characterization of the myocardial inflammatory...

J A C C : B A S I C T O T R A N S L A T I O N A L S C I E N C E V O L . 3 , N O . 6 , 2 0 1 8

ª 2 0 1 8 T H E A U T H O R S . P U B L I S H E D B Y E L S E V I E R O N B E H A L F O F T H E AM E R I C A N

C O L L E G E O F C A R D I O L O G Y F O UN DA T I O N . T H I S I S A N O P E N A C C E S S A R T I C L E U N D E R

T H E C C B Y L I C E N S E ( h t t p : / / c r e a t i v e c o mm o n s . o r g / l i c e n s e s / b y / 4 . 0 / ) .

PRECLINICAL RESEARCH

Characterization of the MyocardialInflammatory Response in Acute
Stress-Induced (Takotsubo) CardiomyopathyHeather M. Wilson, BSC, PHD,a Lesley Cheyne, BSC,a Paul A.J. Brown, MBCHB, PHD,b Keith Kerr, MBCHB,a
Andrew Hannah, MBCHB,c Janaki Srinivasan, RDCS,c Natallia Duniak, MBCHB,b Graham Horgan, BA, MSC, PHD,a

Dana K. Dawson, DM, DPHILa

VISUAL ABSTRACT

IS

F

KcD

w

T

A

s

th

M

Wilson, H.M. et al. J Am Coll Cardiol Basic Trans Science. 2018;3(6):766–78.

SN 2452-302X

rom the aSchool of Medicine, Medical Sciences & Nutrition, University of Aberdeen, Forest

ingdom; bDepartment of Pathology, Aberdeen Royal Infirmary, Foresterhill, Aberdeen, Sc

epartment of Cardiology NHS Grampian, Aberdeen Royal Infirmary, Foresterhill, Aberdeen

ork was supported by grants from NHS Grampian Endowments and British Heart Foundation

he authors have reported that they have no relationships relevant to the contents of this pa

ll authors attest they are in compliance with human studies committees and animal welf

titutions and Food and Drug Administration guidelines, including patient consent where appr

e JACC: Basic to Translational Science author instructions page.

anuscript received June 5, 2018; revised manuscript received August 20, 2018, accepted Au

HIGHLIGHTS

� Takotsubo cardiomyopathy is an acute

heart failure syndrome often triggered by

emotional or physical stress, where no

treatment currently exists, and exact

pathogenic mechanisms are unclear.

� Rats in which takotsubo-like cardiomyopathy

was induced showed localized myocardial

inflammatory changes, including progressive

inflammatory infiltrates and myofiber atro-

phy, that persisted over the 14-day time

course examined.

� Early neutrophil infiltrates were followed

by clusters of myocardial macrophages,

typically of an M1 proinflammatory

phenotype, with no switch to M2

resolving macrophages; individual M2

macrophage levels, however, correlated

with recovery in cardiac function. Human

post-mortem myocardial tissue shared

features of the experimental model

demonstrating M1 macrophage clusters.

� The persistent clinical symptoms and long-

term morbidity/mortality observed in takot-

subo patients may, in part, relate to chronic

nonresolving myocardial inflammation.

https://doi.org/10.1016/j.jacbts.2018.08.006

erhill, Aberdeen, Scotland, United

otland, United Kingdom; and the

, Scotland, United Kingdom. This

Project Grant no. PG/15/108/31928

per to disclose.

are regulations of the authors’ in-

opriate. For more information, visit

gust 21, 2018.


http://www.basictranslational.onlinejacc.org/content/instructions-authors

http://crossmark.crossref.org/dialog/?doi=10.1016/j.jacbts.2018.08.006&domain=pdf

http://creativecommons.org/licenses/by/4.0/

R E V I A T I O N S

J A C C : B A S I C T O T R A N S L A T I O N A L S C I E N C E V O L . 3 , N O . 6 , 2 0 1 8 Wilson et al.D E C E M B E R 2 0 1 8 : 7 6 6 – 7 8 Takotsubo Cardiomyopathy and Myocardial Inflammation

767

SUMMARYAB B

AND ACRONYM S

EF = ejection fraction

IL = interleukin

MHC = major

histocompatibility complex

MI = myocardial infarction

qPCR = quantitative

polymerase chain reaction

TNFa = tumor necrosis factor-

alpha

Takotsubo cardiomyopathy is an acute stress-induced heart failure syndrome for which the exact pathogenic

mechanisms are unclear, and consequently, no specific treatment exists. In an experimental model of stress-

induced takotsubo-like cardiomyopathy, the authors describe the temporal course of a chronic inflammatory

response post-induction, with an initial early influx of neutrophils into myocardial tissue followed by macro-

phages that are typical of a proinflammatory M1 phenotype, and a nonsignificant increase in systemic

inflammatory cytokines. Post-mortem myocardium from the more complex clinical takotsubo patients share

features of the study’s experimental model. These findings suggest modulators of inflammation could be

a potential therapeutic option. (J Am Coll Cardiol Basic Trans Science 2018;3:766–78) © 2018 The Authors.

Published by Elsevier on behalf of the American College of Cardiology Foundation. This is an open access article

under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

SEE PAGE 779

A cute stress-induced (takotsubo) cardiomyop-athy has a dramatic clinical presentation,mimicking a myocardial infarction (MI) (1)

and is often precipitated by major emotional/physicalstress. Despite unobstructed coronary arteries, theleft ventriculogram shows characteristic myocardialballooning, not corresponding to a single coronaryartery territory, and the heart function is severelyreduced. Regardless of a spontaneous process of re-covery as a recognized feature of takotsubo cardio-myopathy, the mortality risk and long-termmortality are also comparable with MI (2,3). There isno current treatment, and the etiology is unclear,therefore its pathophysiology needs to be resolvedbefore targeted therapies can be trialed. We have pre-viously demonstrated severe global edema in bothleft and right ventricular myocardium of takotsubopatients during the acute phase with incomplete res-olution at 4 months (4–7). The substrate of this persis-tent edema is most likely an ongoing inflammatoryresponse, which remains to be characterized in detail.However, this is challenging in patients, where coex-istent cardiac pathologies, lack of access to tissue,very small biopsies that do not reflect global changes,and tissue availability only at a single time point, arethe norm.

