Technical Comments

Shin Takasawa et al. (1) challenge the role of inositol 1,4,5-triphosphate (IP3) as an intracellular second messenger that mobilizes Ca21 in pancreatic P cells. They found that cyclic adenosine diphosphate-ribose (cADP-ribose), but not IP3, releases Ca2+ from islet microsomes. It is difficult to reconcile their results with many studies that establish IP3 as an intracellular Ca2+mobilizing second messenger in pancreatic 1 cells (2). Confronted with such provocative results, we performed a series of experiments to compare the Ca2+-mobilizing actions of the two second messengers in 13 cells. We used clonal insulin-secreting RINm5F cells and cells obtained from ob/ob mice, where more than 95% of the islet cells correspond to normal 13 cells. The cells were permeabilized by high-voltage electric discharges, a technique that creates clean holes in the plasma membrane, but leaves intracellular Ca2+-storing organelles in situ and undamaged (3). We found pronounced Ca2+ release when IP3 was added to insulin-secreting RINm5F cells (Fig. 1A) or to pancreatic 13 cells from ob/ob mice (Fig. 1B). In marked contrast to the results in the report by Takasawa et al., there was no Ca2+ release after the addition of cADP-ribose. In experiments with 13 cells, we first added a low dose of caffeine to sensitize the release mechanism that presumably might respond to cADP-ribose. After maximal Ca2+ release by IP3, further Ca2+ was released from 13 cells by the sulfhydryl reagent thimerosal which, as we have shown before, indicates the possible existence of a Ca2+-induced Ca2+ release mechanism in 1 cells (4). With the use of intact 1 cells, we looked for the caffeinesensitive intracellular Ca2+ pool on which cADP-ribose presumably acts. In small clusters of 1 cells that had been loaded with Fura-2, in the absence of extracellular Ca2+, there was marked Ca2+ release from intracellular stores by IP3-forming agonists, whereas caffeine-induced Ca2+ release was absent (5). Detailed studies using caffeine indicate that, in the 1 cell, caffeine increases intracellular free Ca2+ concentration ([Ca2+1i1) by a mechanism unrelated to its intracellular Ca2+-mobilizing action (5). Furthermore, we used the patch-clamp technique to monitor the Ca2+-sensitive K+ conductance in 13 cells for detection of any small release of Ca2+ following the addition of cADP-ribose. This method is more sensitive than fluorimetric methods, and it has been used to record increases in [Ca2+ ]i after intracellular application of 1P3 and guanosine 5 '-0(3-thiotriphosphate) (GTP-y-S) in the pancreatic 1 cell (6). Even so, we were unable to obtain evidence of Ca2+ release from intracellular stores after the addition of cADP-ribose in 14 out of 14 cells, whereas formation of IP3 potently induced release of Ca2+ (Fig. 2). These results raise several questions. First, what might be the reason for the absence of IP3-induced Ca2+ release as reported by Takasawa et al.? Their procedure of purification of microsomes might have adversely affected the IP3-sensitive Ca2+ stores, which can be easily damaged during fractionation (7). Second, why was cADP-ribose-induced Ca2+ release seen in their preparation but not in ours? We do not have a definitive answer to this question. It is possible that cADP-riboseinduced Ca2+ release in the 13 cell is small in magnitude and requires rigorous experimental conditions to be detected. Alternatively, the source of Ca2+ released in the experiments of Takasawa et al. might be cells other than 1 cells. It should be recalled that the islets used by Takasawa et al. contain a large proportion of cells that are not 13 cells. We avoided this potential problem by using a tumor cell line and an almost pure preparation of normal 13 cells as well as by performing experiments on single mouse 13 cells. A possible explanation for our negative results with cADP-ribose could be that our preparation of the compound was inactive. However, precautions were taken to ensure that this was not the case. By using cADP-ribose from different sources, who verified the activity of the substance in other cell systems, we guarded against the possibility that the lack of effect in our experimental system was not simply a result of an inactive batch of the compound. Moreover, cADP-ribose seems to be a stable compound (8). With the aim of taking a more physiological experimental approach (that is, using cells instead of isolated organelles), we deliberately did not exactly duplicate the experiments conducted by Takasawa et al. Hence, there remains a possibility that some experimental factors might have adversely affected the cADP-ribose-sensitive release mechanism in our system. Takasawa et al. also demonstrate that extracts of islets incubated in a high con-