Obtaining a pure population of fibroblasts is crucial to studying their role in wound repair and fibrosis. Described here is a detailed method to isolate fibroblasts and myofibroblasts from uninjured and injured mouse hearts followed by characterization of their purity and functionality by immunofluorescence, RTPCR, fluorescence-assisted cell sorting, and collagen gel contraction.
Cardiac fibrosis in response to injury is a physiological response to wound healing. Efforts have been made to study and target fibroblast subtypes that mitigate fibrosis. However, fibroblast research has been hindered due to the lack of universally acceptable fibroblast markers to identify quiescent as well as activated fibroblasts. Fibroblasts are a heterogenous cell population, making them difficult to isolate and characterize. The presented protocol describes three different methods to enrich fibroblasts and myofibroblasts from uninjured and injured mouse hearts. Using a standard and reliable protocol to isolate fibroblasts will enable the study of their roles in homeostasis as well as fibrosis modulation.
Cardiac fibroblasts, cells of mesenchymal origin, play a significant role in maintaining the electrical conduction and mechanical forces in the heart in addition to the maintenance of cardiac architecture during homeostasis1. Following injury, these cells are activated, expand, and produce extracellular matrix (ECM) proteins2. Many preclinical studies have revealed fibroblasts as critical cellular regulators that maintain the structural integrity of an injured heart3 as well as main effector cells responsible for unchecked production and deposition of ECM proteins, resulting in stiff scar formation and heart failure4. Fibroblasts are a heterogenous group of cells, making it challenging to dissect their reparative function from pro-fibrotic maladaptive properties. Recently, the functional heterogeneity of two distinct fibroblast subtypes following myocardial injury have been defined, indicating the possibility of isolating different fibroblast subtypes and studying their role in wound healing5.
Obtaining a pure fibroblast population is crucial in delineating their functional role in repair and fibrosis. However, the presence of multiple fibroblast markers that recognize other cell types make it challenging to isolate a substantially pure fibroblast population6. Several elegant studies have devised clever ways to isolate cardiac fibroblasts from uninjured and injured myocardium. The most popular and well-established method of enriching fibroblasts is through selective adhesion following enzymatic tissue digestion7.
Additionally, fluorescence-activated cell sorting (FACS) of fibroblasts based on cell surface antigens has been successfully described8. In the study, following enzymatic digestion, the mesenchymal cells were sorted as lineage-negative (Lin: Ter119−CD45−CD31−) and gp38-positive (gp38+) from mouse hearts. Gp38+ve cells were confirmed to be fibroblasts based on their co-expression of col1α1 and other mesenchymal markers. Although most tissue digestion is completed after dissecting out the ventricle in a Petri dish, a recent study has investigated the use of a direct needle enzyme perfusion of the left ventricle to isolate myocytes and non-myocytes which include fibroblasts9. Fibroblasts were then isolated by selective adhesion in this case.
This protocol describes the isolation and enrichment of fibroblasts using three methods. The first is an already established method involving selective adhesion of fibroblasts following enzymatic digestion. The second method is used to primarily isolate injury-induced alpha smooth muscle expressing myofibroblasts. The third method involves sequential, magnetic depletion of an enzyme-digested cardiac cell suspension of hematopoietic and endothelial cells. Following depletion, fibroblasts/myofibroblasts are isolated based on the presence of the antigen MEFSK4 using magnetic beads. Recently, MEFSK4 has been described as an antigen present on quiescent as well as activated fibroblasts, making it a suitable marker for fibroblast identification and isolation. Naturally, all the methods described here have unique limitations. It is therefore highly recommended to check the purity of the isolated cell population by flow analysis, immunostaining, and semi-quantitative real-time PCR. However, these methodologies can be expanded upon, and additional markers can be added in order to exclude other contaminating populations prior to utilizing the fibroblast and myofibroblast populations for crucial experiments.
This study strictly upholds the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The Vanderbilt University Institutional Animal Care and Use Committee approved the protocol (protocol number: M1600076-01).