Here, we used a previously established and well-reported rodent experimental model of catechol-amine (stress)-induced takotsubo-like cardiacdysfunction (8–12), where the disease process isconsistent and reproducible, and there are no con-founding complications or major limitations to tissueavailability. Although it remains unclear whether a“catecholamine surge” is fundamentally causal intriggering human takotsubo cardiomyopathy, thisexperimental model results in a similar wall motionabnormality that is observed in the human disease inthe absence of an ischemic or infectious insult (8–12).Our aim was to establish the type and time course of

the myocardial inflammatory changes in this stress-induced experimental model and to then determinewhether similar changes existed in viable, nonfibroticmyocardium from necessarily more complex humanpost-mortem cases. We specifically focused ondetecting and characterizing macrophages, a hetero-geneous cell population that infiltrates damaged tis-sue and functionally range between proinflammatory(M1 macrophages) and anti-inflammatory/tissuereparative/profibrotic (M2 macrophages) subtypes(13,14). Our results reveal, for the first time to ourknowledge, a clear, localized, and temporal inflam-matory response with a predominance of M1 proin-flammatory macrophages within rat myocardial tissuefollowing induction of takotsubo-like cardiomyopathyand evidence for a similar process in more complexhuman post-mortem cases. Unlike post-MI, there is noswitch to M2-resolving macrophages, and the per-centage ofM2macrophages correlates with recovery ofcardiac function. This low-level, nonresolvinginflammation provides a rationale for an immuno-mechanistically targeted therapeutic strategy in acondition for which no current therapy is available.

METHODS

ANIMALS. All animal work was performed in accor-dance with the United Kingdom Home Office regula-tions and approved by Aberdeen University AnimalEthics Committee. Female Sprague Dawley rats, 4 to 6months old (due to better survival rates on diseaseinduction than in older rats), weighing 300 � 12 gwere housed in a temperature-controlled (25�C) fa-cility with a 12-h light/dark cycle and given free ac-cess to food and water.

HUMAN POST-MORTEM SAMPLES. Research approvalfor post-mortem hearts was obtained under family

http://creativecommons.org/licenses/by/4.0/

Wilson et al. J A C C : B A S I C T O T R A N S L A T I O N A L S C I E N C E V O L . 3 , N O . 6 , 2 0 1 8

Takotsubo Cardiomyopathy and Myocardial Inflammation D E C E M B E R 2 0 1 8 : 7 6 6 – 7 8

768

consent for 2 female patients (69 and 54 years of age)who died shortly after episodes of acute takotsubocardiomyopathy in Aberdeen Royal Infirmary. Patient1 presented with chest pain, and an electrocardiogramshowed 9-mm ST-segment elevation in inferior leadsand ST-segment depression in the correspondingprecordial leads; cardiac catheterization showednormal coronary arteries and a large area of apical andmid-cavity ballooning (Supplemental Figure 1A). Shedied 5 h after presentation. Autopsy established api-cal inferior myocardial rupture with tamponade and50% stenosis of the right coronary ostium. Patient 2had a previous, documented takotsubo event at afamily funeral 3 years before death, and presented atanother family funeral with chest pain. There wasanterior ST-segment elevation on electrocardiogram,moderate stenosis in the right coronary artery oncatheterization, and apical ballooning on left ven-triculogram (Supplemental Figure 1B). She died onthe fifth day of admission, in torsade de pointes andventricular fibrillation. Autopsy established patentcoronary arteries with moderate atheroma of the rightcoronary artery and an old area of left ventricularmyocardial fibrosis. Both patients had a modest,disproportionately low biomarker (troponin) riserelative to the large area of wall motion abnormality.No recent infarction was noted on microscopy at au-topsy. The myocardial tissue used for analysis wasfrom areas unaffected by fibrosis. Four control heartspecimens with no known clinical or post-mortemfindings of heart disease were obtained fromarchived cases. Formalin-fixed myocardium (5-mmsections) were examined by histology and immuno-histochemistry and scored in a blinded fashion byP.A.J.B. and N.D.INDUCTION OF STRESS-INDUCED CARDIOMYOPATHY.

Rats received a single dose of 100 mg/kg isoprenalineintraperitoneally, a protocol that induces reproduc-ible takotsubo-like cardiac dysfunction (8–12) withwidespread akinesia or hypokinesia of apical seg-ments and hypercontractility of the cardiac base in85% of rats. This model results in supraphysiologicallevels of circulating catecholamines, initial highblood pressure and heart rates, and a high yield ofwall motion abnormality after the induced stress ofisoprenaline and thus provides a stress-inducedmodel in the absence of ischemic and pathogen-induced inflammatory insults. Systolic regional wallmotion abnormalities in this model disappear be-tween days 6 to 15 post-induction. Survival ratesfollowing injection were 90%. Similar volumes ofintraperitoneal saline injections were given to thecontrol rat group. Rats were sacrificed serially onday 0 (6 h after injection) and days 1 to 7 and day 14

after disease induction, with group sizes as indicatedfor each experimental analysis.

ECHOCARDIOGRAPHY. Echocardiography was per-formed before sacrificing rats to ascertain the pres-ence of wall motion abnormality and to quantify thecardiac function. A Vivid Q equipment (GE Health-care, Oslo, Norway) with an 11.5-MHz pediatric probewas used to acquire images. Rats were placed underanesthesia (1.25% isoflurane in oxygen after 4% iso-flurane induction), and chests were shaved. One long-axis and 3 short-axis views (base, mid, and apex) wereacquired. The duration of anesthesia did not exceed 2min in any rat. Heart rate was 290 � 5 beats/minduring echocardiography (normal rate). Ejectionfraction (EF) was calculated using the area–lengthmethod (15). Characteristics of experimental stress-induced cardiomyopathy are shown in SupplementalFigure 2.

