MOLECULAR AND CELLULAR MECHANISMS OF DISEASE

Adult cardiac fibroblast proliferation is modulatedby calcium/calmodulin-dependent protein kinase II in normaland hypertrophied hearts

Tamara P. Martin & Ahmed Lawan & Emma Robinson &

David J. Grieve & Robin Plevin & Andrew Paul &Susan Currie

Received: 29 March 2013 /Revised: 17 June 2013 /Accepted: 2 July 2013# Springer-Verlag Berlin Heidelberg 2013

Abstract Increased adult cardiac fibroblast proliferation re-sults in an increased collagen deposition responsible for thefibrosis accompanying pathological remodelling of the heart.The mechanisms regulating cardiac fibroblast proliferationremain poorly understood. Using a minimally invasive trans-verse aortic banding (MTAB) mouse model of cardiac hyper-trophy, we have assessed fibrosis and cardiac fibroblast prolif-eration. We have investigated whether calcium/calmodulin-dependent protein kinase IIδ (CaMKIIδ) regulates proliferationin fibroblasts isolated from normal and hypertrophied hearts. Itis known that CaMKIIδ plays a central role in cardiac myocytecontractility, but nothing is known of its role in adult cardiacfibroblast function. The MTAB model used here producesextensive hypertrophy and fibrosis. CaMKIIδ protein expres-sion and activity is upregulated in MTAB hearts and, specifi-cally, in cardiac fibroblasts isolated from hypertrophied hearts.In response to angiotensin II, cardiac fibroblasts isolated fromMTAB hearts show increased proliferation rates. Inhibition ofCaMKII with autocamtide inhibitory peptide inhibits prolifer-ation in cells isolated from both sham andMTAB hearts, with a

significantly greater effect evident in MTAB cells. These re-sults are the first to show selective upregulation of CaMKIIδ inadult cardiac fibroblasts following cardiac hypertrophy and toassign a previously unrecognised role to CaMKII in regulatingadult cardiac fibroblast function in normal and diseased hearts.

Keywords Calcium/calmodulin-dependent protein kinaseII . Cardiac hypertrophy . Aortic banding . Cardiacfibroblasts . Proliferation . Fibrosis

AbbreviationsAIP Autocamtide inhibitory peptideBW Body weightCaMKII Calcium/calmodulin-dependent protein kinase IIFS Fractional shorteningGAPDH Glyceraldehyde 3-phosphate dehydrogenaseHW Heart weightLVEDD Left ventricular end-diastolic dimensionLVESD Left ventricular end-systolic dimensionLVEDP Left ventricular end-diastolic pressureMTAB Minimally invasive transverse aortic banding

Introduction

Myocardial hypertrophy is an essential adaptive response bywhich the heart responds to various mechanical, electricaland chemical stresses. However, sustained overloading ofthe heart promotes adverse remodelling of the myocardium,resulting in contractile dysfunction and eventually heart fail-ure [11]. Structural remodelling of the heart is a dynamicprocess involving both contractile (myocytes) and non-contractile cells (fibroblasts, endothelial cells). Cardiac fi-broblasts are the most numerous non-myocyte cell typewithin the heart and are the predominant secretory cells

T. P. MartinInstitute of Cardiovascular and Medical Sciences,University of Glasgow, Glasgow G12 8QQ, UK

A. LawanDepartment of Pharmacology, School of Medicine, YaleUniversity, 333 Cedar Street, New Haven, CT 06520, USA

E. Robinson :D. J. GrieveCentre for Vision and Vascular Science, Queen’s UniversityBelfast, Grosvenor Road, Belfast BT12 6BA, UK

R. Plevin :A. Paul : S. Currie (*)Strathclyde Institute of Pharmacy & Biomedical Sciences,University of Strathclyde, Hamnett Building,Glasgow G4 0RE, UKe-mail: susan.currie@strath.ac.uk

Pflugers Arch - Eur J PhysiolDOI 10.1007/s00424-013-1326-9

within the extracellular matrix, releasing various growthfactors and pro-inflammatory cytokines [3, 21]. Paracrinesignalling between myocytes and fibroblasts, coupled withdirect cell–cell communication, has been reported as essen-tial for efficient contraction of the myocardium [12, 16, 35].However, although these previous studies are critical to ourunderstanding of how these cells interact, this has yet to beconfirmed in situ. Extensive myocardial fibrosis duringremodelling, as a result of excess collagen deposition byhyper-proliferative fibroblasts, disrupts cell–cell interac-tions, leading to alterations in myocyte electrical activityand, hence, contractile dysfunction. Fibrosis is a key factorunderlying reduced compliance and diastolic dysfunctionwhich are important components of the early cardiacremodelling. Cardiac fibroblasts therefore play a key rolein both normal and diseased myocardium, not only regulat-ing structural aspects of cardiac growth but also indirectlyinfluencing contractile aspects via their effects on myocyteactivity. Although this important role is acknowledged, littleis understood of the underlying regulatory mechanisms thatcontrol cardiac fibroblast function.

Calcium/calmodulin-dependent protein kinase IIδ (CaMKIIδ)has been identified as a pivotal regulatory protein in the heart,which is important for normal cardiac function as well as mod-ulating abnormal responses in the diseased myocardium [1, 7,20]. Evidence suggests that CaMKIIδ plays an important role inpathological cardiac remodelling, as expression and activity areupregulated in structural heart disease in patients with end-stageheart failure and animal models with pressure overload-inducedhypertrophy [2, 4, 5, 14, 18, 37]. Transgenic overexpression ofeither the nuclear splice variant CaMKIIδB [38] or the cytosolicsplice variant CaMKIIδC [39] promotes cardiac hypertrophy, andselective inhibition of CaMKII activity by pharmacological ortransgenic approaches protects against cardiac remodelling andmaintains myocardial function [13, 36]. This enzyme has, there-fore, been highlighted as an important target for therapeuticintervention [8]. However, it is important to note that the focusfor the previous research has predominantly been on the role ofCaMKII in cardiac myocyte function. No previous studies haveexamined the role that CaMKII may play in modulating adultcardiac fibroblast activity. This is an important aspect to consider,given the role of fibroblasts in the development of fibrosis duringmyocardial disease.

