Uniform sized cancer spheroids production using hydrogel-based droplet microfluidics: a review

Abstract

Three-dimensional (3D) cell culture models have been extensively utilized in various mechanistic studies as well as for drug development studies as superior in vitro platforms than conventional two-dimensional (2D) cell culture models. This is especially the case in cancer biology, where 3D cancer models, such as spheroids or organoids, have been utilized extensively to understand the mechanisms of cancer development. Recently, many sophisticated 3D models such as organ-on-a-chip models are emerging as advanced in vitro models that can more accurately mimic the in vivo tissue functions. Despite such advancements, spheroids are still considered as a powerful 3D cancer model due to the relatively simple structure and compatibility with existing laboratory instruments, and also can provide orders of magnitude higher throughput than complex in vitro models, an extremely important aspects for drug development. However, creating well-defined spheroids remain challenging, both in terms of throughputs in generation as well as reproducibility in size and shape that can make it challenging for drug testing applications. In the past decades, droplet microfluidics utilizing hydrogels have been highlighted due to their potentials. Importantly, core-shell structured gel droplets can avoid spheroid-to-spheroid adhesion that can cause large variations in assays while also enabling long-term cultivation of spheroids with higher uniformity by protecting the core organoid area from external environment while the outer porous gel layer still allows nutrient exchange. Hence, core-shell gel droplet-based spheroid formation can improve the predictivity and reproducibility of drug screening assays. This review paper will focus on droplet microfluidics-based technologies for cancer spheroid production using various gel materials and structures. In addition, we will discuss emerging technologies that have the potential to advance the production of spheroids, prospects of such technologies, and remaining challenges.

Graphical Abstract

1. Introduction

For the past decades many in vitro models that can better mimic human physiology and functions have been developed in the fields of biology and medicine where better understanding of human physiology and pathophysiology as well as better in vivo-like models for drug discovery are needed. Traditional two-dimensional (2D) cell culture models have been used as standard technologies for many decades [1]. They are simple, user-friendly, and well-established, with many laboratory instruments available for automation and high-throughput testing. Despite decades of scientific advancements made through 2D cell culture approaches, they do not recapitulate the real cell environment in vivo, such as appropriate cell-to-cell communications and three-dimensional structures, and thus cells grown in 2D often do not have the functions or responses of those seen in vivo [2, 3]. In contrast, three-dimensional (3D) cell culture systems better mimic the 3D environment of human tissue in vivo with higher degree of physiological complexity compared to traditional 2D culture [4]. Spheroids (Fig. 1), 3D aggregates of cells, is one of the most widely utilized 3D cell culture models that can resemble in vivo-like structure and functions of tissues, especially as cancer models, displaying many histopathological characteristics seen in vivo [5–7]. The size of spheroids can vary from few tens of micrometers to few millimeters in diameter depending on the cell types, growth condition, and desired tissue models. These spheroids can easily be monitored through fluorescent or confocal microscopy due to their structural simplicity. Importantly, spheroids can provide several orders of magnitude higher throughput than most other 3D cell culture platforms [8], and thus ideal for high-throughput screening applications. Specifically in cancer research, spheroids can serve as superior tumor models in broad range of applications, including investigating carcinogenesis, tumor progressions, and developing anticancer therapeutics [9]. Spheroids can be created in mono-culture or co-culture formats with a wide range of cell types, and additional components such as extracellular matrix (ECM) and hydrogels are often added for better structural integrity. To form spheroids, groups of suspended cells must be placed together in 3D using some sort of physical structure, such as a round-bottom well plate, microstructures in microfluidic devices, or encapsulating in droplets, to form compact aggregates. In this rapidly advancing field, numerous spheroids production methodologies, such as ECM scaffold [10], spinning bioreactor [11], hanging drop [12], and magnetic levitation [13] methods have been developed and utilized [14]. However, most of these techniques have relatively low throughput. Microfluidic technologies that can provide accurate control, can automate labor-intensive processes, and can be high throughput, have been highlighted as a promising technology that can overcome such challenges [15]. Microfluidics have demonstrated its capability in performing many different biological assays with high precision and high throughput. Specifically, droplet-based microfluidics technology, which provides precise controlling over a small volume of biological reagents, chemicals, and cells can be used for rapid spheroids production with uniform sizes [16, 17]. Additionally, in terms of materials used for creating 3D structures using droplet microfluidics, hydrogel such as methacrylic gelatin (GelMA), polyethylene glycol (PEG), and alginate or agarose have been considered as promising 3D scaffolds for creating a tissue-like environment due to their flexibility and porosity. Here, gel scaffold supports long-term culture by allowing exchange of nutrients through the porous structure of gel scaffolds while holding the cells together in 3D format [18]. In this review, we will briefly discuss conventional 3D spheroid production technologies and current state of art in gel droplet microfluidics for producing uniformly sized cancer spheroids, followed by introducing two promising core-shell gel droplet generation techniques and their prospects in spheroid production and applications. Organ-on-a-chip (OOC) models are another promising in vitro 3D cell culture models. However, they are significantly more complicated, relatively low throughput, and difficult in utilizing for any high-throughput assays [19, 20]. Hence, despite its advantage, the low throughput make this technology for suitable for further downstream studies and be used as mechanistic models, while 3D aggregate-based in vitro models can be utilized in applications where high throughput is critical, such as for initial drug screening applications. Many reviews on OOC models have been published in recent years [21–23], and thus this review will not cover OOC models.