RAT HEART TISSUE PROCESSING. After schedule 1sacrifice, rat hearts were fixed in 5% phosphate-buffered formaldehyde and embedded in paraffin.All hearts were sectioned in 3 transverse blocks: apex,mid-heart, and base, and 4-mm sections from eachlevel examined by histology and immunohistochem-istry. Isoprenaline-injected rats for each of the timepoints (day 0 equivalent to 6 h after injection), days 1to 7 and day 14 (n ¼ 7 per group) was compared withsaline-injected control rats (n ¼ 7 per group), that is,70 animals in total. Because some deaths occurred oninjection in our designed groups, the number of ani-mals available for each group was 5 to 7 animals asindicated in Results. In additional experiments,hearts were snap frozen for gene expression analysis(control, saline-injected animals, and animals sacri-ficed on day 3 and day 14 after induction of takotsubocardiomyopathy).

HISTOPATHOLOGY. Hematoxylin and eosin–stainedsections from the mid-heart, where the wall motionabnormality and subendocardial hemorrhage wereobserved, were scored semiquantitatively in a blin-ded manner for histological/pathological changes byan experienced consultant pathologist (K.K.). Arbi-trary scores of 0, 1, and 2 were given for hemorrhage,immune/inflammatory cellular infiltrates, eosino-philic fiber staining areas, and degenerate myofibrils/fiber atrophy where the scoring system was defined as0 ¼ no lesions, 1 ¼ scattered/patchy/little present,and 2 ¼ readily identifiable throughout the section. Insome sections, hemosiderin was identified in macro-phages and confirmed by Prussian blue stain, and wasscored on a scale of 0 to 2 (Supplemental Figure 3).Most insults occurred in regional areas of the sectionas illustrated in Supplemental Figure 3.








769

INFLAMMATORY CELL COMPOSITION: IMMUNOHISTO-

CHEMISTRY. Sections were deparaffinized, micro-waved in citrate buffer (pH 6.0), and endogenousperoxidase activity quenched by incubation with 3%H2O2. Neutrophil counts were determined from he-matoxylin and eosin–stained sections by morphology.Macrophages and T cells were detected by anti-CD68antibodies (ED1, rat and KP1, human) (Abcam, Cam-bridge, United Kingdom) and anti-CD3 antibodies(MCA772GA, rat, Bio-Rad, Hercules, California; andF7.2.38, human, Dako, Glostrup, Denmark), respec-tively, and B cells by anti-CD20 antibody, L26(Abcam) and the DAKO Envision Detection Kit K5007.Specific brown staining was developed by incubationwith Chemate-DAB (Dako, Ely, United Kingdom).Slides were counterstained with hematoxylin, andnuclei blue stained with Scott’s Tap Water. The DAKOG2-Double staining kit, K5361 (DakoCytomation,Glostrup, Denmark) was used for staining with anti-CD68 (ED1, rat and KP1, human) and the M1-macrophage markers anti-major histocompatibilitycomplex (MHC) class II (OX-3, rat, TAI.1B5, human)and inducible nitric oxide synthase (iNOS) (AB15323;Bio-Rad) for rat, and SOCS3 (AB16030, Abcam) forhuman and the M2-macrophage marker, anti-CD163(ED2, Abcam for rat; EDHu-1, Bio-Rad for human),according to the manufacturers’ instructions. Positivestaining was detected using diaminobenzidine/LiquidPermanent-Red (DakoCytomation) (16). A negativecontrol where primary antibody was replaced by anonspecific IgG equivalent was included in allanalyses.

QUANTIFICATION OF NEUTROPHIL, MACROPHAGE,

AND T-CELL COUNTS. The mid-heart sections of rathearts was partitioned into 8 consecutive sections(Supplemental Figure 4), and these were used toquantify macrophage, T-cell, and neutrophil counts.Each section was separated into 3 fields of view(epicardium, mid-wall, and endocardium) to directlycompare areas. For human post-mortem tissue,defined sections approximately 3 cm � 2 cm ofmyocardial tissue was used. Intact CD68-positivecells were counted, and absolute counts representthe mean number of macrophages from each of the 8sections of rat tissue. The entire area of human post-mortem tissue was considered, and the averagemacrophage number per field determined from 15fields per section was determined. Counts wererepeated until 3 consecutive values were obtainedwith a margin of error of �5%. A similar techniquewas used to count the CD3-positive T cells and neu-trophils per area. Both the intraobserver and inter-observer variation were found to be <5% (p < 0.002).

The number of double-positive cells were countedand expressed as a percentage of total macrophages.

QUANTIFICATION OF SERUM AND TISSUE CYTOKINES.

Serum samples collected from blood derived on car-diac puncture at the time of sacrifice were analyzedfor tumor necrosis factor (TNF)-a, interleukin (IL)-6,and IL-10 as determined by DuoSet enzyme-linkedimmunosorbent assay kits from R&D Systems (Min-neapolis, Minnesota) according to the manufacturer’sinstructions. Cytokine levels in the myocardialtissue were determined by quantitative polymerasechain reaction (qPCR). Tissues were homogenized inTRIzol reagent (Sigma-Aldrich, St. Louis, Missouri)and cDNA synthesis carried out from 1 mg of RNAusing a Tetro cDNA Synthesis Kit (Bioline, London,United Kingdom). qPCR was performed using a LightCycler 480 (Roche, Basel, Switzerland), and geneexpression of TNF-a, il-6 and il-10, was determined inrelation to housekeeping gene HPRT. Groupsanalyzed consisted of control, saline-injected animals(n ¼ 10), isoprenaline-injected animals sacrificed 3days after injection (n ¼ 9), and isoprenaline-injectedanimals sacrificed 14 days after injection (n ¼ 5).

STATISTICAL ANALYSES. The Mann-Whitney U testwas used to compare the macrophage, neutrophil,and T-cell counts/heart section. Pearson’s correlationcoefficient determined the significance of a correla-tion between macrophages and their subtypes withheart EF. All values are reported as mean � SD. Avalue of p < 0.05 was considered as significant.

RESULTS

HISTOLOGICALCHARACTERIZATIONOFINFLAMMATIONIN

STRESS-INDUCED TAKOTSUBO-LIKE CARDIOMYOPATHY.

