Miniaturized scalable arrayed CRISPR screening in primary cells enables discovery at the single donor resolution

Patel, Miti A.; Boribong, Brittany P.; Sinha, Hugo; Xiao, Bin; Xie, Keqiang; Vo, Philippe Q. N.; Chin, Andrew B.; Ellouzi, Ayoub; Little, Samuel R.; Shih, Steve C.C.; Wu, Hao; Muller, William J.; Hirukawa, Alison

doi:10.1038/s41598-025-13532-z

Download PDF

Article
Open access
Published: 11 August 2025

Miniaturized scalable arrayed CRISPR screening in primary cells enables discovery at the single donor resolution

Scientific Reports volume 15, Article number: 29350 (2025) Cite this article

3355 Accesses
1 Altmetric
Metrics details

Subjects

A Correction to this article was published on 30 September 2025

This article has been updated

Abstract

High-efficiency gene editing in primary human cells is critical for advancing therapeutic development and functional genomics, yet conventional electroporation platforms often require high cell input and are poorly suited to parallelized experiments. Here we introduce a next-generation digital microfluidics (DMF) electroporation platform that enables high-throughput, low-input genome engineering using discrete droplets manipulated on a planar electrode array. The system supports 48 independently programmable reaction sites and integrates seamlessly with laboratory automation, allowing efficient delivery of CRISPR-Cas9 RNPs and mRNA cargo into as few as 3,000 primary human cells per condition. The platform was validated across diverse primary human cell types and cargo modalities, demonstrating efficient delivery of various cargo, with high rates of transfection, gene knockout via non-homologous end joining, and precise knock-in through homology-directed repair. To showcase its utility in functional genomics, we applied the platform to an arrayed CRISPR-Cas9 screen in chronically stimulated human CD4⁺ T cells, identifying novel regulators of exhaustion, including epigenetic and transcriptional modulators. These findings establish our DMF-based electroporation platform as a powerful tool for miniaturized genome engineering in rare or precious cell populations and provide a scalable framework for high-content genetic screening in primary human cells.

Highly efficient CRISPR-mediated genome editing through microfluidic droplet cell mechanoporation

Article Open access 16 September 2024

Scalable continuous-flow electroporation platform enabling T cell transfection for cellular therapy manufacturing

Article Open access 26 April 2023

Rapid generation of homozygous fluorescent knock-in human cells using CRISPR–Cas9 genome editing and validation by automated imaging and digital PCR screening

Article 20 September 2024

Introduction

Gene editing in primary human cells has become increasingly essential for both therapeutic applications and fundamental research, including immuno-engineering, rare disease modeling, and functional genomics^1,2,3. While a variety of transfection technologies exist, electroporation has been able to produce high-efficiency genome editing. However, conventional electroporation-based delivery systems often require hundreds of thousands to millions of cells per condition⁴ limiting their utility in settings where cell availability is constrained or there is limited proliferative runway, such as with patient-derived samples or rare immune subsets like Regulatory T cells^5,6. Moreover, existing methods are frequently incompatible with large-scale screens because of limitations in throughput, reagent consumption, and process automation^7,8.

Coupling electroporation with microfluidics has emerged as a promising strategy for low-input genome editing^9,10,11. While most microfluidic electroporation systems to date have employed flow-based architectures, which guide cells and cargo through fixed microchannels^12,13 these platforms often lack flexibility in reagent mixing, involve complex fluidic setups that are difficult to integrate into automation, and require relatively high cell input per condition. In contrast, digital microfluidics (DMF)—which manipulate nanoliter- to microliter-scale droplets via electrostatic actuation on a planar electrode array—offers fine control over reaction composition, timing, and localization^14,15,16,17. With no moving parts and plug and play automation system, this discrete droplet paradigm is particularly well-suited for high-throughput and low-volume applications, where conservation of reagents and primary cells is critical.

Recent studies have shown the feasibility of DMF-based genome editing in both prokaryotic and mammalian systems^18,19. For example, Little et al. incorporated CRISPR-based knockout and knock-in editing in primary human cells by implementing Tri-Drop Electroporation^20,21 in which two conductive buffer droplets flank a central droplet of cell suspension (Supplementary Fig. S1A). This tri-droplet structure bridges the anode and cathode electrodes to form a transient, low-current electroporation zone that enables efficient delivery of RNPs and mRNA, while minimizing Joule heating, hydrolysis by-products, and other viability-compromising effects often observed in cuvette-based systems^22,23.

Building on prior demonstrations of DMF for intracellular delivery, we developed a next-generation electroporation platform featuring 48 independently programmable reaction sites, integration with laboratory automation, and scalable device manufacturing. The system retains key advantages of DMF while enabling high-efficiency gene delivery using up to 100-fold fewer cells than standard cuvette-based systems.

In this study, we aimed to validate the platform’s performance across multiple primary human cell types, both adherent and suspension, delivering a range of biomolecular cargos—including mRNA and CRISPR-Cas9 ribonucleoprotein (RNP) complexes—with high transfection and knockout efficiencies using as few as 3,000 cells per edit. To showcase the system’s potential for functional genomics, we applied it to an arrayed CRISPR-Cas9 screen targeting 45 candidate regulators of exhaustion in chronically stimulated human CD4⁺ T cells. By integrating phenotypic markers (e.g., LAG-3 expression), cytokine secretion profiles (IFNγ, TNFα), and viability metrics, we identified multiple perturbations that reversed features of exhaustion, both well-characterized checkpoint molecules and less-explored epigenetic and transcriptional regulators in CD4⁺ T cells.

These findings validate the next generation system as a powerful tool for miniaturized, high-throughput genome engineering in primary cell populations and establish a framework for identifying next-generation targets to further our understanding of disease biology.

Results

Evaluating the impact of cell number on transfection efficiency

Commercially available electroporation methods are widely used for transfecting primary cells but require high input cell numbers, limiting their utility with rare or patient-derived populations. To assess the effect of cell number on transfection efficiency, EGFP mRNA was delivered into two primary human cell types: skeletal muscle myoblasts (adherent) and T cells (suspension), across a range of input cell densities using the 96 well Lonza Nucleofector shuttle. In myoblasts, GFP expression was highest at 200,000 and 100,000 cells/edit at 48 h post-transfection (98.72% ± 0.89 and 75.27% ± 7.19, respectively). Efficiency dropped sharply at lower cell numbers, with 50,000 cells/edit yielding 28.04% ± 7.33 GFP + cells, and < 10% at 10,000 cells/edit or below. At 2,500 cells/edit, expression was negligible and comparable to no-template controls (Fig. 1A). A similar trend was seen in T cells. 48 h post-transfection, 250,000 cells/edit resulted in 84.67% ± 9.68 GFP + cells, while 125,000 and 40,000 cells/edit yielded 46.58% ± 11.43 and 19.03% ± 3.81, respectively. At 10,000 cells/edit, expression dropped to 1.98%, near background levels (Fig. 1B). These findings demonstrate that efficient gene delivery with the Lonza system depends on high cell input, highlighting a need for alternative approaches when working with limited cell numbers.

High-throughput transfection in low cell densities with digital microfluidics

To address low transfection efficiency in low cell densities, we developed a state-of-the-art digital microfluidic (DMF) platform that enables precise delivery of a variety of payloads into mammalian cells with high-throughput, array processing. The system allows users to customize and optimize electrical parameters over a wide range, transfect up to 100x less cells per edit, and simultaneously perform 48 unique electroporation reactions on a single cartridge. Its SBS-format design and compatibility with liquid handlers enable integration with automated workflows and scalable high-throughput screening workflows (Fig. 1C. Supplementary Fig. S1B). The recommended workflow for using the system is to deposit payloads onto the bottom plate substrate, use a liquid handler to load cells and additional payloads onto the cartridge, and run our platform with user defined electroporation parameters. After transfection, cells are to be offloaded for further recovery and culture and transfection efficiency can be determined using a variety of methods, such as flow cytometry, sequencing, and high-content imaging (Fig. 1C).

Analysis of EGFP mRNA transfection efficiencies in primary human myoblasts and T cells using our transfection platform

To evaluate the performance of our transfection platform, we transfected primary human skeletal muscle myoblasts (adherent) and primary human T cells (suspension) with EGFP mRNA at low input cell densities and plated at a high density of 9–31,000 cells/cm² into a 96-well plate. Results were compared to those obtained using the Lonza Nucleofector system, which showed poor transfection efficiency under similar low-input conditions.

