Published: May 14th, 2021
This paper aims to provide a protocol for the isolation and culture of primary cancer-associated fibroblasts from a syngeneic murine model of triple-negative breast cancer and their application for the preclinical study of novel nanoparticles designed to target the tumor microenvironment.
Cancer-associated fibroblasts (CAFs) are key actors in the context of the tumor microenvironment. Despite being reduced in number as compared to tumor cells, CAFs regulate tumor progression and provide protection from antitumor immunity. Emerging anticancer strategies aim to remodel the tumor microenvironment through the ablation of pro-tumorigenic CAFs or reprogramming of CAFs functions and their activation status. A promising approach is the development of nanosized delivery agents able to target CAFs, thus allowing the specific delivery of drugs and active molecules. In this context, a cellular model of CAFs may provide a useful tool for in vitro screening and preliminary investigation of such nanoformulations.
This study describes the isolation and culture of primary CAFs from the syngeneic 4T1 murine model of triple-negative breast cancer. Magnetic beads were used in a 2-step separation process to extract CAFs from dissociated tumors. Immunophenotyping control was performed using flow cytometry after each passage to verify the process yield. Isolated CAFs can be employed to study the targeting capability of different nanoformulations designed to tackle the tumor microenvironment. Fluorescently labeled H-ferritin nanocages were used as candidate nanoparticles to set up the method. Nanoparticles, either bare or conjugated with a targeting ligand, were analyzed for their binding to CAFs. The results suggest that ex vivo extraction of breast CAFs may be a useful system to test and validate nanoparticles for the specific targeting of tumorigenic CAFs.
During the past decades, it has become clear that killing tumor cells is usually not sufficient to eradicate malignancy, as the tumor microenvironment may prompt tumor relapse and induce therapeutic resistance1,2. A novel paradigm has then emerged: targeting tumor stroma to deprive the tumor of supporting factors and thus, boost the efficacy of chemotherapeutics3,4,5. In particular, cancer-associated fibroblasts (CAFs) are an interesting stromal target in many solid tumors6,7. CAFs are a very heterogeneous group of cells that interact with cancer cells and cells of the immune system through the secretion of growth factors, cytokines, and chemokines; build up and remodel the extracellular matrix; and enable metastasis formation8,9,10,11,12. Depending on the tumor type, CAFs show pro-tumorigenic functions, while other subtypes of CAFs seem to have tumor-suppressive functions13,14. To better clarify this dichotomy, a thorough characterization of CAFs from primary and metastatic tumors is important.
In this context, an emerging field of research has focused on the development of nanosized agents designed to target and/or destroy CAFs by delivering active molecules and drugs able to remodel the tumor microenvironment15,16,17,18. Several types of nanoparticles have been designed to achieve CAF ablation by cytotoxic drugs, to induce CAF-targeted photodynamic therapy, or to reprogram CAFs by reverting them to a quiescent state or inducing TNF-related apoptosis induced ligand expression, which induces apoptosis of neighboring cancer cells16,19. Moreover, the potential of many nanoparticles to actively target specific biological markers gives rise to the hope of selecting CAF subsets to target. Although its absolute specificity for CAF is still questioned, fibroblast activation protein (FAP) is one of the most promising targets of pro-tumorigenic stroma and is exploited to steer nanodrug delivery, thus paving the path for the development of CAF-targeting nanotherapeutics20,21,22.
This paper describes the isolation of primary CAFs from a syngeneic model of murine breast cancer and reports their use in the study of the targeting capability of nanoparticles engineered to recognize the CAF marker, FAP. Ferritin nanocages are used as candidate nanosystems to set up the method, as their specificity of delivery may be shaped by the surface exposure of targeting moieties23,24. Moreover, ferritins have been successfully proven to be excellent biocompatible shuttles for antitumor applications, triggering rapid accumulation of the payload in the tumor mass25,26,27. To date, preclinical studies of CAF-targeting nanosystems have involved in vitro testing on fibroblast cell lines stimulated in culture with transforming growth factor-beta to induce cell activation and the expression of some immunophenotypic features of CAFs28,29. This method is usually applied to immortalized cell lines (such as NIH3T3, LX-2) and is quite rapid and simple, yielding activated cells in a few hours or days. A limitation is that although in vitro stimulation induces the expression of some genes attributed to activated myofibroblasts, it cannot entirely recapitulate all the biological features of real CAFs, especially their heterogeneity in vivo.
