Unconventional neurotrophic factors CDNF and MANF: Structure, physiological functions and therapeutic potential

    Unconventional neurotrophic factors CDNF and MANF: structure, physiological functions and therapeutic potential Maria Lindahl, Mart Saarma, P¨aivi Lindholm PII: DOI: Reference:

S0969-9961(16)30171-1 doi: 10.1016/j.nbd.2016.07.009 YNBDI 3798

To appear in:

Neurobiology of Disease

Received date: Revised date: Accepted date:

2 March 2016 29 June 2016 13 July 2016

Please cite this article as: Lindahl, Maria, Saarma, Mart, Lindholm, P¨aivi, Unconventional neurotrophic factors CDNF and MANF: structure, physiological functions and therapeutic potential, Neurobiology of Disease (2016), doi: 10.1016/j.nbd.2016.07.009

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT Unconventional neurotrophic factors CDNF and MANF: structure, physiological functions

T

and therapeutic potential

Institute of Biotechnology, P.O.Box 56, Viikinkaari 5, FI-00014, University of Helsinki, Finland

NU

SC R

a

IP

Maria Lindahla, Mart Saarmaa, Päivi Lindholma,*

MA

*Corresponding author

AC

CE P

TE

D

E-mail address: [email protected] (P. Lindholm)

1

ACCEPTED MANUSCRIPT Abstract Cerebral dopamine neurotrophic factor (CDNF) and mesencephalic astrocyte-derived neurotrophic

T

factor (MANF) promote the survival of midbrain dopaminergic neurons which degenerate in

IP

Parkinson’s disease (PD). However, CDNF and MANF are structurally and functionally clearly

SC R

distinct from the classical, target-derived neurotrophic factors (NTFs) that are solely secreted proteins. In cells, CDNF and MANF localize in the endoplasmic reticulum (ER) and evidence

NU

suggests that MANF, and possibly CDNF, is important for the maintenance of ER homeostasis. MANF expression is particularly high in secretory tissues with extensive protein production and thus

MA

a high ER protein folding load. Deletion of MANF in mice results in a diabetic phenotype and the activation of unfolded protein response (UPR) in the pancreatic islets. However, information about

D

the intracellular and extracellular mechanisms of MANF and CDNF action is still limited. Here we

TE

will discuss the structural motifs and physiological functions of CDNF and MANF as well as their therapeutic potential for the treatment of neurodegenerative diseases and diabetes. Currently

CE P

available knockout models of MANF and CDNF in mice, zebrafish and fruit fly will increase

AC

information about the biology of these interesting proteins.

Keywords: CDNF, MANF, Parkinson’s disease, diabetes, ER stress

2

ACCEPTED MANUSCRIPT 1. Introduction Neurotrophic effects of CDNF and MANF in mammals have mainly been

T

demonstrated in animal models of Parkinson’s disease. Increasing evidence indicates that CDNF and

IP

MANF, when applied as extracellular proteins or delivered by viral vectors can protect and repair

SC R

midbrain dopamine neurons in vivo (Airavaara et al 2012, Back et al 2013, Cordero-Llana et al 2015, Lindholm et al 2007, Voutilainen et al 2011, Voutilainen et al 2009). Neuroprotective effects

NU

of MANF have also been shown in rodent models of cerebral ischemia and spinocerebellar ataxia (Airavaara et al 2010, Yang et al 2014). Importantly, cytoprotective effects of CDNF and MANF are

MA

not restricted to neurons. Infusion of MANF reduced tissue damage in myocardial infarction in mice (Glembotski et al 2012). Deletion of MANF in mice resulted in diabetes indicating the importance of

D

MANF for the functionality of pancreatic insulin-producing beta cells (Lindahl et al 2014).

TE

Although CDNF and MANF show neurotrophic activities, they are structurally clearly

CE P

distinct from the classical secreted NTFs. The latter include glial-cell-line-derived neurotrophic factor (GDNF) family ligands (GFLs) and neurotrophins, which belong to the family of cystine knot cytokines and bind transmembrane receptors to induce intracellular signaling cascades (Airaksinen

AC

& Saarma 2002). GDNF functions by binding to its co-receptor GFRα1, thereby activating receptor tyrosine kinase RET and inducing intracellular signaling which promotes the survival and regeneration of neurons (Paratcha & Ledda 2008). In contrast, the mechanism of CDNF and MANF cytoprotective action is still largely unclear, and their ability to bind transmembrane receptors has not been demonstrated. Although MANF and CDNF can be secreted from cells they are largely retained intracellularly in the ER (Apostolou et al 2008). Interestingly, studies suggest that MANF is important for protein homeostasis in the ER since knockdown of MANF in cultured cells and knockout of MANF in mice and fruit fly results in the activation of UPR, a signaling pathway induced by ER stress (Apostolou et al 2008, Lindahl et al 2014, Palgi et al 2012).

