Selective expansion of target cells using the Enrich TROVO platform

Cell isolation and expansion from mixed populations are crucial steps in various fields of biological research and clinical applications, including cancer research, regenerative medicine and immunotherapy. Obtaining pure populations of target cells is essential for understanding their behavior, function and role in biological processes [1]. By analyzing heterogeneous cell populations in bulk, we can only obtain averaged data about the population, which may lead to the loss of crucial information about small but potentially significant subpopulations.

However, isolating specific cell types from heterogeneous cell populations can be a challenging task, as conventional methods often involve labor-intensive techniques, rigorous timing and a high amount of initial material. More importantly, without combining purification and expansion, purified cells might display compromise cell viability, so functionality thus requires an addition step of verification, which is an outstanding issue for sensitive adherent cells. Serial dilution is still a powerful tool owing to its capacity for integrated isolation and expansion [2]. However, in addition to throughput problems, the extreme low cell concentration used in serial dilution forbids the crosstalk of cells and normally bias toward populations capable of independent growth, largely reducing the sample diversity needed to represent biological heterogeneity.

When cultured in bulk, the uneven growth rate of individual clones in tumor-derived cell pools have led to cellular heterogeneity, with the creation of inconsistencies and failures in many cases [3]:

Challenges in existing technologies on tumor heterogeneity

Clone diversity has been well recognized in genomics. Encapsulation technologies such as emulsion/droplet [9] and plastic/glass chips [9–12] are used to restrain each clone or DNA to a defined physical space, so that the growth of faster growing clones will be limited by self-competition, which has been successful in removing bias during PCR-based bulk amplification processes. Both have been used for cell culture purposes. However, existing encapsulation technologies usually do not suffice in simplicity and flexibility for wide adoption in regular laboratories.

While droplet/emulsion-based technologies [13–16] excel in throughput, the drastically different aeration/nutrition conditions between encapsulated droplets and bulk culture requires extensive optimization. This issue is particularly acute given that many tissue-derived cell populations are sensitive to culture conditions. Second, with droplet/emulsion it is difficult to track individual clones, so that the clonal behaviors of these cells' fate and migration cannot be observed. Third, although cell encapsulation can passively slow down the cellular growth of faster growing clones, they may still overcome such limits and became over-represented. Therefore, an active approach that identifies fast growers and reduces their abundance during culture will be more efficient in maintaining pool diversity. Microwells are most similar to bulk cultures and compatible with existing analysis tools. Without specialized equipment, cells are difficult to retrieve from the premade glass/plastic microwells, preventing its adoption into existing pipelines involving subsequent fluorescence-activated cell sorting, proteomics and sequencing analysis, or as a drug modality. The ‘hard’ wells are also expensive to manufacture and customize. Last but not least, while hydrogel microwells have been reported by individual laboratories, most of these require fixed wafer-based photo-masks [17–21]. However, the ability of customizing shape and dimensions are important to researchers for various research needs that cannot be easily achieved with fixed wafers. For instance, small microwells can be used for short-term single-cell morphology analysis, larger wells can be used for long-term growth studies and elongated wells can be used for cell-mobility measurement.

To address the limitations of existing technologies, we developed an innovative, optically controlled and biocompatible method and commercial machine (Enrich TroVo, Enrich Biosystems Inc., CT, USA) for isolating and expanding target cells from mixed populations. This approach combines the advantages of microfabrication technology with the precision of optical control, offering a versatile and efficient alternative to traditional cell-isolation techniques. By utilizing micro-hydrogel wells (MHWs) to segregate cell populations and an innovative patching technique to selectively eliminate contaminant cells, our method allows for the isolation and expansion of target cells with minimal impact on their viability and proliferation capacity.

This method has the potential to transform the way researchers approach cell isolation and expansion, providing a more efficient and biocompatible solution to a critical challenge in biological research. The versatility and adaptability of this technique makes it suitable for various applications, such as the purification of patient-derived cell lines, cell line construction, and the isolation and expansion of tumor-killing T cells from tumor-infiltrating lymphocytes or endogenous cells. Furthermore, the co-culture setup facilitated by our method enables the investigation of cell–cell interactions in a controlled environment, offering valuable insights into the dynamic behavior of cells within complex systems. Overall, the development of this optically controlled, biocompatible method for target cell isolation and expansion has promising implications for both basic research and clinical applications.

