The authors have declared that no competing interests exist.
Conceived and designed the experiments: GP TAY JSDB. Performed the experiments: GP TAY LP JPD RR CM PAC SM JC KAD KT DGDC LRM SBK UB. Analyzed the data: GP TAY LP AMC JPD RR GA JSDB. Wrote the paper: GP TAY JSDB.
Tumor genomic instability and selective treatment pressures result in clonal disease evolution; molecular stratification for molecularly targeted drug administration requires repeated access to tumor DNA. We hypothesized that circulating plasma DNA (cpDNA) in advanced cancer patients is largely derived from tumor, has prognostic utility, and can be utilized for multiplex tumor mutation sequencing when repeat biopsy is not feasible. We utilized the Sequenom MassArray System and OncoCarta panel for somatic mutation profiling. Matched samples, acquired from the same patient but at different time points were evaluated; these comprised formalin-fixed paraffin-embedded (FFPE) archival tumor tissue (primary and/or metastatic) and cpDNA. The feasibility, sensitivity, and specificity of this high-throughput, multiplex mutation detection approach was tested utilizing specimens acquired from 105 patients with solid tumors referred for participation in Phase I trials of molecularly targeted drugs. The median cpDNA concentration was 17 ng/ml (range: 0.5–1600); this was 3-fold higher than in healthy volunteers. Moreover, higher cpDNA concentrations associated with worse overall survival; there was an overall survival (OS) hazard ratio of 2.4 (95% CI 1.4, 4.2) for each 10-fold increase in cpDNA concentration and in multivariate analyses, cpDNA concentration, albumin, and performance status remained independent predictors of OS. These data suggest that plasma DNA in these cancer patients is largely derived from tumor. We also observed high detection concordance for critical ‘hot-spot’ mutations (
The development of cancer is primarily due to genetic aberrations that drive oncogenesis and determine the clinical manifestations of tumors; these may also impact response to treatment
We set out to optimize benefit for patients with advanced solid tumors referred for Phase I clinical trials by allocating specific targeted therapies to patients who harbor tumor molecular aberrations targeted by the agent in question
An alternative source of tumor DNA is circulating plasma DNA (cpDNA)
Patients with late stage advanced solid tumors who were referred to the Drug Development Unit in the Royal Marsden NHS Foundation Trust between September 2009 and August 2010, and who were eligible for a Phase I trial were included in this study. All patients provided written informed consent for genetic analysis of their tumors and plasma samples prior to participation in this study. Eight mls of peripheral blood were sampled in a BD Vacutainer Cell Preparation Tube (CPT) containing sodium heparin, which permits plasma and mononuclear cell separation during a single centrifugation step. The tube was inverted a minimum of 8 times to ensure thorough mixing of the sample, and then centrifuged at 1800 g for 15 min. The resultant plasma supernatant was transferred to a clean tube and stored at −80°C until analysis. In addition, 20 healthy volunteers provided 8 ml of blood for analysis using this method. Corresponding FFPE samples (primary and/or metastatic sites) for each patient were also requested. The relevant regulatory and independent ethics committee (National Research Ethics Service (NRES) Committee London-Chelsea, United Kingdom) approved this study prior to trial commencement.
For the analyses of tumor samples, hemotoxylin- and eosin-stained slides were reviewed by a board-certified pathologist (K.T.) to ensure adequate viable tumor and to determine the tumoral zone to core. DNA from FFPE specimens was extracted from 1 mm cores when possible or from 10 µm unstained sections with smaller biopsies using the QIAamp DNA FFPE Tissue Kit (Qiagen, Valencia, CA, USA), according to the manufacturer's recommendations. The extracted DNA was subsequently eluted in 30 µl of ATE buffer and stored at −20°C until further analysis. DNA was quantified using the Nanodrop 1000 Spectrophotometer (Thermo Scientific).