Another strategy involves the extraction of primary CAFs from human or mouse tumor samples30,31. This ensures that CAF activation occurs in a physiological context, and that the heterogeneity of CAF subpopulations is maintained. According to the research objective, CAFs may be derived from different sources, thus offering the possibility to study the most reliable condition. The protocol reported here would be valuable for scientists who seek to perform a preliminary evaluation of the functionality of novel nanoparticles designed to target CAFs from a murine breast cancer model. Isolated CAFs would be useful for screening those nanoparticles that are promising enough to proceed for in vivo evaluation in animal models of cancer. This will be relevant during the first steps of nanoparticle production, driving nanotechnologists toward the refinement of nanoparticle design by mainly considering the strategy of ligand immobilization to achieve optimal targeting properties.
1. Establishing a syngeneic 4T1 model of breast cancer
NOTE: The present protocol describes the isolation of primary CAFs from a mouse 4T1 breast tumor. The animal study described here has been approved by the Italian Ministry of Health (aut. number 110/2018-PR).
2. Tumor dissociation into single cells
NOTE: For the following steps, use sterile reagents and disposables in a laminar flow hood. Work with 4 tumors at a time.
3. Extraction of primary CAFs from breast tumor
NOTE: For section 3, use the mouse Tumor-Associated Fibroblast Isolation kit containing Non-Tumor-Associated Fibroblast Depletion Cocktail and Tumor-Associated Fibroblast Microbeads suited for magnetic labeling of the cells (see the Table of Materials).
4. Process validation
5. CAFs targeting by engineered Ferritin nanoparticles
NOTE: A recombinant variant of human ferritin heavy chain (HFn) was used as bare nanoparticle or was conjugated with targeting moieties. Here, HFn nanoparticles functionalized with the variable portion of an anti-FAP antibody (Fab@FAP) were prepared by the NanoBioLab at University of Milano-Bicocca at two HFn:Fab@FAP molar ratios, 1:1 and 1:5, according to a previously described protocol32.
6. Statistical analysis and experimental replicates
In vivo model set-up for optimal CAFs isolation
The injection of 105 4T1-luc cells into the mammary fat pad of female BALB/c mice leads to the growth of a detectable tumor mass at 5 days after implantation. By measuring the tumor volume by calipers and the tumor cell viability by BLI, tumor growth was monitored for one month after implantation. To find a sacrifice window that is adequate for CAFs isolation, an optimal compromise was sought between higher tumor size and BLI on the one hand and an emerging tumor ulceration and necrosis on the other hand (Figure 1). As a necrotic core appears 20 days after implantation, and it enlarges at 25 and 30 days (as documented by BLI images in Figure 1C), day 20 was set as the time point to optimize cell recovery after the isolation process. Even after carefully removing all visible necrotic areas during the first steps of tumor handling ex vivo, a high percentage of dead cells was found at the end of dissociation into single cells (Table 1). As this percentage may be relevant, especially with the increase in tumor size, dead cell removal is always necessary when working with the 4T1 model.
Optimization of CAFs isolation procedure, culture, and characterization
Two more passages are needed to isolate the population of CAFs (CD90.2+ CD45-) from the panel of collected viable cells: the depletion of non-tumor-associated fibroblasts and the enrichment of tumor-associated fibroblasts (Figure 2). The depletion bead cocktail efficiently removes CD45+ cells (accounting for 67.35% and 0.69% of total cells pre- and post-depletion, respectively, Table 2), and was always used to process one single tumor in every column. As shown in Table 1, the number of eluted cells dropped from an average of 3 × 106 to 1 × 105 after the depletion step. Due to this massive decrease in total cell number, it is convenient to pool the collected cells from at least 2 to a maximum of 4 tumors in a single tube before proceeding with the enrichment step. By doing so, an adequate number of cells was obtained for incubation with tumor-associated fibroblast beads and passed through a single separation column to obtain a final average of 93% of CD90.2+ CD45- cells (Table 2).
