− mice exhibit an M2 phenotypic shift in resident cardiac macrophages

    Implications for the Role of Macrophages in a Model of Myocardial Fibrosis: CCR2−/− Mice Exhibit An M2 Phenotypic Shift In Resident Cardiac Macrophages A. Falkenham, T. Myers, C. Wong, J.F. Legare PII: DOI: Reference:

S1054-8807(16)30034-5 doi: 10.1016/j.carpath.2016.05.006 CVP 6925

To appear in:

Cardiovascular Pathology

Received date: Revised date: Accepted date:

18 February 2016 25 May 2016 25 May 2016

Please cite this article as: Falkenham A, Myers T, Wong C, Legare JF, Implications for the Role of Macrophages in a Model of Myocardial Fibrosis: CCR2−/− Mice Exhibit An M2 Phenotypic Shift In Resident Cardiac Macrophages, Cardiovascular Pathology (2016), doi: 10.1016/j.carpath.2016.05.006

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT Implications for the Role of Macrophages in a Model of Myocardial Fibrosis:

T

CCR2-/- Mice Exhibit An M2 Phenotypic Shift In Resident Cardiac Macrophages

RI P

A Falkenham1, T Myers2, C Wong1, JF Legare1,2,3

1

SC

Department of Pathology, Dalhousie University, Halifax, NS, Canada

2

NU

Department of Surgery, Dalhousie University, Halifax, NS, Canada

3

MA

Department of Microbiology and Immunology, Dalhousie University, Halifax, NS, Canada

ED

Running head: M2 shift in CCR2-/- cardiac macrophages

Corresponding author:

PT

Alec Falkenham

Halifax, NS B3H 4R2

AC

5850 College St.

CE

10A Tupper Medical Building, Dalhousie

[email protected] Phone #: (902)494-6451 Fax #: (902)494-5125

1

ACCEPTED MANUSCRIPT Abstract BACKGROUND: Macrophages (MΦ) are functionally diverse and dynamic. Until recently, cardiac

T

MΦ were assumed to be monocyte-derived; however, resident cardiac MΦ (rCMΦ), present at

RI P

baseline, were identified in myocardia and have been implicated in cardiac healing. Previously, we demonstrated that CCR2-/- mice are protected from myocardial fibrosis – an observation

SC

initially attributed to changes in infiltrating monocytes. Here, we re-explored this observation in

NU

the context of our new understanding of rCMΦ. METHODS: Male CCR2-/- and C57BL/6 hearts were digested and purified to a single cell suspension, incubated with fluorophore-linked

MA

antibodies (CCR2, CX3CR1, CD11b, Ly6C, TNF-α, IL-10), and assessed by flow cytometry. Differentiated MΦ were co-cultured with fibroblasts in order to characterize how MΦ

ED

phenotype influences fibroblast activation. Fibroblasts were characterized for their expression of smooth muscle cell actin (SMA). RESULTS: A significant decrease in Ly6C expression was

PT

observed in the CCR2-/- cardiac MΦ population relative to WT, which corresponded with

CE

significantly lower TNF-α expression and significantly higher IL-10 expression. Using in-vitro coculture system classical MΦ promoted fibroblast activation relative to non-classical MΦ.

AC

CONCLUSION: CCR2-/- rCMΦ favour a more anti-inflammatory phenotype relative to WT controls. Moreover, a shift toward the anti-inflammatory promotes proliferation, but not activation in vitro. Together, these observations suggest anti-inflammatory cardiac MΦ populations may inhibit myocardial fibrosis in a pathological setting by preventing the activation of fibroblasts.

2

ACCEPTED MANUSCRIPT News and Noteworthy Here, we provide novel evidence for baseline differences in rCMΦ phenotypes (i.e. classical vs.

T

non-classical) and how these differences could modulate cardiac healing. Importantly, we

RI P

observed differences in how classical vs. non-classical MΦ influenced fibroblast activation,

SC

which could, in turn, affect fibrosis.

NU

Keywords

cardiac; resident macrophages; inflammation; fibrosis; flow cytometry; TNF-alpha; IL-10

MA

Introduction

MΦ are important mediators in tissue development, homeostasis, and disease

ED

progression1. As resident MΦ localize to different tissues during embryonic development, they acquire tissue/organ-specific functions2–5. These tissue/organ-specific MΦ are now known to

PT

exhibit distinct functions and phenotypes in tissues including the central nervous system, liver,

CE

spleen, lungs, peritoneum, skin, and myocardium. This is particularly important in tissues like myocardium, in which the ability to regenerate tissue after injury is assumed to be minimal 6.

AC

Moreover, circulation can contribute to tissue MΦ populations in function-specific manors, such that monocyte-derived MΦ can promote, resolve, or regulate inflammation78. Until recently, cardiac MΦ were thought to be exclusively derived from circulating monocytes during injurious stimuli9. This was supported by extensive research in different cardiac models, including Angiotensin II infusion and myocardial infarction, in which MΦ inundate the heart from circulating monocytes78,10. Using a GFP-linked CX3CR1, Pinto et al. demonstrated a CX3CR1+ MΦ population in the adult murine myocardium that persisted at baseline and was distinct from bone marrow derived monocytes 9. This observation has generated a resurgence in cardiac MΦ research that has focused on their phenotype, function,

3

ACCEPTED MANUSCRIPT turnover, and etiology1–4,7–28. As our understanding of MΦ evolves, researchers have been increasingly able to separate the MΦ into a broad phenotypic spectra13. Notably, MΦ could be

T

differentiated by their expression of the chemokine receptors CCR2 and CX3CR1 8,14. The

RI P

classical MΦ phenotype expressed high levels of CCR2 and low CX3CR1. Conversely, the nonclassical MΦ phenotype was either CCR2 low or negative and expressed high levels of CX3CR1.

