Microelectrode Study of Alternating Responses to Repetitive Premature Excitation in Canine Purkinje Fibers

Microelectrode Study of Alternating Responses to Repetitive Premature Excitation in Canine Purkinje Fibers By Jack P. Bandura and Daniel A. Brody ABSTRACT Microelectrode studies were performed to produce alternation of coupled beats in the specialized conduction system of the heart and to determine the underlying mechanisms of their production. A typical preparation consisted of a superfused Y-shaped bifurcating segment of canine Purkinje tissue, in which the two branches of the Y were approximately symmetrical and terminated in a functionally communicating bridge of myocardium. Transmembrane potentials were recorded simultaneously from a proximal and a distal location in one branch of the Y. Stimulating the stem of the preparation with 60 pulse pairs/min at a critical coupling interval produced alternation of the coupled beats. Alternate reversal of the proximl-to-distal sequence of activation was observed under these circumstances. This finding probably resulted from alternate antegrade block of the instrumental branch accompanied by retrograde conduction through the accessory pathway offered by the other branch and the common-myocardial bridge. The validity of this inference was confirmed by microelectrode exploration and by transection of the retrograde pathway. Analysis of the results demonstrated further that alternating block was also intimately related to the alternation of preceding cycle lengths. KEY WORDS transmembrane action potentials rate-dependent block cycle length and repolarization disparity of refractoriness retrograde activation in orthograde block superfused Purkinje tissue Studies previously performed by Cohen et al. (l) demonstrated an unusual electrocardiographic response to repetitive premature stimulation. Their investigations in both dog and man revealed that appropriately coupled premature stimuli applied to the atria could produce an alternating pattern of ventricular excitation. Using His bundle recording techniques, they verified that this phenomenon was accompanied by constancy of the preceding cycle length, premature coupling interval, and atrioventricular conduction time (1). These authors (1) hypothesized that, during conditions of critical repetitive premature ventricular excitation, a portion of the specialized conduction system was blocked and that this blocked portion had a shorter diastolic recovery period than it did if conduction through the affected segment had been orthograde. They attributed this phenomenon to delayed retrograde activation of the forwardblocked segment. The present study was performed on isolated branching Purkinje fiber com- From the Department of Medicine, Section of Medical Physics, University of Tennessee College of Medicine, Memphis, Tennessee 38163. This study was supported by U. S. Public Health Service Grants HL-01362, HL-09495, HL-14032, and HL-18265 of the National Heart and Lung Institute. Received September 19, 1973. Accepted for publication January 4, 1974. 406 plexes to evaluate single fiber activity during alternating activation sequences produced by repetitive premature stimulation. Cellular electrical events were recorded during a two-to-one orthograde block with retrograde activation of the blocked pathway to demonstrate the parameters that were apparently responsible for the production of this alternating activation sequence and to analyze further the previously proposed supposition. Methods Adult dogs were anesthetized with sodium pentobarbital (30 mg/kg, iv). A lateral incision was made in the fourth intercostal space, and the heart was rapidly excised. Branching complexes of Purkinje tissue were removed from the ventricles and placed in a Lucite chamber that was continuously perfused with oxygenated modified Tyrode's solution aerated with 95% O 2-5% CO 2 and maintained at 37 C. The millimolar composition of the Tyrode's solution was: NaCl 137.0 KC1 2.7, MgCl 2 0.5, NaH 2 PO 4 0.18, NaHCO 3 12.0, CaCl s 2.7, and glucose 5.5. Fifteen Y-shaped anatomic configurations of Purkinje tissue were selected for study; Figure 1 represents a typical preparation within the chamber. The preparations were electrically driven by 60 pulse pairs/min. Microelectrodes filled with 3M KC1, having a resistance of 5-10 megohms, were used to record transmembrane events of single fibers within both branches. The signals were preamplified (NF-1 Circulation Rwwrrfc, VoL XXXIV, March 1974

ALTERNATING PREMATURE EXCITATION 407 FIGURE 2 FIGURE 1 Photograph of the preparation within the chamber, c represents the stem of the Y-shaped Purkinje configuration, a and b represent the two rami of the bifurcated bundle which are interconnected by the distal muscle mass mmd. SE represents the stimulating electrode applied to the stem of the Y-shaped preparation. The muscle mass attached to the stem of the Y was used to pin the proximal end of the preparation for mechanical stability. Bioelectric, Inc.) and displayed on a conventional fourchannel oscilloscope (Tektronix) for photographing with a Polaroid oscilloscope camera (2). The signals were also simultaneously recorded on a four-channel analog tape recorder (Hewlett-Packard 3960). Results Figure 2 presents typical results obtained from 5 of the 15 preparations in response to critically coupled paired stimulation. Section S displays the control activation times of transmembrane