1. Introduction
The cultivation of cells in a laboratory is an essential and fundamental tool used across a wide array of scientific research fields. Traditionally, adherent cells have been cultivated as a monolayer, since protocols for the proper handling and analysis of such cultures are well-established [
1]. However, in recent years it has become increasingly apparent that traditional monolayer cultures simply cannot be induced to mimic the three-dimensional (3D) environment that cells are exposed to in vivo—resulting in critical differences between 2D and 3D cell cultures in terms of their morphology, differentiation, protein expression, functionality, migration, apoptosis, and response to drugs [
1,
2,
3,
4,
5,
6]. As a result, 3D cell cultures are increasingly being favored by researchers due to their superior physiological relevance [
7].
Today, 3D cell cultures are often achieved by using various scaffold-free techniques—for example, via low-adhesion plates which promote the aggregation of cells into spheroids, or via extracellular matrix (ECM) coated plates that facilitate cell differentiation into organoids [
7]. The multilayers of cells obtained through these various mechanisms reflect the in vivo conditions more accurately than traditional 2D cultures, however, they are comparatively much more variable in terms of their size, shape, and composition. Furthermore, tracking and analyzing specific cells over an extended period of time becomes a far more difficult task, due to their potentially shifting position within this complex cell construct [
7,
8].
In vivo, however, most mammalian cells proliferate while surrounded by an ECM which forms a complex scaffold consisting primarily of hydrated proteins [
1]. Taking a cue from nature, scaffold-based 3D culture approaches use fabricated 3D structures to imitate this ECM. Rigid scaffolds (for example, those fabricated from paper or fibrils) need to be laboriously synthesized and manufactured—and even once they are manufactured, seeding cells evenly and homogeneously within these structures often presents a substantial challenge for researchers. These issues make it difficult to truly standardize results between different experiments [
7]. Furthermore, such rigid scaffolds are only suitable for certain types of tissue; they are patently unsuitable for use in myocardial tissue engineering, for example [
9].
Hydrogels are able to mimic this natural ECM system by absorbing high amounts of water, and they also have the benefit of providing viscoelastic strength while facilitating a far more uniform distribution of cells [
7,
9]. Perhaps not surprisingly, there are now many materials available to researchers looking to “fine tune” the porosity, stiffness, and/or degradation of a hydrogel in order to optimize cell proliferation and better mimic organ-specific ECMs [
6,
10,
11]. However, cell structure, function, and behavior are not merely influenced by the 3D arrangement of the cells—they are also influenced by the complex combination of biochemical, physical, and physicochemical properties (e.g., soluble factors, pH, oxygen supply, temperature and osmolality) which form the microenvironment in which the cell develops [
12]. Exercising granular control over this cell microenvironment is thus of critical importance in many cases—particularly as it has been shown that (for example) stem cell development is influenced by this microenvironment and abnormal levels of pH and oxygen tension are associated with the development of various pathologies [
12]. Furthermore, gradients in the cell microenvironment can act as signals influencing the regulation of cell function and behavior [
12,
13].
Microfluidic cell culture devices enable good control over the microenvironment of the cells and even the implementation of stable gradients of various forms in physiologically relevant scales [
12]. Furthermore, microfluidic devices enable tissue-tissue communication, dynamic fluid flow, and the application of normal mechanical stimuli/cues [
14]. 3D cell cultures within microfluidic devices can therefore even be used to mimic organs in their complex microarchitecture and function. These organ-on-chip (OOC) devices can, for example, be applied to improve the understanding of human drug metabolism and toxicity in vitro prior to the initiation of clinical trials [
14,
15,
16,
17]. Such OOC devices have, inter alia, already been successfully implemented for human livers [
15], lungs [
18], and even for modeling specific diseases such as virus-induced kidney disease [
19] and an infected epidermis model, among others [
20].
