Basic Fibroblast Growth Factor 2 Is a Determinant of CD4 T Cell–Airway Smooth Muscle Cell Communication through Membrane Conduits

Once activated, CD4 T cells acquire the ability to migrate to the site of inflammation, a process that enables them to communicate with a variety of immune and structural cell types. A range of molecular mechanisms ensures optimal communication of CD4 T cells with their cellular environment via cytokines and growth factors, exosomes, direct cell-to-cell contact, membrane nanotubes (MNTs), and trogocytosis (1–8). CD4 T cells have been shown in close proximity to airway smooth muscle cells (ASMCs) in vivo, indicating the plausibility that contact-dependent communication of CD4 T cells and ASMCs may be responsible for altered functionality of these cells in airway diseases with increased lymphocyte infiltration (9, 10). In several airway pathologies, including asthma, increased thickness of ASM bundles as a result of hyperplasia and hypertrophy may account for excessive airway narrowing in response to various stimuli (11). The tubular connections that have usually been referred to as MNTs have been shown previously to connect CD4 T cells to ASMCs, T cells, and B cells (6, 7, 12). In vitro, upon activation, primary human CD4 T cells interact with ASMCs via MNTs, allowing myeloid cell leukemia protein-1 (Mcl-1) transfer from ASMCs to CD4 T cells and resulting in suppression of CD4 T cell apoptosis (6). Hence characterization of the structure of CD4 T cell–derived lymphocyte-derived membrane conduits (LMCs) communicating with ASMCs, as well as the mediator(s) inducing their formation, is essential for understanding how intercellular communication between ASMCs and CD4 T cells alters their functionality in airway diseases.

MNTs are a continuum of cell membrane extensions forming bridges between cells. MNTs can be formed by a variety of cell types including neural, cancer, stem, and stromal cells, as well as lymphoid and myeloid immune cells (6, 7, 12–19). MNT functionality is determined by their size and cytoskeletal composition, whether containing either actin alone or also tubulin (20). Aside from the transfer of surface proteins because of membrane fluidity, MNTs can transfer calcium ions, microRNAs, vesicles, organelles, and pathogens (12–14, 21–25). Although there is very limited information concerning the cause and molecular mechanisms involved in the formation of this mode of intercellular communication, a number of conditions including cellular stress and availability of cytokines and specific ligands such as Fas are known to initiate MNT formation (5, 13, 23, 26, 27). However, as a process highly dependent on cell types and microenvironments, mechanisms controlling MNT formation are likely not shared among all cell types.

The fibroblast growth factor (FGF) family, consisting of 22 structurally related polypeptides in human, is involved in physiological processes by controlling cell proliferation, migration, survival, and differentiation, as well as embryonic development. FGFs exert their activity by binding to surface receptor tyrosine kinases, FGF receptors (FGFRs) 1–4. Basic FGF2 (FGF2b) induces chemotaxis, cell adhesion, migration, and neuronal proliferation. Furthermore, FGF2b induces formation of neurite outgrowths and branching mediated by FGFR1 (28–32), suggesting their possible role in MNT formation that shares the same morphological features.

In this study, we hypothesized that the formation of neurite outgrowths may share the same mechanisms and triggering factors by T cells for the formation of MNT-like structures. We have opted to use the term LMCs throughout because the range of sizes of LMCs often exceeds the nanometer scale. We have demonstrated that the formation of LMCs by CD4 T cells, rich in F-actin and tubulin, induces marked changes in the CD4 T cell membrane structure topologically. Importantly, as a factor inducing actin polymerization in CD4 T cells, FGF2b was expressed by ASMCs that induced LMC formation by the T cells. FGFR1 was established as a receptor on the T cell surface with the potential to mediate the effect of FGF2b. LMCs appear to transfer mitochondria from ASMCs to CD4 T cells, a process sensitive to LMC frequency. Asthmatic ASMCs were found to express more FGF2b compared with control ASMCs, inducing more LMC formation and mitochondrial transfer.