Animal models of cardiac disease have proved invaluablefor assessing cardiac function in vivo and for relating dys-function to altered cellular parameters. Transverse aorticconstriction (TAC), pioneered in mouse [25], is used rou-tinely to assess effects of left ventricular (LV) hypertrophy.Recent refinement of TAC, using a minimally invasive sur-gical approach (the minimally invasive transverse aorticbanding (MTAB) model), avoids entering the pleural space,allowing rapid banding of the aortic arch. The MTAB pro-cedure results in a similar rate and extent of hypertrophy and

fibrosis as that observed using more traditional methods andcan be used as a more convenient approach to study progres-sion of hypertrophic remodelling [22].

Here, we have used the MTAB procedure to induce pres-sure overload hypertrophy in mice. For the first time, wehave studied CaMKIIδ in adult cardiac fibroblasts isolatedfrom normal and hypertrophied hearts. Previous work by ourgroup has demonstrated increased fibrosis in hearts fromMTAB mice [22]. This is confirmed in the present studyand corresponds with an increased proliferative capacity ofcultured cardiac fibroblasts isolated from MTAB hearts.Novel findings have shown that following MTAB surgery,CaMKIIδ expression is increased specifically at the level ofthe cardiac fibroblast as well as in whole cardiac homoge-nates. For the first time, we show that the inhibition ofCaMKII reduces proliferation of fibroblasts isolated fromboth control and MTAB hearts with a more significant effectevident in the latter. These results highlight the potential thatmodulation of CaMKII activity may have on the develop-ment of fibrosis in diseased myocardium.

Methods

MTAB surgery

Healthy adult male C57BL/6J mice (Harlan, UK) weighingbetween 25 and 30 g were used for these experiments.Procedures conformed to the UK Animals (Scientific Proce-dures) Act 1986 and were approved by institutional ethicalreview committees. Surgical protocol was an adaptation of aMTAB technique described previously in detail [22]. Briefly,animals were anaesthetised with 3 % isofluorane andmaintained on 1.5 % isofluorane in 100 % oxygen. A smallhorizontal skin incision was made at the level of thesuprasternal notch, trachea was located and a longitudinalcut was made down the sternum. The thymus was retracted,and a 5-0 silk suture was passed under the aortic arch. Thesuture was tied against a 27-G needle between the origin ofthe right innominate and left common carotid arteries. Sham-operated animals went through the same procedure exceptthat the aortic arch was not tied. Animals were housed inheated cages for 24 h following surgery and were left for upto 4 weeks to allow cardiac remodelling to occur.

In vivo assessment of cardiac function

To acquire invasive haemodynamic measurements, micewere anaesthetised as previously described [22, 42] withisofluorane/oxygen, and the right carotid artery was cannu-lated with a high-fidelity 1.2-F pressure-volume catheter(SciSense, Inc.). Aortic pressure was measured, and thecatheter was advanced into the LV for recording of steady-

Pflugers Arch - Eur J Physiol

state function using the PowerLab System (ADI Instru-ments). Absolute volume measurements were corrected forα (derived by simultaneous aortic outflow recording byechocardiography) and parallel conductance (hypertonic sa-line injection) [42].

Transthoracic echocardiography

Echocardiographic assessment of left ventricular functionwas performed 4 weeks after MTAB or sham surgery. Ani-mals were anaesthetised as described previously [22]. Withthe use of the MIUS HDI 3000CVechocardiography systemand a 13 MHz linear array transducer, two-dimensionalshort-axis views and M-mode images were recorded at thelevel of the papillary muscle. LV end-systolic dimension(LVESD) and LV end-diastolic dimension (LVEDD) wereassessed from M-mode traces and per cent fractional short-ening calculated. Fractional shortening (FS) is expressed as[(LVEDD − LVESD)/LVEDD] × 100. An average of threemeasurements of each variable was used.

Tissue homogenisation, sectioning and collagenquantification

For post-mortem measurements 4 weeks after surgery, micewere weighed and euthanised with pentobarbital sodium(200 mg/kg Euthatal). Hearts were rapidly excised, washedin Ca2+-free Krebs, blotted dry and weighed. Whole cardiachomogenates were prepared as previously described [6].Hearts were fixed and processed as described previously[22]. Samples were embedded in paraffin and tissue sec-tioned at 5 μm every 200 μm. For collagen assessment,slides were stained for 1 h in Picro-Sirius Red solution aspreviously described [22]. Slides were mounted inHistomount (National Diagnostics), and five random sec-tions per heart (10 areas of interest per section) werephotographed at ×20 magnification using non-polarised lightwith the Leica DM LB2 microscope and the Leica DFC 320camera (Leica Microsystems, Germany). Quantification ofPicro-Sirius Red staining used the Image-Pro Plus Software(version 5.0; Media Cybernetics, USA), with stained areaexpressed as a percentage of the total area of interest. Valueswere averaged to give one representative value per heart.

Isolation of cardiac fibroblasts

Hearts were rapidly removed under sterile conditions in alaminar flow hood and washed in warmed Ca2+-free Krebssolution (120 mM NaCl, 5.4 mM KCl, 0.52 mM NaH2PO4,20 mM HEPES, 11.1 mM glucose, 3.5 mM MgCl2, 20 mMtaurine, 10 mM creatine, 1 mM EGTA, 1 % Pen/Strep(GIBCO); pH 7.4). Cardiac fibroblasts were isolated fromthe left and right ventricles under sterile conditions as

previously described [17] using an enzyme solution contain-ing 0.8 mg/ml collagenase type 2 (Worthington Chemical,UK) and 0.3 mg/ml protease XIV (Sigma-Aldrich). Cellswere resuspended in Dulbecco's modified Eagle's medium(DMEM; GIBCO 21969) supplemented with 20 % foetalcalf serum (heat-inactivated; GIBCO), 1 % L-glutamine(GIBCO) and 4 % Pen/Strep. Cells were allowed to adhereto culture dishes for ∼3–4 h before removal of mediumcontaining non-adherent cells (endothelial cells andmyocytes), fresh medium was added and fibroblasts weregrown for 5–7 days to 60–80 % confluence at 37 °C in anatmosphere of 5 % CO2. Passage 1 (P1) cells were used in allexperiments, except for quantification, where P0 cells wereused.

