Continuous Online Protein Quality Monitoring during Perfusion Culture Production Using an Integrated Micro/Nanofluidic System

We demonstrate a new micro/nanofluidic system for continuous and automatic monitoring of protein product size and quantity directly from the culture supernatant during a high-cell-concentration CHO cell perfusion culture. A microfluidic device enables clog-free cell retention for a bench-scale (350 mL) perfusion bioreactor that continuously produces the culture supernatant containing monoclonal antibodies (IgG1). A nanofluidic device directly monitors the protein size and quantity in the culture supernatant. The continuous-flow and fully automated operation of this nanofluidic protein analytics reduces design complexity and offers more detailed information on protein products than offline and batch-mode conventional analytics. Moreover, chemical and mechanical robustness of the nanofluidic device enables continuous monitoring for several days to a week. This continuous and online protein quality monitoring could be deployed at different steps and scales of biomanufacturing to improve product quality and manufacturing efficiency.


Table of Contents
• Fabrication of the nanofluidic device • Section 1. Nanofluidic online protein size monitoring system integrated with perfusion culture.
o Figure S1. Perfusion culture of CHO cells using the microfluidic cell retention device.
o Figure S2. High-concentration perfusion culture of CHO cells using the microfluidic cell retention device.
o Method for perfusion culture using the microfluidic cell retention device o Figure S3. Online nanofluidic protein size monitoring system.
o Figure S4. Continuous online buffer exchange and cell clarification.
o Figure S5. Continuous online protein labeling, free dye removal, and protein denaturation.
o Figure S6. Protein size monitoring by the nanofluidic filter array device.
• Section 2. Characterization of protein sizing in the nanofluidic device.
o Figure S7. Offline separation of protein mixture in the nanofluidic device.
o Table S1. Information about the proteins used for Figure S1.  • References S-4

Fabrication of the nanofluidic device
The nanofluidic device (slanted nanofilter array) was fabricated through multiple standard MEMS fabrication methods. To make the nanofilter array on a silicon substrate, photolithography by a stepper with a 5X reduction (NSR2005i9, Nikon Precision Inc.) and dry etching by reactive ion etching (RIE) were used. The access holes to load samples and apply electric field were made by wet etching using potassium hydroxide (KOH). A thermal oxide layer (SiO2) was grown on the silicon substrate for electrical insulation between the silicon substrate and buffer solution. Fusion bonding method was used to bond the silicon and glass (Pyrex) substrates to prevent nanochannel collapse because the aspect ratio of width and depth in the nanochannel was small. As a final step, the bonded substrates were cut by die saw machine.
S-5 Section 1. Nanofluidic online protein size monitoring system integrated with perfusion culture.  Method for perfusion culture using the microfluidic cell retention device CHO cells were grown in a customized spinner flask whose working volume was 350 mL.
The pH (7.0) and dissolved oxygen (40%) of the bioreactor were automatically controlled by a commercial bioreactor controller (BIOSTAT ® A PLUS, Sartorius Stedim North America Inc.).
Sodium bicarbonate (7.5%) solution (S8761, MilliporeSigma) was used as a base solution. The detailed procedures for fabrication of the microfluidic cell retention and perfusion culture were described before. 1 Perfusion began on day 3 with a rate of 700 mL day -1 (two bioreactor volumes per day). Fresh cell culture medium (CD OptiCHO™, 12681011, Thermo Fisher Scientific) was continuously supplied into the bioreactor while cell culture supernatant containing monoclonal antibodies (IgG1) and toxic metabolites were removed by the microfluidic cell retention device.
The culture harvest removed from the bioreactor was collected in a harvest bottle, which was replaced daily. Most of the cells (>98.5%) were maintained in the bioreactor and reached high cell concentration (20-40 million cells mL -1 ). They continuously produced monoclonal antibodies (IgG1). Cell culture was sampled daily and, cell culture parameters, such as cell concentration, viability, live cell diameter, pH, glucose and lactate concentrations, and oxygen level, were measured by the automated cell culture analyzer (FLEX II, NovaBiomedical).

S-10
A customized hollow fiber membrane with reduced internal volume was used to remove free dyes from the protein-dye mixture ( Figure S5). The hollow fiber from the module (C06-E005-05-N, Spectrum Labs) was cut off and inserted into Luer fittings (64-1579, 64-1578, Warner Instruments). Subsequently, a PEEK tubing (1571, IDEX Health & Science) was inserted into the hollow fiber. An adhesive was applied to the tubing and fiber to hold the tubing and prevent leakage during continuous online free dye removal. The complete free dye removal setup was placed at room temperature for more than 24 hours to ensure fully curing of the adhesive.
Finally, the free dye removal setup was connected to the online protein labeling setup. 10X PBS with pH 7.2 (70013032, Thermo Fischer Scientific) was diluted with deionized water by 10-fold to prepare 1X PBS. This 1X PBS was continuously flowed into the free dye removal setup to remove free dyes from the protein-dye mixture. with a metal ceramic resistive heater (HT24S2, Thorlabs) to denature the proteins. The resistance of the heater was controlled by a DC power supply, and the temperature was monitored by a resistance temperature detector (TH100PT, Thorlabs). The final labeled, purified, and denatured protein solution was fed into the nanofluidic filter array device to monitor its size distribution.   Table S1 (see below) shows the detailed information about the proteins. 200 V was applied to the filter array. All the proteins were fluorescently labeled and denatured using SDS and DTT. Error bars, data range (n = 3).

S-14
The signal intensity of the certain protein size marker was dominant in the certain postconcentration channel. For example, the first and second highest peaks for the trypsin inhibitor (20.1 kDa) were in the post-concentration channels #2 and #3, respectively. The postconcentration channels #3 and #4 exhibited the first and second highest peaks, respectively, for both Ovalbumin (44.3 kDa) and standard IgG1 (23.5 kDa of two light chains and 50 kDa of two heavy chains). Moreover, the post-concentration channels #4 and #5 contained most of the fluorescence signals coming from β-Galactosidase from E. coli (116 kDa). Table S1. Information about the proteins used for Figure S1. The purity was measured with offline gel electrophoresis equipment (Bioanalyzer 2100, Agilent).    Bioanalyzer). The bioreactor and post-clarification samples were denatured using an offline denaturation method, while the post-denaturation sample was not additionally denatured, because it was already denatured through the online denaturation step (Figure S11). The result shows that the post-denaturation sample had more HMWP (32.5%) and less MAIN (62.4%) than the bioreactor and post-clarification samples, noting that the online denaturation method was incomplete, compared with the offline method.
To identify the cause of incomplete online denaturation, first, standard IgG1 was tested (see below; Figure S12). Standard IgG1 was denatured through the online system (65 °C using a ceramic heater). In the online sample preparation system, the final DTT concentration after mixing with labeled proteins was 7.9mM. At this DTT concentration, 45.6% MAIN and 48.8%  respectively. These results mean that the higher proportion of HWMP in the perfusion cultures could be due to incomplete denaturation of IgG1 on the online sample preparation system.  Please see the next page. (Online: "the sample is diverted from the process and may be returned to the process stream", Inline: "the sample is not removed from the process stream", at-line: "the sample is removed, isolated from and analyzed near to the process stream", offline: "the sample is tested in a conventional quality control (QC) lab outside of the production area")