3

ACCEPTED MANUSCRIPT In the present review we will discuss the structural motifs of CDNF and MANF that are important for their function, animal models available to unravel MANF and CDNF biological

T

roles in vivo, and summarize preclinical studies on potential therapeutic effects of CDNF and MANF

IP

for the treatment of neurodegenerative diseases. Our aim is to provide an up-to-date insight to the

SC R

biology of CDNF and MANF trophic factors.

NU

2. Molecular structure of CDNF/MANF protein family

MANF (also known as arginine-rich, mutated in early stage tumors; ARMET) is an

MA

evolutionary conserved protein present in vertebrate and invertebrate species, including Drosophila melanogaster and Caenorhabditis elegans (Petrova et al 2003). CDNF is a paralog of MANF found

D

in vertebrates (Lindholm et al 2007). Amino acid sequence of CDNF/MANF family members

TE

reveals no homology with other proteins. MANF and CDNF are relatively small proteins with a

CE P

molecular weight of 18 kDa, highly soluble and monomeric in neutral solution (Hellman et al 2011, Hoseki et al 2010, Latge et al 2015, Lindholm et al 2007, Mizobuchi et al 2007). Their primary sequence contains an amino-terminal (N-terminal) signal peptide that directs them to the ER and

AC

when cleaved, results in a mature protein which can be secreted (Lindholm et al 2007, Mizobuchi et al 2007, Petrova et al 2003) (Fig. 1A). Originally, MANF was discovered as a survival promoting factor for midbrain dopaminergic neurons in vitro derived from the culture medium of rat type-1 astrocyte ventral mesencephalic cell line (Petrova et al 2003). Sequence analysis of the active protein revealed a homology to a predicted human arginine-rich protein (ARP) of 234 amino acids (Shridhar et al 1996). Based on sequence analysis of different organisms it was concluded that the putative arginine-rich region of human ARP is not translated and the protein was renamed MANF. According to the original report, human MANF is 179 amino acids long and contains a predicted

4

ACCEPTED MANUSCRIPT signal peptide of 21 amino acids, cleavage of which results in a mature protein of 158 amino acids (Petrova et al 2003) (Fig. 1A). Still, there is some discrepancy about the start methionine of human

T

MANF. In the UniProt database (http://www.uniprot.org/), a sequence of human MANF (Acc. No.

IP

P55145) is 182 amino acids long with a signal peptide of 24 amino acids. It is unclear whether Met-

SC R

1 or Met-4 is the initiator methionine in the MANF P55145 sequence.

CDNF was identified by analysis of database sequences homologous to MANF,

NU

cloned, purified and characterized (Lindholm et al 2007). CDNF consists of 187 amino acids and contains a predicted signal peptide of 26 amino acids, cleavage of which results in a mature protein

MA

of 161 amino acids. Amino acid identity between the mature forms of human CDNF and MANF is 59%. CDNF and MANF proteins apparently lack the pro sequence for enzymatic activation that is

D

common for classical NTFs including GFLs. Mature GFLs have seven cysteine residues in their

TE

primary structure whereas CDNF and MANF have eight cysteines with conserved spacing

CE P

(Airaksinen & Saarma 2002, Lindholm et al 2007, Petrova et al 2003) (Fig. 1A). Human MANF was not glycosylated when expressed and secreted from transiently transfected cells (Apostolou et al 2008, Lindholm et al 2008). Human CDNF contains an N-linked

AC

glycosylation site (Apostolou et al 2008) and an O-linked glycosylation site (Sun et al 2011) and both glycosylated and non-glycosylated forms of CDNF are detected in overexpressing cells (Apostolou et al 2008). However, glycosylation is not required for the neuroprotective activity of CDNF or its secretion (Lindholm et al 2007, Sun et al 2011).

2.1. Functions of the two domains A characteristic feature of the primary sequence of CDNF/MANF family proteins is eight conserved cysteine residues which form four disulphide bridges (Hoseki et al 2010, Lindholm et al 2008, Lindholm et al 2007, Parkash et al 2009). Solving the crystal structure of human MANF 5

ACCEPTED MANUSCRIPT revealed a two-domain protein in which the N-terminal domain is homologous to saposin-like proteins (SAPLIPs) (Parkash et al 2009). Solution structure of MANF resolved by nuclear magnetic

T

resonance (NMR) spectroscopy showed that the carboxy-terminal (C-terminal) domain of human

IP

MANF is homologous to SAF-A/B, Acinus and PIAS (SAP) protein superfamily (Hellman et al

SC R

2011) (Fig. 1B). The structure of CDNF highly resembles that of MANF. Structure of the saposinlike N-terminal domain of CDNF has been determined by X-ray crystallography and NMR spectroscopy (Latge et al 2013, Parkash et al 2009). Recently a solution structure of full-length

NU

CDNF, including the C-terminal domain, was resolved by NMR (Latge et al 2015). Intriguingly,

MA

the two domains of CDNF/MANF appear to have distinct functions and a flexible linker between the domains allows them a freedom of orientation in relation to each other which might be an important feature for their mechanism of action (Hellman et al 2011, Hoseki et al 2010, Latge et al

TE

D

2015).

