Regional Workload Induced Changes in Electrophysiology and Immediate Early Gene Expression in IntactIn SituPorcine Heart

Regional Workload Induced Changes in Electrophysiology and Immediate Early Gene Expression in IntactIn SituPorcine Heart

J Mol Cell Cardiol 29, 3147–3155 (1997) Regional Workload Induced Changes in Electrophysiology and Immediate Early Gene Expression in Intact In Situ ...

355KB Sizes 0 Downloads 36 Views

J Mol Cell Cardiol 29, 3147–3155 (1997)

Regional Workload Induced Changes in Electrophysiology and Immediate Early Gene Expression in Intact In Situ Porcine Heart Parviz Meghji, Sirfraz A. Nazir, David J. Dick, Mark E. S. Bailey1, Keith J. Johnson1 and Max J. Lab Department of Physiology, Charing Cross and Westminster Medical School, London W6 8RF, UK; 1 (present address) Division of Molecular Genetics, University of Glasgow, Anderson College, 56 Dumbarton Rd, Glasgow G11 6NU, UK (Received 10 March 1997, accepted in revised form 29 July 1997) P. M, S. A. N, D. J. D, M. E. S. B, K. J. J  M. J. L. Regional Workload Induced Changes in Electrophysiology and Immediate Early Gene Expression in Intact In Situ Porcine Heart. Journal of Molecular and Cellular Cardiology (1997) 29, 3147–3155. Cardiac remodelling and hypertrophy induced by chronic haemodynamic overload (stretch) eventually leads to a decrease in cardiac function, an increased incidence of ventricular arrhythmia and mortality. The mechanisms by which myocytes sense haemodynamic stress and activate growth signals are largely unknown. Nuclear immediate early genes may act as third messengers, converting the stretch stimulus into long-term changes of gene expression via cytoplasmic signal transduction. However, previous studies have used cell cultures and isolated hearts, neither of which are ideal models. We have developed a new in situ porcine heart model where local strain (stretch) can be applied, for several hours if required, thus allowing the comparison of changes in electrophysiology and gene expression with unstrained myocardium in the same preparation. A pneumatically controlled stretch-device was attached to a portion of the right ventricle of an anaesthetized animal using suction. Chronic stretch was applied for 30 min or 1 h. Regional loading produced (i) a transient decrease in monophasic action potential duration (3.5±0.8%; P<0.05), followed by (ii) an elongation by 15 min, despite maintained stretch (3.4±1.5%; P<0.05 compared to the pre-stretch situation). A control segment of the right ventricle did not show these changes. Northern blot analysis showed that both c-fos and c-myc were induced in the areas sampled, but they were 12fold and three-fold higher, respectively, in stretched compared with control tissue after 30 min. Thus, prolonged regional stretch can produce complex changes in cardiac electrophysiology and increase expression of some immediate early genes. Our model may be useful for studying the cascade of events that lead to remodelling, hypertrophy, and arrhythmia.  1997 Academic Press Limited K W: Mechanical stretch; Stretch-activated channels; Porcine; Ventricle; Protooncogenes; Immediate early genes; Cardiac hypertrophy.

Introduction Chronic haemodynamic overload of the heart caused by increased volume or pressure, as for example in myocardial infarction, valve disease, or congestive failure, ultimately leads to left

ventricular remodelling and hypertrophy. Although the development of ventricular hypertrophy is a compensatory mechanism initially, it is also a related step in pathological ventricular remodelling, eventually leading to a decrease in cardiac function and an increased incidence of ventricular

Please address all correspondence to: Prof. Max J. Lab, National Heart & Lung Institute, Imperial College of Medicine, Charing Cross Hospital, Fulham Palace Road, London W6 8RF, UK.