Fig. 1.

An overview of the structural and functional characteristics of spheroids, along with relevant applications and technologies used. Imaged created with BioRender.com.

2. Spheroids

Spheroids have been developed and widely utilized in many applications since their introduction in 1970s [24]. Once cells are seeded on culture substrates, they spontaneously aggregate as a group, forming spheroids [25]. Aggregation is influenced by multiple environmental factors including nutrients, oxygen, and presence of ECM [26]. However not all cells form spheroids spontaneously, necessitating the development of specialized techniques to force cells to form spheroids. Their structure and size uniformity are also influenced by the spheroid formation methods used, which in turn also determine in what applications they can be utilized.

2.1. Structure

Spheroids are typically formed with single cell type, but multiple cell types such as fibroblast or immune cells can also be mixed to allow co-culture within spheroids. Once aggregation starts, they turn their shapes from 2D clustering of cells to 3D spheres typically within few days. Depending on the cell type used, spheroids can increase beyond 500 μm in diameter, showing avascular tumor characteristics with an outer layer of proliferation zone and an inner necrotic zone due to different level of oxygen gradient and nutrient diffusion into the spheroids [27–30]. Shaping spheroids rely on several factors, such as cell type, density, substrate, and external stress or stimulation applied such as biomechanical cues [31]. As the culture progresses, spheroids maintain their size or be condensed rather than linearly expand in their size, and this occurs more with the presence of collagens or basement membrane. Understanding this growth dynamics of spheroid is often important when testing the efficacy of drugs in therapeutic development applications or in toxicity testing applications.

2.2. Uniformity

Spheroids are an excellent 3D model for recapitulating the tumor microenvironment, but non-uniform spheroids (mainly in terms of size) are undesirable in many applications, particularly for drug screening and discovery [32]. The presence of oxygen gradients and necrotic cores in spheroids due to size variations and imbalanced nutrient diffusion can lead to, for example, variations in drug penetration, resulting in uncertainty in assessing drug screening outcomes. Thus, the use of uniformly sized spheroids can help reduce batch-to-batch variability and poor quality control. Moreover, uniform spheroids can improve drug stability, bioavailability, and targeting specificity, which are all crucial for developing safe and effective pharmaceutical products [33]. However, achieving uniformity in spheroid size remains a significant challenge, as spheroid size may vary even under the same culture conditions [34, 35]. This is an even bigger issue when large number of uniformly sized spheroids are needed for high throughput screening applications, as generating uniformly sized spheroids at scale remains challenging.

2.3. Applications: Evaluation of drug & therapeutic efficacy

Evaluation of cancer chemotherapy or radiotherapy in 3D culture models that better represent the physiology and responses of cancer are becoming vital in cancer drug development. Recently, cancer spheroids have been widely utilized for new therapeutic screening and discovery applications. Measuring the size of spheroid along the peripheral edge is the most common evaluation method of therapeutic efficacy testing. As spheroids are exposed to drug, cell-to-cell junction of peripheral cells start to be disrupted and consequently results in reduced size and volume of spheroids in a dose-dependent manner [36–38]. Larger spheroids (> 300 μm diameter) have in general shown higher drug resistance than smaller spheroids [39]. From the spheroid compactness perspective, less densely aggregated spheroids are more sensitive to drug compared to more densely aggregated spheroids where drug diffusion to the core region is more challenging [40]. ECM in the spheroids also impact drug efficacy. Once cells are mixed with ECM components, they aggregate stronger, where spheroids become highly compact. Spheroids formed from multiple different cell types also impact how spheroids respond to therapeutics. For example, integration of stromal cells have been shown to affect drug resistance, as stromal cells support survival of cancer cells and induce higher drug resistance than mono-culture spheroids [41].

3. Conventional techniques for spheroid formation

There are many distinct techniques for spheroid formation based on the tendency of cells to aggregate or applying external factors that can facilitate cell aggregation. Here, we provide a brief overview on some of the most commonly used spheroid formation methods and discuss the advantages and challenges of each method.

3.1. Hanging drop

Hanging-drop technology was initially developed to culture nerve fibers [42], then adopted in various areas due to its simplicity and cost effectiveness. A small drop of cell suspension with a defined cell density is placed in a culture plate and then flipped 180°. This way, the drop retains its shape by surface tension while cells settle down to the bottom of the drop by gravity, and the cell aggregation process starts (Fig. 2a) [43–49]. One of the advantages using this method is its relatively high throughput in spheroid formation. For example, thousands of human glioblastoma (LN299) spheroids could be formed using a microfabricated device with an array of silicon microwells, with mineral oil used to avoid evaporation of medium under this flipped drop condition [50]. However, the hanging-drop culture could only be maintained for about 3 days due to the limited supply of medium to the spheroids, thus manual daily medium exchange is needed [50, 51]. As this technology has evolved, variations in this method are emerging to overcome these challenges [43, 52], yet fundamentally these challenges still remain.

Fig. 2.

Conventional methods for spheroid formation. a Hanging drop method. Flip over the culture plate with cell suspension drop and incubate cells on the bottom of the drop accumulated by gravity. b Liquid overlay method. Promoting spontaneous cell aggregation without cell adhesion on the plate surface using low-adhesive material. c Pellet culture method. Centrifuge cells in media suspension to pull down cells to the bottom of the tube to maximize cell aggregation. d Spinning culture / bioreactor. Spinning culture media or container with optimized speed that promote cell aggregation. e Magnetic levitation method. Magnetic nanoparticle-labeled cells in culture plate attracted by external magnetic field to form spheroid. f Free flow microfluidic (microwell) method. Cells fall into the microwells in a microfluidic culture chamber, aggregates formed within each microwell in a high throughput fashion. g Free flow microfluidic (U-shaped trap) method. Trapping cells in a U-shaped microstructure for cell aggregate formation within each microstructure under continuous media flow. Images created with BioRender.com.

