The contribution of refractoriness to arrhythmic substrate in hypokalemic Langendorff-perfused murine hearts

The clinical effects of hypokalemia including action potential prolongation and arrhythmogenicity suppressible by lidocaine were reproduced in hypokalemic (3.0 mM K⁺) Langendorff-perfused murine hearts before and after exposure to lidocaine (10 μM). Novel limiting criteria for local and transmural, epicardial, and endocardial re-excitation involving action potential duration (at 90% repolarization, APD₉₀), ventricular effective refractory period (VERP), and transmural conduction time (Δlatency), where appropriate, were applied to normokalemic (5.2 mM K⁺) and hypokalemic hearts. Hypokalemia increased epicardial APD₉₀ from 46.6 ± 1.2 to 53.1 ± 0.7 ms yet decreased epicardial VERP from 41 ± 4 to 29 ± 1 ms, left endocardial APD₉₀ unchanged (58.2 ± 3.7 to 56.9 ± 4.0 ms) yet decreased endocardial VERP from 48 ± 4 to 29 ± 2 ms, and left Δlatency unchanged (1.6 ± 1.4 to 1.1 ± 1.1 ms; eight normokalemic and five hypokalemic hearts). These findings precisely matched computational predictions based on previous reports of altered ion channel gating and membrane hyperpolarization. Hypokalemia thus shifted all re-excitation criteria in the positive direction. In contrast, hypokalemia spared epicardial APD₉₀ (54.8 ± 2.7 to 60.6 ± 2.7 ms), epicardial VERP (84 ± 5 to 81 ± 7 ms), endocardial APD₉₀ (56.6 ± 4.2 to 63.7 ± 6.4 ms), endocardial VERP (80 ± 2 to 84 ± 4 ms), and Δlatency (12.5 ± 6.2 to 7.6 ± 3.4 ms; five hearts in each case) in lidocaine-treated hearts. Exposure to lidocaine thus consistently shifted all re-excitation criteria in the negative direction, again precisely agreeing with the arrhythmogenic findings. In contrast, established analyses invoking transmural dispersion of repolarization failed to account for any of these findings. We thus establish novel, more general, criteria predictive of arrhythmogenicity that may be particularly useful where APD₉₀ might diverge sharply from VERP.     

Criteria for arrhythmogenicity in genetically-modified Langendorff-perfused murine hearts modelling the congenital long QT syndrome type 3 and the Brugada syndrome

The experiments investigated the applicability of two established criteria for arrhythmogenicity in Scn5a+/Δ and Scn5a+/− murine hearts modelling the congenital long QT syndrome type 3 (LQT3) and the Brugada syndrome (BrS). Monophasic action potentials (APs) recorded during extrasystolic stimulation procedures from Langendorff-perfused control hearts and hearts treated with flecainide (1 μM) or quinidine (1 or 10 μM) demonstrated that both agents were pro-arrhythmic in wild-type (WT) hearts, quinidine was pro-arrhythmic in Scn5a+/Δ hearts, and that flecainide was pro-arrhythmic whereas quinidine was anti-arrhythmic in Scn5a+/− hearts, confirming clinical findings. Statistical analysis confirmed a quadratic relationship between epicardial and endocardial AP durations (APDs) in WT control hearts. However, comparisons between plots of epicardial against endocardial APDs and this reference curve failed to correlate with arrhythmogenicity. Restitution curves, relating APD to diastolic interval (DI), were then constructed for the first time in a murine system and mono-exponential growth functions fitted to these curves. Significant (P < 0.05) alterations in the DI at which slopes equalled unity, an established indicator of arrhythmogenicity, now successfully predicted the presence or absence of arrhythmogenicity in all cases. We thus associate changes in the slopes of restitution curves with arrhythmogenicity in models of LQT3 and BrS.     

