Spatial Nonuniformity of Excitation–Contraction Coupling Causes Arrhythmogenic Ca2+ Waves in Rat Cardiac Muscle

Ca2+ waves underlying triggered propagated contractions (TPCs) are initiated in damaged regions in cardiac muscle and cause arrhythmias. We studied Ca2+ waves underlying TPCs in rat cardiac trabeculae under experimental conditions that simulate the functional nonuniformity caused by local mechanical or ischemic local damage of myocardium. A mechanical discontinuity along the trabeculae was created by exposing the preparation to a small jet of solution with a composition that reduces excitation–contraction coupling (ECC) in myocytes within that segment. The jet solution contained either caffeine (5 mmol/L), 2,3-butanedione monoxime (BDM; 20 mmol/L), or low Ca2+ concentration ([Ca2+]; 0.2 mmol/L). Force was measured with a silicon strain gauge and sarcomere length with laser diffraction techniques in 15 trabeculae. Simultaneously, [Ca2+]i was measured locally using epifluorescence of Fura-2. The jet of solution was applied perpendicularly to a small muscle region (200 to 300 &mgr;m) at constant flow. When the jet contained caffeine, BDM, or low [Ca2+], during the stimulated twitch, muscle-twitch force decreased and the sarcomeres in the exposed segment were stretched by shortening normal regions outside the jet. Typical protocols for TPC induction (7.5 s-2.5 Hz stimulus trains at 23°C; [Ca2+]o=2.0 mmol/L) reproducibly generated Ca2+ waves that arose from the border between shortening and stretched regions. Such Ca2+ waves started during force-relaxation of the last stimulated twitch of the train and propagated (0.2 to 2.8 mm/sec) into segments both inside and outside of the jet. Arrhythmias, in the form of nondriven rhythmic activity, were induced when the amplitude of the Ca2+-wave was increased by raising [Ca2+]o. Arrhythmias disappeared rapidly when uniformity of ECC throughout the muscle was restored by turning the jet off. These results show, for the first time, that nonuniform ECC can cause Ca2+ waves underlying TPCs and suggest that Ca2+ dissociated from myofilaments plays an important role in the initiation of Ca2+ waves.

I schemic and failing hearts are both prone to ventricular arrhythmias and commonly show regional differences in contractile strength caused by heterogeneous impairment of excitation-contraction coupling (ECC). It is generally accepted that lethal arrhythmias are frequently associated with alterations of the excitation step of ECC. 1 It is less well known what role nonuniform ECC 2 plays in initiating arrhythmias. 3,4 We have previously investigated the triggered propagated contractions (TPCs) phenomenon in rat cardiac trabeculae. TPCs probably result from local damage and the ensuing nonuniform ECC. 4 TPCs consist of local sarcomere shortening 5-7 associated with a [Ca 2ϩ ] i transient that propagates in a wave-like manner along the muscle. 8 -10 Ca 2ϩ waves underlying TPCs cause delayed after-depolarizations (DADs) and triggered arrhythmias. 6,7,10 In the model of damaged muscle, TPCs and underlying Ca 2ϩ waves started invariably in regions located near the dissected end of the muscle or near cut branches. 4,5 The regions bordering the damaged areas exhibit elevated cytosolic and sarcoplasmic reticulum (SR)-Ca 2ϩ and constitute a source of nonuniformity in ECC. 6 However, a detailed study of the role of these regions in the initiation of arrhythmogenic Ca 2ϩ waves is hampered by the difficulty in controlling the extent and severity of damage, and as such neither sarcomere length (SL) nor [Ca 2ϩ ] i can be measured reliably.
Here, we developed a novel model of controlled nonuniformity in rat trabeculae. Using this model, we show that controlled initiation of Ca 2ϩ waves underlying TPCs can trigger nondriven regular spontaneous contractions in cardiac muscle. The initiation of arrhythmogenic Ca 2ϩ waves can be explained by nonuniform ECC and Ca 2ϩ -dissociation from the contractile filaments occurring during relaxation of nonuniform cardiac muscle. right ventricle of Lewis Brown Norway rats [5][6][7][8][9][10][11][12][13][14] (Charles River Canada, Saint Constant, QC, Canada) and mounted between a motor arm and force (F) transducer in a bath perfused by HEPES solution on an inverted microscope. SL was measured by laser diffraction techniques 12 ( Figure 1A).
Measurement of [Ca 2ϩ ] i has been described previously. 8 -11 Briefly, Fura-2 salt was microinjected iontophoretically into the trabecula. 11 Excitation light of 340, 360, or 380 nm was used and fluorescence was collected using an image intensified CCD camera (IIC) at 30 frames/s to assess local [Ca 2ϩ ] i ( Figure 1A). 11 We calculated [Ca 2ϩ ] i in a region of interest along the trabeculae from the calibrated ratio of F 360/ /F 380 (see Miura et al for details 9 and Figure Is, available online at http://circres.ahajournals.org).

