Isolated islets are extremely fragile in vitro
due to cell death from both apoptosis and necrosis [16
]. If KU-32 were to be used as a clinical intervention for patients with diabetic peripheral neuropathy, any possible cytotoxicity to islets must be avoided. Initial cytotoxicity studies were performed in 5
mM glucose with doses of KU-32 from 0.03 to 30μ
M. After 24 hours of exposure to KU-32, there was no measureable cell loss to the islets exposed to any of the doses tested (). Replicate studies in high glucose also showed no signs of toxicity (results not shown). These initial toxicology screens, using the alamarBlue assay, provided general information concerning changes in cell numbers, but the natural heterogeneity in the size of native islets (ranging from 50–400μ
m diameters), along with the natural donor-to-donor variability, resulted in large variations, as noted by the error bars in . Thus, we conducted additional viability studies using apoptosis/necrosis fluorophores and confocal microscopy.
Figure 1 KU-32 improved viability of isolated human islets. (a) Toxicity studies using alamarBlue as an indicator of cell number shows no statistically significant toxic effect of KU-32 on human islets after a 24-hour exposure to 7 half-log doses (n = 6 replicates). (more ...)
(left) provides a typical example of cell death associated with the vehicle-treated islets. The green staining indicates cell death due to apoptosis, and red is associated with necrotic cell death. In contrast, islets exposed to KU-32 showed less cell death (, right). Islets maintained in culture have an increasing percentage of cell death, as we and others have shown [16
]. Exposure to KU-32 halted the time-dependent increase in cell death. illustrates the progressive loss of viable cells/islets in the control group, while continued exposure to 1μ
M KU-32 for the duration of the experiment maintained cell death at around 10% in the treated group. Shorter time points were also tested starting with a 1-hour exposure. However, exposure durations of less than 48 hours failed to significantly alter cell viability. Of interest, the majority of cells died due to apoptosis. shows that far fewer apoptotic cells were found in islets exposed to KU-32 compared to the controls. In contrast the amount of cell death due to necrosis was identical in both groups, but relatively low.
Further confirmation that KU-32 protected cells from apoptosis came from oxygen consumption studies. At each time point tested, oxygen consumption was greater per islet when exposed to KU-32 (). Although the oxygen consumption per islet was increased with KU-32, when the oxygen consumption data were normalized to the percentage of viable cells within each islet (using the data shown in ), group differences in oxygen consumption were ablated (data not shown). In summary, the increased oxygen consumption associated with the KU-32 exposed islets was likely due to the increased number of live cells rather than changes in the cellular respiration.
KU-32 increased oxygen consumption per islet. Exposure to KU-32 increased oxygen consumption per islet through an 18-hour period. Measurements were made in triplicate wells with 13–33 islet equivalents (IE)/well.
To determine whether KU-32 altered function in β
cells, perifusion experiments were performed on human islets. Islets were first exposed to low glucose (3
mM) for 90 minutes. Subsequently, the perfusate was switched to high glucose (20
mM) and the supernatant collected. While there was no significant effect by KU-32 on the insulin secreted in low-glucose conditions, there was significantly more insulin released in response to high glucose from the islets preexposed to KU-32 (). Repeated static incubation studies showed similar results with statistically higher insulin release in the KU-32 exposed islets (). In low glucose (3
mM), there was no effect of KU-32 on insulin secretion. The improved insulin secretion was demonstrated with a minimum of 16 hours of exposure prior to the assay. One-hour exposures to KU-32 failed to stimulate an improvement in insulin secretion even in high-glucose conditions (data not shown).
Figure 3 KU-32 increased insulin secretion. (a) In perifusion experiments, exposure to KU-32 increased insulin secretion in response to high-glucose. Initially, islets were maintained in 3mM glucose. At time 0, the perfusate was switched to 17.3mM (more ...)
