Volume 40, Issue 3 e3441
RESEARCH ARTICLE
Open Access

Simplifying stable CHO cell line generation with high probability of monoclonality by using microfluidic dispensing as an alternative to fluorescence activated cell sorting

Lina Chakrabarti

Corresponding Author

Lina Chakrabarti

Cell Culture & Fermentation Sciences, BioPharmaceuticals Development, R&D, AstraZeneca, Gaithersburg, USA

Correspondence

Lina Chakrabarti, Cell Culture & Fermentation Sciences, BioPharmaceuticals Development, R&D, AstraZeneca, 1 MedImmune Way, Gaithersburg, MD 20878, USA.

Email: [email protected]

Contribution: Conceptualization, ​Investigation, Writing - original draft

Search for more papers by this author
James Savery

James Savery

Machine Learning & AI, BioPharmaceuticals Development, R&D, AstraZeneca, Cambridge, UK

Contribution: Methodology

Search for more papers by this author
John Patrick Mpindi

John Patrick Mpindi

Biostatistics, BioPharmaceuticals Development, R&D, AstraZeneca, Cambridge, UK

Contribution: Methodology

Search for more papers by this author
Judith Klover

Judith Klover

Cell Culture & Fermentation Sciences, BioPharmaceuticals Development, R&D, AstraZeneca, Gaithersburg, USA

Contribution: Methodology

Search for more papers by this author
Lina Li

Lina Li

Cell Culture & Fermentation Sciences, BioPharmaceuticals Development, R&D, AstraZeneca, Gaithersburg, USA

Contribution: Methodology

Search for more papers by this author
Jie Zhu

Jie Zhu

Cell Culture & Fermentation Sciences, BioPharmaceuticals Development, R&D, AstraZeneca, Gaithersburg, USA

Contribution: Methodology

Search for more papers by this author
First published: 10 March 2024

Abstract

Single cell cloning is a critical step for cell line development (CLD) for therapeutic protein production, with proof of monoclonality being compulsorily sought in regulatory filings. Among the different single cell deposition technologies, we found that fluorescence activated cell sorting (FACS) offers high probability of monoclonality and can allow selective enrichment of the producer cells. However, FACS instruments are expensive and resource-intensive, have a large footprint, require highly skilled operators and take hours for setup, thereby complicating the cell line generation process. With the aim of finding an easy-to-use alternative to FACS, we identified a flow cytometry-based microfluidic cell dispenser, which presents a single cell sorting solution for biopharmaceutical CLD. The microfluidic cell dispenser is small, budget-friendly, easy-to-use, requires lower-cost consumables, permits flow cytometry-enabled multiparametric target cell enrichment and offers fast and gentle single cell dispensing into multiwell plates. Following comprehensive evaluation, we found that single cell deposition by the microfluidic cell dispenser resulted in >99% probability of monoclonality for production cell lines. Moreover, the clonally derived producer cell lines generated from the microfluidic cell dispenser demonstrated comparable or improved growth profiles and production capability compared to the FACS derived cell lines. Taken together, microfluidic cell dispensing can serve as a cost-effective, efficient and convenient alternative to FACS, simplifying the biopharmaceutical CLD platform with significant reductions in both scientist time and running costs.

1 INTRODUCTION

Biotherapeutic proteins represent a powerful and promising modality of treatment for a multitude of conditions, including cancer, autoimmune diseases, hematological disorders, and more. Developing recombinant cell lines and cell culture processes that produce therapeutic proteins with high productivity, in a safe and time- and cost-effective way is core to the biopharmaceutical industry. Cell line development (CLD) is an important step in the process of drug development and typically uses Chinese hamster ovary (CHO) cells.1-3 To minimize population heterogeneity and ensure reproducible product quality, regulatory authorities require that a production cell line must be derived from a single progenitor cell.4 Therefore, single cell cloning is a critical step for biopharmaceutical production of monoclonal antibodies and other therapeutic proteins.