Enrich TroVo system

The approach utilized in this study was derived from a pending patent by Enrich Biosystems Inc. (US20210172856A1, WO2022235507A1). The TroVo device comprises a vertical 5× microscope with a high-resolution CMOS camera (18MP USB 3.0 Color CMOS C-Mount Microscope Camera) mounted on top for capturing images. A fluorescence assembly, featuring replaceable filter sets, illuminating LED and mirrors, is situated between the microscope and the camera. The sample stage is linked to a C7-precision x-y-z actuator that enables sample movement and focusing. A 5.5-inch TFT LCD screen (2560×1440 resolution) is positioned between the stage and the sample to mask it, with the backlight of the LCD removed, leaving only the polarity filter, the RGB filter and the liquid crystal. The bottom of the sample stage has a coaxial light beam consisting of a white LED and a 405-nm laser beam, which is produced by utilizing a 425-nm high-pass dichroic filter. The white beam illuminates the sample and serves as a backlight for the LCD, while the 405-nm laser beam cures the resin through the LCD mask. The mechanics and lighting are controlled by an Azteeg X3 PRO controller board with additional customized circuits. The desktop computer (Intel i5 processor) is used to display masks and patterns on the LCD via HDMI connection. The computer is also used as the device for image capture and processing. The software for image acquisition, processing and particle manipulation control was developed using Python programming language, incorporating Numpy and SciPy libraries for image processing. The stage movement is controlled via the PySerial library using G-Code communication between the machine and the computer. Four images are acquired for each area of interest to accurately address the underlying LCD pixel under each particle of interest. The serial acquisition steps introduce shifts between each image, which are corrected by alignment.

Cell culture & maintenance

OPM-2 cells were cultured in RPMI-1640 medium supplemented with 10% FBS and 1% Pen-Strep. The OPM-2 cell line was derived from the peripheral blood of a 56-year-old female patient diagnosed with multiple myeloma [22]. The OPM-2 model has been used in many myeloma research studies. Human liver fibroblast cells and C4-2 cells were cultured in DMEM supplemented with 10% FBS and 1% Pen-Strep. C4-2 is a cell line with epithelial-like morphology that was isolated from a human prostate cancer LNCaP cell [23]. Cells are grown in T25 flasks and passaged twice weekly.

MHW generation

MHWs or microwells are generated on optically clear-bottomed cell culture containers, such as petri dishes by hydrogel lithography (TroVo MW Application in software suite, patent pending WO2022235507A1), where the sizes and shapes of microwells can be designed by software and projected onto a 2K TFT monochrome LCD display. A wide-field (4 mm in diameter), narrow-angle 405-nm laser is used to induce the gelation of various photopolymerized hydrogels. To generate microwells, reagents (Enrich Cell Cap gel) were added into desired glass-bottomed dishes (for a 14-mm glass-bottomed dish, 150 µl of reagent required). The LCD cover was removed and the microwell print was initiated. The exposure time for microwell printing varied depending on the machine used and required optimization each time. After printing, the dish was washed with phosphate-buffered saline (PBS) two to three times to remove all residue reagents. Finally, 2 ml of PBS was added to all printed dishes and they were incubated overnight. Microwells can be printed on any types of cell culture dishes, with a preference for glass-bottomed dishes owing to their improved image quality.

The next day, the PBS was removed and fresh media added. Cells were introduced into the microwells and cultured while observing their phenotypes. Desired cells containing target microwells and undesired cells containing nontarget microwells were identified. Then, a biocompatible photosensitive hydrogel solution, a patching reagent, was added into cell culture containers and light was employed to solidify the hydrogel into non-target microwells. This allowed the desired cells to expand in a liquid medium while undesired cells were trapped in semi-solid hydrogels.

