Chromogranin A‐positive hormone‐negative endocrine cells in pancreas in human pregnancy

Abstract Introduction We sought to determine whether chromogranin A‐positive hormone‐negative (CPHN) endocrine cells are increased in the pancreas of pregnant women, offering potential evidence in support of neogenesis. Methods Autopsy pancreata from pregnant women (n = 14) and age‐matched non‐pregnant control women (n = 9) were obtained. Staining of pancreatic sections for chromogranin A, insulin and a cocktail of glucagon, somatostatin, pancreatic polypeptide and ghrelin was undertaken, with subsequent evaluation for CPHN cell frequency. Results The frequency of clustered β‐cells was increased in pregnant compared to non‐pregnant subjects (46.6 ± 5.0 vs. 31.8 ± 5.0% clustered β‐cells of total clustered endocrine cells, pregnant vs. non‐pregnant, p < .05). Frequency of endocrine cocktail cells was lower in pregnant women than non‐pregnant women (36.2 ± 4.0 vs. 57.0 ± 6.8% clustered endocrine cocktail cells of total clustered endocrine cells, pregnant vs. non‐pregnant, p < .01). No difference in frequency of CPHN cells was found in islets, nor in clustered or single cells scattered throughout the exocrine pancreas, between pregnant and non‐pregnant women. The frequency of CPHN cells in pregnancy was independent of the number of pregnancies (gravidity). Conclusions Our findings of no increase in CPHN cell frequency in pancreas of pregnant women suggest that this potential β‐cell regenerative mechanism is not that by which the increased β‐cell mass of pregnancy is achieved. However, an increase in the percentage of clustered β‐cells was found in pregnancy, with decreased frequency of other endocrine cells in clusters, suggesting a compensatory shift from other pancreatic endocrine cell types to β‐cells as a mechanism to meet the increased insulin demands of pregnancy.


| INTRODUC TI ON
We have previously reported a novel endocrine cell type in the pancreas that expresses chromogranin A but none of the other known pancreatic hormones. We adopted the term chromogranin A-positive hormone-negative (CPHN) cells and found that they were increased in diabetes and other pancreatic disorders. [1][2][3][4][5] The distribution of CPHN cells, occurring mainly as single cells or in small clustered groups of 1-3 cells scattered throughout the exocrine pancreas, rather than in established islets, appeared to recapitulate what is found in the pancreas during early human development 1 and led us to conclude that these were newly forming cells.
However, there is an alternative explanation: that CPHN cells are endocrine cells demonstrating an altered identity that occur with increased frequency in the setting of hyperglycaemia in type 1 and type 2 diabetes. It is known that pancreatic endocrine cells can undergo various 'cell plasticity' events in response to physiological stresses. 6 There is also growing evidence that pancreatic endocrine cells may alter their identity (either by dedifferentiation or transdifferentiation) in response to metabolic abnormalities 7,8 with alpha to beta transdifferentiation having been described, albeit mainly in mice. 9 Mammalian pregnancy is marked by a stepwise adaptation of the endocrine pancreas with expansion of maternal insulin secretory capacity as the primary adaptation that is necessary to allow the foetus preferential use of circulating glucose. [10][11][12] The physiological stimulus of pregnancy gives rise to a period of metabolic plasticity, decreasing maternal insulin sensitivity by approximately 50%. 13,14 To maintain euglycaemia, that insulin resistance is offset by an ~250% increase in maternal insulin production. 13,14 These physiological adaptations are accomplished partly through increased workload (insulin synthesis and secretory burden) of the existing β-cells. However, there is also an increase in β-cell numbers in pregnancy. 12 The exact mechanism of adaptive expansion of β-cells in human pregnancy is still under investigation. However, studies in mouse models [15][16][17][18][19] and in vitro studies with human islets 20 have revealed that the morphological changes such as replication and increase of pancreatic β-cell numbers in pregnancy are driven mainly by placental lactogen (PL) and prolactin-mediated gene regulation.
Chromogranin A-positive hormone-negative cells may represent newly forming endocrine cells or pancreatic endocrine cells with an altered identity, perhaps dedifferentiating/transdifferentiating towards a β-cell fate in response to physiological stress conditions. We therefore questioned whether, in pregnancy, a physiological state of β-cell stress without the confounding variable of hyperglycaemia, there would be an increase in the frequency of CPHN cells.
To address this question, we analysed sections of pancreata from human autopsy subjects (14 pregnant and 9 non-pregnant agematched women) and sought to determine CPHN cell frequency in the pregnant vs. control women. Human autopsy pancreas from 14 pregnant women and 9 age and   pre-pregnancy BMI matched non-pregnant control women was   obtained Table 1, Figure S1A, B). Both the pregnant and the nonpregnant women have been included in a previous publication. 21 Subjects were identified in the Mayo Clinic autopsy database (RAR).

