Real-time analysis of live tissue yields important functional and mechanistic data. This paper describes the protocols and critical variables to ensure accurate and reproducible generation of data by a novel and pump-free multi-channel fluidics system that maintains and assesses a wide range of tissue and cell models.
Many in vitro models used to investigate tissue function and cell biology require a flow of media to provide adequate oxygenation and optimal cell conditions required for the maintenance of function and viability. Toward this end, we have developed a multi-channel flow culture system to maintain tissue and cells in culture and continuously assess function and viability by either in-line sensors and/or collection of outflow fractions. The system combines 8-channel, continuous optical sensing of oxygen consumption rate with a built-in fraction collector to simultaneously measure production rates of metabolites and hormone secretion. Although it is able to maintain and assess a wide range of tissue and cell models, including islets, muscle, and hypothalamus, here we describe its operating principles and the experimental preparations/protocols that we have used to investigate bioenergetic regulation of isolated mouse retina, mouse retinal pigment epithelium (RPE)-choroid-sclera, and cultured human RPE cells. Innovations in the design of the system, such as pumpless fluid flow, have produced a greatly simplified operation of a multi-channel flow system. Videos and images are shown that illustrate how to assemble, prepare the instrument for an experiment, and load the different tissue/cell models into the perifusion chambers. In addition, guidelines for selecting conditions for protocol- and tissue-specific experiments are delineated and discussed, including setting the correct flow rate to tissue ratio to obtain consistent and stable culture conditions and accurate determinations of consumption and production rates. The combination of optimal tissue maintenance and real-time assessment of multiple parameters yields highly informative data sets that will have great utility for research in the physiology of the eye and drug discovery for the treatment of impaired vision.
Perifusion systems have a long history in the life sciences. In particular, for the study of secretory function by islets, they have been used to characterize the kinetics of insulin secretion in response to secretagogues1. In addition to collecting outflow fractions for subsequent assay of hormones and metabolites, real-time sensors have been incorporated, predominantly for the detection of oxygen consumption2,3,4. Widespread efforts to better understand mechanisms mediating diseases of the eye have been limited by a lack of physiologically relevant methods to assess metabolic regulation and dysregulation of the various isolated components of the eye, including the retina, retinal pigment epithelial (RPE)-choroid-sclera and cultured RPE cells. Static systems designed for cultured cells have been adapted for tissue5, but tissue requires flow for adequate oxygenation. Flow systems have been successful at accurately and reproducibly measuring real-time responses in oxygen consumption rate (OCR) by the retina and RPE-choroid-sclera, and the tissues stay metabolically stable for more than 8 h allowing highly informative protocols involving multiple test compounds4,6,7,8,9. Nonetheless, the operation of fluidics systems has historically required a custom-made apparatus and trained technical staff in non-standardized methodologies. Such systems have not been adopted as the standard methodology in most laboratories. The BaroFuse is a newly developed fluidics system that does not rely on pumps, but rather on gas pressure to drive flow through multiple channels and tissue chambers (Figure 1). Each channel is continuously monitored for OCR, and the outflow is collected with a plate-based fraction collector for subsequent assay of contents. Importantly, the tissue perifusion chambers for the instrument are designed to accommodate tissues of various geometries and sizes.
The heart of the instrument is the fluidics system, where flow is driven from a sealed, pressurized reservoir through small inner diameter (ID) tubing (contributing the most significant flow resistance in the fluid circuit) up into the glass tissue chambers that house the tissue. Pressure to the media reservoir module (MRM) is supplied by low-pressure and high-pressure regulators connected to a gas cylinder containing a mixture of gases (typically 21% O2, 5% CO2, balance N2), and the reservoir is sealed from the top by the perifusion chamber module (PCM) that holds the tissue chamber assemblies (TCAs). Flow rate is controlled by the length and ID of the resistance tubes and the pressure setting of a low-pressure regulator. Outflow tubes connected to the top of tissue chambers deliver fluid to either a waste receptacle (that is continuously weighed for automatic determination of flow rate) or into wells of a 96-well plate controlled by the fraction collector. The O2 detection system measures the lifetime of an O2-sensitive dye painted on the inside of each of the glass tissue chambers downstream of the tissue. This information is then used to continuously calculate OCR. The entire fluidics system resides in a temperature-controlled enclosure and the gas tank, fraction collector and computer are the major components of the instrument (Figure 2A). Finally, software that runs the instrument serves to control its operation (including the preparation and timing of injected test compounds, flow measurement system, and fraction collector timing), as well as processing and graphing the OCR data and other supplemental measurements.