SC

The non-classical population was associated with a mature phenotype, typically observed in

NU

both late-infiltrating phases and resident populations8,9.

rCMΦ have since been separated into at least 4 distinct populations that can be

MA

distinguished based on their expression of markers including CX3CR1, MHC class II, Ly6C, CD11b, CD64, CD11c, and CD2069,11,15,16. Molawi et al. have classified the populations into (1) CX3CR1-

ED

MHC II-, (2) CX3CR1-MHC II+, (3) CX3CR1+MHC II-, and (4) CX3CR1+MHC II+, which we have also characterized by CX3CR1, CCR2, CD11b, and Ly6C. RCMΦ originate during embryonic

PT

development as CX3CR1+MHC- cells. Notably, these embryonic-derived rCMΦ can confer

CE

protection against and favor resolution of injury by promoting cardiac regeneration and preventing remodeling1612. While still under debate, rCMΦ appear to be replenished primarily

AC

from circulating CCR2+ monocytes that express CX3CR1 and MHC class II16. Despite extensive work to characterize rCMΦ phenotypes, the functions of these different populations remain poorly characterized, particularly in models of disease. It is known that animals lacking CCR2 exhibit impaired ability to turnover rCMΦ from circulating monocytes16. Moreover, work in our laboratory using CCR2-/- mice suggested that these animals were partially protected from AngII-mediated inflammation and fibrosis17,18. When we first described the benefits of CCR2-/- deficiency, we did not appreciate or explore the role of this genetic defect on resident cardiac MΦ. Notably, early MΦ infiltrate did not significantly differ from WT mice and the underlying mechanism(s) for the benefits in CCR2-/-

4

ACCEPTED MANUSCRIPT mice remained unclear. Given the close relationship between MΦ activation and cardiac healing, comparing baseline resident cardiac MΦ populations between CCR2-/- and WT mice

T

warrants investigation.

RI P

In order to elucidate the role of CCR2 on rCMΦ phenotype, we compared myocardial tissue and rCMΦ from CCR2-/- mice relative to WT controls. When we discovered that CCR2-/-

SC

displayed a significant shift in rCMΦ Ly6C expression relative to WT, we investigated the

NU

functional significance of such a shift in terms of MΦ phenotype and how this would influence

MA

fibroblast activation and thus, fibrotic remodeling.

Methods

ED

Mice

Animal experiments were performed in accordance with the Canadian Council on

PT

Animal Care and approved by the Dalhousie University Committee on Laboratory Animals. Male

CE

C57BL/6 mice (8-10wk old; wild type: WT) were purchased from Jackson Laboratory (Bar Harbour, ME, USA) and male CCR2-/- mice (8-10wk old) on a C57BL/6 background were provided

AC

by one of the authors (T.I.) from a routinely genotyped colony. The absence of CCR2 was confirmed using conventional PCR, as previously demonstrated17. Mice were housed in the Carleton Animal Care Facility at Dalhousie University and provided food and water ad libitum. At the time of harvest, animals were weighed and hearts were extracted, flushed with saline, and weighed.

Immunohistochemistry Hearts were processed in 10% buffered formalin overnight and paraffin-embedded. Serial-sections (5µm) were stained for histological analysis. Basic myocardial histology and

5

ACCEPTED MANUSCRIPT cellular infiltrate were assessed using heart sections stained with hematoxylin and eosin (H&E). The area of the heart affected was calculated as previously described 17,19,20. Fibrotic deposition

T

was examined using heart sections stained with picrosirius red (SR) and the counter stain Fast

RI P

Green (FG). Collagen content was semi-quantified by photographing representative SR/FG whole heart-sections at 5x magnification. Using Adobe Photoshop CS6, red pixels were positively

SC

selected and summed for a total number of red (collagen) pixels. Subsequently, non-background

NU

pixels were summed for the total heart pixels. Collagen pixels were divided by the total heart pixels to provide a semi-quantitative measurement of the percent collagen content in the heart.

MA

All tissues were processed simultaneously for SR/FG and the same red colour palette was used to select red pixels.

ED

Immunofluorescence was performed on fixed co-cultures for KI-67 (eBioscience, San Diego, California, USA), collagen type-I (Rockland, Gilbertsville, PA), and α-smooth muscle cell

PT

actin (α-SMA; Sigma Aldrich)15. In brief, cells were permeabilized with 0.1% Triton-X-100/PBS,

CE

blocked with 5% BSA/PBS, and then incubated with primary antibodies overnight at 4°C. Sections were then incubated with AlexaFluor488, AlexaFluor555, or AlexaFluor647. Nuclei were

AC

stained with Hoechst. Slides were visualized using a Zeiss Axiovert 200M and photographed with a Hamamatsu Orca R2 Camera. Slides were visualized using a Zeiss Axiovert 200M and photographed with a Hamamatsu Orca R2 Camera.

Co-culture The MΦ-fibroblast co-culture was conducted as previously described with modification21,2215. Bone marrow-derived MΦ were generated as previously described23. In brief, animals were anaesthetized then sacrificed and the femurs and tibias were isolated. Cells were

6

ACCEPTED MANUSCRIPT flushed from the marrow using a 30G needle attached to a syringe containing DMEM (Dulbecco’s Modified Eagle Media) complete (DMEM-C: DMEM, L-glutamine, Pen Strep, 10%

T

FBS). Following washes and cell straining, bone marrow-derived MΦ were resuspended in

RI P

DMEM complete with 15% L929 conditioned medium and plated in T75 flasks. Media was replaced on day 3. On day 7, media was changed to (1) DMEM-C for control, (2) DMEM-C +

SC

100ng/mL LPS + 10ng/mL IFN-γ for M1 differentiation, or (2) DMEM-C + 10ng/mL IL-4 for M2

NU

differentiation. On day 9, MΦ populations were lifted using 0.25% trypsin, washed with DMEMC, counted, and re-plated in 12-well plates at a density of 2x105/well. Samples of MΦ

MA

populations were screened for F4/80, CD11b, Ly6C, CD206, TNF-α, and IL-10 expression using flow cytometry as described below. NIH/3T3 (ATCC, Manassas, VA) served as fibroblasts in

ED

mono- and co-cultures. NIH/3T3 were maintained in DMEM-C until lifted using 0.25% trypsin, washed with DMEM-C, counted, and re-plated at 2x105/well. Mono- and co-cultures were

PT

incubated for 72hrs, at which point, the supernatant was removed and the cells were washed

CE

with PBS then fixed in 4% PFA. The cells were immunofluorescently labeled as described above and wells were read for fluorescence using a Tecan infinite M200 Pro (Tecan, Männedorf,

AC

Germany) plate reader. α-SMA expression was standardized to DAPI fluorescence intensity for co-cultures and fibroblast monocultures. The co-culture expression was then calculated relative to fibroblast monoculture expression. Immunofluorescence and flow cytometry were used to quantify cell-purity.