0022–2828/97/113147+09 $25.00/0


 1997 Academic Press Limited


P. Meghji et al.

arrhythmia and mortality (James and Jones, 1989; Levy et al., 1990; Almendral et al., 1995). The mechanisms by which myocytes sense haemodynamic stress and activate growth signals remain largely unknown. An increase in mechanical load or stretch probably via some mechanotransduction process, including stretch activated channels, may be the mediator in the acute situation, and perhaps also after chronic myocardial hypertrophy. Myocardial stretch can also influence cardiac electrophysiology by an immediate process, now often termed ‘‘mechanoelectric feedback or transduction’’ (Lab, 1982, 1996). Stretch can produce rhythm disturbances in the short term (Lab, 1982; Hansen et al., 1990; Dick and Lab, 1997). In the long term, the heart wall hypertrophies; and this is also associated with production of rhythm disturbances. Signal transduction during mechanicallyinduced remodelling appears to be through stimulation of GTP binding protein Gq leading to a cascade of events including the hydrolysis of phosphatidylinositol by phospholipase C, production of diacylglycerol (DAG) and inositol-1,4,5(tri)phosphate (IP3), activation of protein kinase C (PKC) and mobilization of intracellular Ca2+ (see Sugden and Bogoyevitch, 1995 for review). Activation of PKC may be the common event in the signalling cascade which results in mediation of gene transcription through phosphorylation of pre-existing transcription factors and/or through induction of novel transcription factors. Proto-oncogenes or immediate early (IE) genes may be transcription factors and it has been proposed that they may act as ‘‘third messengers’’ of long-term cellular responses (Maniatis, 1991). Stretch of quiescent or beating neonatal myocytes grown on deformable silicon dishes stimulated the expression of IE genes such as c-fos, c-myc, c-jun and Erg-1, which peaked 30–60 min after stretch was applied (Sadoshima et al., 1992a, 1992b). Since many genes contain the binding sites for AP-1 (fos and jun complex) and Erg-1, these IE gene products may work as third messengers in this way. At 6–12 h after exposure to a hypertrophic stimulus, there is induction of cardiac fetal genes such as atrial natriuretic factor, b-myosin heavy chain and skeletal muscle a-actin (see Glennon et al., 1995 for review). Much of the existing work in the hypertrophy field uses neonatal heart cells (Sadoshima et al., 1992a, 1992b; Yamazaki et al., 1995). These cells may not be the best model for studying hypertrophy, since they still maintain some ability to divide. Adult cardiac myocytes terminally differentiate and have lost their ability to divide. They respond to increased workload by an increase in cell size.

Furthermore, the neonatal cultured cell system as well as isolated perfused hearts, exclude participation of neural, humoral and vascular factors. We have developed a unique model where we can regionally load the myocardium by applying local strain, and then compare regional changes in electrophysiology and gene expression with a control region remote from the stretched area, but in the same in situ porcine heart. The model can be used to study cardiac stretch and the initial events leading to hypertrophy, and has the following features: an adult intact blood perfused heart (pig) in situ; intact neurohumoral systems; the ability to take stretched and unstretched material from the same heart; the facility to record the electrophysiological consequences of stretch from the same segments of myocardium that we use for the molecular biological studies. The pig heart makes a particularly relevant model because it bears a close resemblance to the human heart anatomically and is likely to become increasingly important for xenotransplantation into man (Dunning et al., 1994; Sachs, 1994; Fryer et al., 1995; Bach et al., 1996). We propose to use our model to see if we can detect a possible commonality in the cascades that produce mechanically induced changes in electrophysiology with arrhythmia, and myocardial hypertrophy. That is, regional mechanical load (stretch) in the intact heart in situ initiates mechanoelectric transduction, which produce arrhythmogenic electrophysiological changes. If this stretch is prolonged, it also starts the mechanotransduction mechanisms regulating changes in gene expression, which can not only influence protein synthesis but, in turn, can also produce further electrophysiological changes.

Materials and Methods Experimental model Landrace/large white pigs of either sex were anaesthetised as described previously (Horner et al., 1994). Briefly, after tracheal intubation, the pigs were mechanically ventilated. The thorax was opened and the heart exposed. An electrode was placed on the right atrium and the heart was paced with square pulses, 2-ms long and twice stimulation threshold at a rate of 20 beats/min above the intrinsic heart rate of approximately 85 beats/min. Intraventricular and arterial pressures were recorded, and fluid balance maintained intravenously.