3,000 myoblasts were transfected per edit with EGFP mRNA using our transfection method and monitored post-transfection for GFP expression for 48 h and proliferation for 108 h. Viability of myoblasts were measured via confluency as consistent cell growth is an indicator of a viable cell population. At 48 h post-transfection, confluence measurements indicated consistent cell growth across conditions, with the off-chip control showing the highest confluence (69.31% ± 7.72), and on-chip groups showing comparable levels (non-electroporated: 59.99% ± 7.22; electroporated without EGFP mRNA: 57.79% ± 13.89; electroporated with EGFP mRNA: 48.91% ± 3.86) (Fig. 1D). GFP expression in the electroporated group reached 76.50% ± 2.42 48 h post-transfection, while all control groups exhibited minimal fluorescence (non-electroporated: 0.26% ± 0.3; no mRNA control: 0.01% ± 0.02; off-chip: 0.01% ± 0.01) (Fig. 1D).

Primary human T cells were similarly transfected with EGFP mRNA using our platform with 10,000 cells per edit. Proliferation was monitored for up to 156 h post-transfection, revealing sustained cell expansion, with a sharp increase beyond the 100-hour mark (Fig. 1E). 48 h post-transfection, GFP fluorescence reached 45.50% ± 11.00 by microscopy (Fig. 1E). Flow cytometry analysis was also conducted at 24 h post-transfection to assess acute cell health and cell density. Representative gating for the flow cytometer analysis can be seen in Supplementary Fig. S2A. T cell counts for on-chip samples were close to the initial cell input of 10,000 cells/edit (non-electroporated: 11,835 ± 1,272; no mRNA control: 10,508 ± 1,469; electroporated with EGFP mRNA: 9,115 ± 1,050). Samples that were electroporated with mRNA showed a significantly lower cell count compared to the non-electroporated samples (p = 0.0027). Cell counts for off-chip controls (18,210 ± 1,400) were significantly higher compared to all on-chip samples (p < 0.0001) (Supplementary Fig. S2B). Acute viability remained high in both electroporated (75.42% ± 2.04) and non-electroporated (87.24% ± 0.51) groups (p < 0.0001) (Supplementary Fig. S2B). Flow cytometry analysis confirmed high transfection efficiency: 90.69% ± 2.18 in CD4⁺ T cells and 92.23% ± 2.12 in CD8⁺ T cells (Fig. 1E). Minimal and significantly lower expression was observed in non-electroporated controls (CD4+: 0.58% ± 0.38; CD8+: 1.62% ± 0.97) (p < 0.0001) (Fig. 1E). Together, these results demonstrate that our transfection system enables highly efficient mRNA delivery in both adherent and suspension primary human cells at low input cell numbers—conditions under which conventional nucleofection methods are largely ineffective. The combination of high transfection efficiency, minimal impact on viability, and compatibility with small cell numbers supports the utility of the presented transfection system for genome engineering in challenging primary cell systems.

Validation of gene knockout in primary human cells

We next validated the delivery of a Cas9 ribonucleoprotein (RNP) complex to generate a gene knockout. For adherent primary human myoblasts, we targeted the muscleblind like splicing regulator 1 (MBNL1) which is abundantly expressed in skeletal muscle cells²⁴ has been shown to have dysregulated function within myotonic dystrophy, and plays a key role in disease progression²⁵. Using our transfection platform, MBNL1 RNP was delivered to 3,000 myoblasts per edit. 96 h post-transfection, analysis of the DNA was performed via Sanger sequencing and quantified via Inference of CRISPR Edits (ICE) analysis to determine knockout of MBNL1. Non-electroporated samples showed no knockout (0.00% ± 0.00) and electroporated samples showed a significantly higher knockout score of 97.83% ± 1.84 for MBNL1 (p < 0.0001) Loss of MBNL1 was further confirmed at the protein level via immunoblot, with an average of 84% KO (84% MBNL1 KO ± 11%, n = 4) (Supplementary Fig. S3A-C). These results validate that our transfection method can generate a high knockout in primary human myoblasts with a low starting number of cells.

We next validated a knockout in primary human suspension cells by targeting the T cell receptor alpha constant (TRAC) gene in various T cell subsets. The TRAC gene is a core component of the T cell Receptor (TCR) complex, which recognizes antigens bound to major histocompatibility complexes, which then activates signaling cascades within T cells, which then undergo a variety of functions such as cytokine production and cell differentiation^26,27,28. TRAC or non-targeting control (NTC) RNP was delivered to 10,000 T cells per edit using our Transfection System. 72 h post-transfection, viability of cells and the surface expression of T cell receptor α/β (TCR α/β) was evaluated using flow cytometry. Representative gating for the flow cytometer analysis can be seen in Supplementary Fig. S4A. The viability of the T cells post-knockout was > 95% in both the samples transfected with the NTC RNP and the TRAC RNP (97.83% ± 0.09 and 97.23% ± 0.28, respectively) (p = 0.0017) (Fig. 2B). TCR α/β knockout was negligible in the NTC RNP samples for both CD4⁺ and CD8⁺ T cells (0.37% ± 0.30 and 1.02% ± 1.02, respectively) (Fig. 2B). However, in the TRAC RNP samples, TCR α/β knockout was significantly higher in both CD4⁺ and CD8⁺ T cells (90.41% ± 3.27 and 91.87% ± 3.13) (p < 0.0001) (Fig. 2B). These results showed we could produce high knockout efficiency in a low number of T cells, and we next wanted to test this protocol on a rare T cell subset.

Regulatory T cells are specialized T cells that maintain homeostasis within the immune system through suppression of inflammatory cells^29,30 and have been shown to prevent autoimmunity through maintaining self-tolerance, making them a potential therapeutic target for treating autoimmune disease³¹. However, Regulatory T cells are rare, comprising only 5–10% of the circulating CD4⁺ T cells^32,33. Since our transfection system requires a small cell input, we wanted to test if we could edit this rare T cell subset. We first confirmed that the Regulatory T cell population was CD4⁺ CD25 + FOXP3 + and found 67% of the population was positive (Supplementary Fig. S4B). We targeted the TRAC gene for knockout and 72 h post-transfection, viability of cells and the surface expression of TCR α/β were evaluated using flow cytometry. Representative gating for the flow cytometer analysis can be seen in Supplementary Fig. S4C. Regulatory T cells showed high viability in both the non-electroporated and electroporated samples (98.66% ± 0.29 and 97.22% ± 0.89, respectively (p = 0.0015) (Fig. 2C). In the non-electroporated samples, there was minimal TCR α/β knockout (3.98% ± 0.61), as expected. In the electroporated samples, the cells showed significantly higher TCR α/β knockout at 76.78% ± 8.62 (p < 0.0001). These results validate and showcase the utility of this platform for effective transfection of rare cell types that are low in cell number.

Efficient intracellular delivery of large DNA templates enables knock-in edits in primary T cells

To assess the ability of the platform to deliver gene editing components for knock-in applications, we performed targeted insertion of a 4 kb GFP template at the TRAC locus in primary human Pan CD3 + T cells. The GFP template consisted of a circular single-stranded DNA (cssDNA) construct, which has been shown to exhibit improved intracellular stability and serve as an effective payload for knock-in experiments³⁴. Fluorescence microscopy confirmed successful integration ten days post-delivery (Fig. 2D). Quantification of GFP + cells was assessed via flow cytometry. Representative gating for the flow cytometer analysis can be seen in Supplementary Fig. S4D. Edited cells exhibited a clear GFP shift relative to controls, with GFP expressions for 500 V, 2 ms, 2 pulses being 52.37% ± 9.70 (0.3 µg) and 41.14% ± 9.45 (0.6 µg) and 500 V, 4 ms, 2 pulses being 42.02% ± 7.56 (0.3 µg) and 47.99% ± 8.24 (0.6 µg), which are all significantly higher than both the no-template condition (11.21% ± 2.83) and the off-chip control (10.62% ± 4.49) (p < 0.0001) (Fig. 2D). These results validate the platform’s capacity to deliver large-format DNA payloads efficiently into primary T cells using minimal input material.

To further explore therapeutic use cases, we delivered a non-reporter 3 kb double stranded linear DNA template encoding a second-generation anti-HER2 chimeric antigen receptor (CAR), flanked by TRAC-targeting homology arms. Representative gating for the flow cytometer analysis can be seen in Supplementary Fig. S4E. Delivery was well tolerated, with all conditions showing viabilities above 55% (Fig. 2E). CAR expression was measured by Protein L staining, which detects single chain fragment variable (scFv) domains and confirmed by co-staining with CD3. HER2 CAR expression was up to 40%, with an average of 27.03%, with significantly elevated levels compared to the no-template condition (0.0375 ng Homology Directed Repair (HDR)/edit: p < 0.001, all other concentrations: p < 0.0001) (Fig. 2E). Notably, no dose-dependent significant differences were observed across template concentrations, indicating that efficient delivery can be achieved with reduced DNA input (Fig. 2E).

