p57^KIP2 Expression in Normal Islet Cells and in Hyperinsulinism of Infancy

Archival tissues from 15 pancreatectomized HI patients were obtained from five clinical centers (Table 1). In all, the diagnosis of HI was made according to accepted criteria (11,12). A total of 11 male and 4 female patients, age range 2 weeks to 13 months, were included in the study. Of these patients, 11 had focal disease; for these patients, analyses compared tissue inside and outside the lesion in each patient. Four patients had the diffuse form of HI. Of the 15 subjects, 12 were previously reported (Table 1).

A total of 15 control pancreatic samples were included in the study; 12 were obtained from autopsies carried out between 1988 and 1998 in seven male subjects and five female subjects aged 17 weeks gestation to 3 years. These samples consist of a random subgroup of the previously published control population (19). All fetuses and infants died as a result of diseases not related to the pancreas, and in all, autopsies were done for medical reasons according to accepted procedures at each individual institution. All dysmorphic subjects were excluded, as were subjects with known chromosomal abnormalities or genetic syndromes. We only included subjects in whom the autopsy was performed within 24 h of death. Adult control pancreas samples were obtained from two pancreas donors and from a patient who underwent partial pancreatectomy for insulinoma. All samples were stained with hematoxylin and eosin and screened for adequate quantity of tissue, normal morphology, and good tissue preservation.

Sections (5 μm thick) were prepared from archival paraffin-embedded tissue, placed on SuperFrost Plus glass slides (Menzel-Glaser, Germany), and left to dry at 37°C overnight. Slides were deparaffinized in xylene and then rehydrated in serial concentrations of alcohol (100, 90, and 80%) and double-distilled water. Antigen retrieval was carried out as described by Cattoretti et al. (33). Briefly, slides were microwaved in 0.01 mol/l citrate buffer (pH 6) for 3 min at full power until boiling and then for 15 min at 20% power. Slides were left to cool at room temperature (RT) for 30 min. They were then blocked by nonimmune serum for 10 min at RT before application of each primary antibody.

Slides were double-stained for p57^KIP2 and each of the four major pancreatic hormones (insulin, glucagon, somatostatin, and pancreatic polypeptide). Antibodies, incubation times, detection systems, and substrates are listed in Table 2. To prevent cross-reactivity of the two detection systems, avidin-biotin blocking kit (catalogue no. 00-4303; Zymed) was used before incubation with anti-hormone antibody. As a negative control, slides underwent the same procedure but were incubated with phosphate-buffered saline without anti-p57^KIP2 antibody. Each batch included a negative control.

Sections were double-stained for IGF-II and insulin. Antibodies, incubation times, detection systems, and substrates are listed in Table 2. Cross-reactivity of the anti–IGF-II antibody with proinsulin or insulin was excluded by preabsorbing the antibody with the two peptides overnight, a procedure that did not affect the intensity of the stain. Pretreatment, incubation times, and conditions were similar for all slides in each step described.

All slides were coded, and ≥1,000 hormone-positive cells were counted under high magnification (400×). The frequency of p57^KIP2/hormone-positive cells was determined and expressed as percent hormone-positive cells (mean ± SE).

Altogether, 11 different fields were assessed under high magnification (400×). Two images were produced from each field, using different filters in the same settings of microscope and camera (L-600 and Coolpix 950, respectively; Nikon)

Images were analyzed using Image-Pro Plus software (Media Cybernetics). Total stained area was expressed in pixels, and total integrated optic density was expressed in arbitrary optic density units. β-Cell IGF-II protein content was expressed as a ratio of IGF-II integrated optical density (IOD) to insulin-stained area. Counting criteria and software settings were identical for all slides.

Results for p57^KIP2 expression in different age-groups were analyzed using the Kruskal-Wallis nonparametric analysis of variance test, whereas the HI groups were compared with control subjects using the Mann-Whitney U test. The Wilcoxon paired nonparametric test was used to compare IGF-II expression inside and outside the lesion in focal HI.

p57^KIP2 Expression was demonstrated as dark brown nuclear staining, whereas pancreatic hormones where stained red in cell cytoplasm (Figs. 1A–H). Within the pancreas, p57^KIP2 was specifically localized to the endocrine cells. Very few p57^KIP2 positive cells were seen in the acinar tissue, and none were found among the duct cells (Fig. 1A). In the control pancreas, β-cells demonstrated the highest frequency of p57^KIP2 expression (34.9 ± 2.7%), whereas other islet cell types stained for p57^KIP2 with much lower frequency (∼1–3%) (Figs. 1B–D and 2). No significant change in p57^KIP2-positive β-cell proportion was observed during the different developmental stages of the human pancreas (Figs. 1E and 3).