Co-culture setup

Adherent cell lines (HCC4-Fibroblasts and C4-2 mCherry) were trypsinized them while suspension cells (OPM2-GFP) could be counted directly. The required number of (10–30% of target cells) cells for each co-culture and OPM2-GFP were counted and plated them to microwell-containing plates. Once cells settled down imaging for day 0 was performed.

Patching of fibroblasts (day 5)

The first step was to prepare the patching reagent by mixing 15% gelMA, 0.5% lithium phenyl-2,4,6-trimethylbenzoylphosphinate (LAP) and 0.5 mM tartrazine. This reagent was then be added to the culture plate. Next, target and non-target microwells were identified based on the images obtained. Finally, the non-target microwells were exposed to a 405-nm laser light to trigger the gelation of the patching reagent.

Retrieving cells (day 18)

Continuous culturing and monitoring of cells in the target microwell should be carried out until they reach confluence. For suspension cells (OPM2-GFP), they were retrieved from the target microwell using gel-loading tips and transferred to suitable cell culture containers for continued culturing. In the case of adherent cells (C4-2 mCherry), once they reached confluence within the microwells, they were trypsinized by adding a drop of warm trypsin to their respective wells. After the cells detached, they were collected using gel-loading tips with the aid of a 3.5× magnifying glass. Subsequently, both suspension and adherent cells were further cultured to achieve the desired cell density.

Results & discussion

Undesired cell overgrowth poses a significant challenge in the purification of desired cells. In order to showcase the effectiveness and robustness of our Enrich TroVo platform, we employed rapidly proliferating liver fibroblasts (HCC4) as the undesired cell population. To simulate different desired cell types in various application scenarios, we utilized OPM2-GPF cells as a suspension cell model and C4-2 cells as an adherent cell model. Both cell types inherently expressed fluorescent markers, but it is important to note that any image-based phenotypes can be utilized for selection purposes.

Figure 1 provides a schematic representation of our image-based, phenotype-driven selective sealing process on a 48-well plate. The figure illustrates the sequential steps involved in our methodology; red dots indicate cells exhibiting the desirable phenotype, while the blue dots represent unwanted cells.

Different cell types are initially seeded into a glass-bottomed dish with microwells printed. The microwells serve as individual chambers to isolate and culture the cells. A suitable culture medium is added to the dish, providing the necessary nutrients and conditions for the cells to grow. It is important to note that media can easily exchange throughout the well, allowing for nutrient exchange between the cells in the microwells. To seal or patch the undesired cells, a liquid hydrogel solution is then added. By employing a narrow-angle 405-nm laser, the liquid hydrogel solution undergoes polymerization, resulting in the selective sealing of unwanted wells. This process helps to trap undesired cells within hydrogel, causing them to gradually perish due to nutrient deficiency. The remaining liquid hydrogel solution is subsequently replaced with a cell culture medium. This transition enables the continued growth and culturing of the desired cells within the remaining open microwells. Last, the desired cells can be retrieved for further expansion or downstream applications. This step ensures the generation of a purified and viable cell population represented by red dots in Figure 1. This visual representation helps illustrate the underlying concept of our methodology.

Our TroVo system is highly suitable for co-culturing of feeder cells to promote the proliferation of target cells. As feeder cells are renowned for their capacity to supply a variety of soluble or membrane-bound growth factors and receptors, which aid in the in vitro survival and growth of certain fastidious cells, the TroVo platform is an ideal choice for this purpose.

Furthermore, our TroVo system can be effectively utilized for the generation of monoclonality validated cell lines. This approach offers a less labor intensive and cost-effective alternative to the use of specialized equipment for monoclonal cell generation such as image cell sorters and limiting dilution techniques.