These findings demonstrate that our novel transfection system supports high-efficiency intracellular delivery of multi-kilobase nucleic acid templates into primary T cells at low input cell numbers. The ability to generate both reporter-expressing and functional CAR-T cells using a miniaturized, array-based system highlights its utility for scalable genome engineering and therapeutic development in primary cell populations.

Arrayed CRISPR KO screen identifies genetic regulators of CD4⁺ T cell exhaustion

Chronic T cell activation in cancer and persistent viral infections often drives T cells into a dysfunctional exhaustion state, characterized by impaired effector function, sustained expression of inhibitory receptors, and widespread transcriptional and epigenetic remodeling^35,36. Exhausted CD4⁺ and CD8⁺ T cells display diminished cytokine production, reduced proliferative capacity, and elevated expression of inhibitory receptors such as PD-1, LAG3, and CTLA-4^37,38,39,40. Although immune checkpoint blockade has shown clinical success in partially reversing exhaustion⁴¹ the molecular circuitry that maintains this state remains unknown.

Functional genomics studies in CD8⁺ T cells have revealed key transcriptional and epigenetic regulators of exhaustion, including TOX, NR4A family members, DNMT3A, and EZH2^38,39,40. However, systematic efforts to map analogous mechanisms in CD4⁺ T cells remain limited despite their critical role in sustaining anti-tumor immunity and orchestrating broader immune responses⁴⁰. Identifying pathways that enable CD4⁺ T cells to resist exhaustion may yield new opportunities to enhance cellular therapies.

To address this gap, we used our platform to perform a miniaturized, arrayed CRISPR-Cas9 knockout screen targeting 45 putative or known exhaustion regulators in chronically stimulated primary human CD4⁺ T cells with an input of 10,000 cells per target (Fig. 3A). Post electroporation, cells were restimulated every three days to model chronic T cell exhaustion. This long, but gentle stimulation schedule was intentionally designed to promote the gradual acquisition of exhaustion phenotypes, rather than inducing acute activation-induced cell death or differentiation. By extending the stimulation period over 28 days, we aimed to more faithfully recapitulate the transcriptional and functional changes that occur during progressive T cell dysfunction in vivo. On Day 28, we assessed exhaustion phenotypes by measuring expression of the inhibitory receptor LAG-3 using flow cytometry (Fig. 3B). Culture supernatants were collected for cytokine profiling by ELISA.

Knockout of several targets modulated LAG-3 expression, a canonical marker of T cell exhaustion (Fig. 3C). As expected, the knockout of LAG3 itself produced the strongest reduction in LAG-3 expression. Interestingly, perturbations such as SIN3A, PRDM1, and ID2 led to significant increases in the frequency of LAG-3-CD4⁺ T cells relative to controls (NTC, off-chip control), suggesting a reversal of exhaustion-associated surface phenotypes (Fig. 3C).

To distinguish true modulators of exhaustion from perturbations that affected cell viability or T cell identity, we assessed acute viability (Supplementary Fig. S5A) and CD4⁺ T cell frequency (Supplementary Fig. S5B) across all conditions. In parallel, we evaluated functional restoration by measuring secretion of IFN-γ and TNF-α (Fig. 3D,E). Supernatants from targets with significantly different LAG-3 expression than the controls were analyzed via sandwich ELISA on Day 28. Knockouts such as PRDM1 and PAF1 enhanced effector cytokine production, while perturbations that increased exhaustion markers were associated with reduced cytokine levels—consistent with a dysfunctional or terminally exhausted phenotype.

To integrate these diverse readouts, we calculated a composite exhaustion escape score by summing z-score normalized values for IFNγ, TNFα, and %LAG3⁻ CD4⁺ T cells and ranked the targets (Supplementary Fig. S5C). Perturbations with poor viability (< 70%) or diminished CD4⁺ frequency (< 90%) were deprioritized due to confounding effects on cell health or lineage integrity.

The screen revealed several knockout targets that reproducibly enhanced effector function while reducing exhaustion markers. Top hits included known regulators like LAG3 and PRDM1, but also identified less characterized genes such as SIN3A, SETBP1, PAF1, and STAT1. These targets were associated with increased cytokine secretion and higher frequencies of non-exhausted (LAG-3-) CD4⁺ T cells. Radar plot analysis highlighted SIN3A and PAF1 as particularly strong multi-dimensional escape mediators, showing robust gains across all phenotypic and functional axes (Fig. 3F).

Collectively, this screen uncovers a diverse set of regulators that modulate CD4⁺ T cell exhaustion, ranging from canonical immune checkpoints to, to our knowledge, the first report of certain chromatin-associated and transcriptional regulators. The enrichment of factors like SIN3A and SETBP1 reinforces emerging evidence that epigenetic control is central to the exhaustion program. Meanwhile, PAF1—a component of the transcriptional elongation complex—may influence exhaustion through effects on chromatin accessibility or transcriptional pausing. Hits such as PRDM1 and STAT1 further support the importance of cytokine signaling and terminal differentiation in exhaustion regulation. Importantly, this screen was enabled by our platform’s ability to perform high-efficiency CRISPR editing using 25-fold fewer cells and ~ 50-fold less Cas9 per condition than conventional methods (Fig. 3G). Additionally, utilizing less cells per edit can reduce experimental labor and costs in consumables and reagents needed for cell culture. The platform’s miniaturized, arrayed format enabled efficient functional screening in primary human T cells with limited cell input, offering a scalable and cost-effective strategy for immune cell engineering and target discovery.

Discussion

In this study, we detail our state-of-the-art DMF transfection platform that can deliver a variety of payloads into mammalian cells, including sensitive primary human cells. We have shown that we can perform genome edits in as few as 3,000 cells per reaction while maintaining high cell viability, whereas more widely used electroporation methods require 10–100× cellular inputs and show reduced cell health when cell density too low^43,44,45. Our transfection platform represents a significant advancement for applications working with rare cell populations or patient-derived cells that are limited in number. We have also shown that the platform can be used with various cell types, showing high transfection efficiency in a primary human adherent and suspension cell. Additionally, our platform can perform 48 edits simultaneously, allowing for high-throughput screenings to be completed with lower cell numbers and lower amounts of reagents than traditional electroporation methods.

We demonstrated that we could perform clinically relevant CRISPR knockouts in adherent and suspension primary cells. In adherent primary human myoblasts, we showed a > 90% knockout in the MBNL1 gene with a cell input of 3,000 cells, a gene that has been implicated in muscle disorders such as myotonic dystrophy²⁵. Current CRISPR screens completed in myoblasts use a viral-based transfection method^46,47,48 which has drawbacks in requiring larger volumes of cells, cost of reagents needed, and potential cytotoxicity and immunogenicity of viral vectors to cells. Further studies can incorporate an arrayed CRISPR knockout screen to identify genes implicated in muscle disorders and potential drug targets across multiple donors.

Performing TRAC knockout in primary suspension human T cells offers a variety of implications in the realm of cancer immunotherapy. The TRAC locus not only results in uniform CAR expression but also enhances T-cell potency⁴⁹. Targeting the CAR to the TRAC locus averts tonic CAR signaling and allows for effective internalization and re-expression of the CAR, delaying effector T-cell differentiation and exhaustion. With the many potential benefits to using the TRAC locus for CAR-T development, our results showing effective TRAC knockout with high viability compared to other methods clearly show the potential for creating CAR-T cells⁵⁰.

Henceforth, we validated the TRAC locus as a clinically relevant targeting locus by using a GFP knock-in construct. Results with our transfection platform demonstrated knock-in efficiencies on par with published results³⁷. Circular single-stranded DNA (cssDNA) enables more efficient targeted genomic integration than double-stranded DNA (dsDNA) in both immortalized and primary cells, while also preserving cell viability by avoiding the immunogenic response typically triggered by dsDNA. Yet, with many CAR constructs still being dsDNA, we similarly performed KI edits with an anti-HER2 CAR construct, demonstrating the efficacy of our system in performing edits utilizing various types of constructs. We successfully created anti-HER2 CAR-T cells, using only 1 million donor cells to streamline process development and test numerous concentrations to find the optimal amount of construct per edit to obtain the highest knock-in efficiencies. Currently, CAR-T immunotherapy is one of the novel approaches to treat tumors, but there haven’t been widespread clinical applications of this therapy due to the inefficiency of CAR-T cells. Furthermore, many modern CAR-T development strategies include viral methods to achieve high KI efficiencies, which is not ideal for clinical applications⁴⁹. Alternatively, our transfection system allows for efficiencies comparable to viral methods, but without the risk that comes with viral-based methods. Moreover, our transfection system allows for testing many different conditions with far fewer reagents to truly optimize CAR-T cells to make them more effective⁴⁷. Additionally, since our system requires a smaller cellular input, this makes it optimal for autologous cell therapies as we can minimize the number of cells isolated from the patient. In the future, our platform can examine how simultaneous gene knockouts can improve the overall efficacy of CAR-T cells⁵⁰. With the gentler approach and miniaturization offered by our transfection system, high-throughput editing has become increasingly possible.