The percentage of p57^KIP2-positive β-cells in diffuse HI was similar to that in the control subjects (Figs. 1F and 4). Complete loss of p57^KIP2 staining was clearly demonstrated inside the affected area of focal HI (Figs. 1G–H). Interestingly, a tendency toward increased p57^KIP2 expression was observed in β-cells outside the affected area of focal HI compared with diffuse HI and control subjects, although this did not reach statistical significance (Fig. 4).

Because the increased proliferation previously documented in focal HI could be caused by increased expression of the maternally imprinted IGF-II gene, we quantitated the IGF-II protein content of β-cells inside and outside of the lesion in eight patients. IGF-II staining was identified exclusively in β-cell cytoplasm, both inside and outside the focal lesion (Figs. 1I–L). Outside the lesion, all cells stained with IGF-II also stained for insulin; however, only ∼27% of the insulin-stained area also stained for IGF-II. In normal β-cells from the same age group, a similar IGF-II distribution was seen (data not shown). To quantify the amount of IGF-II within the β-cell mass, the intensity of IGF-II staining was expressed as the ratio of IGF-II IOD to insulin-stained area. In focal HI, IGF-II staining within the focal lesion was slightly increased when compared with that outside of the lesion in the same patient (7.5 ± 0.9 vs. 5.7 ± 0.6 arbitrary units; P < 0.04) (Fig. 5).

We have shown that in the pancreas, p57^KIP2 is expressed almost exclusively in the endocrine cells, and that within the islets, expression is primarily localized to β-cells. During development, the proportion of β-cells expressing p57^KIP2 does not appear to vary, and in diffuse HI, the proportion of β-cells expressing the protein is not different from control subjects of a similar age group. In contrast, p57^KIP2 is not expressed by β-cells within the focal HI lesion. IGF-II expression was also seen primarily in β-cells, and staining was increased within the lesion of focal HI when compared with β-cells in the unaffected region from the same patient.

p57^KIP2 Was originally described in 1995; it is a Cdk inhibitor causing cell cycle arrest and accumulation of cells in the G₁ phase. It has been shown to bind to cyclin/Cdk complexes in a cyclin-dependent manner and inhibit their activity (26,28). The gene is located within a cluster of imprinted genes in humans and mice, with the maternal allele primarily expressed (27,34).

The finding that p57^KIP2 expression is limited to the endocrine portion of the pancreas explains the low gene expression reported in human whole-pancreas mRNA preparations (26,28). It is also consistent with the nature of the islet cell population, especially β-cells, in that it is postmitotic and terminally differentiated. This may partially account for the failure of β-cell regeneration after exposure to harmful factors such as hyperglycemia and hyperlipidemia, a phenomenon that may have important implications in the pathogenesis of type 2 diabetes. The low expression in other islet cells compared with β-cells suggests that the latter represent a higher differentiation stage. The very low p57^KIP2 expression in acinar and ductular cells may provide a possible explanation for the proliferative capacity those cells retain (35).

The finding that p57^KIP2 expression does not change during different stages of development is unexpected, because the proportion of β-cells undergoing proliferation does change during fetal development, as we previously reported (19). However, the proliferation frequency ranged from ∼5% at gestational week 17 to 0% in the adult pancreas. Because only 30–40% of β-cells are p57^KIP2 positive, it is likely that the methods used are not sufficiently sensitive to detect small absolute differences in the low proportion of cells undergoing proliferation at the different developmental stages. In two samples triple-stained for insulin, p57^KIP2 and Ki67 (a nuclear marker of cell proliferation), we were unable to identify any cells that expressed both nuclear antigens, suggesting that cells expressing p57^KIP2 do not undergo proliferation (data not shown).