action potentials simultaneously recorded at sites 1 and 2 from limb a of the preparation in response to the primary stimulus of the pulse pair. The associated diagram of the Y-shaped preparation displays the location of the stimulating electrodes SE and the recording microelectrodes 1 and 2 and the hypothetical pathways that the impulse normally traclrculatkxi RocnrtA, Vol. XXXIV, Uarch 1974 Alternating patterns of response to repetitive premature stimuli at an S-X coupling interval of 275 msec. The preparation was taken from the right ventricle. S represents the control activation times in response to the primary stimulus of the pulse pair. Xj represents orthogradely propagated impulses through limb a. X2 represents the alternate premature response that was orthogradely blocked in limb a and subsequently conducted through an accessory pathway, namely, limb b and the distal muscle mass miry. In each section a diagram of the Y-shaped preparation illustrates the position of the stimulating electrodes SE and the recording microelectrodes 1 and 2. The apparent pathways of impulse transmission have been drawn in with arrows to indicate the direction of propagation. Heavy black bar(s) indicate a blocked impulse. Recordings 1 and 2 come from microelectrodes 1 and 2, respectively. Heavy arrows indicate the stimulus artifact. The base line, 0, indicates a potential of 0 me. The upstrokes of the microelectrode recordings have been retouched. Voltage and time calibrations for each section are 50 mv and 10 msec, respectively. veled. The activation times are 1 msec and 3 msec for cells impaled by microelectrodes 1 and 2, respectively. The activation times were determined as the time interval from the onset of the stimulus artifact recorded on a separate channel of the tape to the earliest potential deviation associated with the upstroke of the transmembrane response. Sections X, and X2 show activation times of the alternating responses to the premature second stimuli of the pulse pairs. The S-X, and S-X2 coupling intervals, which are identical, are 275 msec. X, demonstrates prolongation of activation times to 2 msec and 10 msec for microelectrodes 1 and 2, respectively, in response to the early premature stimulus; this prolongation reflects increased conduction delay. The response seen in recording 1 arises from a prepotential that indicates block in the vicinity of that electrode (3). Xj represents the alternate response to the premature stimulus. Following this stimulus only an

408 BANDURA.BRODY electrotonic event, which failed to attain threshold, is recorded by the proximal microelectrode 1. The failure of this low-amplitude electrotonic potential to activate the fiber resulted in orthograde block. This finding is represented diagrammatically by the interruption of the transmitted event through limb a. Following the forward block there was a period of delay, because the impulse had to travel through limb b and the distal muscle mass mm d ; then both cells were activated in the reverse sequence. The activation times during retrograde activation were 42.5 msec and 35 msec for microelectrodes 1 and 2, respectively. The activation sequence shown in X, and X 2 consistently alternated at a coupling interval of 275 msec, and the coupling interval was extremely critical. Alteration of the coupling interval by ± 3 msec terminated the alternating response. The membrane potential at the onset of the propagated deflection ( 61 mv) was 2 mv more negative than was the membrane potential at the onset of the nonpropagated deflection ( 59 mv). Following this recording, limb b was transected and, subsequently, only the orthograde propagation sequence was produced in response to early premature stimulation. Figure 3 demonstrates the results following the transection. Section S again represents the activation times of the transmembrane action potentials recorded at sites 1 and 2 from limb a in response to the primary stimulus of the pulse pair. The activation times were identical to those prior to transection of limb b. X[ displays prolongation of the orthograde response resulting from early premature stimulation at an S-X, coupling interval of 283 msec. The activation times are the same as in section X, of Figure 2. Foreshortening of the S-X, coupling interval in an attempt to produce an alternating excitation sequence produced block at an S-X 2 coupling interval of 277 msec (Fig. 3, X 2 ). Again the electrotonic prepotential was observed in microelectrode 1; it failed to produce a transmitted response. The schemata associated with each section depict the pathways of impulse propagation. The retrograde activation sequence shown in section X 2 of Figure 2 could not be produced at any coupling interval, since the alternate route of impulse propagation, namely limb b, had been interrupted by the transection maneuver. Figure 4 provides a similar example of alternation in response to critical premature stimulation. Again section S of Figure 4 displays the activation sequence of action potentials recorded from sites 1 FIGURES Failure to produce alternation following transection of accessory pathway, limb b. S represents control activation times in response to the primary stimulus of the pulse pair. X i represents the response to premature stimulation at an S-Xi coupling interval of 283 msec. X 2 represents complete block produced at an S-X2 coupling interval of 277 msec. The labeling of records and