Isolation of Cells From Myocardium Hearts from WT and KO mice were harvested under sterile conditions and used for cell isolation as previously described with modification17,24. In brief, hearts were mechanically and enzymatically digested in a collagenase solution (1 mg/mL collagenase II; Cedarlane

7

ACCEPTED MANUSCRIPT Laboratories, Ltd., Burlington, ON, Canada) in DMEM-C at 37°C, with agitation for 45 minutes. Cell isolates were twice washed in DMEM-C and purified over a Percoll gradient (30% and 70%).

T

Cells were then counted and used for flow cytometric analysis. Cells were washed with FACS

RI P

buffer (DPBS, 1% BSA, 0.1% NaN3) and incubated the antibodies αCD11b-APC (eBioscience, San Diego, California, US), αF4/80-PerCP-Cy5.5 (Biolegend, San Diego, California, US), αCX3CR1-

SC

PerCP-Cy5.5 (Biolegend), αCCR2-FITC (Biolegend), αTNF-PE (Biolegend), αIL-10-PE (Biolegend),

NU

and αLy6C-PE (Biolegend). Following incubation, cells were twice washed with FACS buffer and fixed with 1% formalin/FACS buffer. Isotype control (Santa Cruz Biotechnology, Santa Cruz, CA,

MA

USA) was used for negative gating. Data was acquired using a BD FacsCalibur and analyzed using FlowJo (FlowJo, LLC, Ashland, OR, USA). MΦ gating was performed by gating on cells using

ED

forward (FSC) and side scatter (SSC), which was then applied to a CD11b x F4/80 dot plot. The F4/80+ events were gated and applied to a CD11b x Ly6C dot plot for the characterization Ly6C

PT

mean fluorescent index (MFI) of mature and immature MΦ populations in the myocardium.

CE

When F4/80 labeling was not performed, CD11b x SSC was used to distinguish

Statistics

AC

polymorphonuclear cells (SSChigh) and monocyte-derived cells (SSClow-mid).

Results are expressed as mean ± SEM. Multiple group comparisons were performed by one-way analysis of variance (ANOVA) followed by Bonferroni post-test for comparing means. One-sample t-tests were used for relative comparisons between experimental groups and standardized controls.

8

ACCEPTED MANUSCRIPT Results Structural differences in myocardium are not present at baseline between WT and CCR2-/-

T

animals

RI P

We first investigated whether gross histological differences existed between CCR2-/- and WT mice at baseline. In order to evaluate baseline histology, 8-10wk (age-matched) CCR2-/- and

SC

WT mice were sacrificed and their hearts were characterized using H&E and SR/FG. Consistent

NU

with previous work from our laboratory, WT hearts did not contain observable areas of mononuclear cell accumulation (Fig. 1A). Similarly, CCR2-/- hearts were also void of mononuclear

MA

cell accumulation (Fig. 1B). Predictably, WT and CCR2-/- hearts did not contain the excess interstitial and perivascular collagen that is normally associated with disease models (Fig. 1C and

ED

D, respectively).

Quantification of cellular infiltration using representative histology sections confirmed

PT

that there were no significant differences in mononuclear cell accumulation between CCR2-/-

CE

and WT animals (Fig. 2A; 0.3% ± 0.3 relative to 0.4% ± 0.4) or collagen deposition (Fig. 2B; 3.7% ± 1.0 relative to 2.1% ± 1.3) between CCR2-/- and WT hearts. As demonstrated by others rCMΦ are

AC

very difficult to identify by routine histology and are best characterized by cell isolation and purification. Using a standardized and well-established technique we proceeded to isolate and purify mononuclear cell population from the heart of WT and CCR2-/- animals. This quantification of mononuclear cells isolated from WT and CCR2-/- hearts was performed prior to flow cytometric characterization. Consistent with visual analysis, the number of purified mononuclear cells isolated between WT and CCR2-/- mice did not significantly differ (Fig. 2C) – [(3.3 ± 0.6 vs. 2.8 ± 0.6) x 105, respectively]. Structural comparison of the hearts by heart weight:body weight ratio further confirmed that WT and CCR2-/- hearts do not significantly differ in hypertrophy score (Fig. 3D).

9

ACCEPTED MANUSCRIPT Together, these results suggest that there are no observable gross cellular or histological differences between WT and CCR2-/- hearts at baseline. As such, we next explored whether

T

inherent differences were present in the baseline cardiac MΦ populations between WT and

RI P

CCR2-/- mice. Specifically, we were interested in identifying if differences in rCMΦ existed

SC

between WT and CCR2-/- animals.

NU

CCR2-/- rCMΦ exhibit an M2 shift in phenotype

In order to better examine potential cardiac differences at the cellular level, rCMΦ were

MA

isolated from the CCR2-/- and WT myocardia and processed for flow cytometry. In the cell isolate, rCMΦ were characterized for their expression of MΦ markers (i.e. CD11b, CCR2, and

ED

CX3CR1) and activation markers (i.e. Ly6C, TNF-α, and IL-10). Total rCMΦ were selected based on their expression of CX3CR1 as previously described, which did not significantly differ between

PT

WT and CCR2-/- (Fig. 3A and C, respectively). Within the CX3CR1 gated populations, however,

CE

there is a prominent CD11bhigh Ly6Chigh population in the WT myocardium (Fig. 3B) that is absent in the CCR2-/- myocardium (Fig. 3D). Moreover, when gated on, the CD11bhigh Ly6Chigh population

AC

also expresses CCR2 (Fig. 3E; representative histogram of n=4) and was the only population to significantly differ between WT and CCR2-/- rCMΦ (Fig. 3F). Thus, WT and CCR2-/- myocardia significantly differ in their baseline populations of MΦ, as characterized by their expression of CX3CR1, CD11b, and Ly6C. Since Pinto’s definitive characterization of a rCMΦ population in the heart, a growing body of literature supports that rCMΦ are not a uniform population, but rather phenotypically and functionally distinct populations. rCMΦ populations are, however, linked by their expression of CX3CR111, which provided an initial gate for excluding non-rCMΦ. We then assessed the frequency of CX3CR1+ populations based on their varied expression of CD11b and