In this paper we describe the protocols for using the fluidics system to perifuse and assess OCR and lactate production rate (LPR) for various isolated components of the eye. LPR is a parameter reflecting glycolytic rate that is highly complementary to OCR, where the pair accounts for the two major branches of energy generation from carbohydrates in the cell10. As preparation of the tissue and loading it into the tissue chambers is best learned by watching the procedure, the video will help illustrate several of the critical steps that are performed during set up and operation that are not easily conveyed by text alone.
The description of the protocol is divided up into 8 sections that correspond to different phases of the experiment (Figure 2B): 1. pre-experimental preparation; 2. preparation/equilibration of the perifusate; 3. instrument set up; 4. tissue equilibration; 5. experimental protocol; 6. instrument break down; 7. data processing; and 8. assays of outflow fractions.
All procedures for harvesting tissue from rats and mice were approved by the University of Washington Institutional Animal Care and Use Committee.
1. Pre-experimental preparation
NOTE: The following tasks are completed at least a day in advance of the experiment.
2. Preparation and equilibration of perifusate (Time: 30 min not including incubation time)
3. Equilibration of temperature and dissolved gas to set up the instrument (Time: 75 min)
4. Tissue loading and equilibration period (Time: 90 min)
5. Experimental protocol (Time: 2-6 h)
NOTE: Once baseline stabilization is underway, the next tasks are injecting the test compounds and changing plates on the fraction collector if more than one will be used.
6. Ending the experiment and breaking down the system (Time: 30 min)
7. Data processing (Time: 15-45 min)
8. Assays of outflow fractions
To illustrate the resolution of the data generated from isolated components of the eye, OCR and LPR was measured with three types of tissue (retina, RPE-choroid-sclera, and RPE cells) following a commonly used protocol (the mitochondrial stress test10; Figure 10, Figure 11, and Figure 12). The amount of tissue used for each tissue is shown in Table 1. Data was processed and graphed using the software package that was developed for the fluidics system. The preparation of retina and RPE-choroid-sclera is relatively straight-forward and takes less than 20 min for each tissue type. OCR was constant during the time that test compounds were injected, indicating stable health and function of the tissue and supporting the validity of the method (Figure 10). Once validated for each tissue type, we have not found it necessary to run controls where no test compounds are injected for each experiment. Consistent with data obtained using more conventional perifusion methods6,8,13, OCR decrease in response to oligomycin and increased OCR in response to FCCP. Changes in LPR were in the opposite direction of those observed for OCR: oligomycin increased LPR, which then decreased (but only slightly) in response to FCCP (Figure 11). To compare the statistical significance of the effect of each sequential test compound, t-tests were performed (which are calculated automatically by the software that comes with the instrument). Since the goal of the paper was to describe how to perform the method, the number of replicates carried was not always high enough to produce statistical significance. In general, though, when the number of replicates were 3 or more, effects of FCCP and oligomycin on both OCR and LPR were significant.