Historically, single cell cloning was performed by limiting dilution plating, and assurance of clonal derivation was assessed based on statistical calculations.5 Although manual dilution methods are gentle on cells due to low sheer stress and provide a high viability after the seeding step, multiple rounds of limiting dilution are required to achieve the desired probability of clonal derivation, thereby making the CLD process time-consuming and inefficient. Several advanced technologies have emerged that can perform the single cell deposition into multiwell plates efficiently. The commonly used methods for the generation of clonally derived cell lines (CDCL) include fluorescence activated cell sorting (FACS)6 and microfluidic cell dispensers7 like the image-guided single cell printer from Cytena8 and the pico-droplet Cyto-Mine system.9 Recently, a nanofluidic technology platform Beacon™ has been developed by Berkeley Lights which allows isolation, maintenance, and analysis of thousands of CDCL in parallel on a single nanofluidic chip.10 Although the Beacon platform offers an opportunity to miniaturize the cell culture system and increase the throughput of the CLD process, the technology is not yet commonly used for isolation of clonally derived CHO cell lines. Traditional FACS instruments are capable of depositing a single cell into microplate wells with both accuracy and efficiency. Previous studies from our laboratory demonstrated >99% probability of monoclonality using a BD Influx™ FACS sorter coupled with high-resolution cell imaging.6 In addition, FACS allows identification and selective enrichment of producing cells based on fluorescent signal intensity of markers predictive of productivity.11-15 However, high pressure inside the sorter nozzle can inflict stress on the cells during the sorting process that may lead to reduced outgrowth of single cells post sorting. Furthermore, these cell sorters are large in size, need significant maintenance, use expensive consumables, require highly skilled operators, and take hours for setup, thereby complicating the cell line generation process. On the other hand, microfluidic cell dispensers have a smaller footprint, lower consumable costs, are easy-to-use, require limited training to operate, and are rapid to set-up. Moreover, the image-based cell dispensers offer increased confidence of the presence of a single cell in a well and provide definitive support for clonal derivation of production cell lines. However, the time taken for image analysis and the use of a decision-making algorithm for each droplet lengthens the single cell deposition procedure significantly compared to FACS, which is less desirable.

In the current market there is limited availability of microfluidic cell dispensers having a combined capability of fluorescence-enabled cell enrichment with fast and efficient single cell deposition. A recent study described the usefulness of a flow cytometry-based microfluidic cell dispenser (Namocell) for cloning human pluripotent stem cells (hPSCs).16 It was demonstrated that the microfluidic cell dispenser not only offers fast single cell dispensing into multiwell plates, its low sorting pressure of <2 psi also provides the gentle cell deposition necessary to preserve the viability and integrity of the hPSCs. In addition, the cIonally derived hPSCs generated using the microfluidic-based cell sorting instrument expressed pluripotency-associated markers and showed normal cell growth.16 As this microfluidic cell dispenser is fully automated, requires no calibration, has low maintenance and with built-in lasers permits multiparametric cell analysis and target cell enrichment, we reasoned that this instrument has the potential to be an efficient and easy-to-use alternative to FACS.

Here we compare Namocell microfluidic cell dispenser Pala™ system head-to-head with FACS using a multiwell plate-based workflow. We demonstrate successful generation of CDCL employing this microfluidic cell dispenser, which also used less resources for the cloning step than FACS. The CDCL from the microfluidic dispenser demonstrated similar or improved growth profiles and production capability compared to the FACS derived cell lines. Importantly, we have also defined a statistical analysis process to calculate the probability of monoclonality from single cell deposition using microfluidic cell dispensing. Therefore, microfluidic cell dispensers, such as the Pala system, can be considered as a suitable and reliable alternative to FACS, which will significantly simplify the single cell cloning step of the CLD process.

2 MATERIALS AND METHODS

A summary of the steps performed in this comparative study between FACS and microfluidic cell dispensing is depicted in the Figure 1 workflow.

Details are in the caption following the image
Workflow to isolate clonally derived cell lines by microfluidic dispensing and fluorescence activated cell sorting, and to compare cell line growth and productivity performance in fed-batch culture.

2.1 Cell culture and stable transfections

A suspension-adapted, proprietary CHO cell line derived from CHO-K1 and a glutamine synthetase selection system was used. Stable transfectant pools were generated by nucleofection of linearized expression plasmids and then selected and maintained in proprietary medium supplemented with 50 μM methionine sulfoximine (Sigma-Aldrich, MO), and 50 mg/L dextran sulfate (Sigma-Aldrich). Suspension cell cultures were grown at 120 rpm on an orbital shaking platform in a humidified incubator set at 37°C and 6% CO2. Cells were passaged every 3–4 days. Measurement of viable cell density and viability was accomplished using a trypan blue exclusion method with a ViCell automated cell counter (Beckman Coulter, CA).