Immunofluorescence

Cardiac fibroblasts and colonic vascular smooth muscle cellswere grown on coverslips until ∼70 % confluent, fixed in 4 %paraformaldehyde and permeabilised with TritonX-100(0.01 %) for 10 min. Non-specific binding was blocked using1 % bovine serum albumin (BSA) in phosphate-bufferedsaline (PBS) for 1 h at room temperature followed by directaddition of primary antibody overnight at 4 °C. The anti-α-smooth muscle actin antibody and anti-vimentin antibody(both from Sigma-Aldrich) were prepared with 1 % BSA insterile PBS at 1:100 dilution. The TRITC anti-mouse IgG (GEHealthcare) and the Alexa Fluor 594 anti-rabbit IgG (Molec-ular Probes) were then applied at 1:200 and 1:400 dilution,respectively, for 1 h at room temperature. After washing,coverslips weremounted using the VECTASHIELD®Mount-ing Medium containing 4′,6-diamidino-2-phenylindole(DAPI) (Vecta laboratory) and stored at 4 °C in the dark untilthey were viewed and photographed. The DAPI counter-stainin the mounting medium stained the cells' nucleus blue. Pic-tures were taken using the Nikon Eclipse™ E600 oil Immer-sion microscope connected to the Photometrics (CoolSnap™FX) digital camera managed by the MetaMorph™ Software(Universal Imaging Corporation, West Chester, PA).

Immunoblotting

Samples were subjected to electrophoresis using the InvitrogenNuPAGE® System as previously described [22]. Proteins weretransferred to nitrocellulose, blocked in 5 % non-fat drymilk/TBST buffer (20 mM Tris base, 137 mM NaCl and0.1 % (v/v) Tween 20, pH 7.6) for 1 h at room temperatureand then incubated overnight at 4 °C with antibodies againsteither CaMKIIδ (rabbit polyclonal antibody 1:1,000, custom-made against the C-terminus of CaMKIIδ; Eurogentec),phospho-Thr286-CaMKIIδ (mouse monoclonal, 1:500; Ther-mo Scientific) or glyceraldehyde 3-phosphate dehydrogenase(GAPDH) (mouse monoclonal, 1:20,000; Thermo Scientific).

Pflugers Arch - Eur J Physiol

Blots were cut horizontally to allow probing of both CaMKIIδand GAPDH from the same blot. This was followed by incu-bation with either anti-rabbit or anti-mouse IgG-horseradishperoxidase conjugate diluted 1:10,000 and 1:5,000 (Sigma-Aldrich and Jackson Laboratory), respectively. Signals werequantified across a range of protein loads by densitometryusing the GS-800 Densitometer and Quantity One ImageSoftware (version 4.5.2, Bio-Rad). For CaMKIIδ, where therewere at least two or three variants expressed, depending on thepreparation, all signals were quantified.

Proliferation assay

Cells were plated onto sterile coverslips in 12-well cultureplates at 5,000 cells per well in DMEM (as described above)and allowed to attach and grow for 2 days before serumstarvation in DMEM (with no foetal calf serum added) for24 h. Cells were then stimulated with 1 μM angiotensin II(Ang II) (in serum-free DMEM), and proliferation wasassessed over 0–72 h. For experiments using inhibitor, cellswere pre-incubated with the CaMKII inhibitor autocamtide 2-related inhibitory peptide (1 μM) for 1 h prior to stimulation.Following stimulation, cells were washed in PBS, fixed with70 % methanol for 30 min and stained with haematoxylin(Surgipath Gill II) for 20 min. Coverslips were mounted ontomicroscope slides using Mowiol (Sigma). Photographs of thecells were taken (10 images per point) using the Leica DMLB2 microscope (Leica Microsystems, Germany) at 5× ob-jective connected to the Leica DFC 320 digital camera (LeicaMicrosystems, Germany), and images were prepared usingthe Adobe Photoshop.

Statistical analysis

Where appropriate, data were expressed as mean values ±standard error of the mean (SEM). Comparisons wereassessed using the student's t test analysed within theGraphPad Prism® (version 5.0) or two-way ANOVA withBonferroni post-test. Differences were considered statistical-ly significant when p<0.05. Where no p value is shown,p>0.05.

Results

MTAB induces cardiac hypertrophy and myocardial fibrosis

A significant increase in heart size was observed 4weeks post-banding as displayed in cross-sections from MTAB hearttissue (Fig. 1a(i)). This occurred in the presence of a sustainedincrease in systolic blood pressure (SBP) assessed invasively1 week post-surgery (77.91±3.41 vs. 112.18±6.77 mmHg,sham vs. MTAB, p=0.0011, n=6) (Fig. 1a(ii)). Heart

weight/body weight (HW/BW) ratio was significantly in-creased 4 weeks following MTAB surgery, indicating hyper-trophic growth of the heart (4.6±0.04 vs. 9.6±0.6 (HW/BW),sham vs. MTAB, respectively, p>0.0001, n=8) (Fig. 1a(iii)).Importantly, body weights were not significantly differentbetween groups. Systolic dysfunction was observed after1 week as a decrease in invasively assessed LVdP/dtmax (indexof LV contractility) and after 4 weeks as a decrease in echo-cardiographic per cent FS (44±2.1 vs. 22.6±2.8, sham vs.MTAB, respectively, p<0.0001, n=8) (Fig. 1b(i)). Diastolicdysfunction was also apparent at 1 week following MTABsurgery (Fig. 1c) as evidenced by compromised invasivelyassessed LV relaxation indicated by increased LVEDP, alteredLVdP/dtmin (index of LV relaxation) and increased tau (relax-ation time constant). This was confirmed at 4 weeks by adecrease in mitral valve (MV) E/A ratio (the ratio betweenearly (E) and late atrial (A) ventricular filling velocity)assessed by echocardiography (Fig. 1b(ii)). It is important tonote that diastolic dysfunction and compromised filling of theheart relate to a stiffer and less compliant myocardium, whichis particularly characteristic of a fibrotic heart. Indeed, assess-ment of collagen deposition in the current study using Picro-Sirius Red staining of ventricular sections confirmed thisobservation. Optical density analysis using Image-Pro PlusSoftware established that interstitial collagen deposition wasmarkedly increased in MTAB hearts (0.41±0.06 vs. 8.03±2.3(per cent stained area) sham vs. MTAB, respectively,p=0.038, n=4) (Fig. 1d). In an attempt to relate collagenproduction in vivo at the tissue level with that produced byisolated cardiac fibroblasts in vitro, additional experimentswere performed using immunocytochemistry. Using antibod-ies against collagen types I and III (Abcam), cardiac fibro-blasts from sham and MTAB hearts were probed following∼10 days in culture. However, it proved too difficult to mea-sure specific collagen signals since the overall backgroundsignal was very high. For this reason, quantification was notpossible, and these results have not been included in thepresent study.

