We present a protocol for growing high-reproducible spheroids and their phenotypic characterization using image capture and proteomics.
We present a protocol that describes the properties and advantages of using a standalone clinostat incubator for growing, treating, and monitoring 3D cell cultures. The clinostat mimics an environment where cells can assemble as highly reproducible spheroids with low shear forces and active nutrient diffusion. We demonstrate that both cancer and non-cancer hepatocytes (HepG2/C3A and THLE-3 cell lines) require 3 weeks of growth prior to achieving functionalities comparable to liver cells. This protocol highlights the convenience of utilizing incubators for 3D cells with cameras monitoring the cell growth, as snapshots can be taken to count and measure spheroids upon treatment. We describe the comparison of THLE-3 and HepG2/C3A cell lines, showing how non-cancerous cell lines can be grown as well as immortalized cancer cells. We demonstrate and illustrate how proteomics experiments can be conducted from a few spheroids, which can be collected without perturbing cell signaling, i.e., no trypsinization required. We show that proteomics analysis can be used to monitor the typical liver phenotype of respiratory chain metabolism and the production of proteins involved in metal detoxification and describe a semi-automated system to count and measure the spheroid’s area. Altogether, the protocol presents a toolbox that comprises a phenotypic characterization via image capture and a proteomics pipeline to experiment on 3D cell culture models.
In vitro cell cultures have been proven to be necessary and invaluable in establishing fundamental knowledge in biology. Much of the scientific understanding in biology and cancer specifically has come from the 2D culture system that is cells growing in a monolayer. Although 2D culture has been the dominant cell culture system, it has many disadvantages that can potentially stifle further biological progress. For example, 2D cultures lack cell-cell interactions important for cell signaling and proliferation1. To date, 3D culture systems have been shown to better model differentiation, drug response, tumor invasion, and biology2,3,4,5. 3D modeling of malignant cancers is especially vital due to the rise in the aging population and cancer mortality. Hepatocellular carcinoma (HCC) is one of the leading causes of cancer-related mortality worldwide and frequently has an abysmal prognosis6. HCC is known to have a low cure rate, poor drug response, and high rate of reoccurrence6,7,8. Several 3D models for normal liver and HCC have been developed that mimic the physiology of in vivo normal and malignant liver tissue9,10.
Some of the current 3D systems include liquid overlays, bioreactors, hydrogel, scaffolds, and 3D-printed structures. Spheroids generated in bioreactors specifically provide unique advantages because they mimic tumor distribution of nutrient exposure, gas exchange, and cell proliferation/quiescence11. Bioreactors are especially suitable for cancer models due to their ease of use, large scalability, nutrient diffusion, and accessibility11. In addition, bioreactors can allow for high throughput experiments, greater reproducibility, and decreased human error. The bioreactor used in this study is unique because it simulates a system of reduced gravity, which minimizes disruptive shearing forces applied in typical bioreactors allowing for better reproducibility12. The omnidirectional gravity and reduction in shearing forces allow for the cells to develop in a more physiological manner. As evidence, HepG2/C3A cells grown under this methodology develop spherical organelles that produce in vivo levels of ATP, adenylate kinase, urea, and cholesterol13,14. In addition, drug treatments in this 3D system are more advanced and automated in comparison to 2D cultures. In 2D cultures, drug treatments must often have a short time course due to the need to trypsinize and maintain cell health. Yet, in our case, we can perform long-term drug treatments of spheroids without the need to disrupt the structure and physiology of the cells. Therefore, a shift from 2D to 3D cultures is necessary to better model in vivo biological phenomena and further scientific development.
This paper presents a methodology for growing high-reproducible spheroids (Figure 1 and Figure 2) and shows a semi-automated system to phenotypically characterize 3D structures (Figure 3). At the image level, we provide information on counting and measuring spheroids' area (Figure 3). By using mass spectrometry methods, we show how proteomics can be used to assess specific biological functions (Figure 4). By collecting and analyzing this data, we hope to improve the understanding of the biology behind 3D cell culture systems.
1. Buffers and reagents
2. Preparation of spheroids
NOTE: Figure 1A represents the initial steps for preparing and culturing 3D spheroids from cell lines.