2.2 Single cell sorting by FACS and flow cytometry-enabled microfluidic cell dispensing

Cells were fluorescently labeled with 1 mM of CellTracker™ Deep Red (CTDR, Thermo Fisher Scientific, MA) that was reconstituted with dimethyl sulfoxide (DMSO) (Sigma-Aldrich). Cell suspensions at a concentration of 1 × 106 cells/mL in a sterile tube were mixed with 2 μL of DMSO-CTDR solution. Cells were incubated at room temperature, 120 rpm shaker for 30 min. Cells were centrifuged at 300×g for 5 min, the medium was decanted, and cells were resuspended in 1 mL fresh medium. The cell suspension was passed through a 40 μm cell strainer cap (Becton Dickinson) before sorting. The Influx cell sorter (BD Bioscience, CA, USA) equipped with a 140 micron size nozzle was used for FACS6 and the Pala (Namocell, a Bio-Techne brand) was used for microfluidic dispensing. The fluorescently labeled cells at a concentration of 1 × 106 cells/mL for FACS and 1 × 104 cells/mL for microfluidic cell dispensing were sorted by depositing one cell per well into individual wells of 384-well plates containing proprietary conditioned medium. The sorting conditions for the Influx instrument were optimized to ensure high assurance of single cell isolation and were set as proprietary parameters.

2.3 Single cell imaging and colony outgrowth measurement

Following cell deposition, the plates were centrifuged at 2000×g for 5 min, thereby ensuring that all cells settled at the bottom of the wells.6 Immediately after centrifugation, wells of the 384-well plates were imaged using a Cellavista imager (Synentec, Germany) according to the method described by Evans et al.6 Plates were incubated at 37°C in a humidified atmosphere with 6% CO2 for outgrowth. Confluence per well in 384-well plates was measured using the same imager. The results were analyzed using Nyone software (Synentec, Germany).

2.4 Fed-batch culture at small scale

Antibody production was evaluated by fed-batch culture in 125 mL Erlenmeyer flasks or 96 deep well (DW) plates. The production cultures were grown at 35.5°C in a humidified 6% CO2 atmosphere for 14 days unless otherwise mentioned. Shaker speed was maintained at 120 rpm for flasks and 350 rpm for 96DW plates. Cell density and viability were monitored during cultivation in flasks, but not for the 96DW cultures. Proprietary feed was added periodically to the production cultures. Antibody titers in the culture supernatant were determined using Protein A biosensors in an Octet QK384 (Pall ForteBio, CA) for 96DW and by HPLC (Agilent Technologies, CA) for flask cultures.

2.5 Statistical analysis

Probability of a single cell per well after sorting: To account for experimental variability and the number of wells evaluated (sample size), a one-sided upper 95% confidence interval for the probability of more than one cell per well P(d) was calculated as a conservative estimate of the sorting efficiency. The single-sided upper confidence interval was calculated using the Clopper–Pearson “Exact” method.17 This can be calculated from the 1 − α quantile of the beta distribution given in equation below:
P d = 95 % Upper Conf . Lim . Prob 2 cells well = B 1 α , x + 1 , n x ,
where n is the total number of nonempty observed wells; x is the number of observations with ≥2 cells/well; 1 − α is the target confidence level (1 − α = 0.95).
Probability of cell settling into the focal plane of the Cellavista imager: To account for experimental variability and the number of wells evaluated (sample size), a one-sided upper 95% confidence interval on the probability of observing additional cells settling over time P(i) was calculated as a conservative estimate of the deposition efficiency as described by Evans et al.6 Each centrifuged plate was scored for the presence or absence of cells on the well bottom. Wells that were determined to have no cells were examined after 2 weeks of incubation at 37°C, 6% CO2 for cell outgrowth. The 95% confidence interval was also calculated using the Clopper–Pearson “Exact” method17 as shown in equation below:
P i = 95 % Upper Conf . Lim . Prob cell not settling = B 1 α , x + 1 , n x ,
where n is the total number of empty wells observed after deposition; x is the number of empty wells with cell outgrowth after 2 weeks; 1 − α is the target confidence level (1 − α = 0.95).
The probability of having these two events occur was determined by the probability of having greater than one cell per well, P(d), multiplied by the probability of a cell not settling and deposited into the focal plane for imaging, P(i):
P d × P i = P nonclonal

All other statistical analysis: Data are presented as mean ± SD. Comparisons of mean differences between groups were made by unpaired two-tailed Student's t-test unless otherwise stated. A probability level of p < 0.05 was considered to be statistically significant.