CaMKIIδ expression and activity is increasedin hypertrophied hearts

Whole ventricular homogenates, obtained 4 weeks post-surgery from both sham-operated and MTAB hearts, wereanalysed for expression of CaMKIIδ using a custom-madeantibody against the C-terminus of the protein. Quantitativeimmunoblotting using densitometry analysis across a rangeof protein loads demonstrated that CaMKIIδ expression(relative to GAPDH) was increased in MTAB compared tosham animals (0.41±0.05 vs. 0.27±0.03, p=0.038, n=8)(Fig. 2a(i)). In addition, assessment of the activation stateof CaMKII in whole homogenates was monitored by quan-tifying autophosphorylation at Thr286 using a specific

Pflugers Arch - Eur J Physiol

phospho-CaMKII antibody (Thermo Scientific). Quantita-tive immunoblotting, again performed across a range ofprotein loads, revealed that the phosphorylation status ofCaMKIIδ was significantly increased in MTAB hearts(0.38±0.06 vs. 0.24±0.03, MTAB vs. sham, respectively,p=0.04, n=8) (Fig. 2b(i)), thus providing evidence that theactivity of CaMKIIδ was increased in hypertrophied hearts.Scatter plots show individual samples plotted from sham andMTAB groups. Data has been plotted for both CaMKIIδ andphospho-CaMKII protein expression versus either HW/BWratio or per cent FS, and both experimental groups were com-pared on the same scatter plot. MTAB data show a clearcorrespondence of an increased expression and phosphorylation

status of CaMKII with increased HW/BWratios and decreasedper cent FS (Fig. 2a(ii), b(ii)).

Cardiac fibroblasts express CaMKIIδ with increasedexpression evident in cells isolated from MTAB hearts

Cardiac fibroblasts were isolated under sterile conditionsusing a chunk dissociation method as described in the‘Methods’ section. Isolation via Langendorff perfusion al-ways resulted in cells that became infected in culture due to alack of sterility during the isolation procedure; hence, the useof this protocol was avoided. A typical image of fibroblastsafter one passage is shown (Fig. 3a). Lack of contamination

A

B

(i) (iii)(ii)

C

(ii)(i)

Sham MTAB

5X5X

Sham (n=6) MTAB (n=6)

Heart rate (bpm) 560 ± 21 562 ± 32

LVEDP (mmHg) 5.89 ± 0.52 12.62 ± 1.18 **

LVdP/dtmin (mmHg s- 1) -7324 ± 474 -5421 ± 386*

LVdP/dtmax(mmHg s- 1) 8442 ± 476 6280 ± 556*

(ms) 5.98 ± 0.36 8.10 ± 0.69*

0

25

50

75

100

125

150

MTABSham

SB

P(m

mH

g)

*

0

2

4

6

8

10

12

MTABSham*

Fib

ros

isS

co

re(o

ve

rall

co

llag

en

co

nte

ntin

%)

D

Sham MTAB

Fig. 1 Evidence for cardiac hypertrophy, contractile dysfunction andfibrosis following MTAB surgery. a, i Representative tissue cross-sections from 4-week sham and MTAB hearts (×5 magnification). iiHistogram showing mean SBP after 1 week, n=6. iii Mean values forheart weight/body weight ratios after 4 weeks, n=8. b, i Histogramshowing fractional shortening (per cent) after 4 weeks, n=8. ii Histo-gram showing MV E/A ratio after 4 weeks, n=6. c Table showing heart

rate data and data measuring contractile function of the heart after1 week, n=6. d Representative images of Picro-Sirius Red stained heartsections prepared from sham and MTAB animals and viewed usingnon-polarised light. Histogram representing mean collagen deposition,quantified as described in the ‘Methods’ section, in sham and MTABhearts 4 weeks after MTAB surgery, n=4. All data are expressed asmean values ± SEM. *p<0.05; **p<0.001; ***p<0.0001

Pflugers Arch - Eur J Physiol

of fibroblast preparations by smooth muscle cells ormyofibroblasts was demonstrated by immunostaining withα-smooth muscle actin (Fig. 3b), and vimentin was used as apositive marker for fibroblasts (Fig. 3c). Cells were kept inculture for up to 2 weeks with only one passage (subsequentpassages show phenotypic changes and myofibroblast ap-pearance, data not shown). Importantly, for quantificationanalysis, P0 cells were used which had only been in culturefor 5–7 days and had not been subjected to trypsinisation. As

adult cardiac fibroblasts have not been used previously toexamine CaMKIIδ, it was important to verify first that thesecells express the protein. Immunoblot analysis using thesame custom-made anti-CaMKIIδ antibody as previouslyused for ventricular homogenate preparations, demonstratedthat CaMKIIδ was highly expressed in murine adult cardiacfibroblasts and that there may be more than one splice variant(denoted by several bands) (Fig. 3d(i)). Interestingly, therewas a significant increase in CaMKIIδ protein expression

A(i) (ii)

B(i) (ii)

CaMKIIδ

GAPDH

ShamMTAB

GAPDH

Phospho-CaMKIIδ

Sham MTAB

60 kDa

50kDa

40 kDa

60 kDa

50kDa

40 kDa

0 5 10 150.0

0.2

0.4

0.6

0.8

1.0

1.2ShamMTAB

HW/BW Ratio

OD

(nor

mal

i se d

toG

AP

DH

)

0 10 20 30 40 500.0

0.2

0.4

0.6

0.8

1.0

1.2ShamMTAB

% Fractionalshortening

OD

(nor

mal

ise d

toG

AP

DH

,AU

)

0 5 10 150.0

0.2

0.4

0.6

0.8ShamMTAB

HW/BW Ratio

OD

(nor

mal

ised

toG

AP

DH

,AU

)

0 10 20 30 40 500.0

0.2

0.4

0.6

0.8ShamMTAB

% Fractionalshortening

OD

(nor

mal

ised

toG

AP

DH

,AU

)

