations in carcinoma derived tumor samples, such as the Ion AmpliSeqâ„¢Cancer Hotspot Panel v2 and the TruSeq Amplicon-Cancer Panel, also claim detection of base substitutions in the when the mutant allele frequency is above 5%. Since the mutant allele frequency in stool DNA in patients with colorectal neoplasia is often much less than 5%, see column 7 of table 1 in (Traverso et al., 2002), these sequencing platforms should not have the sensitivity required for this application. Single Molecule Real Time (SMRT) sequencing is a time dependent sequencing-by-synthesis method (US Pat Nos 7,033,764, 7,052,847, 7,056,661 , and 7,056,676) that can be used to sequence PCR products once these products are transformed into circular single stranded molecules through the ligation of hairpin loops at each end. Due to the multiple kilobase long read lengths that are possible with this SMRT sequencing, a short DNA test sequence on the order of hundreds of bases in length can be sequenced many times by simply allowing the reaction to continue around the circular template and collecting multiple sub-reads. The circular consensus sequence (CCS) obtained from these sub-reads has a higher accuracy than that of the sub-reads alone because random errors in one sub-read are not likely to be repeated on subsequent sub-reads and will not be part of the consensus (US Pat No 8,535,822).

SMRT sequencing has an error rate of approximately 15%(85% accuracy), however, when CCS is performed the errors cancel out and the accuracy of the consensus for the single molecule of DNA being read can be quite high and depends on the number of sub-reads. When the sequence is read 6 times, the accuracy of the consensus sequence approaches 99.99% (see figure 7 of (Travers et al., 2010)), and further improvements in accuracy can be obtained by further increasing the number of sub-reads. Accuracy in this range should be sufficient to detect the rare mutations expected in stool and blood DNA of patients with colorectal neoplasia.

The SMRT cell that is used for SMRT sequencing has roughly 150,000 zero mode waveguides (ZMW) where a single molecule of DNA can bind, and when loaded with template/polymerase roughly half of the ZMWs, or 75,000, are correctly loaded. Since only a fraction will give read lengths that are long enough to provide accurate consensus sequences, the number of useful reads will be less than 75,000, and is often in the range of 45,000-55,000. However, even if only 50,000 molecules are read with 99.99% accuracy, this should be sufficient to provide an accurate and sensitive assay for noninvasive detection of cancer as described below.

For colorectal cancer screening, the APC gene is considered the ideal biomarker because it is usually the first gene to be mutated in the multistep process that leads to metastatic tumors. Another gene in the Wnt signalling pathway, the beta catenin gene, is often mutated in those polyps where the APC gene is not mutated. Mutations in the KRAS, BRAF and the TP53 genes occur later in tumor development and can also be detected in stool DNA or in cf-DNA. A highly sensitive assay detecting mutations in these 5 genes should therefore be useful in for detecting CRC and the polyps that lead to CRC using stool DNA or cf-DNA as starting material.

Previous attempts at developing a sensitive and accurate non-invasive test for CRC using stool DNA as analyte that focused on the APC gene used assays such as the digital protein truncation test (see US Pat No 7,910,300), the Elisa Protein Truncation Test (see US Pat No 8,1 14,587), a droplet Protein Truncation Test (see US Pat Application 20100304378), and a Protein Truncation Test that uses MALDI-TOF mass spectroscopy for detection (see US Pat Nos 8,507,648, 7,807,407, 7,563,598, and 6,329,180). Furthermore, US patent application 20130012410 describes using melting curve analysis and direct sequencing to detect mutations in the APC gene, although no mention was made concerning how the sensitivity and specificity of direct sequencing would be sufficient for this application. Additional effort has also been spent developing stool-based tests for CRC including work done to stabilize the genomic DNA in a stool sample (US Pat No 6,177,251 ), assays that detect variation in DNA methylation patterns (US Pat Nos 7,785,772 and 8,053,189) and an assay that relies on DNA quantitation (US Pat No 8,343,722).

Apparently no previous attempt has been described that uses massively parallel sequencing to scan the APC gene for mutations in order to detect CRC, although work has been done to improve the sensitivity and specificity of one type of massively parallel sequencing (the HiSeq lllumina platform) that could be used for analysis of patient samples (Kinde et al., 201 1 ). One potential reason that existing methods of performing massively parallel sequencing (MPS) have not been considered for this application is that the error rate in the sequence generated by the most common forms of MPS is too high to allow sensitive and accurate detection of mutations in stool DNA.