Once seeded onto tissue culture plates, these recovered cells attached to the plastic and revealed a large spindle-shaped morphology typical of fibroblasts (Figure 3A,B) and different from 4T1 tumor cells (Figure 3D). Not pooling tumors after depletion causes the cellular yield with tumor-associated fibroblast microbeads to be too low to establish a culture. In other cases, when the duration and temperature of the incubations with microbeads are not carefully maintained, some non-specific binding may occur. In such cases, the enrichment steps were less efficient and higher percentages of CD90.2- CD45+ and CD90.2- CD45- cells were recovered along with the CD90.2+ CAFs (5.97 ± 1.5 and 16.75 ± 1.1, respectively) (Figure 4 and Table 3). These contaminant cells were likely to be responsible for the presence in culture of small clones with different morphology (Figure 4B, black arrowheads) that grew faster than CAFs and prevailed over the primary CAF culture (Figure 4C). These suboptimal results confirmed the importance of always double-checking both CD90.2 and CD45 expression, as well as cell morphology at the end of the enrichment process and during cell growth in culture.
Use of isolated cells to evaluate CAF-targeting potential of engineered nanodrugs
The freshly isolated CAFs can be used for several applications ranging from basic research to pharmacological studies. This group’s aim is to develop HFn nanocages that can specifically target CAFs. HFn was functionalized with a specific anti-FAP antibody fragment (HFn-FAP) at two different protein:antibody ratios (a lower 1:1 and a higher 1:5), and their binding with CAFs was tested. As FAP was used as a surface biomarker of pro-tumorigenic CAFs, it was of fundamental importance to check FAP expression on isolated CAFs (Figure 3C). FAP expression was followed over 5 passages in culture to confirm that the primary CAF culture maintained its original characteristics.
FAP functionalization on HFn was found to contribute to a significant shift toward CAF targeting as compared to bare HFn, and the lower amount of antibody (1:1) was enough to observe this effect (Figure 5A). However, this was not observed with the tumor 4T1 cells used to set up the in vivo tumor model, wherein bare HFn showed higher binding than functionalized HFn (Figure 5B). This was most likely due to the absence of FAP overexpression in 4T1 cells and the preferential interaction of HFn with TfR1, which regulates HFn uptake in cells, as widely reported by this group27,33. These results confirm the usefulness of using primary cultures of breast CAFs to preliminarily screen the targeting capability of nanoparticles designed to tackle the tumor microenvironment.
Figure 1: Establishment of the 4T1 tumor model. Cells (105 cells/mouse) were injected into the mammary fat pad. Tumor growth was followed at days 5, 10, 15, 20, 25, and 30 after implantation (A) by measuring tumor volume with calipers and (B) by bioluminescence imaging. Tumor volumes and BLI are expressed as mm3 and counts, respectively. Results are reported as average ± SEM (n=6). (C) BLI representative images obtained 5, 10, 15, 20, 25, and 30 days after cell implantation confirm tumor growth until day 25, when it seems to reach a plateau. At the last time point of analysis (30 days), the BLI does not increase as compared to day 25. Starting at day 20, areas of necrosis and ulceration start to become visible in the central part of the tumor. Color Scale: Min = 1,194, Max = 20,462. Abbreviations: BLI = bioluminescence imaging; SEM = standard error of the mean. Please click here to view a larger version of this figure.