This study also demonstrated the feasibility of performing a high-throughput, arrayed CRISPR-Cas9 screen in primary human CD4⁺ T cells to examine T cell exhaustion using our transfection platform. We utilized prolonged stimulation of T cells that allowed for the induction of a gradual, physiologically relevant exhaustion phenotype, enhancing the translatability of findings. For this study, we focused on cell surface marker of exhaustion, LAG-3, and leveraged functional readouts to define exhaustion. However, T cell exhaustion is a spectrum, not a binary state. A limitation of our study is that by focusing solely on LAG-3 expression, nuanced phenotypes or early versus terminal exhaustion subsets are missed. TIM3, TIGIT, PD-1, and transcriptional markers like TOX or EOMES often co-express variably and offer better resolution when analyzed together and would be interesting to examine in future screens^37,42,51.

We were able to carry out multiple downstream phenotyping assays, including surface marker profiling and cytokine secretion analysis. This multiplexed approach enabled the identification of several gene knockouts that partially reversed exhaustion-associated dysfunction, highlighting the power of low-input, high-throughput screening to uncover regulators of T cell exhaustion in physiologically relevant models. Importantly, all top hits retained high viability and CD4 expression, suggesting that their impact on function is not attributable to off-target toxicity or loss of lineage identity. Well-known regulators of exhaustion such as PRDM1⁵² and LAG3 were recovered, validating the sensitivity of our approach. Interestingly, among the top hits were SIN3A, SETBP1, and PAF1—factors not traditionally linked to immune regulation, suggesting under-explored roles for chromatin remodeling and transcriptional elongation in maintaining the exhausted state.

SIN3A is a transcriptional corepressor that recruits histone deacetylase complexes; its loss may derepress effector genes typically silenced in exhausted T cells. While SIN3A has not been directly shown to repress IFN-γ or TNF-α in T cells, these loci are known to be epigenetically silenced in exhausted T cells⁵³ and HDAC inhibition has been shown to be important for T cell function^54,55,56,57. Given that SIN3A serves as a scaffold for HDAC-containing repressor complexes, it is plausible that SIN3A contributes to the suppression of effector programs during T cell exhaustion. This hypothesis warrants further investigation, for example by using SIN3A ChIP-seq in exhausted versus non-exhausted T cells to map direct genomic binding sites and assess overlap with silenced effector loci. SETBP1 has not been previously implicated in T cell exhaustion, but its known molecular functions suggest a potential role in maintaining the exhausted state. In myeloid malignancies, SETBP1 interacts with the SET oncoprotein and chromatin regulators to repress transcription of differentiation and tumor suppressor genes⁵⁸. These effects are mediated, in part, through recruitment of histone methyltransferases and the establishment of repressive chromatin marks⁵⁹. Given that T cell exhaustion is stabilized by an epigenetically repressive landscape—including factors like EZH2, DNMT3A, and TOX^42,53—it is plausible that SETBP1 may similarly enforce transcriptional silencing of effector genes in exhausted T cells. Our data showing increased cytokine secretion and LAG-3- cells upon SETBP1 knockout support this hypothesis and warrant further investigation using ChIP-seq to map SETBP1 occupancy in exhausted versus functional T cells.

Together, these findings demonstrate the versatility and power of our DMF-based electroporation platform for genome engineering in primary human cells. By combining low cell input requirements with high editing efficiencies and compatibility with arrayed functional assays, this system enables broad applications ranging from therapeutic cell engineering to mechanistic interrogation of gene function. While this study focused on myoblasts and T cells, broader validation across additional primary cell types—such as CD34⁺ hematopoietic stem cells or epithelial cells—would further support platform generalizability. Thus, future work will focus on expanding the repertoire of compatible cargos and cell types, as well as integrating multi-omic readouts to deepen insights into cellular state and function following genome perturbation.

Materials and methods

Primary human skeletal muscle myoblast culturing

Primary human skeletal muscle myoblasts from a male, non-smoking donor was obtained from Zen-Bio and cultured according to protocols provided by the vendor. Briefly, cells were cultured at 37 °C, 5% CO₂ in Skeletal Muscle Cell Growth Media (Zen-Bio) with media changes every 2–3 days. Cells were cultured until 70% confluent, then passed and expanded. Cells were not grown past 90% confluent to prevent fusing and differentiation of cells and were not used past Passage 10. One myoblast donor was used for both the GFP mRNA transfection and the knockout experiment.

Primary human T cell procurement and isolation

Pan CD3 + T cells were derived from Peripheral Blood Mononuclear Cells (PBMCs) that were isolated from a fresh leukopak (Charles River Laboratories) via density gradient separation using Histopaque-1077 (Sigma Aldrich). From the isolated PBMC population, Pan CD3 + T cells were isolated via negative selection using the EasySep Human T Cell Enrichment Kit (STEMCELL Technologies) following manufacturer’s protocol. After isolation, T cells were aliquoted and cryopreserved in liquid nitrogen until use. One leukopak donor was used for the GFP mRNA transfection and knockout experiments. Another leukopak donor was used for the knock-in experiments.

Human natural T regulatory cells (CD4⁺ CD25+) derived from the peripheral blood (iQ Biosciences) and Peripheral Blood derived CD4⁺ Helper T Cells, Negatively Selected (AllCells) were purchased directly from the listed vendors. One regulatory T cell donor and one CD4⁺ T cell donor were used for their respective experiments.

Primary human T cell activation for transfection or knockout experiments

Pan CD3 + T cells were thawed and cultured at a concentration of 1 × 10⁶ cells/mL in ImmunoCult™-XF T Cell Expansion Medium (STEMCELL Technologies) with ImmunoCult™ Human CD3/CD28 T Cell Activator (STEMCELL Technologies) at manufacturer’s recommendation concentration of 25 µL/mL and 200 IU/mL IL-2 (PeproTech).

CD4⁺ Helper T cells were thawed and cultured at a concentration of 1.25 × 10⁶ cells/mL in Basal T Cell Media (RPMI 1640 with Glutamax (Gibco) + 10% FBS + 1% penicillin-streptomycin) with 400 IU/mL IL-2, 25 ng/mL IL-7 (Miltenyi Biotec), and 50 ng/mL IL-15 (Miltenyi Biotec).

T regulatory cells were thawed and cultured at a concentration of 1 × 10⁶ cells/mL in with Basal T Cell Media with 1 µg/mL soluble anti-human CD28 antibody (BioLegend) and 300 IU/mL IL-2. T regulatory cells were plated in a 96-well round-bottom plate that was coated with 10 µg/mL anti-human CD3 antibody (BioLegend) overnight at 4 °C.

All T cells were cultured and activated at 37 °C, 5% CO₂ for 3 days prior to transfection experiments.

Primary human T cell activation for knock-in experiments

Pan CD3 + T cells were thawed and cultured at a concentration of 2.5 × 10⁵ cells/mL in Basal T Cell Media with 200 IU/mL IL-2, 10 ng/mL IL-7, 5 ng/mL IL-15, and Dynabeads™ Human T-Activator CD3/CD28 (Gibco) at a bead to cell ratio of 2:5 in a 96-well round-bottom plate. Cells were cultured and activated at 37 °C, 5% CO₂ for 3 days prior to transfection experiments. Prior to transfection, Dynabeads™ were removed from cells.