Restitution analysis of alternans and its relationship to arrhythmogenicity in hypokalaemic Langendorff-perfused murine hearts

Alternans and arrhythmogenicity were studied in hypokalaemic (3.0 mM K⁺) Langendorff-perfused murine hearts paced at high rates. Epicardial and endocardial monophasic action potentials were recorded and durations quantified at 90% repolarization. Alternans and arrhythmia occurred in hypokalaemic, but not normokalaemic (5.2 mM K⁺) hearts (P < 0.01): this was prevented by treatment with lidocaine (10 μM, P < 0.01). Fourier analysis then confirmed transition from monomorphic to polymorphic waveforms for the first time in the murine heart. Alternans and arrhythmia were associated with increases in the slopes of restitution curves, obtained for the first time in the murine heart, while the anti-arrhythmic effect of lidocaine was associated with decreased slopes. Thus, hypokalaemia significantly increased (P < 0.05) maximal gradients (from 0.55 ± 0.14 to 2.35 ± 0.67 in the epicardium and from 0.67 ± 0.13 to 1.87 ± 0.28 in the endocardium) and critical diastolic intervals (DIs) at which gradients equalled unity (from −2.14 ± 0.52 ms to 50.93 ± 14.45 ms in the epicardium and from 8.14 ± 1.49 ms to 44.64 ± 5 ms in the endocardium). While treatment of normokalaemic hearts with lidocaine had no significant effect (P > 0.05) on either maximal gradients (0.78 ± 0.27 in the epicardium and 0.83 ± 0.45 in the endocardium) or critical DIs (6.06 ± 2.10 ms and 7.04 ± 3.82 ms in the endocardium), treatment of hypokalaemic hearts with lidocaine reduced (P < 0.05) both these parameters (1.05 ± 0.30 in the epicardium and 0.89 ± 0.36 in the endocardium and 30.38 ± 8.88 ms in the epicardium and 31.65 ± 4.78 ms in the endocardium, respectively). We thus demonstrate that alternans contributes a dynamic component to arrhythmic substrate during hypokalaemia, that restitution may furnish an underlying mechanism and that these phenomena are abolished by lidocaine, both recapitulating and clarifying clinical findings.     

Empirical correlation of triggered activity and spatial and temporal re-entrant substrates with arrhythmogenicity in a murine model for Jervell and Lange-Nielsen syndrome

KCNE1 encodes the β-subunit of the slow component of the delayed rectifier K⁺ current. The Jervell and Lange-Nielsen syndrome is characterized by sensorineural deafness, prolonged QT intervals, and ventricular arrhythmogenicity. Loss-of-function mutations in KCNE1 are implicated in the JLN2 subtype. We recorded left ventricular epicardial and endocardial monophasic action potentials (MAPs) in intact, Langendorff-perfused mouse hearts. KCNE1 ^−/− but not wild-type (WT) hearts showed not only triggered activity and spontaneous ventricular tachycardia (VT), but also VT provoked by programmed electrical stimulation. The presence or absence of VT was related to the following set of criteria for re-entrant excitation for the first time in KCNE1 ^−/− hearts: Quantification of APD₉₀, the MAP duration at 90% repolarization, demonstrated alterations in (1) the difference, ∆APD₉₀, between endocardial and epicardial APD₉₀ and (2) critical intervals for local re-excitation, given by differences between APD₉₀ and ventricular effective refractory period, reflecting spatial re-entrant substrate. Temporal re-entrant substrate was reflected in (3) increased APD₉₀ alternans, through a range of pacing rates, and (4) steeper epicardial and endocardial APD₉₀ restitution curves determined with a dynamic pacing protocol. (5) Nicorandil (20 μM) rescued spontaneous and provoked arrhythmogenic phenomena in KCNE1 ^−/− hearts. WTs remained nonarrhythmogenic. Nicorandil correspondingly restored parameters representing re-entrant criteria in KCNE1 ^−/− hearts toward values found in untreated WTs. It shifted such values in WT hearts in similar directions. Together, these findings directly implicate triggered electrical activity and spatial and temporal re-entrant mechanisms in the arrhythmogenesis observed in KCNE1 ^−/− hearts.     

心力衰竭家兔心室动作电位时程整复性变化对室性心律失常的影响

 目的：探讨压力负荷诱导的兔慢性心力衰竭(CHF)模型离体心室动作电位时程整复性(APDR)变化对室性心律失常(VA)的影响。方法：雄性新西兰大耳兔20只,随机分为对照(CTL)组和CHF组,每组10只。CHF模型制备采用经腹主动脉缩窄术,造模结束4周后行心脏超声检查评价造模结果。在整体心脏Langendorff灌流条件下行离体电生理研究,分别记录和测量心室不同位点的单相动作电位(MAP)及有效不应期(ERP),并绘制APDR曲线;对2组心脏进行快速电刺激,观察室性心律失常（VA）的诱发。结果：与CTL组相比,CHF组心室相同部位90%单相动作电位时程(MAPD90)、ERP及APDR曲线最大斜率(Smax)均明显增大(均P<0.01),且VA更容易诱发(均P<0.05);此外,CHF组动物APDR曲线Smax的变异系数(COV-Smax)均较CTL组增大(均P<0.05)。结论：CHF时心室APDR曲线Smax及COV-Smax均增大,促进室性心律失常的发生。     