The molecular mechanism responsible for the neuroprotective effects of KU-32 depends on inhibition of Hsp90 and subsequent expression of Hsp70 [4
]. Thus, we measured Hsp70 levels in human islets exposed to KU-32 for 24 hours. presents a typical western blot recognizing Hsp70 and the housekeeping protein, actin. Surprisingly, there was no difference in Hsp70 levels between the vehicle- (V) and KU-32-treated (T) islets when normalized to actin levels (). Isolated islets normally express a high level of Hsp70; therefore, KU-32 may not have had an effect on quiescent islets. We stressed the islets by incubating them in hypoxic conditions with acidic media while exposed to either vehicle or KU-32. Control and KU-32 treated islets were placed in stress media (see methods) and cycled between 5 and 20% CO2
. Under these conditions, islets had more Hsp70 in the stressed conditions (although not statistically different from basal conditions), but instead of KU-32 increasing the Hsp70 levels further, it returned Hsp70 back toward baseline levels ().
Figure 4 KU-32 failed to induce an increase in Hsp70 levels. (a) A typical western blot is shown with Hsp70 from samples of islets treated with vehicle (V) or KU-32 (T). MW size indicated on left. (b) Densitometry measurements of repeated western blots indicate (more ...)
In order to determine whether the protective action of KU-32 on islets existed in vivo, diabetic Lepr
mice were treated with weekly injections of KU-32. Elevations in blood glucose levels in these mice generally occur at ages 4 to 8 weeks of age. Therefore, once chronic hyperglycemia was confirmed, weekly KU-32 injections were administered starting at 10 weeks of age and continuing for 8 weeks. Blood glucose levels were not different between the vehicle and KU-32-treated animals (). Likewise, there was no difference in the serum insulin levels in the KU-32 versus vehicle-treated animals at the termination of the study (). Likewise, there were no signs of toxicity in the KU-32-treated animals.
Figure 5 KU-32 did not alter blood glucose and insulin levels in vivo. (a) Weekly blood glucose measurements demonstrate that both groups of animals were diabetic, and levels were not affected by KU-32 administration (n = 12 vehicle- and 16 KU-32-treated mice). (more ...)
At the termination of the study, the pancreata were removed and analyzed using immunofluorescence. Islet morphology was clearly different in the diabetic mice. Islets were either intact, or they had large areas devoid of α
, or δ
cells, termed disrupted islets [21
]. Finally, there was a third category of islets that was composed of scattered endocrine cells, defining an area that had once been an islet. These morphological categories could be viewed with both immunofluorescence and immunohistochemistry techniques (). Both vehicle- and KU-32-treated animals had islets that fell into the three categories and there was no difference in the distribution of islets within each category between groups ().
Figure 6 Lepr
mice had islets that were categorized into three types. (a) Immunofluorescence (left) and immunohistochemistry (right) illustrate the 3 types of islets that were identified in the Lepr
mice. Upper panels show intact islets, middle panels (more ...)
From the serial sections, islet area was measured, and there was a statistically significant increase in total area of the endocrine-stained cells in the KU-32-treated mice (). In summary, more total pancreatic area was composed of α, β, or δ cells with KU-32 treatment. However, the density of islets per microscopic field in the pancreas tail decreased in the KU-32-treated animals ().
Figure 7 KU-32 improved the total islet area, but not the islet density or cell composition. (a) Treatment with KU-32 increased the total area of the pancreas identified by immunohistochemistry as endocrine cells. (*indicates P < 0.007, n = 40 islet from (more ...)
We hypothesized that KU-32 would spare some of the diabetes-induced loss of endocrine cells within the islets, by sparing the β-cells specifically. Thus, we compared the percentage of α, β, and δ cells within the islets. These measurements were only made on intact and disrupted islets (not scattered cells). shows that there was no difference between the two groups, with both groups of islets showing a dominant percentage of β-cells ().
While there was no difference in the percentage of β-cell to other endocrine cells with KU-32 administration, there was a clear difference in the intensity of the insulin staining, which is indicative of the insulin level/cell. Insulin staining intensity was greatest in the KU-32-treated animals when calculated per islet () and per β-cell ().
Figure 8 Insulin staining intensity per cell was higher with KU-32 treatment. Representative examples are shown of insulin immunohistochemistry for islets from vehicle-treated and KU-32 administered mice. (b) Insulin staining intensity values (brown) were background-subtracted (more ...)