DMEM, RPMI 1640, FBS, penicillin, streptomycin, and amphotericin B (PSA), PBS, Cell light Mitochondria-RFP, BacMan 2.0, MitoTracker Red CM-H2Xros, calcein AM, FITC-conjugated phalloidin, and recombinant human FGF2b were purchased from Invitrogen. Anti-human CD4 Abs, conjugated with PerCP, Alexa Fluor 647, and Brilliant Violet 421, were purchased from BioLegend. Rabbit anti-human FGF2b neutralizing Ab, rabbit anti-human FGFR1, rabbit anti-human α/β-tubulin, Alexa Fluor 647–conjugated donkey anti-rabbit IgG, and Fluoroshield mounting medium with DAPI were purchased from Abcam. Viability dye, eFluor 780, was purchased from eBioscience. PMA and ionomycin were purchased from Sigma-Aldrich. MACS kit, negative selection (MACS), was obtained from Miltenyi Biotec. PD 173074 and PD 161570 were purchased from Cayman Chemical.

Primary human ASMCs were obtained from transplant-grade lungs procured by the International Institute for the Advancement of Medicine. Protocols were approved by the McGill University Health Center Ethics Review Board, and all methods were performed in accordance with the relevant guidelines and regulations. Tissues were treated with elastase and collagenase for 20 min at 37°C with gentle agitation to digest the tissues. Cells were plated in six-well plates in DMEM supplemented with 10% FBS and 1% PSA in the density of 5 × 10⁴ cells per well.

Following informed consent, human peripheral blood was collected from healthy volunteers in heparin-coated tubes and was diluted with PBS at a 1:2 ratio (blood/PBS). Diluted blood was deposited on Ficoll-Paque PLUS (GE Healthcare) and centrifuged at 1400 rpm for 35 min. The PBMC layer was isolated from the centrifuged blood. PBMCs were cultured in RPMI 1640 supplemented with 10% FBS and 1% PSA, 20 ng/ml PMA, and 250 nM ionomycin for 48 h. CD4 T cells were isolated from PBMCs by MACS kit. Jurkat cells (TIB-152; ATCC) were activated and cocultured in conditions similar to CD4 T cells.

ASMCs were seeded 24 h before coculture. After isolation with an average purity of 80%, CD4 T cells were cocultured either unstained or stained with calcein AM 7 nM in PBS incubated in CO₂ chamber at 37°C for 30 min. After three washes with serum-free RPMI, stained CD4 T cells were cocultured with ASMCs at 5 × 10⁵ cells per well for 24 h in a mixture of RPMI 1640 and DMEM (1:1) supplemented with either 10 or 0.1% FBS and 1% PSA. Where appropriate, cells were treated with the FGFR inhibitors, 10 and 30 nM PD 173074, 100 and 200 nM PD 161570, 1 μg/ml anti-FGF2b neutralizing Ab, or 50 ng/ml human recombinant FGF2b, at the time of coculture in the presence of starvation medium (0.1% FBS).

FGF2b, brain-derived neurotrophic factor (BDNF), nerve growth factor (NGF), neurotrophin (NT) 3, insulin-like growth factor 1 (IGF-1), and NT-4/5 were measured by real-time quantitative PCR (qPCR) in human ASMCs and T cells. Total RNA was extracted, using RNeasy mini kit (Qiagen) according to manufacturer’s instructions. Reverse transcription was performed with an AffinityScript qPCR cDNA synthesis kit and using oligo dT primers (Agilent Technologies). qPCR was performed using iTaq SYBR green supermix (Bio-Rad). cDNA was amplified in the StepOnePlus real-time PCR system (Applied Biosystems). Relative mRNA expression was calculated using the ΔΔ cycle threshold method.

ASMCs were treated with FGF2b and control small interfering RNA (siRNA), as well as siRNA transfection reagent according to the manufacturer’s protocol (Santa Cruz). The highest efficiency of FGF2b knockdown was measured based on FGF2b mRNA expression by qPCR 72 h after siRNA transfection (data not shown). After 48 h of transfection, ASMCs were collected and seeded in six-well tissue culture plates and cocultured with calcein AM–stained CD4 T cells the following day.

ASMCs were cultured on coverslips and cocultured with calcein AM prestained CD4 T cells. Cocultured cells were either fixed in 4% paraformaldehyde or fixed and permeabilized with Triton X-100, 0.3% in PBS, and stained with phalloidin, rabbit anti-FGFR1 primary Ab, and rabbit anti–α/β-tubulin primary Ab. Anti-rabbit IgG-Alexa Fluor 647 was used as a secondary Ab. Coverslips were placed on slides using Fluoroshield mounting medium with DAPI. Cells were visualized using either a fluorescence microscope (Olympus) with 40× immersion oil objective or confocal microscope (LSM 700; Zeiss) with a 63× immersion oil objective. LMCs were quantified in a blinded manner with ×40 magnification. The proportion of CD4 T cells bearing LMCs/total CD4 T cells was determined.

