2022 Volume 69 Issue 12 Pages 1407-1414
The development of pancreatic cancer (PC) is associated with worsening of glucose tolerance. However, there is limited information about the effects of PC on islet morphology. The aim of this study was to elucidate changes in alpha and beta cell mass in patients with PC. We enrolled 30 autopsy cases with death due to PC (9 with diabetes; DM) and 31 age- and BMI-matched autopsy cases without PC (controls, 12 with DM). Tumor-free pancreatic sections were stained for insulin and glucagon, and fractional beta cell (BCA) and alpha cell area (ACA) were quantified. In addition, expression of de-differentiation markers, i.e., ALDH1A3 and UCN3, was qualitatively evaluated. The pancreas of subjects with PC showed atrophic and fibrotic changes. There was no significant difference in BCA in subjects with PC compared to controls (1.53 ± 1.26% vs. 0.95 ± 0.42%, p = 0.07). However, ACA and ACA to BCA ratio were significantly higher in subjects with PC compared to controls (2.48 ± 2.39% vs. 0.53 ± 0.26% and 1.94 ± 1.93 vs. 0.59 ± 0.26, respectively, both p < 0.001). Increased ACA to BCA ratio was observed in subjects with PC irrespective of the presence of DM. Qualitative evaluation of ALDH1A3 and UCN3 expression showed no significant difference between the groups. In conclusion, in subjects with PC, alpha to beta cell mass ratio is increased, which may contribute to the increased risk of worsening glucose metabolism. Further studies are warranted to elucidate the mechanisms of increased alpha to beta cell mass in patients with PC.
IN GENERAL, pancreatic cancer (PC) is a malignant disease with a poor prognosis, with a 5-year survival rate of 3–10% [1-6]. Development of PC is often associated with worsening glucose metabolism and onset of diabetes, so-called pancreatic cancer associated diabetes [7-9], while in subjects with type 2 diabetes, the risk of PC has also been shown to be 1.5 to 2 times higher than in those without diabetes [7, 8, 10, 11].
It has been reported that patients with PC exhibit insulin resistance and beta cell dysfunction [12-14], and it is speculated that a complex combination of islet blood flow dysfunction, microthrombosis and perivascular fibrosis suppresses insulin secretion dynamics and leads to the development of diabetes [8, 15].
Previous histological studies have shown decreased beta cells and increased alpha cells in patients with PC [16, 17], while others have reported no change in beta and alpha cells in PC patients [18], and there is no certainty about the changes in endocrine cells in PC patients.
Moreover, a recent histological study has shown that the percentage of dedifferentiated beta cells was increased in non-diabetic PC patients, suggesting that paraneoplastic beta cell dysfunction precedes hyperglycemia [19]. However, there are limited data available on histological changes in the pancreatic islets in subjects with PC, and it remains unclear how PC affects beta cell mass (BCM) and alpha cell mass (ACM).
Therefore, in the present study, we aimed to elucidate the effects of PC on ACM and BCM and its effects on dedifferentiation markers in subjects with and without diabetes, using autopsy specimens.
The characteristics of the subjects are shown in Table 1 and Supplementary Table 1. Autopsy specimens of pancreas were obtained with the permission of the bereaved family. This study was approved by the Review Committee of the Faculty of Medicine, Keio University (approval number 20120475). Potential cases that died of PC and age- and BMI-matched cases without pancreatic disease (controls) were identified by retrospective analysis of the autopsy database of Keio University. To assess the effects of diabetes, both cases with and without diabetes were included.