For cpDNA extraction, plasma was thawed at ambient temperature and cpDNA extracted from 2 ml of plasma using a QIAamp DNA Blood Midi Kit (Qiagen, Valencia, CA, USA), according to the manufacturer's instructions, with the following modifications: for each 2 ml sample of plasma, an additional centrifugation step (16000 g, 5 min, RT) was added before the extraction procedure in order to eliminate cellular debris from the plasma. At the end of the procedure, the DNA was eluted in 100 µl of AE elution buffer. DNA concentration was measured with fluorescent staining, using the Quant-iT™ Pico-Green® double stranded DNA (dsDNA) Assay Kit (Invitrogen, Carlsbad, CA) and the SynergyHT microplate reader (Biotek). DNA from the cancer cell lines analyzed was extracted from pellets using the QIAamp DNA Mini Kit (Qiagen, Valencia, CA, USA), according to the manufacturer's recommendations. For purposes of comparison, all cpDNA concentrations presented in this manuscript are expressed as ng/ml of plasma.
The OncoCarta panel (v1.0) consists of 24 pools of primer pairs and extension primers, and has the capacity to detect 238 mutations in 19 genes. The protocol provided by Sequenom (San Diego, CA) was followed with minor modifications. The amount of DNA added to the polymerase chain reaction (PCR) was 20 ng per reaction for FFPE DNA samples. For plasma DNA samples, 30 µl of DNA were added to 30 µl of pure water, and used for the OncoCarta panel (v1.0) processing. DNA was amplified using the OncoCarta PCR primer pools, unincorporated nucleotides were inactivated by shrimp alkaline phosphatase (SAP), and a single base extension reaction was performed using extension primers that hybridize immediately adjacent to the mutations and a custom mixture of nucleotides. Salts were removed by the addition of a cation exchange resin. Multiplexed reactions were spotted onto SpectroCHIP II arrays, and DNA fragments were resolved by MALDI-TOF on the Compact Mass Spectrometer (Sequenom, San Diego, CA).
Data analysis was performed using MassArray Typer Analyzer software 4.0.4.20 (Sequenom), which facilitates visualization of data patterns and the raw spectra. Typer automates the identification of mutants by comparing ratios of the wild type peak to that of all suspected mutants and generates an OncoMutation report detailing specific mutations and the ratios of wildtype and mutation peaks. All mutations from the Oncomutation report were reviewed manually by 2 blinded operators, with selected reviewed mutations from the OncoMutation report compared and confirmed to be concordant. Manual review of mutations on all OncoCarta spectra was performed to identify “real” mutant peaks from salt peaks or other background peaks. Statistical analyses are detailed in the Supplemental Methods S1.
A total of 105 patients referred for phase I trial participation were enrolled between September 2009 and August 2010 (
Parameter | No. of patients |
|
|
Male | 45 (43.3%) |
Female | 59 (56.7%) |
|
56 (range 22–75) |
|
|
Colorectal cancer | 25 (24.0%) |
Breast cancer | 19 (18.3%) |
Melanoma | 15 (14.4%) |
Castration resistant prostate cancer | 11 (10.6%) |
Ovarian cancer | 15 (14.4%) |
Other |
19 (18.3%) |
|
|
|
36 (34.6%) |
|
62 (59.6%) |
|
6 (5.8%) |
Min | Median | Max | Mean | sd | |
|
0.3 | 5.8 | 17.5 | 6.1 | 3.7 |
|
|
|
|
|
|
No. of patients | 4 | 30 | 41 | 19 | 10 |
|
23 | 34 | 43 | 34.4 | 4.2 |
|
100 | 202.5 | 3531 | 300.3 | 370.2 |
|
0.5 | 17.3 | 1600 | 55.4 | 196.1 |
One patient was subsequently found to be ineligible for this study as he had not exhausted all lines of available antitumor treatments.
Includes non-small cell lung cancer (NSCLC), mesothelioma, sarcoma, glioblastoma, adenocarcinoma of unknown primary (ACUP), cholangiocarcinoma, and cervical, endometrial, duodenal, esophageal, pancreatic and renal cancers.
cpDNA was collected from 101 (97%) patients; it was not possible to draw blood from 1 patient for technical reasons and blood was not collected from 2 patients due to logistical errors.