Total PBMCs were incubated with PMA and ionomycin for 48 h. Cells were collected and incubated with or without recombinant human FGF2b for 15, 20, and 30 min. After each time point, cells were fixed and permeabilized with fixation/permeabilization buffer (BD Biosciences) and stained with anti-CD4 Ab and phalloidin. Samples were acquired by FACSCanto (BD) and analyzed by FlowJo software.

Cocultured cells were fixed with 4% paraformaldehyde and were postfixed in 1% osmium tetroxide and 1.5% potassium ferrocyanide. Samples were dehydrated in a graded series of ethanol and critical point dried with 100% ethanol. Samples were coated by platinum using Leica EM ACE600 Sputter Coater, and imaging was performed with a FEI Inspect F50 FE-Scanning electron microscope at an accelerating voltage of 5 kV.

Cocultured cells were fixed with 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer and were postfixed in 1% osmium tetroxide and 1.5% potassium ferrocyanide. Samples were stained with tannic acid, 1% in 0.1 M sodium cacodylate, and dehydrated in a graded series of concentrations of ethanol. Samples were infiltrated with Epon 812 and ethanol mixture, 1:1 and 3:1, followed by 100% Epon 812. Samples were embedded in low-viscosity, thermally curing Epon resin. Ultrathin sections (70–100 nm) were cut from the resin blocks by a Reichert-Jung Ultracut E ultramicrotome with a Diatome (Biel, Switzerland) diamond knife. The sections were transferred onto 200-mesh Cu transmission electron microscopy (TEM) grids with Formvar support film. Imaging was carried out on an FEI Tecnai 12 TEM equipped with an AMT XR80C CCD camera at an accelerating voltage of 120 kV.

ASMCs were stained with either Cell light Mitochondria-RFP, BacMan 2.0, or MitoTracker Red CM-H2Xros following manufacturer’s instruction before coculture with CD4 T cells and were cocultured with either unstained CD4 T cells or CD4 T cells stained with calcein AM for 18 h in RPMI 1640 and DMEM (1:1) supplemented with 10% FBS and 1% PSA. Cells were cocultured either directly or were separated by Transwell (CoStar). Transfer of mitochondria was either visualized and imaged after fixation of cells with paraformaldehyde (4%) by confocal microscopy under a 63× oil immersion objective or quantified by flow cytometer (either LSRII or FACSCanto; BD) after staining with the viability dye eFluor 780 and anti-human CD4 Ab. Data were analyzed by FlowJo software.

Statistical analysis was performed using GraphPad Prism 5 software. Data are presented as the mean ± 1 SEM, unless otherwise specified. Either Student t test or one-way ANOVA followed by Tukey’s posttest was used to compare either the means of two or more groups, respectively. The p values <0.05 were considered significant.

It has been shown that MNTs can be formed by either one or both communicating cells (33). We examined the distribution of length and diameter, as well as cytoskeletal composition of LMCs. Scanning electron microscopy was used to measure the range of diameters, 1.3–2 μm, and lengths, 2.9–60.5 μm, of LMCs. The length and width were related as a power function (Fig. 1A–D). Confocal microscopy and selective staining of CD4 T cells established that LMCs connecting CD4 T cells to ASMCs were of CD4 T cell origin (Fig. 1E).

Jurkat cells, an immortalized line of human CD4-expressing T lymphocytes, also generated LMCs with the same morphological features as primary CD4 T cells connecting to ASMCs (Fig. 1F).

To respond to stimuli that induce changes in the cell shape and morphology, it has been shown that cells use their surface membrane reservoir, in the form of membrane microvilli (34–36). Consistent with this mechanism of membrane recruitment, the scanning electron microscopy images indicated that CD4 T cells have a distinct cell membrane morphology and density of membrane microvilli when forming no LMC (Fig. 1A), medium-sized LMC (Fig. 1B), or long LMC (Fig. 1C). Longer LMCs were associated with a smoother cell membrane.