To determine initially the type and time course of themyocardial inflammatory changes in the experi-mental model of stress-induced cardiomyopathy,sections of rat heart were characterized histologically.Six hours after administration of isoprenaline to rats,histopathological examination revealed localizedsubendocardial hemorrhage in the mid-ventricularregion of the heart (maximum histopathologicalscore 1.9) (Figure 1A) and post-day 5 showed anincrease in hemosiderin-containing phagocytic cells(confirmed by Prussian blue stain) (SupplementalFigure 3D), supporting a resolution of injury. An im-mune cell/inflammatory infiltrate was evident after 6h, peaking at days 3 to 4 showing scattered, but later,obvious infiltrates in regional areas, subsiding by day5 (Figure 1B, Supplemental Figure 3). This infiltratecorrelated with prevalence of degenerate myofiberswhere histological scores peaked at day 4, subsiding





FIGURE 1 Histological Characterization of Mid-Heart Sections of Control Rats and From Rats After Induction of

Takotsubo Cardiomyopathy

Histological characterization of hematoxylin and eosin–stained mid-heart sections of control rats and from rats over the time course (0 to 14

days) after induction of takotsubo cardiomyopathy show a temporal immune/inflammatory cellular infiltrate. Sections of heart tissue were

semiquantitatively scored by a consultant histopathologist (K.K.) blinded to the study. Arbitrary scores of 0, 1, and 2 were given for (A)

hemorrhage (gray bars) or presence of hemosiderin-containing phagocytic cells (white bars), (B) immune/inflammatory cellular infiltrates,

(C) degenerate myofibrils/myofiber atrophy, and (D) eosinophilic myofibers. The number of animals per group were as follows; control n ¼7, day 1 n ¼ 6, day 2 n ¼ 6, day 3 n ¼ 7, day 4 n ¼ 6, day 5 n ¼ 7, day 6 n ¼ 6, day 7 n ¼ 5, and day 14 n ¼ 5 animals. Statistical difference in

histological scores between control rats and those injected with isoprenaline were calculated with p values denoted above each time point bar.

Values in bold indicate statistically significant differences compared with the control group.



770

at day 5. Beyond day 5, fiber atrophy was moreevident, then decreased by day 14 (Figure 1C).Hypereosinophilic myofibers were observed, peakingat day 6 when nuclear pyknosis was observed(Figure 1D), although this change was found across the14-day period examined. Fibrosis was also identifiedin late-stage sections (days 6, 7, and 14), consistentwith the sequelae of earlier inflammatory changes.Vascular proliferation was found in 50% of animals byday 14. No gross changes in monocyte numbers orhistology of the spleen could be detected between theanimal groups.

DETAILED CHARACTERIZATION OF IMMUNE CELL

INFILTRATE OVER TIME. The increase in immune/inflammatory cell infiltrates observed by histologicalanalysis over the time course of disease was exam-ined in more detail by specific cell characterizationdefining the macrophages, neutrophils, and T-cellcounts within the rat mid-heart sections. Six hoursfollowing isoprenaline injection, early neutrophilicinfiltrates predominated, peaking at day 1 to 2 andsubsiding thereafter, typical of an inflammatoryresponse (Figure 2). Macrophage counts peaked at day3 and were typically found in regional clusters.

FIGURE 2 Immunohistochemical Analysis Following Induction of Stress-Induced Takotsubo-Like Cardiomyopathy

Immunohistochemical analysis demonstrates progressive neutrophil and macrophage infiltration into myocardial tissue (day 0 to 14) following

induction of stress-induced takotsubo-like cardiomyopathy. Representative examples of neutrophil (A) and macrophage (CD68-positive

cells) (B) staining in mid-region rat myocardial tissue, original magnification �400. Neutrophil (C) and macrophage (D) cell counts in the mid-

region of saline injected (control) and isoprenaline injected rat hearts (day 0 to 14). Values represent mean counts � SD in 24 fields of view

from defined areas as described in the Methods section. The number of animals per group were as follows: control n ¼ 7, day 1 n ¼ 6, day 2

n ¼ 6, day 3 n ¼ 7, day 4 n ¼ 6, day 5 n ¼ 7, day 6 n ¼ 6, day 7 n ¼ 5, and day 14 n ¼ 5 animals. Statistical difference in mean number of

neutrophil and macrophage counts between control rats and those injected with isoprenaline were calculated with p values denoted above

each time point bar. Values in bold indicate statistically significant differences compared with the control group.


771

Macrophage infiltrates subsided by day 5, but werestill evident at day 14 and did not return to baselinelevels, consistent with the time scale of the immunecell infiltrate noted on histological assessment. Weobserved no detectable T-cell infiltrates over the timecourse examined.

Cardiac function as determined by the change inpercentage EF pre- and post-isoprenaline injection,deteriorated immediately after disease induction,and partially recovered thereafter, although not tobaseline (Supplemental Figure 2). The most signifi-cant decline in cardiac function was observed at the6-h post-induction time point and thus precededmacrophage infiltration and the time of peak inflam-matory cell infiltrate.

COMPARISONS OF THE PROPORTION OF M1 AND M2

MACROPHAGE SUBTYPES OVER THE COURSE OF

DISEASE. Given the increase in macrophage numbersin heart sections of our experimental model oftakotsubo-like cardiomyopathy, it was important tocharacterize their inflammatory phenotype. The pro-portions of M1 (proinflammatory tissue destructivemacrophages expressing markers iNOS and MHC classII) and M2 (CD163-positive anti-inflammatory, tissuereparative/profibrotic macrophages) (11,12) weretherefore examined (Figure 3A).

The percentage of MHC class II–positive macro-phages and iNOS-positive macrophages, indicative ofM1 cells, increased over time after disease induction(Figures 3B and 3C), peaking at day 4. The percentages


FIGURE 3 Double Immunostaining of Mid-Heart Sections Showing M1-Activated and M2-Positive Macrophages

Double immunostaining of mid-heart sections show a greater predominance of M1-activated macrophages over time with little change in percentage of M2-positive

macrophages. Total macrophages were detected by anti-CD68 and liquid permanent red staining (red), whereas macrophage activation markers were identified by the

relevant antibodies and DAB staining (brown). Macrophage-rich areas were analyzed as outlined in Methods. Examples of double-stained macrophages are highlighted

by arrows in A. Original magnification �400. Changes in percentages of double-stained macrophages over time for inducible nitric oxide synthase (iNOS) (B), major

histocompatibility complex (MHC) class II (C), and CD163 (D). The number of animals per group were as follows; control n ¼ 6, day 1 to 14 n ¼ 5 for each time point.