Transient EGFP mRNA transfection for primary human myoblasts and T cells with the Lonza Nucleofector

Primary Human Myoblasts and Pan CD3 + T cell electroporation were performed using the Amaxa™ 96-well Shuttle™ with the 4D Nucleofector (Lonza). EGFP mRNA (TriLink Biotechnologies) was added to Nucleofector^® Solution at a final working concentration of 0.1 mg/mL for both cell types. The SF Cell Line 96-well Nucleofector^® Kit (Lonza) was used for cell preparation. For primary human myoblast cells, 2,500 − 200,000 cells per reaction were used with program EY-100. For primary human T cells, 10,000−250,0000 cells per reaction and were used with program EO-115. After electroporation, myoblasts were offloaded into Skeletal Muscle Cell Growth Media (Zen-Bio) in a flat-bottom 96-well plate and T cells were offloaded in Basal T Cell Media with 200 IU/mL IL-2. Plates were imaged on the Incucyte Live-Cell Analysis System (Sartorius). Phase contrast and fluorescent images were collected at 20X every 6 h for 2 days. %GFP/%confluency and %confluency was analyzed using the Incucyte Live-Cell Imaging and Analysis software.

Cell preparation and offloading using the Dropgenie transfection system

Cell count and cell viability were assessed using a NucleoCounter^® NC-202™ (ChemoMetec) and cells were used for experiments if viability was > 80%. Before transfection experiments, cells were washed in DMFection Buffer (DropGenie) and resuspended in Transfection Buffer (0.01% Surfactant F (ITW Reagents) in DMFection Buffer). Cells with payload are loaded onto the DropGenie cartridge using an 8-channel 125 µL Integra Voyager pipette on an Assist Plus. Post-transfection, cells are offloaded from the DropGenie cartridge using the Assist Plus into a 96-well plate.

Transient EGFP mRNA transfection in primary human myoblasts and T cells using the Dropgenie transfection system

Myoblasts were loaded onto the cartridge at a concentration of 3.0 × 10⁶ cells/mL in Transfection Buffer with EGFP mRNA (TriLink Biotechnologies) at a final concentration of 200 µg/mL in 100 µL. The experimental parameters used for electroporation are voltage: 500 V, pulse duration: 3 ms, number of pulses: 3. Post-transfection, cells were offloaded into Skeletal Muscle Cell Growth Media in a 96-well flat-bottom plate and imaged on the Incucyte Live-Cell Analysis System (Sartorius) at 37 °C, 5% CO₂. Phase contrast and fluorescent images were collected at 20X every 2 h for 9 days. %GFP/%confluency and %confluency was analyzed using the Incucyte Live-Cell Imaging and Analysis software.

T cells were loaded onto the cartridge at a concentration of 1.0 × 10⁷ cells/mL in Transfection Buffer with EGFP mRNA (TriLink Biotechnologies) at a final concentration of 100 µg/mL in 100 µL. The experimental parameters used for electroporation for Incucyte analysis are voltage: 500 V, pulse duration: 3 ms, number of pulses: 2. Post-transfection, cells were offloaded into Basal T Cell Media with 200 IU/mL IL-2 in a flat-bottom 384-well plate and imaged on the Incucyte Live-Cell Analysis System (Sartorius) at 37 °C, 5% CO₂. Phase contrast and fluorescent images were collected at 10X every 4 h for 12 days. %GFP/%confluency and %confluency was analyzed using the Incucyte Live-Cell Imaging and Analysis software. The experimental parameters used for electroporation for flow cytometer analysis are voltage: 500 V, pulse duration: 2 ms, number of pulses: 2. Post-transfection, cells were offloaded into Basal T Cell Media with 200 IU/mL IL-2 in a 96-well round-bottom plate and cultured for 24 h at 37 °C, 5% CO₂ prior to flow cytometer analysis (Attune NxT Flow Cytometer, Thermo Fisher). Flow cytometry was performed to determine viability, CD4⁺ and CD8⁺ T cells, and GFP positivity percentage using the flow cytometry protocol detailed below. Flow cytometry data can be accessed in the following link: https://doi.org/10.6084/m9.figshare.28817420.

Dropgenie cartridge preparation for knockout experiments

Guides (Supplementary Table S1) were resuspended to 100 pmol/µL (T cell experiments) and 200 pmol/µL (myoblast experiments) and Poly-L-glutamic acid (PGA) (Sigma Aldrich) was reconstituted to 100 mg/mL in nuclease-free water. Guides and PGA were then combined to achieve a final guide concentration of 55.6 pmol/µL (T cells experiments) or 111.1 pmol/µL (myoblast experiments) and added to an Echo^® Qualified 384-well Low Dead Volume plate (Beckman Coulter). The guide-PGA mixtures were then deposited onto the DropGenie substrate in 40 nL (T cell experiments) or 65 nL (myoblast experiments) spots using an Echo 655 (Beckman Coulter) (Supplementary Fig. S1B). Cartridges were assembled when the guides were fully dehydrated.

MBNL1 knockout in primary human myoblasts

Myoblasts were loaded onto the cartridge at a concentration of 3.0 × 106 cells/mL in Transfection Buffer with SpCas9 Nuclease (Aldevron) at a final concentration of at a final concentration of 300 pmol in 100 µL. Each edit contained 3,000 myoblasts, 7 pmol of guide, and 3 pmol of Cas9. The experimental parameters used for electroporation are voltage: 500 V, pulse duration: 3 ms, number of pulses: 3, incubation time: 10 min. Post-transfection, cells were offloaded into Skeletal Muscle Cell Growth Media in a flat-bottom 96-well plate and cultured for 96 h at 37 °C, 5% CO2.

Total genomic DNA was collected, and PCR amplification was performed using the Platinum™ Direct PCR Universal Master Mix Kit (Invitrogen), following manufacturer’s protocol. PCR products were sent to Quintara Bio for Sanger Sequencing. Sequencing data was analyzed using the Inference of CRISPR Edits (ICE) analysis tool (Synthego Performance Analysis, ICE Analysis. 2019. v3.0. Synthego). Primers are detailed in Supplementary Table S1. Sequencing data can be accessed in the following link: https://doi.org/10.6084/m9.figshare.28817111.

Western blotting for MBNL1 in primary human myoblasts

Primary human myoblasts were lysed in RIPA buffer (Thermo Scientific) supplemented with Halt Protease & Phosphatase inhibitors (Thermo Scientific). Protein concentrations were quantified using a BCA assay (Sigma), and 20–40 µg of total protein per sample was resolved on a 12% Tris-Glycine SDS-PAGE gel and transferred to a PVDF membrane. Membranes were blocked in 5% BSA (Thermo Scientific) in TBST + 0.2% Tween (Thermo Scientific) for 1 h at room temperature, then incubated overnight at 4 °C with polyclonal anti-MBNL1 antibody (Cell Signaling (94633) 1:500 dilution). After washing, membranes were incubated with HRP-conjugated anti-rabbit secondary antibody (1:1000) for 1 h at room temperature. Signal was detected using ECL substrate (Pierce) and imaged on a chemiluminescence system (iBright). Tubulin (Cell Signaling) was used as a loading control.

CRISPR KO in primary human T cells using the Dropgenie transfection system

Activated T cells were loaded onto the cartridge at a concentration of 1.0 × 10⁷ cells/mL in Transfection Buffer with SpCas9 Nuclease (Aldevron) at a final concentration of at a final concentration of 42 pmol in 100 µL. Each edit contained 10,000 T cells, 2.1 pmol of guide, and 0.42 pmol of Cas9. Additionally, regulatory T cells included 0.5 pmol of Alt-R™ Cas9 Electroporation Enhancer (Integrated DNA Technologies) per edit.

The experimental parameters used for transfection for Pan CD3 + T cells are voltage: 500 V, pulse duration: 3 ms, number of pulses: 2, incubation time: 5 min. The experimental parameters used for transfection for T regulatory and CD4⁺ T cells are voltage: 550 V, pulse duration: 2 ms, number of pulses: 2, incubation time: 5 min.

Post-transfection, Pan CD3 + T cells and regulatory T cells were offloaded into Basal T Cell Media with 200 IU/mL IL-2 in a 96-well round-bottom plate and cultured for 3 days at 37 °C, 5% CO₂. Flow cytometry was performed to determine viability, CD4⁺ and CD8⁺ T cells in the Pan CD3 + T cells and CD + CD25 + T cells in the regulatory T cells, and TRAC knockout percentage using the flow cytometry protocol detailed below. Flow cytometry data for the TRAC knockout in Pan CD3 + T cells can be accessed in the following link: https://doi.org/10.6084/m9.figshare.28818038. Flow cytometry data for the TRAC knockout in regulatory T cells can be accessed in the following link: https://doi.org/10.6084/m9.figshare.28817729.