3.2. Liquid overlay

Another approach to from spheroid is liquid overlay (Fig. 2b). This method can be performed using commercialized products such as multi-well plates, making this method simple and popular. In this method, cell suspensions are placed to low-adhesive surface plate [53–57] or plate coated with non-adhesive materials such as agarose [58, 59] or polyacrylamide [60, 61], which enable spontaneous cell aggregation without anchoring of cells. Constant rocking with shaker often assists the aggregation process as well. However, the biggest disadvantage is that size, shape, and aggregate formation varies depending on cell types as well as limited controllability of the aggregate formation process [62].

3.3. Pellet culture

Pellet culture utilizes centrifugal force to pull cells to the bottom of a centrifuge tube to maximize cell aggregation (Fig. 2c) [63, 64]. Cell aggregates formed by this method can be transferred to plates, coated with non-adherent materials, for further cultivation to complete spheroid formation. For example, periodontal ligament stem cell spheroids were formed by centrifugation at 180 × g in a conical tube [64], while 96 V-shaped non-treated well plates were also used to form human primary osteoblast spheroids for faster spheroid production [63]. However, due to the relatively large shear stress applied to cells during the centrifugation process, cells can be easily damaged during the process, negatively impacting the formed spheroids. Spheroids formed by this method are also usually large in size (typically > 800 μm) [65], and thus limited in their applications.

3.4. Spinning bioreactor

Cell spheroids can also be formed by a spinning bioreactor under optimized stirring condition (Fig. 2d). In this method, cells are suspended in media under a controlled rotating environment to promote cell aggregation while oxygen and nutrient are delivered [66–70]. Bioreactors (rotating or non-rotating) have also been utilized with continuous recirculation of the culture medium that provides dynamic culture with proper mixing. In this method, controlling shear stress is critical for cell lines that has low cohesiveness. For example, human breast cancer cell (MDA-MB-231) spheroids disaggregated when the shear stress reached 20 dyne/cm² [71]. For mass production of spheroids, large-size bioreactor can be used [68]. However, in this method the spheroid formation process cannot be visualized, lowering the process controlability, and the size distribution of the spheroids are also random [72], providing challenges to this method.

3.5. Magnetic levitation

Magnetic levitation is another method used for spheroid formation (Fig. 2e). In this method, cells are first labeled with magnetic nanoparticles and loaded on a plate with low-adhesive surface. When external magnetic field is applied to the culture plate, cells are attracted by the magnetic field, resulting in aggregate formation [73, 74]. In some applications, this method has been further utilized to engineer multicellular constructs, for example using murine embryonic fibroblast cell (NIH 3T3) and murine preadipocyte (3T3-L1) in different shapes and spatial arrangement [75] using this method’s high controllability. However, this method is often limited by its low throughput in spheroid production, as well as not always compatible with cells due to the addition of magnetic nanoparticles [76].

3.6. Free flow microfluidics

Conventional methods for spheroid formation are less flexible when modifying culture environment and less efficient in spheroid production, which hampers their broader use for many applications, including for drug screening. Microfluidics is capable of better controlling the culture environment, such as fluid dynamics and microenvironment as well as controlling spatial geometry [77–80]. Integration with electrical and mechanical components such as sensors for real-time monitoring or microvalves for sophisticated fluidic operation, enabling in situ analysis [81–84]. Thus, it has the potential to be used for improved spheroid production. For example, a microwell array has been used to generate cell spheroids and to supply culture medium continuously (Fig. 2f) [85–87]. In this method, cell suspension is first loaded into the microfluidic chamber that containing an array of microwells at the bottom, where cells are settled down into individual microwells while excessive cells outside of the microwells can be flushed away. Cells inside microwells can then be aggregated and forming spheroids, while continuously applying culture media supply is also easy [85–87]. In another report, a microfluidic device with integrated microvalves has been developed for multiplexed cancer treatment testing on primary ovarian cancer cell aggregates, where the device enabled parallel testing of several different drug concentrations [86]. This approach offers easy and simple use, however the relatively high flow rate required to retrieve spheroids formed within the microwells for further manipulations poses challenges. In another example, an U-shaped microwell structure array was utilized, with the openings positioned against the flow of cell suspension from the inlet (Fig. 2g). The U-shaped microwells act as a trapping site to capture cells when cells pass though the chamber, where the cells inside the trap form aggregates [88]. Culture media can then be continuously infused into the chamber at a controlled flow rate, which allow perfusion of nutrients throughout the spheroid culture. Another advantage is that the size of cell spheroids formed could be controlled by the size of the microstructures. For easier media exchange, the microfluidic device can be also integrated with a pneumatic control function to easily turn on and off the flow [89]. Table 1 summarizes several examples taken from a large literature data set in regard to spheroid formation.

Table 1.

Summary of conventional spheroid generation methods.

