Identifying scientific markers of cancer is key to early disease detection and treatment planning – often possible long before a diagnosis has been made or symptoms develop. It can also help doctors to track disease progression and guide treatment in established cases.

However, searching healthy cells for small changes to DNA in cancer cells is akin to searching for needles in haystacks. That’s because while some changes are significant – and relatively easy to detect – others are so small they evade analysis. And it’s these changes that can give cancer DNA an edge and affect patient outcomes – early detection can boost survival rates by as much as 90% in some cancers. In what could signal a major breakthrough for oncology research, a team from the Department of Biomedical Engineering at Johns Hopkins University has developed an innovative device to help make identifying these small changes easier and more efficient.

HYPER-Melt – or high-density profiling and enumeration by melt – is a digital, microfluidics-based platform that allows researchers to manipulate and analyse minute volumes of fluid. By separating blood samples into smaller and smaller chunks, the device makes them easier to individualise and analyse. This makes it easier to identify disease-carrying DNA and to separate it from healthy DNA.

HYPER-Melt is the brainchild of Chrissy O’Keefe, a PhD student who has been under the mentorship of Professor Jeff Wang, a microfluidics expert in the university’s Department of Mechanical Engineering and at the Institute for NanoBioTechnology (INBT), and senior INBT research scientist Tom Pisanic.

O’Keefe and her team have published their work on HYPER-Met in a Science Advances paper entitled ‘Facile profiling of molecular heterogeneity by microfluidic digital melt’. The research is the culmination of more than five years’ work in the field.

O’Keefe says the device, which is based on hardware and software components, evolved from an interest in studying the heterogeneity of a sample on a moleculeby- molecule level.

“Cancer, along with many other diseases, is highly adaptable and contains many heterogeneous cell types,” she says. “It is constantly changing and evolving to ensure it can survive and thrive in its environment – in the body. We wanted to develop a technology that could detect and quantify this variability so we could better understand tumour evolution. We wanted to look at this at its earliest stages, as well as at later stages where the cancer adapts and avoids treatment or causes a relapse. In particular, we were interested in studying cases where molecules may be very rare – such as cell-free DNA from a blood sample.”

Collecting signatures

HYPER-Melt works by investigating small sequence variations of tens to thousands of single molecules of DNA. It digitises molecules into 4,096nL chambers and uses a thermal-optical platform to simultaneously amplify and obtain the high-resolution melt signatures of each chamber.

In practice, the workflow is similar to a standard bench top PCR experiment. The user prepares their mastermix and target on the bench top, and loads the mixture into a syringe. When the syringe punctures the seal of the chip, the mixture enters the device by vacuum-assisted loading, and a partitioning fluid is pressurised through the channels to digitise each chamber. This process is fast and efficient, taking no longer than five minutes.

The final step involves the device being placed on the flatbed heater to undergo digital PCR and highresolution melt (dPCR-HRM) using the same time frame as a microtiter plate experiment.

O’Keefe says there are a number of reasons why the technology could revolutionise cancer detection. “HYPER-Melt allows highly parallelised molecular profiling of rare molecules,” she says. “The platform uses a simple, all-in-one workflow and demonstrates highly sensitive detection that is roughly two to three orders of magnitude higher than current techniques.

“For example, digital analysis in a microtiter-plate system is limited to tens of molecules, which limits the amount of information that can be collected. Microfluidic manipulation, on the other hand, enables the acquisition of more information per sample from thousands of individual molecules in parallel, while minimising reagent usage. In addition, the ability to assess various sequence changes of a given locus provides more in-depth information than current PCR-based gold standards.”

She says HYPER-Melt could be used to study the genetic and epigenetic make-up of a population and pave the way for an increased understanding of disease development and progression.

“HYPER-Melt is readily extendible to applications like precision medicine. The ability to profile disease populations within a patient can help identify genetic or epigenetic therapeutic targets, as well as small clonal populations that may be resistant, which could help in developing and adjusting treatment regimens.”

The device is best suited to samples with rare and heterogeneous molecules. Liquid biopsies, which contain trace amounts of cell-free DNA in the plasma, could be a key focus of HYPER-Melt-based research.

“Many diseases, such as cancer, are highly variable so that they can adapt and grow in their environment,” says O’Keefe. “HYPER-Melt can be used to assess this variability throughout the development of disease by simply taking a routine blood sample.”

In particular, it could benefit doctors working to detect cancers that ordinarily require invasive procedures like biopsies, endoscopies and colonoscopies to secure accurate diagnoses.

She said the device could also be used to monitor cancer evolution in response to therapy. “The ability to detect small, rare sequence changes could give an early indication of evolving therapeutic resistance.”