3. Spheroids culture into bioreactors (Figure 2)
NOTE: To preserve the structure of spheroids, use wide bore tips when handling 3D spheres.
4. Image capture and counting of spheroids
NOTE: The simplified pipeline for spheroids count is shown in Figure 3A. For spheroids count and area determination (section 5), it is critical to evaluate the compactness of the 3D structures. This will contribute to a more enhanced color contrast, which is necessary for the method to be accurate.
5. Planimetric determination of the spheroid area
NOTE: The simplified pipeline for determining the spheroid area is shown in Figure 3B.
6. Collection of spheroids
NOTE: It is strongly advised that spheroids are collected using wide bore tips to preserve their 3D structure. The collection can be done by using the plug in the front of the bioreactor (Figure 2A).
The size of spheroids at collection may vary depending on the cell line, starting number of cells, and splitting process (days in culture, number of spheroids per bioreactor, and splitting ratio).
7. Viability of spheroids
NOTE: Spheroid viability was determined by measuring the activity of the adenylate kinase (AK) released by damaged cells (Figure 4A). Due to the diffusion gradient, AK measurement is efficient when spheroids are smaller than 900 µm in diameter12. If the spheroids become larger, or if there is any doubt regarding the viability measurement, an ATP assay can be performed15.
8. Protein extraction
NOTE: Figure 1B represents the workflow for spheroids processing and protein extraction.
9. Sample cleanup
NOTE: Before proceeding to the proteomics analysis, it is necessary to remove the salt present in the samples. Salts can interfere with high-performance liquid chromatography-mass spectrometry (HPLC-MS) analysis as they ionize during electrospray, suppressing the signal from peptides. The setup for de-salting used in this study was previously demonstrated by Joseph-Chowdhury and colleagues16.
10. Proteomics analysis via liquid chromatography coupled with mass spectrometry
NOTE: To generate the data for this manuscript, a nLC-MS/MS system with a two-column system setup with a 300 µm ID x 0.5 cm C18 trap column and a 75 µm ID x 25 cm C18-AQ (3 µm) analytical nano-column packed in-house was used.
11. Data analysis
In this protocol, we describe the properties of an innovative stress-free 3D cell incubator, a system designed specifically for the culturing of 3D spheroids (Figure 2). We optimized the protocol for 3D culturing of THLE-3 and HepG2/C3A cell lines. The protocol described here is simple to use and allows for the reproducibility and cost-effective culturing of > 100 spheroids per bioreactor. Once in the bioreactor, spheroids are treated similarly to cells maintained in 2D culture. Optimal growth conditions are achieved by exchanging the media two to three times a week (Figure 2B) and adjusting the rotation speed according to spheroids' growth and size (Figure 2C). This system, in which 3D spheroids are cultured in rotating bioreactors (Figure 2E,F), provides an optimal growth environment for 3D structures by exposing spheroid to an equal and very low amount of shear force.
We have previously shown how spheroids can be used for the analysis of chromatin modification16. Here, we demonstrate in detail how to obtain liver spheroids and how proteomics experiments can be conducted for the analysis of the full proteome (Figure 1). In brief, the protocol was initiated by using THLE-3 or HepG2/C3A flat cells until the culture reached 80% confluency. To culture cells as spheroids, approximately 2,000 cells were plated in an ultra-low attachment plate containing microwells to allow them to self-aggregate, and then, formed spheroids were transferred to a bioreactor (Figure 1A). Although they are functionally active after 3 weeks in culture, as previously demonstrated17, we show results from spheroids collected after 36 days in culture for this protocol. After collection, spheroids were spun down, and both pellet and supernatant were stored for analysis. Cell viability was assessed from the supernatant for the quantification of adenylate kinase released by damaged cells, as described previously17. Cellular proteins were extracted from the cell pellet, and the full proteome was analyzed by high-resolution mass spectrometry (Figure 1B).