Sr. No	Methods	Platform	Flow condition	Advantage	Disadvantage	Cell type	Spheroid size (diameter)	Max. culture days	Ref.
1	Hanging drop	Silicon microwells partitioned by mineral oil	Static	Simple equipment, Low shear stress, good control over number of cells	Media evaporation, labor intensive	Human glioblastoma (LN229)	100 to 500 μm	9	[50]
		Universal 96-well plate	Static			Human colorectal cancer cell (HCT116)	1.5 mm	10	[51]
2	Liquid overlay	Ultra-low attachment 96-well round-bottomed plate	Static	Convenient, simple equipment, low shear stress	Variation in size and shape	Human brain glioblastoma (U-87 MG), Human glioma cell (KNS42), laryngeal squamous cancer cell (HN4), Human breast cancer cell (MDA-MB-231)	300 to 500 μm	14	[90]
3	Pellet culture	96 V-shaped noncoated well plate	Static	Convenient, simple equipment	Low reproducibility, relatively large spheroid size, not suitable for shear stresssensitive cells	Human primary osteobla-sts (hOBs)	300 to 400 μm	7	[63]
4	Spinning bioreactor	75 ml polycarbonate culture chamber integrated with buoyant vortices	20 to120 mL/min	More productive, long term culture	Not suitable for cells with low adherence, variation in size, no real-time monitoring	Human lung cancer cell (Calu-3)	200 μm	5	[68]
		Rotating wall vessel with bubble capturing	10 rpm			Human lung cancer cell (A549)	200 to 300 μm	3	[67]
5	Magnetic levitation	Magnetized cells with 40–50 nm Fe3O4 nanoparticles	Static	Faster spheroid formation	Requires magnetization of cells	Bone marrow human mesenchymal stem cell (hMSC)	200 μm	10	[73]
		Magnetized cells with 4.5 μm carboxylate magnetic particles	Static	High controllability in size and shape		Embryonic mouse fibroblast cell (NIH 3T3 & 3T3-L1)	800 μm	10	[75]
6	Free flow Microfluidics	PDMS microwells treated with Pluronic F-127 & 2% Matrigel	Static	High throughput, high reproducibility, accurate control of number of cells in spheroids	Complicated device fabrication, difficult to operate, difficult for spheroid harvesting	Ovarian cancer cell	160 μm	14	[86]
		PDMS U-shaped trap	0.05 to 10 μL/min			Human breast cancer cell (MCF-7)	50 μm	1	[88]

4. Droplet microfluidics for high-throughput spheroid production

3D cell culture models have begun to be integrated into many conventional cell culture and screening systems, such as multi-well plates and robotic liquid handling systems, necessitating ways of generating 3D spheroids in high throughput. Droplet microfluidics enables encapsulation of cells within water-in-oil emulsion droplets and forming spheroids with these droplets, and can do so with high precision and high throughput, leading to rapid and cost-efficient drug screening and development [16, 91–96]. Importantly, droplet microfluidics enables generating uniformly sized spheroids. In this section, we will discuss several commonly utilized droplet microfluidics-based single-layered spheroid generation technologies and their applications. Table 2 summarizes all conditions and factors that contribute to spheroid formation. Core-shell (also called multi-layer) droplet microfluidics-based spheroid formation will be discussed in Section 5.

Table 2.

Summary of spheroid formation techniques that utilized both single core and core-shell structured droplet generations.

Sr. No	Platform	Type of droplet	Droplet composition	Polymerization	Cell type	Spheroid formation time (days)	Max. culture days	Spheroid size (diameter)	Application	Ref.
1	Flow focusing generator	Single core	N/A	N/A	Human brain glioblastoma (U-87 MG)	2	7	98.6 to 126.4 μm	Photothermal therapy (PTT) with rGO-BPEI-PEG nanocom posite	[16]
2	Tree-branched gradient droplet generator	Single core	N/A	N/A	Human breast cancer cell (MCF-7) / Embryonic mouse fibroblast cell (NIH-3T3)	1	5	200 μm	Drug screening (doxorubicin)	[99]
3	Glass capillary generator	Single core hydrogel	0.1% HLECM + 5% Dextran-tyramine (Dex-Ta) + 250 U/mL horseradish peroxidase (HRP)	Enzymatic crosslinking with 30% H2O2	Intrahepatic cholangiocyte (IC) / cholangiocar cinoma (CCA)	7	21	112 μm	Drug screening (gemcita bine & cisplatin)	[104]
4	Flow focusing generator with docking array	Single core hydrogel	2% Sodium alginate	Ionic crosslinking	Human breast cancer cell (MCF-7 & -7R;doxorubicin resistant) / bone marrow stromal fibroblast (HS-5)	1	4	200 μm	Drug screening (doxorubicin & paclitaxel)	[105]
5	Flow focusing generator with magnet	Single core hydrogel	0.5% Sodium alginate	Ionic crosslinking	Human cervical carcinoma (HeLa) / human embryonic kidney (HEK293)	3	5	150 μm	On-chip oil rinsing & co-culture effect on spheroid formation	[94]
6	Flow focusing generator with cell scattering loop	Single core hydrogel	2% Sodium alginate	Ionic crosslinking	Human cervical carcinoma (HeLa)	2	6	138 μm	Effect of ECM on spheroid formation & drug screening (vincristine)	[106]
7	Double flow focusing generator	Single core hydrogel	Methacrylic gelatin (MGel)	Photo polymerization	Human breast cancer cell (MCF-7) / Macrophage (RAW264.7)	3	14	100 μm	Stiffness & co-culture effect on spheroid formation	[132]
8	Flow focusing generator	Core-shell hydrogel	Core: Mixture of hydrogel (collagen, Matrigel, alginate) Shell: 2% Sodium alginate	Ionic crosslinking	Human breast cancer cell (MCF-7)	4	8	135 to 235 Lim	Effect of ECM on spheroid formation & drug screening (tamoxifen & docetaxel)	[113]
9	3D (non-planar) flow focusing generator	Core-shell hydrogel	Core: 0.1% Alginic acid + 1% Carboxymethyl cellulose / Shell: 0.5% agarose	Core: Ionic crosslinking, Shell: temperaturmediated crosslinking	human embryonic kidney (HEK 293T)	4	14	70 μm	Stiffness effect on spheroid formation & cell-to-cell communication (JAG1-NOTCH signaling by xenografting)	[112]
10	Flow focusing generator	Core-shell hydrogel	Core: DMEM / Shell: 1, 1.5 and 2% Alginate + Ca-EDTA /	Core: n/a, Shell: ionic crosslinking	Human breast cancer cell (MCF-7) / Human mammary fibro blast (HMF)	2	7	70 to 180 Lim	Drug screening (paclitaxel & curcumin)	[102]
11	Flow focusing generator	Core-shell hydrogel	Core: 2% methyl cellulose / Shell: 8, 12 and 16% GelMA	Core: n/a, Shell: photo polymerization	Human liver cancer cell (HepG2) / human umbilical vein endothelial cell (HUVEC)	7	15	120 to 300 Lim	Method development - characterization of Core-shell droplet generation	[107]

