Pleiotrophin (PTN) is a heparin-binding cytokine that is highly conserved across mammalian species (1, 2). It is expressed in multiple tissues during embryonic and fetal development and is important for tissue growth and morphogenesis in organs that undergo branching morphogenesis (3, 4). The pancreas is one such organ where branching ductal morphogenesis gives rise to endocrine, acinar and ductal cell lineages (5). In particular PTN contributes to cell proliferation and angiogenesis and, in the embryonic pancreas, Ptn is expressed in areas of vasculogenesis adjacent to the differentiating ductal epithelium. In mouse embryonic pancreatic primordia explants, inhibition of Ptn expression resulted in a failure of endocrine precursors to fully differentiate and express insulin or glucagon (6). By E11-13 Ptn expression becomes limited to the pancreatic epithelium and PTN is localized to the basement membranes, especially associated with islet budding from the ductal epithelium.

Ptn expression is retained following birth in multiple tissues and increased expression is found following tissue damage and inflammation (2, 7, 8). Conversely, Ptn ^-/- mice develop insulin resistance associated with a failure to accumulate fat in white adipose, increased lipolytic activity and systemic cytokine-induced inflammation (9). Moreover, late pregnant Ptn ^-/- mice are glucose intolerant and have an altered metabolism of both glucose and lipids, which is associated with a diabetogenic state in the pregnant mother. During pregnancy, PTN is expressed within the cytotrophoblasts of the placenta and possibly released into the maternal circulation since circulating levels more than doubled between first and third trimester of human pregnancy (10). Despite the association of PTN with metabolic control the presence and actions of PTN in the endocrine pancreas have not been described. Gene expression libraries of adult human or rodent pancreatic islets or isolated β-cells have often failed to detect substantial expression of PTN mRNA (11, 12). However, PTN was detected in islets of 3-month-old rats and was down-regulated in diabetic animals (13). Also, in humans PTN is one of only eleven circulating proteins to be significantly associated with advancing chronological age, suggesting a retention of expression in adult tissues (14).

A major cell membrane receptor for PTN is the receptor phosphotyrosine phosphatase-β/ζ (RPTP β/ζ) encoded by the protein tyrosine phosphatase receptor Type Z1 (PTPRZ1) gene (2). PTPRZ1 is strongly expressed in adult mouse and rat pancreatic islets, mouse β-cell lines, and in human β-cells (15), raising the possibility that β-cells may be a target for PTN actions. PTN undergoes proteolytic modification in the extracellular fluid resulting in proteins with differing biological activities. For instance, cleavage by plasmin or metalloproteinase 2 (MMP2) can result in multiple peptides which differentially alter angiogenesis and can alter the actions of VEGF-A on endothelial cells (16, 17). In addition to RPTP β/ζ, a number of other PTN-binding proteins are associated with pancreatic islets such as the αvβ3 integrin (18).

In this study we have examined the expression and presence of PTN in young and adult mouse pancreas, how this might alter during pregnancy and the actions of PTN on β-cell DNA synthesis and glucose-stimulated insulin release.

We examined β-cells isolated by FACS from the young mouse pancreas for Ptn and Ptprz1 mRNA presence and found both ( Table 1 ). Previously, we had identified a sub-population of β-cells in mouse and human pancreas that have little Glut2 expression but have greater proliferative potential (21). Thus, we examined if a differential expression of Ptn and Ptprz1 existed between Ins⁺Glut2^LO and the more abundant and functional Ins⁺Glut2^HI cells. Ptn was five-fold more abundantly expressed within Ins⁺Glut2^LO cells and Ptprz1 expression three-fold greater when compared to Ins⁺Glut2^HI cells ( Table 1 ). Other PTN binding proteins differentially expressed between Ins⁺Glut2^LO and Ins⁺Glut2^HI cells were the αvβ3 integrin sub-units, and heparin sulphate proteoglycan 2 (Hsgp2). Messenger RNA for PTN and RPTP β/ζ was also readily detectable in the INS1E β-cell line ( Figure 1A, B ), when compared to the house-keeping gene β-actin ( Figures 1F ), and Ptn continued to be expressed in adult mouse pancreas ( Figure 1G ).