Statistical difference in the percentage iNOS, MHC class II, and CD163-positive macrophages between control rats and those injected with isoprenaline were calculated

with p values denoted above each bar. Values in bold indicate statistically significant differences compared with the control group.



772

decreased between day 4 and 5 and then demon-strated a second peak of increase at days 6 to 14consistent with the increase in total macrophagecounts at that time and a classic inflammatoryresponse.

The percentage of CD163-positive M2 macrophagesshowed a very different pattern. They increased fromday 0 to day 1, suggesting increased macrophageactivation. The percentage did not change signifi-cantly from day 1 to day 6, but fell on day 7 of disease(Figure 3D).

To determine whether there was a relationshipbetween the total macrophage counts or the per-centage of M1 and M2 macrophages and recovery ofcardiac function, we correlated the macrophagenumbers or M1/M2 percentages with percentage leftventricular EF determined before culling (Figure 4).There was no significant correlation between thelevel of total myocardial macrophage infiltration or

percentage M1 macrophages and the heart EF(R2 ¼ 0.003; p ¼ 0.885 total macrophages; R2 ¼ 0.114;p ¼ 0.259, iNOS-positive M1 macrophages; and R2 ¼0.049; p ¼ 0.469, MHC class II–positive M1 macro-phages). There is a significant correlation (R2 ¼ 0.383;p ¼ 0.018) between the percentage of M2 macro-phages and percentage EF, indicating that a greaterpreponderance of M2 cells improved recovery of car-diac function.LOCAL AND SYSTEMIC CYTOKINE CHANGES OVER

THE COURSE OF DISEASE. The levels of pro- andanti-inflammatory cytokines were also analyzed post-isoprenaline injection and compared with control ratsto determine whether local and systemic inflamma-tion was apparent and could contribute to theunderlying macrophage polarization. Systemiccytokine levels were low (IL-6 and IL-10) or unde-tectable (TNF-a) with or without induction oftakotsubo-like cardiomyopathy. At day 3 and day

FIGURE 4 Correlations Between the Total Macrophage Counts or the Percentage M1 and M2 Macrophages and Recovery of

Cardiac Function

Correlations between the total macrophage counts or the percentage M1 and M2 macrophages and recovery of cardiac function as determined

by percentage left ventricular ejection fraction (EF). The total number of macrophages (A) or percentage of iNOS (B) and MHC class II–

positive (C)M1 macrophages and CD163-positive (D)M2-like macrophages was determined from immunohistochemical analysis and correlated

with the percentage ejection fraction before cull. A total of 13 animals were analyzed. The coefficient of determination (R2) and p value were

determined by Pearson’s correlation test.


773

14 post-induction, IL-6 and IL-10 systemic levels wereelevated over control animals (Figures 5A and 5B),although this did not reach statistical significance(IL-6 day 3 p ¼ 0.0952, IL-6 day 14 p ¼ 0.0754; IL-10day 3 p ¼ 0.2103, IL-10 day 14 p ¼ 0.1904). Geneexpression levels for IL-6 and IL-10 in myocardialtissue were also greater post-disease induction(Figures 5C and 5D), but again, this increase did notreach statistical significance (IL-6 day 3 p ¼ 0.3951,IL-6 day 14, p ¼ 0.4739; IL-10 day 3 p ¼ 0.4815, IL-10day 14 p ¼ 0.5000). Gene expression levels for TNFawere on the limit of detection by qPCR for both con-trol and takotsubo-like cardiomyopathy animals (datanot included).

Overall, these results in our experimental modelsuggest a characteristic ongoing inflammatoryresponse within the rat myocardium, with early andprominent neutrophil infiltration, followed by M1proinflammatory macrophages.

CHARACTERIZATION OF INFLAMMATORY RESPONSES IN

HUMAN POST-MORTEM TAKOTSUBO CARDIOMYOPATHY

PATIENTS. To determine whether such an inflamma-tory component observed in a rat model could also beidentified in post-mortem samples of human takot-subo cardiomyopathy patients, myocardial tissuefrom 2 patients (Patient #1 deceased 5 h and Patient#2 deceased 5 days post-presentation) was assessed.The myocardium used for assessment was unaffectedby any visible macroscopic fibrotic changes orinfarction. Occasional small areas of lymphocytic andmacrophage infiltration were evident in the inter-stitium and within areas of myocytes underlying thepericardial fat. Focal subendocardial hemorrhage wasalso present. For control hearts with no known heartcondition and no post-mortem findings of heart dis-ease, less prominent lymphocytic and macrophageinfiltration was observed histologically in all cases(n ¼ 4).

FIGURE 5 Analysis of Systemic and Myocardial Cytokine Levels in Control Animals and Those With Takotsubo-Like Cardiomyopathy

Analysis of systemic and myocardial cytokine levels in control animals and those with Takotsubo-like cardiomyopathy at day 3 (d3) and day 14

(d14) are shown. Serum levels of IL-6 (A) and IL-10 (B) are mean pg/ml levels � SD; gene expression levels of IL-6 (C) and IL-10 (D). Groups

analyzed consisted of control, saline-injected animals (n ¼ 10), isoprenaline-injected animals sacrificed 3 days after injection (n ¼ 9), and

isoprenaline-injected animals sacrificed 14 days after injection (n ¼ 5). Values are mean fold change over control animals � SD. Values above

each bar indicate p values compared with control animals. IL ¼ interleukin.



774

Immunohistochemical analysis demonstrated ascattered distribution of CD3-positive cells in bothtakotsubo cardiomyopathy patients’ post-mortemmyocardial tissues, whereas control tissues demon-strated sparse cell positivity and 1 case that revealedscattered groups of CD3-positive cells. Control tissueshowed no positive staining for CD20 (B lympho-cytes), whereas both patients revealed a scatteredCD20 cell positivity. Patient #2, deceased 5 daysfollowing presentation, demonstrated a scatteredmacrophage distribution with occasional aggregatedclusters of macrophages in the subendothelium andsubpericardial regions (Figure 6), whereas aggregatedclusters were less obvious in Patient #1, deceased 5 hafter onset of symptoms or in control subjectmyocardium.