SAPLIPs are versatile proteins with abilities to interact with lipids and membranes

CE P

(Bruhn 2005). The saposin-fold of MANF/CDNF N-terminal domain consists of five alpha helices and a 310 helix in a globular “closed leaf” conformation with a hydrophobic core and three cysteine

AC

bridges stabilizing the structure (Hellman et al 2011, Latge et al 2013, Latge et al 2015, Parkash et al 2009). Porcine NK-lysin and human granulysin having cytolytic activity were identified as the closest structural homologs for the N-terminal domain of MANF and CDNF. Two patches of positively charged lysine and arginine residues are located on the surface of MANF N-terminal domain which could mediate interactions with negatively charged phospholipids (Parkash et al 2009). However, no published data exist about studies of the potential lipid interaction of MANF and CDNF. A dynamic structural region on the surface of CDNF N-terminal domain that might be an important functional site for protein-protein interaction or catalytic activity was located to the helix α4. Another dynamic region formed by residues in the very N-terminal region of CDNF might be important for the signal peptide cleavage (Latge et al 2015). Alternative pairings of cysteine 6

ACCEPTED MANUSCRIPT residues have been shown by mass spectrometry analysis in the N-terminal domain of aphid Armet (Wang et al 2015) and mouse Armet (Mizobuchi et al 2007) (Fig. 1A) which, if verified, may be

T

important for MANF functionality (Wang et al 2015).

IP

SAP domains are putative DNA binding motifs found in a variety of nuclear proteins

SC R

and are predicted to be involved in chromosomal organization (Aravind & Koonin 2000). According to the NMR solution structure, the C-terminal domain of human MANF contains three

NU

alpha helices α6, α7 and α8, of which α6 is loosely formed, whereas the helices α7 and α8 form a helix-loop-helix DNA binding motif (Hellman et al 2011). CDNF C-terminal domain contains two

MA

alpha helices, α6 and α7, and its structure closely resembles the structure of MANF C-terminal domain (Latge et al 2015). The SAP domain of Ku70 (Ku autoantigen p70 subunit; also known as

D

XRCC6) protein was identified as the closest structural homolog of MANF C-terminal domain

TE

(Hellman et al 2011). Ku70, via its SAP domain, is able to interact with the cytoplasmic proapoptotic BCL2 associated X (BAX) protein, thereby keeping it inactive and preventing apoptosis

CE P

(Sawada et al 2003). When MANF or its C-terminal domain was microinjected as a complementary DNA (cDNA) to the nuclei or as a protein to the cytoplasm it protected mouse superior cervical

AC

ganglion (SCG) sympathetic neurons against BAX-mediated apoptosis. Thus, it was speculated that MANF could prevent apoptosis by interacting with BAX (Hellman et al 2011). However, no interaction between MANF and BAX has been demonstrated so far, suggesting that the antiapoptotic activity of MANF is not mediated by a direct interaction between MANF and BAX (Matlik et al 2015). Of note, no survival promoting effects on SCG neurons were observed when MANF was added to the culture medium (Hellman et al 2011). An important functional motif in the C-terminal domain of MANF is a cysteine bridge linking helices α7 and α8. The cysteine bridge is located in a CXXC motif 127CKGC130 in mature human MANF (corresponding motif 132CRAC135 in human CDNF) (Fig. 1A, B). It is now evident that this cysteine bridge is important for MANF activity. Mutating the CKGC motif of human 7

ACCEPTED MANUSCRIPT MANF abolished its intracellular survival-promoting activity in cultured neurons as well as its extracellular cytoprotective activity in a rat model of focal cerebral ischemia (Matlik et al 2015).

T

Although the CXXC motif can be found in thiol-disulphide oxidoreductases, MANF evidently does

IP

not have oxidoreductase activity (Hartley et al 2013, Matlik et al 2015, Mizobuchi et al 2007).

SC R

At the very end of the C-terminus, human CDNF and MANF have KTEL and RTDL sequences, respectively, which resemble the canonical Lys-Asp-Glu-Leu (KDEL) sequence for ER

NU

retention (Raykhel et al 2007) (Fig. 1A, B). KDEL receptors (KDELRs) in the Golgi recognize KDEL and KDEL-like sequences and mediate protein retrieval from the Golgi back to the ER.

MA

Indeed, removal of the RTDL sequence causes MANF to re-localize from the ER to the Golgi and increases MANF secretion from cell lines and primary neurons in vitro (Glembotski et al 2012,

TE

D

Henderson et al 2013, Matlik et al 2015, Oh-Hashi et al 2012).