3 RESULTS AND DISCUSSION

3.1 Efficiency and probability of single cell deposition using a microfluidic cell dispenser

To determine the accuracy and efficiency of the microfluidic cell dispenser for depositing a single cell, we used stable transfectant pools of cells expressing three different recombinant proteins, namely two easy-to-express (ETE) mAbs and one difficult-to-express (DTE) non-mAb therapeutic protein. The pools of cells were fluorescently labeled with CTDR and the fluorescent cells were deposited into medium-filled 384-well microplates at a frequency of one cell/well, then the plates were centrifuged and imaged. The image files for each plate were analyzed for the presence of one or more fluorescent cells in each well. Twenty plates from cell deposition events over the span of 4 months were evaluated. No difference was seen in deposition efficiency between cell populations. The frequency of the microfluidic cell dispenser placing only one cell/well was determined to be 96.3% from 7466 wells examined from 20 plates (Table 1). This efficiency was calculated by dividing the number of wells containing a single cell by the total number of wells containing any number of cells. Using these observed sample data shown in Table 1, statistical analysis determined that at the upper limit of the 95% confidence interval, no more than 4.09% of the wells will contain more than one cell.
95 % Upper Conf . Lim . Prob 2 cells well = B 1 α , x + 1 , n x ,
TABLE 1. Efficiency of microfluidic cell dispenser single cell deposition.
# wells sorted # wells empty # wells with any cell # wells with 1 cell # wells with >1 cell frequency of 1 cell/well frequency of >1 cell/well # empty wells with outgrowth after 2 weeks
7466 700 6766 6516 250 96.3% 3.7% 0
where n is the total number of nonempty observed wells (n = 6766 from Table 1); x is the number of observations with ≥2 cells/well (x = 250 from Table 1); 1 − α is the target confidence level (1 − α = 0.95).
95 % Upper Conf. Lim . Prob 2 cells well = B 0.95 , 250 + 1 6766 250 = 0.0409 ,
P d = 4.09 % .

3.2 Probability of cells failing to settle into the focal plane of the imager

All plates were incubated and examined after 2 weeks for outgrowth. Using the observed sample data shown in Table 1, statistical analysis determined that at the upper limit of the 95% confidence interval, no more than 0.043% of the wells will have cells that fail to settle to the bottom of the well after centrifugation for 5 min at 2000×g.
95 % Upper Conf . Lim . Prob cell not settling = B 1 α , x + 1 , n x ,
where n is the total number of empty wells observed after deposition (n = 700 from Table 1); x is the number of empty wells with cell outgrowth after 2 weeks (x = 0 from Table 1); 1 − α is the target confidence level (1 − α = 0.95).
95 % Upper Conf. Lim . Prob(cell not settling) = B 0.95 , 0 + 1 700 0 = 0.00043 ,
P i = 0.043 % .

3.3 Combined overall probability of monoclonality

The probability of monoclonality was estimated by combining the probability of more than one cell per well obtained using the microfluidic cell dispenser P(d) with the probability of cells failing to deposit or settle into the Cellavista focal plane P(i).

Using the observed sample data from Table 1 concerning the efficiency of the microfluidic cell dispenser, we have shown that at the upper limit of the 95% confidence interval, no more than 4.09% of the wells will contain more than one cell.

Thus, P(d) = 0.0409.

Moreover, examining the number of empty wells that contained growing cells at 2 weeks after deposition showed that no empty wells on the day of sorting gave rise to growing colonies of cells after 2 weeks incubation. Using these observed sample data from Table 1, statistical analysis determined that at the upper limit of the 95% confidence interval, no more than 0.043% of the wells will have cells that fail to settle to the well bottom after centrifugation for 5 min at 2000×g.

Therefore, P(i) = 0.00043.

For a production line to be derived from a nonclonal population, the cell sorter must deliver more than one cell in a well and one or more of the cells must fail to settle into the focal plane during centrifugation. Thus, the probability of having these two events occur is the probability of having greater than one cell per well, P(d), multiplied by the probability of a cell not settling into the focal plane for imaging, P(i):
P d × P i = P nonclonal
0.0409 × 0.00043 = 0.0000176 or
4.09 % × 0.043 % = 0.00176 %
P nonclonal = 0.00176 %

Therefore, P(clonal) = 100% − P(non-clonal) or (100% − 0.00176%) = 99.998%.