10

ACCEPTED MANUSCRIPT Ly6, which are markers of maturity and may reflect functional divergence. Largely, the MΦ populations did not significantly differ between WT and CCR2-/- mice (Table 1). Consistent with

T

the appearance of the CD11b x Ly6C dotplot, the only group that exhibited a significant change

RI P

was the CD11b+ Ly6C++ gate (Fig. 3F). CD11b++ Ly6C++ cells represented 2.2 ± 0.3% and 0.8 ± 0.2% of myocardial CX3CR1+ cells in WT and CCR2-/- mice, respectively – a 2.75 fold increase in WT

SC

over CCR2-/-. Thus, there are distinct differences between WT and CCR2-/- rCMΦ populations at

NU

baseline. Quantification of rCMΦ phenotypes

Ly6C+

WT

22.9 ± 1.4

14.3 ± 1.2

CCR2-/-

23.0 ± 0.8

14.0 ± 1.3

CD11b++

Ly6C++

Ly6C-

Ly6C+

Ly6C++

4.0 ± 0.7

17.2 ± 2.6

10.2 ± 1.0

2.2 ± 0.3***

13.3 ± 1.4

10.2 ± 1.1

0.8 ± 0.2

ED

Ly6C-

MA

CD11b+

PT

5.2 ± 1.0

CE

***p<0.001, n=6-8 (data represented as percentages of CX3CR1+ events)

Table 1 – Quantification of rCMΦ phenotypes. Populations were separated based on their

AC

expression of CD11b – an indicator of phenotypic immaturity. Of the CD11b high rCMΦ, CCR2-/- exhibited a significantly lesser CD11b high Ly6C high population relative to their WT counterpart (n=4-6; ***p<0.001).

While Ly6C is an indicator of MΦ phenotype, it does not necessarily correlate to function. As such, we next analyzed the expression of 2 functional markers of opposing MΦ polarities: inflammatory TNF-α and anti-inflammatory IL-10. The classical MΦ phenotype, which expresses higher levels of Ly6C, is typically associated with a pro-inflammatory functional state.

11

ACCEPTED MANUSCRIPT In contrast, the non-classical MΦ phenotype expresses lower levels of Ly6C and is instead associated with an anti-inflammatory, pro-fibrotic, and/or pro-angiogenic functional state(s).

T

Intracellular flow cytometry was used to characterize the expression of the cytokines.

RI P

rCMΦ populations were again defined as CX3CR1+ cells isolated from the myocardium. We then separated the populations based on their expression of CD11b into CD11blow and CD11bhigh

SC

gates. The representative contour plots for WT (Fig. 4A) and CCR2-/- (Fig. 4B), reveal a leftward

NU

shift toward lower TNF-α expression. Reflective of baseline differences, CD11blow rCMΦ – often considered the predominant resident population – did significantly differ in its expression of

MA

TNF-α (Fig. 4C; *p<0.05). WT CD11blow rCMΦ displayed higher TNF-α relative to the CD11blow population in CCR2-/-, as measured by MFI. Interestingly, the CD11bhigh population did not differ

ED

in TNF-α expression between WT and CCR2-/-, despite the WT CD11bhigh group containing the Ly6Chigh population described above (Fig. 5D). Similarly, an inverse trend – increased IL-10

PT

expression – is observable in the representative contour plots for WT (Fig. 5A) and CCR2-/- (Fig.

CE

5B) rCMΦ. In contrast to TNF- α expression, CCR2-/- rCMΦ in both (Fig. 5C) CD11blow and (Fig. 5D) CD11bhigh populations trended toward increased IL-10 expression, which reached

AC

significance in the former population (*p<0.05). While significant findings were limited to the CD11blow population for both TNF-α and IL10, the data suggests that CCR2-/- rCMΦ favor a lesser pro-inflammatory and stronger antiinflammatory phenotype than their WT counterpart. This novel observation now offers some insight into potential mechanisms by which rCMΦ can influence the myocardial inflammatory response and in turn fibrosis.

Classical MΦ increase fibroblast activation in vitro

12

ACCEPTED MANUSCRIPT In previous work, we demonstrated a difference in the AngII-mediated myocardial fibrotic responses between CCR2-/- and WT mice, with the former exhibiting reduced fibrosis. In

T

addition, we had also shown that early mononuclear infiltration does not significantly differ in

RI P

the AngII response, suggesting potential baseline differences could account for the differences in fibrosis. In this manuscript, we have supported that theory by showing significant differences

SC

in CCR2-/- and WT rCMΦ populations at baseline; CCR2-/- and WT rCMΦ exhibit an inverse trend

NU

in their expression of TNF-α and IL-10 that may favour a more protective phenotype in CCR2-/mice. To support our findings, we used an in vitro co-culture system designed to evaluate

MA

fibroblast activation in the presence of different MΦ phenotypes. Bone marrow-derived MΦ were grown in culture to represent 3 different phenotypes as

ED

previously described– (1) non-induced MΦ, (2) classical MΦ using IFN-γ and LPS, and (3) nonclassical MΦ using IL-4 – and all groups were then co-cultured with 3T3 fibroblasts for 24hrs25.

PT

Induced MΦ populations were characterized for their expression of CD11b, CD206, Ly6C, F4/80,

CE

IL-10 and TNF-α (Fig. 1S). Generally, classically induced MΦ exhibited dichotomous CD11b expression, increased Ly6C and TNF-α expression and reduced CD206 and IL-10 expression. In

AC

contrast, non-classically induced MΦ exhibited uniform CD11b expression, reduced Ly6C and TNF-α expression and increased CD206 and IL-10 expression. F4/80 expression remained consistent across groups. Mono- and co-cultured fibroblasts were labeled for collagen-I as indicators of fibroblast phenotype and (Fig. 6A-C) αSMA or (Fig. 6D-F) KI-67 to demonstrate their activation and proliferation, respectively. Representative micrographs demonstrate the co-localization (yellow) of collagen-I (red) and αSMA (green) and the co-localization (pink) of DAPI (blue) and KI-67 (red) on collagen-I+ (green) cells. While collagen-I expression did not significantly differ between any

13

ACCEPTED MANUSCRIPT groups (Fig. 6G), fibroblasts co-cultured with M1 MΦ exhibited significantly greaterαSMA expression, as measured by fluorescence emission (Fig. 6H, *p<0.05).

