Intracellular pH Plays a Critical Role in Glucose-Induced Time-Dependent Potentiation of Insulin Release in Rat Islets

Chemicals were purchased from the following companies: Fisher Scientific (Fair Lawn, NJ), Sigma (St. Louis, MO), Mallinkrodt Chemicals (Paris, KY), Molecular Probes (Eugene, OR), and Bio Rad (Hercules, CA). Radioactive insulin was purchased from New England Nuclear Life Science Products (Boston, MA), and cylinders of 95% O₂/5% CO₂ were purchased from Empire Airgas (Elmira, NY).

Male Wistar rats aged 2–3 months were obtained from Harlan (Indianapolis, IN). Pancreatic islets were isolated according to the collagenase digestion method described by Lacy and Kostianovsky (21), with slight modifications. The animals were fed ad libitum until they were killed.

Krebs-Ringer bicarbonate (KRB) solution of pH 7.4 buffered with 10 mmol/l HEPES (KRBH) was used for islet isolation and for the major part of the incubations. A bicarbonate-buffered solution of pH 7.4 containing 25 mmol/l HCO₃⁻ and constantly gassed with 95% O₂ and 5% CO₂ (KRB) was used during the priming period where indicated to demonstrate the difference in glucose-induced priming in the two different buffer systems. The standard composition of KRBH solution was 128.8 mmol/l NaCl, 4.8 mmol/l KCl, 1.2 mmol/l KH₂PO₄, 1.2 mmol/l MgSO₄, 2.5 mmol/l CaCl₂, 5 mmol/l NaHCO₃⁻, and 10 mmol/l HEPES. The composition of the KRB solution was similar except that it contained 25 mmol/l NaHCO₃⁻ and no HEPES. All media were maintained at pH 7.4, except in the experiments in which the medium pH was raised to 8.3 to produce intracellular alkalinization. In the Ca²⁺-free KRB or KRBH solutions used, Ca²⁺ was omitted from the medium, and 5 mmol/l EGTA was added to ensure stringent Ca²⁺-free conditions. The basal glucose concentration used was 2.8 mmol/l, whereas the priming glucose concentration was 11.1 mmol/l in Ca²⁺-free conditions and 16.7 mmol/l in the presence of Ca²⁺. These are the optimal concentrations of glucose for the stimulation of insulin secretion in the absence and presence of Ca²⁺, respectively (22). Stimulation of islets in the test period, as distinct from the priming period, was always done with 16.7 mmol/l glucose in the presence of Ca²⁺. Bovine serum albumin (0.1%) was present in the medium during all incubations. To produce intracellular acidification, we added DMA to the KRB/KRBH solution to a final concentration of 40 μmol/l where indicated. Acetate (at a final concentration of 40 mmol/l) was used as an alternative acidifying agent. Intracellular alkalinization was produced, where indicated, by 1) increasing the medium pH to 8.3, 2) adding 5 mmol/l ammonium chloride to the medium, or 3) substituting the Cl⁻ in the medium with gluconate⁻. After adding the acidifying/alkalinizing agents to the media, the medium pH was readjusted to 7.4 (except when intracellular alkalinization was produced by increasing the medium pH). Similar media were used during the measurement of intracellular pH, except that the KRB media were not gassed and contained either 5 or 25 mmol/l HCO₃⁻ as indicated.

Freshly isolated islets were preincubated for 30 min in basal KRBH solution with 2.8 mmol/l glucose. Islets were then subjected to a 45-min priming period, with or without Ca²⁺ as indicated. The concentration of glucose, the buffer system, and the presence of acidifying or alkalinizing agents during the priming period varied according to the experiment, as detailed in the text. After the priming period, the islets were washed and rested in basal KRBH for 20 min, followed by the test stimulation with 16.7 mmol/l glucose in KRBH for 20 min. At the end of this test period, samples were collected to measure insulin secretion and islet insulin content. In each static incubation experiment, three to five replicate batches of five size-matched islets were used for each priming condition. The number of times that each experiment was repeated is denoted by n in the figure legends.

Radioimmunoassay was used to measure the amount of insulin released during the final stimulation and the total content of insulin in the islets. Values are expressed as fractional release, as a percentage of the total content per 20 min. Absolute release rates (pg · islet⁻¹ · min⁻¹ during the 20-min period) are presented in the figure legends.