Since RPTP β/ζ was expressed by β-cells, PTN may be part of the cell surface proteome in adult mouse through sequestration by the receptor or other binding proteins in addition to a local expression. PTN was localized within adult mouse pancreas sections by fluorescence immunohistochemistry. PTN immunoreactivity was absent from the acinar tissue and pancreatic ducts but was present in the islets of Langerhans and smaller extra-islet endocrine clusters. Staining was predominantly localized to β-cells within islets, as judged by cellular co-localization with insulin (84.7 ± 2.5% of β-cells, n=6 mice), and was less abundant in α-cells assessed from the co-localization with glucagon ( Figure 2A ). The specificity of the PTN staining was confirmed by the absence of signal following omission of the primary PTN antibody ( Figure 2B ), omission of the secondary PTN antibody, or prior pre-absorption of the primary antibody with excess recombinant PTN. Substantial co-localization was observed between Glut2 and PTN within islets suggesting a presence in mature β-cells ( Figure 2C ) while little co-localization was seen between somatostatin, as a marker of δ-cells, and PTN ( Figure 2D ). PTN was also observed to co-localize with β-cells within the islets of human pancreas ( Figures 2E, F ), with immunopositive cells being uniformly distributed throughout the islets. However, the presence of PTN-immunopositive cells was lower in human islets than in mouse and differed depending on whether the islets were small or larger (small < 10,000 µm², 60.1 ± 2.3% of islet cells; large >10,000 µm², 34.9 ± 3.2%, p<0.001). These results suggest that PTN is present on pancreatic β-cells in both adult mouse and human.

We next examined if PTN presence was altered in pancreas during the mild metabolic stress represented by pregnancy. Quantitative PCR showed that the expression of Ptn did not significantly change over the course of pregnancy ( Figure 1G ). Pregnant mouse pancreata at GD 9, 12 and 18 were also co-stained for insulin, glucagon, Glut2 and PTN compared with non-pregnant animals ( Figure 3 ). A similar co-localization of PTN to β-cells was observed during pregnancy with the percentage of β-cells that co-stained for PTN not differing from the 85% calculated in non-pregnant animals at GD 9 or 12 (n=6). At GD 9 and 12 ( Figures 3B, C ), when total β-cell mass is beginning to increase to counter the insulin resistance of pregnancy (23), PTN was largely localized to the mature β-cells expressing insulin and Glut2 ( Figure 3B ). However, PTN was also localized to immature insulin-immunoreactive cells that lacked staining for Glut2 (Ins⁺Glut2^LO cells) ( Supplementary Figure 1 ).

The relative abundance of PTN-containing β-cells compared to α-cells was quantified in the pancreas of non-pregnant vs. pregnant female mice. Approximately 60% of PTN-immunoreactive cells within islets co-stained with insulin and 20% with glucagon in non-pregnant animals ( Figure 4A ). This did not differ at GD 12 of pregnancy. These results show that the distribution of immunoreactive PTN between pancreatic endocrine cell types did not significantly differ during the adaptive increase in β-cell mass that occurs during mouse pregnancy. Further, we examined the proportion of β-cells within islets that contained PTN immunoreactivity when Ins⁺Glut2^HI and Ins⁺Glut2^LO cells were considered separately, and how this might change across pregnancy. This relative distribution of PTN between β-cell phenotypes did not significantly change over the course of pregnancy ( Figure 4B ).

To examine a more pronounced metabolic stress than pregnancy, we partially depleted β-cells by treatment with STZ. Two weeks following STZ administration the number of insulin-staining β-cells were substantially reduced compared to control mice, a time point associated with a regenerative response of β-cells in young animals (20). These animals still had relatively elevated mean blood glucose values at this time relative to vehicle-treated controls (STZ 11.9 ± 1.4 mM, control 8.4 ± 0.6 mM, p<0.05). Strong PTN immunoreactivity was associated with the remaining β-cells compared with islets from vehicle-treated mice ( Figures 2G, H ). However, the proportion of β-cells that contained immunoreactive PTN did not differ after STZ treatment (control 88.7 ± 2.5%, STZ 90.5 ± 1.6%, n=8, 3 pancreatic sections per animal).

The possibility that PTN might contribute to β-cell proliferation was examined using the INS1E cell line, a model of active β-cell replication as occurs in vivo in early life (20). Incubation with PTN (0.1 and 1μg/ml) resulted in a significant increase in the number of cells undergoing DNA synthesis as determined by immunoreactivity for Ki67 ( Figure 5A ). Analysis of mRNA expression using qPCR in INS1E cells showed a significant increase in the expression of the transcription factor Pdx1 and of insulin and Glut-2 mRNA following incubation with PTN ( Figures 1C–E ). The expression of Pdx1 was chosen since it is mechanistically associated both with β-cell proliferation as well as insulin biosynthesis. Although both Ptn and Ptprz1 were expressed by INS1E cells their expression was not altered in the presence of exogenous PTN ( Figures 1A, B ). The increased expression of insulin suggested a possible additional action on insulin release. This was explored by measuring basal and glucose-stimulated insulin release from INS1E cells following incubation with PTN, and the remaining cellular content of insulin. Incubation in the presence of 28.8 mM glucose increased insulin release, but without changing the remaining cellular content, compared to 2.8 mM glucose. However, neither insulin release nor content were altered in the presence of 0.1 μg/ml PTN ( Figure 6 ). Thus, it appears that exogenous PTN did not alter glucose-stimulated insulin secretion from INS1E cells.