There is reported variability in the content of cytoskeletal components, F-actin and tubulin, among LMCs originating in different cell types (33). Confocal micrographs confirmed the presence of both F-actin and α/β-tubulin in all LMCs from CD4 T cells selectively stained with calcein AM (Fig. 2).

Next, we explored the mechanism(s) by which ASMCs induce LMC formation by CD4 T cells. Given the morphological similarity of the LMCs observed in this study to neurite outgrowths, we examined mRNA expression by ASMCs of a number of factors well-known to induce neurite outgrowth formation including FGF2b, NGF, IGF1, BDNF, NT3, and NT4/5 (28, 37–40). Although all of the earlier factors tested were expressed by ASMCs, only FGF2b expression was increased in ASMCs when cocultured with CD4 T cells (Fig. 3A).

To examine the contribution of FGF2b to LMC formation, we administered a rabbit polyclonal neutralizing Ab against FGF2b in coculture conditions. There was a significant reduction in the ratio of CD4 T cells with LMCs to total CD4 T cells, enumerated using fluorescence microscopy (Fig. 3B). Rabbit IgG isotype control did not affect the LMC formation by CD4 T cells compared with non–Ab-treated cells (Supplemental Fig. 1). Neutralizing FGF2b Ab did not affect the viability of CD4 T cells assessed by flow cytometry and eFluor780 (Supplemental Fig. 2B).

To determine whether FGF2b synthesized from ASMCs was the inducer of LMC formation, we knocked down FGF2b mRNA expression in ASMCs, using siRNA, and cocultured them with CD4 T cells, stained with calcein AM. The ratio of CD4 T cells with LMCs to total CD4 T cells was reduced by the suppression of FGF2b mRNA expression by ASMCs, and LMC formation was restored by the addition of exogenous FGF2b (Fig. 3C). Compared with control, control siRNA and FGF2b siRNA did not affect the total number of T cells (Supplemental Fig. 2A).

Based on our observation showing F-actin is contained within LMCs and the requirement of actin polymerization for LMC formation, we wished to confirm the effect of FGF2b on MNT formation by quantifying actin polymerization. Exogenous FGF2b increased mean fluorescence intensity (MFI) of F-actin in the CD4 T cell population at the time points 15, 20, and 30 min postexposure (Fig. 3D).

Having confirmed the effect of ASMC-derived FGF2b on LMC formation and based on the prior demonstration of the involvement of the FGF2b/FGFR1 axis in the formation of neurite outgrowths (28), we studied the role of FGFR1 in LMC formation. We performed immunofluorescent staining of cocultured cells with an anti-FGFR1 Ab and visualized the cells by confocal microscopy. Our observations confirmed the expression of FGFR1 protein by CD4 T cells, present in the cell body as well as along the length of the LMCs (Fig. 4A). In addition, we measured mRNA expression of FGFR1 by CD4 T cells, cocultured with ASMCs. Compared with CD4 T cells alone, coculture increased mRNA expression of FGFR1 in CD4 T cells (Fig. 4B).

To further investigate the role of FGFR1 in LMC formation, we used two chemical compounds known to selectively inhibit the tyrosine kinase activity of FGFR1, PD161570 and PD173074. Cocultured cells treated with either of the inhibitors had a decreased ratio of CD4 T cells with LMCs to total CD4 T cells, again demonstrating the involvement of FGFR1 in LMC formation (Fig. 4C, 4D). Compared with control, PD17307 and PD161570 did not affect the viability of CD4 T cells measured by flow cytometry stained by viability dye eFluor780 (Supplemental Fig. 2B).

Because LMCs formed between T cells and ASMCs contain microtubules that are potentially necessary for organelle movement, we used TEM on cocultured cells to visualize the organelles present in these structures. We observed mitochondria in LMC structures (Fig. 5A); hence we examined the possibility of mitochondrial transfer from ASMCs to CD4 T cells. We transduced ASMCs with baculoviral particles loaded with a fusion construct of the leader sequence of E1 α pyruvate dehydrogenase tagged with red fluorescent protein (Cell light Mitochondria-RFP, BacMan 2.0). Coculture of transduced ASMCs with calcein AM–stained CD4 T cells was performed, and cells were visualized by confocal microscopy (Fig. 5B). After visualization, transferred mitochondria were quantified by flow cytometry after staining of the cells with a viability dye and an anti-CD4 Ab (Fig. 5C). To confirm mitochondrial transfer and the necessity of cell contact for this process, we stained ASMCs with MitoTracker Red CM-H2Xros. Mitochondrial transfer to cocultured CD4 T cells, either physically separated by Transwell or in direct contact, was quantified by flow cytometry. There was no detectable transfer of mitochondria when CD4 T cells were physically separated but shared the same medium with ASMCs; however, mitochondrial transfer was detected in a substantial fraction of T cells (15.3 ± 2.2%) when cells were cocultured directly (Fig. 5D).