T

In contrast to the αSMA findings, M2 MΦ were the only subset to significantly increase

RI P

fibroblast proliferation relative to fibroblast monocultures (Fig. 6I; ***p<0.001). In monocultures, 17.9 ± 1.9% of fibroblasts, characterized by co-localization of collagen-I and KI-67,

SC

were undergoing proliferation. MΦ+ Fibroblasts and M1 + Fibroblasts exhibited similar levels of

NU

proliferation at 28.95 ± 2.9% and 29.2 ± 1.9% of KI67+ fibroblasts, respectively. In contrast, M2+Fibroblasts exhibited 40.2 ± 4.4% of fibroblasts undergoing proliferation – a 2.25-fold

MA

increase over the fibroblast monocultures. Thus, while M1 MΦ are able to promote fibroblast

ED

activation, M2 MΦ are able to promote fibroblast proliferation.

PT

Results Summary

The findings of this manuscript are summarized below in Table 2. In brief, we

CE

observed an increase in Ly6C expression in WT relative to CCR2-/- mice. This increase is

AC

likely attributable to the presence of a CD11b++ Ly6C++ population that was absent in CCR2-/-. In addition, the increased Ly6C expression in WT relative to CCR2-/- mice was associated with an increase in rCMΦ TNF-. Overall, this is representative of a shift in rCMΦ toward the M1 phenotype relative to CCR2-/- and the M1 phenotype was shown to favour fibroblast activation in vitro. Conversely, CCR2-/- mice displayed a reduction in rCMΦ Ly6C expression relative to WT mice. This decrease was associated with increased IL-10 and a shift toward the M2 phenotype. Lastly, the M2 phenotype was shown to promote fibroblast proliferation in vitro.

14

ACCEPTED MANUSCRIPT

Summary of findings

IL-10

rCMΦ shift

How rCMΦ shift affects fibroblasts

T

TNF-a

RI P

Ly6C

SC

WT M1 Activation ⬆ ⬆ ⬇ CCR2-/M2 Proliferation ⬇ ⬇ ⬆ Arrows are relative to other genotype (i.e. WT relative to CCR2-/- and CCR2-/- relative to WT)

NU

Table 2 – This table represents a summary of our findings in this manuscript.

MA

Discussion

A growing body of literature has expanded our understanding of MΦ population

ED

composition, diversity, and activation1–4,12,15,16. We now know and must consider how the

PT

complex activation, recruitment, and etiology of MΦ in the heart influences tissue development, homeostasis, and disease8,11,15,16. Recently, our laboratory demonstrated that rCMΦ do not

CE

necessarily maintain a constant phenotype and that their phenotype can be associated with

AC

differences in the expression of CX3CR115. Moreover, we also previously demonstrated that CCR2-/- mice are in part protected from AngII-mediated cardiac fibrosis in a process that is not dependent on early infiltrating MΦ17. In the current manuscript, we built upon that observation by investigating rCMΦ phenotypes in the context of CCR2-/- mice. There is no debate that MΦ can influence fibroblast function; however, the association is likely only partially responsible for the observed differences in cardiac fibrosis between CCR2-/-and WT mice22,26. In the current manuscript, we assessed the baseline nonmyocyte cell count, the myocardial collagen content, and the hypertrophy to determine whether there were gross myocardial differences prior to physiological stress or disease such as initiating AngII infusion. The lack of gross differences at baseline suggests that any inherent

15

ACCEPTED MANUSCRIPT differences in rCMΦ between CCR2-/- and WT mice do not significantly affect normal cardiac function and development. Important in the context of this manuscript is whether potential

T

underlying cellular differences between CCR2-/- and WT mice, not obvious at baseline, could

RI P

affect cardiac responses to stress and as such healing.anti-inflammatory/pro-healing CX3CR1+ rCMΦ are initially derived from embryonic and fetal MΦ populations, but are progressively

SC

replaced with age from circulating monocytes in a process largely dependent on CCL2-CCR21–4,7– . In addition, baseline CCR2+ rCMΦ exhibit upregulated genes involved with inflammasome

NU

28

activation and become important contributors of pro-inflammatory IL-1 in the AngII model of

MA

myocardial fibrosis. As such, the contribution of the CCR2+ rCMΦ could change how the heart responds to stressors11. In support of this process, we identified a CCR2+CD11bhighLy6Chigh MΦ

ED

population in WT hearts that was absent in CCR2-/- hearts. This is particularly interesting in the context of recent observations in CX3CR1-/- mice: at baseline, CX3CR1-/- mice show a greater

PT

contribution of CCR2+CD11bhighLy6Chigh to their rCMΦ relative to WT controls. Moreover,

CE

CX3CR1-/- mice exhibited exacerbated inflammation and fibrosis during AngII infusion. The process by which this occurs is summarized in a diagram representing the theoretical

AC

contribution of M1 and M2 MΦ to the total rCMΦ population (Fig. 7; WT values adapted from Molawi et al.)16. Together, this supports the premise that at the onset of stressors, CCR2-/- mice may be partially protected from inflammation while CX3CR1-/- mice may be predisposed to inflammation. In turn, the corresponding chemokines, CCL2 and CX3CL1, likely have competing roles in preserving and replacing the anti-inflammatory rCMΦ population, respectively. In order to provide a mechanistic link between MΦ phenotype/activation and myocardial fibrosis, we studied a co-culture system in which differently activated bone marrowderived MΦ were incubated with fibroblasts in vitro. While we acknowledge that isolated rCMΦ would have provided a better representation of the in vivo interactions with fibroblasts, rCMΦ

16

ACCEPTED MANUSCRIPT cannot realistically be isolated in sufficient quantities to perform such an experiment. As such, bone marrow-derived MΦ were chosen as an appropriate analog. Consistent with previous

T

findings, MΦ were able to stimulate fibroblast activation into myofibroblasts15,26. A growing

RI P

body of research suggests that inflammation can promote fibroblast activation through processes that involve TNF-, fibroblast activation protein (FAP), and IL-6 signaling22,27.