The impact of high-resolution melt technology

Similar current devices are based on digital PCR platforms, with the most popular version being droplet digital PCR (ddPCR), which can isolate and detect single molecules. It’s a sensitive technique that can find rare biomarkers; however, it is limited to a few known or targeted sequences.

HYPER-Melt, on the other hand, offers greater insights through its ability to assess unknown sequences, and identify a spectrum of modifications using high-resolution melt.

“Typically, digital PCR produces a binary yes/no result for each chamber. However, high-resolution melt allows us to acquire additional sequence information about each amplicon,” says O’Keefe. “High-resolution melt measures sequence variations by using an intercalating dye to detect changes in fluorescence during a thermal ramp. For example, in this study we looked at methylation changes – an epigenetic modification that affects gene expression. Highresolution melt allows us to assess the whole range of potential methylation changes, even those relating to unknown targets. One of the largest technical challenges in developing this platform was detecting real-time variations in fluorescence from thousands of nanolitre-volumes in parallel. Once acquired, though, we can obtain the sample’s entire sequence profile.”

Hyper-MELT could offer clinicians significant cost and time savings compared with other current technologies, such as sequencing, says O’Keefe.

“Sequencing, which is commonly used for genetic profiling, can provide exact sequence information,” says O’Keefe. “However, it is an expensive technology and it also takes a long time to achieve the requisite sensitivity (<0.1%) for challenging samples with rare molecules, so it is not particularly suited to clinical use.

“By comparison, HYPER-Melt can achieve much higher sensitivity – 0.00005% – with much lower cost; in fact, for roughly the same cost as a microtiter plate. In addition, it has a simple, rapid workflow with few hands-on steps, and provides an easy-to-interpret result, making it an efficient and relatively simple means for molecular profiling.”

“Cancer is constantly changing and evolving to ensure it can survive and thrive in its environment – in the body. We wanted to develop a technology that could detect and quantify this variability so we could better understand tumour evolution.”

O’Keefe said HYPER-Melt was aimed at achieving the outcomes of digital PCR and sequencing, and then building on these outcomes in terms of developing better sensitivity, efficiency and reliability.

O’Keefe says the pace of development in microfluidics is presenting exciting opportunities for the scientific community. The improved performance and reliability of commercial microfluidic technologies, such as ddPCR, have sparked a boom in research.

“One simple advantage of digital techniques like microfluidics is absolute quantification,” she says. “Researchers can simply count positive chambers to exactly quantify their sample without using standards, and this enables more accurate measurements for applications such as copy number variation.

“Another powerful advantage is the ability to detect extremely rare molecules. As a result, we are seeing more and more digital microfluidics being used in cancer studies to detect single nucleotide polymorphisms and rare mutations, especially in liquid biopsies.”

Bringing the device to market

Each of the prototype devices consists of a single polydimethylsiloxane (PDMS) pattern layer sandwiched between glass, and is made using an ultra-thin soft lithography technique. PDMS has several advantages, such as ease of fabrication, biocompatibility and optical properties, says O’Keefe.

“In addition, the gas permeability of PDMS allows rapid, vacuum-assisted loading. Glass, which can be easily bonded to PDMS, serves to provide a flat, rigid surface for heating while maintaining optical transparency.”

Having a beta-ready version of HYPER-Melt available within three years is a key target.

“There are still many improvements that we need to make before a packageable version of HYPER-Melt is ready for market,” O’Keefe concedes. “One major milestone would be to establish sufficient sensitivity and specificity of the platform in a clinical setting. If the platform demonstrates sufficient performance, then we can start to tackle scalable manufacturing methods and high-throughput assessment of samples.

“Currently, one of the device’s biggest drawbacks is the relatively low throughput of the system. HYPERMelt can analyse roughly five samples at a time, which limits its use in large studies that require the assessment of hundreds to thousands of samples. We’re therefore looking to improve the throughput for larger-scale studies by focusing on developing a more advanced thermal-optical platform.”

O’Keefe and her team are working with partners at Johns Hopkins University and other medical institutions to advance and refine the HYPER-Melt technology and usability to prepare it for market.

The team hopes to evolve the technology so it can be used as a diagnostic tool to detect tumour-derived DNA in liquid biopsies. “We believe that the high sensitivity of this platform, coupled with the multidimensional data set, can improve early detection of some types of cancer, especially those without reliable screening techniques,” she says.

Follow-up studies are focused in particular on lung cancer and ovarian cancer. Using lung cancer as an example, O’Keefe says the technique could be used following a CT scan to provide secondary, non-evasive analysis, potentially replacing the need to run a biopsy. In the case of ovarian cancer, which does not currently have a reliable diagnostic technique that can be used prior to symptom evolution, there is real potential to develop new biomarkers that could pave the way for early detection, diagnosis and treatment.