| PC group | Controls | Total | |||||
|---|---|---|---|---|---|---|---|
| Total | DM (–) | DM (+) | Total | DM (–) | DM (+) | ||
| N (male/female) | 30 (19/11) | 21 (12/9) | 9 (7/2) | 31 (21/10) | 19 (12/7) | 12 (9/3) | 61 (40/21) |
| Age, years | 68.8 ± 9.7 | 68.4 ± 10.3 | 69.7 ± 8.4 | 65.6 ± 7.0 | 63.7 ± 6.4 | 68.8 ± 7.0 | 67.1 ± 8.5 |
| BMI, kg/m2 | 21.4 ± 4.2 | 21.1 ± 3.3 | 22.2 ± 6.0 | 22.2 ± 4.1 | 22.3 ± 4.0 | 22.0 ± 4.3 | 21.8 ± 4.1 |
| HbA1c, %1) | 6.7 ± 1.1 | 5.2 ± 0.2 | 7.1 ± 0.85†‡ | 6.3 ± 1.0 | 5.6 ± 0.7 | 7.3 ± 0.5†‡ | 6.5 ± 1.0 |
Data are mean ± SD or n (%). PC; pancreatic cancer, DM; diabetes.
1) HbA1c was obtained in 36 subjects.
† p < 0.05 vs. PC group without DM.
‡ p < 0.05 vs. controls without DM.
Eligible cases 1) underwent a complete autopsy within 24 hours of death, 2) had available medical information prior to death, 3) were not receiving obvious glucocorticoid therapy, and 4) had sufficient size and quality of preserved tumor-free pancreatic tissue. Cases in which the pancreatic tissue was autolytic were excluded.
After reviewing about 2,000 autopsy cases from 1992 to 2017, 39 cases of PC were identified. Among them, 5 cases were excluded because of lack of pancreatic specimens and 4 cases were excluded because of insufficient quality of the specimen, resulting in inclusion of a total of 30 cases with PC (9 with diabetes; DM) in this study. One subject in the PC group was diagnosed with DM at the time of PC diagnosis; the rest had been diagnosed with type 2 DM at least 2 years prior to their PC diagnosis.
The histological type of PC in all cases was pancreatic ductal adenocarcinoma (PDAC). Among them, 4 cases were mucinous carcinoma and the rest were tubular adenocarcinoma. The location of the tumor was the pancreatic body or tail in 12 cases and the pancreatic head in 18 cases.
As controls, 31 age- and BMI-matched cases without pancreatic disease (with and without type 2 DM) were selected from the database. Of those, 14 cases died of malignancy other than PC (Supplementary Table 1). The presence of DM was confirmed by review of medical records and autopsy reports. In 41 cases (20 PC, 21 controls), specimens were collected from the body or tail of the pancreas, and in the remaining cases, specimens were collected from the head of the pancreas.
Glycated hemoglobin (HbA1c) level within 1 year prior to death was obtained from the medical records in 36 cases (Table 1) [20]. HbA1c values were expressed as National Glycohemoglobin Standardization Program (NGSP) values [21].
Tissue processing of pancreasThe pancreas was fixed with formaldehyde at autopsy and embedded in paraffin for subsequent analysis. In cases of PC, we selected sections of tumor-free pancreatic tissue for histological analysis. Subsequently, 5-μm sections were stained as follows: 1) hematoxylin and eosin, 2) insulin (peroxidase staining) with hematoxylin, and 3) glucagon with hematoxylin, and viewed under an optical microscope. For immunohistochemical staining, guinea pig polyclonal antibody against porcine insulin and rabbit polyclonal antibody against human glucagon were used (DAKO Japan).
As dedifferentiation markers, chromogenic immunohistochemical staining for aldehyde dehydrogenase 1 family member A3 (ALDH1A3) and urocortin 3 (UCN3) expression was evaluated using an automated staining system (Bond Max, Leica Biosystems) with a Bond Polymer Refine Detection Kit (Leica Biosystems). For primary antibodies, an anti-ALDH1A3 rabbit polyclonal antibody (dilution, 1:1000; Novus, NBP2-15339) and an anti-UCN3 rabbit monoclonal antibody (dilution, 1:200; Sigma, HPA-038281) were used.