Intracellular pH in islets was measured under various conditions using the pH-sensitive dye carboxy-SNARF-AM (carboxy seminaphthorohodofluor-acetyloxymethyl ester, or 126208-13-7 Spiro[7H-benzo[c]xanthene-7,1′(3H)-isobenzofuran]-ar′-carboxylic acid, 3-(acetyloxy)-10-(dimethylamino)-3′-oxo-,(acetyloxy)methyl ester), an agent that has been used successfully to measure intracellular pH in multicellular systems (23,24,25,26). As described by Muller-Borer et al. (23), this dye can be used for accurate quantitative measurement of intracellular pH in thick tissue specimens up to a depth of 300 μm. Even though emission fluorescence intensity may vary as a result of possible attenuation of emitted light in the deeper cell layers and/or possible differences in dye concentration among cells, the ratio between the two emission fluorescence intensities remains constantly proportionate to the pH (23), making SNARF-AM a suitable agent to measure pH in thick tissues. In our experiments, preferentially picked small islets of <100 μm in diameter were used to ensure uniform dye loading. Islets were loaded with carboxy-SNARF-AM at 37°C for 1 h, at a concentration of 5 μmol/l. Dye-loading in islets was monitored visually, and uniformly stained islets were picked for pH measurements. Using a Nikon Diaphot fluorescence microscope, the loaded islets were excited at 514 nm, and the emission fluorescence was measured at two different wavelengths, 580 and 640 nm. Intracellular pH was calculated from the ratio between the fluorescence at 580 and 640 nm. The calibration curve was derived using islets with their intracellular pH fixed at known values ranging from 6.0 to 8.0 in KRBH containing 100 mmol/l K⁺ and 20 μmol/l nigericin. The pH of the media in which islets were maintained during the pH measurements and of the solutions added (e.g., HEPES, DMA) had been adjusted to 7.4. A representative recording for each experiment is shown in the figures. The number of such recordings done for each condition is denoted by n, and mean intracellular pH values from n recordings are indicated in the figure legends.

Values were expressed as mean ± SE, and groups were compared using the paired t test.

In HEPES-buffered medium in the absence of Ca²⁺, islets previously exposed to 11.1 mmol/l glucose produced higher insulin release in response to subsequent stimulation with 16.7 mmol/l glucose than the islets previously exposed to 2.8 mmol/l glucose. In other words, glucose induced strong TDP under stringent Ca²⁺-free conditions (Fig. 1). Because this result was unexpected and contradicted the consensus in the literature, follow-up studies were performed to compare the ability of glucose to prime in the HEPES-buffered media (KRBH) and HCO₃⁻-buffered media (KRB).

When paired islets were pretreated with 11.1 mmol/l glucose in KRBH and KRB under strict Ca²⁺-free conditions, TDP was observed in KRBH but not in KRB (Fig. 2). Thus, the ability of glucose to induce priming under Ca²⁺-free conditions seems to be determined by the difference between HEPES and bicarbonate as buffer systems. As demonstrated later and described in the discussion, the presence of HEPES results in a lower intracellular pH in the islets than that maintained in HCO₃⁻-buffered media. This has been reported before in mouse islets (8,9,10). We demonstrated a similar pH-lowering effect of HEPES in rat islets by measuring intracellular pH in the presence and absence of HEPES. Intracellular pH measurements were performed in whole islets using the fluorescent dye carboxy SNARF-AM, an agent that has been used successfully to measure intracellular pH in multicellular systems (23,24,25,26). Islets loaded with 5 μmol/l SNARF-AM were maintained in KRBH (or other media as indicated), and the intracellular pH was recorded at 37°C during a 20-min period. The presence of HEPES in the medium uniformly resulted in a lower intracellular pH. The basal intracellular pH in islets maintained in regular KRBH solution (containing 10 mmol/l HEPES, 5 mmol/l HCO₃⁻, and 2.8 mmol/l glucose) ranged from 7.0 to 7.2, with a mean value of 7.1 ± 0.02 (n = 60). In HEPES-free KRB with 5 mmol/l HCO₃⁻, the basal pH was higher (7.4 ± 0.06; n = 25). When 10 mmol/l HEPES was added to HEPES-free KRB with 5 mmol/l HCO₃⁻, the intracellular pH decreased to between 6.9 and 7.1, indicating the acidifying property of HEPES (Fig. 3A). A similar change was observed when 25 mmol/l HEPES was added to a HEPES-free KRB medium containing 25 mmol/l HCO₃⁻ (Fig. 3B). The acidifying effect of HEPES was similar in both the presence (Figs. 3A and B) and absence of Ca²⁺ (Figs. 3C and D). Previous studies in mouse islets (8,9) suggested that glucose-induced Ca²⁺ influx results in a decrease of intracellular pH, which is evident in the presence of HEPES, even though an earlier study reported no change in intracellular pH on removal of Ca²⁺ or when the glucose concentration was raised (10). In our hands, glucose did produce a decrease in islet pH but to a lesser degree (data not shown), possibly as a result of a difference in the intracellular buffering mechanisms between rat and mouse islets.