This provides an overall probability of 0.00176% that multiple cells will be present in a well without being captured in an image. Therefore, the combined overall probability that a cell will be clonally derived using single cell deposition by the microfluidic cell dispenser and photo documentation imaging with verification is greater than 99%, at the 95% confidence interval. Previous studies have shown that single cell deposition by using a FACS instrument also provides greater than 99% probability of monoclonality at the 95% confidence interval.6 Taken together the results indicate that the microfluidic cell dispenser matches the FACS single cell deposition efficiency and hence can be considered as a suitable alternative to FACS.

3.4 Cell culture performance of microfluidic cell dispenser derived cell lines

Development of clonally derived manufacturing cell lines is a crucial step toward ensuring reproducible cell culture performance and generating consistent product quality for biopharmaceuticals. To elucidate the suitability and applicability of a microfluidic cell dispenser for successful generation of cell line, we evaluated the performance of CDCL isolated from the stable pools expressing two different recombinant proteins: an ETE mAb and a DTE non-mAb therapeutic protein. Following deposition of single cells into 384-well plates from the pools for the ETE and DTE molecules, the CDCL were expanded and evaluated in an automated small-scale productivity screen in a 96DW format.18 FACS derived CDCL from the same parental pools were assessed in parallel for comparison. We observed no difference in 384-well outgrowth of the CDCL isolated by FACS and microfluidic cell dispensing (Figure 2a,b). Likewise, 96DW titer evaluation revealed no difference in the median cell line titer achieved between the FACS and microfluidic cell dispenser for either molecule (Figure 2c,d).

Details are in the caption following the image
Characterization of fluorescence activated cell sorting and microfluidic dispenser (Pala) derived cell lines producing easy-to-express (ETE) and difficult-to-express (DTE) molecules. (a, b) Colony outgrowth in 384-well plates at 14 days after single cell deposition. Data represent mean ± SD (n = 4–7). (c) Harvest titer of the cell lines producing an ETE mAb after 3 days in batch culture in 96DW plates. (d) Harvest titer of the cell lines producing a DTE protein after 10 days in fed-batch culture in 96DW plates. Each data point in c and d represents an individual recombinant protein expressing cell line and the horizontal lines in each scatter plot represent the median value.

To better understand the production capability of the cell lines generated by microfluidic cell dispensing compared to those from FACS, cell lines producing the ETE and DTE molecules derived from both cloning methods were cultured for 12–14 days in shake flasks under fed-batch conditions representative of a small-scale bioreactor production process. The top 8–10 CDCL based on 96DW titer data were identified and chosen for this fed-batch evaluation. For the ETE mAb, our results demonstrated a significant increase in the median cell specific productivity for the CDCL from the microfluidic cell dispenser with no difference in the median volumetric titer when compared to the CDCL from FACS (Figure 3a,b). The increased specific productivity of the microfluidic cell dispenser derived cell lines for the ETE mAb can potentially be attributed to their slower growth rate and reduced peak cell density (Figure 3c,d) during the 14-day fed-batch process. No significant difference was observed in the end-of-run (EoR) viability (Figure 3e). It is well established that there is a reciprocal relationship between cell growth and cell specific productivity in CHO cells and many growth-inhibiting approaches have been utilized to enhance protein productivity.19-22 Here for the ETE mAb, the slower growth rate of the CDCL from the microfluidic dispenser suggests that microfluidic cell deposition may enable selection of cell lines with different growth phenotypes which, if sustained, may provide benefits for manufacturing processes. On the other hand, for the DTE molecule, the CDCL isolated by microfluidic cell dispensing and FACS showed no significant difference in growth, viability, and productivity (Figure 4a–e). Interestingly, although the cell lines from both the cloning methods showed a similar wide diversity in titer and cell specific productivity (Figure 4a,b) for the DTE molecule, the distribution of the values indicates an increased frequency of higher producing cell lines being generated from the microfluidic cell deposition (Figure 4a,b). Taken together, the results indicate that recombinant protein producing monoclonal CHO cell lines can be isolated, cultured and screened at high efficiency, thereby qualifying the microfluidic cell dispenser as being suitable for performing key CLD work with agility.