Morphometric analysisFor quantification of beta cell area (BCA), a single cross-sectional pancreatic section of each subject was used; it was photographed using a NanoZoomer-XR slide scanner at 200× (20× objective), displayed with NDP.view2 software (Hamamatsu Photonics, Shizuoka, Japan) and then the percentage of BCA to the total pancreas area was digitally measured using Image Pro Plus software (Media Cybernetics, Silver Spring, MD, USA). Similarly, the ratio of alpha cell area (ACA) to total pancreas area was also digitally measured, and the ratio of ACA to BCA was determined in each case. These measurements were performed by one researcher (T.T.), and the inter-observer and intra-observer coefficients of variation were approximately 6% and 5%, respectively. All measurements were performed twice, and the mean of the two measurements was used. At the time of measurement, the researcher was blinded to the PC and glucose metabolism status of each sample.
In addition, mean islet size was quantified using NDP.view2 in a randomly selected region of the pancreas containing approximately 100 islets in each case (101 ± 17 islets, 6,191 islets in total) [20, 22]. For evaluation of dedifferentiation markers, we selected 20 cases in the PC group (12 with high ACA to BCA ratio and 8 with low ACA to BCA ratio) and 10 cases in the controls (with low ACA to BCA ratio) and compared them qualitatively (Supplementary Table 2). Two researchers (T.T. and Y.M.) evaluated each slice using the same criteria under double-blind conditions. The expression of these markers within the islets was qualitatively evaluated as four levels, [none (–) to very high (+++)], depending on the degree of positive staining.
Statistical analysisData showing a non-normal distribution were expressed as median (interquartile range) and those showing a normal distribution were expressed as mean ± SD. Mann-Whitney U test was used to analyze differences between the two groups, and Kruskal-Wallis test was used to analyze differences between the four groups. Statistical analysis was performed using SPSS 26 software (IBM, Chicago, IL, USA). For all analyses, a value of p < 0.05 was regarded as significant.
Characteristics of patients in the PC group (n = 30, 9 with DM) and controls (n = 31, 12 with DM) are shown in Table 1 and Supplementary Table 1. HbA1c levels in the DM group with and without PC were significantly higher compared to those in subjects without DM with and without PC, respectively.
Pancreatic morphology in subjects with PCRepresentative microphotographs of the pancreas of subjects with and without PC are shown in Fig. 1 Atrophic and fibrotic changes of the pancreatic parenchyma were more prominent in the tumor-free pancreas of those with PC compared with the pancreas of those without PC. In contrast, islet morphology was relatively intact in those with PC compared with those without PC.
Representative microphotographs of pancreas of subjects with (b) and without (a) pancreatic cancer. Tumor-free pancreatic sections were immunostained for insulin (brown) and with hematoxylin. Scale bar = 500 μm.
While there was no significant difference in BCA between subjects with and without PC (1.53 ± 1.26% vs. 0.95 ± 0.42%, p = 0.07) (Figs. 2 and 3a, Supplementary Table 2), ACA was significantly greater in subjects with PC compared with those without PC (2.48 ± 2.39% vs. 0.53 ± 0.26%, p < 0.001) (Fig. 3b). ACA to BCA ratio was also significantly higher in those with PC compared with those without PC (1.94 ± 1.93 vs. 0.59 ± 0.26, p < 0.001) (Fig. 3c). Increases in ACA and ACA/BCA ratio were observed in PC patients with and without DM, while BCA did not differ significantly between patients with and without DM or PC (Supplementary Fig. 1). Increases in ACA and ACA to BCA ratio were observed in samples from both the pancreatic head and body/tail (Supplementary Fig. 2). There was no significant difference in mean islet size between the groups with and without PC (6,973 ± 3,508 vs. 5,819 ± 2,180 μm2, p = 0.25).
Representative photomicrographs of pancreas immunostained for glucagon (brown) (a, c, e, g, i, k) or insulin (brown) (b, d, f, h, j, l) and with hematoxylin. a–d: A case without pancreatic cancer (c and d: higher power images). e–l: Cases with pancreatic cancer (g, h, k and l: higher power images). Scale bars, 500 μm for lower power images and 100 μm for higher power images.