SC

Moreover, when MΦ are co-cultured with fibroblasts, their interactions promote feedback loops

NU

that favour a pro-fibrotic response. Notably, MΦ production of TNF- directly promotes fibroblast proliferation, as well as other pro-inflammatory cytokine production. In addition, pro-

MA

fibrotic fibroblasts also produce TNF-. Thus, even in the co-culture system these interactions are not uni-directional, but rather a cohesive response to promote fibrosis during inflammation.

ED

This may serve a protective role for replacing tissue lost during early inflammatory processes. At the same time, this reparative process can become corrupted in chronic inflammation, leading

PT

to normal tissue being degraded and replaced with scar tissue28. In our co-culture, we observed

CE

pro-inflammatory M1 MΦ increasing fibroblast activation, as demonstrated by increased αSMA expression. Interestingly, M2 MΦ – not M1 MΦ – promoted fibroblast proliferation, suggesting

AC

that the biphasic MΦ response observed in cardiac healing is much more complex than a single cell type. Moreover, it suggests that the healing/reparative response traditionally associated with the M2 phenotype may (i) begin with the M1 phenotype and (ii) have fibroblasts bridging the gap in the biphasic response. Thus, inflammation may serve as a precursor to fibrosis through fibroblast activation, which is temporally followed by fibroblast proliferation during the reparative phase. Together, our findings support the importance of rCMΦ in fibroblast function and in turn, cardiac healing. Specifically, we have provided additional support for how changes in rCMΦ phenotype and/or activation can influence cardiac healing through the susceptibility to

17

ACCEPTED MANUSCRIPT inflammation and the subsequent fibrotic response. Future studies could examine how systemic changes in inflammation, such as that which occurs with obesity or metabolic syndrome, alter

T

rCMΦ phenotype and/or activation and in turn, also susceptibility to cardiac injury. We have

RI P

also provided a novel mechanism by which CCR2-/- mice may be protected from stressors such as AngII-mediated myocardial fibrosis. While previous data still supports the role of reduced

SC

proliferation in the protecting CCR2-/- mice from myocardial fibrosis, our novel findings support

NU

the link between MΦ and fibroblasts in this process. Notably, MΦ phenotype may be linked to a biphasic fibroblast response (i.e. activation followed by proliferation). Lastly, in concert with our

MA

previous findings, we have raised the question of whether CCR2 and CX3CR1 may have competing roles in the replacement, preservation, and/or survival of rCMΦ. This manuscript

ED

paves the way for future studies to explore the seemingly opposing roles of the chemokine receptors CCR2 and CX3CR1 in rCMΦ homeostasis and how this balance impacts the

CE

PT

development of myocardial fibrosis.

Limitations

AC

In this study, we were limited to examining rCMΦ at a single age (~8wks), which, in light of recent evidence, can change the proportion of different rCMΦ subsets. Moreover, our characterizations were performed at baseline and further work would be required to determine whether the significance of our findings can be extrapolated to cardiac disease models.

Acknowledgments We’d like to thank Pat Colp and Mari Vaci for their technical skills and expertise that contributed to this manuscript. We’d also like to thank Dr. Thomas Issekutz for providing the CCR2-/- mice for our studies.

18

ACCEPTED MANUSCRIPT

Grants

T

Canadian Institute of Health Research 45980

RI P

REACH 36243 Bridge Fund, Dalhousie 48552

NU

SC

Dept. of Surgery, Dalhousie 49488

Disclosures

MA

The authors have declared that no competing interests exist

Wynn, T. a, Chawla, A. & Pollard, J. W. Macrophage biology in development, homeostasis

PT

1.

ED

References

and disease. Nature 496, 445–55 (2013). Perdiguero, E. G., Klapproth, K., Schulz, C., Busch, K., Azzoni, E., Crozet, L., Garner, H.,

CE

2.

AC

Trouillet, C., de Bruijn, M. F., Geissmann, F. & Rodewald, H.-R. Tissue-resident macrophages originate from yolk-sac-derived erythro-myeloid progenitors. Nature 518, 547–551 (2014). 3.

Gordon, S., Plüddemann, A. & Martinez Estrada, F. Macrophage heterogeneity in tissues: phenotypic diversity and functions. Immunol. Rev. 262, 36–55 (2014).

4.

Epelman, S., Lavine, K. J. & Randolph, G. J. Origin and Functions of Tissue Macrophages. Immunity 41, 21–35 (2014).

5.

Gautier, E. L. & Yvan-Charvet, L. Understanding macrophage diversity at the ontogenic and transcriptomic levels. Immunol. Rev. 262, 85–95 (2014).

19

ACCEPTED MANUSCRIPT 6.

Choi, W.-Y. & Poss, K. D. Cardiac regeneration. Curr. Top. Dev. Biol. 100, 319–44 (2012).

7.

Sunderkötter, C., Nikolic, T., Dillon, M. J., Van Rooijen, N., Stehling, M., Drevets, D. A.,

T

Leenen, P. J. M. & Sunderkotter, C. Subpopulations of mouse blood monocytes differ in

Nahrendorf, M., Swirski, F. K., Aikawa, E., Stangenberg, L., Wurdinger, T., Figueiredo, J.-L.

SC

8.

RI P

maturation stage and inflammatory response. J. Immunol. 172, 4410–7 (2004).

L., Libby, P., Weissleder, R. & Pittet, M. J. The healing myocardium sequentially mobilizes

NU

two monocyte subsets with divergent and complementary functions. J. Exp. Med. 204,

9.

MA

3037–47 (2007).

Pinto, A. R., Paolicelli, R., Salimova, E., Gospocic, J., Slonimsky, E., Bilbao-Cortes, D.,

ED

Godwin, J. W. & Rosenthal, N. a. An abundant tissue macrophage population in the adult

e36814 (2012).

van der Laan, A. M., Ter Horst, E. N., Delewi, R., Begieneman, M. P. V, Krijnen, P. A. J.,

CE

10.