Regional Workload Induced Changes Vacuum or pressure

Solenoid valve



Motion MAP Suck


Figure 1 The pneumatically operated device for applying regional stretch in intact heart A small piston pivots the proximal ends of three levers around appropriately positioned fulcrums. The levers formed a tripod. Suction-operated feet on the distal ends of the levers attached the tripod to the epicardium of the ventricle of an anaesthetised pig. A downward movement of the piston caused the legs of the device to splay out, thus stretching the myocardium. Monophasic action potentials (MAP) were recorded using a contact electrode placed on the area to be stretched and a reference suction electrode placed adjacent to one of the suction feet.

Blood gases and pH were monitored throughout and kept within physiological limits.

The stretch device The stretch device consisted of a small pneumatically driven piston which pivots the proximal ends of three levers around appropriately positioned fulcrums (Fig. 1). The levers had suction-operated feet on the distal ends, which formed a tripodal base. The feet attached to the epicardium; thus, the epicardium should experience a large proportion of the imposed mechanical force. The device weighed 14 g. We selected a degree of stretch that we felt was pathologically relevant, yet within physiological limits. A reasonable yardstick was the increase in end-diastolic volume during exercise. We geometrically converted this volume change to a segment length change on the epicardial surface and then chose a comparable ratio. The device produced a stretch along a chord, whereas our calculations applied to stretch along a circumference, and the precise radius of curvature of the selected area of the right ventricle is unknown. However, the area chosen was relatively flat and any small inaccuracy


is physiologically insignificant. The distance between adjacent feet of the stretch device was 23 mm in the unstretched position and 29 mm in the stretched position. However, as we were interested in the centre of the area subtended by the device (for the molecular biological studies), we needed to confirm that the device stretched this area significantly, the stretch was within physiological limits, and that the stretch was maintained. In a separate series of five experiments, a strain gauge mechanical recording device (Lab and Woollard, 1978) was placed in the middle of the stretched area. We did not use these experiments for the molecular biological studies, because the recording device per se might have influenced the gene expression of the underlying tissue.

Electrophysiology Epicardial monophasic action potentials and electrograms were recorded before, during, and after stretch using atraumatic contact electrodes (Franz, 1983) placed on the stretched area, and a reference suction electrode placed on the adjacent area. Suction electrodes (Lab and Woollard, 1978) were also used in the separate five experiments above. All the recordings produced action potentials comparable in shape to transmembrane action potentials using microelectrodes, i.e. they have reliable time courses of repolarisation and we can measure their overall duration. Their amplitudes, however, are some 30– 60% of the transmembrane action potential. We place less reliance, in this study, on the diastolic membrane potential or the rise time of the depolarisation spike. Placement of the contact type of electrode (1 mm in diameter) in the stretched area, produces little if any surface damage. The suction electrode produces a small petechia (blood blister; 2–3 mm in diameter). We always used the atraumatic contact electrode when tissue biopsies were required for RNA extraction. The contact electrode allowed longer recording times than with the suction version (Franz, 1983).

Protocol In a series of six experiments, the stretch device was sucked onto the right ventricle and 5 min equilibration allowed. Subsequently, the electrical recordings were started. After a further 5 min, we pneumatically applied stretch and maintained it for either 30 min or 1 h. Tissue biopsies were taken


P. Meghji et al.

Segment length (mm)





5.5 0

5 10 Time (min)


(b) 310







Stretch (0.5 min)

Stretch (15 min)

Action potential duration (ms at 70%)

End diastolic length (mm)


300 290 280 270 260 250

Figure 2 Changes in end diastolic length (EDL) during stretch of the porcine right ventricle. (a) Segment length recorded using a strain gauge mechanical recording device (see text) before (control), and after stretch was applied. (b) Mean end diastolic length (EDL)±.. are shown before (control), 0.5 and 15 min after stretch was applied. These results are representative of five experiments [∗, P<0.05 compared to control. Analysis of variance for repeated measures was employed to determine whether stretch (loading conditions) influenced the segment length recordings for the various experimental stages. When a significant result was indicated by the F statistic, Student’s t-test was used to analyse the data].

from stretched and control areas of the right ventricle, avoiding the petechiae produced by the suction feet of the stretch device. These were frozen immediately in liquid nitrogen and stored at −70°C.

