This article has been updated with quotes from the authors given at a press conference on Wednesday.
NEW YORK – Broad Institute researchers have developed a CRISPR-based molecular diagnostics platform using microfluidics chips for the detection of viruses in human samples. A single chip has the capacity to detect a single virus in more than 1,000 samples at a time, or to search a smaller number of samples for more than 160 different viruses — including SARS-CoV-2, the virus responsible for the COVID-19 pandemic.
In a study published on Wednesday in Nature, the researchers described their development of Combinatorial Arrayed Reactions for Multiplexed Evaluation of Nucleic acids (CARMEN), a platform for scalable, multiplexed pathogen detection. In the CARMEN platform, nanoliter droplets containing CRISPR-based nucleic acid detection reagents self-organize in a microwell array to pair with droplets of amplified samples, testing each sample against each CRISPR RNA (crRNA) in replicate, the researchers wrote.
They combined CARMEN with the Cas13 nuclease (CARMEN-Cas13) to test more than 4,500 crRNA-target pairs on a single array. Using CARMEN-Cas13, the team developed a multiplexed assay that simultaneously differentiated all 169 human-associated viruses with 10 or more published genome sequences and rapidly incorporated an additional crRNA to detect the causative agent of the 2020 COVID-19 pandemic. CARMEN-Cas13 further enabled comprehensive subtyping of influenza A strains and multiplexed identification of dozens of HIV drug-resistance mutations.
"CARMEN's intrinsic multiplexing and throughput capabilities make it practical to scale, as miniaturization decreases reagent cost per test [more than] three-hundredfold," the authors wrote. "Scalable, highly-multiplexed CRISPR-based nucleic acid detection shifts diagnostic and surveillance efforts from targeted testing of high-priority samples to comprehensive testing of large sample sets, greatly benefiting patients and public health."
The microfluidic technology that the researchers used to build CARMEN was first developed in 2018 in the lab of Paul Blainey, a Broad core member and MIT associate professor, and co-senior author of the new study. The researchers created rubber chips, slightly larger than a smartphone, with tens of thousands of microwells each designed to hold a pair of nanoliter-sized droplets, according to the Broad. In each droplet pair, one droplet contains viral genetic material from a sample, and the other contains reagents meant to detect a virus.
"The microwell chips are made like a stamp — it's rubber poured over a mold. We're easily able to replicate and share this technology with collaborators," co-first author and Broad postdoctoral fellow Cheri Ackerman said in a statement.
"This miniaturized approach to diagnostics is resource-efficient and easy to implement," Blainey added.
The specific CRISPR platform the researchers used for viral detection is a version of the SHERLOCK (Specific High Sensitivity Enzymatic Reporter UnLOCKing) platform developed in the lab of the Broad's Feng Zhang. SHERLOCK was first described in a paper in Science in April 2017 by members of the Zhang lab, and the same team published a second paper in March 2018 describing new improvements on the technology, including a 3.5-fold increase in signal sensitivity through a combination of Cas13 with auxiliary CRISPR-associated enzyme Csm6.
In April 2018, researchers in the lab of Pardis Sabeti at the Broad published a paper in Science demonstrating that SHERLOCK could detect Zika virus (ZIKV) and dengue virus (DENV) in patient samples at concentrations down to 1 copy per μl. The team further showed that SHERLOCK could distinguish between the four DENV serotypes as well as region-specific strains of ZIKV from the 2015 to 2016 pandemic.
"The current pandemic has only underscored that rapid and sensitive tools are critical for diagnosing, surveilling, and characterizing an infection within a population. The need for innovative diagnostics that can be applied broadly in communities has never been more urgent," the Broad's Sabeti, a professor at Harvard University and co-senior author of the new Nature study, said in a statement.
"CRISPR-based diagnostics are an attractive tool for their programmability, sensitivity, and ease of use," added co-first author and Broad postdoctoral fellow Cameron Myhrvold. "Now, with a way to scale up these diagnostics, we can explore their potential for comprehensive approaches — for example, enabling clinicians to see if patients are harboring multiple infections, to rule out a whole panel of diseases very quickly, or to test a large population of patients for a serious infection."
This team's work is not the first instance of SHERLOCK being used as the basis for a SARS-CoV-2 diagnostic. In February, engineering biology firm Sherlock Biosciences told GenomeWeb that it had been developing and testing a series of assays for SARS-CoV-2, and that it was almost ready to release what would likely be the first CRISPR-based diagnostic test for the virus.