We further investigated whether PTN could alter either DNA synthesis or glucose-stimulated insulin release from isolated adult mouse islets. However, PTN (0.1 and 1μg/ml) had no significant effect on the percentage of β-cells undergoing DNA synthesis ( Figure 5B and Supplementary Figure 2 ). Islets were also pre-incubated with 1 μg/ml PTN before exposure to either a low (2 mM) or high (25 mM) glucose stimulus in vitro. A seven-fold increase in insulin release was observed with 25 mM glucose compared to 2 mM, but neither the basal nor glucose-stimulated insulin release was altered by pre-incubation with PTN ( Figure 6C ). When islets were incubated with 2 mM glucose in the presence of the sulfonylurea tolbutamide, as a positive control for insulin release, a five-fold increase was observed with or without PTN. The insulin content of islets following glucose challenge was no different in the presence of PTN or tolbutamide, indicating a substantial reserve of cellular insulin that was not altered in the presence of PTN ( Figure 6D ). The findings support a lack of effect of PTN on acute glucose-stimulated insulin release. Also, in the adult mouse pancreas where the rate of β-cell turnover is low (0.2%, Figure 5B ) compared to INS1E cells, PTN alone was unable to provide a mitogenic stimulus.

The presence of PTN was localized to the islets of Langerhans of adult male and female mice and also in human pancreas. All sizes of islets showed a predominant co-localization of PTN with insulin, although a minority showed co-localization to glucagon-containing α-cells. Little staining was associated with δ-cells. The strong association between PTN and β-cells was further supported by the co-distribution with cells expressing Glut2. Although Ptn has not been shown previously to be prominently expressed in adult β-cells (11, 12), unlike the readily detectable levels of mRNA in early life and in proliferative β-cell lines, we show here that PTN mRNA is detectable in the adult mouse pancreas. The presence of PTN as part of the adult β-cell surface proteome could also be related to the abundant expression of the Ptprz1 receptor (15), as well as an association with other cell surface binding proteins such as HSPG2 and the αvβ3 integrin. The distribution of PTN-immunoreactive β-cells throughout human islets corresponds to the more uniform distribution of β-cells as compared with rodents (24). However, significant differences were found in the proportion of human islet cells containing PTN depending on islet size, with PTN-positive cells being more abundant in smaller islets. This mirrors a similarly greater proportional presence of β-cells in smaller vs. larger islets in humans (25), although the percentage of islet cells represented by β-cells overall is lower than in mouse islets.

We and others have previously described a sub-population of β-cells that in mouse and human express insulin but little Glut2 glucose transporter, and therefore exhibit poor glucose-stimulated insulin secretion (21, 26, 27). These Ins⁺Glut2^LO cells have multi-potential islet endocrine cell lineage potentiality, and a continued proliferative capacity. Furthermore, they are more abundant during the β-cell mass expansion that occurs during pregnancy (23) and during the limited regeneration of β-cells that occurs following their depletion with STZ in young mice (21, 23). Ptn was expressed at five-fold higher levels in Ins⁺Glut2^LO cells than in the presumably more functionally mature Ins⁺Glut2^HI cells in the young mouse, and a similarly elevated expression of the Adamts1 protease was seen, which cleaves PTN into a variety of bioactive fragments (28), and an increased expression of the Ptprz1 receptor, Hspg2 and subunits of the αvβ3 integrin. The visualization by immunohistochemistry of PTN in Ins⁺Glut2^LO cells as well as mature β-cells suggests that Ins⁺Glut2^LO progenitor cells may represent a focal point of PTN expression and paracrine action within pancreas in early life.

When mice were exposed to the metabolic stress of pregnancy there were no significant changes in the relative pancreatic expression, or the distribution of PTN between sub-populations of β-cells, despite the 2-3-fold increase in β-cell mass that occurs over the time course of normal mouse pregnancy (23). However, since the circulating levels of maternal PTN increase with the progression of pregnancy, at least in human (10), the PTN available for receptor activation on β-cells during pregnancy may include peptide derived from other tissues. The presence of PTN has been linked to tissue damage and subsequent repair, and its expression is upregulated in macrophages by inflammatory cytokines such as IFN-β and -γ (29, 30). Following depletion of β-cells with STZ in young mice intense staining for PTN was associated in the remaining β-cells 14 days following administration of STZ, at a time when β-cell regeneration is occurring (21).