This protocol also demonstrates a semi-automated method for spheroids counting (Figure 3A) using the public image processing program FIJI (Fiji Is Just ImageJ)18. A good-quality picture of the spheroid should be taken for the analysis, and some parameters should be considered as mentioned in section 5. Then, after preparing the picture for analysis, a macro script (Supplementary File 1) is used for counting the spheroids. The macro works by first making a folder called FIJI Spheroids counting, inside the folder where the spheroid pictures are located. In this folder, all information from the analysis is saved; this includes a picture of the spheroids that were counted, with an ID number on each spheroid. It also includes an Excel file called spheroid counting. This file contains the pixel area and ID number for each spheroid that was counted. The data corresponding to one analyzed picture is presented in each tab of the file. The tab is labeled according to the name of the picture analyzed. As spheroid size can be impaired by many factors, including the number of structures within a vessel and drug treatment, it is also important to monitor their surface area (planimetry). The macro script presented here (Supplementary File 2) works by measuring the black areas, which correspond to spheroids in the picture (Figure 3B). The output is gathered in a file called planimetry.xlsx, which contains the measured area, perimeter, and diameter of each spheroid. There is also a measurement called Feret, used to calculate the diameter. Feret is the longest possible diameter, while minFeret is the shortest. The diameter is the average of these two. Inside the output folder, besides the planimetry.xlsx file, there is also a picture of the spheroids that were measured.
Before proceeding to the proteome analysis, spheroids' viability was evaluated over culture time. Levels of AK increase up to day 17, reaching approximately 7% of cell death, and then the death decrease to levels below 5% (Figure 4A), which is in accordance with previously published work17. This protocol also shows the full proteome analysis for monitoring the cell phenotype. Firstly, the proteomes of THLE-3 and HepG2/C3A flat cells and spheroids were compared. By analyzing the first principal component (PC1), it is evident that there is a strict separation of spheroid samples from flat cell cultures, and it seems that the correlation of cell type (THLE-3 and HepG2/C3A) is not relevant (Figure 4B). Although THLE-3 and HepG2/C3A spheroids do not cluster together, they share similar profiles consistent with liver function. We demonstrate in this protocol the example of metallothioneins, which have a role in metal detoxification performed by the liver. We identified in the proteomics analysis 2 isoforms overexpressed in spheroids in comparison to flat cells (MT1E and MT1X) (Figure 4C). We also show the Gene Ontology (GO) enrichment of both cell lines grown as spheroids. The carbohydrate metabolic process, which comprises the tricarboxylic acid cycle (TCA cycle), the electron transport chain (cellular respiration), and pyruvate metabolism, is a frequent term and is enriched in both HepG2/C3A and THLE-3 spheroids (Figure 4D,E). Cellular detoxification, fatty acid, and cholesterol metabolism are other functions enriched in both spheroids. Together, these functions are known to be crucial for the liver function.
Figure 1: Workflow for spheroids culture and sample preparation. (A) 3D cell culture experimental approach. Flat cell cultures at desired confluency were trypsinized and seeded on an ultra-low attachment 24 well plate containing microwells, where cells self-assemble into spheroids. After 24 h, spheroids were transferred to a bioreactor and cultivated until they are ready for analysis. (B) After collection, spheroids were pelleted and both pellet and culture supernatant were stored until processing. Histones16and non-histones proteins were extracted, digested into peptides, and analyzed by high-resolution mass spectrometry. Raw files obtained from the mass spectrometry were searched against the human database, and the data were further processed. Please click here to view a larger version of this figure.
Figure 2: 3D cell culture system. (A) Bioreactor parts. The bioreactor is composed of a gas exchange and humidification chamber containing water beads and an openable cell chamber with two plugs for media exchange and spheroids collection. (B) Bioreactor media exchange. The bioreactor is filled with 10 mL of growth media by using a syringe with a needle. (C) The system control app. The rotation speed, CO2 level, temperature, alarm log, and other functionalities can be controlled using the control unit. (D) Placing the bioreactor in the 3D incubator. Each bioreactor has an associated motor which can spin the bioreactor slowly. (E) Bioreactor in movement with the speed (rpm) controlled by a (C) tablet. The speed (rpm) is adjusted according to the spheroids' size. (F) Bioreactors inside de 3D incubator. The 3D incubator can fit up to 6 individually controlled bioreactors. Photo courtesy of Jason Torres Photography. Please click here to view a larger version of this figure.