There is a downside, however: the fabrication of microfluidic devices using conventional methods can be very challenging, time consuming and expensive. One way to overcome these challenges is 3D printing, which is becoming increasingly popular within the field of biotechnology [
21,
22,
23,
24,
25]. 3D printing not only offers researchers the ability to engage in rapid prototyping, but also permits the fabrication of highly customized complex structures [
21]. It enables easy chip-to-world interfacing (such as 3D-printed Luer-lock-systems) for medium supply, as well as the use of removable support materials to facilitate the fabrication of overhanging structures and cavities (such as hollow microfluidic channels) [
21,
26].
As researchers continue to push for ever-more-realistic in vivo-like conditions within the laboratory, growing ambitions to construct more complex microenvironments for cell cultures will undoubtedly continue to spur on the development of novel hydrogels and the integration of various cell types. This will in turn create an ever-growing demand for screening systems that allow for more reliable evaluations of different cultivation conditions—including hydrogel compositions, cell densities, media supplements, oxygen concentrations, and other crucial factors which influence the cellular microenvironment. The 3D-printed microfluidic perfusion system presented in this study introduces a novel solution for the integration of hydrogels and parallel cultivation of four separate 3D cell cultures. By using a porous membrane for separation of the hydrogel compartment from the microfluidic perfusion channel system, the hydrogel of each chamber is constructed to be of an equal thickness and is protected from flow-induced detachment. In addition, this design also provides for comparatively easy monitoring of the cells. Finally, due to the high degree of customizability offered by additive manufacturing, this microfluidic cultivation system can be rapidly scaled to a multi-chamber device or a parallel perfusion system in a single device.
2. Materials and Methods
2.1. 3D Printing and Post-Processing of Cultivation Device and Perfusion System Parts
After computer-aided design (CAD) using SolidWorks 2022 (Dassault Systems Deutschland GmbH, Stuttgart, Germany), the 3D-printed parts of the cultivation device and perfusion system were fabricated using a high-resolution 3D printer AGILISTA-3200 W (Keyence Deutschland GmbH, Neu-Isenburg, Germany) which manufactures objects via inkjet technology using an ultraviolet (UV) curing process—resulting in a resolution of 635 × 400 dots per inch and a layer thickness of 15 µm [
27]. The clear polyacrylate 3D printing material AR-M2 (Keyence Deutschland GmbH, Neu-Isenburg, Germany) was used as model material, and AR-S1 (Keyence Deutschland GmbH, Neu-Isenburg, Germany) was used as support material during the printing process. In its cured form, the model material shows biocompatibility in accordance with ISO 10993:12 [
28]. Subsequent to the printing process, objects were scraped from the printing platform and the support material was removed via incubation in an ultrasonic water bath (Bandelin electronic, Berlin, Germany) for 30 min at 60 °C—twice with detergent (Fairy Ultra Plus, Procter and Gamble, Bethel, CT, USA), and then once more with ddH
2O (Arium
® Sartorius Stedim Biotech GmbH, Göttingen, Germany)). Small channels were thoroughly rinsed after every incubation step by attaching a cleaning syringe. Finally, the objects were incubated in ethanol (70%
v/
v; VWR International GmbH, Darmstadt, Germany) on a SSM3 gyratory rocker (Cole-Parmer Instrument Company Ltd., St Neots, UK) at 70 rpm for 1 h, thoroughly rinsed with ddH
2O, and then completely dried.
2.2. Assembly of the Cell Cultivation Device
The cultivation device consists of two 3D-printed parts separated by polyester membranes cut from Transwell®-Clear Inserts (pore size: 3 µm; Corning, Kaiserslautern, Germany), which are sealed via standard O-rings (Landefeld Druckluft und Hydraulik GmbH, Kassel, Germany; 6 × 1 mm, FKM). A transparent 0.25 mm thin polycarbonate sheet (Modulor GmbH, Berlin, Germany) was used as the bottom plate of the device. The polycarbonate sheet and the lower 3D-printed part of the system were bonded using an adhesive medical tape (3M 9877, 3M Medical Solutions Division, Healthcare Business Group, Neuss, Germany) and connected to the upper 3D-printed part using standard M2 metal screws and a custom-built metal frame. The adhesive medical tape was cropped using a cutting plotter (Cameo 4, Silhouette America, Inc., Lindon, UT, USA).
