NEW YORK — Researchers at the Dana-Farber Cancer Institute have developed a PCR technique called "inter-Alu-PCR" that, in combination with targeted sequencing and a custom algorithm, they believe can detect microsatellite instability (MSI) in colorectal cancer patients' blood samples.
The group initially aims to use the PCR-based method to monitor for minimal residual disease (MRD) in late-stage cancer patients after surgery.
Oncologists typically identify MSI through the presence of one to five National Cancer Institute-designated biomarkers in a patient's tumor tissue. If one of the biomarkers in the patient's tumor DNA has been mutated, the tumor is classified as "MSI-Low" (MSI-L). If two or more biomarkers have been mutated, then the tumor is termed "MSI-High" (MSI-H).
"Those standard markers have been very good and reliable to detect colorectal cancer, if you have the MSI-H phenotype," Mike Makrigiorgos, professor of radiation oncology at the DFCI and Harvard Medical School, explained. "Groups have done multiplex PCR and capillary electrophoresis, which works very nicely on tumors as long as MSI is not at lower levels."
His team therefore wanted to develop a method that combined the speed of PCR and breadth of next-generation sequencing (NGS) to rapidly identify indels that are typical for MSI-H tumors in other cancer types besides colon cancer.
"With NGS, you can survey thousands of microsatellites at the same time and obtain a global view of the indels in the genome," Makrigiorgos said. "NGS allows for high sensitivity, as you can score many targets at the same time by [making] the targets bigger, which help inform you [as] you look at a few nanograms of DNA."
The group's new method, inter-Alu-PCR, begins by extracting DNA from a patient's blood sample using modified primers that capture Alu poly-A tails with amplicons generated from adjacent Alu elements. Makrigiorgos explained that the method uses Alu elements because they have been linked to several inherited human diseases and various cancer types.
Sequencing libraries are then prepared using the amplified products via index PCR, followed by targeted NGS sequencing. The data is then subjected to an algorithm called "MSI-Tracer" that identifies MSI-caused indels at low tumor purities in the sample.
"We estimate that you'd need a finger prick of blood, which would provide at least 100 picograms of circulating DNA for inter-Alu-PCR," Makrigiorgos said. "You only need to survey about 1,000 Alu elements, containing thousands of microsatellite [locations], to identify the MSI-H phenotype if it's in the blood sample."
After sample collection, inter-Alu-PCR requires about two to three hours before overnight sequencing on Illumina's MiSeq instrument to establish a patient's MSI-status, Makrigiorgos said.
In a proof-of-concept study published last month in Nucleic Acids Research, Makrigiorgos and his colleagues demonstrated Inter-Alu-PCR's feasibility on cancer tissue and liquid biopsy samples.
Makrigiorgos' team first analyzed the MSI status of inter-Alu-PCR-captured microsatellites by evaluating tissue biopsy-derived DNA from colon cancer patients and matched normal samples. After characterizing the patients' MSI status, the researchers sequenced the inter-Alu-PCR products and compared the data using MSIsensor and MSI-Tracer algorithms.
Overall, the data indicated that combining inter-Alu-PCR with NGS could accurately identify MSI status using ultra-low-pass sequencing without the need for matched normal tissues.
The group then used inter-Alu-PCR on samples with pre-defined MSI status in colon adenocarcinoma, corpus endometrial carcinoma, and stomach adenocarcinoma. By extracting inter-Alu regions from whole-genome sequencing data, the team spotted cases of clustering in all MSI-H and microsatellite stable (MSS) patients.
To establish the method's limit of detection in tissue samples, the researchers tested multiple scenarios using serial dilutions of MSI-H tumor DNA from colon cancer into matched normal DNA. His team used droplet digital PCR to validate the dilution approach using tumor-specific somatic mutations such as KRAS for a subset of the mutations.
When a paired normal sample was not present, the researchers found that that the method had a limit of detection of 0.15 to 0.5 percent for somatic indels using a low-tumor purity clinical sample. When matched normal tissue was available, inter-Alu-PCR had a somatic limit of detection between 0.05 to 0.5 percent.
Makrigiorgos' team also showed how inter-Alu-PCR could be potentially used to detect MSI-related poly-adenine deletions in cell-free DNA (cfDNA) from a patient's blood sample. Analyzing cfDNA from colon cancer patients and healthy samples, the researchers saw that MSI-H patients produced a higher MSI-Tracer score compared to MSS or normal samples.
Overall, the study authors found that inter-Alu-PCR could classify MSI using as low as 0.1 ng of input DNA from a patient's blood sample.
Makrigiorgos now plans to improve inter-Alu-PCR by modifying the method to capture higher numbers of Alu-associated microsatellites in colon cancer tissue and liquid biopsy samples, which he believes will increase its sensitivity and help detect MRD. However, he noted that his team will also need to conduct larger scale studies to define the degree of MSI in other solid tumor types.
Several firms in the cancer space plan to develop or currently offer their own tissue and liquid biopsy-based tests to establish a patient's MSI status. Notably, Promega offers its OncoMate MSI Dx Analysis System, which applies a pentaplex PCR system, targeting five mononucleotide-repeat markers and two pentanucleotide markers to measure MSI status in solid tumors.
Makrigiorgos acknowledged that targeted re-sequencing and exome sequencing methods are being employed to diagnose MSI in different tissue types. However, he argued that one of inter-Alu-PCR's major advantages for tracing MRD in blood samples is that users do not need to perform hybrid capture to identify mutations of interest, minimizing the cost of sample prep.
He also highlighted that inter-Alu-PCR does not require as much sequencing as other MSI-detection methods to spot the MSI-positive phenotype.
"You already have your candidate agent with inter-Alu-PCR, where you can extract the mononucleotides and examine them," he said. "Without having to know the deletions in the tumor, you can use the MSI-H phenotype to trace MRD, as the indels tend to be there if the tumor is MSI-H."
Makrigiorgos believes that inter-Alu-PCR could also eventually be used for early cancer detection in high-risk groups. He noted that patients with Lynch syndrome — a genetic condition associated with a high likelihood of developing colon, endometrial, stomach, and other cancers — could potentially have the test done as a preliminary step prior to more invasive and costly procedures. He also believes that Inter-Alu-PCR could be used to allow repeated testing for tumor load in MSI-positive patients on treatments such as chemotherapy, radiation, and brachytherapy.
"Patients diagnosed with Lynch syndrome that have mismatched repair deficiency often have to undergo repeated … expensive testing via colonoscopy," Makrigiorgos said. "If the inter-Alu-PCR test is positive, then you can go to colonoscopy."
However, he acknowledged that the inter-Alu-PCR and MSI-Tracer workflow will need to demonstrate high analytical sensitivity and specificity in these applications before the method is used in the clinical space.
Makrigiorgos and his colleagues have filed patents for inter-Alu-PCR, as well as "a subsequent sequencing method" that he said will obtain about "10 to 30 times" more microsatellites. While he believes inter-Alu-PCR is "practical enough" to be used in the clinical space, the group does not have any current plans to commercialize the method through a startup or an existing firm.