CE P

2.2. MANF expression and secretion is regulated by ER stress A number of cellular insults including pharmacological perturbations, reduction in ER

AC

calcium stores, viral infections, altered protein glycosylation, and increased or unbalanced protein expression can disrupt protein folding and cause accumulation and aggregation of unfolded proteins in the ER lumen causing ER stress and if prolonged, leading to ER stress induced apoptosis (Szegezdi et al 2006). Chronic ER stress is known to associate with many pathophysiological conditions including diabetes mellitus, ischemia, and neurodegenerative diseases such as PD, Alzheimer´s disease (AD) and amyotrophic lateral sclerosis (ALS) (Eizirik et al 2008, Fonseca et al 2011, Lindholm et al 2006, Szegezdi et al 2006), as well as with glomerular and tubular kidney disease (Inagi et al 2014). ER stress triggers the activation of UPR (Fig. 2A, B), a cellular defense mechanism to alleviate the stress by suppression of protein translation, degrading misfolded proteins through ER8

ACCEPTED MANUSCRIPT associated protein degradation (ERAD), and activating signaling pathways that lead to the production of molecular chaperons involved in the protein folding (Hetz 2012, Szegezdi et al 2006).

T

UPR is mediated through ER transmembrane receptors pancreatic ER kinase-like ER kinase

IP

(PERK), activating transcription factor 6 (ATF6) and inositol-requiring enzyme 1 (IRE1) (Cox et al

SC R

1993, Harding et al 1999, Haze et al 1999, Mori et al 1993); reviewed in (Hetz 2012, Szegezdi et al 2006). Binding of ER chaperone glucose regulated protein 78 (GRP78, also known as BiP) to the ER luminal domain of the receptors helps to maintain them inactive. Upon ER stress, misfolded and

NU

unfolded proteins can sequester GRP78 from the receptors leading to UPR activation (Bertolotti et

MA

al 2000, Dudek et al 2009, Zhou et al 2006b) (see Figure 2 for details). The promoter-region of MANF contains an ER stress response element II (ERSEII)

D

sequence that is recognized by ER stress-inducible transcription factors, cleaved ATF6 and spliced

TE

XBP1 (Lee et al 2003, Mizobuchi et al 2007, Oh-Hashi et al 2013). MANF mRNA and protein levels are induced by ER stress in various cell types in vitro and in vivo (Apostolou et al 2008,

CE P

Glembotski et al 2012, Hartley et al 2013, Lee et al 2003, Lindholm et al 2008, Mizobuchi et al 2007, Tadimalla et al 2008). Under normal conditions MANF mostly localizes to the luminal side

AC

of the ER (Apostolou et al 2008, Glembotski et al 2012, Matlik et al 2015, Mizobuchi et al 2007, Sun et al 2011) (Fig. 2A). However, ER stress has been shown to increase MANF secretion which is unconventional when comparing to other ER stress induced proteins which are usually retained in the ER (Glembotski 2011, Glembotski et al 2012, Tadimalla et al 2008) (Fig. 2B). It has been suggested that MANF functions as a cardiomyokine, a heart-derived secreted factor that is able to affect cardiovascular function (Glembotski 2011). MANF expression was increased in the heart myocytes in a mouse model of myocardial infarction in vivo. Furthermore, ER stress increased the amount of secreted MANF in the culture medium of neonatal rat ventricular myocytes, and extracellularly applied MANF promoted the survival of myocytes in serum starvation and in simulated ischemia (Tadimalla et al 2008). In a mouse model of congenital 9

ACCEPTED MANUSCRIPT nephrotic syndrome and in tunicamycin- and ischemia-reperfusion induced renal tubular ER stress, MANF levels were increased in podocytes and tubular cells. Importantly, in these animal models of

T

kidney disease, MANF was shown to be secreted into the urine coinciding with podocyte and

IP

tubular ER stress (Kim et al 2016). Thus MANF seem to be an ER stress dependent and secreted

SC R

protein.

Molecular mechanisms regulating MANF secretion have been investigated and

NU

current evidence suggests that MANF retention in the ER is dependent of GRP78 and KDELRs. Expression of Manf and Grp78 is similar in mouse tissues (Mizobuchi et al 2007), and

MA

overexpression of MANF together with GRP78 caused accumulation of intracellular MANF, even when the C-terminal RTDL sequence was deleted (Oh-Hashi et al 2012). Using cultured

D

cardiomyocytes and HeLa cells it was demonstrated by cross-linking and immunoprecipitation that

TE

MANF interacts with GRP78. The interaction was shown to be dependent on calcium and depletion of ER calcium e.g. by thapsigargin, an inhibitor of SERCA (sarco/endoplasmic reticulum calcium

CE P

transporting ATPase), triggered MANF secretion that was not dependent on the RTDL motif. It was suggested that under normal ER calcium concentrations MANF is retained in the ER by GRP78 and

AC

also by KDELR (Fig. 2A). If calcium is depleted from the ER, MANF may dissociate from GRP78 leading to increased secretion of MANF, whereas MANF interaction with KDELR remains unaffected (Fig. 2B). The RTDL motif of MANF was hypothesized to function as a weak retention motif interacting with KDELR that fine-tunes MANF secretion (Glembotski et al 2012). A different model suggested that thapsigargin-induced secretion of MANF is dependent on the C-terminal RTDL sequence and MANF interaction with KDELR via the RTDL sequence regulates its secretion from SH-SY5Y neuroblastoma cells (Henderson et al 2013). The same study demonstrated that KDELRs localize to the cell surface, suggesting that MANF binds KDELRs also on the cell surface (Henderson et al 2013). However, in an in vivo model of stroke the RTDL sequence was not needed for the neuroprotective effect of extracellularly applied MANF on cortical neurons (Matlik et al 10

ACCEPTED MANUSCRIPT 2015). Harvey’s group also developed a reporter assay of secreted ER calcium monitoring proteins which contain the C-terminal ASARTDL sequence of MANF and are secreted in response to the

T

ER calcium depletion (Henderson et al 2014).