Signal processing All signals were fed into high impedance amplifiers (Lectromed, UK, MT8P’s) and were split via a patchpanel system to allow simultaneous digitisation (1000 Hz) using a PC, CHART software (CED, UK) and CED 1401 AD/DA converter, and storage of the analogue signal on magnetic tape (TEAC, XR501). Digital and analogue signals were continuously monitored using the PC and software system, and oscilloscopes, respectively.


Stretch (0.5 min)

Stretch (15 min)

Figure 3 Changes in monophasic action potentials (MAPs) durations during stretch of the porcine right ventricle. (a) Superimposed MAPs recorded from porcine right ventricle before (control) (–––), 0.5 (···) and 15 min (–––) after stretch was applied. Action potential amplitudes have been normalised to illustrate respective duration differences from control. (b) Mean (±..) action potential durations measured at 70% repolarisation before (control), 0.5 and 15 min after stretch was applied. These results are representative of five experiments.

Northern blot analysis Total RNA was extracted in guanidinium isothiocyanate buffer as described previously (Chomczynski and Sacchi, 1987) from tissue stored in liquid nitrogen. This was then purified to give poly A+ RNA (mRNA) using oligo(dT)-cellulose spin columns (Clontech). Poly A+ RNA from control and stretched tissue was compared using Northern blotting (Sambrook et al., 1989). The mRNA (3–6 lg) was denatured by heating (65°C) in 50% (v/v) formamide and 2.2  formaldehyde, then run by electrophoresis through a 1.2% agarose gel containing 2.2  formaldehyde, and transferred by capillary blotting onto nylon membranes (Hybond N, Amersham). The RNA was baked onto the filter for 2 h at 80°C. The membranes were prehybridised at 42°C for 3 h in a solution containing 5X SSPE,


Regional Workload Induced Changes

50% formamide, 1% sodium dodecyl sulphate, and 100 lg/ml sheared salmon sperm DNA, and were then hybridized with [a-32P]-labeled cDNA probe in the same solution at 42°C for 24 h. cDNA probes [c-fos, pc-fos1 (Van Straaten et al., 1983); c-myc, J558 (Adams et al., 1983), and glyceraldehyde 3phosphate dehydrogenase (GAPDH)] were labeled by random priming with [a-32P]dCTP (50 lCi; 3000 Ci/mmol). After hybridisation, the filter was finally washed in 0.1±0.5X standard saline citrate buffer/0.1% sodium dodecyl sulphate at 50°C and autoradiographed with intensifying screens at −70°C for 2–3 days. Blots were stripped after probing with one cDNA and then rehybridised with another probe. To normalise for the loaded amounts and transfer efficiencies, the same membranes were always rehybridised with GAPDH cDNA. Autoradiographs were scanned with a densitometer, and the ratio of the levels of c-fos or c-myc mRNA with GAPDH mRNA was calculated.



Statistical analysis Data are presented as means±.. Significant differences between groups were determined by using the Student’s t-test. A value of P<0.05 was considered as significant. The investigation was performed in accordance with the Home Office ‘‘Guidance on the operation of the Animals (Scientific Procedures) Act 1986’’, published by Her Majesty’s Stationery Office, London.

Results Electromechanical studies We selected a degree of stretch that we felt was pathologically relevant, yet within physiological limits (see Materials and Methods). The stretch along a chord was 6 mm. The distance between adjacent feet of the stretch device increased from 23 to 29 mm. As the central area subtended by the device was sampled for the molecular biological studies, we needed to confirm that the stretch there was within physiological limits, and that it was maintained. In a separate series of five experiments, a strain gauge mechanical recording device (Lab and Woollard, 1978) was placed in the middle of the stretched area. This device was calibrated as described previously (Murphy, 1994). We monitored myocardial segment movement [Fig. 2(a)].



30 min stretch

Figure 4 Northern blot analysis of mRNA (6 lg/lane) obtained from stretched and control porcine right ventricle. Autoradiogram is shown of a blot which was sequentially probed to the corresponding cDNAs for c-fos, c-myc, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH), stripping between each. GAPDH was used to normalise for loading and transfer efficiencies. The results illustrated are representative of three experiments.