2.3. Assembly of the Perfusion System
Perfusion of the cell cultivation device was achieved from a medium reservoir with an IP-4 peristaltic pump (Ismatec, Wertheim, Germany) using Tygon® pump tubing (IDEX Health and Science GmbH, Wertheim, Germany; inner Ø 1.22 mm), connecting standard chromatography PTFE tubing (Ø 0.8 mm, Bohlender GmbH, Grünsfeld, Germany) and fittings. A TubeSpin® Bioreactor (TPP Techno Plastic Products AG, Trasadingen, Switzerland) was used as a medium reservoir. A 3D-printed adapter was designed to connect the tubing to the TubeSpin® Bioreactor. For some experiments, a bubble trap with an internal volume of 97 µL (Darwin Microfluidics, Paris, France) was integrated into the perfusion setup and then used in passive mode.
2.4. Cell Line and Cell Culture Conditions
L-929 cells (CLS Cell Lines Services GmbH, Eppelheim, Germany) were routinely cultivated in Dulbecco’s Modified Eagle’s Medium (DMEM; Sigma-Aldrich Chemie GmbH, Steinheim, Germany), supplemented with 10% fetal calf serum (Sigma-Aldrich Chemie GmbH, Steinheim, Germany) and 1% Penicillin/Streptomycin (Sigma-Aldrich Chemie GmbH, Steinheim, Germany) within 175 cm2 cell culture flasks (Corning, CellBind Surface, Corning, NY, USA) in a 5% CO2 humidified atmosphere at 37 °C (Heracell 240 incubator, Thermo Fisher Scientific Inc., Waltham, MA, USA). At 70–85%, confluency, cells were harvested via Trypsin/EDTA solution treatment (Biochrom GmbH, Berlin, Germany). Experiments were performed with cells of passage numbers up to 10.
2.5. Hydrogel Preperation
For all experiments, an in situ cross-linkable alginate hydrogel described by Dahlmann et al. was used [
9]. Briefly, the alginate hydrogel formation is based on the spontaneous condensation of an alginate hydrazide and an alginate aldehyde. Alginate aldehydes were generated via oxidation of vicinal diols within the monomeric units according to the Malaprade reaction. The carboxylate functions of alginate were directly transferred into the corresponding acyl hydrazides using standard carbodiimide chemistry [
9]. Aldehydes were thoroughly dialyzed over a period of 3–5 days against distilled water, with a minimum repeated water exchange of three times per day. Hydrazides were dialyzed against 100 g·L
−1 sodium chloride (NaCl; VWR International BVBA, Leuven, Belgium) for one day, followed by 50 g·L
−1 NaCl for one day and two days against distilled water. Lyophilization was performed with a Christ Alpha 2–4 (Christ Osterode, Osterode am Harz, Germany) freeze dryer [
9]. To begin a cultivation, both lyophilized hydrogel precursors were each dissolved in a 0.9% (
w/
v) NaCl solution at 70 °C to a concentration of 1% (
w/
v) using a Thermomixer (Thermomixer comfort, Eppendorf, Hamburg, Deutschland) at 1000 rpm, and then sterile filtered (0.2 µm PES syringe filter, Filtropur S, Sarstedt AG & Co. KG, Nümbrecht, Germany). Subsequently, volume fractions of 40% collagen I solution from bovine skin (3 mg·mL
−1 aqueous solution in 0.01 M HCl, Sigma-Aldrich, Co., St. Louis, MO, USA), 5% 0.9% (
w/
v) NaCl solution and 5% 0.1 M sodium hydroxide (NaOH; Sigma-Aldrich Chemie GmbH, Steinheim, Germany) were mixed thoroughly with 25% sterile hydrazide-derivatives of alginate, and then kept on ice. Polymerization was initiated by adding 25% sterile alginate aldehyde and thereby obtaining a final gel concentration of 0.5% (
w/
v). Samples were thoroughly mixed and immediately transferred to the respective chamber or well. For cell-containing hydrogels, the required volume of cell suspension was centrifuged at 300×
g for 5 min (MiniSpin
® plus, Eppendorf SE, Hamburg, Germany), the supernatant was discarded, and the cells were then resuspended in the sterile alginate aldehyde solution to obtain a final cell concentration of 0.75 million cells per mL.