These results demonstrate that islets experience a significant decrease of intracellular pH in the presence of HEPES. As shown previously, glucose induces significant priming in the presence of HEPES even under stringent Ca²⁺-free conditions, whereas no such priming is observed in the absence of HEPES (Fig. 2). To test whether it is the low intracellular pH that acts as the major permissive factor for glucose-induced priming, we experimentally modulated the intracellular pH in islets and determined its effects on priming. DMA, an inhibitor of the Na⁺/H⁺ exchanger in the β-cell, was selected as a suitable acidifying agent. The Na⁺/H⁺ exchanger is one of the major buffering mechanisms in the β-cell, which acts by extruding excess H⁺ ions in exchange for medium Na⁺. Inhibition of this exchanger by DMA results in accumulation of H⁺ ions within the cell, producing a rapid decrease in intracellular pH. The ability of DMA to decrease intracellular pH in islets is demonstrated in Fig. 4. The addition of 40 μmol/l DMA to KRBH or KRB steadily decreased the islet intracellular pH in both the presence and absence of Ca²⁺. In KRBH, the pH changed from 7.1 ± 0.03 to 6.9 ± 0.04 in the presence of Ca²⁺ (n = 21) and from 7. 1 ± 0.02 to 6.9 ± 0.03 (n = 12) in the absence of Ca²⁺. In KRB containing 5 mmol/l HCO₃⁻, the pH changed from 7.4 ± 0.04 to 7.2 ± 0.02 (n = 6). In Ca²⁺-free KRB containing 25 mmol/l HCO₃⁻ (the medium in which glucose does not induce priming), the pH changed from 7.7 ± 0.06 to 7.2 ± 0.03 upon the addition of DMA (n = 7).

In previous studies (3,4,5,6,7) as well as ours, glucose failed to induce TDP in Ca²⁺-free KRB (Fig. 2). However, when the intracellular pH in islets was decreased using 40 μmol/l DMA during the Ca²⁺-free priming period, TDP was observed in KRB where it was formerly absent. A comparison of the effects of DMA on glucose-induced priming in Ca²⁺-free KRBH and Ca²⁺-free KRB is shown in Fig. 5. In islets primed with 11.1 mmol/l glucose, the presence of DMA produced a significant difference in the primed response, although no significant change was observed in unprimed islets maintained at basal glucose. The presence of DMA resulted in the development of glucose-induced TDP in Ca²⁺-free KRB. In KRBH where glucose normally produces priming, the presence of DMA significantly enhanced the magnitude of the response (Fig. 5). Acetate, a weak acid that produces a reversible decrease in the intracellular pH, also enhanced the glucose-induced priming in KRBH (data not shown). Thus, a low intracellular pH is favorable for glucose-induced priming. Neither DMA nor acetate alone induced TDP (in the absence of a stimulatory concentration of glucose), indicating that this effect of intracellular pH on priming is dependent on the presence of a priming agent. In the presence of Ca²⁺ where glucose induces priming in both KRBH and KRB, the degree of TDP was higher in KRBH (Fig. 6), providing additional evidence of the favorable effect of low intracellular pH on priming.

