Drug-based mobilisation of mesenchymal stem/stromal cells improves cardiac function post myocardial infarction

ABSTRACT There is an unmet need for treatments that prevent the progressive cardiac dysfunction following myocardial infarction. Mesenchymal stem/stromal cells (MSCs) are under investigation for cardiac repair; however, culture expansion prior to transplantation is hindering their homing and reparative abilities. Pharmacological mobilisation could be an alternative to MSC transplantation. Here, we report that endogenous MSCs mobilise into the circulation at day 5 post myocardial infarction in male Lewis rats. This mobilisation can be significantly increased by using a combination of the FDA-approved drugs mirabegron (β3-adrenoceptor agonist) and AMD3100 (CXCR4 antagonist). Blinded cardiac magnetic resonance imaging analysis showed the treated group to have increased left ventricular ejection fraction and decreased end systolic volume at 5 weeks post myocardial infarction. The mobilised group had a significant decrease in plasma IL-6 and TNF-α levels, a decrease in interstitial fibrosis, and an increase in the border zone blood vessel density. Conditioned medium from blood-derived MSCs supported angiogenesis in vitro, as shown by tube formation and wound healing assays. Our data suggest a novel pharmacological strategy that enhances myocardial infarction-induced MSC mobilisation and improves cardiac function after myocardial infarction.

. Study design for assessment of the effect of mirabegron+AMD3100 treatment on cardiac function after MI.One rat from each group died within 24h of surgery, reducing the n number from 8 to 7 for each group.One rat from the vehicle group was excluded due to lack of significant infarct (<10%), reducing the vehicle group to n= 6. (Created with BioRender.com)Fig. S4.Gating strategy for the flow cytometry analysis of BM and blood derived culture expanded CFU-Fs.From left to right, top to bottom: first we gated into whole cells to exclude debris (FSC-A, SSC-A), then the single cells were gated into (FSC-H, FSC-W), followed by the live cells (Zombie Aqua™ negative).These were then gated into the CD45-negative fraction, to exclude haematopoietic cells, followed by exclusion of the HIS48+ and CD31+ CD43+ to remove any granulocytes, endothelial cells, and monocytes.Finally, we looked at the CD90 and CD29 expression to determine the rat MSC cell surface marker expression.(B) Example images from the wound healing assay are given for HUVECs treated with DMEM (negative control), EGM-2 (positive control), conditioned medium from blood-derived rMSCs from the MA group (BLOOD MA) and conditioned medium from blood-derived rMSCs from the Vehicle group (BLOOD V).The time after producing the scratch ('wound') is indicated as 0 h (immediately after) and 24 h (24 h after).

Disease Models & Mechanisms • Supplementary information
out in time.Thus, we decided to include additional animals in the MA group to maximise the probability of them being more evenly distributed throughout the study and to also account for potential loss of animals, as this was the main group of interest.The surgeon was blinded to the group assignment.To the MA group, mirabegron (YM178, ApexBio) was given at 3 mg/kg/day using the gavage method for five days post MI starting at 24h after surgery, and after 5 days AMD3100 (AMD3100 octahydrochloride hydrate, Sigma) was administered, one hour after the last dose of mirabegron, at 5 mg/kg using the intraperitoneal injection method (Fig. S2a).The A group received only the single dose of AMD3100 (5 mg/kg) (Fig. S2b), the M group received only the daily gavage of mirabegron (3 mg/kg) (Fig. S2c), and the SR group received daily gavage of the β3AR antagonist SR59230A (SR 59230A hydrochloride, Tocris) (5 mg/kg) for the first five days after MI (Fig. S2d).

Animal studies: experimental design and drug administration
For the study investigating the natural stem cell mobilisation post MI, Lewis rats in the MI group underwent left anterior descending artery (LAD) ligation surgery, and the animals in the sham group underwent a sham surgery, which did not include tying of the suture around the vessel.For each timepoint we assessed 4 animals in each group (MI and sham), and we had a total of four timepoints: day 1, 3, 5, and 10 (Fig. S1).The rats were culled at each of these endpoints for collection of peripheral blood (PB) by cardiac puncture and hind legs for bone marrow (BM) extraction. For

Disease Models & Mechanisms • Supplementary information
The final animal study looked at the effect of mirabegron+ AMD3100 (MA) on cardiac function at 5 weeks after MI surgery.The experimental design is summarised in Figure S3 below.The drug administration regimens for the MA and the vehicle groups were the same as in the drug selection study described above (Fig. 2a, e).
Briefly, MI inducing surgery was performed on Lewis rats, which were randomly assigned to two groups: vehicle and MA.The animal surgeon was blinded to group assignment at the time of surgery.Each group then received the corresponding treatment and at day 5 after the last drug administration they underwent the first cardiac MRI (acute stage), which was then followed with the second cardiac MRI 5 weeks later (chronic stage).The animals were culled at the 5-week timepoint following the second MRI scan and the PB, BM and hearts were collected for analysis.