| Autopsy cases
To be included in the study, each subject had to fulfil the following criteria: complete autopsy within 24 h of death, with pancreas of adequate size and quality collected. If the pancreas tissue quality was suboptimal, due to autolysis or pancreatitis, this was an exclusion criterion. Women included in the study did not have a history of diabetes. Subject demographic data and cause of death are listed in Fasting blood glucose values were not available in the medical record; therefore, non-diabetic status was based upon the lack of a history of diabetes.

| Pancreatic tissue processing
Autopsies were undertaken at Mayo Clinic. Pancreas tail was sampled (~2.0 × 1.0 × 0.5 cm size) and fixed in formaldehyde together with a sample of spleen before being embedded in paraffin. 4 μm sections were cut from the tissue blocks.

| Immunofluorescence staining and quantification of chromogranin A-positive hormonenegative [CPHN] cells
Sections were stained for Chromogranin A, insulin, glucagon, somatostatin, pancreatic polypeptide and ghrelin to detect CPHN cells as described previously. 1,3 In brief, slides were incubated overnight at 4°C with a cocktail of primary antibodies prepared in antibody solution (3% bovine serum albumin in Tris-buffered saline with 0.2% Tween 20). The primary antibodies were detected by a cocktail of appropriate secondary antibodies (Jackson ImmunoResearch, Westgrove, PA). The information for both primary and secondary antibodies is detailed in Table S1. Mounting medium containing DAPI (Vectashield,Vector Laboratories, Burlingame, CA) was used for counterstaining the slides to mark the nuclei and slides were viewed using a Leica DM6000 microscope (Leica Microsystems, Deerfield, IL). All images were acquired using the ×20 objective (×200 magnification) and using a Hamamatsu Orca-ER camera (catalog no. C4742-80-12AG,Indigo Scientific, Bridgewater, NJ) and Openlab software (Improvision, Lexington, MA).  To calculate the % of total clustered endocrine cells, the denom-

| Statistical analysis
Data are presented as means ±SEM. Statistical analysis was performed using GraphPad Prism 8 (GraphPad Software, San Diego, CA, USA). (Table S2) Consistent with our previous report, 21

| The comparison of CPHN cells in islets and clustered cells in pregnant vs. non-pregnant women Table S3
Table

| Effect of gravidity on the frequency of CPHN cells
In our pregnancy cohort, there were 5 women who were gravida 1 (G1) and 3 women who were gravida 2 (G2). In addition, there was