Reduction of Local Contraction
To produce nonuniform ECC, a restricted region was exposed to a small jet of solution (Ϸ0.06 mL/min) that had been directed perpendicularly to a small muscle segment (300 m; Figure 1) using a syringe pump connected to a glass pipette (Ϸ100 m diameter) ( Figure 1A and 1B; see also supplemental Movie 1 in the online data supplement). The jet was positioned with respect to the muscle using a neutral colorant ( Figure 1B) or fluorescein (Ͻ0.01 mg/mL). To reduce contraction in the exposed region by modified ECC, the jet solution was composed of standard HEPES solution containing either: (1) Caffeine (CF; 5 mmol/L); (2) 2,3-butanedione monoxime (BDM; 20 mmol/L); or (3)

Data Analysis
Data were expressed as meanϮSEM. Statistical analysis was performed using ANOVA followed by a Post-hoc test. Differences were considered significant when PϽ0.05 (see online data supplement).

Nonuniformity and Sarcomere Mechanics
The jet reached one short muscle segment (Ϸ300 m) ( Figure 1A and 1B). Figure 1C shows that the fluid flow from the pipette using a solution with composition similar to the bath solution (HEPES) had no effect on F or SL by itself.
When a jet containing either BDM ( Figures 1C and 2A), caffeine (Figure 2A), or low [Ca 2ϩ ] (Figure 2A) was applied to the stimulated trabeculae, sarcomere stretch rapidly replaced the normal active shortening in the exposed segment ( Figure 2B), whereas peak force (F/F max ) decreased (Ϫ11Ϯ3% with low [Ca 2ϩ ] jet (nϭ5), Ϫ28Ϯ5% with caffeine (nϭ6), and Ϫ36Ϯ7% with BDM (nϭ5) ( Figure 2C; Table). Low [Ca 2ϩ ] jet solution did not affect resting SL (SL o ), whereas caffeine and BDM slightly decreased (Ϫ2.2Ϯ0.7%) and increased (ϩ3.1Ϯ0.6%) SL o , respectively. All effects were rapidly reversible ( Figure 1C). Sarcomere dynamics along muscles exposed to a jet revealed 3 distinct regions during the twitch ( Figure 2B): (1) a region located Ͼ200 m from the jet where sarcomeres exhibited typical shortening (see [1] in Figure 2B); (2) the segment exposed to the jet where sarcomeres were stretched (see [2] in Figure 2B); (3) in a region between [1] and [2] sarcomeres shortened early during the twitch and then were stretched although less than in segment [2]; we denoted this region [3] the Border Zone (BZ). The BZ extended 1 to 2 cell lengths (100 to 200 m) beyond the jet-exposed region ( Figure 2B). The diffraction pattern of sarcomeres in BZ illuminated by a Ϸ150 m diameter laser beam showed a clear single peak during both shortening and lengthening; similar changes in regional sarcomere dynamics were observed in caffeine and low [Ca 2ϩ ] jet experiments (data not shown).