Differential effects of extracellular cesium on early afterdepolarizations in ventricular myocytes and arrhythmogenesis in isolated hearts of rats and guinea pigs

CsCl has been shown to be arrhythmogenic in-vivo and to cause early afterdepolarizations (EADs) in isolated cardiac preparations, but the underlying electrophysiological mechanisms are ill-defined. To elucidate these actions further, the effects of extracellular solutions containing 3 mM CsCl and either 2 mM KCl (Cs2K solution) or 5 mM KCl (Cs5K solution) on membrane potential and ionic currents in rat and guinea-pig ventricular myocytes were compared. Cs2K solution rapidly and reversibly inhibited outward I_K1, and reduced other K⁺ currents by about 20%. Current-clamped myocytes were rapidly hyperpolarized by this solution and action potentials were prolonged, but EADs were not observed. In contrast, EADs were triggered by E-4031, H₂O₂, and the pyrethroid tefluthrin. Membrane-potential changes reversed after replacing Cs2K with Cs5K solution, with the recovery of 50% of outward I_K1. These results suggest that Cs2K solution inhibited I_K1 and caused a late prolongation of the action-potential duration, but the affected membrane potentials were too negative to elicit EAD mechanisms. In isolated hearts perfused with modified Tyrodes, Cs2K, and Cs5K solutions, bradycardia and arrhythmias were evoked by both CsCl-containing solutions. A comparison of such results with the effects of these solutions on myocytes suggests that I_K1 inhibition and EADs in ventricular myocytes are unlikely to be involved in arrhythmogenesis under our conditions.     

Differential analysis in the time domain of the baroreceptor cardiac reflex sensitivity as a function of sequence length

Analysis of baroreceptor sensitivity (BRS) in the time domain through the spontaneous sequence method used a measure of BRS based on the average of all sequences detected, without making any distinction in cardiac cycle length. In this article, we study differentially the functioning of the baroreflex as a function of the length of the cardiac sequences (3, 4, 5, or 6 cardiac cycles). One hundred and four students performed three mental stress tasks: mental arithmetic, memory, and visual attention. The results show that (1) as sequence length decreases, the relationships between BRS and indexes of vagal cardiac control increase, (2) the BRS associated with the more short sequences (3 and 4 beats) is the most vulnerable to mental stress, particularly the mental arithmetic task, and (3) BRS increases progressively as sequence length decreases. These results suggest that the nature and functioning of the baroreflex differ as a function of the length of the cardiac sequences.     

A Model Study of Intramural Dispersion of Action Potential Duration in the Canine Pulmonary Conus

Regional gradients of action potential duration (APD) due to electrophysiological differences between endocardial, midmyocardial, and epicardial myocytes may exist across the ventricular wall. In addition, activation sequence-induced gradients of APD may occur if intramural fiber rotation accelerates or decelerates the depolarization wave front. To investigate relative contributions of regional and activation sequence-induced gradients to intramural APD dispersion, we simulated action potential propagation in two-dimensional models with idealized geometries representing the canine pulmonary conus. Ionic currents for endocardial myocytes were described using the Luo–Rudy membrane equations. Modifications to I_Ks approximated action potentials of epicardial and midmyocardial cells. Spatial coupling was modeled with a bidomain representation of tissue structure that included unequal anisotropic conductivity ratios. Activation sequence-induced gradients reached 69 ms cm–1 during a nonuniform activation sequence where the change in orientation between endocardial and epicardial fibers accelerated the depolarization wave front. Regional gradients reached 133 ms cm^–1 at the boundary between endocardial and midmyocardial cells. When regional and activation sequence-induced gradients were oriented in opposite directions, overall APD dispersion decreased. When the gradients were oriented in the same direction, overall dispersion measured as high as 202 ms cm^–1. This gradient exceeded values previously estimated as sufficient to induce cardiac arrhythmia during premature stimulation and suggests that regional and activation sequence-induced gradients increase arrhythmia vulnerability in the presence of other arrhythmogenic conditions. © 1998 Biomedical Engineering Society. PAC98: 8710+e, 8722Fy     