The motion was, as expected, somewhat attenuated compared with control areas, but the phase relation with the intraventricular pressure was normal. Activating the stretch device (splaying its legs) stretched the area, as shown, by an increase in measured end diastolic length [Fig. 2(b)]. Contractile excursion was also further restricted. These observations show that the loading conditions of

Normalised against GAPDH (arbitrary units)


P. Meghji et al. c-fos


20 15


10 5 0 15




in expression of IE genes, c-fos and c-myc (Figs 4 and 5) compared to control. There was no difference in expression of GAPDH, which was used for normalising data. The bar chart in Figure 5 shows that expression of c-fos or c-myc was increased by 11.8- and 3.4-fold, respectively, after 30 min of stretch and 7.6- and 3.0-fold after 60 min of stretch. This was despite the fact that the stretch stimulus was maintained suggesting that the expression of these genes increases transiently, reaching a peak before 60 min.

5 0

Discussion Control


30 min



60 min

Figure 5 Bar graph showing results of densitometric analyses of Northern blot experiments. A portion of the right ventricle was stretched either for 30 min or 1 h. Autoradiograms with mRNA signals were scanned using a densitometer, and the values (arbitrary absorbance units) were normalised against those obtained for glyceraldehyde-3-phosphate dehydrogenase (GAPDH) to express mRNA as ratios (±..). The results illustrated are derived from three experiments (three pigs) for each time point (∗, P<0.05 compared to control using the Student’s t-test).

the experimental area was increased (P<0.05) compared with that before stretch and that the increase was maintained. There were no obvious changes in the control area (not shown). Regional stretch significantly shortened the right ventricular monophasic action potential duration at 70% depolarisation by 3.5±0.8% within 5 s [P<0.05; n=5; Figs 3(a) and (b)] in the stretched area. The shortening of the action potential during stretch, however, was transient. Fifteen minutes after application of stretch, the action potential duration at 70% depolarisation significantly relengthened by 3.4±1.5% compared to the prestretched value (P<0.05; n=5). The changes in duration of monophasic action potentials were not due to alterations in heart rate, as the heart was paced.

Gene expression studies Low expression of c-fos and c-myc was observed for all control samples of tissue at 30 and 60 min (Figs 4 and 5). The control levels at 60 min were not markedly different from the control levels at 30 min. Stretch applied for 30 or 60 min induced an increase

This study shows, in parallel, regional workloadinduced changes in electrophysiology and protooncogene expression from the same area of the right ventricle of intact pig heart in situ. To our knowledge, this type of electrical–molecular study is unique. The action potential duration shortens within seconds, then re-lengthens over the next few minutes. IE genes are expressed within 0.5 h. Similar rapid increases in IE gene mRNA levels including c-fos and c-myc, in response to stretching cultured cardiac myocytes on silicone membranes, have been shown (Komuro et al., 1990; Sadoshima et al., 1992a, b; Sadoshima and Izumo, 1993). However, these are immature cells which are devoid of influences from other cell types. Our study reported here is different from earlier studies, in that we studied control and stretched tissue from the same in situ heart. This has the additional advantage that the myocardium is studied in a ‘‘normal’’ environment, with interaction with neighboring cells and, furthermore, there is exposure to neural, humoral and vascular factors.

Limitations of the model, and possible pitfalls in interpretation Although the study system confers advantage over isolated myocytes or isolated perfused hearts, there are some intrinsic disadvantages. It is not possible to know which cell type is responsible for the changes in gene expression, since RNA was isolated from samples of right ventricle comprising a mixture of cell types. In a previous study on isolated cells, induction of IE genes by stretch was observed both in the myocyte and non-myocyte fractions (Sadoshima et al., 1992a). It is likely that a variety of cell types are involved in whole heart hypertrophy. That notwithstanding, one could argue that this is

Regional Workload Induced Changes

the ‘‘real’’ situation, and, in addition, the nonmyocyte fractions, for example fibroblasts (Butt et al., 1995) are important in remodelling. It is conceivable that the contact electrode itself may have induced gene changes. We think this highly unlikely. The electrode was applied gently to the myocardium, and the volume of tissue sampled was several times larger than the volume of tissue subtended by the electrode.