Beta cell area (BCA, a), alpha cell area (ACA, b) and ACA to BCA ratio (c) in subjects with and without pancreatic cancer. Bars indicate mean.
To assess the role of dedifferentiation as a mechanism of increased ACA to BCA ratio in subjects with PC, sections of pancreas from controls and PC were stained for ALDH1A3 (Fig. 4a and 4c) and UCN3 (Fig. 4b and 4d) and the expression patterns of these markers within the islets were qualitatively evaluated (Supplementary Table 2). However, we could not find any certain relationship between the ACA to BCA ratio and the degree of staining. We also did not find any qualitatively significant difference in expression of dedifferentiation markers between subjects with and without PC.
Representative microphotographs of islets immunostained for ALDH1A3 (brown, a and c) or UCN3 (brown, b and d) and with hematoxylin. Subjects with (c and d) and without (a and b) pancreatic cancer. Scale bar = 250 μm.
In this study, we report that 1) ACA and ACA to BCA ratio were increased in subjects with PC compared with those without PC, 2) an increase in ACA to BCA ratio was observed in subjects with PC irrespective of the presence or absence of DM, and 3) there was no significant change in expression of dedifferentiation markers in subjects with PC.
Type 1 and type 2 DM are both characterized by a deficit of BCM [23-26]; however, the effects of PC on BCM remain unclear. Similarly, ACA/BCA ratio has been reported to be significantly correlated with glucose metabolism indices such as fasting blood glucose level, blood glucose level 2 hours after a load, and HbA1c, but there are few reports on ACA/BCA ratio in PC [27]. In our prior study, we reported that BCA but not ACA was reduced in subjects with PC compared with those without PC [25], while in this study, we observed no reduction in BCA but increases in ACA and ACA to BCA ratio in subjects with PC. Several reasons for this inconsistency including differences in patient characteristics between the studies can be postulated. Especially, subjects who had undergone surgical resection of PC were included in our prior study, while autopsy cases that died of PC, i.e., with advanced clinical stage of PC, were included in this study. That is, the subjects in the previous study had operable early-stage cancer, whereas the patients in the present study had so-called end-stage cancer in which the cause of death was PC in all of them. Indeed, the pancreases of subjects with PC in this study were markedly atrophic and fibrotic, which are major histological changes often observed in subjects with advanced PC [19, 28]. Due to the atrophic and fibrotic changes in the exocrine pancreas, BCA and ACA expressed as fractional area to the exocrine area may be overestimated in subjects with advanced PC. In this study, thus, the increase in ACA to BCA ratio in subjects with PC suggests two possibilities; reduced BCM and/or increased ACM, although it was not possible for us to calculate the actual BCM or ACM due to the lack of information on pancreas weight in each individuals. Two of the subjects with PC in the present study had a history of surgery for PC, but when we excluded these subjects and performed the analysis, the results remained the same.
DM secondary to pancreatic cancer is associated with a reduced response of beta cells to an oral glucose load, hyperglycemic clamp, and glucagon stimulation [29-32]. Basso et al. conducted a glucagon-stimulation study in PC, type 1 DM, type 2 DM patients, and healthy controls to evaluate beta cell function in PC patients. They found no significant increase in C-peptide level in type 1 DM and PC patients after glucagon stimulation, but a significant increase in controls and type 2 DM patients [31].
Previous studies showed that plasma islet amyloid polypeptide level was elevated in PC patients with DM, and it was speculated that this might contribute to the pathogenesis of the disease. However, this hypothesis has not been explored further, and the mechanism of altered beta cell function remains unknown [33-35]. On the other hand, Permert et al. found, in a series of studies, insulin resistance in subjects with PC, which was more pronounced in subjects with DM, but was also present in normoglycemic subjects with PC [36, 37].