4.1. Single-layer water-in-oil emulsion droplet techniques

Water-in-oil emulsion droplet is the most common form of droplets that can hold cells with supporting culture medium separated from the surrounding carrier oil. To form monodispersed cell-laden droplet, co-flowing, T-junction flow-focusing, or pinched flow-focusing geometries have been utilized, where the flow rate of aqueous and continuous phase as well as fluidic channel geometry can be adjusted to achieve a certain size/shape of spheroids [97, 98]. For example, Lee et al. used a flow-focusing microfluidic device to generate 3D brain tumor spheroids in water-in-oil emulsion droplets to evaluate the photothermal effect (PTT) of nanocomposites on tumor spheroids (Fig. 3a) [16]. Here, human brain glioblastoma cells (U-87 MG) were encapsulated in droplets and the generated cell-laden droplets incubated for 2 days for spheroid formation. After incubation, spheroids were replated and treated with reduced graphene oxide-branched polyethyleneimine-polyethylene glycol (rGO-BPEI-PEG) nanocomposites (60 μg/mL) for 4 h and then exposed by near infrared (NIR) laser with 1 W/cm² dose to activate the nanocomposites. After treatment, viability of brain tumor spheroid was reduced from 91% to 55%, indicating that nanocomposite can reduce cancer cell proliferation and metastasis. In another report, a tree-like droplet generator was introduced for drug screening (Fig. 3b) [99]. Human breast cancer cell (MCF-7) and embryonic mouse fibroblast (NIH 3T3) cells were co-encapsulated with various mixing ratio through flow gradient, thus homotypic (only MCF-7 or NIH 3T3) or heterotypic (20 to 80% fibroblast cells) spheroids could be generated, confirmed through fluorescent expression analyses. This represents a more realistic tumor microenvironment where stromal cells occupy a major portion. After 5 days of incubation, spheroids were treated with a standard chemotherapeutic drug (2 to 10 μg/mL of Doxorubicin), resulting in stromal cell-mediated drug resistance. In these systems, microfluidic droplet generator platform demonstrated mass production of spheroids with high generation speed. Nevertheless, these systems only allow short-term spheroid culture (< 7 days) since the spheroids are encapsulated within a water-in-oil emulsion droplet environment, limiting nutrient availability.

Fig. 3.

Various single-core droplet generation techniques for spheroid formation. a Glioblastoma tumor spheroid formation for photothermal (PTT) effect study using nanoparticles. b Homotypic or heterotypic spheroid generation with MCF-7 and NIH 3T3 via tree-shaped gradient flow focusing droplet generator for chemotherapeutic study. c Generation of hybrid hydrogel droplet with intrahepatic cholangiocyte using a co-flow glass capillary droplet generator for drug screening. All figures are reproduced with kind permission by authors.