Gene expression Our results are in keeping with previous findings of IE gene induction in stretched cultured cells (Komura et al., 1991; Sadoshima et al., 1992a, 1992b; Yamazaki et al., 1995) and Langendorff hearts which were pressure- or volume-overloaded (Bauters et al., 1988; Schunkert et al., 1991; Kolbeck-Ruhmkorff et al., 1993). Thus, as far as expression of IE genes is concerned, the in situ heart behaves essentially in the same way as these in vitro models. Perhaps more relevantly put, we can confirm that these in vitro models behave essentially in the same way as the in situ heart.

Electrophysiology The initial stretch or load-induced shortening of action potential duration we see is in keeping with several other studies in a variety of preparations (see review; Lab, 1996), including the in situ porcine heart. These studies show that increased load shortens the action potential duration, and that load affects the T wave (see reviews by Lab, 1982, 1996; Dean and Lab, 1988; Franz, 1996). One widely accepted mechanism for the early shortening of action potential is an increase in the open probability of stretch-activated channels (Hansen et al., 1990, 1991). However, the action potential duration in our present study then re-lengthened. In this case, invoking the stretch-activated channel mechanism means opening may have been transient to explain the re-lengthening while the stretch was still on. Adaptation has been previously demonstrated for stretch-activated channels (Gustin et al., 1988; Hamill and McBride, 1992; Small and Morris, 1994). Single channel analysis has suggested that this adaptive behavior occurs because of a reduction in open channel probability rather than a decrease in channel conductance, and further application of pressure or stretch is required to achieve the same open probability (Hamill and McBride, 1992). Alternatively, there could have


been adaptation at a more ‘‘gross’’ level: for example, slippage or stress relaxation over time of the extracellular matrix or cytoskeleton (Dick and Lab, 1997). This would transmit less pulling tension on the stretch activated channel.

‘‘Electromolecular’’ aspects The observation that the initial electrical response to stretch is transient has, at least, two interpretations in relation to the induction of the protooncogenes: (1) the stretch activated channel is not causally related, (or may not be the only parameter involved); and (2) the transient opening of the channels is a trigger that activates a cascade leading to the induction of IE genes. In a recent study, skeletal muscle was stretched for different lengths of time and induction of IE genes was measured (Dawes et al., 1996). The authors concluded that it was necessary for the stretch to be applied continuously over 1 h in order to observe the full induction of c-fos and c-jun mRNA. If stretch activated channel opening is transient, the Dawes et al. (1996) observation would support the suggestion that shortening of the action potential duration (i.e. stretch-activated channel opening) is not the only parameter that is involved in the induction of the protooncogenes. This is also supported by the observation that gadolinium at a concentration that inhibited opening of stretch-activated ion channels failed to prevent either the increase in protein synthesis or the stretch-induced IE gene expression in cultured neonatal cardiac ventricular myocytes (Komura et al., 1991; Sadoshima et al., 1992b). The mechanism for the re-lengthening of the action potential may provide clues for the events downstream of the electrophysiological changes. For example, possible mechanisms for stretch-activated channel re-closure could include extracellular matrix or cytoskeletal ‘‘slippage’’. One could regard this as an early manifestation of remodelling. It is not clear yet whether extracellular matrix, integrins, cytoskeleton and growth factors do play a role in stretch activated channel function and remodelling (Sadoshima et al., 1992b; Juliano and Haskill, 1993; Dick et al., 1995; Sackin, 1995; Dick and Lab, 1997). Release of angiotensin II and endothelin-I from cardiac myocytes has been demonstrated using in vitro models of load (stretch)-induced cardiac hypertrophy (Sadoshima et al., 1993; Yamazaki et al., 1996). The characteristic stretch-induced hypertrophic responses such as immediate-early genes and fetal genes, as well as an increase in protein synthesis, were blocked by antagonists of


P. Meghji et al.

angiotensin II and endothelin I receptors (Sadoshima et al., 1993; Yamazaki et al., 1996). Stretch could also involve a number of other factors including cell signal cascades intimately related to the cytoskeleton (Wang et al., 1993). In summary, we have demonstrated that our intact in situ porcine heart model can not only be used for studying electrophysiological and molecular biological changes between stretched and control tissue from the same heart, but it can be useful for gaining insight into the cascade of events that lead to hypertrophy and remodelling. Furthermore, the use of stretch-activated channel blockers, such as gadolinium will allow us to ascertain the role of these channels in the responses that we have observed. Our experience with anaesthetised pigs so far suggests that stretch could be applied for periods up to 12 h, so it would even be possible to study the re-expression of fetal gene programme. Finally, appropriate timing of stretch on a beat-to-beat basis, which we can do with our stretch apparatus, may allow the model to be used to assess the relative importance of systolic v diastolic load.