More quantitative analysis of macrophage countsshowed the average macrophage number per fieldin patients was significantly higher than in controlsubjects (10.6 � 2.8 Patient #1; 27.01 � 16.1 Patient#2, vs. 6.7 � 5.4 control tissue; p < 0.05). Tocompare the M1/M2 phenotypes of macrophagesbetween groups and relate results to that of ourexperimental model, double immunohistochemistrywas performed. As with our experimental model,there was a greater proportion of M1 macrophages(CD68/HLADR or CD68/SOCS3 positive) in myocar-dial tissue from takotsubo cardiomyopathy patientswhen compared with control tissue, again reflectinga potential proinflammatory response (Figure 6).By contrast, there was no obvious difference inthe proportion of M2, CD163-positive myocardial

FIGURE 6 Post-Mortem Myocardial Tissue From Takotsubo Patients

Post-mortem myocardial tissue from takotsubo patients show the presence of macrophages and demonstrate an increase/predominance of M1

macrophages. Macrophage staining (CD68-positive cells) showing, as indicated by arrows: (A) scattered macrophages in the subendothelial

area of the myocardium in a control subject with no underlying heart condition, (B) very few subendothelial aggregates of macrophages

in patient 1 deceased 5 h after presentation with takotsubo cardiomyopathy, and (C) subpericardial aggregates of macrophages in the

myocardial tissue from patient 2 with takotsubo cardiomyopathy deceased 5 days after presentation. Original magnification �400.

(D) Percentage of CD68-positive macrophages staining for M1 markers HLADR and SOCS3, and M2 marker CD163 in patients with

takotsubo cardiomyopathy (TTC) (n ¼ 2) and age/sex-matched control subjects (control) (n ¼ 4).


775

tissue macrophages between patients and controlsubjects.

DISCUSSION

Several mechanisms have been implicated in patho-genesis of takotsubo cardiomyopathy, including cor-onary artery spasm, microvascular dysfunction, andexcessive release of catecholamines in response tostress (17,18). None of these mechanisms, however,has been conclusively proven to be responsible forthe disease solely or in combination, accounting forthe lack of targeted therapies. Inflammation has beenimplicated from noninvasive imaging, because of theintense myocardial edema observed on cardiac mag-netic resonance imaging (3,5), and inflammatorychanges in the heart have been reported from

biopsies of patients, albeit these were taken from theright ventricular myocardium (19). Proteomic analysisof apical cardiac tissue in an experimental model alsoproposes inflammation might be involved (12), and anincrease in macrophage infiltration has been observed24 h after disease induction (20). To characterize fullythe inflammatory aspects of the disease, we used awell-published experimental model of stress-inducedtakotsubo-like cardiomyopathy, initiated in theabsence of ischemic or pathogen-induced inflamma-tory insults, to follow a detailed time course of in-flammatory changes in myocardial tissue, focusing onthe infiltration of inflammatory macrophages and thesubtypes present, and whether inflammation waslocalized or systemic. Our experimental model elim-inates the complications of any confounding clinicalconditions and comorbidities that are found with



776

patients and, for the first time, provides a modelwhereby we can discriminate and directly comparespecific inflammatory responses over a controlledtime period. First, we demonstrate a classic inflam-matory response upon disease induction, with earlyinfiltration of neutrophils followed by an increase inmacrophages and an increase in cytokine geneexpression. Second, the increase in myocardial mac-rophages corresponded to the increase in the per-centage of those with a proinflammatory M1phenotype. Third, the percentage of M2 macrophagesdid not increase significantly over time, but individ-ual levels correlated with recovery in cardiac func-tion. Fourth, the inflammation appears localized withno substantial increase in circulating cytokine levels;and lastly, the inflammatory cell characterization inour experimental rat myocardial tissue was recapit-ulated in post-mortem myocardial tissue from takot-subo cardiomyopathy patients.

Macrophage functions help maintain cardiovascu-lar health; however, following an insult, M1 macro-phages, which are associated with tissue damage,dominate, and their inflammatory actions canunleash functions that promote disease (13,14).Subsequently, M2 macrophages produce anti-inflammatory, proangiogenic, and proreparative fac-tors (e.g., vascular endothelial growth factor, andTGF-b1) and engulf apoptotic cells to facilitate repair(13,14). Following induction of stress-induced takot-subo-like cardiomyopathy, the infiltrating macro-phages show a temporal increase in their M1activation and proinflammatory characteristics peak-ing at day 4 with a second peak at day 7, a time pointwhere there was a corresponding decrease in thepercentage of M2-activated cells. A normal heart hasits own tissue-resident cardiac macrophages thatdisplay a M2-like, CD163-positive phenotype (21). Thedrop in percentage of CD163-positive macrophagesmay reflect exhaustion of these tissue-resident cellsafter myocardial injury and the replacement bymonocyte-derived macrophages infiltrating from thecirculation (22,23). The failure of an up-regulation inthe percentage of M2 macrophages at days 5 to 14strongly contrasts with that observed with models ofMI where there is a peak of M1 macrophages and arobust increase in systemic inflammatory cytokinelevels. This is followed by increased M2 macrophages(days 5 to 7) that initiates myocardial remodeling andpotentially scarring (24,25). By contrast, in ourtakotsubo-like experimental model, we observe apeak of M1 macrophages, but no secondary peak ofM2 reparative macrophages, and low and no signifi-cant increase in circulating cytokine levels. This lack

of the timely suppression and spatial containment ofthe inflammatory response to the myocardium in ourtakotsubo-like model may perpetuate into a long-term low-grade chronic inflammatory state. Thissupports our recently demonstrated clinical conceptof continued and persistent symptoms rather thancomplete resolution of disease observed in takotsubopatients (26,27). Indeed, we show that the increase inpercentage EF, as a measure of cardiac function,correlated positively with the percentage of M2macrophages. This suggests that shifting the balancefrom M1 to M2 macrophages by specific therapies maybe beneficial by promoting myocardial repair andfunctional improvement post-takotsubo. These ideasneed to be validated in future pre-clinicalexperiments.