4.2. Single-layer hydrogel droplet techniques

Hydrogels can be employed for mono- or co-culture of cells in a droplet format, and have been utilized with different combination of crosslinking strategies to create cell-encapsulated hydrogel droplets. Hydrogels are biocompatible and can provide higher relevance to the natural in vivo environment, resulting in good cell proliferation and spheroid formation environment [100–103]. For example, an intrahepatic cholangiocyte (IC) and patient-derived cholangiocarcinoma cells (CCA) were encapsulated in hydrogel droplets through a co-flowing glass capillary microfluidic droplet generator to demonstrate its capability for a high throughput cancer drug assessment assay (Fig. 3c) [104]. In detail, a fused silica capillary (200 μm inner diameter) was used to fabricate the nozzle that is inserted into a semi-permeable silicone tubing having a borosilicate capillary spacer. For polymerization, 5% Dextran-tyramine (Dex-Ta) was mixed with 0.1% human liver-derived ECM and 250 U/mL horseradish peroxidase (HRP) as hydrogel precursor, where the diffusion of H₂O₂ into the precursor droplets resulted in enzymatic crosslinking. The polymerized droplets were then washed through N-hexadecane and phosphate buffer silane (PBS). Although the human liver-derived ECM improved the formation of spheroids, there was high variability in cell-to-cell junction, thus non-uniformly sized spheroids were created, which eventually led to inefficiency in drug screening applications. Sabhachandani et al. introduced a more controlled single-layer hydrogel droplet for spheroid production [105]. In this system, the droplet generator was composed of a T-junction geometry, with a perpendicular cross linker channel and a spheroid docking area that contains 1,000 spheroid docking sites for drug screening assays. Alginate was used, which provided robust mechanical strength for maintaining cellular functions. To crosslink alginate, the ratio of calcium ions need to be precisely controlled when mixed with alginate, where the degree of crosslinking significantly affects nutrient diffusion into spheroids forming inside the hydrogel droplets. In this report, MCF-7 was suspended in 2% w/v alginate solution and encapsulated by mineral oil with 3% w/v surfactant, then loaded into the docking sites individually, which avoids the fusion of multiple hydrogel droplets and allowoing single-spheroid-resolution drug testing. Once droplets are settled down into the array of docking sites, the calcium chloride fluid channel was opened for calcium chloride to reach the alginate droplets. Despite the on-chip spheroid generation, crosslinking, and drug screening, spheroid size was still not uniform and crosslinking took too long due to the long travel distance between the calcium chloride inlet and the large area of docking sites. Additionally, since the spheroids are still covered by oil, incubation and nutrient supply were limited as well. To improve this, Yoon et al. suggested an on-chip washing technology by using magnetic nanoparticle (MNP) [94]. In this work, human cervical carcinoma (HeLa) cells and MNPs were mixed with alginate before flown to the flow-focusing channel, while being separated from the calcium chloride flow by a stream of culture medium in the center to avoid instant diffusion and crosslinking. In the downstream part of the device, a neodymium magnet was placed, 1mm away from the culture medium flow channel, to induce magnetic attraction. Once droplets were attracted by the applied magnetic force, cells could break through the culture media / oil boundary, enablig the spheroids to be transferred into water phase from oil phase, and then collected. This process eliminated additional process of spheroid production, significantly reducing the deformation of gel droplet by centrifugal force and loss of droplets. However, this magnetic attraction process and the location of a spheroid inside an alginate droplet is highly depending on the concentration of MNPs, which led to non-uniform shape in the formed spheroids. When considering 3D cell culture, selection of biomaterials is vital as cells will behave differently depending on their microenvironment. Mixed hydrogel with natural components, such as basement membrane or collagen, reconstruct better microenvironment to cells and eventually support in vivo-like inherent characteristics. Wang et al. utilized gel mixture and a double flow-focusing droplet generator to improve the location and shape of core [106]. Here, alginate and Matrigel were mixed with different volume ratio by controlling flow rate for encapsulating HeLa cells, and demonstrated better spheroid formation than using pure gel material based droplets due to the matrix environment which is typically seen in vivo by ECM.

5. Core-shell structured hydrogel droplets for uniformly sized spheroid production

Although single layer hydrogel droplets mimic physiological environment and enable cell proliferation and long-term cultivation, there are many difficulties in uniform sized spheroid production due to the external alterations such as shear force and difficulties in unwanted stiction between spheroids that leads to sample loss [107]. Core-shell structured droplets are composed of two different materials that forming distinct layer separation. Recently, core-shell structured droplets have attracted a huge attention from researchers due to their structural and chemical properties and high flexibility in choosing material based on the research purpose [108–110]. Given such advantages, core-shell hydrogel droplet offers better protection of spheroids from surrounding environment while allowing nutrients perfusion and minimized loss of samples in controlled manner due to the outer porous shell layer of droplet [102, 110–113]. Core-shell hydrogel droplets are promising tool for a broad range of applications such as targeted drug delivery [110, 114–116], cell biology [117, 118], bio imaging and sensing [114, 119, 120], and bio printing [111]. For example, they can encapsulate therapeutic molecules in the core area and be delivered to the desired location in human body. Through a designed triggering method, the contents in the core can be released and activate its biological or chemical reaction at the targeted organ or tissue. In this case, the shell layer protects the contents inside core area from unexpected triggering [110].

5.1. Materials of core-shell droplets

Core-shell droplet can be engineered by various materials for core and shell layer, respectively. The core can be liquid, gas, or solid depending on the purpose and applications [108, 109]. For example, liquid core is often utilized for drug delivery and catalysts applications due to the high controllability of contents release. Drugs or catalyst in the core retain their functions until releasing at targeted location [121, 122]. Furthermore, liquid core enables isolation of pluripotent stem cells and allows them to maintain its stemness under highly viable condition thanks to the high efficiency of nutrient diffusion and oxygen exchange [123]. Recently, gas phase core or purely hollow core have been highlighted due to the high surface area and superior optical property [124]. Core-shell droplet with gas phase core provides controlled release of cargo, such as cosmetics, inks, and dye thereroe they can be applied in cosmetic, paint and food industries [125]. However, gas phase is not suitable material for cell encapsulation due to the lack of culture medium. Among all materials, solid core, especially polymer, is the most common material for cell culture and spheroid formation as it provides tissue-like environment thus enables better cell proliferation with sufficient nutrient and oxygen supply through the porous structure. Solid core initially starts as liquid but solidified based on utilized platform or cell types through various solidification methods, such as photo polymerization, ionic crosslinking, and thermal polymerization. Taking advantage of it, organic polymers, such as gelatin, agarose, alginate, PEG, etc., can also be utilized as shell material of the core-shell droplet. It protects the cargo by surrounding the core and blocking from corrosion or penetration of particles [108]. Wang et al. used GelMA as a core material to entrap liver cells through photo polymerization and for 15-day culture in the core [107]. This proves that polymeric core immobilize the cells and consequently providing better support to cell aggregation than the liquid core. Other polymeric materials, such as polystyrene [126], poly lactic-co-glycolic acid (PLGA) [127], isotactic polypropylene (iPP) [128], and silica [129, 130] are also utilized as solid core material. In this review, organic polymers that contribute to the cancer spheroid formation will be discussed.

