NEW YORK (360Dx) – Five-year-old startup Selux Diagnostics intends to parlay the pairing of a novel reagent and microplate reader detection into low-cost, high-throughput antimicrobial susceptibility testing (AST) and minimum inhibitory concentration (MIC) determination.
With a recent scientific publication under its belt, the company hopes to compete in the space of rapid automated phenotypic AST. SeluxDx expects to begin clinical trials of its technology in the second half of 2019 to support 510(k) submission to the US Food and Drug Administration.
CTO Eric Stern cofounded Selux in 2014 in a lab in his attic. In the early days, the company was funded by angel investors, but it has since grown to 28 employees and was recently awarded milestone-based funding from the Biomedical Advanced Research and Development Authority (BARDA) potentially worth $45 million. It also raised $25 million in a private Series B round, Stern said in a recent interview.
Boston-based Selux described its technology for single-day AST in a proof-of-concept study published in Scientific Reports last month.
Overall, the Selux approach is to make it quick and easy for microbiology labs to ascertain antimicrobial susceptibility and MIC.
"We feel very strongly that, in order to fit into the laboratory's paradigm for cost, throughput, menu, and speed, the right way to do this is with a test that essentially uses a bulk measurement of the bacteria, rather than looking at the bugs on an individual basis," Stern said.
For rapid clinical diagnosis, "what the labs are looking for is to have a system that is capable of delivering rapid results, but also high throughput," Stern said. The Selux system enables random access loading and will have a throughput of 100 samples per shift, he said.
Furthermore, given that AST is essentially "a companion diagnostic for antibiotics," Stern said, the number of antibiotics that can be tested easily is important, as is having a minimum cost of goods, because infectious disease testing in general is very price sensitive. The menu of drugs on the Selux consumables is five times greater than any competitor, Stern said, and will also include newly approved antimicrobials.
The system will also be unique in that it will be the only platform capable of providing AST results from positive blood cultures as well as isolated bacterial colonies, with results in about six hours for the majority of samples. And, although a separate ID step would be required before results are released, labs could perform the AST and get susceptibility information in parallel.
The Selux test uses consumables that are compatible with standard fluorescent microplate readers. The system will ultimately have two types of microplate consumables, Stern said, one for all gram-positives and one for all gram-negatives. These will be stable at room temperature, making it easier for labs to store them.
In the study, the firm describes building the new assay for determining bacterial surface area and showing that it could accurately predict antibiotic susceptibilities at a four-hour time point, Stern said. The study also demonstrates that a small-molecule fluorophore can be used to distinguish growth modes that are not distinguishable by volumetric assays, he said.
As outlined in the Scientific Reports study, the Selux scientists began with the hypothesis that measuring total bacterial surface area during antibiotic exposure could distinguish resistant cells that that are happily growing and dividing in the presence of antibiotics from cells in their death throes. Importantly, the latter might be in growth modes where they swell or lengthen, but are not actually replicating. They also reasoned that they could garner information on minimum inhibitory concentration, or MIC, that is needed for patient treatment, using this method.
The first reagent they used was a simple IgG antibody that bound most bacteria, which was then detected with horseradish peroxidase. However, some bacterial species bound the antibody better than others, so they next moved on to a more novel approach.
Bacterial cells have a surface charge — a fact that firms like Qvella and QuantuMDx are also exploiting in their technologies, albeit for different ends — and that means that certain chemicals will electrostatically bind bacterial membranes.
SeluxDx chose a chemical called a europium-cryptate-diamine chelate as their reagent to universally paint the outside of all bacteria. In general, cryptates and cryptands form 3D structures that act like tiny cages which can interact with ions. Cryptates incorporating rare earth elements, like europium, are also fluorescent. The study notes that the molecule Selux is using can yield signal amplification comparable that from the HRP/IgG scenario, in which multiple HRP molecules typically bind a single IgG and the HRP enzyme in turn catalyzes many chromogenic substrates, boosting the signal.