Fig. 2 CaMKII expression and activity is increased in hearts followingMTAB surgery; a, i Immunoblot analysis of total CaMKIIδ expressionin whole heart homogenates from sham and MTAB obtained 4 weeksafter MTAB. A representative immunoblot (cut to allow probing forboth CaMKIIδ and GAPDH) is shown with increasing loads (8, 10 and12 μg total protein) for each sample. Densitometry analysis wasperformed on individual immunoblots, data were normalised toGAPDH and mean normalised data were shown in the accompanyinghistogram, n=8, *p<0.05. ii Correlation plots are shown for CaMKIIδexpression relative to both HW/BW ratios and per cent FS for sham and

MTAB. b, i Immunoblot analysis of phospho-Thr286-CaMKII expres-sion with a representative immunoblot (cut to allow probing for bothCaMKIIδ and GAPDH) showing increasing total protein loads (8, 10and 12 μg) for each preparation. Densitometry analysis was performedon individual blots, data were normalised to GAPDH and meannormalised data were shown in the accompanying histogram, n=8,*p<0.05. Data are expressed as mean ± SEM. ii Correlation plots areshown for phospho-Thr286-CaMKII expression relative to both HW/BW ratios and per cent FS for sham and MTAB

Pflugers Arch - Eur J Physiol

(relative to GAPDH) between fibroblasts isolated from shamandMTAB hearts (0.95±0.03 vs. 0.81±0.02,MTAB vs. sham,p<0.01, n=6) (Fig. 3d(ii), (a, b)). This increased expression incardiac fibroblasts mirrors the increase that is observed inwhole cardiac ventricular homogenates (Fig. 3d(iii)), previous-ly quantified in Fig. 2.

Proliferation of cardiac fibroblasts is increasedfollowing cardiac hypertrophy and is differentially regulatedby CaMKII in cells from MTAB versus sham animals

To investigate whether cardiac fibroblast proliferation rateswere altered following hypertrophy, cells isolated from bothsham and MTAB hearts were quiesced for 24 h prior tostimulation with Ang II (1 μM) for 0–72 h. Representativeimages of fibroblasts stained with haematoxylin are shown ateach time point (Fig. 4a). Results indicate that in response toAng II, total cardiac fibroblast number increased over 72 hstimulation with a significant difference between control and

MTAB cells evident after 48 h and maintained at 72 h(59.6±2.5 vs. 41.8±2.3, MTAB vs. sham, respectively,p=0.006) (Fig. 4b). When analysed using a repeated mea-sures, two-factor ANOVA as MTAB vs. sham/time,p<0.01 at both 48 h and 72 h. To determine whether CaMKIIis involved in cardiac fibroblast proliferation induced by AngII, quiesced cells were pretreated with the CaMKII inhibitor,autocamtide inhibitory peptide (AIP) (1 μM), for 1 h follow-ed by stimulation with Ang II (1 μM) for 0–72 h as describedabove. Ang II-induced fibroblast proliferation was markedlyreduced following pretreatment with AIP in cells isolatedfrom either sham-operated or MTAB hearts (Fig. 5a, b).Proliferation was significantly decreased at 24 h, and thisdecrease was maintained at 48 and 72 h in both sham (72 h;50.9±3.3 (untreated) vs. 29.8±0.6 (AIP treated), respective-ly) andMTAB cells (72 h; 71.0±7.2 (untreated) vs. 37.5±1.5(AIP treated), respectively). Importantly, when a repeatedmeasures, two-factor ANOVA was performed comparingboth sets of data (i.e. (MTAB+AIP) vs. (sham+AIP)/time,

A B

CCardiac fibroblasts

D (i).

Cardiac fibroblasts Smooth muscle cells

CaMKIIδ

GAPDH

Sham MTAB (WH)

(ii) (a) (iii)

0.0

0.5

1.0

1.5

MTABSham

*O

ptic

alD

en

si ty

(no

rma

lise

dto

GA

PD

H;A

U)

60 kDa

50kDa

CaMKIIδ

60 kDa

50kDa

WH CF

Sham MTAB (CF)

40kDa

CaMKIIδ

GAPDH

(b)

Fig. 3 Cardiac fibroblast preparations from MTAB hearts show in-creased expression of CaMKIIδ. a Image of a typical cardiac fibroblastpopulation (passage 1) (20× objective). b Cardiac fibroblasts werestained for α-smooth muscle actin (40× objective) after one passageof cells in culture. Colonic smooth muscle cells (isolated and cultured inparallel) served as a positive control for the presence of this protein. cCardiac fibroblasts stained positively for vimentin (40× objective). d, iImmunoblot analysis of CaMKIIδ expression in cardiac fibroblasts(CF) (∼104 cells). Whole cardiac homogenate (WH) (5 μg) served as

a positive control for CaMKIIδ expression. ii Equivalent loads (104

cardiac fibroblasts) from either sham or MTAB hearts were compared,and blots were cut to probe for both CaMKIIδ and GAPDH andquantified using GAPDH as an internal control. The accompanyinghistogram shows mean data ± SEM, n=6, *p<0.05. iii Equivalent loadsof WH (5 μg total protein) from either sham or MTAB hearts werecompared, and blots were cut to probe for both CaMKIIδ and GAPDH,n=3

Pflugers Arch - Eur J Physiol

p<0.05 at 48 h and p<0.01 at 72 h, showing that AIP had asignificantly greater effect upon proliferation in MTAB cellsthan sham cells (Fig. 5c). AIP pretreatment had no obviouseffect on cell number at time zero (AIP-treated vs. untreatedcells, sham p=0.270 and MTAB p=0.105). Taken together,these findings demonstrate for the first time that (1) adultcardiac fibroblast proliferation is increased during cardiachypertrophy, (2) CaMKII modulates fibroblast proliferationin both normal and hypertrophied hearts and (3) inhibition ofCaMKII has a significantly greater effect on reducing pro-liferation rates in fibroblasts isolated from MTAB heartscompared with the effect on fibroblasts isolated from shamhearts.