5.2. Generation of homogeneous sized spheroid

The formation of uniform-sized hydrogel droplet is directly connected to homogeneous spheroid production. For the past few decades, research related to cancer spheroid formation and their use in drug screening applications have highlighted the importance of uniformly sized spheroids, where heterogenicity gives rise to different levels of hypoxia within the spheroids and consequently leads to large variations in drug responses. To mitigate these challenges, Kim et al. [131] and Yu et al. [113] have suggested core-shell structured cell-encapsulated hydrogel droplet generation as a means to create more uniformly sized spheroids. Yu et al. mixed MCF-7 with alginate and two different ECM components (Matrigel and collagen) for more uniformly sized spheroid formation. To avoid a rapid gelation, the core solution containing cells and Matrigel is chilled in a bath maintaining 4°C. In the shell, alginate solution is mixed with calcium carbonate (CaCO₃) and flow to embrace the core at the first flow focusing channel. Then, droplets are exposed to the mineral oil mixed with acetic acid. Acetic acid then diffuses into the generated droplet to lower the pH and react with calcium ions within the alginate shell leading to gelation. To ensure gelation of shell layer, they put additional inlet for acidified mineral oil at the downstream of second flow focusing channel. This study demonstrates two different gelation methods (thermal polymerization and ionic crosslinking) to create core-shell structure with high mechanical stiffness. Additionally, they tested multiple combination of ECM components, agarose, and cells. By looking into the cell proliferation, Matrigel + alginate + collagen mixture showed the best spheroid formation indicating better tumor environment. However, temperature control-based gelation is slower than instant crosslinking, such as photo polymerization thus cells are randomly distributed until 4 days incubation which makes relatively less homogeneous spheroid formation. In other reports, similar core solidification method has demonstrated (Fig. 4a and 4b) [102, 112] with stromal cells (fibroblast) mixed in the shell region mimicking tumor physiology and modeling cell-to-cell interaction. Sun et al. investigated the anticancer drugs response using paclitaxel and curcumin through higher throughput core-shell droplet generation (>200 spheroids/min). Zhu et al. suggested two-step crosslinking approach through 3D nonplanar droplet for central positioning of core, thus creating more uniform sized spheroid within core-shell droplet. In this report, a mechanical property of shell material has been evaluated and concluded that high stiffness of gel material showed high circularity of formed spheroid. Although alginate-calcium provide gelation through ionic bonding, it is still difficult to precisely control the degree of the diffusion of calcium ion, thus positioning of cells in the core region remains challenging. Lee et al. utilized two different concentrations of gelatin methacryloyl (GelMA) that provides much higher mechanical stiffness than alginate and better controllability on generating well defined structure, to create MCF-7 spheroids for investigating the effect of stiffness on cell viability and cell proliferation through co-culturing the spheroids with macrophage or fibroblasts (Fig. 4c) [132]. A double flow-focusing microfluidic device was utilized to produce spherical microgel droplets supporting 3D culture of encapsulated breast cancer cells, followed by 2 min ultraviolet (UV) crosslinking at 200 mW/cm². Specifically, two aqueous phase channels were used to form the soft core and harder shell layers, surrounded by mineral oil with surfactant. The core layer is composed of cancer cells with supporting cells while the outer layer contains stiffer gel material than the core, allowing uniformly sized spheroid formation. From culturing the spheroids in gel materials with different stiffness, it was found that the formation of spheroids can be accelerated under higher mechanical stiffness. In another report, Wang et al. utilized GelMA and methyl cellulose (MC) to make core-shell droplet (Fig. 4d) [107]. In the core, human liver cancer cells (HepG2) were mixed with 2% methyl cellulose (MC) and encapsulated through flow focusing droplet generator. MC is biocompatible material and widely used in cell culture applications as it supports long-term cell incubation. Moreover, high viscosity of MC solution guarantees formation of laminar flow with GelMA solution that showing ‘sandwich-like’ configuration in the upstream of droplet generation junction. By changing each flow rate, the thickness of shell could be controlled that result in different size of spheroid. Importantly, forming the circular core lumen highly depends on geometry of droplet generator, such as, length of traveling channel as well as flow rate because the core solution flux is key component of creating uniform sized core shape.

Fig. 4.

Core-shell droplets for advanced spheroid formation and culture. a 3D non-planar double focusing droplet generator that produces better encapsulation compared to 2D planar structure. b Tumor and stromal fibroblast co-culture system with alginate core-shell structure for high throughput drug screening. c MCF-7 spheroid formation with double-concentrated hydrogel for studying the effect of stiffness on spheroid formation d Generation of core-shell droplets with higher mechanical stiffness of shell structure for better core formation. All figures are reproduced with kind permission by authors.