IP

Unlike MANF, expression of CDNF was not upregulated by tunicamycin-induced ER

SC R

stress in osteosarcoma U2OS cells suggesting that CDNF is expressed constitutively (Apostolou et al 2008). Similarly to MANF, secretion of CDNF was reported to decrease in vitro by

NU

overexpression of GRP78 and KDELR indicating that the mechanisms regulating CDNF and

MA

MANF secretion are fundamentally similar (Norisada et al 2016).

TE

D

3. Tissue expression

CE P

3.1. MANF expression in mouse and human Manf mRNA and protein is widely expressed both in neuronal and non-neuronal tissues shown by in situ hybridization (ISH), reverse transcription polymerase chain reaction (RT-

AC

PCR), Western blot and immunohistochemistry (IHC) analyses. MANF expression in the rodent brain is mainly neuronal as studied by IHC (Lindholm et al 2008, Wang et al 2014). In the central nervous system (CNS) of postnatal and adult mouse, high Manf mRNA and protein expression was detected in several brain areas including the olfactory bulb, cerebral cortex layers II-VI, piriform cortex, hippocampal CA1-CA3 pyramidal regions and dentate gyrus, several hypothalamic nuclei, and spinal cord. In the nigrostriatal dopamine system, low to intermediate levels of MANF were expressed in the striatum and substantia nigra (SN), where MANF-positive neurons were most abundant in the pars reticulata. Some of the MANF-positive neurons in the SN also expressed tyrosine hydroxylase (TH), indicating MANF expression in dopamine neurons (Lindholm et al 2008). In the rat brain, MANF protein expression was suggested to be developmentally regulated 11

ACCEPTED MANUSCRIPT with the expression being highest within two weeks after birth and declining in adult brain (Wang et al 2014).

T

MANF is also widely expressed in the peripheral tissues of adult mouse. Especially

IP

high levels of MANF have been detected by IHC in glandular cells within secretory tissues such as

SC R

pancreatic exocrine acinar cells and endocrine islet beta cells, in the salivary glands and in the testicular spermatocytes (Lindahl et al 2014, Lindholm et al 2008, Mizobuchi et al 2007),

NU

suggesting important roles for MANF in cells with high protein production and secretion. Broad MANF expression was also detected during mouse embryonic development. In

MA

the brain of mouse embryonic day E12.5-E17 embryo, high MANF expression was revealed in the roof of neopallial cortex, median sulcus and in the non-neuronal cells of choroid plexus in lateral

D

ventricles. Low to moderate expression was detected in the developing striatum and SN by IHC. In

TE

the peripheral nervous system (PNS), high MANF expression was detected in the dorsal root

CE P

ganglia, trigeminal ganglia and SCG by IHC (Lindholm et al 2008). In peripheral embryonic tissues MANF expression was high in the developing pancreas, salivary glands, liver and in the cartilage

AC

primordia of vertebra.

MANF expression in human tissues has not been extensively studied. Similarly to mouse, MANF mRNA is expressed in several brain areas and in most peripheral tissues by RT-PCR (Lindholm et al 2008). In accordance with the RT-PCR and rodent data described above, The Human Protein Atlas (www.proteinatlas.org/ENSG00000145050-MANF/tissue) and BioGPS (www.biogps.org) database describe high MANF expression especially in secretory tissues. MANF is also present in human blood serum (Galli, et al. in press).

3.2. CDNF expression in mouse and human

12

ACCEPTED MANUSCRIPT Based on ISH and IHC analyses, mouse Cdnf mRNA and protein levels are generally lower compared to MANF. Cdnf mRNA expression was detected, although at a low level, in most

T

regions of embryonic, postnatal and adult mouse brain as well as in all regions of adult human

IP

brain studied by RT-PCR. Although Cdnf mRNA expression was undetectable in embryonic mouse

SC R

brain by ISH, expression was detected in many brain areas at postnatal day 10 (P10). Similarly with MANF, CDNF protein expression was mainly neuronal in the adult mouse brain by IHC. However, CDNF immunostaining was not detected in TH-positive dopamine neurons in the SN. In fact, only a

NU

few CDNF-positive neurons were detected in the adult mouse SN. CDNF immunoreactivity was

MA

observed in the cerebral cortex of adult mouse brain through layers II-VI, in the hippocampal CA1 and CA3 pyramidal regions and dentate gyrus, in the locus coeruleus of brain stem and in the

D

Purkinje cells of the cerebellum (Lindholm et al 2007).

