Introduction
Insulin-secreting pancreatic beta cells are long-lived and thus vulnerable to various metabolic and cytotoxic assaults. In the case of type 1 diabetes, beta cells are selectively targeted by autoimmune attack and die mostly through apoptosis. Proinflammatory cytokines such as IL-1β, IFN-γ, IL-17 and TNF-α have been shown to be crucial players in this process [
1,
2]. There is a great need to identify factors that could protect beta cells against apoptosis, stimulate beta cell proliferation and thus prevent or reverse the development of diabetes.
Mesencephalic astrocyte-derived neurotrophic factor (MANF) is a small protein with a molecular mass of 18 kDa. It contains an amino-terminal signal peptide that directs it to the endoplasmic reticulum (ER) and, when cleaved, results in a mature protein that can be secreted [
3]. Interestingly, at the carboxy-terminal end MANF has an RTDL motif that is responsible for its retention in the ER and fine-tunes MANF secretion [
3]. MANF was originally discovered as a survival-promoting factor for brain dopaminergic neurones in vitro [
4] and in vivo [
5]. Since then it has also been shown to protect several other cell types, including neuronal cells, cardiomyocytes and HeLa cells [
3,
6‐
8]. Despite extensive studies, the receptors for MANF remain unknown [
3]. It has been suggested that MANF binds via its C-terminal KDEL-like motif RTDL to the KDEL receptor on the cell surface [
9]. It has recently been reported that MANF binds to lipid sulfatides (e.g. 3-
O-sulfogalactosylceramide) and promotes MANF uptake and cytoprotection [
10]. The protective effect of MANF probably depends on its ability to alleviate ER stress [
5,
7]. We previously reported that
Manf knockout mice used as a model of diabetes develop the condition owing to a progressive postnatal reduction of beta cell mass caused by reduced beta cell proliferation and increased beta cell apoptosis [
11]. Additionally, both in vitro and in vivo, MANF was identified as a mitogen for mouse beta cells. Furthermore, a recent study by Cunha et al [
12] showed that thrombospondin 1 protects rat, mouse and human beta cells against cytokine-induced cell death by maintaining the expression of MANF.
Unresolved ER stress and chronic activation of the unfolded protein response (UPR), a cell signalling pathway involved in the restoration of ER homeostasis, are involved in beta cell dysfunction and death in the pathogenesis of both type 1 and type 2 diabetes [
13,
14]. We demonstrated increased expression of UPR markers and sustained phosphorylation of the eukaryotic initiation factor 2 alpha (eIF2α), which leads to global protein synthesis arrest, in islets from
Manf knockout mice [
11]. The mechanism by which lack of MANF induces sustained ER stress in beta cells remains elusive, as does the potential protective effect of this growth factor, particularly when administered as an extracellular protein.
In this study, we tested whether human MANF protein could protect primary and clonal human beta cells against death induced by proinflammatory cytokines. Global transcriptomic analysis was performed to identify molecular mechanisms behind the observed partially protective effects of MANF.
Discussion
We previously showed that global knockout of
Manf in mice results in early-onset diabetes owing to increased beta cell apoptosis and reduced proliferation via persistent activated ER stress-induced UPR pathways [
11]. This led us to investigate the role of MANF as a potential regenerative factor in human beta cells. In the present study, we demonstrate that MANF expression and secretion is stimulated in cytokine-stimulated human beta cells. Furthermore, exogenous MANF protein protects primary and clonal human beta cells against cytokine-induced cell death and may even induce beta cell proliferation.
MANF is expressed at high levels in human beta cells and exocrine cells. Interestingly, we did not detect MANF protein in alpha or pancreatic polypeptide cells. This is in line with the immunohistochemical pattern of MANF expression in mouse islets (T. Danilova, T. Otonkoski and M. Lindahl, unpublished results). The alpha cells were unaffected in the
Manf knockout mice, supporting the idea that MANF is not crucial for alpha cell function [
11]. However, single-cell RNA-seq has shown similar
MANF mRNA levels in human alpha and beta cells [
35]. There are two possible explanations for this: either
MANF mRNA is not translated in alpha cells, or the antibody used in this study does not recognise the form of MANF produced by these cells. We also detected MANF protein throughout the pancreatic epithelium during embryonic development. At this stage, MANF was present also in the insulin and glucagon double-positive cells. MANF seems to have no obvious function in pancreatic organogenesis, since there were no evident defects in embryonic pancreas development in the
Manf knockout mice.