Generation of hydrogel microwells & trapping cells

MHWs are generated by hydrogel lithography, where the size and shape of microwells can be designed using software (Figure 2A) and projected onto a 2K TFT monochrome LCD display (Figure 2B). A wide-field (4 mm in diameter), narrow-angle 405-nm laser is used to induce the gelation of various photopolymerization hydrogels (Figure 2C), most of which are enzymatically digestible. Using gelatin-MA and dextran-MA, which are manufactured in Enrich, we have ascertained the optimal composition of hydrogels to form optically transparent grids that bind wells to regular polystyrene or glass-based petri dish/microtiter plates (Figure 2D). This is achieved by screening and optimizing hydrogel material, the concentration of gel-MA, a photoinitiator (LAP) and light-absorption agent (tartrazine). The microwell generation software is fully parameterized and the generating scanning speed is around 20 mm²/min with the resulting wall of ∼80 µm in height. Based on our preliminary experiments, 10K microwells can be automatically generated within 30 min, which can subsequently be washed with PBS prior to cell culture use. Minimal hands-on time is involved. A 12K monochrome LCD will be tested to generate higher density wells.

Owing to the dimensional difference between fluidic movement and the height of microwell walls, we noticed that cells settled at the bottom of microwells were not affected by macroscopic disturbance. The microtrapping property of MHW can be used to segregate each clone, adherent or suspensive, into a high well-density system, and to register individual clone behavior by its microwell index. More importantly, such trapping capacity can resist common cell-culture operations such as medium change or plate loading for imaging, effectively converting any off-the-shelf cell culture container into a single cell visualization tool without the need for modifying established protocols.

Biocompatibility of hydrogel solutions & microwells

Hydrogels such as matrigels have been widely used to encapsulated tumor cells/primary cells for 3D cell culture. Although constructed by gelatin, other protein factors and optimal gelatin concentrations are needed to support in gel cell growth. In contrast, we have designed a skim hydrogel formula, that although displays good biocompatibility when not polymerized, will starve the encapsulated cells during days of culturing. Provided with the high-precision microgelation technology, we have derived a negative selection method to purify target cells via selective nutrition starvation, and to investigate this using photo-mediated gelatin hydrogels, an experiment was conducted. Initially, 1000 cells were evenly seeded onto a 35-mm petri dish and incubated for 24 h at 37°C to allow for initial cell attachment. Subsequently, the cell culture medium was switched to a solution containing 10% gelatin, 0.5% LAP (a photosensitizer) and 0.5 mM tartrazine (a photoinitiator) in PBS. Using this gelatin solution, a hydrogel disk with a diameter of 500 µm was formed, selectively covering approximately 50 cells. Fresh medium was supplied into the petri dish, and cell growth was monitored over a period of 2 weeks.

On day 13, 0.1 U/ml collagenase was added to the growth medium. This step aimed to dissolve the hydrogel, thereby eliminating the cells that were located underneath it. The process of selective growth starvation was assessed by monitoring the remaining cells and evaluating their response to the removal of the hydrogel. This experiment, as described in Figure 3, provides insights into the implementation of photo-mediated gelatin hydrogels for selective growth control and subsequent elimination of cells.

Selective expansion of suspension cells (OPM2-GFP) from fast-growing contaminant cells (fibroblasts)

Recent explosive applications of cell therapies isolating TILs or other immune cells from surgically excised tumors for further therapeutics purposes presents a need to enrich lymphocytes (mostly in suspension) from complicated tissues and based on a variety of phenotypes, that is, tumor killing or cell proliferation. Therefore, we devised a system to evaluate TroVo's capacity to selectively expand targeted suspension cells. We spiked in 100 fluorescent OPM2 cells with 200 human liver fibroblasts and seeded them to TroVo-generated grid systems (Figure 4A & B). The grid-carrying cell culture plates were then incubated to a common CO₂ incubator for several days and imaged daily. Micrograph time series were segmented and generated for each OPM2 clone and used to determine whether or not they could be a target clone (Figure 4C). On day 5, capture solution was injected to replace the culture medium. Microwells containing proliferated, spherical, green cells were chosen on the stitched image as the target well, while microwells containing unwanted cells were sealed automatically on TroVo (Figure 4D). Cell culturing was resumed and monitored daily for growth of target cells. The cell-retrieval day depended on the confluency status of cells on respective wells. Cell density was closely monitored and whichever well looked confluent had cells retrieved from it. Gel-loading tips were used to retrieve cells with the help of a 3.5× magnifying glass. Furthermore, the retrieved cells continued to grow to obtain fully expanded and highly purified target cells (Figure 4E).