diagrams is the same as it is in Figure 2. Upstrokes of microelectrode records have been retouched. Voltage and time calibrations are 50 mv and 10 msec, respectively. and 2 in response to the primary stimulus of the pulse pair. The activation times are 5 msec and 9 msec for microelectrodes 1 and 2, respectively. Sections X[ and X 2 represent alternating patterns of response to the premature second stimulus of the pulse pair at an S-X coupling interval of 268 msec. The activation times for the records in X, are 20 msec and 26 msec, respectively, and 62 msec and 60 msec, respectively, in X 2. The diagram in each section indicates the hypothetical pathways of impulse propagation. Proof for the transmission of the impulse through the alternate pathway was derived from multiple impalements of limb b and the endocardial surface of the distal muscle mass with a third microelectrode during the reverse activation sequence seen in X 2. These data are shown in Figure 5 with an associated diagram depicting the location from which they were taken. It is evident that as the electrode was removed to a progressively more distal location in limb b and in the distal muscle mass toward the insertion of limb a, the activation time became progressively longer. Thus, the impulses that had been orthogradely blocked in limb a were traveling through limb b and the distal muscle mass and entering the distal end of limb a to activate it in reverse fashion. Examination of the action potential configurations in limb b also revealed that their durations were slightly less than those in limb a, and the configura- Circulatkm Research, Vol. XXXIV, March 1974

ALTERNATING PREMATURE EXCITATION TORS 409 3C 3B FIGURE 4 Alternating patterns of response to repetitive premature stimuli at a coupling interval of 268 msec. The preparation was taken from the right ventricle. S represents the control activation times of cells impaled by microelectrodes 1 and 2 in response to the primary stimulus of the pulse pair. X\ represents the orthograde conducted beat in limb a. X2 represents the alternate orthograde block with activation via the accessory pathway (limb b and distal muscle mass mmd). The activation times of microelectrodes 1 and 2 are 5 and 9 msec, 20 and 26 msec, and 62 and 60 msec for S, Xj, and X2, respectively. For nomenclature of data and diagrams see Figure 2. Upstrokes have been retouched in all records except for 1 in Xi. Voltage and time calibrations are 50 mv and 20 msec for each section. tion exhibited less plateau; thus, the cell generally appeared to repolarize more rapidly. The occurrence of this shorter action potential duration appeared to be important in the initiation and maintenance of the alternating response. Figure 6 provides evidence to explain the mechanism for alternating responses. Figure 6 illustrates a strip recording of successive pulse pairs taken from the experiment shown in Figure 2; the S-X coupling interval was 275 msec. There were apparent alternations of the preceding cycle length associated with the alternating activation sequences of the premature responses X, and X 2. The cycle length preceding the forward-blocked response X 2 (256 msec) was 30 msec longer than was the cycle length preceding the premature impulse X, that conducted orthogradely (226 msec). This alternation in cycle length was a result of the forward block in limb a with its subsequent delayed retrograde activation. Hence, the onset of activation began later than it would have begun if conduction had been orthograde. Consequently, its recovery encroached on the subsequent phase-4 interval. The shorter preceding cycle length altered the refractory period of the subsequent response sufficiently to permit the next premature response to be transmitted orthogradely. This find- Circulatim Research, Vol. XXXIV. March 1974 FIGURE S Sequential recordings from the accessory pathway during an orthograde block with delayed retrograde activation of limb a (see section X2 of Fig. 4). Tracings 3A, 3B, and 3C represent recordings taken from limb b and mmd during the X2 conduction pattern seen in Figure 4. The activation times of 3A, 3B, and 3C are 8, 17.5, and 40 msec, respectively. The corresponding diagram depicts the approximate location of the third microelectrode for each impalement. Arrows under each recording indicate the stimulus artifact. Upstrokes in 3B and 3C have been retouched. Voltage and time calibrations are 50 mv and 20 msec, respectively. ing agrees with the previously demonstrated "Ashman phenomenon" (4, 5). Discussion Using a simplified model of the multibranching S X, S X2 S X, S X2 S X, FIGURE 6 Continuous recording of transmembrane action potentials from the preparation illustrated in Figure 2. The record demonstrates alternation of the duration of electrical diastole, as is described in the discussion section of the text. Complexes (S) show response to the primary stimulus of the pulse pair. Xi represents the orthogradely conducted response in limb a to premature stimulation. X2 represents the alternate orthogradely blocked beat with delayed retrograde activation. Arrows represent points of stimulus application. The solid bars under each Xi and X2 response indicate the phase-4 recovery period from 95% repolarization to the subsequent electrical stimulation. The alternating diastolic recovery times are 256 and 226 msec, respectively. Voltage and time calibrations are 50 mv and 400 msec, respectively.