The effect of muscle length on motor unit discharge characteristics in human tibialis anterior muscle

Summary Muscle length influences the contractile properties of muscle in that when muscle is lengthened the relaxation phase of the muscle twitch is prolonged and when muscle is shortened, the relaxation phase is shorter in duration. As a result, the force exerted by active motor units varies with muscle length during voluntary contractions. To determine if motoneuron spike trains were adjusted to accommodate for changes in the contractile properties imposed by shortened and lengthened muscle, motor unit action potentials were recorded from the tibialis anterior muscle at different muscle lengths. Twenty subjects performed isometric ramp contractions at ankle angles of 20° dorsiflexion, neutral between dorsiflexion and plantar flexion, and 30° plantar flexion, which put the tibialis anterior muscle in a shortened, neutral, or lengthened condition, respectively. During isometric contractions where torque increased at 5% MVC/s, motor unit discharge rate at recruitment was greater in shortened muscle than in lengthened muscle (P<0.05). Brief initial interspike intervals (<40 ms) occurred more frequently in shortened muscle than in either neutral length or lengthened muscle. During steady contractions, motor unit discharge rate was greater per unit torque (N.m) in shortened muscle than in neutral length or lengthened muscle (P<0.05). These findings indicate that muscle length does influence the discharge pattern of motor unit spike trains during isometric ramp contractions. Spike trains with higher discharge rates at recruitment in shortened muscle may take advantage of the catch-like properties in muscle and be useful in taking up the slack in the passive elements of the muscle and tendon. During steady submaximal contractions, the higher discharge rate per unit torque (N.m) in shortened muscle is likely due to the decreased peak tension and shorter one-half relaxation time observed in shortened muscle, and may indicate that the tibialis anterior muscle is operating on the steep portion of the length-tension curve when the ankle is fully dorsiflexed.     

Effects of auditory stimulus timing in the respiratory cycle on the evoked cardiac response in man at rest

Summary The present experiments were carried out in 21 healthy adults to study the effects of auditory stimulus timing within the respiratory cycle on evoked cardiac response. The stimulus (80 dB white noise) was started by the first and finished by the second R-wave after change in respiratory phase, and presented in different series in either early inspiration or early expiration. The spirogram and eight sequential interbeat intervals (IBI) after respiratory phase change were recorded. The mean of IBI and standard deviation (SD) were calculated separately for each IBI of 20 trials for each subject, during both the prestimulus and poststimulus phases. The stimulus effects were expressed as changes from prestimulus conditions, in terms of IBI and SD. The mean of each of the eight IBI and its SD were found to vary consistently in the same direction, i.e., SD increased with increasing mean IBI. Stimulation during early inspiration did not produce any effect during this respiratory phase. It was not before the beginning of the following expiration that a significant deceleration was evoked, which was associated with an enhanced SD, whereas stimulation during early expiration promptly evoked a biphasic cardiac response of the deceleration — acceleration pattern and an increase and decrease in SD, respectively. While SD was found to be a function of age, no such finding was obtained for IBI and SD. These results are discussed in terms of the vagal gating hypothesis.     

A diffractometer using a lateral effect photodiode for the rapid determination of sarcomere length changes in cross-striated muscle

A simple method is devised to record rapid sarcomere length changes of muscle fibres using a lateral effect diode. In the standard position the diffractometer records length changes between 1.65 and 3.8 m, the output being linear 1 V/m with a frequency response of –3 dB at 1.2 kHz. The absolute error is<0.05 m between 1.65 and 2.80 m and <0.1 m between 2.81 and 3.30 m. The resolution of length changes is<0.005 m over the whole range. By varying the detector position the length range can be extended to either side, and spatial resolution can be improved at the expense of length range.     