Discussion

This is the first study to show that CaMKII plays a regulatoryrole specific to adult cardiac fibroblast function. We haveprovided novel evidence to show that (1) CaMKIIδ isexpressed in adult cardiac fibroblast preparations (Fig. 3d(i)),(2) CaMKIIδ expression is significantly elevated in cardiacfibroblasts isolated from MTAB hearts (Fig. 3d(ii)), (3) in-hibition of CaMKII (using AIP) significantly reduces adultcardiac fibroblast proliferation in vitro and (4) the extent of

reduction in proliferation following AIP treatment is signif-icantly different between fibroblasts isolated from sham andMTAB hearts (Fig. 5c). These are important new findingssince increased cardiac fibroblast proliferation is directlylinked to increased collagen production and fibrosis in thediseased heart [9]. We have shown that fibroblast prolifera-tion is increased in cells that have been isolated from heartssubjected to aortic banding (Fig. 4). In the present study, wehave used Ang II to stimulate proliferation above basal ratesin culture. This has allowed a more rapid assessment ofproliferative function. It is worth noting that we do observeincreased proliferation over time in the absence of Ang II incells isolated from both sham and MTAB hearts as well ashigher basal rates of proliferation in cells isolated fromMTAB hearts; however, this is over a much longer timescale(7–10 days) (data not shown). The fact that both basal andstimulated responses are elevated in cells isolated fromMTABhearts clarifies for the first time that the hyperproliferativephenotype associated with the disease state is maintained inculture. Importantly, evidence for a less compliant, fibroticheart is apparent from MTAB data (Fig. 1c), and although wewere unable to assess collagen production quantitatively, spe-cifically in cardiac fibroblasts, collagen deposition is signifi-cantly increased inMTAB hearts (Fig. 1d). Importantly, for thefirst time, these data provide a link between hyperproliferative

A

B

Time (h)

MTAB

0 24 48 72

Sham

100 µM

Fig. 4 Proliferation of cardiacfibroblasts is increased in cellsisolated from MTAB hearts.Cardiac fibroblasts isolated fromcontrol and MTAB hearts andplated onto sterile coverslipswere stimulated with Ang II(1 μM) for 0–72 h beforestaining with haematoxylin andcell counting. a Representativeimages (20× objective) of cellsfrom sham and MTAB heartsstained with haematoxylin ateach time point. Data are fromone experiment, typical of threeothers, and each time point wasperformed in triplicate. b Meancell numbers for sham andMTAB preparations werecounted as described in the‘Methods’ section. Each datapoint represents mean ± SEM,n=3. **p<0.01; ***p<0.001

Pflugers Arch - Eur J Physiol

fibroblasts in vitro, leading to excess collagen depositionin vivo in the same model of hypertrophy.

It is now widely acknowledged that increased expressionand activity of CaMKII is associated with the hypertrophiedand failing heart [1, 2, 4, 5, 14, 18, 37–39], and it has beenproposed that this protein could be a useful target for thera-peutic intervention [8]. In the current study, we have providedfurther evidence that in line with previous studies using dif-ferent models, CaMKII protein expression and activity areboth upregulated in MTAB hearts (Fig. 2). These increasesin protein expression and activity show a strong correlationwith hypertrophic growth (HW/BW ratios) and compromisedcontractile function (decreased per cent FS) (Fig. 2). Impor-tantly, new results examining CaMKIIδ protein expression in

cardiac fibroblasts isolated from MTAB hearts show for thefirst time that there is upregulation of the protein specifically atthe level of these cells (Fig. 3d(ii)).

A clear connection between increased levels and activityof CaMKII and the altered cardiac calcium handling andcontractility that is a characteristic of diseased heart has beenestablished in the previous work [10, 26, 30], and this linksdirectly with cardiac myocyte function. However, until veryrecently, the possibility that CaMKII may modulate cardiacfibroblast function remained largely ignored. A recent studyusing neonatal cardiac fibroblasts suggested that CaMKIIcould play a role in proliferation [40]; however, this hasnever been examined in adult fibroblasts, nor has the poten-tial role of CaMKII in hyperproliferative fibroblasts from

A

B

Time (h)

0

72h

Sham Sham + AIP MTAB MTAB + AIP

0 24 48 720

10

20

30

40

50Sham + AIPMTAB + AIP

* **

Time (h)

% In

hib

itio

n w

ith

AIP

100 µM

Fig. 5 CaMKII inhibition reduces proliferation of cardiac fibroblastsisolated from both sham-operated and MTAB hearts. a Representativeimages of fibroblasts isolated from sham-operated and MTAB hearts, at0 and 72 h, stained with haematoxylin in the presence and absence ofAIP pretreatment. b Mean cell numbers taken over a period of 72 h of

stimulation from sham-operated and MTAB preparations and in thepresence and absence of AIP. cHistogram comparing per cent inhibitionof cell growth following AIP treatment between fibroblasts isolatedfrom sham and MTAB hearts. Each data point represents mean ± SEM,n=3. *p<0.05; **p<0.01; ***p<0.001

Pflugers Arch - Eur J Physiol

diseased hearts been addressed. Interestingly, other work innon-cardiac derived fibroblasts has highlighted a role forCaMKIIδ in proliferation. In human fibroblasts derived andcultured from skin biopsies, CaMKII activation was shownto be central to insulin-mediated proliferation and inhibition ofCaMKII, resulted in inhibition of thymidine incorporation[23]. In a separate study, the role of CaMKII in balloonangioplasty-induced injury was examined in rat carotid arter-ies. In this model, there was evidence of significantly elevatedlevels of CaMKII along with increased vascular smooth mus-cle cell and fibroblast proliferation. Suppression of CaMKIIδusing adenoviral siRNA infusion into the lumen of the carotidartery inhibited fibroblast proliferation following injury [15].Evidence is therefore emerging that CaMKII can modulatefibroblast function. Given the multiplicity of targets forCaMKII in cardiac myocytes, it seems likely that a similarrange of protein targets exists in cardiac fibroblasts. The targetsresponsible for CaMKII-mediated effects on fibroblast prolif-eration have yet to be identified; however, it is likely that thisinvolves fundamental regulation of fibroblast signalling pro-teins and ion channels by CaMKII. A crucial novel finding inthe current study is the differential modulation of fibroblastproliferation by CaMKII between cells isolated from normaland hypertrophied hearts (Fig. 5). The findings suggest thatselective inhibition of elevated CaMKII inhibits the increasedproliferation associated with hypertrophic hearts to a greaterextent than proliferation of fibroblasts from normal hearts. Thishas implications for treating the fibrotic phenotype where agreater effect would be needed to reverse the hyperproliferativephenotype.