In our study, we did not set out to determinewhether the inflammatory response observed wasthe initial cause of the disease or a response toan insult. Given that the greatest drop in cardiacfunction (as determined by percentage EF) is imme-diately after disease induction and precedes the peakof neutrophil and macrophages into the myocar-dium, we consider that the nonresolving inflamma-tion is a secondary feature of disease. It remainsunclear what would initiate the inflammatoryresponse in the human disease although mechanicalor catecholamine-induced myocyte stress in theakinetic/dyskinetic myocardial regions has beensuggested as a potential cause (17–19). Regardlessof the instigating trigger, the persistent presence ofproinflammatory cells such as M1-type macrophageswould be expected to contribute to the longer-termpathology.

As with rat myocardium, myocardial tissue fromPatient #2, deceased 5 days after diagnosis, revealedmore prominent infiltrates of macrophages whencompared with Patient #1 deceased 5 h after diag-nosis, suggesting a primary insult in humans couldalso result in a progressive inflammatory response,albeit a greater number of post-mortem studies arerequired to confirm this. Our takotsubo post-mortemhearts exhibit a predominance of M1 proin-flammatory macrophages that could initially be dis-missed as being associated with other cardiacpathologies, such as coronary artery disease. In thesetting of MI, it is well accepted that macrophageinfiltrates track to a characteristic localized area ofischemia-induced injury/scarring. By contrast, ourfindings suggest that in takotsubo, localized M1macrophage infiltrates occur within viable myocardialareas. This may be a potential marker of takotsuboeven when other well-documented cardiac

PERSPECTIVES

COMPETENCY IN MEDICAL KNOWLEDGE: Low-grade

myocardial inflammation is a distinguishing feature following

both experimental and clinical stress-induced takotsubo cardio-

myopathy. Our results establish that a proinflammatory (M1

macrophages) phenotype prevails following a decline in cardiac

function, which may contribute toward the morbidity and

increased mortality associated with this condition in a subset of

patients.

TRANSLATIONAL OUTLOOK: No current therapy exists for

takotsubo cardiomyopathy. Our study suggests that targeting

the myocardial inflammatory process at the onset of disease may

have merit as a clinical intervention to alleviate the long-term

persistent symptoms that occur in a number of patients.


777

pathologies are present. Imaging of macrophages byultrasmall superparamagnetic iron oxides in themyocardium of takotsubo patients in the acute phasecould provide further indications.

Myocardial biopsies from takotsubo cardiomyop-athy patients have previously supported the idea ofa potential inflammatory response. In a study byWittstein et al. (28), of the 5 patients with takotsubocardiomyopathy who underwent right ventricularendomyocardial biopsy, 4 had interstitial infiltratesconsisting primarily of mononuclear lymphocytesand macrophages and contraction bands withoutmyocyte necrosis. A further histopathological anal-ysis of a myocardial biopsy sample obtained from a71-year-old woman, who died 8 h after onset oftakotsubo, revealed multiple necrotic lesions,monocytes and neutrophil infiltration, and hemor-rhage (29), whereas the presence of inflammatorycells (mainly monocytes) and disrupted myofibershas also been reported (30,31). Unlike our currentstudy, however, none of these studies definitivelycompared biopsies with those of control subjectswith no underlying heart pathology, or defined thenature of the inflammatory cells or phenotype ofmacrophages.

There are no consensus recommendations for long-term management of takotsubo cardiomyopathy toprevent recurrences, and the current clinical man-agement of these patients presents significant clinicalequipoise. Long-term treatment with b-blockers, orcalcium channel blockers and aspirin did not improveleft ventricular function in a multicenter retrospec-tive study (32). From a translational perspective,our results point to an inflammatory component,suggesting that immune-modulation therapies maybe a therapeutic avenue in this condition, and inter-vention studies should be considered.

STUDY LIMITATIONS. The experimental model usedis a surrogate and may not fully reflect clinical humantakotsubo cardiomyopathy where there are multiplepredisposing and/or precipitating factors. Catechol-amines have been described as a precipitating factorin human takotsubo development, but they are notthe only one. This is currently the only well-published model for this stress-induced disorder,providing us with a unique opportunity to temporallyassess the inflammatory response in the absenceof multiple predisposing factors and comorbiditiesthat can complicate the analysis in patients.The number of post-mortem cases with Takotsubo

cardiomyopathy in this study was limited to2 patients, and therefore will require confirmationin larger post-mortem datasets.

CONCLUSIONS

We demonstrate for the first time that a classicinflammatory response with predominant M1 macro-phage infiltration of the myocardium, without aconventional switch to M2 macrophages is a charac-teristic of experimental takotsubo-like cardiomyopa-thy. This results in low-grade chronic inflammationin myocardial tissue that has potential to drivethe long-term and persistent symptoms observed intakotsubo patients. Similar inflammatory changeswere observed in otherwise macroscopically normalmyocardium of post-mortem takotsubo humanhearts, pointing to a potential new marker of thiscondition and suggesting inflammatory modulators,for example, switching the balance from M1 to M2macrophages may be a therapeutic option.

ACKNOWLEDGMENTS The authors acknowledge thehelp of Dr. Marc van Bilson and Chantal Munt,Fysiologie, School for Cardiovascular Diseases, Fac-ulty of Health, Medicine and Life Sciences, MaastrichtUniversity, in cDNA preparation.

ADDRESS FOR CORRESPONDENCE: Dr. Heather M.Wilson, School of Medicine, Medical Sciences &Nutrition, University of Aberdeen, Foresterhill,Aberdeen, AB25 2ZD, Scotland, United Kingdom.E-mail: [email protected].

mailto:[email protected]



778

RE F E RENCE S

1. Prasad A, Lerman A, Rihal CS. Apical ballooningsyndrome (Tako-Tsubo or stress cardiomyopathy):a mimic of acute myocardial infarction. Am Heart J2008;155:408–17.

2. Tornvall P, Collste O, Ehrenborg E, Jarnbert-Petterson H. A case-control study of risk markersand mortality in takotsubo stress cardiomyopathy.J Am Coll Cardiol 2016;67:1931–6.

3. Stiermaier T, Moeller C, Oehler K, et al. Long-termexcess mortality in takotsubo cardiomyopathy: pre-dictors, causes and clinical consequences. Eur JHeart Fail 2016;18:650–6.

4. Dawson DK, Neil CJ, Henning A, et al. Tako-Tsubo cardiomyopathy: a heart stressed out ofenergy? J Am Coll Cardiol Img 2015;8:985–7.