These results clearly demonstrate the pH-lowering effect of HEPES and the ability of glucose to induce better priming in the presence of a low intracellular pH. In situations in which glucose induces strong priming, such as in KRBH (with or without Ca²⁺) or KRB with DMA, the intracellular pH was maintained at or below 7.2. If a low intracellular pH is critical for glucose-induced priming, then an increase in the intracellular pH should inhibit the ability of glucose to induce TDP. To explore this possibility, we monitored glucose-induced priming in KRBH under conditions of intracellular alkalinization. During the priming period, the intracellular pH in islets was increased by raising the medium pH to 8.3. Increasing the medium pH to 8.3 resulted in a corresponding intracellular pH of 7.7 ± 0.04 (n = 7). Increasing the medium pH to 8.0, 8.1, or 8.2 did not produce significant changes in the intracellular pH, indicating the effective buffering capacity in the β-cell. As expected, glucose-induced priming was significantly inhibited when the intracellular pH was increased to 7.7 (Fig. 7). Alkalinization produced by other traditional methods such as the addition of 5 mmol/l NH₄Cl to the medium or the substitution of the medium Cl⁻ with gluconate⁻ also produced significant inhibition of glucose-induced priming (data not shown).

The results of this study demonstrate that the induction of TDP by glucose is critically dependent on the intracellular pH and not directly on the presence of extracellular Ca²⁺. The first observation that eventually led to this conclusion was the ability of glucose to induce TDP under stringent Ca²⁺-free conditions, in surprising contrast to the results of all previous studies on the subject. In previous investigations, conducted with rat islets in Ca²⁺-free HCO₃⁻-buffered medium, the use of 0.5–1.0 mmol/l EGTA during the priming period was sufficient to completely abolish TDP (3,4,5,6,7). In our study, however, glucose consistently induced TDP in a Ca²⁺-free HEPES-buffered medium, despite the presence of EGTA in as high a concentration as 5 mmol/l. The fold-change over the unprimed response was higher in the absence of Ca²⁺, despite the possible attenuation of β-cell performance caused by a 45-min exposure to Ca²⁺-free conditions.

An important difference between HEPES and HCO₃⁻ buffering systems is that HEPES buffer maintains a lower intracellular pH in the islets than does HCO₃⁻ buffer, as reported in previous studies (8,9,10,11). This effect is particularly evident in the presence of a stimulatory concentration of glucose. Shepherd and Henquin (8) demonstrated that the intracellular pH in mouse islets was 0.03 units lower in a HEPES-buffered medium than in an HCO₃⁻-buffered medium, and this was further decreased by 0.15 units when stimulated by high glucose. Similar results were reported in a previous study (10) in which the addition of HCO₃⁻ to a HEPES-buffered medium resulted in the increase of intracellular pH in mouse islets by 0.08–0.09 units. In rat islets, glucose-induced increase in intracellular pH was reversed by the absence of HCO₃⁻ (11). The proposed mechanism for this difference in intracellular pH between HEPES and HCO₃⁻-buffered media is as follows. The initial rise in intracellular pH caused by glucose as a result of the mitochondrial H⁺ uptake during the oxidation of pyruvate (27) is later reversed by the influx of Ca²⁺, which tends to lower the pH. The β-cell attempts to compensate for this fall in intracellular pH by activating the intracellular buffering systems, predominantly the Na⁺/H⁺ exchanger and the Cl⁻/HCO₃⁻ exchanger. In a HEPES-buffered medium in which the concentration of HCO₃⁻ ions is significantly low, the buffering is carried out mainly by the Na⁺/H⁺ exchanger. In a HCO₃⁻-buffered medium, however, both exchangers are effective, resulting in overcompensation for the Ca²⁺-induced acidification and stabilizing the intracellular pH at a higher point (8).