For the CRISPR KO screen, post-transfection, CD4⁺ T cells were offloaded into Basal T Cell Media with 200 IU/mL IL-2 in a 96-well round-bottom plate and cultured for 3 days at 37 °C, 5% CO₂. To induce exhaustion in vitro, CD4⁺ stimulated in 50 µL Basal T Cell Media with ImmunoCult™ Human CD3/CD28 T Cell Activator (STEMCELL Technologies) at manufacturer’s recommendation concentration of 25 µL/mL and 10 IU/mL IL-2, with media changes every 3 days for 3 weeks. Exhaustion status was confirmed via flow cytometry expression of PE-LAG3 (BioLegend) using the flow cytometry protocol detailed below. Flow cytometry data can be accessed in the following link: https://doi.org/10.6084/m9.figshare.28821560.

Evaluation of secreted factors from exhausted CD4⁺ T cells via ELISA

On Day 26 of exhaustion stimulation, samples were re-stimulated with 50 µL Basal T Cell Media with ImmunoCult™ Human CD3/CD28 T Cell Activator (STEMCELL Technologies) at manufacturer’s recommendation concentration of 25 µL/mL and 10 IU/mL IL-2). After 48 h of culture, cell supernatants were collected and preserved at − 80 °C until ELISA processing. TNFα and IFNγ Sandwich Elisa kits were used according to manufacturer’s instructions (RayBiotech), with supernatants diluted in Assay Diluent B (RayBiotech) with dilution factors of either 3.84 or 2.5 depending on available sample. Parameter Logistic (4PL) Regression was used to interpolate unknown sample concentrations from standard curves. Media only (MO) conditions were used to subtract the background presence of TNFα and IFNγ, and the concentration was normalized to event counts enumerated via flow cytometry and adjusted based on dilution factors.

Generation of dsDNA donor template for homology-directed insertion of anti HER2 chimeric antigen receptor (CAR)

Design of homology arms was performed as described by Roth et al.¹. Cloning of a 3000 bp insert size was performed with multiple fragment In-Fusion cloning according to manufacturer’s instructions (Clontech). Briefly, synthesis of dsDNA sequences homologous to the target’s locus were ordered (Alt-R™ HDR Donor Block, Integrated DNA Technologies). Inserts containing a P2A self-cleaving peptide, a CAR/reporter transgene, and a polyadenylation signal derived from bovine growth hormone were designed and synthesized (Alt-R™ HDR Donor Block, Integrated DNA Technologies). A HER2-specific CAR was designed based on a sequence from Priceman et al.⁶⁰.

Stable genomic integration of full circles’ GFP construct in primary T cells using our transfection system

TRAC RNP was assembled using 110 pmol TRAC sgRNA and 33 pmol of SpCas9 complexed on ice for 20 min. TRAC GFP construct from Full Circles Therapeutics was tested at concentrations of 0.06 and 0.03 µg per edit.

T cells were loaded onto the cartridge at a concentration of 1.0 × 10⁷ cells/mL in Transfection Buffer with the described RNP and knock-in construct. The experimental parameters used for transfection are voltage: 500 V, pulse duration: 2 ms or 4 ms, number of pulses: 2. Each edit contained 10,000 T cells, 1.1 pmol of guide, 0.33 pmol of Cas9, and either 0.06–0.03 µg of the knock-in construct.

Post-transfection, T cells were offloaded into Basal T Cell Media with 300 IU/mL IL-2 in a 96-well round-bottom plate and cultured for 10 days at 37 °C, 5% CO₂. Flow cytometry was performed to determine CD4⁺ and CD8⁺ T cells, TRAC knockout, and GFP positivity percentage using the TRAC KO flow cytometry protocol detailed below. Flow cytometry data can be accessed in the following link: https://doi.org/10.6084/m9.figshare.28817957 Images were also taken on day 10 post-transfection using the EVOS Cell Imaging Systems (Thermo Fisher).

Stable genomic integration of an anti-HER2 CAR in primary T cells using our transfection system

TRAC RNP was assembled using 110 pmol TRAC sgRNA and 33 pmol of SpCas9 complexed on ice for 20 min. Anti-HER2 CAR from Integrated DNA Technologies was tested concentrations of 10 ng, 5 ng, 0.625 ng, 0.0375 ng, and 0 ng per edit.

T cells were loaded onto the cartridge at a concentration of 1.0 × 10⁷ cells/mL in Transfection Buffer with the described RNP and Anti-HER2. The experimental parameters used for transfection are voltage: 500 V, pulse duration: 4 ms, number of pulses: 2. Each edit contained 10,000 T cells, 1.1 pmol of guide, 0.33 pmol of Cas9, and a concentration of the Anti-HER2 CAR as described above.

Post-transfection, T cells were offloaded into Basal T Cell Media with 300 IU/mL IL-2 and 15 µM Alt-R™ HDR Enhancer V2 (Integrated DNA Technologies) in a 96-well round-bottom plate and cultured for 10 days at 37 °C, 5% CO₂. Flow cytometry was performed to determine viability, CD3T cells, TRAC knockout, and Anti-HER2 CAR-T cells using the flow cytometry protocol detailed below. Flow cytometry data can be accessed in the following link: https://doi.org/10.6084/m9.figshare.28824383.

Flow cytometry protocols

Flow cytometry panels can be found in Supplementary Table S2.

Samples were transferred to a 96-well conical-bottom plate (Sarstedt) and washed twice with Flow Cytometry Buffer (1X PBS with 2% FBS and 2mM EDTA (Invitrogen)). Samples were incubated at 4 °C in the appropriate surface staining master mixes of stains diluted in Flow Cytometry Buffer for 20 min in the dark.

After the cells were stained with the appropriate surface staining master mix, they were washed twice with 1X PBS and stained with a viability dye (Ghost Dye™ Violet 510, Cytek Biosciences) at a concentration of 1:1000 diluted in 1X PBS for 10 min at room temperature. The cells were then washed twice with 1X PBS and then resuspended in Flow Cytometry Buffer prior to analysis.

For the characterization of T regulatory cells, surface and viability staining followed the protocol as detailed above. Following surface staining, cells were fixed and permeabilized using the True-Nuclear™ Transcription Factor Buffer Set (BioLegend) according to the manufacturer’s instructions. Intracellular staining was performed using FITC-FoxP3 (BioLegend). The cells were then resuspended in Flow Cytometry Buffer prior to analysis. Flow cytometry data can be accessed in the following link: https://doi.org/10.6084/m9.figshare.28817822.

For the staining of the Anti-HER2 CAR-T cells, the cells were first washed twice in 4% BSA/PBS and then stained with FITC-Protein L (GenScript) at a concentration of 1:300, diluted in a 4% BSA solution, for 20 min at room temperature in the dark. Cells were washed twice in Flow Cytometry Buffer and followed the surface and viability staining protocol as detailed above.

Data availability

All data supporting the findings of this study are available within the article and its Supplementary Information files or data repositories. Flow cytometry data and source data for amplicon sanger sequencing data are available on Figshare, Project ID 245501 and linked here: https://figshare.com/projects/Miniaturized_Scalable_Arrayed_CRISPR_Screening_in_Primary_Cells_Enables_Discovery_at_the_Single_Donor_Resolution/245501 . The anti-HER2 CAR construct sequence is available upon request from the corresponding author.