TE

Resembling the expression pattern of MANF, CDNF mRNA was detected in most of the peripheral non-neuronal human and mouse tissues analyzed. In agreement with mRNA

CE P

expression analysis, CDNF protein was highly expressed in the adult mouse heart, skeletal muscle, salivary glands and testis as studied by Western blot (Lindholm et al 2007). Thus, MANF and

AC

CDNF expression in human and mouse seem to partly overlap in the brain and peripheral organs, however, their intensity of expression differ. CDNF expression is high in the skeletal muscle, heart, lung, and testis compared to the expression in other tissues. MANF expression is very high in endocrine and exocrine tissues compared to other tissues. The broad expression pattern of these factors in mammals suggest distinctive roles and knockdown of Manf and Cdnf genes in mice will reveal some of these functions.

4. In vivo knockout models

13

ACCEPTED MANUSCRIPT 4.1. MANF knockout mice suffer from diabetes To understand the physiological role of MANF in vivo, we generated MANF

T

knockout mice. As knockout of Manf in Drosophila (Palgi et al 2009) and knockdown in zebrafish

IP

(Chen et al 2012) lead to neuronal and dopaminergic phenotypes we were expecting a neuronal

SC R

phenotype in the knockout mouse as well. Surprisingly, conventional Manf -/- and pancreas-specific Pdx-1 Cre::Manf fl/fl mice developed severe insulin-dependent diabetes due to progressive postnatal reduction of beta cell mass, caused by decreased beta cell proliferation and increased apoptosis

NU

(Lindahl et al 2014). We also demonstrated that pancreatic islets of Manf -/- mice displayed

MA

activation of UPR genes and sustained activation of the ER stress marker eukaryotic initiation factor 2 alpha (eIF2α), implicating unresolved ER stress as one possible cause of beta cell failure in the

D

mice. Beta cell proliferation might be attenuated in Manf -/- mice because of translation repression of

TE

cyclin D1 by phosphorylation of eIF2α leading to subsequent cell cycle arrest in G1 and G2/M

CE P

phases (Bourougaa et al 2010, Brewer & Diehl 2000). Alterations in several genes important for ER homeostasis are associated with diabetic phenotypes. Importantly, mutations in the PERK gene associated with Wolcott-Rallison syndrome

AC

(WRS) result in permanent neonatal diabetes in human (Delepine et al 2000). Similarly, Perk mutations in mouse recapitulate many of the defects found in human WRS patients including hyperglycemia and growth retardation (Harding & Ron 2002, Harding et al 2001). Mutations in WFS1 gene in human cause Wolfram syndrome, characterized by neonatal diabetes, deafness and optic atrophy (Inoue et al 1998). Furthermore, knockout studies of UPR genes including eIF2a, p58IPK, Wfs1, Ire1α, Atf4, Atf6a and Xbp1 result in beta cell loss and subsequent diabetes in mice (Hetz 2012, Scheuner & Kaufman 2008). Taken together, many of the UPR proteins, and especially MANF, seem indispensable for the survival of beta cells. Diabetes mellitus is a group of metabolic disorders characterized by hyperglycemia caused by the inability of the endocrine pancreas to maintain sufficient levels of circulating 14

ACCEPTED MANUSCRIPT bioactive insulin. Type 1 diabetes (T1D) results from the autoimmune destruction of insulin producing pancreatic beta cells leading to total insulin deficiency, whereas type 2 diabetes (T2D)

T

develops when the beta cells are no longer able to respond to an increased insulin demand caused

IP

by insulin resistance. Pancreatic beta cells, producing large quantities of secretory insulin, are

SC R

especially sensitive to ER stress (Kaufman et al 2010). Increasing evidence indicates that ER stress and prolonged UPR are the major causes of beta cell destruction both in T1D and T2D. In T2D, chronic high blood glucose, fatty acid exposure and insulin resistance may contribute to unresolved

NU

ER stress in beta cells and beta cell death. Activation of the UPR associated with lipid metabolite accumulation has also been linked to the pathology of insulin resistance in target tissues (Cnop et al

MA

2012). In T1D, pro-inflammatory cytokines may induce activation of NF-κB, which has predominantly pro-inflammatory and pro-apoptotic roles in beta cells. NF-κB activates inducible

TE

D

nitric oxide synthase (iNOS) expression and NO production, triggering ER stress via inhibition of the SERCA pump and consequent ER calcium depletion. Increased UPR leads to increased

CE P

inflammation and upregulation of the production and secretion of pro-inflammatory cytokines and chemokines which further increases local inflammation by recruitment of immune cells leading to

AC

increased insulitis, beta cell death, and in susceptible individuals progression to T1D (Eizirik et al 2013, Kaufman et al 2010). Recent evidence support MANF as a protective factor for beta cells in ER stress. Using a transgenic mouse model with an increased basal load of unfolded proteins and ER stress in a T1D non-obese (NOD) background devoid of autoimmune diabetes, authors demonstrated reduced levels of transcription factor GLIS3 along with defective upregulation of MANF in the UPR, followed by beta cell senescence and death (Dooley et al 2016). Genome-wide association studies has revealed variations in the GLIS3 gene that are associated with risk of both T1D and T2D in humans. In addition, mutations in the GLIS3 have been connected to a rare congenital form of diabetes (Senee et al 2006). Reduced levels of GLIS3 in Glis3 heterozygous mice lead to increased 15

ACCEPTED MANUSCRIPT beta cell apoptosis and diabetes including defective upregulation of Manf expression in genetically diabetes-resistant NOD background. In addition, diabetes was observed in NOD mice fed with high

T

fat diet which caused downregulation of Glis3 expression, followed by decreased Manf expression.