行走步长影响老年人下肢肌力的仿真分析

背景：目前人口老龄化日趋严重,老年人不耐受高强度锻炼,骨骼、肌肉退化,常由于跌倒导致伤残或死亡,如何提高老年人的下肢肌力成为当前的研究热点问题。目的：分析行走步长对老年人（> 60岁）下肢肌力的影响,探索高效的行走方式,为日常锻炼提供指导。方法：选择30名健康老年人为受试者,均为男性,年龄65-75岁,测量其形态参数,导入仿真软件Opensim 4.0,分别建立各受试者骨肌模型。应用Vicon 2.3三维动作捕捉系统,实时记录各受试者正常步长、大步长行走的运动轨迹;同时采用足底三维测力台记录步行中足底受到的地面反作用力;将各受试者运动数据导入Opensim 4.0软件建立运动模型,进行动力学仿真,结合足底测力台数据验证其有效性。模型经验证后仿真各受试者在不同步长行走过程中,髂肌、腰大肌、股直肌、臀大肌、臀中肌、股二头肌、半腱肌、大收肌的肌力变化;将各受试者2次不同步长行走肌力数据分别导入SPSS 20.0进行统计分析,行配对t检验。结果与结论：（1）建立了各受试者骨骼肌肉模型及不同步长行走的运动模型,导出各受试者不同步长行走过程中下肢重要肌肉的时间-肌力变化曲线;（2）经统计学分...     

The impact of noise on the reliability of heart-rate variability and complexity analysis in trauma patients

This study focused on the impact of noise on the reliability of heart-rate variability and complexity (HRV, HRC) to discriminate between different trauma patients and to monitor individual patients. Life-saving interventions (LSIs) were chosen as an endpoint because performance of LSIs is a critical aspect of trauma patient care. Noise was modeled and simulated by modifying original R–R interval (RRI) sequences via decimation, concatenation, and division of RRIs, as well as R-wave detection using the electrocardiogram. Results showed that under increasing simulated noise, entropy and autocorrelation measures can still effectively discriminate between LSI and non-LSI patients and monitor individuals over time.     

Beat-to-beat measurement and analysis of the R-T interval in 24 h ECG Holter recordings

This study assesses the feasibility of beat-to-beat measurement of the R-T interval in Holter ECG recordings. The low sampling rate of the Holter system was increased by a specific interpolating filter, and the precision and accuracy of two T-wave fiducial point (T-wave maximum: T_m, T-wave end: T_e) detection algorithms were compared. The results of the validation tests show better performance of the T_m measurement procedure in the presence of high noise levels. The overall process for the beat-to-beat R-T interval measurement was then tested on ECG Holter recordings collected during free and controlled respiration. Finally, the R-T_m and the corresponding R-R intervals were measured on 24h ECG recordings of healthy subjects and the spectral analysis was applied to the constructed series. Both R-R and R-T_m spectra show two main frequency components (low-frequency ∼0·1 Hz, high-frequency ∼0·25 Hz) changing in their power ratios continuously throughout the 24h period. The method described seems to provide a dynamic index of the sympatho-vagal balance at the ventricle that can be useful for a deeper understanding of ventricular repolarisation duration variability.     

The cardiac cycle time effect revisited: temporal dynamics of the central-vagal modulation of heart rate in human reaction time tasks

Lacey and Lacey (1974) suggested that during reaction time tasks higher brain centers dynamically adjust efferent vagal nerve pulses to the sino-atrial node of the heart, inducing phase-dependent heart rate changes. Since then, animal and human neuro-physiological results have provided evidence for this hypothesis. Higher subcortical and cortical brain centers may have reciprocal interactive pathways relating to autonomic control comparable to those at the level of peripheral autonomic changes and brain stem reflexes. In humans such central effects may be observed in the short latency vagal control of heart rate that has been studied mostly in reaction time (RT) tasks. RT task parameters modulate vagal pulses to the cardiac sino-atrial node (SAN), which in turn exerts a phase-dependent change in the ongoing cardiac interbeat interval. Simulations of human RT task effects in an animal model of heart rate change support this hypothesis. The current study examined evidence for vagal control of three human phasic heart rate responses in RT tasks. The evidence indicates that the initiation of an RT response triggers a reflexive shift from vagal activation to vagal inhibition. This shift is cardiac cycle phase dependent. Graded anticipatory cardiac deceleration during the warning interval of an RT task varies with task relevance and time uncertainty. This response may be part of a control process engaged in time keeping. Hence, temporal variables mediate the central-autonomic-vagal modulation of heart rate.     

