Geert Van Raemdonck (global field support expert at PharmaFluidics) and Koen Sandra (scientific director of the Research Institution for Chromatography, RIC) teamed up for a 10 October 2019 “Ask the Expert” webinar to introduce micro Pillar Array Column (μPAC™) technology for liquid chromatography–mass spectrometry (LC–MS) for host-cell protein (HCP) detection. Van Raemdonck explained that μPAC technology approaches chromatography differently than does packed-bed technology. Microfluidic channels with arrays of free-standing pillars are etched lithographically into a silicon wafer. The resulting permeability and low “on-column” dispersion provided by the perfect order of the separation bed makes μPAC-based chromatography truly unique.
Van Raemdonck’s Presentation
Current protein research relies on packed-bed LC columns of 50–75 cm coupled to high-resolution mass spectrometers to analyze protein samples from tissues, body fluids, and cell lysates. The stationary phase of a conventional column consists of a bed of microparticles packed into a cylinder or capillary tube. Those particles have a certain size distribution. On top of that heterogeneity, their location in the column is more or less random, which leads to additional dispersion as a sample migrates through a column. Because not all flow paths in the column are identical, sample molecules have different retention times, inducing additional dispersion. Besides that problem, packed-bed column technology gradually is reaching its limits in terms of what can be achieved by reducing particle diameter and increasing column length.
To overcome those limitations, PharmaFluidics developed an LC support structure in which separation beds are micromachined into silicon wafers. That presents two key benefits: First, nearly perfect order is achieved in the column, reducing eddy dispersion to an absolute minimum. Second, the fact that the pillars are free standing and positioned at a defined distance from one another improves column permeability, enabling construction of long columns that can be operated at moderate pressures.
Sandra’s Case Study
Often, protein biopharmaceuticals are produced recombinantly in mammalian, yeast, and bacterial expression systems. As well as the therapeutic protein, those cells produce endogenous HCPs. Because such impurities can remain despite multiple purification steps, they must be monitored closely. Enzyme-linked immunosorbent assays (ELISAs) are the gold standard for measuring HCPs because of their high sensitivity and throughput. But ELISAs cannot detect HCPs for which no antibodies are raised. Moreover, those assays establish HCP totals but do not give insight into individual HCPs. In a multicomponent setup, ELISAs also have poor quantitation power. MS in combination with LC separation complements ELISA to offer qualitative and quantitative data on individual HCPs.
A team at RIC combined µPAC technology with high-end orbitrap MS for HCP monitoring downstream. Proteinaceous samples collected at different purification steps were reduced and alkylated prior to overnight trypsin digestion. Then 2 µg of digested sample was loaded onto a 200-cm µPAC column and separated using a 116-min solvent gradient. Semiquantitation of HCPs revealed that more than one peptide-spectrum match was achieved by spiking predigested proteins (e.g., antidiuretic hormone, enolase, polyhydroxyalkanoates, and bovine serum albumin) at known concentrations in every sample. Efficient clearance of impurities (1,812 protein groups) to the final purification step (two groups) was illustrated, corresponding to HCP concentrations of 2.0 × 10⁶ (harvest) and 9.9 ppm (purification step 3). Despite the enormous dynamic range (1.0 × 1010 – 5.0 × 10⁶ on the MS1 level) over which tryptic peptides were present, the combination of µPAC-based separations and high-resolution orbitrap MS/MS obtained acceptable MS/MS spectra for HCPs below 5 ppm.
Questions and Answers
Do the pillars clog, for example during column overloading? Any system can become clogged, but μPAC units are less prone to that due to their small size and permeability.
Do you use static exclusion lists for MS? We did not in this case, but that approach can improve HCP detection. Exclusion lists already are sufficient for peptides, but static exclusion lists merit further study.
How do you quantify data-dependent acquisition (DDA)? MS1 data can be used for semiquantitation. Tracking several data points, we calculate the analytes’ universal response factors, then quantify against four spiked-in proteins using label-free methods.
How many injections can I perform on one chip? It is difficult to predict because each case is sample dependent, but our teams ran a column for six months and performed more than 3,500 injections without loss of performance.
The full presentation of this webcast can be found on the BioProcess International website at the link below.