Details are in the caption following the image
Evaluation of fed-batch culture performance of fluorescence activated cell sorting (FACS) and microfluidic dispenser derived cell lines producing an easy-to-express (ETE) molecule. (a) Day 14 titers. (b) Specific productivity (Qp). (c) Cell doubling time (h). (d) Maximum viable cell density (VCDmax). (e) End of fed-batch run (EoR) viability. Each data point represents an individual ETE protein producing cell line (n = 10 cell lines from each of FACS and microfluidic dispenser, Pala) and the horizontal lines in each plot represent the median value. *p < 0.05 by two-tailed Student's t-test considered to be statistically significant.
Details are in the caption following the image
Evaluation of fed-batch culture performance of fluorescence activated cell sorting (FACS) and microfluidic dispenser derived cell lines producing a difficult-to-express (DTE) molecule. (a) Day 12 titers. (b) Specific productivity (Qp). (c) Cell doubling time (h). (d) Maximum viable cell density (VCDmax). (e) End of fed-batch run (EoR) viability. Each data point represents an individual DTE protein producing cell line (n = 8 cell lines from FACS and n = 10 cell lines from microfluidic dispenser, Pala) and the horizontal lines in each plot represent the median value.

3.5 Microfluidic cell dispenser platform requires reduced resources

In addition to the performance of the microfluidic cell dispenser, we compared several other factors pertaining to the time required for equipment set up and single cell deposition in 384-well plates, the cost of consumables and the ease of use between Namocell Pala microfluidic cell dispenser and the BD Influx system. Table 2 shows a comparison of key features of the operation of the two systems. The microfluidic system has benefits in terms of its small footprint, rapid setup, fast single cell dispensing, and budget-friendly consumables.

TABLE 2. Comparison of key features of FACS and microfluidic cell dispensing instrument.
FACS Microfluidic cell dispenser
Dimension 87 × 55 × 94 in. 25 × 14 × 9 in.
Initialization/calibration time 1.5–2 h 5 min
Shutdown time 20 min 2 min
384-well plate sort time 5 min 6 min
Sheath fluid PBS, potential clogging Water, no clogging
Sheath consumption ~4 L/project ~100 mL/project
List price of sterile disposable fluidics kit $3500 N/A
List price of sterile cartridge N/A $110
Highly trained operator Yes No
Maintenance High Low
  • Abbreviations: FACS, fluorescence activated cell sorting; PBS, phosphate buffered saline.

4 CONCLUSION

It is important for biopharmaceutical manufacturing to demonstrate consistent cell culture performance that results in consistent product quality. Accordingly, ICH established guidelines, standards, and control strategies for cell line generation.4 Regulatory agencies require evidence that the production cell line has a high probability of being derived from a single cell. Here, we utilized the single cell deposition capability of the microfluidic cell dispenser together with the monoclonality documentation of the Cellavista imager to demonstrate >99% probability of a single-cell progenitor for production cell lines. Moreover, the CDCL derived from the microfluidic cell dispenser demonstrated similar or improved cell growth and production performance compared to the FACS derived CDCL. The sterile microfluidic disposable cartridges provide advantages in handling numerous samples, mitigating concerns related to cross-contamination between cell lines and waste cleanup. Furthermore, the laser-based flow cytometry technology with typical forward and side scatter detectors and multiple laser configurations allows detection and selective enrichment of a variety of cell phenotypes without the complexity of the FACS process. The simple operation of the microfluidic cell dispenser system not only saves hands-on time, but also makes the device accessible for more end-users to operate independently with less training. Also, when compared to FACS, the microfluidic cell dispenser showed benefits in terms of consumables costs. Taken together, microfluidic cell dispenser can serve as a suitable alternative to FACS by simplifying the CLD platform with a significant reduction in running costs.

AUTHOR CONTRIBUTIONS

Lina Chakrabarti: Conceptualization; investigation; writing – original draft. James Savery: Methodology. John Patrick Mpindi: Methodology. Judith Klover: Methodology. Lina Li: Methodology. Jie Zhu: Methodology.

ACKNOWLEDGMENTS

The authors thank Diane Hatton for critical review of the manuscript. The authors also thank Tom Albanetti for help with the Cellavista imager. The work was supported by AstraZeneca Gaithersburg Flow Cytometry Core Facility.

    FUNDING INFORMATION

    This research received no external funding.

    CONFLICT OF INTEREST STATEMENT

    The authors work for AstraZeneca and own AstraZeneca stock.

    PEER REVIEW

    The peer review history for this article is available at https://www.webofscience.com/api/gateway/wos/peer-review/10.1002/btpr.3441.

    DATA AVAILABILITY STATEMENT

    Data sharing is not applicable to this article as no new data were created or analyzed in this study.