Figure 2: Tumor sample preparation and flow cytometry characterization of isolated CAFs. Tumors were excised and reduced into (A) small pieces of approximately 1–2 mm with the help of a scalpel; (B) single cell suspension after tissue digestion and mechanical dissociation of an excised tumor; (C) cell pellet obtained after lysis of red blood cells; (D) flow cytometry analysis of CD45 and CD90.2 expression in cells obtained after removal of red blood cells and dead cells, (E) after depletion of non-cancer-associated fibroblasts, and (F) after enrichment of cancer-associated fibroblasts, where the majority of cells are CD90.2+ CD45- (blue rectangle). Abbreviations: CAFs= cancer-associated fibroblasts; CD = cluster of differentiation; PE-A = area of phycoerythrin; FITC-A = areas of fluorescein isothiocyanate. Please click here to view a larger version of this figure.
Figure 3: Morphological analysis of CAFs and FAP expression. (A, B) CAFs morphology was checked throughout all passages of culture by optical microscopy (different levels of confluence at passage 2 and passage 5) and compared to (D) 4T1 tumor cells; scale bars = 10 µm. (C) Fibroblast activation protein (FAP) expression was evaluated by flow cytometry at the end of the isolation process on the collected CD90.2+ CD45- cells to confirm their molecular characteristics. A fluorescence threshold was set on unstained control cells (red graph) to quantify mean fluorescence intensity and the percentage of positive cells (FAP+) among antibody-stained cells (blue graph). Abbreviations: CAFs= cancer-associated fibroblasts; FAP = fibroblast activation protein; CD = cluster of differentiation; MFI = mean fluorescence intensity; Ab = antibody; FITC-A = area of fluorescein isothiocyanate; FAP+ = FAP-positive cells. Please click here to view a larger version of this figure.
Figure 4: Example of a sub-optimal CAFs isolation process. (A) Flow cytometry evaluation after the final isolation step revealed the presence of contaminant CD90.2- CD45+ and CD90.2- CD45- cells. (B) These cells can be seen as clones with small round morphology upon seeding (black arrowheads and inset in the bottom left corner of the panel) that (C) prevailed over CAFs after the third passage in culture. Scale bars = 10 µm. Abbreviations: CAF = cancer-associated fibroblast; CD = cluster of differentiation; PE-A = area of phycoerythrin; FITC-A = areas of fluorescein isothiocyanate. Please click here to view a larger version of this figure.
Figure 5: Binding of HFn nanocages on CAFs and 4T1. HFn nanocages were fluorescently labeled with FITC, functionalized with an anti-FAP antibody fragment (HFn-FAP) at two different protein-antibody ratios (1:1 and 1:5), and incubated with (A) target CAFs and (B) 4T1 cells at 4 °C for 2 h. Binding was evaluated by flow cytometry. (A) HFn-FAP binding with CAFs is significantly increased at both antibody fragment concentrations by three-fold as compared to bare HFn. (B) in contrast, a significantly higher binding of bare HFn was observed in 4T1 cells, where binding is not enhanced by FAP recognition. Results are reported as average ± standard deviation of three independent experiments. *** p = 0.0003; °° p = 0.0021 ; °°° p = 0.0008. Abbreviations: CAFs= cancer-associated fibroblasts; FAP = fibroblast activation protein; HFn = recombinant variant of human ferritin heavy chain used as bare nanoparticle or conjugated; HFN-FAP = HFn nanoparticles functionalized with the variable portion of an anti-FAP antibody prepared at two HFn:Fab@FAP molar ratios, 1:1 and 1:5. Please click here to view a larger version of this figure.
|Cell Number (average ± SD)
|Extraction Yield Total (per passage)
|Excised tumor (Post red blood cell removal)
|1.27 x 108 ± 9.81 x 107
|Excised tumor (Post dead cell removal)
|3.01 x 106 ± 9.61 x 105
|Post depletion cocktail
|1.15 x 105 ± 4.95 x 104
|3.00 x 104 ± 1.2 x 104
Table 1: Total cell count after each step of isolation of cancer-associated fibroblasts.