RPE cells have not been previously analyzed with flow systems but responded similarly to RPE-choroid-sclera (consistent with the view that a large fraction of OCR is due to RPE cells; Figure 11). These illustrative examples highlight ability of the system to maintain tissue viability as reflected by the stability of OCR in the control channels, and the high signal to noise ratio for changes in OCR of the magnitude induced by oligomycin and FCCP, which was more than 100 to 1. In addition, assays of outflow fractions can be used to correlate the rate of uptake or production of a wide array of compounds exchanging with the extracellular fluid are complementary to OCR (in this case, LPR). These features of the instrument allowed accurate quantification of characteristic differences in tissue responses between tissue types performed in parallel. OCR by RPE-choroid-sclera and RPE cells are consistently more sensitive to oligomycin than retina (Figure 11 and Figure 12), although for the RPE-choroid-sclera the duration of exposure to FCCP was not long enough to reach steady state. A point to consider arose when using DMSO as a solvent. At higher concentrations, (0.2%) DMSO had a transient effect on OCR by retina (presumably reflecting an effect of a change in osmotic pressure brought about by DMSO's effect on membrane permeability).
Based on the assumption that KCN completely inhibits respiration by its direct action on cytochrome c oxidase, OCR at the end of the KCN exposure is set to 0 and all OCR values are calculated based on the change relative to the KCN value. OCR can occur independent of the respiratory chain and cytochrome c oxidase. However, the magnitude of this contribution to overall OCR is generally not more than a few percent (data not shown) and the extended length of time that tissue is exposed to KCN ensures that substrates of oxidases that are not part of the electron transport chain have been depleted.
Statistical analysis
Single experiments were shown as indicated in the figures, but with multiple channels that were averaged. Data was then graphed as the average ± the standard error (SE; calculated as SD/√n).
Figure 1. Schematic of the fluidics/assessment system. Major components include the enclosure, temperature control elements, fluidics and tissue chamber systems, regulation of gas pressure in the head space above perifusate, fraction collector/flow rate monitoring, and O2 detectors. Abbreviations: MRM = Media Reservoir Module, PCM = Perifusion Chamber Module, TCA= Tissue Chamber Assemblies. Please click here to view a larger version of this figure.
Figure 2. (A) Picture of the major components of the instrument. The major components consist of gas tank (pressure regulators), enclosure, fraction collector and computer. (B) Experimental flow chart showing the major categories of steps and the time it takes to complete them. Please click here to view a larger version of this figure.
Figure 3. View of the MRM. The MRM is shown with an MRM insert (left) and stir bars (right) placed into the bottom of the MRM inserts (placed in each side of the MRM Divider). Please click here to view a larger version of this figure.
Figure 4. Tubing assembly and purge tubing assembly in the MRM. (A) Test compound injection tubing assembly and purge tubing assembly attached to ports on the MRM. (B–C) The test compound injection assembly and purge tubing assembly (B) are placed in the groove in the front of Enclosure (C). Please click here to view a larger version of this figure.
Figure 5. Powering up the MRM temperature controller. Please click here to view a larger version of this figure.
Figure 6. Tissue chambers and gas tank. Positioning the O2 detector on the detector stand (which also supports the MRM and PCM), and placement of the band around the fins of the PCM that help secure the tissue chambers in place. Please click here to view a larger version of this figure.
Figure 7. (A) High- and low-pressure regulators on the gas tank. (B–C) Purge tube. Purge tube allows the headspace in the MRM to clear of air and fill with gas from the supply tank. Pictures showing open purge tube (B) and close purge tube (C). Test compound injection assembly stays closed through the purge process. Please click here to view a larger version of this figure.
Figure 8. Tissue chamber and the outflow setup. (A) Dimensions of the Transwell membrane after it is cut into three strips of equal width. (B) Outflow multi-tube support. (C) Outflow multi-tube support positioned on the lip of the enclosure with the tubing adapters near the tissue chambers. (D) Picture of outflow tubing assemblies attached to the tissue chambers. (E) Aerial view of the enclosure without the Lid. Please click here to view a larger version of this figure.
Figure 9. Injection of compound in MRM. Injecting a test compound through the injection port into the MRM using a 5 mL syringe. Please click here to view a larger version of this figure.
Figure 10. OCR and LPR curves in response to test compounds. OCR and LPR by retina isolated from mice (1 retina/channel) in response to the presence or absence (control) of test compounds as indicated. Each curve is the average of 6 replicates from a single experiment (error bars are SE; p-values are calculated by performing paired t-tests comparing steady state values for each test agent with that of the previous test agent). Please click here to view a larger version of this figure.