The acidifying effect of HEPES was confirmed in our study, in which the addition of HEPES to a HEPES-free HCO₃⁻ medium produced a dramatic decrease in the intracellular pH. The basal intracellular pH in our HEPES-buffered medium was 7.1 ± 0.02, significantly lower than that in an HCO₃⁻-buffered medium. Islets in HEPES-free KRB maintained an intracellular pH above 7.4, which was consistently decreased by the addition of HEPES. As in previous studies, we also noticed the tendency of high glucose to increase the intracellular pH in an HCO₃⁻-buffered medium and to decrease it in a HEPES-buffered medium, albeit to a lesser degree. The presence of Ca²⁺, however, did not seem to produce any significant change in the intracellular pH of unstimulated or glucose-stimulated islets. The presence of HEPES uniformly produced a lower intracellular pH in islets, perhaps to a slightly greater extent in the presence of Ca²⁺. Because the acidifying effect of HEPES is consistently seen even in the absence of glucose stimulation, it is evident that HEPES lowers intracellular pH by mechanisms other than the inadequate compensation for glucose-induced acidification described in earlier studies (8,20). Strong evidence for such a pH-lowering property of HEPES comes from the fact that the addition of HEPES brings the intracellular pH down to 7.0–7.1 even in the presence of 25 mmol/l HCO₃⁻, which would maintain a fully operative HCO₃⁻/Cl⁻ exchanger (Figs. 3B and D). Hence, although an inoperative HCO₃⁻/Cl⁻ exchanger contributes to the lower islet pH maintained in media buffered by HEPES alone, there must be another major mechanism whereby HEPES decreases intracellular pH independent of glucose-stimulation, medium Ca²⁺, or HCO₃⁻.

In view of this major difference between the two buffer systems, we considered the possibility that the lower intracellular pH in islets incubated in a HEPES-buffered medium may be permissive for the induction of TDP by glucose. If this hypothesis were true, then lowering the intracellular pH by experimental means in islets maintained in a HCO₃⁻-buffered medium should allow glucose to induce TDP. Hence, we investigated the priming ability of glucose in the presence of DMA, a powerful acidifying agent that decreases the intracellular pH by inhibiting the Na⁺/H⁺ exchanger. As shown in Fig. 4, DMA induced a persistent decrease in the intracellular pH in islets by 0.2 units or more, indicating this to be a suitable acidifying agent for this purpose. Glucose-induced priming is not normally evident in HCO₃⁻-buffered media in the absence of Ca²⁺, as shown in several previous studies (3,4,5,6,7) as well as this one (Fig. 2). In the presence of DMA, however, priming was induced by 11.1 mmol/l glucose in a stringently Ca²⁺-free HCO₃⁻-buffered medium containing 5 mmol/l EGTA (Fig. 5). In HEPES buffer in which glucose consistently induces priming regardless of the presence of extracellular Ca²⁺, the magnitude of the primed response was significantly enhanced by the presence of DMA (Fig. 5) as well as by the acidification produced by acetate (data not shown). Furthermore, glucose-induced priming was consistently inhibited by intracellular alkalinization produced by increasing the medium pH (Fig. 7) or other methods in both the presence and absence of extracellular Ca²⁺. All of these findings are strong evidence for the dependence of glucose-induced TDP on the intracellular pH and establish that glucose requires an appropriately low intracellular pH to induce TDP.

Under Ca²⁺-free conditions, the critical intracellular pH below which glucose can induce TDP in rat islets is close to 7.2. In situations in which glucose can induce strong priming in the absence of Ca²⁺, such as in KRBH or in KRB acidified by DMA, the islet intracellular pH is always maintained at or below 7.2 (equivalent to an H⁺ concentration of 63 nmol/l). In the presence of Ca²⁺, glucose can induce TDP in KRB as well as KRBH, as long as the islet intracellular pH is maintained at an appropriate level. When the intracellular pH is driven above 7.7 (equivalent to an H⁺ concentration of 20 nmol/l) by experimental alkalinizing methods, the TDP normally present in KRBH is inhibited irrespective of the presence of Ca²⁺. When the H⁺ ion concentration in the cell is maintained above 63 nmol/l (i.e., at pH 7.2 and lower), TDP occurs irrespective of the presence of Ca²⁺. Therefore, it can be concluded that a major factor that allows glucose to induce TDP is an adequate concentration of intracellular H⁺ (at or higher than 63 nmol/l). In the range of H⁺ concentration between 20 and 63 nmol/l (pH 7.7–7.2), however, Ca²⁺ plays a major role in priming. In a HEPES-free HCO₃⁻-buffered medium, when the islet H⁺ concentration is maintained within the range of 20–63 nmol/l, glucose induces TDP in the presence of Ca²⁺ but not in its absence, albeit to a lesser degree than in the presence of HEPES. In this situation, the presence of Ca²⁺ overrides the deficit in H⁺ ions. This brings us to the conclusion that depending on the circumstances, either a high intracellular H⁺ concentration or normal intracellular Ca²⁺ concentrations are required for priming to occur. In other words, either one of the two cations is permissive for glucose-induced TDP and both may be acting on the same site(s). Furthermore, the combination of normal Ca²⁺ conditions and an experimentally increased H⁺ concentration (e.g., in the presence of DMA) results in enhanced glucose-induced TDP. This suggests that they both can act simultaneously to optimize TDP.