Change history

30 September 2025
A Correction to this paper has been published: https://doi.org/10.1038/s41598-025-21748-2

References

Roth, T. L. et al. Reprogramming human T cell function and specificity with non-viral genome targeting. Nature 559, 405–409 (2018).
Article ADS CAS PubMed PubMed Central Google Scholar
Hendel, A. et al. Chemically modified guide RNAs enhance CRISPR-Cas genome editing in human primary cells. Nat. Biotechnol. 33, 985–989 (2015).
Article CAS PubMed PubMed Central Google Scholar
Stadtmauer, E. A. et al. CRISPR-engineered T cells in patients with refractory cancer. Science 367, eaba7365 (2020).
Article CAS PubMed PubMed Central Google Scholar
Li, L. H., Mccarthy, P. & Hui, S. W. High-efficiency electrotransfection of human primary hematopoietic stem sells. FASEB J. 15, 586–588 (2001).
Article CAS PubMed Google Scholar
Sather, B. D. et al. Efficient modification of CCR5 in primary human hematopoietic cells using a MegaTAL nuclease and AAV donor template. Sci. Transl Med. 7, 307ra156 (2015).
Article PubMed PubMed Central Google Scholar
Cromer, M. K. et al. Global transcriptional response to CRISPR/Cas9-AAV6-Based genome editing in CD34 + Hematopoietic stem and progenitor cells. Mol. Ther. J. Am. Soc. Gene Ther. 26, 2431–2442 (2018).
Article CAS Google Scholar
Tian, R. et al. CRISPR Interference-Based platform for multimodal genetic screens in human iPSC-Derived neurons. Neuron 104, 239–255e12 (2019).
Article CAS PubMed PubMed Central Google Scholar
Parnas, O. et al. A genome-wide CRISPR screen in primary immune cells to dissect regulatory networks. Cell 162, 675–686 (2015).
Article CAS PubMed PubMed Central Google Scholar
Lissandrello, C. A. et al. High-throughput continuous-flow microfluidic electroporation of mRNA into primary human T cells for applications in cellular therapy manufacturing. Sci. Rep. 10, 18045 (2020).
Article ADS CAS PubMed PubMed Central Google Scholar
Geng, T. & Lu, C. Microfluidic electroporation for cellular analysis and delivery. Lab. Chip. 13, 3803–3821 (2013).
Article CAS PubMed PubMed Central Google Scholar
Chang, A. Y. et al. Microfluidic electroporation coupling pulses of nanoseconds and milliseconds to facilitate rapid uptake and enhanced expression of DNA in cell therapy. Sci. Rep. 10, 6061 (2020).
Article ADS CAS PubMed PubMed Central Google Scholar
Wang, S., Zhang, X., Wang, W. & Lee, L. J. Semicontinuous flow electroporation chip for high-throughput transfection on mammalian cells. Anal. Chem. 81, 4414–4421 (2009).
Article CAS PubMed PubMed Central Google Scholar
Sharei, A. et al. A vector-free microfluidic platform for intracellular delivery. Proc. Natl. Acad. Sci. U S A. 110, 2082–2087 (2013).
Article ADS CAS PubMed PubMed Central Google Scholar
Choi, K., Ng, A. H. C., Fobel, R. & Wheeler, A. R. Digital microfluidics. Annu. Rev. Anal. Chem. Palo Alto Calif. 5, 413–440 (2012).
Article ADS CAS PubMed Google Scholar
Pollack, M. G., Fair, R. B. & Shenderov, A. D. Electrowetting-based actuation of liquid droplets for microfluidic applications. Appl. Phys. Lett. 77, 1725–1726 (2000).
Article ADS CAS Google Scholar
Barbulovic-Nad, I., Yang, H., Park, P. S. & Wheeler A. R. Digital microfluidics for cell-based assays. Lab. Chip. 8, 519–526 (2008).
Article CAS PubMed Google Scholar
Cho, S.K., Moon, H., & Kim, C.-J. Creating, transporting, cutting, and merging liquid droplets by electrowetting-based actuation for digital microfluidic circuits. J. Microelectromechanical Syst. 12, 70–80 (2003).
Article ADS Google Scholar
Quach, A. B. V., Little, S. R. & Shih, S. C. C. Viral generation, packaging, and transduction on a digital microfluidic platform. Anal. Chem. 94, 4039–4047 (2022).
Article CAS PubMed Google Scholar
Sinha, H., Quach, A. B. V., Vo, P. Q. N. & Shih, S. C. C. An automated microfluidic gene-editing platform for deciphering cancer genes. Lab. Chip. 18, 2300–2312 (2018).
Article CAS PubMed Google Scholar
Little, S. R. et al. A digital microfluidic platform for the microscale production of functional immune cell therapies. (2024). https://doi.org/10.1101/2024.09.03.611092
Little, S. R. et al. A Tri-Droplet liquid structure for highly efficient intracellular delivery in primary mammalian cells using digital microfluidics. Adv. Mater. Technol. 8, 2300719 (2023).
Article CAS Google Scholar
Handbook of Electroporation. (Springer International Publishing, Cham, (2017). https://doi.org/10.1007/978-3-319-32886-7
Li, Y. et al. Electroporation on microchips: the harmful effects of pH changes and scaling down. Sci. Rep. 5, 17817 (2015).
Article ADS CAS PubMed PubMed Central Google Scholar
Ho, T. H. et al. Muscleblind proteins regulate alternative splicing. EMBO J. 23, 3103–3112 (2004).
Article CAS PubMed PubMed Central Google Scholar
Yadava, R. S., Mandal, M. & Mahadevan, M. S. Studying the effect of MBNL1 and MBNL2 loss in skeletal muscle regeneration. Int. J. Mol. Sci. 25, 2687 (2024).
Article CAS PubMed PubMed Central Google Scholar
Mariuzza, R. A., Agnihotri, P. & Orban, J. The structural basis of T-cell receptor (TCR) activation: an enduring enigma. J. Biol. Chem. 295, 914–925 (2020).
Article PubMed Google Scholar
Shah, K., Al-Haidari, A., Sun, J. & Kazi, J. U. T cell receptor (TCR) signaling in health and disease. Signal. Transduct. Target. Ther. 6, 412 (2021).
Article PubMed PubMed Central Google Scholar
Hwang, J. R., Byeon, Y., Kim, D. & Park, S. G. Recent insights of T cell receptor-mediated signaling pathways for T cell activation and development. Exp. Mol. Med. 52, 750–761 (2020).
Article CAS PubMed PubMed Central Google Scholar
Corthay, A. How do regulatory T cells work? Scand. J. Immunol. 70, 326–336 (2009).
Article CAS PubMed PubMed Central Google Scholar
Vignali, D. A. A., Collison, L. W. & Workman, C. J. How regulatory T cells work. Nat. Rev. Immunol. 8, 523–532 (2008).
Article CAS PubMed PubMed Central Google Scholar
Goswami, T. K. et al. Regulatory T cells (Tregs) and their therapeutic potential against autoimmune disorders - Advances and challenges. Hum. Vaccines Immunother. 18, 2035117 (2022).
Article Google Scholar
Jiang, T. J. et al. Percentage and function of CD4⁺ CD25 + regulatory T cells in patients with hyperthyroidism. Mol. Med. Rep. 17, 2137–2144 (2018).
CAS PubMed Google Scholar
Saito, S., Sasaki, Y. & Sakai, M. CD4⁺ CD25high regulatory T cells in human pregnancy. J. Reprod. Immunol. 65, 111–120 (2005).
Article CAS PubMed Google Scholar
Xie, K. et al. Efficient non-viral immune cell engineering using circular single-stranded DNA-mediated genomic integration. Nat. Biotechnol. https://doi.org/10.1038/s41587-024-02504-9 (2024).
Article PubMed PubMed Central Google Scholar
Wherry, E. J. et al. Molecular signature of CD8⁺ T cell exhaustion during chronic viral infection. Immunity 27, 670–684 (2007).
Article CAS PubMed Google Scholar
Pauken, K. E. & Wherry, E. J. Overcoming T cell exhaustion in infection and cancer. Trends Immunol. 36, 265–276 (2015).
Article CAS PubMed PubMed Central Google Scholar
Blackburn, S. D. et al. Coregulation of CD8⁺ T cell exhaustion by multiple inhibitory receptors during chronic viral infection. Nat. Immunol. 10, 29–37 (2009).
Article CAS PubMed Google Scholar
Yao, C. et al. Single-cell RNA-seq reveals TOX as a key regulator of CD8⁺ T cell persistence in chronic infection. Nat. Immunol. 20, 890–901 (2019).
Article CAS PubMed PubMed Central Google Scholar
Scott, A. C. et al. TOX is a critical regulator of tumour-specific T cell differentiation. Nature 571, 270–274 (2019).
Article CAS PubMed PubMed Central Google Scholar
Borst, J., Ahrends, T., Bąbała, N., Melief, C. J. M. & Kastenmüller W. CD4⁺ T cell help in cancer immunology and immunotherapy. Nat. Rev. Immunol. 18, 635–647 (2018).
Article CAS PubMed Google Scholar
Topalian, S. L. et al. Safety, activity, and immune correlates of anti-PD-1 antibody in cancer. N Engl. J. Med. 366, 2443–2454 (2012).
Article CAS PubMed PubMed Central Google Scholar
Khan, O. et al. TOX transcriptionally and epigenetically programs CD8⁺ T cell exhaustion. Nature 571, 211–218 (2019).
Article CAS PubMed PubMed Central Google Scholar
Zhang, C. P. et al. Efficient non-viral delivery of macromolecules in human primary hematopoietic stem cells and lymphocytes. J. Mol. Cell. Biol. 15, mjad018 (2023).
Article PubMed PubMed Central Google Scholar
Zhang, Z., Qiu, S., Zhang, X. & Chen, W. Optimized DNA electroporation for primary human T cell engineering. BMC Biotechnol. 18, 4 (2018).
Article PubMed PubMed Central Google Scholar
Covello, G., Siva, K., Adami, V. & Denti, M. A. An electroporation protocol for efficient DNA transfection in PC12 cells. Cytotechnology 66, 543–553 (2014).
Article CAS PubMed Google Scholar
Liu, S. et al. CRISPR/Cas9 screening reveals Zfp607b as a novel transcription factor regulating myogenesis. Genes Dis. 12, 101444 (2025).
Article ADS CAS PubMed Google Scholar
Liu, W. et al. CRISPR screen identifies the RNA-Binding protein Eef1a1 as a key regulator of myogenesis. Int. J. Mol. Sci. 25, 4816 (2024).
Article CAS PubMed PubMed Central Google Scholar
Phelps, M. P., Bailey, J. N., Vleeshouwer-Neumann, T. & Chen, E. Y. CRISPR screen identifies the NCOR/HDAC3 complex as a major suppressor of differentiation in rhabdomyosarcoma. Proc. Natl. Acad. Sci. U. S. A. 113, 15090–15095 (2016).
Eyquem, J. et al. Targeting a CAR to the TRAC locus with CRISPR/Cas9 enhances tumour rejection. Nature 543, 113–117 (2017).
Article ADS CAS PubMed PubMed Central Google Scholar
Liu, X. et al. CRISPR-Cas9-mediated multiplex gene editing in CAR-T cells. Cell. Res. 27, 154–157 (2017).
Article ADS PubMed Google Scholar
Philip, M. et al. Chromatin States define tumour-specific T cell dysfunction and reprogramming. Nature 545, 452–456 (2017).
Article ADS CAS PubMed PubMed Central Google Scholar
Hwang, S., Cobb, D. A., Bhadra, R., Youngblood, B. & Khan, I. A. Blimp-1-mediated CD4 T cell exhaustion causes CD8 T cell dysfunction during chronic toxoplasmosis. J. Exp. Med. 213, 1799–1818 (2016).
Article CAS PubMed PubMed Central Google Scholar
Sen, D. R. et al. The epigenetic landscape of T cell exhaustion. Science 354, 1165–1169 (2016).
Article ADS CAS PubMed PubMed Central Google Scholar
Ibrahim, M. L. et al. Histone deacetylase inhibitors directly modulate T cell gene expression and signaling and promote development of effector-exhausted T cells in murine tumors. J. Immunol. Baltim. Md. 1950. 212, 737–747 (2024).
CAS Google Scholar
Philips, R. L. et al. HDAC3 restrains CD8-lineage genes to maintain a bi-potential state in CD4⁺ CD8⁺ thymocytes for CD4-lineage commitment. eLife 8, e43821 (2019).
Article PubMed PubMed Central Google Scholar
Liu, S. et al. CK2B induces CD8⁺ T-Cell exhaustion through HDAC8‐Mediated epigenetic reprogramming to limit the efficacy of Anti‐PD‐1 therapy in Non‐Small‐Cell lung cancer. Adv. Sci. 2411053 https://doi.org/10.1002/advs.202411053 (2025).
Boucheron, N. et al. CD4(+) T cell lineage integrity is controlled by the histone deacetylases HDAC1 and HDAC2. Nat. Immunol. 15, 439–448 (2014).
Article CAS PubMed PubMed Central Google Scholar
Cristóbal, I. et al. SETBP1 overexpression is a novel leukemogenic mechanism that predicts adverse outcome in elderly patients with acute myeloid leukemia. Blood 115, 615–625 (2010).
Article PubMed Google Scholar
Piazza, R. et al. Recurrent SETBP1 mutations in atypical chronic myeloid leukemia. Nat. Genet. 45, 18–24 (2013).
Article CAS PubMed Google Scholar
Priceman, S. J. et al. Regional delivery of chimeric antigen receptor-engineered T cells effectively targets HER2 + Breast cancer metastasis to the brain. Clin. Cancer Res. Off J. Am. Assoc. Cancer Res. 24, 95–105 (2018).
Article CAS Google Scholar