1. Heart Dissection
2. Isolation of Fibroblasts from Single Cell Suspension
3. Purity and Functionality Analysis of Isolated Fibroblast Population
Flow gating scheme demonstrating myofibroblast isolation using αSMA-GFP reporter mice
Uninjured hearts showed no detectable GFP+ cells in αSMA-GFP reporter mouse model; hence, they were used to establish a gate for the background signal of the GFP channel post-compensation (Figure 2). αSMA+ cells were sorted based on the presence of GFP expression from the injured left ventricle 10 days following MI. A small percentage of endothelial (GFP+/CD31+ cells; SD = 3.8% ± 0.0164; n = 5) and hematopoietic (GFP+/CD45+ cells; SD = 3.18% ± 0.0112; n = 5) cells also expressed GFP in the injured αSMA-GFP mouse hearts (Figure 2A). However, GFP+/CD31-/CD45- cells did not express AN2, a pericyte marker.
Uninjured (quiescent) and injured (activated, αSMA+GFP+) cells expressed fibroblast markers
GFP+ cells isolated from αSMA-GFP mice expressed αSMA, collagen type 1 alpha-1 chain (COL1α1), vimentin, and periostin when analyzed by IF analysis. Uninjured fibroblasts isolated by selective adhesion expressed vimentin but did not demonstrate expression of the activated fibroblast markers: αSMA, periostin, and COL1α1 (Figure 3A). Both uninjured and activated fibroblasts expressed the MEFSK4 antigen when analyzed by flow analysis (Figure 3B). Magnetically isolated MEFSK4+ve cells from uninjured mice hearts expressed markers of fibroblasts: Col1a1, pdgfrα, and periostin. In contrast, magnetically isolated CD45 and CD31 positive cells had negligible expression of fibroblast markers.
Fibroblasts and myofibroblasts demonstrated the ability to contract collagen
In cell culture on stiff plastic, fibroblasts have been shown to contract collagen gels in the presence of TGFβ, demonstrating their functional capability of contraction13,14. This in vitro characteristic of fibroblasts is very similar to the connective tissue contraction that happens during tissue repair as well as other biological processes. Both uninjured fibroblasts, isolated by selective adhesion, and myofibroblasts, isolated and sorted from αSMA-GFP mice, demonstrated an ability to contract collagen (Figure 4).
Figure 1: Schematic of fibroblast isolation using three different approaches. (A) differential plating, (B) GFP+ cell sorting of αSMA positive cells, and (C) magnetic bead based isolation of fibroblasts. Representative bright field of the cells in culture following differential plating. Scale bar = 50 µM. Please click here to view a larger version of this figure.
Figure 2: FACS analysis of single cells isolated from αSMA-GFP mice hearts following MI. (A) Representative FACS gating scheme demonstrating GFP+ cells co-expressing CD31, CD45, or AN2 from αSMA-GFP mice injured hearts 10 days after myocardial infarction (MI). (B) Graphical quantification of the presented FACS data for post-MI hearts; n = 5 experiments were performed independently (***p < 0.0001 as calculated using one-way ANOVA with Tukey's multiple comparisons test). This figure is adapted from Saraswati et al.10. Please click here to view a larger version of this figure.
Figure 3: Purity analyses of fibroblasts isolated from uninjured and injured mice hearts. (A) Immunofluorescence staining of cell populations (P0) from heart of uninjured, or injured αSMA-GFP mice sorted by FACS. Both uninjured and activated cells express fibroblast (FB) markers, such as COL1α1, and vimentin, but not hematopoietic marker CD45 or the endothelial marker CD31. Cells isolated from injured αSMA-GFP mice heart expressed activated fibroblast markers, αSMA, and periostin, which were not present in the cells isolated from uninjured mice hearts. Nuclei were stained with DAPI, n = 3 experiments were performed independently. Scale bar = 100 µm. (B) Representative FACS overlay histogram of uninjured and activated fibroblasts (P3–P5) showing the expression of the fibroblast marker MEF-SK4. For a negative control, rat IgG was used, n = 2 experiments were performed independently. (C) Relative fold change of Col1α1, Pdgfrα, and Postn transcripts in uninjured MEKSK4+ve fibroblasts, n = 3 experiments were performed independently (*p < 0.05 as calculated using two-way ANOVA with Tukey's multiple comparisons test). (A) and (B) are adapted from Saraswati et al.10. Please click here to view a larger version of this figure.
Figure 4: Functional characterization of fibroblasts isolated from uninjured and injured mouse hearts. Representative figure of collagen gel contraction in the presence of uninjured and injured activated αSMA+ fibroblasts (P3–P5). The graph represents the percentage change in the initial gel area after 24 h and 48 h of contraction when incubated with uninjured and injured activated αSMA+ fibroblasts, n = 2 experiments were performed independently. This figure is adapted from Saraswati et al.10. Please click here to view a larger version of this figure.