Nonuniformity and [Ca 2؉ ] i Transients
In contrast to the similarity of the effects of the various jet solutions on sarcomere dynamics, jets of caffeine, BDM, or low [Ca 2ϩ ] o solution had distinct effects on [Ca 2ϩ ] i ( Figure  3A, 3B, and 3C respectively). Robust electrically driven   Figure 3C). Average [Ca 2ϩ ] i -transients decreased by Ϫ36Ϯ6, Ϫ17Ϯ3, and Ϫ37Ϯ7%, and [Ca 2ϩ ] diast changed by ϩ82Ϯ28, ϩ48Ϯ13, Ϫ44Ϯ14% respectively in segments exposed to caffeine (nϭ9), BDM (nϭ9), and low [Ca 2ϩ ] jet (nϭ9), compared with [Ca 2ϩ ] i -transients and [Ca 2ϩ ] diast outside the jet (see Table and online data supplement). The [Ca 2ϩ ] iϪ changes were smaller in BZ, consistent with a gradient between regions caused by diffusion of the contents of the jet. Ca 2ϩ waves (caffeine: nϭ9; BDM: nϭ9; and low [Ca 2ϩ ] jet : nϭ9 [15 muscles]) started systematically in the BZ after the decline of the last stimulated Ca 2ϩ transient. These waves propagated into the regions outside and, in the cases of BDM and low [Ca 2ϩ ] jetϪ , inside the jet exposed region ( Figure 3; supplemental Movie 2 in the online data supplement). Figure 3B (BDM jet) clearly shows 2 initiation sites of Ca 2ϩ waves in the BZ and symmetric propagation into regions outside and inside the jet.
Lowering [Ca 2ϩ ] to 0.2 mmol/L in the jet also triggered Ca 2ϩ waves if [Ca 2ϩ ] o in the main solution was slightly increased (from 2 to 2.5 mmol/L). These waves started in the BZ and propagated inside and outside the jet exposed region at different velocities, with waves in the jet region being the slowest ( Figure 3C). Similar observations were made in low [Ca 2ϩ ] jet -exposed muscles. In caffeine-exposed muscles, Ca 2ϩ waves did not propagate into the jet region ( Figure 3A). This precluded determination of the site of origin of Ca 2ϩ waves, but the earliest Ca 2ϩ surge was again observed in the BZ. These observations suggest strongly that the initial surge in [Ca 2ϩ ] i in the BZ initiated Ca 2ϩ waves. Ca 2ϩ waves induced by exposure to either BDM, caffeine, or low [Ca 2ϩ ] jet started late during relaxation, ie, 40Ϯ5 ms after the maximal rate of sarcomere shortening in the stretched segment and 35Ϯ15 ms after twitch force had decline below 30% of peak force (F 30 in Figure 4C; Table).

Propagation of Ca 2؉ Waves
Propagation velocity of the Ca 2ϩ waves (V prop ) ( Figure 5), outside (nϭ27) and inside (nϭ13) the jet region, ranged from 0.2 to 2.8 mm/s, ie, comparable to Ca 2ϩ waves observed in our damaged muscle studies (0.34 to 5.47 mm/s). 8 -10 V prop correlated with the [Ca 2ϩ ] i increase seen in the BZ during the initial Ca 2ϩ surge (⌬C W ) (rϭ0.66, PϽ0.0001; nϭ40; Figure  5C). Furthermore, V prop correlated with the amplitude of the waves both inside and outside the jet 8 (data not shown). These Effect of local jet exposure on F and SL. A, Typical F and SL tracings in the jet-exposed segment before (gray) and during (black) exposure to a jet of HEPES (control), low [Ca 2ϩ ] jet (LC), caffeine (CF), and BDM solution. F is normalized to the maximal force (F max ). B, Spatial effects of local exposure to BDM on SL patterns; SL tracings recorded from 3 different segments along the muscle: [1] outside the jet, [2] inside the jet, [3] in a border zone (BZ) between [1] and [2] (BZ). The traces compare effects of BDM exposure on resting SL (Ⅲ) and SL during peak twitch (F) in [1], [2], and [3]. C, Summary of effects of HEPES (nϭ7), caffeine (CF; nϭ6), BDM (nϭ5), and low [Ca 2ϩ ] jet (LC; nϭ5) on Force (F/Fmax), resting SL (SL 0 ), and SL at peak-twitch (DSL peak ) in the absence and presence of the jet-flow. F/Fmax and SL in the segment that had been exposed to the jet for 5 minutes (ON) are compared with F and SL before exposure to the jet-flow (OFF pre ; †PϽ0.05) and compared with F and SL both (*PϽ0.01) before and 5 minutes after cessation of the jet (OFF- Frequently, waves propagating inside the jet region collided with the wave arriving from the opposite BZ and then terminated (5/7 waves; see white arrow in Figure 4B). All waves moving inside low [Ca 2ϩ ] jet propagated slowly over 100 to 200 m with a steep decline in amplitude (see b and c, in Figure 5A and 5B), sometimes gradually slowing down before terminating. This contrasted with waves observed in the region outside the low [Ca 2ϩ ] jet ; such waves propagated rapidly with little decline in amplitude (see a in Figure 5A and 5B).