| DISCUSS ION
Pancreatic non-hormone expressing endocrine cells (CPHN cells) are found in abundance in normal human pancreas during development. 1 A resurgence of CPHN cells has been documented in both type 1 diabetes (T1D) 4 and type 2 diabetes (T2D) 13 where there is a deficit in the endocrine, specifically the beta cell, complement. 22,23 As a consequence of these observations, we have previously proposed that CPHN cells are an immature type of endocrine cell representing, in the setting of a pathological beta cell deficiency, an attempted, albeit insufficient, regenerative response, in an effort to restore the beta cell complement.
Pregnancy represents a unique physiological state of beta cell stress. In this study, we report that the frequency of CPHN cells In pregnancy, maternal pancreatic β-cells undergo morphological adaptation to compensate for the excess insulin demand. 19 Previous studies suggest that rodent and human β cells adapt differently to pregnancy. 12,21,24,25,26 The mechanisms responsible for this adaptation involve increased replication and decreased apoptosis in rodents, 25,27 but possibly involve islet neogenesis in humans. 21 Recently, it has been proposed that 'altered identity' of pancreatic β-cells is another player in the stress-induced adaptation of pancreatic β-cells in humans (reviewed in. 28 ) We previously reported that the adaptive increase in β-cells in human pregnancy is achieved by formation of new islets, which is again confirmed by our current observation that the per cent of clustered β-cells was increased in pregnancy. Interestingly, our current data also indicated a decreased proportion of endocrine cocktail cells in pregnancy, suggesting that adaptive changes of endocrine pancreas might drive transdifferentiation of other endocrine cells to β-cells to cope with the excess insulin demand of human pregnancy.
Therefore, it is possible that the CPHN cells present in pregnancy may be indicative of 'endocrine cells in transition' with altered identity rather than new cell formation as was observed in humans with T1D and T2D. 3,4 Here we report that the frequency of CPHN cells in pancreas of pregnant women is not increased in comparison with agematched non-pregnant control women. One possible explanation for the lack of an increase in CPHN cells in pregnant women might be the 'gestational window' of the selected cases in our study.
The average gestational age of the pregnant women in our study was 24.5 ± 2.7 weeks (ranging from 10 to 40 weeks). Early human pregnancy (the first two trimesters) is an anabolic phase, during when pancreatic β-cell activity increases; as such, an enhanced insulinotropic effect of glucose is seen during this phase, whereas insulin sensitivity is either not altered or enhanced. 29 In contrast, the decrease in insulin sensitivity that occurs in later gestation (known as the catabolic phase) may enhance the workload of β-cells to compensate for the increased insulin demand. 29  that β-cell neogenesis or stress-induced dedifferentiation would also peak at those stages. Our study, however, included only a few cases that corresponded to those specific gestational windows.
Therefore, the insufficient number of appropriate cases may have resulted in a failure to detect a surge in CPHN cells during those critical time points.
In conclusion, our study found no increase in the frequency of CPHN cells in pancreas of women in response to the increased metabolic workload of pregnancy when compared with non-pregnant control women. Our finding suggests that this potential beta cell regenerative mechanism is not that by which the increased beta cell mass of pregnancy is achieved, although further studies utilizing pancreas from pregnant women during the critical early and late gestational windows are warranted. Interestingly, however, an increase in the percentage of clustered beta cells was found in pregnancy, with a decrease in frequency of other endocrine cells in clusters, suggesting that a compensatory shift occurs to bolster the beta cell component in order to compensate for the increased insulin demand of pregnancy.

ACK N OWLED G EM ENTS
None.

CO N FLI C T O F I NTE R E S T
The authors report no conflicts of interest in this work.

AUTH O R CO NTR I B UTI O N
ASMM and KZ performed the studies analysed the data and wrote the first draft of the manuscript. RAR procured the autopsy pancreas samples and edited the manuscript. SD and AEB contributed to study design, data interpretation and the writing of the manuscript.
All authors reviewed and approved the final version of the manuscript. Alexandra E Butler is the guarantor of this work.

E TH I C A L A PPROVA L
Institutional Review Board (IRB) approval from the Mayo Clinic and University of California Los Angeles (UCLA) was obtained. Both the Mayo Clinic IRB and the UCLA IRB waived the requirement for written informed consent for this study due to the following reasons: (1) According to the FDA definition of human subjects, a human subject means a living individual and therefore (2) research on human autopsy subjects is considered by the IRB to be 'exempt' as defined by federal regulation 45 CFR 46, as autopsy subjects do not meet the FDA definition of human subjects.

CO N S E NT TO PA RTI CI PATE
Not applicable.

CO N S E NT FO R PU B LI C ATI O N
All authors gave their consent to publish this work.

CO D E AVA I L A B I LIT Y
Not applicable.

DATA AVA I L A B I L I T Y S TAT E M E N T
All relevant data are within the manuscript and its Supporting Information files.