These results suggest that insulin resistance and altered beta cell function seen in PC patients may lead to hyperglycemia [31]. However, previous studies on the amount of beta cells in PC have not yielded consistent results. Meier et al. reported no significant difference in BCA between subjects with and without primary pancreatic disease, including PC, and Yoon et al. reported no significant difference in the relative volume of alpha- or beta-cells between patients with and without PC [18, 38]. On the other hand, Katsumichi et al. reported a decrease in islet size and beta cell count in subjects with PC, but the number of samples was small and as the PC group did not include subjects without DM, the influence of DM on islet morphology could not be excluded [39].
There have been previous reports showing increased plasma glucagon level in patients with PC [30, 40, 41]. Jin et al. reported higher plasma glucagon level, lower insulin level, and higher glucagon/insulin ratio during 75 g OGTT in patients with PC [41]. Also, Kolb et al. reported that there was no difference in insulin-expressing area or islet size between subjects with and without PC, but glucagon-expressing area and glucagon/insulin expression ratio were higher in PC patients [40], consistent with our findings. Schmied et al. reported that PC specimens also showed a significant decrease in beta cells and a significant increase in alpha cells within and around the tumor [16, 17].
Recently, Wang et al. reported that ALDH1A3, a marker of dedifferentiation, is highly expressed and UCN3 expression is low in PC patients [19]. ALDH1A3 is highly expressed in progenitor cells and functions as a marker of beta cell dedifferentiation. In contrast, UCN3 is a beta cell-enriched gene product and its loss is an early event in beta cell dedifferentiation in DM. Therefore, UCN3 is considered a functional maturation marker [42]. However, in this study, we did not find any significant change in expression of dedifferentiation markers in subjects with PC, although there was no significant difference in mean islet size between the groups. Further studies are needed to clarify the role of beta cell dedifferentiation in subjects with PC.
There are limitations of this study. First, we cannot rule out the possibility that other factors such as duration and treatment of PC, cause of death and family history of DM might have influenced the results of this study. Secondly, different portions of the pancreas were sampled, although the proportion of endocrine cells has been shown to be relatively consistent regardless of the pancreatic site, except for the ventral portion of the pancreatic head [26]. Third, with respect to subjects with PC, pathological changes in pancreatic exocrine tissue may have affected endocrine tissue changes, but it is difficult to determine whether the changes in fibrosis and atrophy of the pancreatic parenchyma observed in pancreatic cancer specimens are a direct effect of PC or an effect associated with complicated pancreatitis. Finally, all the subjects in this study were Japanese, so our results may not be as relevant for other ethnic groups.
In conclusion, ACA to BCA ratio was increased in subjects with PC compared to those without PC. Since an increase in ACA to BCA ratio was observed in subjects with PC irrespective of the presence or absence of DM, a relative increase in ACM to BCM may be one of the mechanisms of worsening glucose metabolism associated with PC. Further research is needed to clarify the mechanisms of the relative increase in ACM in subjects with PC.
We thank Yuko Madokoro, Department of Pathology, Keio University School of Medicine, for technical assistance and Dr. Wendy Gray, self-employed, for editing the manuscript.
None of the authors has any potential conflict of interest associated with this study.
The authors were supported by grants from the Ministry of Education, Culture, Sports, Science and Technology (Grant Nos. 20K17542 and 21K08535).
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.
Not applicable.
T.T. researched and analyzed the data and wrote the manuscript. Y.S. designed the study, analyzed the data and wrote the manuscript. J.I., H.S., M.S. and M.N. contributed to discussion and reviewed and edited the manuscript. Y.M. and T.Y. researched the data, contributed to discussion and reviewed and edited the manuscript. H.I. contributed to discussion and reviewed and edited the manuscript. Y.S. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
This study was carried out with the approval of the Keio University School of Medicine Ethics Committee (approval number 20120475, approval date 2022/03/28, initial approval date 2013/02/25).
Pancreatic specimens obtained at autopsy were obtained with the permission of the bereaved families.
This study was conducted with the permission of the bereaved families.