Nonuniformity and Arrhythmias
Nonuniformity of ECC created by the jet induced nondriven rhythmic activity. The arrhythmia consisted of spontaneous twitches at regular intervals starting after an after-contraction that followed the last stimulated contraction. The arrhythmia continued until the next stimulus train (7.5 s; Figure 6 [muscle exposed to BDM]). The intervals between nondriven contractions were usually slightly longer than those of the preceding stimulus train. As shown in Figure 6, these arrhythmias terminated abruptly when the jet was turned off and the uniformity of ECC restored.

The Model of Nonuniform ECC
We have used in this study a novel model of nonuniform ECC in cardiac muscle to study arrhythmogenic Ca 2ϩ waves underlying TPCs in cardiac muscle. We created nonuniformity in ECC by exposing a small segment of the muscle to caffeine, BDM, or low [Ca 2ϩ ] o . We expected that (1) low [Ca 2ϩ ] jet would reduce Ca 2ϩ current and the Ca 2ϩ transient attributable to ECC 15 despite increased SR-Ca 2ϩ content, 16 (2) caffeine would open SR-Ca 2ϩ release channels and thereby deplete the SR, 15,17,18 and (3) BDM would modestly affect Ca 2ϩ transients attributable to ECC 19,20 because of a reduced SR-Ca 2ϩ content 21,22 and potentiation of RyR 23 and inhibited cross-bridge cycling. 20 Consistent with these expectations, the amplitude of stimulated Ca 2ϩ transients decreased dramatically in regions exposed to caffeine and low [Ca 2ϩ ] jet but only slightly with BDM ( Figure 3).
Each of these perturbations reduced muscle force because of creation of a muscle segment which developed less twitch force than the normal cells remote from the jet, as is witnessed by stretch of the weakened sarcomeres in the jet by the fully activated sarcomeres outside the exposed region ( Figure 2C). These regions were connected mechanically by a border zone of 1 to 2 cells, where the sarcomeres first contracted and, then, were stretched ( Figure 2B, region [3]. The diffraction pattern of sarcomeres in the BZ illuminated by an Ϸ150 m diameter laser beam (ie, 1.5 cell lengths) showed a clear single peak during both shortening and lengthening, strongly suggesting that sarcomere contraction in BZ was also partially suppressed, probably owing to diffusion of the contents of each jet solution.
These observations confirm that this method causes nonuniform ECC along the muscle and affects specifically a F onset , force at the moment of onset of initial ͓Ca 2ϩ ͔ i rise in the BZ (expressed as percentage of the last stimulated twitch force during relaxation). t(C W ϪC T ), latency between peaks of stimulated Ca 2ϩ -transient (C T ) and initial Ca 2ϩ rise (C W ) during wave initiation in the BZ; t(ϪdF/dt max ϪF onset ), latency between time of ϪdF/dt max and that of onset of the Ca 2ϩ wave (F onset ); t(F30ϪF onset ), latency between time of 30% twitch force during twitch relaxation (͓F 30 ͔: at which the rate of sarcomere shortening in the stretched segment was maximal; see online data supplement) and F onset . ⌬C W, increment of ͓Ca 2ϩ ͔ I , and C W /C T , relative amplitude of ͓Ca 2ϩ ͔ i in BZ during initiation of Ca 2ϩ waves. V prop , propagation velocity of Ca 2ϩ -waves inside or outside the jet-exposed segment. Values are expressed as meanϮSEM. *nϭ7;
selected region of the trabeculae, which results in regional decrease of contractile force of the sarcomeres.