5. Scally C, Ahearn T, Rudd A, et al. Right ven-tricular involvement and recovery after acutestress-induced (tako-tsubo) cardiomyopathy. Am JCardiol 2016;117:775–80.

6. Schwarz K, Ahearn T, Srinivasan J, et al. Alter-ations in cardiac deformation, timing of contrac-tion and relaxation, and early myocardial fibrosisaccompany the apparent recovery of acute stress-induced (takotsubo) cardiomyopathy: an end tothe concept of transience. J Am Soc Echocardiogr2017;30:745–55.

7. Neil CJ, Nguyen TH, Singh K, et al. Relation ofdelayed recovery of myocardial function aftertakotsubo cardiomyopathy to subsequent qualityof life. Am J Cardiol 2015;115:1085–9.

8. Sachdeva J, Dai W, Kloner RA. Functional andhistological assessment of an experimental modelof Takotsubo’s cardiomyopathy. J Am Heart Assoc2014;3:e000921.

9. Shao Y, Redfors B, Scharin Täng M, et al. Novelrat model reveals important roles of b-adrenor-eceptors in stress-induced cardiomyopathy. Int JCardiol 2013;168:1943–50.

10. Paur H, Wright PT, Sikkel MB, et al. High levelsof circulating epinephrine trigger apical car-diodepression in a b2-adrenergic receptor/Gi-dependent manner: a new model of Takotsubocardiomyopathy. Circulation 2012;126:697–706.

11. Willis BC, Salazar-Cantú A, Silva-Platas C, et al.Impaired oxidative metabolism and calcium mis-handling underlie cardiac dysfunction in arat model of post-acute isoproterenol-induced

cardiomyopathy. Am J Physiol Heart Circ Physiol2015;308:H467–77.

12. Oras J, Redfors B, Ali A, et al. Anaesthetic-induced cardioprotection in an experimentalmodel of the Takotsubo syndrome: isoflurane vs.propofol. Acta Anaesthesiol Scand 2017;61:309–21.

13. Murray PJ, Allen JE, Biswas SK, et al. Macro-phage activation and polarization: nomenclatureand experimental guidelines. Immunity 2014;411:14–20.

14. Sica A, Mantovani A. Macrophage plasticityand polarization: in vivo veritas. J Clin Invest 2012;122:787–95.

15. Dawson D, Lygate CA, Saunders J, et al.Quantitative 3-dimensional echocardiography foraccurate and rapid cardiac phenotype character-ization in mice. Circulation 2004;110:1632–7.

16. Shaikh S, Brittenden J, Lahiri R, Brown PA,Thies F, Wilson HM. Macrophage subtypes insymptomatic carotid artery and femoral arteryplaques. Eur J Vasc Endovasc Surg 2012;44:491–7.

17. Nef HM, Möllmann H, Akashi YJ, Hamm CW.Mechanisms of stress (Takotsubo) cardiomyopa-thy. Nat Rev Cardiol 2010;7:187–93.

18. Goico A, Chandrasekaran M, Herrera CJ. Noveldevelopments in stress cardiomyopathy: Frompathophysiology to prognosis. Int J Cardiol 2016;223:1053–8.

19. Nef HM, Möllmann H, Kostin S, et al. Tako-tsubo cardiomyopathy: intraindividual structuralanalysis in the acute phase and after functionalrecovery. Eur Heart J 2007;28:2456–64.

20. Surikow SY, Nguyen TH, Stafford I, et al.Nitrosative stress as a modulator of inflammatorychange in a model of takotsubo syndrome. J AmColl Cardiol Basic Trans Science 2018;3:213–26.

21. Pinto AR, Paolicelli R, Salimova E, et al. Anabundant tissue macrophage population in theadult murine heart with a distinct alternatively-activated macrophage profile. PLoS One 2012;7:e36814.

22. Pinto AR, Godwin JW, Rosenthal NA. Macro-phages in cardiac homeostasis, injury responsesand progenitor cell mobilisation. Stem Cell Res2014;13:705–14.

23. Hulsmans M, Sam F, Nahrendorf M. Monocyteand macrophage contributions to cardiac remod-eling. J Mol Cell Cardiol 2016;93:149–55.

24. Yan Yan X, Anzai A, Katsumata Y, et al. Tem-poral dynamics of cardiac immune cell accumula-tion following acute myocardial infarction. J MolCell Cardiol 2013;62:24–35.

25. Libby P, Nahrendorf M, Swirski FK. Leukocyteslink local and systemic inflammation in ischemiccardiovascular disease: an expanded “cardiovascularcontinuum”. J Am Coll Cardiol 2016;67:1091–103.

26. Templin C, Ghadri JR, Diekmann J, et al.Clinical features and outcomes of takotsubo(stress) cardiomyopathy. N Engl J Med 2015;373:929–38.

27. Scally C, Rudd A, Mezincescu A, et al. Persis-tent long-term structural, functional, and meta-bolic changes after stress-induced (takotsubo)cardiomyopathy. Circulation 2018;137:1039–48.

28. Wittstein IS, Thiemann DR, Lima JA, et al.Neurohumoral features of myocardial stunningdue to sudden emotional stress. N Engl J Med2005;352:539–48.

29. Elsokkari I, Cala A, Khan S, Hill A. Takotsubocardiomyopathy: not always innocent or predict-able. A unique post mortem insight. Int J Cardiol2013;167:e46–8.

30. Iacucci I, Carbone I, Cannavale G, et al.Myocardial edema as the sole marker of acuteinjury in Takotsubo cardiomyopathy: a cardiovas-cular magnetic resonance (CMR) study. RadiolMed 2013;118:1309–23.

31. Hassan SY. Myocarditis is an essential featurerather than an exclusion criterion for takotsubosyndrome: case report. Int J Cardiol 2015;187:304–6.

32. Fazio G, Pizzuto C, Barbaro G, et al. Chronicpharmacological treatment in takotsubo cardio-myopathy. Int J Cardiol 2008;127:121–3.

KEY WORDS histopathology,inflammation, macrophage, pathophysiology,takotsubo cardiomyopathy

APPENDIX For supplemental figures, pleasesee the online version of this paper.

http://refhub.elsevier.com/S2452-302X(18)30225-0/sref1




































































































































characterization of the myocardial inflammatory...

Documents