Acknowledgments This work was supported by grants from the British Heart Foundation, the Garfield Weston Trust, the Muscular Dystrophy Association and the Wellcome Trust. We are grateful to Robin Price for help in designing and for making the stretch device, and to Barbara Coen for technical assistance.

References A JM, G S, W E, C LM, C S, 1983. Cellular myc oncogene is altered by chromosome translocation to an immunoglobulin locus in murine plasma-cytomas and is rearranged similarly in human Burkitt lymphomas. Proc Nat Acad Sci USA 80: 1982– 1986. A J, V JP, A A, T L, M JL, D JL, 1995. Evidence favoring the hypothesis that ventricular arrhythmias have prognostic significance in left ventricular hypertrophy secondary to systemic hypertension. Am J Cardiol 76: 60D–63D. B FH, R SC, W CJ, S K, F C, W H, 1996. Genetic engineering of endothelial cells to ameliorate xenograft rejection. Clin Transplant 10: 124–127. B C, M JM, B J, M C, ER R, S S, S B, 1988. Coronary flow as a determinant of c-myc and c-fos protooncogene expression in an isolated adult rat heart. J Mol Cell Cardiol 20: 97–101.

B RP, L GT, B JE, 1995. Mechanical load and polypeptide growth factors stimulate cardiac fibroblast activity. Ann NY Acad Sci 752: 387–393. C P, S N, 1987. Single-step method of RNA isolation by acid guanidinium thiocyanatephenol-chloroform extraction. Anal Biochem 162: 156– 159. D NJ, C VM, P KS, N H, G DF, 1996. The induction of c-fos and c-jun in the stretched Latissimus Dorsi muscle of the rabbit: responses to duration, degree and re-application of the stretch device. Exp Physiol 81: 329–339. D JW, L MJ, 1988. The effect of load changes on the Q-T interval and T-wave of the epicardial ECG in the anaesthetized pig. J Physiol 406: 217 (Abstract). D DJ, L MJ, 1997. Mechanical modulation of stretch induced premature ventricular beats: induction of a mechanoelectric adaptation period. (Manuscript submitted). D DJ, C BA, L MJ, 1995. The effect of cytochalasin-B on stretch induced arrhythmia in the Langendorff perfused isolated rabbit heart. J Physiol 489: 166P (Abstract). D JJ, W DJ, W J, 1994. The rationale for xenotransplantation as a solution to the organ shortage. Pathol Biol (Paris) 42: 231–235. F MR, 1983. Long-term recording of monophasic action potentials from human endocardium. Am J Cardiol 51: 1629–1634. F MR, 1996. Mechano-electric feedback in ventricular myocardium. Cardiovasc Res 32: 15–24. F JP, L JR, M AJ, 1995. The emergence of xenotransplantation. Transpl Immunol 3: 21–31. G PE, S PH, P-W PA, 1995. Cellular mechanisms of cardiac hypertrophy. Br Heart J 73: 496–499. G MC, Z X-L, M B, K C, 1988. A mechanosensitive ion channel in the yeast plasma membrane. Science 242: 762–765. H OP, MB DW, 1992. Rapid adaptation of single mechanosensitive channels in Xenopus oocytes. Proc Natl Acad Sci USA 89: 7462–7466. H DE, C CS, H LM, 1990. Stretchinduced arrhythmias in the isolated canine ventricle: evidence for the importance of mechano-electrical feedback. Circulation 81: 1094–1105. H DE, B M, S GP J, T LK, 1991. Dose-dependent inhibition of stretch-induced arrhythmias by gadolinium in isolated canine ventricles. Evidence for a unique mode of antiarrhythmic action. Circ Res 69: 820–831. H SM, L MJ, M CF, D DJ, Z B, H FG, 1994. Mechanically induced changes in action potential duration and left ventricular segment length in acute regional ischaemia in the in situ porcine heart. Cardiovasc Res 28: 528–534. J MA, J JV, 1989. Ventricular arrhythmia in untreated newly presenting hypertensive patients compared with a matched normal population. J Hypertension 7: 409–415. J RL, H S, 1993. Signal transduction from the extracellular matrix. J Biol Chem 120: 577–585. K-R C, H A, Z HG, 1993. Effect of pressure and volume overload on proto-oncogene expression in the isolated working rat heart. Cardiovasc Res 27: 1998–2004.