Figure 11. OCR curves. OCR by RPE-choroid-sclera and retina isolated from mice (1 retina or 2 RPE-choroid-sclera/channel) measured in parallel in response to test compounds as indicated. Data is the average of replicates from a single experiment (n = 2 and 4 for RPE-choroid-sclera and retina respectively; p-values are calculated by performing paired t-tests comparing steady state values for each test agent with that of the previous test agent). Please click here to view a larger version of this figure.
Figure 12. OCR and LPR curves from RPE cells. OCR and LPR from RPE cells attached to transwell membranes that were cut into strips and loaded into the perifusion chambers. Data is the average of replicates from a single experiment (n = 3, with 1.5 membranes/channel (360,000 cells/channel); p-values are calculated by performing paired t-tests comparing steady state values for each test agent with that of the previous test agent). Please click here to view a larger version of this figure.
TISSUE/CELL | Amount/Channel | FLOW RATE: mL/min |
Retina (mouse) | 1 | 0.025 |
RPE-choroid-sclera (mouse) | 2 | 0.02 |
RPE Cells on Transwell Membranes | 360,000 Cells (4 x 1/3 filter strips) | 0.016 |
Table 1. Recommended operating specifications for different tissue.
Supplementary Figure 1. Graphical representation of experimental design. Timing and composition of exposure to test compounds, and timing of fraction collection. Concentration increment (Conc Inc) is the change in concentration to be implemented. Please click here to download this File.
Supplementary Figure 2. User interface at startup. UI of the startup window of the O2 detection software that monitors the O2 in the tissue chambers inserted into the PCM. Please click here to download this File.
Supplementary Figure 3. User interface for experiment settings. UI for entering experimental information (left) and selecting times for collection of outflow fractions (right). Please click here to download this File.
Supplementary Figure 4. User interface of the injection page. Injection page which calculates injection volumes based on desired concentrations of test compound and volume left in the MRM. Please click here to download this File.
Supplementary File 1: Methods for tissue sample preparation. Please click here to download this File.
Due to the importance of bioenergetics in all aspects of cell function and maintenance of various components of the eye, there is a critical need for methods to study its regulation. In particular, neural retina and RPE depend on metabolism for both generation of energy as well as intra- and inter-cellular signaling14,15,16,17. Because of their high oxidative capacity, isolated tissues of the eye are not well-maintained under static conditions18,19 and therefore study of isolated components of the eye require flow systems that can both maintain and assess metabolic processes. The fluidics system was developed to generate OCR and LPR data from a wide range of tissue types and in this paper we presented detailed protocols that were found to produce optimal results.
The major determinant for generating robust data using the flow system includes pre-equilibration of CO2-based media/buffer at 39 °C (to ensure perifusate is not supersaturated with dissolved gas that would degas during the experiment). In particular, media or KRB buffer stored at 4 °C will be supersaturated relative to 37 °C and will degas during the experiment if pre-equilibration times are insufficient. In addition, tissue loaded into the tissue chambers must not be traumatized by improper isolation of tissue due to tearing or incomplete separation of tissue, or by exposing tissue in low amount of bicarbonate-based buffer to atmospheric air for too long. The temperature control, flow stability and reliability of O2 detection have little variability and these factors do not contribute significantly to failure rate.
The instrument has eight flow channels/tissue chambers that run simultaneously which are supplied with perifusate from two reservoirs, four tissue chambers for each reservoir. To get the most accurate time-courses of OCR, kinetic curves are baseline corrected by chambers that are not loaded with tissue. Thus, a typical experimental protocol would involve two groups of three tissue chambers. Protocols in general fall into two categories: one is the different test compound protocols on each side (for instance drug/vehicle on one side of the MRM, and just vehicle on the other); the second is same test compound injection protocol on both sides of the MRM, but different tissue or tissue model on each side of the MRM. In this paper, the effects of oligomycin and FCCP on retina were compared to OCR by tissue that were not exposed to any test compounds, and two tissues were concomitantly assessed under the same protocol and conditions to identify tissue-specific behavior. The latter was illustrated in this study by showing increased dynamic range of metabolic rate by RPE-choroid-sclera relative to retina in parallel in the same experiment. Other reports have described a wider range of study designs including measuring the effects varying O2 levels on OCR and LPR, and concentration dependencies of fuels, drugs, and toxins20,21. In addition, although we have limited the analysis of outflow fractions to the measurement of lactate and calculation of LPR, the information content of an experiment increases greatly if multiple compounds and classes of compounds in the outflow fractions are assayed such as hormones, neurotransmitters, cell signals, and metabolites that can exit the cells20,22,23.