Flow cytometry analysis
Flow cytometry analysis was done on cultured plastic adherent BM and blood CFU-F derived cells to assess their expression of the cell surface markers CD45, HIS48, CD31, CD43, CD90 and CD29, which were selected based on previous work that shows CD90+ CD29+ CD45-CD31-HIS48-CD43-to be a rat MSC marker panel [2][3][4][5] .The list of antibodies we used is included in Table S1 below.All steps were done on ice and flow buffer (2% FBS in PBS) was used for all staining except the live/dead stain Zombie Aqua™, which was done in PBS only.Cells were pelleted in a 96-well plate for the staining procedure.Unspecific binding was prevented using Fc block (1:200) prior to staining.After this, samples were stained with Zombie Aqua™ (1:200)   for 10 min to separate dead cells from the analysis.Finally, the fluorophore-conjugated antibodies were added to the cells and incubated in the dark for 30 min.The cells were washed and resuspended for analysis on the LSR Fortessa II.We used the following controls: compensation (single stains) and 'fluorescence minus one' (FMO) controls.
Compensation was done manually using the single stained cells.The gating strategy is summarised in Fig. S4 and the FMO controls are shown in Fig. S5.We used the FMO controls to set the negative gates for the analysis.
Sandwich ELISAs for CXCL12, TNF-alpha, IL-6, and IL-10 The protein concentrations of the chemokine CXCL12 and the cytokines IL-6, TNFalpha and IL-10 were analysed using the sandwich ELISA method.All antibodies were acquired from Peprotech.Briefly, the plates were coated with the capture antibodies overnight and were washed using washing solution (WS, 0.5% Tween in PBS) prior to blocking with blocking solution (BS,10% FBS in PBS) for an hour.The plates were then washed with WS, and the plasma or BM supernatant samples were added and incubated overnight.Following another wash step, the primary antibodies conjugated

Disease Models & Mechanisms • Supplementary information
to biotin were added for an hour.After this, plates were washed and the streptavidin-HRP was added for 30 min.Plates were washed again and the TMB substrate was added until full colour development, after which the reaction was stopped with sulphuric acid (1 mol/L).Results were read in a plate reader at 450 nm.

Histology
Following the sacrifice of the animal, the heart was excised, washed in heparinised PBS, and fixed in 4% Histofix overnight.Following fixation, the hearts were stored in 70% ethanol.The hearts were embedded in wax, 4 µm sections were obtained and Masson's trichrome staining was performed.Immunofluorescent staining was performed on heart sections according to the following protocol: sections were dewaxed in neoclear for 20 min, then rehydrated in increasing percentages of ethanol -100% (10 min), 95% (5 min), 70% (5 min), 30% (5 min), distilled water (5 min).
Following this, antigen retrieval was done by boiling the slides in citrate buffer (pH= 6) for 20 min.The slides were then permeabilised using 0.1% Triton/PBS for 10 min and following this were rinsed in tap water for 5 min.The sections were coated with blocking solution for 1h prior to addition of the primary antibodies.These were diluted in the same blocking solution (10% normal goat serum in 0.1% BSA in PBS).Antibodies were allowed to incubate overnight at 4 o C. For the isolectin-B4 staining we used isolectin-B4 conjugated to biotin.Three 10-minute washes in PBS were performed prior to adding the secondary antibodies (streptavidin-Alexa Fluor™ 555) were added and incubated overnight in the same way.Finally, the sections were stained with DAPI and mounted using ProLong™ Gold Antifade Reagent.Images were acquired using AxioObserver Z1 and the data was analysed using ImageJ, with a customised macro for blood vessel density quantification.For the scar size and interstitial fibrosis quantification, we used the Masson's trichrome images of the 5 weeks post MI heart sections, and a macro was written for both types of analysis.For the scar size quantification, the whole section was analysed, and for the IF, the percentage collagen was calculated against the whole tissue measured: LV and border zone regions were selected manually.For all histology, we performed blinded analysis using a total of 6 levels of the heart, spaced 250 µm.