What therefore might the biological actions of PTN on β-cells comprise? The turnover of β-cells in the adult mouse outside of pregnancy is slow and regeneration following β-cell loss is limited. We did not find an effect of PTN alone on the DNA synthetic rate in adult mouse β-cells. However, the increase in β-cell mass during pregnancy requires prior activation of the prolactin receptor by lactogenic hormones which cannot easily be reproduced in short-term cultures of isolated islets. It is therefore possible that in β-cells already ‘primed’ for replication PTN could have a contributory action. This is supported by the increased DNA synthesis seen when INS1E mouse β-cells, that have an endogenous replicative capacity, were incubated with PTN, and this was associated with an increased expression of insulin and Pdx1, the latter being required for β-cell proliferation. The increased expression of insulin and Glut-2 suggested that PTN might also modulate insulin release but no change in glucose-stimulated insulin release was found in the presence of PTN. Similarly, incubation in the presence of PTN did not alter glucose-stimulated insulin release from isolated mouse islets in vitro, or the islet content of insulin.

A previous study using Ptn knock-out mice showed these animals to become insulin-resistant in later life with an associated increase in plasma insulin levels at 12 months age (9), but that glucose metabolism was comparable to wild type controls at earlier ages. This suggests that PTN presence is not essential for insulin biosynthesis and release in adult life. Moreover, late pregnant (GD18) Ptn knock-out mice are hypoinsulinemic, hyperglycemic and glucose intolerant and have an altered liver metabolism of both glucose and lipids, which is associated with a diabetogenic state in the pregnant mother (31). The present study using animals less than three months of age further shows that the addition of exogenous PTN was unable to increase insulin release from isolated islets of a β-cell line. The increased levels of insulin mRNA in INS1E cells following exposure to a high dose of PTN may not represent a physiological response since circulating levels are two orders of magnitude lower, at least in the human circulation (10). The local concentration of PTN within the islets as a result of endogenous expression is not known. These findings suggest that the actions of PTN in islets are more likely to be supportive of β-cell replication, as seen when β-cell mass is expanding prior to adulthood, or during pregnancy, rather than insulin release. This agrees with actions in other tissues such as an ability of PTN to increase the number of hematopoietic stem cells (32), microglia (33) and the growth of hepatic and multiple other cancers (7).

While the intracellular pathways enabling PTN-induced β-cell replication are not known a number of pathways have been demonstrated for other cell types. The full length PTN-18 peptide was shown to enhance cell migration whilst a shorter form of PTN, PTN-15 activated mitogenesis in glioblastoma cells (34). PTN-induced cell replication in endothelial cells has been shown to involve both mitogen-activated protein kinase (MAPK) and the phosphatidylinositol 3-kinase (PI3K)-Akt pathways (35). Also, in glioblastoma cells PTN has been shown to cause a dephosphorylation of PTPRZ1, which is normally constituently active, resulting in an increase in phosphorylated β-catenin. This resulted in increased levels of cytoplasmic β-catenin, translocation to the nucleus, and binding to the transcription factor TCF/LEF with a subsequent expression of Myc and cyclin D1-enabled entry into the cell replication cycle (36).

In summary, the present findings show that Ptn is expressed within the endocrine pancreas in the young mouse, predominantly within β-cells, and pancreatic expression is retained in adults. PTN protein is present on β-cells, possibly by interaction with the RPTP β/ζ receptor or other PTN-binding molecules and has the potential to contribute to β-cell expansion.

The original contributions presented in the study are included in the article/ Supplementary Material . Further inquiries can be directed to the corresponding author.

The studies involving human participants were reviewed and approved by the Western University. The patients/participants provided their written informed consent to participate in this study.

The studies were conceived and planned by JS, TH, MR-A and DH. Experimental work was undertaken by AL, BS, SS and TH. All authors contributed to the writing and review of the manuscript. The present address of TH is the Oxford Centre for Diabetes, Endocrinology and Metabolism, Radcliffe Department of Medicine, Churchill Hospital, Oxford OX3 7LE, United Kingdom. All authors contributed to the article and approved the submitted version.

This work was supported by grants from the Lawson Foundation, Toronto, ON, Canada, by the Canadian Institutes of Health Research (MOP-15263), by the Ministerio de Economía y Competitividad of Spain (RTI2018-095615-B-I00) and by Comunidad de Madrid (B2017/BMD- 3684). SS was a recipient of the Ontario Graduate Scholarship. JS was a recipient of CEU-Santander mobility grant.

The authors declare that they have no conflicts of interest that could be perceived as prejudicing the impartiality of the research reported.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.