The loading of isolated retina or RPE-choroid-sclera is straightforward, and once isolated these tissues are simply placed into the top of the tissue chambers with forceps and allowed to sink down to the frit. RPE cells cultured on filter inserts develop appropriate polarization and markers of RPE maturity after 4-8 weeks in culture. It is not feasible to remove the RPE for live cell analysis once attached to the transwell membrane, if RPE maturity and polarization are to be maintained24. The perifusion chamber can accommodate strips of the transwell membrane that are cut with a scalpel while submerged in buffer and rapidly inserted into the tissue chambers. Although cutting filter strips has been placed into a static system24, no other fluidics method to assess these important cell types are available. The responses of RPE cells were rapid and more dynamic than either the retina or the RPE-choroid-sclera, likely in part due to immediate access of both the apical and basal aspects of the RPE cells configured as a monolayer on the membrane insert.
Another factor in assuring data has the highest signal to noise is selecting the optimal ratio of tissue loaded into the perifusion chambers relative to the flow rate. Too little tissue relative to the flow rate results in a difference of dissolved O2 concentration between inflow and outflow that is very small and difficult to measure reliably. In contrast, if the flow is too slow, then the concentration of O2 becomes so low that the tissue is affected by hypoxia. Nonetheless, the gas pressure-driven liquid flow can be maintained at flow rates down to 5 mL/min requiring only small amounts of tissue for accurate OCR and LPR measurements. In the experiments shown here, about 20 mL/min/channel was used which was suitable for either one retina, two RPE-choroid-scleras, or 360,000 RPE cells. To minimize the system effects that delay and disperse the exposure of the tissue to the injected test compound, multiple sizes of the tissue chambers are supplied, so that the amount of tissue (and flow rate) is matched with the appropriate size of the chamber.
Data from the analyses shown in this paper were represented in two ways: absolute magnitude with respect to rate, or fractional changes relative to a steady-state or baseline. The focus was on illustration of measurement of responses to test compounds. However, the fluidics system is well-suited to assess and compare effects of tissue treatment prior to the perifusion analysis such as genetic modifications. Testing whether a treatment is different from control is most robust if the effects of the treatment on normalized responses of test compounds is analyzed. If the analysis requires absolute magnitudes, the statistical power of the analyses of specimens that are pretreated is maximized if their assessment and controls are carried out in the same perifusion experiment.
Except for the stirrer, all parts that come in contact with liquid are supplied by the manufacturer as consumables and have been sterilized. These parts should not be reused, as experiments will occasionally be lost due to incomplete cleaning and contaminated surfaces. The system at the outset of setup is sterile. However, media is added to the MRM, and tissue is loaded in the chambers under non-sterile conditions. We have measured OCR in the system that is assembled with parts that are sterile, but where the experiment itself is carried out under non-sterile conditions. It takes about 14 h for bacteria to accumulate to the point of having measurable OCR (unpublished results). If protocols are used that are less than 10 h or so, then accumulation of bacteria and any effects due to these will be negligible.