Example of a SMRT-CCS Assay for CRC

Drawing 1 lists 5 genes that are frequently mutated in CRC tumors and that have been chosen for a SMRT-CCS based test for CRC. The APC gene is the most important gene to assay in such a test because it is the first gene to be mutated in the multi-step process of CRC tumorigenesis (Levy et al., 1994, Parsons et al., 2005, Powell et al., 1992, Sparks et al., 1998) and hence able to detect the disease at its earliest stage, and it is the gene mutated in the highest percentage of CRC patients, 80%, which is far more than any other gene. Unfortunately, the APC gene is also the most difficult gene to assay because mutations can be found in a large region of the gene (over 1 ,000 bp) and the mutations that are present in neoplasia can be either insertions, deletions, or base substitutions. In all cases, the mutation in the APC generates a premature stop codon, either by reading frame shift of nonsense mechanisms.

In the assay described here, the entire region including and between codon 840 and codon 1 581 are covered using 8 overlapping amplicons of between 324 to 344 bases each. This test sequence size (2,226 bp) is exactly twice the size of the standard APC test sequence region referred to as the Mutation Cluster Region (MCR), which is often defined as being 1 ,1 1 3 bp in length covering codons 1210 to 1 581 of the APC gene (Laurent-Puig et al., 1998). Four other genes known to be frequently mutated in CRC, Beta Catenin (Sparks et al., 1998), KRAS (Sidransky et al., 1992), P53 (Liu and Bodmer, 2006), and BRAF (Benlloch et al., 2006), are also tested. These genes (with the exception of Beta Catenin) are mutated after the APC gene during the process of tumor formation and at lower frequencies than the APC gene.

The oligonucleotide primers listed in Drawing 2 were synthesized in order to amplify 1 5 sequences from these 5 genes for SMRT-CCS. All the amplicons are of similar size (324-344 bp) so that each will be equally represented (smaller amplicons are overrepresented when pooling products for SMRT sequencing), provided that equal molar amounts of PCR product are used for each amplicon. Each amplicon is amplified separately and the concentration and size of each product is determined using a Bioanalyzer (Agilent Corporation). Drawing 3 shows amplification products from genomic DNA derived from cell lines, a stool sample, and cell free DNA from blood plasma. Equal molar amounts of the 15 amplicons are pooled, and then used to generate a library for SMRT-CCS using the standard kit provided by Pacific Biosciences, the DNA Template Prep Kit 2.0.

The library is sequenced using the standard protocol for Pacific Biosciences SMRT sequencing with 1 SMRT cell used per library and the data are analysed using the SMRT Analysis program. Approximately 50,000 reads are generated with sub-reads greater or equal to 6, and the raw accuracy of the consensus sequences is as high as 99.9999% (QV 60). As equal molar amounts of each amplicon is pooled to make the library, and since each amplicon is of approximately the same size, amplicon coverage is uniform and each amplicon is sequenced about 3,000 times. When a mutation is present at the 1 % level, one would expect 30 mutant reads, which is more than enough to obtain an unambiguous positive signal, provided that the background due to sequencing errors is well below 1 %.

The reference sequence is obtained from Medline, and a control sequence obtained using 100% wildtype DNA from the K562 cell line is compared with the test sequence of 1 % mutant DNA. A P value is calculated that indicates the probability of obtaining the given number of variant reads when the test sequence is wildtype. Drawing 4 shows that all 3 mutations present in the DLD1 cell line were detected using this method when the mutant DNA was present at the 1 % level. The first false positive (a T insertion at c2806 of the APC gene) to be detected has a P value 28 times higher than the highest P value for a true mutant, demonstrating that the method was both sensitive and specific. Importantly, the assay scanned 2,226 bases of the APC gene and was able to detect the single base deletion in the APC gene present at the 1 % level in this sample without generating any false positive calls above the 0.3% level.

REFERENCES

BENLLOCH, S., PAYA, A., ALENDA, C, BESSA, X., ANDREU, M., JOVER, R., CASTELLS, A., LLOR, X., ARANDA, F. I. & MASSUTI, B. 2006. Detection of BRAF V600E mutation in colorectal cancer: comparison of automatic sequencing and real-time chemistry methodology. The Journal of molecular diagnostics : JMD, 8, 540-3.