2.6. Microscopic Analysis and Live/Dead Staining
Microscopic imaging was performed using a Cytation 5 Cell Imaging Multi-Mode Reader (BioTek Instruments GmbH, Bad Friedrichshall, Germany) at 37 °C. Compatibility of the cell cultivation device with standard imaging systems was ensured using a 3D-printed adapter in well plate format, which was designed using SolidWorks 2022 and fabricated from polylactic acid (PLA) filament (1.75 mm, Das Filament Inh. Roman Stieben, Emskirchen, Germany) with a Prusa i3 MK3 (Prusa Research a.s., Prague, Czech Republic). For imaging of the cultivation device, it was first removed from the perfusion system in a sterile environment, closed with 3D-printed blind plugs, and finally placed on its adapter in well plate format and transferred to the imaging system. For brightfield imaging, the intrinsic auto-exposure function of the Gen5 imaging software (Version 3.10.06, BioTek Instruments GmbH, Bad Friedrichshall, Germany) was used with 4× and 20× objectives. Live/dead staining of the cultivated cells was performed with calcein AM (Merck Chemicals GmbH, DE, USA) and propidium iodide (PI; Merck KGaA, Darmstadt, Germany). Cells inside the cultivation device were stained for 2 h, and cells cultivated in well plates were stained for 20 min with calcein AM (1 μM) and PI (1 μg·mL−1) containing phosphate-buffered saline (PBS; Life Technologies Limited, Paisley, UK) solution at 37 °C in the dark. Subsequently, the staining solution was removed, and samples were covered with dye-free PBS solution. For image analysis, stitching, and channel overlay pictures, the intrinsic functions of the Gen5 imaging software were used.
2.7. Perfusion Experiments
Assembly of the cultivation device at the beginning of a perfusion experiment and handling of the adapter in well plate format for imaging of the hydrogel-embedded cells is shown in the
Video S5 in the supplementary material. Prior to an experiment, all tubing, fittings, and the metal screws/metal frame/sealings of the device were sterilized via autoclaving (30 min, 121 °C; Systec VX-150, Systec GmbH, Linden, Germany). The 3D-printed parts of the cultivation device as well as all parts of the bubble trap were disinfected via incubation in ethanol (70%
v/
v; VWR International GmbH, Darmstadt, Germany) on a SSM3 gyratory rocker (Cole-Parmer Instrument Company Ltd., St Neots, UK) at 70 rpm for 1 h, thoroughly rinsed with sterile ddH
2O (Arium
® Sartorius Stedim Biotech GmbH, Göttingen, Germany), and completely dried in a sterile environment. For some experiments, the whole perfusion system filled with 25 mL cell culture medium and the assembled cell cultivation device, without membranes, were placed in the incubator (Heracell
TM VIOS 160i CO
2 incubator, Thermo Fisher Scientific Inc., Waltham, MA, USA) overnight prior to beginning the experiment, in order to avoid the formation of gas bubbles inside the cultivation device during the cultivation process. To begin an experiment, the prepared cell cultivation device was disassembled in a sterile environment and every cultivation chamber was filled separately with hydrogel as described in
Section 2.5. Subsequently, the membranes were placed on the hydrogel and the cultivation device was reassembled. The perfusion medium chambers were filled with the cultivation medium via a syringe before microscopic imaging of the hydrogel embedded cells was performed as described in
Section 2.6. Afterwards, the cultivation device was connected to the perfusion system, placed in the incubator, and perfused with 0.25 mL·min
−1 cultivation medium for three days. As a control experiment, 50 µL of the same cell containing hydrogel were cultivated in 96-well plates (Sarstedt AG and Co. KG, Nümbrecht, Germany) with 150 µL cultivation medium, and cells from the same cell suspension were seeded in the 96-well plate at a density of 7500 cells per well in 200 μL cell culture medium. All control experiments were conducted in at least triplicate.