Of the 104 patients analyzed in the study, FFPE primary tumor samples were obtained for 69 (66%) subjects, with FFPE nodal and/or metastatic tumor samples being available for a further 31 (30%) patients. cpDNA was collected from 101 (97%) patients; it was not possible to draw blood from 1 patient for technical reasons and blood was not collected from 2 patients due to logistical errors. A total of 60 patients died during follow up, while data for 44 patients were censored for purposes of this publication. The median follow up time was 5.8 months (range 0.3–17.5) (
Dilutions of DNA extracted from the
Parameters | n (%) |
|
|
Male | 7 (35%) |
Female | 13 (65%) |
|
34 (range 25–52) |
n | min | median | max | mean | sd | |
|
20 | 4.5 | 6.4 | 13.3 | 7.4 | 2.9 |
The overall median cpDNA concentration was 17 ng/ml in these patients with advanced tumors (range: 0.5–1600) (
Box and whisker plots showing 25th, 50th and 75th percentiles, upper and lower adjacent values (whiskers) and Tukey outliers (•). P value is for a two-sided unpaired t-test on log10 DNA concentrations using Welch's correction for unequal variances.
Matched plasma and FFPE were available for analysis from 84 patients. A total of 42 mutations were detected in either or both FFPE tumor and cpDNA specimens obtained from these patients (
|
|
|
|
|
|
|
|
|
|
3/3 (100%) | 7/10 (70%) | - | - | - | 1/1 (100%) | 1/3 (33.3%) | - |
|
3/5 (60%) | - | 2/3 (66.7%) | - | 1/1 (100%) | - | - | - |
|
- | - | - | - | - | 1/1 (100%) | 3/4 (75%) | - |
|
- | - | 0/1 (0%) | 0/1 (0%) | - | 1/1 (100%) | 1/1 (100%) | - |
|
- | 0/2 (0%) | - | - | - | - | 0/1 (0%) | 0/1 (0%) |
|
- | - | 1/1 (100%) | - | - | - | - | - |
|
- | - | 0/1 (0%) | - | - | - | - | - |
|
- | 0/1 (0%) | - | - | - | - | - | - |
|
6/8 (75%) | 7/13 (54%) | 3/6 (50%) | 0/1 (0%) | 1/1 (100%) | 3/3 (100%) | 5/9 (55.6%) | 0/1 (0%) |
The median overall survival (OS) for all patients was 7.9 months (95% CI 5.8, 9.2). Patients were categorised into low and high cpDNA concentration groups based on the maximum healthy volunteer cohort DNA concentration of 13.3 ng/ml; 61 patients were classified as having high cpDNA concentrations with 40 having low levels. The median OS in patients categorised as having low cpDNA concentrations was 10.5 months (95% CI 6.0, NC), while those in the high cpDNA concentration group had a median OS of 6.5 months (95% CI 4.5, 8.4) (logrank p = 0.0383) (
(
We have recently prospectively validated a prognostic score (RMH score) for patients participating in Phase I clinical trials based on the combination of three prognostic factors: serum albumin less than 35 g/L; lactate dehydrogenase (LDH) greater than the upper limit of normal (ULN); and two or more sites of metastases. The presence of each of these variables associated with worsening outcome
Scatterplot showing the relationship between cpDNA concentration and RMH prognostic score. There was a significant positive linear trend between log10(cpDNA) and RMH score (beta = 0.252, p<0.0001).