|Cell Distribution (%)
|Excised tumor (Post dead cell removal)
|1.81 ± 0.98
|10.78 ± 4.51
|56.57 ± 14.05
|28.20 ± 17.57
|Post depletion cocktail
|69.33 ± 16.75
|0.14 ± 0.13
|0.55 ± 0.63
|39.89 ± 30.31
|93.14 ± 3.3
|0.09 ± 0.1
|0.09 ± 0.08
|6.69 ± 3.4
Table 2: Cell distribution according to expression of CD90.2 and CD45 after each passage of isolation of cancer-associated fibroblasts.
|Cell Distribution (%)
|Post Enrichment (no pooling)
|75.80 ± 2.9
|1.49 ± 0.4
|5.97 ± 1.5
|16.75 ± 1.1
Table 3: CD90.2 and CD45 expression of cells collected after a sub-optimal cancer-associated fibroblast isolation experiment.
CAFs are emerging as key players in remodeling extracellular matrix, promoting metastasis progression, and limiting drug access to the tumor site34. However, due to their heterogeneity, their roles are still controversial—some CAFs are tumorigenic, whereas some other subtypes seem to have a tumor-suppressive role. In this context, their isolation can be of extreme interest to shed more light on their much-debated role in cancer progression, which will have important clinical implications12,31. Moreover, successful CAF extraction from human and preclinical mouse tumor models would also facilitate the development of new CAF-targeting drugs. This paper reports a method to efficiently isolate and culture primary CAFs from a syngeneic preclinical model of breast cancer. Based on experience, three experimental steps mostly influence the success of the protocol.
The first one is working fast to avoid aggregation of cell suspensions associated with the risk of clotting in the column and slowing cell efflux: in fact, when cells that passed through the column were clustered together, the probability of a sub-optimal isolation process increased, and the final cultures contained “contaminant” cell clones. The second one is pooling cells from different tumors after the depletion passage to guarantee enough cells to be passed through in the final enrichment step. The final critical issue is plating the extracted cells at high cell densities, most likely starting from a single well of the 24-well plate to boost cell growth and colony expansion for up to 4–5 passages. If the cells were seeded at low densities, they expanded on the free surface and quickly stopped replicating.
Several studies have already described CAF extraction methods of both human and mouse origin, to study their role in promoting cancer development and invasiveness. This has been done in many types of tumors, including breast cancer, melanoma, cholangiocarcinoma, and pancreatic adenocarcinoma35,36,37,38. However, due to the lack of specific CAF markers and to their heterogeneity, the results of these processes are still sub-optimal.
Here, the isolated cells were used to validate the CAF-targeting ability of HFn nanocages functionalized with anti-FAP targeting moieties. The use of primary cultures of CAFs was a significant advantage, allowing the validation of the nanostrategy ex vivo with a much simpler and cost-effective binding experiment than doing it directly in vivo at this preliminary phase of nanodrug optimization.
In this sense, ex vivo testing on CAFs can precede in vivo animal experiments by allowing the screening and selection of the most promising nanoparticles for the optimal targeting of CAF. The CAF model presented here may also be used for preliminary efficacy studies, when loading nanoparticles with active drugs. Animals will be only used later to evaluate biodistribution, pharmacokinetics, and efficacy studies.
Several CAF-targeting nanodrugs are being developed, such as ferritin nanocages for photodynamic therapy in breast cancer and peptide-based particles to deliver doxorubicin in prostate cancer models39,40. However, not many studies have focused on the isolation of CAFs as cell platforms for nanodrug optimization.
This protocol has some limitations. First, it is a time-consuming protocol, requiring the preparation and purchase of several reagents and materials. Second, due to the scarcity of CAFs in the 4T1 breast tumor, their yield is low. Nanoparticle experiments should be planned and performed as soon as the required cell number is achieved, as primary CAFs undergo senescence and cannot be maintained in culture for a long time. In conclusion, this method of CAFs isolation, culturing, and characterization can be a powerful tool to accelerate the development of new targeted nanomedicines in the fight against cancer.
The authors have nothing to disclose.