Figure 3: Phenotypic characterization of spheroids via image capture. (A) Semi-automated spheroid count. After taking snapshots of the spheroids in the bioreactor, the image is prepared for analysis in FIJI. Each spheroid is counted, and an ID number is provided for each of them. A macro is used, and the results are displayed showing the ID for the counted spheroid, the label (name of the picture that was analyzed), and the area (the number of pixels counted in the spheroid). (B) Planimetric determination of spheroid area. Using a macro, the area, perimeter, and diameter of a specific spheroid are determined. Please click here to view a larger version of this figure.
Figure 4: Proteome analysis of liver spheroids. (A) Viability of spheroids was calculated based on the release of adenylate kinase (AK) on culture supernatant. Results are the means of duplicate data points ± SD. (B) Principal component analysis (PCA) was performed to compare the proteome of THLE-3 and HepG2/C3A flat cells and spheroids.(C) Relative abundance of metallothioneins, which are proteins expressed by the human liver. Data are represented as means ± SEM. (D) Functionally grouped network show GO enrichment for HepG2/C3A spheroids and (E) THLE-3 spheroids, where only the label of the most significant term per group is shown. The network was constructed using ClueGo19. The node size represents the term enrichment significance. Please click here to view a larger version of this figure.
Supplementary File 1: Macro script for spheroid counting. Please click here to download this File.
Supplementary File 2: Macro for spheroids planimetric determination. Please click here to download this File.
Understanding the biology behind three-dimensional (3D) cellular structures is extremely important for a more comprehensive knowledge of their functionalities. There is a growing interest in using 3D models for studying complex biology and performing toxicity screening. When cultivating cells in 3D many factors need to be considered, including the phenotypic assessment of the model system. A phenotype is defined as a group of observable characteristics of a specific organism, such as morphology, behavior, physiological and biochemical properties20.
In this protocol, we demonstrate how proteomics experiments can be conducted from a few spheroids and can be used to monitor the typical liver phenotype. Mass spectrometry has become an extensively applied method for 3D cell characterization, allowing the investigation of a variety of biological questions12,16,21,22. For a comprehensive proteome analysis, it is recommended to use at least 20 µg of protein starting material, from which 1 µg is injected into the mass spectrometer. It is important to mention that adding less sample might lead to loss of sensitivity, and adding more would gradually worsen the quality of the chromatography and eventually lead to blocking the column. In this study, we showed that the HepG2/C3A and THLE-3 spheroids are enriched with important proteins from glycolysis and TCA cycle, which are specific liver pathways and are critical for maintaining blood glucose levels and for energy production23,24. Actually, mass spectrometry analysis provides not only information at the protein level but also allows the investigation of protein post-translational modifications, as shown previously by our group16.
Another aspect to be considered in 3D phenotypic studies is the number and size of spheroids. Besides making experiments more reproducible, counting the number of spheroids and determining their size is essential to determine when to split the culture into multiple bioreactors, as the number of 3D structures within a vessel can impact spheroids' size and metabolic activity levels. However, it is important to highlight that the number and size of spheroids depend on the cell line, starting number of cells, splitting process, and time of collection. Details of HepG2/C3A spheroid culture, such as number of cells per spheroids, protein content, and size as a function of age, were provided by Fey, Korzeniowska, and Wrzesinski25. For accurate and successful analysis using the semi-automated method described here, the most critical step is a good spheroids' picture. For simplicity, the picture can be taken with a phone or tablet, but its resolution should be kept as high as possible. As images are quick to acquire, they allow large-scale screening experiments to visualize specific phenotypic features or investigate responses to drug treatment. Therefore, due to the increasing number of cell-based assays, a number of open-source software have been developed over the past 10 years for image analysis26. In this protocol, we describe a semi-automated system using the software FIJI18 to count and measure spheroids' size. We presented scripts (simple programming commands) to define a sequence of algorithmic operations that can be applied to an image collection, making the analysis an easy and quick process. However, depending on the characteristic of the spheroid, a manual measurement should be employed. For instance, if the spheroids are too translucent, the FIJI script will be imprecise. By the way, one of the most important criteria for this method to work is the compactness of the spheroids. This characteristic will contribute to a more enhanced color contrast between the spheroids and the background, which is necessary for the method to be accurate.