Nonuniform ECC and Initiation of Ca 2؉ Waves
Ca 2ϩ waves and TPCs have been closely related to Ca 2ϩ overload in damaged regions and the resultant nonuniformity of muscle contraction. 4 However, in that model it is difficult to investigate underlying mechanisms because damage is difficult to control. This study shows clearly that Ca 2ϩ waves are reversibly initiated in regions without damage and, more specifically, from the BZ, in which contraction is partially suppressed. The common effect of the 3 protocols was to suppress contraction and reduce sarcomere force ( Figure 2) Figure 4A). Once initiated, these Ca 2ϩ waves traveled from the region with the localized [Ca 2ϩ ] i rise proving that this region constitutes the initiation site for Ca 2ϩ waves.
The initiating Ca 2ϩ -surge took place late during twitch relaxation when both F and free Ca 2ϩ in the cytosol had decayed by 70% to 80% ( Figure 4B). By this time the SR-Ca 2ϩ channels have partially recovered 24 and are able to support Ca 2ϩ -induced Ca 2ϩ release (CICR) and Ca 2ϩ wave generation. 15 However, the delay between the stimulusmoment and Ca 2ϩ -surge makes it highly unlikely that Ca 2ϩ entry via L-type Ca 2ϩ -channels causes CICR from the SR and the initial Ca 2ϩ surge. Furthermore, Ca 2ϩ waves never started in jet-exposed regions where sarcomeres were maximally stretched even if the amplitude of the stimulated Ca 2ϩ transients witnessed a robust SR-Ca 2ϩ content. 15 Ca 2ϩ waves never started simultaneously with the peak of stretch ( Figure  4C) making it unlikely that a stretch-related mechanism such as activation of Gd 3ϩ -sensitive stretch-activated channels 10 -14 is involved in the initial Ca 2ϩ -surge.
It has been shown that caffeine 25 and BDM increase the open probability of SR-Ca 2ϩ release channels (RyR), 21,23 whereas low [Ca 2ϩ ] jet could theoretically do so by increasing the SR-Ca 2ϩ load. 16 However, several observations make it unlikely that potentiation of RyR caused spontaneous Ca 2ϩ release in the BZ: (1) the same interventions cause no spontaneous Ca 2ϩ release in uniform muscle 26 or myocytes 21 ; (2) the effect of these interventions must have been maximal in the jet exposed region, whereas Ca 2ϩ waves never started in this region.

A Novel Mechanism Underlying Arrhythmias
We suggest that Ca 2ϩ that is bound to Troponin C (TnC) during this phase of twitch underlies the Ca 2ϩ -surge that initiates Ca 2ϩ waves. 27 It is well known that a quick release of Ca 2ϩ -activated cardiac muscle induces a surge of Ca 2ϩ ions dissociating from myofilaments 11,28 because of rapid reduction of the TnC affinity for Ca 2ϩ owing to a reduction in the number of Ca 2ϩ -activated cross-bridges. 29 The concept of quick-release-induced Ca 2ϩ dissociation from TnC, demonstrated in uniform cardiac muscle, is applicable to the chain of cells in the nonuniform muscle exposed to the jet. Rapid sarcomere shortening during the force decline occurred both in the jet region and in the BZ, but led only to a Ca 2ϩ surge and Ca 2ϩ wave initiation in the BZ (Figure 4B), making it probable that quick-release-induced Ca 2ϩ dissociation from TnC caused by the decline of force in the shortening BZ sarcomeres led to the local Ca 2ϩ surge. 11,28 -32 The region inside the jet, where ECC was all but abolished, probably contained either little TnC-Ca 2ϩ (caffeine or low [Ca 2ϩ ] jet ) or only few Ca 2ϩ activated force generating cross-bridges (BDM), which would render a quick release of this region unable to generate a Ca 2ϩ surge and Ca 2ϩ wave. The BZ, on the other hand, could generate a Ca 2ϩ surge that is large enough to induce local CICR and thus a Ca 2ϩ wave even if only a fraction of TnC 33,34 were occupied with Ca 2ϩ . 4 A quantitative analysis of the relation between quick release dynamics and onset of the Ca 2ϩ surge or waves may shed further light on this mechanism of initiation of Ca 2ϩ waves. Our present study allows for an estimate of the latency (Ϸ40 ms) between rapid sarcomere shortening and the initial rise of [Ca 2ϩ ] i during the Ca 2ϩ surge in a small BZ region Ϸ100 m ( Figure 4A), although the precision of this estimate  is limited by the temporal resolution of the IIC camera (33 ms/frame). Furthermore, focusing the laser beam to Յ100 to 200 m for SL measurements reduced signal to noise ratio thereby precluding more accurate measurement of the time of quick release of BZ sarcomeres. Force derived parameters of twitch relaxation (ϪdF/dt max and F 30 ; see Figure 4B and 4C) also preceded the onset of Ca 2ϩ waves in the BZ by minimally Ϸ35 ms (Table). This measure probably still overestimated the true latency because the onset of the Ca 2ϩ surge was measured from the nadir between the last stimulated Ca 2ϩ transient and the initial [Ca 2ϩ ] i rise in the BZ (Figure 4 and 5). Nevertheless, both estimates allow a minimum delay of 30 to 40 ms between Ca 2ϩ dissociation from TnC and SR-Ca 2ϩ release. Despite methodological limitations, the inverse relationship between amplitude of the initiating Ca 2ϩ transient (C W ) and latency of the Ca 2ϩ transient (t[C T ϪC W ]) (see Table  and the online data supplement) is consistent with the hypothesis that magnitude and rate of SR-Ca 2ϩ release depends on the amount of triggering Ca 2ϩ released from TnC and responsiveness of the SR-Ca 2ϩ release channels. 8