Download references

Acknowledgements

We thank John Fuller, Nick Morgan, and Beckman Coulter Life Sciences (BCLS) for logistical support and protocol development with the ECHO Acoustic Dispenser. We are grateful to Morgan Sturgeon and Integrated DNA Technologies (IDT) for assistance with guide design and reagents. We thank Danna Lee from Full Circles Therapeutics and Patrick Zhang from Quintara Bioscience for molecular cloning and production of circular single stranded DNA cargo production. We thank Mitchell Kozakoff at the ICCB-Longwood Screening Facility at Harvard medical School for infrastructure and technical resources. The KNMRC facility at Northeastern University for cleanroom services. We thank Scott Robinson from MicroQuin for his assistance with the immunoblot and analysis. The authors would like to thank Mark Flanagan and James Paolino from CCG and Scitus Engineering, as well as the group at Shakotis Ltd. We also gratefully acknowledge MITACS for supporting BX’s internship through their Globalink program. Funding for student internships was generously provided by the Massachusetts Life Sciences Center (Mass Life Sciences). This work was also supported in part by MEDTEQ+, whose contribution helped advance development and validation of the platform.

Author information

Miti A. Patel and Brittany P. Boribong contributed equally to this work.

Authors and Affiliations

DropGenie, Cambridge, MA, USA
Miti A. Patel, Brittany P. Boribong, Hugo Sinha, Philippe Q. N. Vo, Andrew B. Chin, Ayoub Ellouzi & Alison Hirukawa
Department of Biochemistry, McGill University, Montreal, Canada
Bin Xiao & William J. Muller
Rosalind and Morris Goodman Cancer Institute, McGill University, Montreal, Canada
Bin Xiao & William J. Muller
Full Circles Therapeutics, Cambridge, MA, USA
Keqiang Xie & Hao Wu
Department of Electrical and Computer Engineering, Concordia University, Montréal, QC, Canada
Samuel R. Little & Steve C.C. Shih
Centre for Applied Synthetic Biology, Concordia University, Montréal, QC, Canada
Samuel R. Little & Steve C.C. Shih
Department of Biology, Concordia University, Montreal, QC, Canada
Steve C.C. Shih

Authors

Miti A. Patel
View author publications
Search author on:PubMed Google Scholar
Brittany P. Boribong
View author publications
Search author on:PubMed Google Scholar
Hugo Sinha
View author publications
Search author on:PubMed Google Scholar
Bin Xiao
View author publications
Search author on:PubMed Google Scholar
Keqiang Xie
View author publications
Search author on:PubMed Google Scholar
Philippe Q. N. Vo
View author publications
Search author on:PubMed Google Scholar
Andrew B. Chin
View author publications
Search author on:PubMed Google Scholar
Ayoub Ellouzi
View author publications
Search author on:PubMed Google Scholar
Samuel R. Little
View author publications
Search author on:PubMed Google Scholar
Steve C.C. Shih
View author publications
Search author on:PubMed Google Scholar
Hao Wu
View author publications
Search author on:PubMed Google Scholar
William J. Muller
View author publications
Search author on:PubMed Google Scholar
Alison Hirukawa
View author publications
Search author on:PubMed Google Scholar

Contributions

M.P, B.P.B and A.H were responsible for conducting and designing experiments and co-wrote the manuscript. B.X, H.S, K.Q contributed to experimental design, performing experiments and manuscript review. P.Q.N.V, A.B.C, and A.E facilitated the experiments and assisted with device fabrication. H.W and WJM contributed to experimental planning and manuscript review. S.C.C.S and S.M.L contributed to process development and manuscript review.

Corresponding author

Correspondence to Alison Hirukawa.

Ethics declarations

Conflict of interest

M.A.P, B.P.B, H.S, B.X, P.Q.N.V, A.B.C, A.E, S.C.C.S, and A.H are either current or former employees, or shareholders of DropGenie. K.X. and H.W. are either current or former employees, or shareholders of Full Circles Therapeutics Inc. Patents related to this study have been filed. The other authors declare no competing interests.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

The original online version of this Article was revised: The original version of this Article contained an error in the name of the author Steve C.C. Shih, which was incorrectly stated as Steve Shih. As a result, the Author contributions, Conflict of interest and the Supplementary Information have been corrected.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary Material 1

Supplementary Material 2

Supplementary Material 3

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Patel, M.A., Boribong, B.P., Sinha, H. et al. Miniaturized scalable arrayed CRISPR screening in primary cells enables discovery at the single donor resolution. Sci Rep 15, 29350 (2025). https://doi.org/10.1038/s41598-025-13532-z

Download citation

Received: 09 April 2025
Accepted: 24 July 2025
Published: 11 August 2025
Version of record: 11 August 2025
DOI: https://doi.org/10.1038/s41598-025-13532-z