PT

murine heart with a distinct alternatively-activated macrophage profile. PLoS One 7,

Hirsch, A., Lavaei, M., Nahrendorf, M., Horrevoets, A. J., Niessen, H. W. M. & Piek, J. J.

AC

Monocyte subset accumulation in the human heart following acute myocardial infarction and the role of the spleen as monocyte reservoir. Eur. Heart J. (2013). doi:10.1093/eurheartj/eht331 11.

Epelman, S., Lavine, K. J., Beaudin, A. E., Sojka, D. K., Carrero, J. A., Calderon, B., Brija, T., Gautier, E. L., Ivanov, S., Satpathy, A. T., Schilling, J. D., Schwendener, R., Sergin, I., Razani, B., Forsberg, E. C., Yokoyama, W. M., Unanue, E. R., Colonna, M., Randolph, G. J. & Mann, D. L. Embryonic and adult-derived resident cardiac macrophages are maintained through distinct mechanisms at steady state and during inflammation. Immunity 40, 91– 104 (2014).

20

ACCEPTED MANUSCRIPT 12.

Lavine, K. J., Epelman, S., Uchida, K., Weber, K. J., Nichols, C. G., Schilling, J. D., Ornitz, D. M., Randolph, G. J. & Mann, D. L. Distinct macrophage lineages contribute to disparate

T

patterns of cardiac recovery and remodeling in the neonatal and adult heart. Proc. Natl.

Mosser, D. M. & Edwards, J. P. Exploring the full spectrum of macrophage activation. Nat.

SC

13.

Rev. Immunol. 8, 958–69 (2008).

Nahrendorf, M., Pittet, M. J. & Swirski, F. K. Monocytes: protagonists of infarct

NU

14.

RI P

Acad. Sci. U. S. A. 111, 16029–34 (2014).

15.

MA

inflammation and repair after myocardial infarction. Circulation 121, 2437–45 (2010). Falkenham, A., de Antueno, R., Rosin, N., Betsch, D., Lee, T. D. G., Duncan, R. & Légaré, J.-

ED

F. Nonclassical resident macrophages are important determinants in the development of myocardial fibrosis. Am. J. Pathol. 185, 927–42 (2015). Molawi, K., Wolf, Y., Kandalla, P. K., Favret, J., Hagemeyer, N., Frenzel, K., Pinto, A. R.,

PT

16.

CE

Klapproth, K., Henri, S., Malissen, B., Rodewald, H.-R., Rosenthal, N. A., Bajenoff, M., Prinz, M., Jung, S. & Sieweke, M. H. Progressive replacement of embryo-derived cardiac

17.

AC

macrophages with age. J. Exp. Med. 211, 2151–8 (2014). Falkenham, A., Sopel, M., Rosin, N., Lee, T. D. G., Issekutz, T. & Légaré, J.-F. Early fibroblast progenitor cell migration to the AngII-exposed myocardium is not CXCL12 or CCL2 dependent as previously thought. Am. J. Pathol. 183, 459–69 (2013). 18.

Xu, J., Lin, S.-C. C., Chen, J., Miao, Y., Taffet, G. E., Entman, M. L. & Wang, Y. CCR2 mediates the uptake of bone marrow-derived fibroblast precursors in angiotensin IIinduced cardiac fibrosis. Am. J. Physiol. Heart Circ. Physiol. 301, H538–47 (2011).

19.

Rosin, N. L., Sopel, M., Falkenham, A., Myers, T. L. & Legare, J. F. Myocardial migration by

21

ACCEPTED MANUSCRIPT fibroblast progenitor cells is blood pressure dependent in a model of angII myocardial fibrosis. Hypertens Res 35, 449–456 (2012). Rosin, N. L., Falkenham, A., Sopel, M. J., Lee, T. D. & Legare, J. F. Regulation and Role of

T

20.

RI P

Connective Tissue Growth Factor in AngII-Induced Myocardial Fibrosis. Am J Pathol

21.

SC

(2012). doi:10.1016/j.ajpath.2012.11.014

Ma, F., Li, Y., Jia, L., Han, Y., Cheng, J., Li, H., Qi, Y. & Du, J. Macrophage-stimulated

NU

cardiac fibroblast production of IL-6 is essential for TGF β/Smad activation and cardiac

22.

MA

fibrosis induced by angiotensin II. PLoS One 7, e35144 (2012). Holt, D. J., Chamberlain, L. M. & Grainger, D. W. Cell-cell signaling in co-cultures of

23.

ED

macrophages and fibroblasts. Biomaterials 31, 9382–94 (2010). Weischenfeldt, J. & Porse, B. Bone marrow-derived macrophages (BMM): isolation and

Sopel, M., Falkenham, A., Oxner, A., Ma, I., Lee, T. D. G., Legare, J. F. & Légaré, J.-F.

CE

24.

PT

applications. Cold Spring Harb. Protoc. (2008). doi:10.1101/pdb.prot5080

AC

Fibroblast progenitor cells are recruited into the myocardium prior to the development of myocardial fibrosis. Int J Exp Pathol 93, 115 (2012). 25.

Ying, W., Cheruku, P. S., Bazer, F. W., Safe, S. H. & Zhou, B. Investigation of macrophage polarization using bone marrow derived macrophages. J. Vis. Exp. e50323 (2013). doi:10.3791/50323

26.

Zeng, Q. & Chen, W. The functional behavior of a macrophage/fibroblast co-culture model derived from normal and diabetic mice with a marine gelatin-oxidized alginate hydrogel. Biomaterials 31, 5772–81 (2010).

27.

Brokopp, C. E., Schoenauer, R., Richards, P., Bauer, S., Lohmann, C., Emmert, M. Y.,

22

ACCEPTED MANUSCRIPT Weber, B., Winnik, S., Aikawa, E., Graves, K., Genoni, M., Vogt, P., Lüscher, T. F., Renner, C., Hoerstrup, S. P. & Matter, C. M. Fibroblast activation protein is induced by

T

inflammation and degrades type I collagen in thin-cap fibroatheromata. Eur. Heart J. 32,

Flavell, S. J., Hou, T. Z., Lax, S., Filer, A. D., Salmon, M. & Buckley, C. D. Fibroblasts as

SC

28.