Propagation of Ca 2؉ Waves
Ca 2ϩ waves propagated in this model at slightly lower velocity (0.2 to 2.8 mm/s) than those of previous studies of regionally damaged muscles. 9,10,35 In this study the amplification of the Ca 2ϩ signal by SR-Ca 2ϩ release required for propagation may have been lower in the absence of Ca 2ϩ loading of the muscle by damaged areas. 4 -6,8,13,34,36 -38 A lower cellular Ca 2ϩ load would explain why Ca 2ϩ waves were both smaller and V prop lower and propagated with a gradual decline of their amplitude and disappeared after a few hundred m. Differences of propagation pattern of Ca 2ϩ waves in caffeine, BDM, or low [Ca 2ϩ ] jet (Figure 3) are consistent with the assumption that regional SR function and SR Ca 2ϩ content determine V prop . 8 Ca 2ϩ waves did propagate into regions with normal SR-Ca 2ϩ release such as inside the BDM jet ( Figure 3B) and propagated slowly inside the low [Ca 2ϩ ] jet where [Ca 2ϩ ] i is reduced ( Figure 3C). 8 Ca 2ϩ waves did not propagate through regions exposed to caffeine ( Figure  3A), which is consistent with the effect of caffeine to deplete SR-Ca 2ϩ required for wave propagation. 8

Implication: Nonuniform ECC, Ca 2؉ Waves, and Arrhythmias
One striking finding of this study is the arrhythmogenic nature of mechanically nonuniform myocardium independent of any damage, which is clearly caused by the induction of Ca 2ϩ waves ( Figure 6). Diastolic [Ca 2ϩ ] i transients are known to cause transient depolarizations caused by electrogenic Na ϩ -Ca 2ϩ exchange and Ca 2ϩ -sensitive inward currents 7,10,39 -41 and alter action potential configuration. 32 Several compartments which normally only release Ca 2ϩ during the cardiac cycle in response to the action potential can release Ca 2ϩ spontaneously during diastole. Such compartments include the SR in which abnormal Ca 2ϩ storage may be arrhythmogenic. [42][43][44] In addition, increased open probability of the SR-Ca 2ϩ release channels 45 owing to channel gene mutation 40,42,46 or to posttranslational channel changes in heart failure may cause arrhythmias. 43 In this study of controlled nonuniformity of muscle contraction, we identify nonuniform ECC in cardiac muscle for the first time as a possible arrhythmogenic mechanism. Whether this mechanism plays a role in the wall of the ventricles remains to be proven, although the arrangement of the cardiac wall in muscle fascicles, which transmit force longitudinally and therefore are subject to comparable constraints as trabeculae in this study, makes this possibility highly likely. This mechanism may contribute to arrhythmogenesis in diseased heart where nonuniform segmental wall motion 2,46 may result from ischemia, nonuniform electrical activation, or nonuniform adrenergic activation. 46 Figure 6. Nonuniform ECC causes arrhythmias. A continuous chart recording of force showing that stimulus trains during local exposure to BDM (gray bars above the tracings) repeatedly induced arrhythmias. An expanded force tracing showing that spontaneous contractions were both preceded and followed by after-contractions induced by the stimulus train. OFF (arrow) indicates when the jet was turned off; S, stimulus trains (2.5Hz-7.5s) repeated every 15 s. [Ca 2ϩ ] o ϭ3.5 mmol/L; temperature 25.8°C. Exp000519ArBDM.