Focal HI is caused by specific loss within affected β-cells of a portion of the maternal allele of Ch11p, which contains the p57^KIP2 gene (20,36). p57^KIP2 Has been shown to be paternally imprinted in several tissues (27). Our finding of loss of p57^KIP2 expression within the focal HI lesion suggests that the gene is also imprinted in human β-cells. This relatively simple immunohistologic stain can be used to confirm loss of heterozygosity (LOH) of the maternal allele in these lesions and may be of use in differentiating focal HI from other forms of hyperinsulinism. Similarly, the same stain may be useful in confirming LOH for this region in other tissues in diseases such as BWS and in certain tumors, such as Wilms’ tumor (37), adrenocortical tumors (31), and lung cancers (30), as long as normal expression and imprinting is confirmed for each tissue type.

The finding of loss of p57^KIP2 expression in focal HI may explain the increased β-cell proliferation in the adenomatous portion compared with unaffected pancreas and diffuse HI (19,38). Other adjacent genes, some imprinted, located in the chromosomal region lost in focal HI may play an additive or synergistic role in inducing β-cell proliferation. Furthermore, other mechanisms must also regulate β-cell proliferation because even within the focal lesion, proliferation rates are only ∼6% (19).

p57^KIP2-Positive cells tended to be more frequent outside the focal lesion compared with cells from control and diffuse HI subjects of the same age group, although this difference did not reach statistical significance. β-Cells outside the lesion are exposed to hypoglycemia and high insulin concentration released from the lesion. This leads to suppressed metabolic activity and decreased cytoplasmic volume (18) and may also result in decreased proliferation mediated by high p57^KIP2 expression.

IGF-II is located in the same region on chromosome 11 but is maternally imprinted and has been associated with increased β-cell proliferation (23). Others have demonstrated that pancreatic IGF-II expression is limited to β-cells (39). In our samples, positive staining was also seen exclusively in β-cells, and a similar staining pattern was found in pancreases from all age groups (data not shown). To estimate the IGF-II content of affected and unaffected β-cells in focal HI, we developed a method using quantitative image analysis of immunofluorescence. By comparing IGF-II IOD to insulin area, we obtained an estimate of the quantity of IGF-II protein as a function of β-cell area. Because we previously showed that IGF-II is expressed exclusively in β-cells, this calculation defines the amount of protein within the β-cells. Insulin content (as defined by insulin IOD) was not used because this reflects the secretory state of the β-cells, which clearly differs inside and outside the lesion. β-Cells that have undergone complete degranulation were not included in this calculation; however, examination of the sections indicated that very few if any of the cells within the lesion are completely degranulated, and those outside the lesion are uniformly heavily granulated, reflecting the suppressed secretion in these functionally normal cells.

Our finding of increased IGF-II within the focal lesion, relative to outside the lesion, supports the hypothesis that IGF-II may be involved in the regulation of focal proliferation. The mechanism causing this increased IGF-II expression is unknown. Paternal disomy has been documented in focal HI (40), and this increase in gene dosage may be responsible for increased expression. Alternatively, H19, a paternally imprinted gene thought to regulate IGF-II expression (29), may play a critical role. However, H19 expression and imprinting in β-cells has not yet been proven. It is also possible that p57^KIP2 may have a direct effect on IGF-II expression, although such a connection has not yet been established. Furthermore, the relatively small absolute difference raises the possibility that this may be a secondary phenomenon.

In conclusion, we have demonstrated that p57^KIP2 is expressed and is paternally imprinted in human pancreatic β-cells. Levels of expression do not appear to parallel changes in rates of β-cell proliferation during development, whereas decreased expression in focal HI is associated with increased rates of proliferation and increased IGF-II expression. Manipulation of p57^KIP2 expression in β-cells may provide a mechanism by which the rate of proliferation can be modulated, and thus this gene may be a potential therapeutic target for reversing the β-cell failure observed in diabetes.

The authors would like to thank Drs. Feilberg Jorgensen and Claus Fenger for providing us with tissue preparations from some of their HI patients. This study was supported by Israel Science Foundation Grant 512-98 and the Hadassah Diabetes Center. Grant BMH4-CT98-3284, awarded as part of the European Commission Biomed 2 program, provided support for international collaborations. Updated information on neonatal hyperinsulinism is available on the European Network for Research into Hyperinsulinism (ENRHI) website at http://www.phhi.u-net.com.