Antibody | Cell Target | Dilution |
7AAD | Dead cells | 1:1000 |
Ghost dye violet 510 | Dead cells | 1:1000 |
APC-CD45 | Hematopoietic cells | 1:200 |
PE-CD31 | Endothelial cells | 1:200 |
anti-AN2/NG2 | Pericytes | 1:11 |
Donkey anti-rat alexa fluor 405 | Secondary Antibody | 1:100 |
Anti-feeder cells-APC (MEFSK4) | Fibroblasts | 1:100 |
Rat IgG-APC | Isotype Control | 1:100 |
Table 1: FACS dyes and antibodies.
Primary Antibody | Dilution | |
α-smooth muscle actin (αSMA) | 1:1000 | |
Fibroblast specific protein 1 (FSP1)1:250 | 1:100 | |
COL 1α1 | 1:1000 | |
Periostin | 1:100 | |
Vimentin | 1:200 | |
CD31 | 1:250 | |
CD45 | 1:250 | |
Secondary Antibody | Dilution | |
Goat anti-mouse Alexa Fluor 488 | 1:200 | |
Goat anti-rabbit-FITC | 1:200 | |
Goat anti-rabbit-Cy3 | 1:200 | |
Goat anti-rat Alexa Fluor 488 | 1:200 | |
Goat anti-rat Alexa Fluor 647 | 1:200 |
Table 2: Immunofluorescence primary and secondary antibodies.
Fibroblasts are a heterogenous group of cells, identified by diverse set of markers. The protein markers that have been used to identify fibroblasts are discoidin domain receptor 2 (DDR2), fibronectin, vimentin, collagen I and III, and Thy115,16,17,18,19,20. Whereas vimentin has been used to identify uninjured quiescent cardiac fibroblasts, fibroblast specific protein 1, αSMA, and periostin have been shown to identify injury-induced activated fibroblasts, with αSMA being the most common marker to detect activated fibroblasts6,12,21. Additionally, Tcf21 and MEFSK4 proteins have gained recent recognition in recognizing both quiescent fibroblasts found in uninjured cardiac tissue as well as activated fibroblasts including myofibroblasts found in injured mouse hearts21,22.
This protocol utilizes three different approaches to isolate and enrich fibroblasts and activated fibroblasts including myofibroblasts. The fibroblast's ability to preferentially adhere to plastic is used in the first approach for isolation. Following enzymatic digestion with liberase, the single cell suspension of cells is seeded on a plastic dish to preferentially adhere. The inability of many non-fibroblast cells to adhere to the polystyrene surface of Petri dishes allows us to remove all media from the dish and remain with a relatively pure population of fibroblasts. A caveat of using this technique is although fibroblasts will preferentially adhere to the polystyrene dish, some contaminating non-fibroblast cells may also attach, leaving a non-homogenous population of cells.
The second isolation technique utilizes FACS to separate αSMA expressing myofibroblasts from other cells. In the transgenic mouse model employed here, GFP is exclusively expressed along with αSMA, so myofibroblasts containing αSMA can be detected by a FACS machine through the fluorescent capabilities of GFP. This isolation procedure enables us to obtain a population of cells that is approximately 99% myofibroblast. The purity analyses of these cells have been extensively described by Saraswati et al.10.
The third isolation technique is an efficient way to isolate both uninjured and activated fibroblasts by magnetic bead-based separation of MEFSK4 expressing cell. By allowing a single cell suspension to bind to anti-CD45 and anti-CD31 magnetic beads and become immobilized in a matrix due to magnetic field effects, this allowed the separation of any hematopoietic as well as endothelial cells that may have contaminated the fibroblast isolation. As MEFSK4 has been recently used as a reliable marker to identify fibroblasts, an antibody that will bind to MEFSK4 expressing cells is able to be applied. After binding a magnetic bead to the antibody, creating a complex that allows isolation of fibroblasts, the magnetic bead-cell complex is passed through a matrix in a magnetic field, and a highly enriched fibroblast population is obtained. The purity of the isolated fibroblast population should be assessed by immunostaining, RTPCR, and flow cytometry analyses.