6. Proposed systems for advanced core-shell hydrogel droplet generation

Although the well-established platforms and techniques have been reported in the field of anticancer drug development and screening, there are still necessities in improvement of core-shell droplet generation that can further decrease operational drawbacks and increase experimental efficiency, for example higher throughput production, well-defined spatial separation of core and shell region, and downstream droplet manipulation. Here, new core-shell hydrogel droplet production methods are proposed. More specifically, pre-polymerized single layer hydrogel droplets are generated through typical T-junction droplet generator, then collected for UV polymerization followed by oil rinsing process through centrifugation (Fig. 5a). Afterwards, they were re-introduced to the secondary droplet generator that consist of double flow focusing channels. In this device, different type of hydrogel precursor solution embraced the re-introduced pre-polymerized droplet forming double layers after polyermerization (Fig. 5b). Since the core is pre-polymerized, two different layers are well-confined, suggesting potential coculture of two different types of cells. Hence, cells can be secured only in the desired region which provides improved positioning of cells and consequently lead to uniform sized spheroid formation and monitoring of the formation. Another proposed method utilizes an interdigitated electrodes (IDE or IDT) [133, 134] that significantly increase a throughput of core-shell droplet generation. In this method, pre-polymerized cell-laden droplets are re-introduced to secondary droplet merging device without oil rinsing process that can reduce entire operation time and avoid unwanted droplet loss. In the device, pre-polymerized single layer hydrogel droplets are spaced out by oil spacing, then flow down individually to the angled junction. From another inlet, different type of hydrogel droplets (also described as non-polymerized droplets) are generated and then paired (1:1) with pre-polymerized single layer hydrogel droplet, controlled by oil spacer (Fig. 5c). In droplet microfluidics, oil spacing technique is commonly used and significantly improve droplet manipulation efficiency such as sorting and merging [133]. After paring, the paired two droplets are travelled to downstream of the device and passed by the merging window, located on the top of IDEs pattern. When voltage is applied, IDE induced an electric field that destabilize the surrounding oil of droplets [135], then the non-polymerized droplet eventually become a shell layer by embrassing the core droplet. Importantly, by utilizing the IDE, highly localized droplet merging is possible compared to other types of electrodes [136] and this eventually makes more efficient core-shell droplet generation and spheroid production in fully controlled manner and this could potentially allows automated anti-cancer drug screening system. Core-shell droplets produced by the two proposed methods will be exposed to secondary UV light irradiation after collection in culture plate to complete the double layered hydrogel droplet structure without harmful effect by double irradiation (Fig. 5d).

Fig. 5.

Proposed microfluidic systems for advanced core-shell droplet generation. a Cell encapsulation in hydrogel precursor via T-junction droplet generator and photo crosslinking at downstream by UV light. b Reflow of washed single layer hydrogel droplet and re-encapsulation by secondary hydrogel precursor. c Reflow and pairing of non-washed pre-polymerized single layer hydrogel droplet and merging with another hydrogel droplet via IDE. d Secondary UV photo polymerization of core-shell droplets produced by both proposed methods. Images created with BioRender.com.

7. Challenges and prospects of proposed systems

Although the proposed models suggest significant improvement in cancer spheroid production, there are still some spaces to modify. 1) During the reflow of pre-polymerized droplet after oil rinsing, there will be surface frictional force in between droplet and glass substrate. In this case, droplet reflow will be out of control, and affect the re-encapsulation rate. To resolve this, different substrate that made of plastic, for example polystyrene can be utilized to reduce the frictional force. 2) Another potential drawback is double UV exposure. Both proposed techniques require two polymerization processes during entire operation for solid gel structure. Therefore, cells in the core might be over-exposed by UV light which can be harmful to certain cell type [137, 138]. To mitigate harmful effect, hydrogel concentration and polymerization conditions should be optimized. 3) As typical microfabrication does, it requires specialized facility for fabricating micro-scaled device and precise droplet manipulation. 4) Depending on the hydrogel type, the effect of electric field can vary, thus merging two droplets made of different materials will behave differently. Following the successful development of the proposed models, a fully automated spheroid manufacturing system would be the next step. Such automated model can include all aspects of droplet manipulation (e.g., generation, on-chip rinsing, on-chip crosslinking, on-chip incubation, merging, and sorting and test the efficacy of potential anticancer drugs). Recently, advances in robotic control and microfabrication technology offer ultra-throughput production and accurate analysis of targets by accelerating preclinical and pharma therapeutic research [139–141]. Importantly, by employing such technologies, development of novel therapeutics against cancer will be highly advanced with significantly reduced cost and time.

8. Conclusion

A 3D form of cancer cells, called spheroids, represent an advanced cancer physiology. Although the conventional methods have been standardized through realiable outcomes, higher throughput with consistent antidrug responses has limited due to the low controllability, resulting in depleted understanding. Through droplet-based microfluidic technology, recent studies could obtain cancer spheroids in controlled manner which contributing improved drug screening. Notably, core-shell structured droplet produces superior spheroid model and enables highly improved drug testing by having uniform sized spheroids in the restricted region which allowing better understanding of cancers. Yet not many studies utilizing core-shell structured hydrogel droplet for cancer spheroid formation have been report. Hence, the proposed technologies could significantly accelerate this area by suggesting highly advanced spheroid production strategy. As more advanced spheroid production and manipulation technologies emerge, we expect drastic changes in clinical trials and therapeutic research related to cancer.