4. Discussion
The presented 3D-printed microfluidic perfusion system was successfully applied in a proof-of-concept cultivation of hydrogel embedded fibroblasts. The permeable membrane separating the hydrogel compartment from the microfluidic perfusion channel system prevented detachment of the hydrogel, while also facilitating adequate gas exchange and nutrient supply to the cells (as indicated by the microscopic analysis and the cell staining experiments described above). This design also permitted easy microscopic monitoring and staining of the hydrogel-embedded cells, and significantly reduced the required volumes of hydrogel and cell suspension. Since the cultivation device can be easily disassembled, the hydrogels were quickly recovered post-cultivation for further analysis. In principle, this property could even allow researchers to reuse this cultivation device.
To ensure a minimal focal distance in the presented design of the cultivation device, the hydrogel-embedded cells were supplied with perfusion medium from one side only. In the proof-of-concept cultivation of 0.75 million L929 cells per mL embedded in the alginate-based hydrogel presented by Dahlmann et al. [
9], no dependency of cell morphology, viability, or growth on the special distance from the perfusion medium chamber was observed. However, the formation of oxygen or nutrient gradients in the hydrogel resulting from the device design could be possible and should be taken into account when using the device. Gradients occur naturally within in vitro tissues of biological organisms, and therefore they may be desirable in some 3D cell culture applications [
12,
30]. Where such gradients are not desired for an intended application of the cultivation device, however, the course of the perfusion medium channels could instead be rapidly adjusted. The perfusion rate can, of course, also be quickly adapted to suit the purposes and parameters of an intended application.
In order to reduce manual intervention and simplify handling, additional efforts to further automate and miniaturize this system in the future are envisioned. The tubing, medium reservoir, and bubble trap elements could all potentially be fabricated as part of the 3D-printed microfluidic cultivation device—thereby decreasing the required amount of culture medium. This would also potentially mitigate the need for removal of the perfusion system prior to microscopic analysis, thereby also reducing manual efforts and contamination risks at that stage of the process. One proposed approach for integrating a bubble trap into a monolithically 3D-printed cultivation device has already been presented by Beckwith et al. [
31].
In many OOC devices presented in the literature, as well as in commercially available systems, tissue barriers are constructed by integrating porous membranes between two compartments of the device [
16,
32,
33,
34]. Many simplified models are based on the perfusion of hydrogel-based 3D cultures in cell culture inserts (e.g., Transwell
® Inserts). In these systems the exchange of nutrients, metabolites, and gases is not restricted to the medium flow, which is in contrast to physiological conditions in tissues. Furthermore, adhesion forces on the wall of the cell culture insert result in an undefined shape and thickness of the culture. In combination with cell-induced deformation of the hydrogel, this can lead to detachment of the culture from the membrane of the cell culture insert. On the other hand, recovery of cells from completely closed systems, such as those based on multiple channels for integration of hydrogel and culture medium, is often very challenging without destruction of the hydrogel [
34]. The cultivation device presented here overcomes these limitations due to the design decisions described above. Another key advantage of the presented device over commercially available chip–based 3D cell culture systems is that its production process allows for rapid prototyping of customized versions of the device. For instance, the additive manufacturing process used to fabricate this device allows rapid modification of the course of the media channels, the hydrogel volume, and the number of enclosed hydrogel chambers. In addition, the integration of sensors, e.g., for the detection of cell metabolites, as previously demonstrated by Siller et al. [
35] and Arshavsky-Graham et al. [
36], is also feasible. Last but not least, the device presents a promising starting point for the realization of complex fluid flow patterns by integrating valve systems into the setup for automated addition of media supplements, for example, based on the work of Winkler et al. [
37].