410 BANDURA, BRODY conduction system, we have successfully analyzed alternating patterns of conduction in response to bigeminal excitation. The data provide strong electrophysiological support for the previously proposed explanation for electrical alternations in response to premature stimulation (1). Production of a two-to-one forward block with retrograde activation appeared to depend on several factors, (l) The necessity of an alternative anatomic route is evidenced by Figures 3 and 5. Following transection of the subsidiary pathway, limb b (Fig. 3), the alternating response ceased and could not be produced under any conditions of premature stimulation. Figure 5 also elucidates the pathway of impulse propagation through the alternate route, limb b, during orthograde block in limb a. This figure is particularly important with respect to longitudinal dissociation. Longitudinal dissociation in limb a could produce a similar recorded event and such a mechanism might be responsible in some instances; however, in this series of experiments, entry from an alternative route rather than from longitudinal dissociation was apparently responsible for the reversal of the activation sequence. (2) The dependence on the duration of the preceding cycle length is clearly seen in Figure 6. This relationship has been implicated in the production of aberrant conduction (4, 6) and is very important in the genesis of the alternating response (l). Our data provide conclusive electrophysiological proof that the preceding cycle length plays a definitive role. Delayed retrograde activation following forward block produced a sufficiently delayed response in that portion of the specialized conduction system which resulted in encroachment on the subsequent diastolic rest period. Thus, the next premature impulse arriving at that region entered tissue that had a slightly shortened refractory period because of the foreshortened preceding cycle length and elicited a forward-propagated response. The conditions required to produce this phenomenon and the S-X coupling interval must be carefully adjusted to produce forward block in limb a without producing block in limb b. These conditions must be met before the alternating response will occur. (3) Disparity of recovery between two adjacent legs of the conduction system is required for this phenomenon to occur. This phenomenon has been previously reported (7, 8) and was demonstrated at the fascicular level in our preparation. It is an absolute necessity that some fiber(s) within the branching network have a sufficiently short refractory period to permit propagation of a premature impulse that is blocked by other branches. Finally, the system is very sensitive temporally. Furthermore, we limited our model preparation to a single bifurcating bundle to reduce the number of fibers that need to be examined as conduction pathways. Therefore, the phenomenon is not easily produced in a simple Y-shaped preparation. In some instances, maneuvers such as stretch or cooling to 30 C were attempted to produce asynchronous activity; however, we were unable to induce alternating responses with these techniques. Orthograde block with retrograde invasion was more easily produced in systems with 5-10 rami; however, we were unable to localize the routes of the alternate response without causing excessive movement and, thus, dislodging the microelectrode(s) impaled at the sites of interest. References 1. COHEN, S.E., LAU, S.H., SCHERLAG, B.J., AND DAMATO, A.N.: Alternate patterns of premature ventricular excitation during induced atrial bigeminy. Circulation 39:819-829,1969. 2. ANDERSON. G.J., GREENSPAN, K., BANDURA. J.P., AND FISCH, C: High fidelity recording of cardiac depolarization J Appl Physiol 29:401-405, 1970. 3. WENNEMARK.J.R., AND BANDURA,].P.: Microelectrode study of Wenckebach periodicity in canine Purkinje fibers. Am J Cardiol, in press. 4. ASHMAN, R., AND HILL. E.: Essentials of Electrocardiography, 2d ed. New York, Macmillan, 1941, p 23. 5. GREENSPAN. K., EDMANDS, R.E., AND FISCH, C: Effect of cycle length alteration on canine cardiac action potentials. Am J Physiol 212:1414-1420, 1967. 6. COHEN, S.E., LAU. S.H., HAFT, J.I., AND DAMATO, A.N.: Experimental production of aberrant ventricular conduction in man. Circulation 36:673-685,1967. 7. HOFFMAN, B.F., CRANEFIELD, P.F., AND STUCKEY, J.H.: Concealed conduction. Circ Res 9:194-203,1961. 8. BANDURA, J.P.: Non-uniformity of impulse propagation in the specialized Purkinje fiber system of the canine heart. Ph.D. Thesis, Indiana University, Indianapolis, 1972. Circulation Research, VoL XXXIV, March 1974