As with any other technique, there are limitations with the techniques described in this manuscript. The limitation of the selective adhesion protocol and magnetic bead-based isolation is that these methods do not differentiate between quiescent and activated fibroblasts. In order to enrich activated fibroblasts, the isolation should be performed 8–10 days following myocardial infarction. Additionally, it is important to check the purity of the isolation with other fibroblast markers. MEFSK4-positive fibroblast purity has been demonstrated only by RTPCR, it is recommended to test (by immunostaining and flow cytometry analysis) with other fibroblast markers and markers that recognize contaminating cell types including hematopoietic (CD45), endothelial (CD31), and pericytes (AN2). If possible, other fibroblast specific markers could be used to further sort or magnetically isolate the fibroblast population.
Using αSMA-GFP mice to isolate and sort myofibroblasts is a reliable technique to obtain an activated fibroblast population. However, a negligible percentage of hematopoietic and endothelial cells has been observed in the flow analysis. To improve upon this technique, CD45+ve/CD31+ve and AN2+ve cells should be excluded from the GFP+ve/αSMA+ve cell sorting. Since αSMA is a widely accepted marker of myofibroblasts, the αSMA-GFP reporter mouse model is a valuable tool that should be exploited to study myofibroblasts in the context of myocardial injuries.
There are several crucial troubleshooting steps that must be taken into account. Digestion time can be decreased if cell viability and yield is affected. A stir bar should not be used to stir the digestion mixture, as this affects cell viability. The tube should be secured on a rocker or in a shaking incubator to agitate the digestion mixture gently. Resuspending the digested tissue 10x with a 5 mL or 10 mL pipette is crucial for proper dissociation of cells.
Proper red blood cell lysis of the single cell suspension must be utilized if cells are going to be sorted or analyzed by flow cytometry. For magnetic bead isolation, degassing of the buffer is essential to prevent the introduction of any air bubbles in the column. The column used for magnetic bead cell isolation should not be reused between different magnetic bead-conjugated cells. For example, a new column should be used to separate CD45+ cells, which should be discarded after elution of the CD45+ cells, then another new one for the CD31+ cell isolation. In our hands, we have not seen contamination of pericytes in isolated/sorted fibroblasts. However, MEFSK4 has been shown to recognize pericytes22. It is therefore recommended to use an additional step to sort out/magnetically deplete pericytes (AN2) from the single cells. Although this protocol is validated in 12 week-old mice, the technique may be used for younger or older mice.
The authors have nothing to disclose.
The authors want to thank Dr. Ivo Kalajzic for the αSMA-GFP mice. Research reported in this publication was supported by the National Institute of General Medical Sciences of the National Institutes of Health (NIH) under Award Number R01GM118300 (S.S.), National Institute of Biomedical Imaging and Bioengineering of the NIH under Award Number R21EB019509 (P.P.Y.), and Scientist Development Grant of the American Heart Association under Award Number 17SDG33630187 (S.S.). Flow cytometry analyses were performed at the VUMC Flow Cytometry Shared Resource which is supported by the Vanderbilt Ingram Cancer Center (P30 CA68485) and the Vanderbilt Digestive Disease Research Center (DK058404).
Reagents | |||
Acetone | |||
Anti-fungal (Amphotericin B-solubilized; Fungizone) | Sigma Aldrich | A9528 | |
Bovine Serium Albumin (BSA) | Sigma | 9048-46-8 | |
Calcium chloride | |||
Citrate Buffer | |||
Collagenase blend (Liberase Blendzyme 3 TH) | Roche Applied Science | ||
DAPI | |||
DDI water | |||