A comparative scale of autonomic function with age through the tone-entropy analysis on heart period variation

Tremendous numbers of heart rate variability studies have aimed to elucidate age-associated alterations of autonomic function in the past decades. However, the studies, far from clarifying ageing mechanisms, fell into confusion by a lack of common scales. The purpose of this study is to show a possibility to establish a comparative scale of autonomic function through a method, tone-entropy (T-E) analysis on heart period variation, whose validity has been already examined on typical physiological cases (Oida et al. in J Appl Physiol 82:1794–1801, 1997; Oida et al. in J Gerontol 54A:M219–M224, 1999a; Oida et al. in Acta Physiol Scand 165:129–134, 1999b; Oida et al. in Acta Physiol Scand 165:421–422, 1999c; Amano et al. in Eur J Appl Physiol 94:602–610, 2005). In this study, 276 subjects from teens to seventies were examined at rest by T-E analysis together with conventional time and frequency domain analyses. The tone (negativity represents vagal predominance) became significantly high [−0.174 ± 0.026 (teens) to −0.024 ± 0.004 (seventies), P < 0.05 for one-way ANOVA], and the entropy (total autonomic activity), significantly low [4.40 ± 0.12 (teens) to 2.90 ± 0.09 bit (seventies), P < 0.05] with advancing age. The result, plotted in 2-D T-E space, showed that the ageing traced a curvi-linear relation from right-bottom to left-top, and was consistent with previously studied typical physiological cases. The conventional analyses showed almost the same autonomic reduction as T-E did, but failed in detecting delicate alteration of autonomic balance. The results, showing that autonomic activity reduced in both pathways impairing vagal predominance significantly with ageing, suggested a possibility to assess autonomic function in 2-D T-E space in a comparative way.     

A new algorithm developed based on a mixture of spectral and nonlinear techniques for the analysis of heart rate variability

In this paper, an algorithm based on a joint use of spectral and nonlinear techniques for heart rate variability (HRV) analysis is proposed. First, the measured RR data are passed into a trimmed moving average (TMA)-based filtering system to generate a lower frequency (LF) time series and a higher frequency (HF) one that approximately reflect the sympathetic and vagal activities, respectively. Since the Lyapunov exponent can be used to characterize the level of chaos in complex physiological systems, the largest Lyapunov exponents corresponding to the complex sympathetic and vagal systems are then estimated from the LF and HF time series, respectively, using an existing algorithm. Numerical results of a postural maneuver experiment indicate that both characteristic exponents or their combinations might serve as a set of innovative and robust indicators for HRV analysis, even under the contamination of sparse impulses due to aberrant beats in the RR data.     

Effect of testosterone and the oestrous cycle on neuronal refractory periods and firing rates of stria terminalis neurones in the female rat

Summary Absolute refractory periods of a subpopulation of corticomedial amygdala (CMA) neurones which project to the medial preoptic/ anterior hypothalamic junction (MPH) via the stria terminalis were recorded in the female rat. Previous experiments have shown that this sub-population of CMA neurones is testosterone-sensitive in the male rat. In the ovariectomised female testosterone propionate (TP, 200 g/day for 18–22 days) significantly reduced the mean absolute refractory period of these CMA neurones compared to oil treated controls (from 1.34 ms to 0.87 ms). In a second experiment the absolute refractory periods of these CMA neurones were measured during the pro-oestrus and di-oestrus stages of the oestrous cycle as well as in ovariectomised controls. The mean absolute refractory period of these neurones was significantly shortened at pro-oestrus (0.99 ms) compared both to animals in di-oestrus 1 (1.45 ms) and ovariectomised controls (1.42 ms). The median firing rate of these CMA neurones was also significantly increased at pro-oestrus (1.24 Hz) compared to di-oestrus 1 (0.13 Hz) and ovariectomised controls (0.09 Hz). No firing rate increase was observed after TP treatment of ovariectomised animals in the first experiment.Results show that the testosterone-sensitive CMA neurones originally discovered in the male rat have a similar sensitivity in the female rat. They also show that the refractory periods of these neurones are shortened at pro-oestrus. Further, these same neurones also show firing rate increases at this time, although such an increase has not been observed in gonadally intact male rats when compared to castrated ones. Results are interpreted in terms of a possible functional sexual dimorphism in the output of these CMA neurones.Supported by the Medical Research Council