The potential for cardiac fibroblasts to modulate cardiacmyocyte function and, therefore, to regulate indirectly car-diac electrophysiology and ultimately contractility shouldnot be understated. A growing body of evidence has demon-strated the capacity for intercellular coupling betweenmyocytes and fibroblasts, and the possibility of electricalcoupling has been highlighted [32]. Importantly, recent workhas suggested that this functional coupling is altered in cellsthat have been isolated from infarcted hearts with reducedconduction velocities observed in cultures where fibroblastswere derived from infarcted hearts [31]. In isolated cellpreparations, the absolute number of fibroblasts has a stronginfluence on cardiac myocyte conduction with increasednumbers, such as that seen in hyperproliferative conditions,creating an obstacle and reducing conduction velocity [24,34]. In addition to physical changes in extracellular matrixproduction, increased numbers of fibroblasts will result inincreased growth factor secretion which will ultimately in-fluence hypertrophic growth. For these reasons, the conceptof therapeutic targeting directed at the level of the cardiacfibroblast is gaining appeal. This concept was strengthenedrecently with evidence showing that cardiac fibroblasts areessential for the adaptive response of the heart to stress

following pressure overload. Findings from a mouse modelof pressure overload-induced hypertrophy showed that Klf5deletion in fibroblasts ameliorated cardiac hypertrophy andfibrosis when the same deletion in myocytes did not. Thisprovided the first evidence for a key role for fibroblasts asmediators of hypertrophy [29]. This has been followed withmore recent studies suggesting that therapy aimed atinhibiting collagen synthesis by fibroblasts in hypertrophiedhearts leads to attenuation of cardiac remodelling and sys-tolic dysfunction [19, 41].

It is important to recognise that the phenotype of fibroblastspresent in diseased hearts can change, and there is evidencefor myofibroblast presence in injured hearts [28]. It has alsobeen suggested that valve fibroblasts may originate fromhaematopoietic stem cells [33], and the suggestion has beenmade that different progenitor populations could serve impor-tant roles in therapeutic targeting [27]. In the present study, wewere unable to detect any obvious visible phenotypic changesbetween sham and MTAB cells isolated from whole mouseventricle. However, specific regional changes in fibroblastsub-populations throughout the ventricle were not tested,and this would be an important consideration for future work.

Findings presented here showing that CaMKII can mod-ulate adult cardiac fibroblast proliferation open up a wholenew aspect to our understanding of the role of CaMKII inmodulating cardiac function. There has been a somewhatblinkered view in our appreciation of the contribution thatnon-contractile cells may have in regulating cardiac contrac-tility through their proliferative, migratory and secretoryactions. The recognition that cardiac fibroblasts are an im-portant component of contractile function and that CaMKIIexpressed in fibroblasts modulates the proliferative responsewill have a pivotal role in changing this view. Future workbuilding upon the current data will provide a crucial mech-anistic understanding of how CaMKII signalling at the levelof the cardiac fibroblast impacts on the remodelling processaccompanying cardiac hypertrophy.

Acknowledgments We would like to thank Margaret MacDonald forher expert technical assistance with mouse surgery and DavidBlatchford for his expertise in cell imaging. This work was supportedby the British Heart Foundation (grant no: FS/06/066/21409).

Conflict of interest The authors declare they have no conflict ofinterest.

References

1. Anderson ME, Heller Brown J, Bers DM (2011) CaMKII in myo-cardial hypertrophy and heart failure. J Mol Cell Cardiol 51:468–473

2. Backs J, Backs T, Neef S, Kreussner MM, Lehmann LH, PatrickDM et al (2009) The delta isoform of CaM kinase II is required for

Pflugers Arch - Eur J Physiol

pathological cardiac hypertrophy and remodeling after pressureoverload. Proc Natl Acad Sci USA 106:2342–2347

3. Baudino T, Carver W, Giles W, Borg TK (2006) Cardiac fibro-blasts: friend or foe? Am J Physiol Heart Circ Physiol 291:H1015–H1026

4. Colomer JM, Mao L, Rockman HA, Means AR (2003) Pressureoverload selectively up-regulates Ca2+/calmodulin-dependent pro-tein kinase II in vivo. Mol Endocrinol 17:183–192

5. Currie S, Smith GL (1999) Calcium/calmodulin-dependent proteinkinase II activity is increased in sarcoplasmic reticulum from cor-onary artery ligated rabbit hearts. FEBS Lett 459:244–248

6. Currie S, Quinn FR, Sayeed RA, Duncan AM, Kettlewell S, SmithGL (2005) Selective down-regulation of sub-endocardial ryanodinereceptor expression in a rabbit model of left ventricular dysfunc-tion. J Mol Cell Cardiol 39:309–317

7. Currie S (2009) Ryanodine receptor phosphorylation by CaMKII:getting the balance right. Front Biosc 14:5134–5156

8. Currie S, Elliott EB, Smith GL, Loughrey CM (2011) Two candi-dates at the heart of dysfunction: the ryanodine receptor andCaMKII as potential targets for therapeutic intervention—anin vivo perspective. Pharm Ther 131:204–220

9. Diez J, Lopez B, Gonzalez A, Querejeta R (2007) The role of themyocardial collagen network in hypertensive heart disease. CurrHyperten Rev 3:1–7

10. Dybkova N, Sedej S, Napolitano C, Neef S, Rokita AG, Hunlich Met al (2011) Overexpression of CaMKIIδC in RyR2R4496C+/−knock-in mice leads to altered intracellular Ca2+ handling andincreased mortality. J Am Coll Cardiol 57:469–479

11. Frey N, Olson EN (2003) Cardiac hypertrophy: the good, the bad,and the ugly. Ann Rev Physiol 65:45–79

12. Gaudesius G, Miragoli M, Thomas SP, Rohr S (2003) Coupling ofcardiac electrical activity over extended distances by fibroblasts ofcardiac origin. Circ Res 93:421–428

13. Hempel P, Hoch B, Bartel S, Karczewski P (2002) Hypertrophicphenotype of cardiac calcium/calmodulin-dependent protein kinaseII is reversed by angiotensin converting enzyme inhibition. BasicRes Cardiol 97:96–101

14. Hoch B, Meyer R, Hetzer R, Krause EG, Karczewski P (1999)Identification and expression of delta-isoforms of the multifunc-tional Ca2+/calmodulin-dependent protein kinase in failing andnonfailing human myocardium. Circ Res 84:713–721

15. House S, Singer HA (2008) CaMKIIδ isoform regulation ofneointima formation after vascular injury. Arterioscler ThrombVasc Biol 28:441–447

16. LaFramboise WA, Scalise D, Stoodley P, Graner SR, GuthrieRD, Magovern JA et al (2007) Cardiac fibroblasts influencecardiomyocyte phenotype in vitro. Am J Physiol Cell Physiol292:C1799–C1808