Univariate testing was used to determine significant predictors of overall survival, which included cpDNA concentration as a continuous variable (HR 2.4 per 10-fold increase, 95% CI 1.4–4.2), albumin <35 g/L (logrank p = 0.0003), and ECOG performance status equal to 2 (logrank p = 0.0007). When cpDNA, albumin and performance status were incorporated into a multivariate model, all three parameters were found to be independent predictors of survival (
Univariate logrank | Multivariate Cox regressionn = 101 | |||
Variable | p | HR | 95% CI | p |
|
(Cox regression) 2.43 (1.39–4.25) p = 0.002 | 1.98 | 1.01–3.88 | 0.045 |
|
0.0003 | 1.86 | 1.01–3.42 | 0.047 |
|
0.0007 | 8.05 | 2.53–25.65 | <0.0005 |
Of 25 patients with CRC, cpDNA samples were obtained from all patients, while FFPE tumor samples were available for analysis for 22 patients. Overall, mutations were detected in 15 of 22 (68.2%) available FFPE tumors and 14 of 25 (56%) cpDNA specimens (
Concordance in the detection of mutations between matched FFPE archival tumors and cpDNA specimens by Sequenom OncoCarta analyses was 70% (7 of 10 patients) for
There was 90% (9 of 10
Of the 15 patients with melanoma, FFPE tumor samples were available for analysis for 10 patients, while cpDNA samples were obtained from all 15 patients. Overall, mutations were detected in 8 of 10 (80.0%) available FFPE tumors and 6 of 15 (40%) cpDNA specimens (
There was 100% concordance (5 of 5 samples) for the
FFPE tumor samples and cpDNA samples were available for analysis for all 19 patients with breast cancer. Overall, mutations were detected in 5 of 19 (26.3%) FFPE tumors and 4 of 19 (21.1%) cpDNA specimens (
The
Of the 11 patients with CRPC, cpDNA samples were obtained from all patients, while FFPE tumors were available for 8 patients. Overall, mutations were detected in 3 of 8 (37.5%) FFPE tumors, and 3 of 11 (27.3%) cpDNA specimens (
Of the 15 patients with advanced ovarian cancer, cpDNA samples were obtained from all patients, while FFPE tumor samples were available for 14 patients. Overall, mutations were detected in 5 of 14 (35.7%) FFPE tumors, and 0 of 14 (0%) cpDNA specimens (
Of the remaining 19 patients with a range of tumor types, cpDNA samples were obtained from 17 patients, while FFPE tumors were available for 12 patients.
The
When considering all patients with matched samples, including those with no mutations detected (n = 83), the concordance in detecting mutations between FFPE and cpDNA was higher in metastases (83.3% of 18 specimens) compared with primary tumor (78.5% of 65 specimens). When considering only patients with mutations detected in at least blood and/or primary tumor (n = 40), the concordance in detecting mutations between FFPE and cpDNA was again higher in metastases (70.0% of 10 specimens) compared with primary tumor (53.3% of 30 specimens). However, because of the difference in the number of primary tumor (n = 65) and metastatic (n = 18) specimens obtained, we are unable to draw any statistical conclusions from these data.
This study has demonstrated, for the first time, the feasibility of multiplex detection of tumor DNA mutations utilizing the multiplex OncoCarta panel from both DNA extracted from FFPE archival tumor tissue and cpDNA. We have shown that total cpDNA levels in patients with advanced cancers are, in general, significantly higher than those in healthy volunteers, with the highest concentrations found in patients with advanced prostate and breast cancers, although this difference was not significant in melanoma and ovarian cancer (
The Sequenom OncoCarta panel has also enabled us to analyze more than 230 known mutation ‘hot-spots’ mutations in over a hundred patients in a high throughput fashion. The OncoCarta panel covers a large and increasing number of oncogenes and can be adapted to include additional genes of interest. It allows tumor mutation detection even with minimal amounts of tumor DNA, poor tissue preservation and the presence of significant amounts of normal DNA. Next generation sequencing technology will allow more DNA coverage and data acquisition, allowing the sequencing of hundreds of full length genes, which will be critical to the study of genes where mutations can be found in multiple disparate locations, as is the case for many tumor suppressor genes such as BRCA1, BRCA2, p53 and PTEN.
As we move towards the development of molecularly targeted agents for selected populations of patients, it is crucial that the molecular characterization of tumors for the prediction of efficacy to targeted therapies is incorporated into early clinical trials
The origin of DNA in the plasma of patients with cancer remains undefined, but may be derived from tumor cell fragments, microparticles or exosomes, or indeed from circulating free DNA
The discordance in results observed between archival tumors and cpDNA specimens could be due to several reasons: 1) Poor mutation detection in cpDNA due to low cpDNA concentrations; 2) Poor DNA quality in the FFPE archival sample; 3) potential false positive results in either sample type; 4) true disease heterogeneity. Overall, however, there was high concordance rate for FFPE
In conclusion, we envision that biomarker studies such as that described above can have a potential impact on molecular stratification and patient care. We therefore recommend a concerted effort by the cancer community in order to develop analytically validated assays on cpDNA that can be clinically qualified for more broad utilization
Complete data set including list of FFPE tumor and cpDNA mutations, cpDNA concentrations and tumor characteristics.
(XLS)
(JPEG)
(TIFF)
(DOC)