In summary, besides presenting a methodology for growing high-reproducible spheroids, a semi-automated system coupled with phenotypic characterization via image capture and proteomics was also described. We expect this toolbox for analyzing 3D cells to become more robust with full-automated image analysis software and next-generation mass spectrometers.
The authors have nothing to disclose.
The Sidoli lab gratefully acknowledges the Leukemia Research Foundation (Hollis Brownstein New Investigator Research Grant), AFAR (Sagol Network GerOmics award), Deerfield (Xseed award), Relay Therapeutics, Merck, and the NIH Office of the Director (1S10OD030286-01).
1.5 mL microcentrifuge tubes | Bio-Rad | 2239480 | |
10 mL syringe | Fisher Scientific | 1481754 | Luer lock tip, graduated to 12 mL |
1000 µL wide bore pipet tips | Fisher Scientific | 14222703 | |
200 µL wide bore pipet tips | Fisher Scientific | 14222730 | |
96-well Orochem filter plate | Orochem | OF1100 | |
96-well skirted plate | Axygen | PCR-96-FS-C | |
96-well vacuum manifold | Millipore | MAVM0960R | |
Ammonium bicarbonate | Sigma | A6141-25G | |
Bronchial Epithelial Cell Growth Medium (BEGM) | Lonza | CC-3170 | |
Cell culture grade water | Corning | 25-055-CV | |
ClinoReactor | CelVivo | N/A | Bioreactor for 3D cell culture |
ClinoStar incubator | CelVivo | N/A | CO2 incubator for 3D cell culture |
DTT | Sigma | D0632-5G | |
Dulbecco's Modified Eagle's Medium (DMEM) | Fisher Scientific | MT17205CV | |
Elplasia 24-well round bottom ultra-low attachment plate containing microwells | Corning | 4441 | |
Fetal Bovine Serum | Fisher Scientific | MT35010CV | |
Formic acid | Thermo | 28905 | |
Hank's Balanced Salt Solution (HBSS) | Fisher Scientific | MT21022CV | |
hEGF | Corning | 354052 | |
HERAcell vios 160i | Thermo | 51033557 | CO2 incubator for 2D cell culture |
HPLC grade acetonitrile | Fisher Scientific | A955-4 | |
HPLC grade methanol | Fisher Scientific | A452-1 | |
HPLC grade water | Fisher Scientific | W5-4 | |
Iodoacetamide | Sigma | I1149-5G | |
L-glutamine | Fisher Scientific | MT25015CI | |
Non-essential amino acids | Fisher Scientific | MT25025CI | |
Oasis HLB Resin 30 µm | Waters | 186007549 | |
Orbitrap Fusion Lumos Tribrid mass spectrometer | Thermo | IQLAAEGAAPFADBMBHQ | High resolution mass spectrometer |
PAULA microscope | Leica | ||
Penicillin-Streptomycin | Fisher Scientific | MT3002CI | |
PerkinElmer Victor X2 multilabel microplate reader | PerkinElmer | ||
pH paper | Hydrion | 93 | |
Phosphoetanolamine | Sigma | P0503 | |
Phosphoric acid | Fisher Scientific | A260-500 | |
Pipette gun | Eppendorf | Z666467 (Milipore Sigma) | |
Refrigerated centrifuge | Thermo | 75-217-420 | |
Reprosil-Pur resin | MSWIL | R13.AQ.003 | 120 Å pore size, C18-AQ phase, 3 μM bead size |
SDS | Bio-Rad | 1610301 | |
Sequencing grade modified trypsin | Promega | V511A | |
SpeedVac vacuum concentrator (96-well plates) | Thermo | 15308325 | Savant SPD1010 |
Sterile hood | Thermo | 1375 | |
Sterile serological pipettes | Fisher Scientific | 1367549 | |
S-trap | Protifi | C02-micro-80 | |
Syringe needle (18 G) | Fisher Scientific | 14817100 | 3" length, 0.05" diameter |
Trifluoroacetic acid (TFA) | Thermo | 28904 | |
Trypsin-EDTA | Gibco | 25300-054 | |
Vortex | Sigma | Z258415 | |
Water bath | Fisher Scientific | FSGPD10 |