The company, which launched in 2019 with licenses to foundational CRISPR and synthetic biology technology from the Broad and Harvard's Office of Technology Development, said it planned to use engineering biology tools, including CRISPR and synthetic biology, to develop a new generation of molecular diagnostics that could rapidly deliver results for a wide range of needs at low cost.
Sherlock Bio spent some time developing a series of assays for SARS-CoV-2 based on SHERLOCK, and then spent some time looking for a collaborative partner and the appropriate platform on which to release the test. It eventually partnered with Cepheid to use SHERLOCK to design tests to run on Cepheid's GeneXpert systems.
For the new study, the Broad researchers aimed to combine the advantages of several pathogen detection methods, while eliminating their problems. Sequencing or microarray hybridization can provide significant information about pathogen genotypes and evolution, but they're difficult to implement on a broad scale because of cost considerations and logistical demands of sample preparation. On the other hand, rapid, low-cost detection methods, such as CRISPR-based approaches, antigen-based tests, PCR, or LAMP can detect only one or a handful of pathogens in a given reaction. So, they turned to miniaturized and self-organizing microfluidic technology, which enables massive multiplexing of biochemical and cellular assays.
In their analyses, the researchers found that CARMEN-Cas13 was sensitive, specific, and statistically robust. The platform detected Zika sequences with attomolar sensitivity, matching the sensitivity of SHERLOCK and PCR-based assays. Further, as the COVID-19 outbreak emerged during the study's manuscript review process, the investigators rapidly incorporated a new test for SARS-CoV-2 into a coronavirus panel for the CARMEN platform, demonstrating the power of this modular master set to be adapted to real-world challenges. Using a single mChip, more than 400 samples can be tested in parallel against the coronavirus panel, the researchers said.
At a press conference following the release of the study on the Wednesday, Myhrvold said that this modularity and the ability to rapidly incorporate new assays is one of CARMEN's biggest advantages. "It's important to recognize that you get to ask a very specific set of questions by designing an assay for CARMEN, and that's the difference from sequencing," he said. Although sequencing provides a significant amount of information, the sheer amount of data can be overwhelming and challenging to work through. CARMEN, however, provides specific information without extraneous data.
To test CARMEN in a more challenging context, the researchers also evaluated the panel of 169 human-associated viruses against 58 plasma, serum, and throat or nasal swab samples from patients with a variety of confirmed infections. Each clinical sample was treated as an unknown and compared to next-generation sequencing, which was performed with more than 2 million reads per sample. Of the 11,268 tests that were interpretable by both methods, 11,236 (99.7 percent) were concordant.
"We found that CARMEN identified the known infection in the majority of samples where NGS detected any sequences from these viruses, including complete concordance between CARMEN and NGS for dengue and Zika tests," the authors wrote. "CARMEN and NGS can also be compared based on their ability to detect the sequence locus targeted by the CARMEN crRNA, revealing that CARMEN is more sensitive than NGS on a per-sequence locus basis among the crRNA targets tested. CARMEN's overall sensitivity of detection, especially for diverse viruses, can be increased by the addition of crRNAs to cover additional loci and/or loci with sequence diversity."
In the future, the team concluded, it's possible that region- and outbreak-specific detection panels could be deployed to test thousands of samples from selected populations, including animal vectors, animal reservoirs, or symptomatic patients.
During the press conference, Blainey said the team is working to deploy CARMEN, both in the context of the COVID-19 pandemic and more generally as a viral diagnostic platform, though it isn't yet ready to share details about possible regulatory or commercialization strategies.
"We're motivated by this pandemic, and we have to move much faster because of the circumstances," Sabeti added. She noted that the technology is currently being used as a research tool, and that the team is seeking Emergency Use Authorization from the US Food and Drug Administration for CARMEN. "We hope to share news soon on that front," she said.
Meanwhile, the researchers are still considering what kinds of refinements would be needed to make the technology deployable in different contexts. Right now, Ackerman said, CARMEN still requires the use of a microscope to read the fluorescent signal from the chip that indicates which virus a patient has. It also requires the preparation of the nanoliter droplets and nucleic acid extraction. This makes CARMEN hard to use in the field — in an airport, for example — or in a doctor's office at the moment.
But according to Blainey, the platform's biggest advantage is its flexibility. "Having that flexibility and adaptability is going to be a huge asset as we think about deploying this technology for all sorts of use cases," he said.