To investigate the dependence of mitochondrial transfer on LMCs and to confirm that the process was sensitive to FGF2b and FGFR1, we demonstrated decreased mitochondrial transfer from ASMCs stained with Mito Tracker to CD4 T cells in the presence of either anti-FGF2b neutralizing Ab or PD161570 (Fig. 5E). Transfer of mitochondria from CD4 T cells to ASMCs was not detected, using confocal microscopy (images not shown), suggesting a possible polarity to the mitochondrial transfer.

For the study of mitochondrial transfer, CD4 T cells and ASMCs were cocultured in medium supplemented with 10% FBS rather than starvation medium, which induces a significant increase of FGF2b mRNA expression relative to S9 by ASMCs compared with medium supplemented with 0.1% FBS, 2.23 ± 0.16 and 1.4 ± 0.1, respectively (n = 8; p < 0.001). The enriched medium also resulted in LMC formation by an increased proportion of CD4 T cells, 0.21 ± 0.03 and 0.44 ± 0.01, respectively (n = 5–12; p < 0.001).

To study the communication of CD4 T cells with asthmatic ASMCs in the context of a disease in which increased infiltration of CD4 T cells in the area of ASM is seen (9), we explored first the expression of FGF2b mRNA by asthmatic ASMCs compared with healthy ASMCs and found that it was increased (Fig. 6A). We quantified the ability of CD4 T cells to make LMCs in contact with asthmatic ASMCs by fluorescence microscopy and found an elevated ratio of CD4 T cells with LMCs to total CD4 T cells (Fig. 6B).

To examine the significance of increased LMC formation induced by asthmatic ASMCs, we assessed the frequency of viable CD4 T cells containing mitochondria from asthmatic ASMCs, in comparison with healthy ASMCs. ASMCs from asthmatic subjects transferred mitochondria to a greater fraction of CD4 T cells. Mitochondrial transfer from both healthy and asthmatic ASMCs was diminished by neutralizing FGF2b (Fig. 6C).

In this study, we have demonstrated an increase in FGF2b mRNA expression by ASMCs when in coculture with CD4 T cells. There was a concomitant increase in the expression of the FGFR1 in CD4 T cells. Our findings suggest a substantive role for ASMC-derived FGF2b and FGFR1 in the induction of LMC formation by CD4 T cells. The formation of LMCs was associated with topological changes in the CD4 T cell plasma membrane, and the LMCs were structurally equipped to transfer mitochondria. The movement of mitochondria occurred from ASMCs to CD4 T cells. ASMCs harvested from asthmatic subjects expressed a higher level of FGF2b mRNA, induced more CD4 T cells to form LMCs, and increased the fraction of CD4 T cells with transferred mitochondria.

CD4 T cells have been reported to connect to other structural and immune cells by LMCs (6, 7, 12). Due to the fact that the formation of LMCs is highly influenced by the microenvironment, and they do not share the same structure and function among different cell types (33), we further characterized these intercellular connections. The surface of lymphocytes is covered with short microvilli, dynamic structures that become flat during cell morphogenesis (34–36). We also noted on scanning electron microscopy images a loss of surface villi in association with the formation of long LMCs. This observation suggests that the formation of LMCs, ranging in length from 2.9 to 60.5 μm, and therefore in some instances extending to several times the diameter of the cell body, requires the recruitment of cell membrane from the plasma membrane microvilli.

