NEW YORK (GenomeWeb) – Belgian chromatography firm PharmaFluidics is looking to make inroads into markets including proteomics and clinical research with its microchip-based liquid chromatography technology.
PharmaFluidics' LC system, called µPAC, uses patterned microchips for sample separation upfront of instrumentation like mass spectrometry. According to Katrien Vanhonacker, the company's vice president of business development and sales, the system offers advantages over conventional LC technology in terms of separation power, peak sharpness, and robustness and reproducibility.
PharmaFluidics launched sales of the system in September of 2017 and this month closed a €7.3 million ($8.8 million) funding round. The company formed in 2010 as a spinout from the University of Brussels (VUB) and raised its first external funding in 2014, said Managing Director Johan Devenyns, calling this "when the real life of the company started." It announced a €2.7 million funding round at the beginning of 2017.
Conventional LC columns use tightly packed particles coated with different chemicals depending on the separation application. Sample is pumped through the particles under pressure with the different components of the sample separating according to their physical and chemical characteristics. These components can then be analyzed by mass spectrometry or another kind of instrument as they come off the LC column.
Extensive sample separation is especially necessary for applications like proteomics where researchers are typically looking at highly complex samples like plasma. However, Vanhonacker said, conventional LC systems have several limitations.
The separation power of a column is linked to how tightly packed it is, with more tightly packed particles offering a higher level of sample separation. However, Vanhonacker noted, the more tightly packed a column is, the higher the pressure required to push sample through it. Longer columns also provide more separation, but in both cases there are practical limitations on how long and high pressure a conventional LC separation can be.
Additionally, she said, the fact that conventional LC column particles are packed essentially at random can create variation in the path sample components travel through them.
"You can imagine if you have three identical peptides starting at the beginning of a column they might follow slightly different path lengths because of the random packing of the beads," she said. "So at the end you don't have a sharp peak. It is dispersed because of these different path lengths."
PharmaFluidics µPAC columns address these issues by using freestanding pillars etched onto a silicon wafer to separate samples. According to Vanhonacker, this allows the company to create columns with ordered and consistent separation structure, eliminating the randomness inherent in packed particle-based approached and thereby providing sharper LC peaks.
Additionally, because the pillars are freestanding, they are more permeable than a conventional column, which means they can be at lower pressure. The company is also able to make very long columns (2 m compared to around 300 mm for a long conventional column) by winding the path back and forth over the silicon wafer, which further improves separation power.
The controlled nature of production of the µPAC columns also improves their robustness and reproducibility, Vanhonacker said. Stability of chromatography over time is major concern for researchers who use LC. It is particularly a challenge for proteomics, where large-scale projects can take months to years.
"If you talk with people about LC and traditional columns, they are always [worried about reproducibility, and they are always afraid to change their columns in the middle of an experiment or project," she said. "That is because every column is different due to the fact that it is packed randomly. You never have exactly the same column."
With the µPAC columns the etched pillar system ensures "it is always the same pattern," Vanhonacker said.
The reproducibility of the columns potentially "opens up different perspectives" for proteomics research Devenyns suggested.
Projects can be more easily standardized "across different facilities and over time, which means that much more data from proteomic projects can be collected in a standardized way, which allows access to very large data series," he said.
Devenyns added that the company saw its technology having the most immediate advantage for proteomic experiments looking to go as deep as possible into a sample or experiment where sample size is very limited.
"I think [in those applications] we really have a core advantage, and then we will expand from there," he said.
LC performance has long been a key factor in proteomics research, but it has become even more so with advances in mass spec technology. A 2016 analysis by University of Wisconsin-Madison researcher Josh Coon and co-authors argued that LC performance is now the limiting factor in certain kinds of proteomic experiments.
That observation stemmed from Coon's finding that his lab's move from Thermo Fisher Scientific's Orbitrap Fusion to its Orbitrap Fusion Lumos provided only minimal improvements in the depth of coverage they achieved in a one-hour analysis of yeast cell lysate despite the latter instrument offering twice the speed of the former.