SC R

thereby determining the fate of beta cell survival in ER stress.

IP

Thus, GLIS3, involved both in T1D and T2D, seem to regulate the expression levels of Manf

In a clinical exome sequencing study of patients with neurocognitive phenotypes in a consanguineous population revealed a patient with a homozygote mutation in the MANF gene. In

NU

contrast to T1D in Manf -/- mice, this patient suffers from T2D and obesity, but also from

MA

microcephaly, mild cognitive disability and short stature, hypothyroidism and primary hypogonadism, recapitulating some of the features found in Manf -/- mice (Lindahl et al 2014,

D

Yavarna et al 2015). We recently found significantly increased levels of serum MANF at the

TE

clinical presentation of T1D in prepubertal children, whereas in older children and adolescents with recent-onset T1D and in adults with longer-term T1D MANF levels were comparable to controls

CE P

(Galli, et al. in press). Whether extremely high levels of serum MANF are related to compromised

AC

beta cell function at the onset of T1D in children remains to be investigated.

4.2. Does MANF regulate UPR? As described earlier, MANF has been found to calcium-dependently interact with ER chaperone GRP78, but is released from GRP78 and secreted upon ER stress caused by calcium depletion from the ER (Glembotski et al 2012, Henderson et al 2013). On the other hand, upon ER stress GRP78 is released from the luminal domains of ER transmembrane receptors PERK, ATF6 and IRE1, activating the UPR pathways (Bertolotti et al 2000, Dudek et al 2009, Zhou et al 2006b). As pancreatic beta cells in Manf -/- mice at E18.5 show increased activation of UPR pathways already before any signs of beta cell failure, we hypothesize that MANF-deficiency in beta cells

16

ACCEPTED MANUSCRIPT leads to chronic activation of the UPR receptors (Lindahl et al 2014). Therefore we speculate that MANF is needed for GRP78 to bind UPR receptors and thereby reduce their activity (Fig. 2A).

T

However, there is no direct evidence that MANF is drawn with GRP78 to the same complexes with

IP

UPR receptors, and would need to be further investigated.

SC R

Alternatively, MANF could be needed to maintain the calcium homeostasis in the ER. MANF-deficiency could cause dysregulation of SERCA pump activity and thereby lead to depletion of ER calcium stores, leading to chronic ER stress and UPR activation in pancreatic beta cells.

NU

MANF binding to GRP78 in a calcium-dependent manner and MANF secretion resulting from

D

levels in the ER (Glembotski et al 2012).

MA

depletion of ER calcium (i.e., by thapsigargin) suggests that MANF action is controlled by calcium

TE

4.3. MANF KO mice suffer from growth defect

CE P

Manf -/- mice show a severe diabetes-independent growth defect the reason of which is unclear (Lindahl et al 2014). As MANF is highly expressed in the mouse hypothalamus and

AC

adenohypophysis, deficiency of MANF may cause impairment in the hormonal control of growth. Alternatively, as observed in Perk -/- mice, Manf -/- mice may have defects in neonatal bone formation (Li et al 2003, Wei et al 2008). Levels of insulin-like growth factor 1 (IGF-1) in the liver of neonatal Perk -/- mice were severely reduced, implicating that PERK is a major regulator of IGF1-dependent growth (Li et al 2003, Wei et al 2008). MANF was also found upregulated and secreted into the extracellular matrix from chondrocytes in cartilage growth plates in two ER stressrelated genetic skeletal disease mouse models of multiple epiphyseal dysplasia (Hartley et al 2013), suggesting that MANF might be important for proper bone formation.

4.4. MANF promotes the proliferation of mouse beta cells 17

ACCEPTED MANUSCRIPT Reports on the effects of extracellular MANF on naïve neurons, cells or cell lines have been limited. We found that recombinant MANF significantly increased the proliferation of mouse

T

beta cells in vitro. Using a multiple low dose streptozotocin (MLD-STZ) induced mouse model of

IP

diabetes, we found increased beta cell proliferation and decreased beta cell death in pancreatic

SC R

regions overexpressing MANF by adeno-associated virus serotype 6 (AAV6)-mediated transduction (Lindahl et al 2014). In addition to its cytoprotective effect, MANF is a potent stimulator of beta cell proliferation in vitro, thus constituting a novel protein with therapeutic potential for stimulating

NU

beta cell renewal. Our work provides novel tools to reveal the mechanism of MANF action by

MA

identifying its receptor(s) and signaling pathways in beta cells.