The dimensions of LMCs, as well as their cytoskeletal components, predict the ability of LMCs to transfer materials, including ions, vesicles, and organelles (33). Hence, in our study, the diameter of LMCs and their complement of both F-actin and tubulin, confirmed by immunostaining and confocal microscopy, suggested that these tubular structures might also transfer organelles. Besides, we have previously reported that Mcl-1, a mitochondrial antiapoptotic protein, was present in LMCs and when expressed in ASMCs appeared in LMC-connected CD4 T cells (6). We speculate that this protein may have been transported in conjunction with mitochondria. Using two different tools to track ASMC mitochondrial movement, by expression of a mitochondrial-associated peptide tagged with RFP and MitoTracker Red, we found a population of CD4 T cells containing ASMC mitochondria. Physically separating ASMCs and CD4 T cells by Transwells (pore size, 400 nm) inhibited transfer of mitochondria (500 nm to 10 μm in diameter). However, whether other mechanisms of intercellular exchange such as trogocytosis and microvesicles might also be involved in mitochondrial transfer was not excluded by our study.

There appeared to be a unidirectional transfer of mitochondria from ASMCs to CD4 cells, although because T cells contain relatively few mitochondria compared with ASMCs our detection technique may have missed the transfer of a small number of mitochondria from T cells to ASMCs. The rich mitochondrial network in ASMCs is likely to have favored the direction of mitochondrial flow from these cells to T cells.

Although there is a dearth of information concerning the mechanisms of LMC formation, a few mechanisms have been proposed (5, 13, 23, 26, 27). The morphological similarity of T cell–derived LMCs to neurite outgrowths prompted us to examine ASMCs for factors known to induce neurite outgrowth formation. Of those tested, only the mRNA expression of FGF2b was increased in ASMCs in the presence of CD4 T cells. Confirmed by different approaches, FGF2b from ASMCs induced LMC formation by CD4 cells, possibly by increasing actin polymerization. Although there are four known receptors for FGF2b, FGFR1-4, in our search for the receptor on CD4 T cells mediating LMC formation we focused our exploration on FGFR1, given its role in neurite outgrowth formation in response to FGF2b (28). Consistent with this idea, pharmacological inhibition of the FGFR1 with two selective tyrosine kinase inhibitors, PD173074 and PD161570, suppressed LMC formation. Reduced mitochondrial transfer in the presence of either neutralizing FGF2b Ab or PD161750, which reduced LMC formation, supports the idea that LMCs are involved in organelle transfer. However, the inhibition of the FGF2b/FGFR1 pathway caused only a partial inhibition of both LMC formation and mitochondrial transfer, suggesting other important residual molecular mechanisms.

ASMCs harvested from asthmatic subjects have been reported to have different characteristics that distinguish them from cells harvested from healthy subjects including enhanced proliferation and increased mitochondrial biogenesis (41, 42). We explored the possibility that asthmatic ASMCs might therefore have differences in their propensity to form intercellular connections via LMCs with CD4 T cells. Consistent with the data on ASMCs harvested from nonasthmatic airway tissues, FGF2b also appeared to be important in the interaction between ASMCs and CD4 T cells via LMCs. Asthmatic ASMCs expressed more FGF2b mRNA and induced more CD4 T cells to make LMCs, leading to a greater frequency of CD4 T cells containing mitochondria of ASMC origin. The magnitude of the difference in mRNA expression for FGF2b between healthy and asthmatic ASMCs was greater than the difference in frequency of CD4 T cells bearing LMCs. There are several possible explanations. First, we did not confirm the level of FGF2b protein in the medium of the two sets of cells, and second, there may not be a linear relationship between FGF2b and the number of LMCs formed. Indeed, detection of FGF2b in medium was problematic (data not shown) in all probability because of the high isoelectric point of FGF2b and its avid binding to negatively charged tissue culture plates. The increased transfer of mitochondria from asthmatic ASMCs to T cells compared with nonasthmatic ASM may relate to the reported increased quantity of mitochondria in asthmatic ASMCs (41). However, we did not assess the mitochondrial mass in the cells we studied.

In conclusion, ASMC-derived FGF2b as well as CD4 T cell–derived FGFR1 regulate the communication between these two cell types by promoting LMC formation by T cells and mitochondrial transfer. The biological significance of LMC-mediated mitochondrial transfer to T cells awaits further elucidation.

We wish to acknowledge Dr. Hojatollah Vali for assistance with electron microscopic studies. We also acknowledge Leora Simon for valuable comments on the manuscript and Dr. Alice Panariti for data on mRNA expression of FGF2b in ASM response to FBS.

The authors have no financial conflicts of interest.