We demonstrated the utility of our optically controlled, biocompatible method for target cell isolation by selectively expanding OPM2-GFP suspension cells from a mixed population (with an initial OPM2 ratio of 10–30%) containing fast-growing contaminant fibroblasts. The target OPM2-GFP cells were successfully harvested with minimal impact on their viability and expansion capacity (Figure 4E).

Selective expansion of adherent cells (C4-2 mCherry) from fast-growing contaminant cells (fibroblasts)

In a second example, we utilized our method to selectively expand C4-2 mCherry adherent cells from a mixed population also containing fast-growing contaminant fibroblasts. The process was similar to the OPM2 example, with seeding and culturing of cells, phenotype observation, target MC selection and hydrogel-based cell isolation (Figure 5A–D). Once the target C4-2 mCherry adherent cells reached confluence within the microwells, they were successfully harvested and expanded (Figure 5E).

In both examples, our optically controlled, biocompatible method for target cell isolation successfully isolated target cells (OPM2-GFP suspension cells and C4-2 mCherry adherent cells) from mixed populations containing fast-growing contaminant fibroblasts. The target cells exhibited minimal impact on their viability and expansion capacity, demonstrating the method's effectiveness in maintaining the health and expansion capacity of isolated cells. These results highlight the method's potential for various applications, including but not limited to: purification of patient-derived cell lines, cell line construction, and the isolation and expansion of tumor-killing T cells from tumor-infiltrating lymphocytes or endogenous cells.

Conclusion

In conclusion, our study presents the successful development and evaluation of the Enrich TroVo system, a novel and innovative approach for efficiently isolating and expanding clonal cell populations from heterogeneous cell groups. By combining microfabrication and optical technologies, including small hydrogel wells and an innovative patching technique, we have demonstrated the system's capacity to selectively eliminate undesired cells while allowing for the growth and expansion of desired cells with minimal impact on their viability and proliferation. The Enrich TroVo system offers a promising alternative to conventional cell isolation methods, particularly for sensitive adherent cells like patient-derived cells. The successful isolation and expansion of target cells, as shown with suspension cells OPM2-GFP and C4-2 mCherry adherent cells, highlight the system's effectiveness in maintaining cell health during the process. Our finding underscores the potential of the Enrich TroVo system as a versatile and efficient platform for diverse biological research and clinical applications, contributing to advancements in cell-based therapies, disease modeling and precision medicine.

Future perspective

Looking ahead, we anticipate that the field of cell isolation and expansion will witness exciting advancements in the next 5–10 years, driven by the integration of artificial intelligence (AI) and the utilization of patient-derived cells. AI-powered algorithms have the potential to revolutionize the identification and isolation of specific cell populations, enabling more precise and efficient cell-sorting processes. By leveraging AI, we can expect automated and intelligent decision-making processes that optimize the isolation and expansion of patient-derived cells, leading to improved experimental outcomes and clinical applications.

Furthermore, the increasing utilization of patient-derived cells will provide a more relevant and personalized approach to biomedical research and therapy development. By incorporating patient-specific cells into our Enrich TroVo system, we can create disease models that closely resemble the individual patient's physiology, enabling better understanding of disease mechanisms and more accurate evaluation of therapeutic interventions. This personalized approach will empower researchers and clinicians to tailor treatments based on the unique characteristic of each patient, leading to improved therapeutic outcomes and reduced side effects.

In conjunction with AI, the integration of advanced imaging techniques and high-throughput analysis will further enhance the capabilities of our TroVo system. Real-time imaging and analysis of cellular behavior within microwells, combined with AI algorithms, will enable the identification of subtle phenotypic changes and the discovery of novel cellular subpopulations. This comprehensive analysis will contribute to a deeper understanding of cellular heterogeneity, disease progression and treatment response, paving the way for personalized and precision medicine approaches. We anticipate that these advancements will revolutionize the field, enabling personalized therapies, accelerating drug discovery and ultimately improving patient outcomes. We are excited to witness the transformative impact of these technologies in the years to come.