DI water | |||
DMEM-F12 with L-Glutamine and HEPES | Life technologies | 11330057 | |
Dnase I(20U/mL) | BioRad | 7326828 | |
Dulbecco's Phosphate-Buffered Saline (dPBS) without Ca2+ and Mg2+ | Gibco | 13190-144 | |
70% Ethanol | |||
FC Blocker (Purified anti-mouse CD16/CD32) | Tonbo Biosciences | 70-0161 | |
Fetal Bovine Serum (FBS) | Life technologies | 16000044 | |
10% goat serum | |||
Hank's Balanced Salt Solution (HBSS) with Ca2+ and Mg2+ | Corning | 21-023-CV | |
1M HEPES | Corning | 25-060-Ci | |
Krebs-Henseleit Buffer powder | Sigma | K3753 | |
Mycoplasma prophylactic (Plasmocin) | Invivogen | ant-mpp | |
Penicillin/Streptomycin | Thermo Fisher Scientific | 15140122 | |
1x Phosphate-Buffered Saline (PBS) | |||
10x Red Blood Cell Lysis Buffer | Miltenyi | 130-094-183 | |
Slow-fade Mounting Media | |||
Sodium azide | |||
Sodium bicarbonate | |||
TGFβ | |||
Trypan Blue Stain (0.4%) | Gibco | 15250-061 | |
Type 1 Rat Collagen | |||
Antibodies | |||
7AAD (stock: 1 mg/mL solution in DMSO) | Molecular Probes | A1310 | dilution = 1:1000; RRID = |
CD45-APC | BD Bioscience | 559864 | dilution = 1:200; RRID = AB_398672 |
CD31-PE | BD Bioscience | 553373 | dilution = 1:200; RRID = AB_394819 |
CD31 | BD Biosciences | 553370 | dilution = 1:250; RRID = AB_394816 |
CD45 | BD Biosciences | 553076 | dilution = 1:250; RRID = AB_394606 |
COL 1α1 | MD Bioproducts | 203002 | dilution = 1:1000; RRID = |
Ghost dye violet 510 (Formulation: 1 uL/test in DMSO) | Tonbo Biosciences | 13-0870 | dilution = 1:1000; RRID = |
Goat anti-mouse Alexa Fluor 488 | Molecular Probes | A11029 | dilution = 1:200; RRID = AB_138404 |
Goat anti-rabbit-Cy3 | Southern Biotech | 4050-02 | dilution = 1:200; RRID = AB_2795952 |
Goat anti-rabbit-FITC | Jackson Immunoresearch Laboratories | 711-165-152 | dilution = 1:200; RRID = AB_2307443 |
Goat anti-rat Alexa Fluor 488 | Molecular Probes | A11006 | dilution = 1:200; RRID = AB_2534074 |
Goat anti-rat Alexa Fluor 647 | Thermo-Fisher | A21247 | dilution = 1:200; RRID = AB_141778 |
Periostin | Santa Cruz | SC67233 | dilution = 1:100; RRID = AB_2166650 |
Vimentin | Sigma Aldrich | V2258 | dilution = 1:200; RRID = AB_261856 |
α-smooth muscle actin (αSMA) | Sigma Aldrich | A2547 | dilution = 1:1000; RRID = AB_476701 |
Fibroblast specific protein 1 (FSP1) | Millipore 07-2274 | 07-2274 | dilution = 1:100; RRID = AB_10807552 |
CD45 Magnetic Beads | Miltenyi Biotec | 130-052-301 | |
CD31 Magnetic Beads | Miltenyi Biotec | 130-087-418 | |
Anti-feeder cells-APC (MEFSK4) | Miltenyi Biotec | 130-102-900 | dilution = 1:100; RRID = AB_2660619 |
anti-APC Beads | Miltenyi Biotec | 130-090-855 | |
Rat IgG-APC | Miltenyi Biotec | 130-103-034 | dilution = 1:100; RRID = AB_2661598 |
Donkey anti-rat Alexa Fluor405 | Abcam | ab175670 | dilution = 1:100 |
anti-AN2/NG2 | Miltenyi Biotec | 130-097-455 | dilution = 1:11; RRID = AB_2651235 |
Other Materials | |||
0.22 µm Filter | Thermo Scientific | 723-2520 | |
10 cm2 Cell Culture Dish | Corning | 430167 | |
10 mL Pipet | Fisherbrand | 13-678-11E | |
40 µm Cell Strainer | Fisherbrand | 22363547 | |
5 mL Pipet | Fisherbrand | 13-678-11D | |
50 mL Conical Tube | Falcon | 352070 | |
6-well Plate | Corning | 3506 | |
Flow Cytometry Tubes | Falcon | 352058 | |
Forceps | |||
Rocker | |||
Single Edge Blade | PAL | 62-0177 | |
Surgical Scissors | |||
GFP-αSMA Reporter Mice | |||
MACS Separator Magnetic Field | |||
MACS Separation Column | |||
Coverslips | |||
Qaigen Rneasy Mini Kit | Qaigen | 74104 | |
Ambion RNAqueous Micro Total Isolation Kit | Ambion | AM1931 | |
BioRad iScript cDNA Syntehsis Kit | BioRad | 1708891 | |
48-well Plate | |||
30G Needle | |||
3 Laser Flow Cytometry Machine (BD LSRFortessa) | BD Biosciences | ||
4 Laser Flow Cytometry Machine (BD FACSAria III) | BD Biosciences | ||
Flow Data Acquiring Software (BD FACSDiva Software v8.0a) | BD Biosciences | ||
Flow Data Analysis Software (FlowJo Software) | BD Biosciences |