A group of researchers in France were among the initial developers of cryptates for use in diagnostics, particularly with a method called time-resolved amplification of cryptate emission, or TRACE. These researchers are now part of Cisbio, a firm that custom develops reagents and that supplied the Eu-cryptate-diamine used by Selux in the Scientific Reports study. The Cisbio technology also uses the rare earth cryptate technology in an application called homogeneous time resolved fluorescence, or HTRF.
In the Selux study, the cryptate reagent was shown to be effective in detecting resistance and determining MIC for a variety of bacteria and antibiotics, including ones that are known to have unique behaviors when exposed to antibiotics. For example, before the cell death phase — which typically begins hours after antibiotic exposure, limiting same-shift testing with death as the endpoint measure — cells sensitive to a particular antibiotic might swell, elongate, form filaments, or otherwise contort when interacting with particular antibiotics. Bacteria that are resistant, however, would continue to divide, and so over time there would be more cells, and therefore more cellular surface area to bind the cryptate. The Selux test was shown to be broadly reliable, even across these different states, and was also able to accurately predict MICs.
Competing in AST
Automated or semi-automated systems for bacterial identification and antimicrobial testing include the Accelerate Diagnostics Pheno, BioMérieux Vitek 2, Becton Dickinson Phoenix, Thermo Fisher Scientific Sensititre, and Beckman Coulter Microscan.
The Accelerate Pheno, in particular, was cleared in 2017 with a sepsis test. It was touted as the first system that could perform ID and rapid phenotypic AST in one instrument directly from positive blood cultures, with ID-AST results in about eight hours.
Accelerate's phenotypic AST is accomplished through imaging to capture contortions of single bacterial cells over time. The single-cell measurements can detect subtle changes in bacterial morphology that indicate a sensitivity to antibiotic, but also tend to limit the overall throughput of the system. The Pheno also tests fewer drugs than the proposed Selux system and its consumables require refrigeration while Selux's do not, Stern said.
Systems like Vitek 2 and Phoenix, on the other hand, can run higher throughputs than the Accelerate Pheno, but not directly from blood culture bottles. Testing in as little as five hours is reported with Vitek 2, but that is after a one-day culture step, Stern noted, and the average reporting time published in the literature is more on the order of eight to 12 hours.
"Since laboratories staff according to eight to 10 hour shifts, and expert personnel are required both to set up AST tests and to interpret the data before they are sent to the clinic, this means the AST results cannot be reported the same day the test is started" in most labs using this system, Stern said, so it is typically run overnight, as is the Beckman Coulter MicroScan.
In terms of footprint in the lab, the Selux system size is also "roughly equivalent to that of a MicroScan," although Selux's is a floor-standing unit while Beckman Coulter's is a benchtop unit, Stern said.
Overall, Stern said Selux's platform "will roughly match the throughputs and random-access loading capabilities" of other lab systems, which can run 100 tests per shift, but it will offer a menu up to five times greater than any other current system, and will have "a highly competitive cost," requiring only room-temperature consumable storage.
However, the aforementioned competing instruments also do ID testing, and Selux does not. Stern does not see this as a major obstacle.
"The two most significant innovations of the past decade in clinical micro have been new ID platforms" he noted. Multiplex ID panels from BioFire and Verigene, for example, have enabled ID directly from positive blood samples, Stern said, and mass spectrometry "has become the new gold standard for ID for all other samples."
In general, multiplex genetic ID platforms can provide results at least two days before standard AST is finished, and mass spec IDs results are often available one day before AST results in standard laboratory workflows, Stern said. Given many labs are now adopting these ID technologies, the Selux test will thus be able to provide them AST results on the same day as their ID results if the tests are run in parallel.
"What we're trying to do is help hospitals with de-escalation and provide a system for laboratories that is basically one-stop shopping — you get the Selux system and you have best-in-class speed and performance for every different sample type that you have," Stern said.