17. Lawan A, Al-Harthi S, Cadalbert L, McCluskey AG, ShweashM, Grassia G et al (2011) Deletion of the DUSP-4 gene revealsan essential non-redundant role for MAP kinase phosphatase-2(MKP-2) in proliferation and cell survival. J Biol Chem286:12933–12943

18. Ling H, Zhang T, Pereira L, Means CK, Cheng H, Gu Yet al (2009)Requirement for Ca2+/calmodulin-dependent kinase II in the tran-sition from pressure overload-induced cardiac hypertrophy to heartfailure in mice. J Clin Invest 119:1230–1240

19. Ma Y, Chen Y, Yang Y, Chen B, Liu D, Xiong Z et al (2013)Proteasome inhibition attenuates heart failure during the late stagesof pressure overload through alterations in collagen expression.Biochem Pharmacol 85:223–233

20. Maier L, Bers DM (2002) Calcium, calmodulin and CaMKII:heartbeat to heartbeat and beyond. J Mol Cell Cardiol 34:919–939

21. Manabe I, Shindo T, Nagai (2002) Gene expression in fibroblastsand fibrosis: involvement in cardiac hypertrophy. Circ Res91:1103–1113

22. Martin TP, Robinson E, Harvey A, Grieve DJ, MacDonald M, PaulA, Currie S (2012) Characterisation and optimisation of a minimal-ly invasive aortic banding procedure to induce cardiac hypertrophy.Exptl Physiol 97:822–832

23. Monaco S, Illario M, Rusciano M, Gragnaniello G, Di Spigna G,Leggiero E et al (2009) Insulin stimulates fibroblast proliferationthrough calcium-calmodulin-dependent kinase II. Cell Cycle8:2024–2030

24. Nguyen TP, Xie Y, Garfinkel A (2010) Fibroblast-myocyte cou-pling promotes cardiac arrhythmias. Hear Rhythm 7:S348

25. Rockman HA, Ross RS, Harris AN, Knowlton KU, SteinhelperME, Field LJ et al (1991) Segregation of atrial-specific and induc-ible expression of an atrial natriuretic factor transgene in an in vivomurine model of cardiac hypertrophy. Proc Natl Acad Sci USA88:8277–8281

26. Sag CM, Wadsack DP, Khabbazzadeh S, Abesser D, Grefe C,Neumann K et al (2009) Calcium/calmodulin dependent proteinkinase II contributes to cardiac arrhythmogenesis in heart failure.Circ Heart Fail 2:664–675

27. Souders CA, Bowers SLK, Baudino TA (2009) Cardiac fibroblast:the renaissance cell. Circ Res 105:1164–1176

28. Sun Y, Weber KT (2000) Infarct scar: a dynamic tissue. CardiovascRes 46:250–256

29. Takeda N, Manabe I, Uchino Y, Eguchi K, Matsumoto S et al(2010) cardiac fibroblasts are essential for the adaptive responseof the murine heart to pressure overload. J Clin Invest 120:254–265

30. van Oort RJ, McCauley MD, Dixit SS, Pereira L, Yang Y, RespressJL et al (2010) Ryanodine receptor phosphorylation by CaMKIIpromotes life-threatening ventricular arrhythmias in mice withheart failure. Circulation 122:2669–2679

31. Vasquez C, Mohandas P, Louie KL (2010) Enhanced fibroblast–myocyte interactions in response to cardiac injury. Circ Res107:1011–1020

32. Vasquez C, Benamer N, Morley GE (2011) The cardiac fibroblast:functional and electrophysiological considerations in healthy anddiseased hearts. J Cardiovasc Pharmacol 57:380–388

33. Visconti RP, Ebihara Y, Larue AC, Fleming PA, McQuinn TC,Mauya M et al (2006) An in vivo analysis of hematopoietic stemcell potential:hematopoietic origin of cardiac valve interstitial cells.Circ Res 98:690–696

34. Xie Y, Garfinkel A, Weiss JN (2009) Cardiac alternans induced byfibroblast-myocyte coupling: mechanistic insights from computa-tional models. Am J Physiol 297:H775–H784

35. Zhang P, Su J, Mende U (2012) Cross talk between cardiacmyocytes and fibroblasts: from multiscale investigative approachesto mechanisms and functional consequences. Am J Physiol HeartCirc Physiol 303:H1385–H1396

36. Zhang R, Khoo MS, Wu Y, Yang Y, Grueter LE, Ni G et al (2005)Calmodulin kinase II inhibition protects against structural heartdisease. Nat Med 11:409

37. Zhang T, Heller Brown J (2004) Role of calcium/calmodulin-de-pendent protein kinase II in cardiac hypertrophy and heart failure.Cardiovasc Res 63:476–486

38. Zhang T, Johnson EN, Gu Y, Morisette MR, Sah VP, Gigena MSet al (2002) The cardiac-specific nuclear δB isoform of Ca2+/calmodulin-dependent protein kinase II induces hypertrophy anddilated cardiomyopathy associated with increased protein phospha-tase 2A activity. J Biol Chem 277:1261–1267

39. Zhang T, Maier LS, Dalton ND, Miyamoto S, Ross J, Bers DM(2003) The deltaC isoform of CaMKII is activated in cardiachypertrophy and induces dilated cardiomyopathy and heart failure.Circ Res 92:912–919

40. Zhang W, Chen D, Feng Q, Wang J, Xiao W, Zhu W (2010)Inhibition of CaMKII suppresses cardiac fibroblast proliferationand extracellular matrix secretion. J Cardiovasc Pharmacol55:96–105

Pflugers Arch - Eur J Physiol

41. Zhang Y, Elsik M, Edgley AJ, Cox AJ, Kompa AR, Wang B et al(2012) A new anti-fibrotic drug attenuates cardiac remodelling andsystolic dysfunction following experimental myocardial infarction.Int J Cardiol. doi:10.1016/j.ijcard.2012.11.067

42. Zhao Y, McLaughlin D, Robinson E, Harvey AP, Hookham MB,Shah AM et al (2010) Nox2 NADPH oxidase promotes pathologic

cardiac remodelling associated with doxorubicin chemotherapy.Cancer Res 70:9287–9297

The experiments comply with the current laws of the country in whichthey were performed.

Pflugers Arch - Eur J Physiol