Looking into this finding, Coon and his colleagues determined that the Lumos's scan speed outstripped the speed at which a typical LC column produced peptides for analysis, suggesting that the instrument's capacity could not be fully utilized with standard chromatography systems.
To demonstrate this point, the researchers built their own higher-performance LC columns in house and found that this more than doubled the number of peptide sequences they were able to identify.
Max Planck Institute of Biochemistry Professor Matthias Mann is an early user of the µPAC columns and said that he sees great potential for the technology in proteomics work. Like Vanhonacker, he cited their reproducibility as highly attractive.
"In principle, when using the same gradients, the peptides should elute at exactly the same times, regardless of the age of the column, what column it is, and even if it is in a lab in Europe, US or Asia," he said. "This would make peptide or small molecule retention time collections truly useful [and] could, in general, make experiments much more comparable if widely adopted."
He added that the technology also appears to be much more robust than normal columns, noting that the company has "done thousands of injections without change" and that his lab's experience "has also been that we have not seen any change over many injections." Additionally, Mann noted that the columns' ability to run at relatively low pressure should make the system more durable and easier to maintain.
The column design also leads to "very little [sample] absorption to the columns and hence a low carryover and high performance at very small amounts," he said.
"Altogether, the promise of these columns would be that they make the LC part of LC MS/MS 'digital' in that it can be predicted and that the performance should stay constant," Mann said.
He noted that for shorter gradients, his lab has found that its own homemade columns still top the µPAC columns in terms of resolution, but, he added that these homemade columns also outperform all other commercially available columns. The µPAC columns are also more expensive than conventional LC columns, Mann said, but he suggested that their potentially longer life span and higher robustness could make up for the increased upfront cost.
Large LC vendors like Agilent offer microchip-based LC systems, but Devenyns said these systems typically still rely on packed particles for doing separations. The notion of an etched microchip-based system is not original to PharmaFluidics, he said, noting that researchers like Fred Regnier, formerly a professor of chemistry at Purdue University and an expert in proteomics sample prep and separations, have previously explored the idea.
PharmaFluidics, however, has developed intellectual property around implementation of this concept. Most notably, Devenyns said, the company has determined chip architectures capable of incorporating long flow paths without creating significant peak dispersion.
"The wafer is eight inches in diameter at best, and we put a two-meter column on it," he said. "So there are a lot of serpentine turns, and a lot of the art is in designing appropriate turn structures so that the peak dispersion [is minimized]. That is where we have substantially novel and patented IP."
"Then there are some other ideas in terms of surface treatment of the pillars and a few other things that are either patent protected or proprietary know-how," he added.
The current version of the µPAC column is designed for nanoflow separations, which are commonly used in proteomics and other areas where depth of coverage is key. Proteomics researchers have increasingly embraced higher flow rates in recent years, though, with the field moving toward microflow separations. Vanhonacker said the current chips can operate at a microflow rate if run at higher pressure. She added that they can also be upgraded for microflow applications.
The columns are available globally, with the bulk of sales thus far coming from Europe and Australia, Vanhonacker said. Europe and the US are the company's initial areas of focus, she added.
PharmaFluidics currently has 12 fulltime employees and plans to double that number using money from its recent funding round, which it said will be put toward developing additional products and expanding its commercial reach.
In addition to marketing the µPAC technology on its own, the company is also looking into forming agreements with other instrument vendors to sell the columns for certain applications, Devenyns said. He declined to name any specific companies that PharmaFluidics is in discussions with but did cite a poster at the 2017 annual meeting of the Human Proteome Organization that the company presented with Thermo Fisher.
That poster demonstrated that a µPAC column coupled to a Thermo Fisher Fusion Lumos mass spec was able to measure more than 3,000 proteins in a sample of roughly 50 HeLa cells.