RI P

2713–22 (2011).

novel therapeutic targets in chronic inflammation. Br. J. Pharmacol. 153 Suppl , S241–6

MA

NU

(2008).

ED

Fig. 1 – Baseline myocardia were assessed for gross histology. Representative H&E stained baseline myocardial sections from (A) WT and (B) CCR2-/- mice. Representative SR/FG stained baseline myocardial sections from (C) WT and (D) CCR2-/- mice. (n=6-8)

AC

CE

PT

Fig. 2 – Baseline myocardia were characterized for MΦ accumulation, collagen content, cell counts, and hypertrophy. (A) Myocardial MΦ accumulation was assessed using H&E sections from each group. (B) In support of the histology findings, hearts were mechanically and enzymatically digested and the isolated leukocytes were counted and compared between groups. (C) Collagen content, as quantified in SR/FG stained myocardial sections, did not significantly differ between WT and CCR2-/- mice. (D) Myocardia were weighed at the time of harvest to generate a hypertrophy score, which did not significantly differ between WT and CCR2-/- myocardia (n=6-8). Fig. 3 – Flow cytometric characterization of isolated myocardial leukocytes labeled with CX3CR1, Ly6C and CD11b. (A) CX3CR1+ cells, consistent with a MΦ phenotype, were selectively gated on in order to encompass the entirety of the myocardium’s MΦ population. (B) Further characterization by Ly6C and CD11b identified a population of CD11b++ Ly6C++ cells that were absent in CCR2-/- myocardia. (C) A CX3CR1+ population in CCR2-/myocardia indicated the preservation of an rCMΦ population. (D) Further characterization, we were able to observe all but one rCMΦ population (i.e. CD11b++ Ly6C++) in CCR2-/-. This population was also CCR2+ and the only population that significantly differed in percentage between CCR2-/- and WT mice. (n=4-6). Fig. 4 – Individual rCMΦ populations were characterized for their expression of TNF-a. Representative flow cytometry for TNF-a expression in CX3CR1+ cells in (A) WT and (B) CCR2-/- myocardia. Gates were applied to CD11b low and high populations for quantitative analysis. TNF-a MFI in CD11b (C) low and (D) high rCMΦ. (n=4-6; *p<0.05). Fig. 5 – Individual rCMΦ populations characterized for their expression of IL-10. Representative flow cytometry for IL-10 expression in CX3CR1+ cells in (A) WT and (B) CCR2-/-

23

ACCEPTED MANUSCRIPT myocardia. Gates were applied to CD11b low and high populations for quantitative analysis. IL-10 MFI in CD11b (C) low and (D) high rCMΦ. (n=4-6; *p<0.05).

SC

RI P

T

Fig. 6 – MΦ-fibroblast co-culture. Representative fields of view stained for Collagen type I, αSMA, and DAPI are shown for the different MΦ co-cultures: (A) undifferentiated MΦ + fibroblasts, (B) M1-differentiated MΦ + fibroblasts, and (C) M2-differentiated MΦ + fibroblasts. Similarly, Representative fields of view stained for Collagen type I, KI-67, and DAPI are shown for the co-cultures: (D) undifferentiated MΦ + fibroblasts, (E) M1differentiated MΦ + fibroblasts, and (F) M2-differentiated MΦ + fibroblasts. Quantification of the (G) collagen type-I, (H) αSMA, and (I) KI-67 are shown below the representative immunofluorescence. Fibroblast monocultures served as controls (images not shown; n=4-6; *p<0.05, ***p<0.001).

CE

PT

ED

MA

NU

Fig. 7 – A schematic of how rCMΦ. turnover and contribution of M1 vs. M2 could be influenced in (A) WT, (B) CCR2-/-, and (C) CX3CR1-/- mice. (A) In WT mice, it is known that classical monocytes recruitment and local proliferation contribute to the CCR2+ portion of rCMΦ. There is evidence that CX3CR1+ cells undergo proliferation to replenish the CX3CR1+ portion of rCMΦ. It is also possible that this population is partially replenished from nonclassical monocytes, but this has yet to be shown. (B) In contrast, we hypothesize that CCR2-/- exhibit impaired recruitment of classical monocytes, thus providing a mechanism by which the CCR2+ M1 population would contract relative to CX3CR1+ M2 rCMΦ. (C) As a complement to the CCR2 data, we have previously shown that CX3CR1-/- mice exhibit an increased CCR2+ rCMΦ population. As such, we hypothesize that the normally CX3CR1+ M2 population contracts in CX3CR1-/-, thus increasing the proportion of M1 to M2 rCMΦ relative to rCMΦ in WT and CCR2-/- mice. (Symbols: - indicates inhibition of a pathway and + indicates enhancement of a pathway)

AC

Table 1 – Quantification of rCMΦ phenotypes. Populations were separated based on their expression of CD11b – an indicator of phenotypic immaturity. Of the CD11b high rCMΦ, CCR2-/- exhibited a significantly lesser CD11b high Ly6C high population relative to their WT counterpart (n=4-6; ***p<0.001). Table 2 – This table represents a summary of our findings in this manuscript.

Fig. 1S – Representative flow cytometry for M cultures post-differentiation. Shown are SSC x ‘marker of interest’ in the different culture conditions and the calculated MFI for said markers.

24

AC

Figure 1

CE

PT

ED

MA

NU

SC

RI P

T

ACCEPTED MANUSCRIPT

25

AC

Figure 2

CE

PT

ED

MA

NU

SC

RI P

T

ACCEPTED MANUSCRIPT

26

AC

Figure 3

CE

PT

ED

MA

NU

SC

RI P

T

ACCEPTED MANUSCRIPT

27

AC

Figure 4

CE

PT

ED

MA

NU

SC

RI P

T

ACCEPTED MANUSCRIPT

28

AC

Figure 5

CE

PT

ED

MA

NU

SC

RI P

T

ACCEPTED MANUSCRIPT

29

AC

Figure 6

CE

PT

ED

MA

NU

SC

RI P

T

ACCEPTED MANUSCRIPT

30

MA

NU

SC

RI P

T

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

Figure 7

31