Many investigators use instruments that are designed to measure OCR under static incubation of a monolayer of cells with a relatively high throughput25,26. In contrast, the fluidics instrument we have tested and described in this paper maintains tissue by ensuring adequate O2 delivery which is critical for the greater diffusion distances that are present in tissue specimens. In addition, it is able to collect fractions allowing assessment of multiple parameters in parallel with OCR which greatly enhances the ability to study relationships between them. Finally, dissolved gas concentrations (such as O2 and CO2) can be controlled, increasing the duration of experiments with bicarbonate-based media and buffer, enabling the user to study the effects of O2. It should be pointed out, a limitation for both methodologies is the inability to study the washout of test compounds, a functionality that other perifusion systems have4,27,28. Another consideration when determining the optimal analysis modality is the fact that fluidics systems use more media and test compounds than static systems. The extra expense is minimized with the current fluidics systems though due to the low flow rates that the system can be used.
Overall, a detailed description of the protocols to perform experiments with a new flow/assessment instrument is described. Data generated with retina and RPE-choroid-sclera recapitulated previous results obtained with systems that are much more difficult to use (and not readily available). It was also demonstrated that the system can maintain and assess RPE cells attached to transwell membranes, a very important cellular model that has not previously been analyzed with flow systems due to the fragility of the cells. The main parts of the protocol consist of a 75 min setup time, followed by a 90 min equilibration period and the experimental protocol making it suitable for routine use by laboratories that do not specialize in the operation of fluidics systems. Although we focused on measuring the acute response of tissue to test compounds, the system is very suitable to comparing tissue from various sources such as animal models or cells models that have been genetically altered or undergone test treatments/conditions. In addition, the scope of assays that can be conducted on the outflow fractions are wide-ranging and include metabolites, cell signaling molecules and secreted hormones/neurotransmitters as well as multi-component analysis generated by mass spectrometry on the fractions as well as the tissue.
The authors have nothing to disclose.
This research was funded by grants from the National Institutes of Health (R01 GM148741 I.R.S.), U01 EY034591, R01 EY034364, BrightFocus Foundation, Research to Prevent Blindness (J.R.C.) and R01 EY006641, R01 EY017863 and R21 EY032597 (J.B.H.).
BIOLOGICAL SAMPLES | |||
C57BL/6J mice | Envigo Harlan (Indianapolis, IN) | N/A | |
REAGENTS | |||
FCCP | Sigma-Aldrich | C2920L9795 | |
Glucose | Sigma-Aldrich | G8270G | |
KCN | Sigma-Aldrich | 60178 | |
Lactate | MilliporeSigma | L6661 | |
Oliigomycin A | Sigma-Aldrich | 75351L9795 | |
CELL CULTURE AND TISSUE HARVESTING | |||
Beuthanasia-D | Schering-Plough Animal Health Corp., Union, NJ | N/A | |
Bovine serum albumin | Sigma-Aldrich | A3059 | |
Euthasol, 390 mg/ml sodium pentobarbital | Virbac | RXEUTHASOL | |
Fetal bovine serum | Sigma-Aldrich | 12303C | |
Hank’s Buffered Salt Solution | GIBCO | 14065056 | |
Krebs Ringer Bicarbonate (KRB) | Thermo Fisher Scientific | J67795L9795 | |
Matrigel | ThermoFisher | #CB-40230 | |
Penicillin-streptomycin | ThermoFisher Scientific | 15140122 | |
ROCKi | Selleck Chemicals | Y-27632 | |
Trypsin-EDTA | ThermoFisher | #25-200-072 | |
SUPPLIES | |||
Gas Cylinders: 21% O2/5% CO2/balance N2 | Praxair Distribution, Inc | N/A | |
Transwell filters | MilliporeSigma | 3470 | |
COMMERCIAL ASSAYS | |||
Amplex Red Glucose/Glucose Oxidase Assay Kit | ThermoFisher | A22189 | |
Glucose Oxidase from Aerococcus viridans | Invitrogen (Carlsbad, CA) | A22189L9795 | |
Lactate Oxidase | Sigma-Aldrich | L9795 | |
EQUIPMENT | |||
BaroFuse Multi-Channel Perifusion system | EnTox Sciences, Inc (Mercer Island, WA | Model 001-08 | |
Synergy 4 Fluorometer | BioTek (Winooski, VT) | S4MLFPTA |