Regional Workload Induced Changes K I, K T, S Y, K M, K Y, H E, T F, Y Y, 1990. Stretching cardiac myocytes stimulates protooncogene expression. J Biol Chem 265: 3595–3598. K I, K Y, K T, S Y, K M, H E, T F, Y Y, 1991. Mechanical loading stimulates cell hypertrophy and specific gene expression in cultured rat cardiac myocytes. J Biol Chem 266: 1265–1268. L MJ, 1982. Contraction-excitation feedback in myocardium. Physiological basis and clinical relevance. Circ Res 50: 757–766. L MJ, 1996. Mechanoelectric feedback (transduction) in heart: concepts and implications. Cardiovasc Res 32: 3–14. L MJ, W KV, 1978. Monophasic action potentials, electrocardiograms and mechanical performance in normal and ischaemic epicardial segments of the pig ventricle in situ. Cardiovasc Res 12: 555–565. L D, G RJ, S DD, K WB, C WP, 1990. Prognostic implications of echocardiographically determined left ventricular mass in the Framingham heart study. N Engl J Med 322: 1561– 1566. M T, 1991. Mechanisms of alternative pre-mRNA splicing. Science 251: 33–34. M C, 1994. Investigation of the relationship between electromechanical alternans and mechanoelectric feedback: relevance to cardiac arrhythmia. MD thesis. Charing Cross and Westminster Medical School, University of London. S DH, 1994. The pig as a potential xenograft donor. Vet Immunol Immunopathol 43: 185–191. S H, 1995. Mechanosensitive channels. Annu Rev Physiol 57: 333–353. S J, I S, 1993. Mechanical stretch rapidly activates multiple signal transduction pathways in cardiac myocytes: potential involvement of an autocrine/paracrine mechanism. EMBO J 12: 1681– 1692. S J-I, J L, T T, K TJ, I S, 1992a. Molecular characterization of the stretchinduced adaptation of cultured cardiac cells. An in


vitro model of load-induced cardiac hypertrophy. J Biol Chem 267: 10551–10561. S J-I, T T, J L, I S, 1992b. Roles of mechano-sensitive ion channels, cytoskeleton, and contractile activity in stretch-induced immediateearly gene expression and hypertrophy of cardiac myocytes. Proc Natl Acad Sci 89: 9905–9909. S J-I, X Y, S HS, I S, 1993. Autocrine release of angiotensin II mediates stretchinduced hypertrophy of cardiac myocytes in vitro. Cell 75: 977–984. S J, F EF, M T, 1989. Molecular Cloning: a laboratory manual, second edition. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press. S H, J L, I S, A CS, L BH, 1991. Localization and regulation of c-fos and c-jun protooncogene induction by systolic wall stress in normal and hypertrophied rat hearts. Proc Natl Acad Sci 88: 11480–11484. S DL, M CE, 1994. Delayed activation of single mechanosensitive channels in Lymnaea neurons. Am J Physiol 267: C598–C606. S PH, B MA, 1995. Intracellular signalling through protein kinases in the heart. Cardiovasc Res 30: 478–492. V S F, M R, C T, V B C, V IM, 1983. Complete nucleotide sequence of a human c-onc gene: deduced amino acid sequence of a human c-fos protein. Proc Nat Acad Sci USA 80: 3183–3187. W N, B JP, D EI, 1993. Mechanotransduction across the cell surface and through the cytoskeleton. Science 260: 1124–1127. Y T, K I, Y Y, 1995. Molecular mechanism of cardiac cellular hypertrophy by mechanical stress. J Mol Cell Cardiol 27: 133–140. Y T, K I, K S, Z Y, S I, H Y, M T, M K, K H, A R, T H, Y Y, 1996. Endothelin-1 is involved in mechanical stress-induced cardiomyocyte hypertrophy. J Biol Chem 271: 3221–3228.