Generation of the human beta cell line EndoC-βH1 provides a valid model of human beta cells for in vitro studies [
16]. These cells mirror the physiological characteristics of human primary beta cells and have also been shown to be sensitive to proinflammatory cytokines [
36]. Recombinant human MANF reduced cytokine-induced cell death in EndoC-βH1 cells. This is in line with a recent study [
12] showing similar protection in mouse beta cells and EndoC-βH1 cells from exogenously administered MANF. Furthermore, in our study, the addition of MANF reduced the expression of ER stress-associated genes and, interestingly,
MANF expression was also reduced. Since
MANF expression is regulated by activating transcription factor 6 (ATF6) and spliced X-box binding protein 1 (sXBP1), which bind to the ER stress element in the
MANF promoter, it is likely that a lower level of UPR can explain the reduced
MANF expression. The knockdown of MANF in these cells significantly aggravated ER stress responses. This is in line with data from knockout mice showing persistent ER stress in beta cells characterised by late embryonic upregulation of the UPR markers
sXbp1 and
Chop followed by increased neonatal expression of
Atf4 and
Chop and sustained pEIF2α activation in the protein kinase RNA-like endoplasmic reticulum kinase (PERK) pathway [
11]. Recently, the timing, amplitude and kinetics of inositol-requiring enzyme 1 (IRE1)α and PERK activation has been found to be important for determining the cellular outcome of the UPR in beta cells [
37]. Our results and those from other groups suggest that MANF is critical for determining the transition from physiological to apoptotic UPR in beta cells where MANF favours the alleviation of ER stress but only if unresolved beta cells switch to the apoptotic PERK/CHOP pathway and apoptosis. However, the exact role for MANF in the regulation of UPR remains to be determined.
NF-κB pathway activation is considered to be an important component of beta cell death induced by inflammatory cytokines [
32]. It has recently been shown that MANF negatively regulates the NF-κB pathway by inhibiting p65-mediated transcriptional activation in fibroblast-like synoviocytes [
38]. MANF is a secreted protein, but it is mostly localised to the lumen of the ER. In inflammation, it has been suggested that it translocates to the nucleus, where it could interfere with the binding of p65 to its target gene promoters, thus suppressing the expression of NF-κB target genes [
38]. However, we did not observe nuclear translocation of MANF in EndoC-βH1 cells under cytokine treatment (data not shown). Analysis of the RNA-seq data identified repression of the NF-κB signalling pathway after addition of MANF, and specifically the downregulation of
BCL10, an inducer of apoptosis acting upstream of NF-κB. In line with this, MANF clearly reduced the nuclear localisation and phosphorylation of RELA/p65 component of the NF-κB complex after the addition of cytokines. Furthermore, si
MANF knockdown exacerbated
BCL10 expression in cytokine-treated EndoC-βH1 cells. These observations indicate that MANF interferes with the NF-κB pathway activation in beta cells upon cytokine treatment.
Interestingly, RNA-seq data showed a global increase in gene expression when the islets were stimulated with cytokines together with MANF. However, addition of MANF alone did not have an effect on the amount of mRNA measured. One of the potential mechanisms could be a reduction of mRNA degradation via suppression of the IRE1-dependent decay of mRNA. Supporting this, prenatal IRE1 activation has been detected in the MANF knockout mice since
sXbp1 expression is already upregulated at E18.5 [
11]. Further studies will be necessary to elucidate the relationship between MANF and IRE1-dependent decay of mRNA regulation.
Most growth factors that stimulate beta cell proliferation in mice have not been effective in human cells. One important confounding factor is the age of the organ donors. It is well characterised that the proliferative capacity of beta cells declines with age due to the accumulation of the cell cycle inhibitor p16
Ink4a [
27,
28]. It was recently shown that TGF-β inhibitors repress the Ink4a/Arf locus, resulting in increased beta cell proliferation in mice and transplanted human islets [
29]. Interestingly, we saw an increase in beta cell proliferation when adding both TGF-β inhibitor and MANF, but not when either one alone was used. One explanation for this could be that the TGF-β inhibitor rejuvenates beta cells, making these cells more responsive to mitogens. Importantly, we were also able to verify the mitogenic effect of MANF in the human EndoC-βH3 model.
Several studies have shown that MANF expression and secretion are increased in situations of cellular stress [
38,
39]. Furthermore, it has recently been shown that MANF levels are elevated in individuals newly diagnosed with type 1 diabetes [
19]. However, based on studies in INS-1E cells, it was recently reported that inflammatory cytokines induce proteasomal degradation of MANF, making the beta cells more prone to apoptosis [
12]. In our study, we found that cultured human beta cells persistently expressed and secreted up to threefold more MANF when exposed to cytokines associated with the pathogenesis of type 1 diabetes. In addition, MANF protein levels were increased despite a global inhibition of protein synthesis that normally occurs during ER stress, suggesting that MANF protein synthesis somehow bypasses PERK–eIF2α-induced translational inhibition. It is likely that secreted MANF exerts its protective effect in an autocrine/paracrine manner. Taken together, it appears that the amount of MANF is crucial for beta cell survival. Further supporting this, MANF expression is reduced in diabetes-susceptible mice compared with diabetes-resistant mice [
40].
In conclusion, we have shown that recombinant MANF partially protects human pancreatic beta cells against proinflammatory-cytokine-induced cell death. Mechanistically, this is at least partially mediated through MANF inhibiting the NF-κB pathway and BCL10. Furthermore, MANF may enhance human beta cell proliferation. Our study elucidates the role of MANF as a protective and mitogenic growth factor for adult human beta cells that could potentially be used to prevent or reverse beta cell loss in diabetes.