FRAMPTON, G. M., FICHTENHOLTZ, A., OTTO, G. A., WANG, K., DOWNING, S. R., HE, J., SCHNALL-LEVIN, M., WHITE, J., SANFORD, E. M., AN, P., SUN, J., JUHN, F., BRENNAN, K., IWANIK, K., MAILLET, A., BUELL, J., WHITE, E., ZHAO, M., BALASU BRAMAN I AN , S., TERZIC, S., RICHARDS, T., BANNING, V., GARCIA, L, MAHONEY, K., ZWIRKO, Z., DONAHUE, A., BELTRAN, H., MOSQUERA, J. M., RUBIN, M. A., DOGAN, S., HEDVAT, C. V., BERGER, M. F., PUSZTAI, L, LECHNER, M., BOSHOFF, C, JAROSZ, M., VIETZ, C, PARKER, A., MILLER, V. A., ROSS, J. S., CURRAN, J., CRONIN, M. T., STEPHENS, P. J., LIPSON, D. & YELENSKY, R. 2013. Development and validation of a clinical cancer genomic profiling test based on massively parallel DNA sequencing. Nature biotechnology.

GRUTZMANN, R., MOLNAR, B., PILARSKY, C, HABERMANN, J. K., SCHLAG, P. M., SAEGER, H. D., MIEHLKE, S., STOLZ, T., MODEL, F., ROBLICK, U. J., BRUCH, H. P., KOCH, R., LIEBENBERG, V., DEVOS, T., SONG, X., DAY, R. H., SLEDZIEWSKI, A. Z. & LOFTON- DAY, C. 2008. Sensitive detection of colorectal cancer in peripheral blood by septin 9 DNA methylation assay. PloS one, 3, e3759.

KINDE, I., WU, J., PAPADOPOULOS, N., KINZLER, K. W. & VOGELSTEIN, B. 201 1 . Detection and quantification of rare mutations with massively parallel sequencing. Proceedings of the National Academy of Sciences of the United States of America, 108, 9530-5.

LAURENT-PUIG, P., BEROUD, C. & SOUSSI, T. 1998. APC gene: database of germline and somatic mutations in human tumors and cell lines. Nucleic acids research, 26, 269-70. LEVY, D. B., SMITH, K. J., BEAZER-BARCLAY, Y., HAMILTON, S. R., VOGELSTEIN, B. & KINZLER, K. W. 1994. Inactivation of both APC alleles in human and mouse tumors. Cancer research, 54, 5953-8. LIU, Y. & BODMER, W. F. 2006. Analysis of P53 mutations and their expression in 56 colorectal cancer cell lines. Proceedings of the National Academy of Sciences of the United States of America, 103, 976-81 . PARSONS, D. W., WANG, T. L, SAMUELS, Y., BARDELLI, A., CUMMINS, J.

M., DELONG, L, SILLIMAN, N., PTAK, J., SZABO, S., WILLSON, J. K., MARKOWITZ, S., KINZLER, K. W., VOGELSTEIN, B., LENGAUER, C. & VELCULESCU, V. E. 2005. Colorectal cancer: mutations in a signalling pathway. Nature, 436, 792. POWELL, S. M., ZILZ, N., BEAZER- BARCLAY, Y., BRYAN, T. M., HAMILTON, S. R., THIBODEAU, S. N., VOGELSTEIN, B. & KINZLER, K. W. 1992. APC mutations occur early during colorectal tumorigenesis. Nature, 359, 235-7.

SIDRANSKY, D., TOKINO, T., HAMILTON, S. R., KINZLER, K. W., LEVIN, B., FROST, P. & VOGELSTEIN, B. 1992. Identification of ras oncogene mutations in the stool of patients with curable colorectal tumors. Science, 256, 102-5.

SPARKS, A. B., MORIN, P. J., VOGELSTEIN, B. & KINZLER, K. W. 1998.

Mutational analysis of the APC/beta-catenin/Tcf pathway in colorectal cancer. Cancer research, 58, 1 130-4. TRAVERS, K. J., CHIN, C. S., RANK, D. R., EID, J. S. & TURNER, S. W. 2010. A flexible and efficient template format for circular consensus sequencing and SNP detection. Nucleic acids research, 38, e159.

TRAVERSO, G., SHUBER, A., LEVIN, B., JOHNSON, C, OLSSON, L, SCHOETZ, D. J., JR., HAMILTON, S. R., BOYNTON, K., KINZLER, K. W. & VOGELSTEIN, B. 2002. Detection of APC mutations in fecal DNA from patients with colorectal tumors. The New England journal of medicine, 346, 31 1 -20.

WINAWER, S. J., ZAUBER, A. G., HO, M. N., O'BRIEN, M. J., GOTTLIEB, L.

S., STERNBERG, S. S., WAYE, J. D., SCHAPIRO, M., BOND, J. H., PANISH, J. F. & ET AL. 1993. Prevention of colorectal cancer by colonoscopic polypectomy. The National Polyp Study Workgroup. The New England journal of medicine, 329, 1977-81 .