In vitro angiogenesis assays
Human umbilical vein endothelial cells (HUVECs) were purchased from Lonza and were used between passage 1-6.For the tube formation assays, we used the Ibidi µ-Slide Angiogenesis slides as per manufacturer's instructions.The wound healing assay was performed as described previously [6] .Images were taken on a widefield microscope (AxioObserver Z1) and analysed using ImageJ.For the tube formation assay we used an Angiogenesis Analyzer macro developed for ImageJ [7] .
To obtain the conditioned media (CMs), the blood and BM derived culture expanded CFU-Fs (MSCs) were allowed to become confluent in a T25 flask and were then washed with serum-free DMEM and fed serum free DMEM for 24 hours.The CMs were then aspirated and filtered through a 0.

Fig. S2 .
Fig. S2.Schematic representation of the drug administration regimens following MI surgery for the study of pharmacological modulation of stem cell mobilisation.A) Mirabegron+AMD3100 (MA) drug administration.B) AMD3100 only (A) drug administration.C) Mirabegron only (M) drug administration.D) SR59230A (SR) drug administration.E) Vehicle (V) group drug administration.Created with Biorender.com

Fig.
Fig.S3.Study design for assessment of the effect of mirabegron+AMD3100 treatment on cardiac function after MI.One rat from each group died within 24h of surgery, reducing the n number from 8 to 7 for each group.One rat from the vehicle group was excluded due to lack of significant infarct (<10%), reducing the vehicle group to n= 6. (Created with BioRender.com)

Fig. S5 .
Fig. S5.Fluorescence minus one (FMO) controls for the flow cytometry analysis.Here we show the gates we used in our analysis, based on the FMO controls, which included all but one marker to ensure that the appropriate negative population is being gated out.From left to right, top to bottom we show the gates for Zombie Aqua (live/dead), Pacific Blue (CD45), FITC (HIS48), PE (CD31, CD43), APC (CD90), PE-Cy7 (CD29).

Fig. S7 .
Fig. S7.Representative histograms and scatter plots of the flow cytometry data.Here we have shown representations of the flow cytometry data for the BM and blood derived culture expanded CFU-Fs (MSCs) from both groups: vehicle and MA.The gating is shown in the Supplementary Methods.
Fig. S8.CXCL12 levels in conditioned medium (CM) from blood and BM derived MSCs.This graph shows the CXCL12 concentration (ng/mL) in the CMs from culture expanded blood and BM MSCs from the vehicle and MA groups.The data were obtained using the sandwich ELISA method.(Ordinary One-Way ANOVA, Tukey post-test, * p< 0.05, n= 3)

Fig. S11 .
Fig. S11.Example images of tube formation assay and wound healing assay.(A) Example images from the tube formation assay are given for each group: HUVECs treated with DMEM (negative control), EGM-2 (positive control), conditioned medium from blood-derived rat MSCs (rMSCs) from the MA group (BLOOD MA), conditioned medium from blood-derived rMSCs from the Vehicle group (BLOOD V), conditioned medium from bone marrow-derived rMSCs from the MA group (BM MA), conditioned medium from BM-derived rMSCs from the Vehicle group (BM V).
sagittal and coronal orientations followed by pseudo two-and four-chamber gradient echo scans.Left ventricular EF, mass, ESV and EDV were quantified from a stack of ECG and respiratory-gated CINE gradient echo images in the short-axis plane (11-12 slices).The acquisition parameters for CINE measurements were: repetition time (TR)=RR interval/number of frames (∼9 ms for 20 frames), TReffective=RR interval, echo time (TE)=2.2ms, flip angle=18°, slice thickness=1.5 mm (continuous slices), acquisition matrix=190×190, field of view = (38.5×38.5)mm2, leading to a spatial inplane resolution of (202×202) μm2, scan time: 18-20 min.Multi slice late gadolinium enhancement (LGE) data were acquired using an inversion recovery gradient echo sequence with a single TI point and flip angle of 90°.The TI was selected to effectively null the healthy myocardium and to provide the best contrast enhancement of the area of infarction.Infarction size (acute stage) and fibrosis (chronic stage) was quantified by tracking the LGE signal for each LV axial slice using the in-built function in the open-source software Segment (version 2.2, Medviso AB, Lund, Sweden).The following acquisition parameters were used: TE = 2.4 ms, TRs = 3.85 ms, FOV = (38 ×38) mm2, matrix size = 180 × 180, spatial resolution (211×211) µm2, 1.5 mm slice thickness, 9-10 slices, scan time: 3 min /slice.All MRI data underwent blinded analysis.Disease Models & Mechanisms: doi:10.1242/dmm.049630:Supplementary information Disease Models & Mechanisms • Supplementary information

Table S1 .
List of fluorophore-conjugated antibodies used in our analysis of culture expanded bone marrow and blood derived CFU-F.