In seeking the mechanism by which glucose induces TDP, it seems appropriate to look for reactions that are both pH- and Ca²⁺-sensitive. It is widely known that certain mitochondrial enzymes are activated by Ca²⁺ and that energy and/or signal production after exposure to stimulatory glucose concentrations would be impaired in the absence of extracellular Ca²⁺. Among such enzymes are isocitrate dehydrogenase (ICDH) and α-ketoglutarate dehydrogenase (αKGDH). Although ICDH and αKGDH can act in the absence of Ca²⁺, their activity is markedly enhanced by Ca²⁺. Importantly, αKGDH is also activated by low pH (28,29,30), and glutamate dehydrogenase, another enzyme that enhances nutrient metabolism by providing anaplerotic input, is also sensitive to pH (31). Thus, one possibility for the results observed in this study is that Ca²⁺- and pH-activated enzymes are responsible for the production of the signals that control TDP. Under normal physiological conditions, Ca²⁺ plays the major role in driving these reactions. Under nonphysiological conditions in the absence of Ca²⁺, activation by low pH can take over the role of Ca²⁺ and provide a sufficient activating force to maintain the signals to TDP. Activation by both Ca²⁺ and H⁺ leads to optimal TDP.

The importance of intracellular pH in the induction of priming by glucose is not surprising, considering all of the documented effects of intracellular pH on nutrient-induced acute insulin release. The influence of intracellular pH on insulin release has been the subject of a number of studies, the majority of the reports indicating that a lower intracellular pH favors increased insulin release. In studies on rat islets, Best et al. (14,15,16) used weak acids (acetate, propionate, and formate) and amiloride to decrease intracellular pH and NH₄Cl to increase it, while measuring intracellular pH with BCECF. The weak acids caused a rapid fall in the pH corresponding to 0.2 pH units, followed by a slow return to the basal value. Amiloride, an inhibitor of the Na⁺/H⁺ exchanger in the β-cell, produced a rapid fall in intracellular pH without recovery, and NH₄Cl induced a rise in pH of ∼0.15 pH units. Nutrient-induced insulin release was observed to be enhanced with intracellular acidification, as were inositol lipid metabolism (14), reduction of 45Ca²⁺ and 86Rb⁺ outflow (15), and increase in intracellular Ca²⁺ (16). The opposite effects were observed on intracellular alkalinization with NH₄Cl. Similar results were reported in a number of other studies performed on mouse islets and insulinoma cell lines as well as on rat islets (12,13,17,18,19,20).

As demonstrated by these studies, nutrient-induced insulin release and related functions are favorably influenced by a decrease in intracellular pH. The data from our investigation show that priming, another important nutrient-induced effect in the β-cell, is also dependent on the intracellular pH. Work is currently under way to identify the specific pH-sensitive messengers involved in priming. Understanding the mechanisms of TDP is important in designing therapeutic approaches to type 2 diabetes. In certain patients who have diabetes and whose acute insulin secretory response is severely impaired, TDP remains intact and the impaired insulin response to glucose can be corrected by inducing TDP (32,33,34). In other patients in whom both the TDP and acute secretory response are impaired, correction of the defect in TDP may lead to correction of the acute secretory response. Hence, documenting the mechanisms of TDP could help in the design of new therapeutic approaches to type 2 diabetes, and the pH-sensitivity of TDP described in this study would be a useful tool in uncovering the underlying mechanisms.

This work was supported by National Institutes of Health Grants DK-42063 and DK-54243 to G.W.G.S. and a Graduate Research Assistantship from Cornell University College of Veterinary Medicine to S.C.G. (on study leave from the University of Peradeniya, Sri Lanka). Our special thanks go to Dr. Thomas Schermerhorn, Dr. Mitsuhiko Noda, Dr. Mitsuhisa Komatsu, and Dr. Susanne G. Straub for valuable help in developing the research techniques and solving problems.