This study was supported by Associazione Italiana per la Ricerca sul Cancro (AIRC) under IG 2017-ID. 20172 project – P.I. Corsi Fabio. SM acknowledges Pediatric Clinical Research center “Romeo and Enrica Invernizzi” that supports her position. AB thanks AIRC (ID. 20172 project) and University of Milan for research fellowship. LS and MS postdoctoral and doctoral fellowships are supported by University of Milan.
|ACK Lysing buffer
|Store at +15° to +30 °C.
|Alexa Fluor 488 goat anti-human antibody
|Protect from light.
|Anti-FAP Fab fragment (3F2)
|7 weeks old female mice
|Bovine Serum Albumin
|Store at 2-8 °C.
|FITC-conjugated fluorescent antibody; clone REA 737; Store protected from light at 2–8 °C.
|PE-conjugated fluorescent antibody; clone 30-H12; Store protected from light at 2–8 °C.
|Cell strainers 70 µm
|laser configuration B4-R2-V0
|Dead cell Removal Kit
|contains 10 mL of Dead Cell Removal MicroBeads; 25 mL of 20× Binding Buffer Stock Solution. Store protected from light at 2−8 °C.
|reagent for in vivo imaging
|DMEM High Glucose w/o L-Glutamine w/ Sodium Pyruvate
|Warm at 37 °C in a water bath before use.
|DPBS w/o Calcium, w/o Magnesium
|Keep at room temperature.
|Ethylenediaminetetraacetic acid (EDTA)
|Fetal Bovine Serum
|Before use, heat at 56 °C for 30 min to kill all the complement proteins.
|Fluorescein Istothiocyanate isomer I (FITC)
|Store protected from light at 2–8 °C.
|GentleMACS C Tubes
|They are used in combination with the gentleMACS Dissociator.
|Two samples can be processed in parallel. Special protocols have been developed for various tissues.
|Ham's F12 w/o L-Glutamine
|Warm at 37 °C in a water bath before use.
|IVIS Lumina II imaging system
|Caliper Life Sciences
|This imaging system is easy to use for both fluorescent and bioluminescent imaging in vivo. Equipped with a Living Image Software for analysis
|Composed of ferromagnetic spheres, designed for stringent depletion of unwanted cells. Position the required number of columns on the magnetic Separation Unit and equilibrate with buffer before use
|200 mM stock
|Light/fluorescence microscope equipped with camera
|DM IL LED Fluo/ ICC50 W CameraModule
|inverted microscope for live cells with camera
|MACS Separation Unit
|Position the magnet on appropriate stand before use.
|MACS Tissue Storage Solution
|Composed of ferromagnetic spheres, designed for positive selection of cells. Position the required number of columns on the magnetic Separation Unit and equilibrate with buffer before use.
|Mycoplasma Removal Agent
|Dilute in culture medium 1:100.
|Thermo Fisher Scientific
|Non essential aminoacids
|Recombinant Human apoferritin H-homopolymer (HFn)
|RPMI 1640 w/o L-Glutamine
|Warm at 37 °C in a water bath before use.
|Tumor-Associated Fibroblast Isolation kit
|Contains 1 mL of Non-Tumor-Associated Fibroblast Depletion Cocktail, mouse; 1 mL of CD90.2 (Tumor-Associated Fibroblast) MicroBeads,mouse. Store protected from light at 2-8 °C.
|Tumor Dissociation Kit
|Contains lyophilized enzymes (D, R, A) and buffer A; reconstitute enzymes according to the manufacturer's instructions, aliquot and store at - 20 °C.
|Trypan blue 0.4%
|dilute 1:1 with a sample of cell suspension before counting the cells
|use at room temperature
|Warm at 37 °C in a water bath before use
|T75 Primo TC flask
|Zeba Spin Desalting Columns
|Thermo Fisher Scientific
|7K MWCO, 2 mL
|1.5 mL tubes
|15 mL tubes
|Mouse mammary gland cancer cell line stably transfected with firefly luciferase gene
|50 mL tubes
Copyright © 2024 MyJoVE Corporation. All rights reserved