-/-

mice, Cdnf

TE

In contrast to Manf

D

4.5. CDNF knockout mice are viable and fertile -/-

conventional knockout mice are viable and fertile

CE P

and show no obvious defects in growth, life span or glucose metabolism (Lindahl, Chalazonitis,

AC

Saarma, et al. manuscript in preparation).

4.6. Knockdown of the manf mRNA in Danio rerio affects dopamine system Similar to mouse and human, manf mRNA was widely expressed in zebrafish from early embryonic development. The manf transcript had the highest expression level in whole embryos at 2 hours post-fertilization but then gradually declined at later stages, indicating that manf mRNA was maternally contributed during early development. In adult zebrafish, manf expression was high in liver, but lower in kidney and eyes. ISH and IHC revealed expression of Manf mostly in neurons of the developing forebrain, basal ganglia, optic tectum and periventricular grey zone, habenula, thalamic, hypothalamic regions and cerebellum. Interestingly, the main expression

18

ACCEPTED MANUSCRIPT regions were largely restricted to ventricular zones which include proliferative neural stem cells involved in neurogenesis in zebrafish brain. Only some dopaminergic neurons expressed Manf

T

whereas Manf-immunoreative neurons were located close to Th-positive cells in the brain of adult

IP

fish (Chen et al 2012). Expression patterns of Cdnf in zebrafish have not been reported.

SC R

A surprising degree of functional conservation has been demonstrated between human genes implicated in neurodegenerative diseases and their zebrafish orthologues. Pathological

NU

features of PD, AD, Huntington´s disease and ALS have been recapitulated in zebrafish. Both manf and cdnf genes are found in zebrafish, but only manf-deletion has been reported (Chen et al 2012).

MA

Combination of two antisense splice-blocking morpholino oligonucleotides were used to reduce the manf expression in zebrafish. At 3 days post-fertilization (dpf) approximately 75% of manf pre-

D

mRNA and about 85% of the Manf protein was lost. Manf expression was successfully knocked

TE

down in the brain, skin cells, anterior and posterior neuromasts and liver at 5 dpf. Zebrafish with Manf knock-down developed without any apparent defects in the head or any distinguishable gross

CE P

phenotype. However, the number of dopamine neurons in diencephalic region was significantly lower in manf morphants at 3 dpf. Furthermore, dopamine levels measured by high-performance

AC

liquid chromatography were reduced in manf morphants compared to controls. In contrast, the number of GABAergic, serotonergic, orexinergic or histaminergic neurons were not reduced. Importantly, the dopaminergic phenotype in morphants could be rescued by manf mRNA coinjection. Bromodeoxyuridine staining revealed that Manf was not involved in the early proliferation phase of dopaminergic neuronal lineage. Pax2a and Nr4a2 have an important role in the survival and differentiation of dopaminergic progenitor cells (Blin et al 2008, Luo et al 2008, Simon et al 2003). Both pax2a and nr4a2 genes were upregulated in manf morphants which were rescued by zebrafish manf mRNA co-injection. Furthermore, manf mRNA reversed the decrease of th1-positive cells in the diencephalon of the Nr4a2 deficient larvae. These results suggested that

19

ACCEPTED MANUSCRIPT Manf might interact with Pax2 and Nr4a2 in the dopaminergic progenitor cells, and thus play an

IP

T

important role for the differentiation and survival of dopaminergic neurons in zebrafish.

SC R

4.7. Manf knockout in Drosophila causes defects in dopamine system

The proteome of D. melanogaster contains a single homolog of the mammalian MANF/CDNF proteins that is more similar to MANF than CDNF (Palgi et al 2009). The primary

NU

structure of Drosophila Manf (DmManf) shows 52% identity to human MANF and 49% identity to human CDNF, has a signal sequence of 22 amino acids followed by a mature protein of 151 amino

MA

acids, eight conserved cysteines and a putative C-terminal ER retention signal, RSEL.

D

In the nervous system, DmManf expression was detected in larval stages

TE

predominantly in glia, excluding midline glia. Interestingly DmManf is expressed in astrocyte-like glia that surround dopamine neurons but not in neurons (Palgi et al 2009). However, in adult

CE P

Drosophila brain, DmManf was expressed in glia but also shown to partly co-localize with panneuronal marker and dopaminergic marker in cell somas, suggesting that DmManf was expressed in

AC

some dopaminergic cells (Stratoulias & Heino 2015). Outside the nervous system, during larval stages, DmMANF was predominantly expressed in garland cells, salivary glands, Malpighian tubules, fat body, in the gut and in the ring gland which is the major endocrine organ of Drosophila. Subcellular expression of DmManf partially overlapped with markers for the ER, the Golgi, and endosomes (Palgi et al 2012). D. melanogaster is an attractive model organism to study function of genes in vivo. Larvae with a loss of function DmManfmutant allele lacking the first two coding exons hatch normally but stop feeding and show delay in growth. After maternally contributed DmManf has worn out, the DmManfmutants die as late first instar larvae failing to undergo the first larval molt. Interestingly, a selective loss of axonal bundles but not cell bodies of dopamine neurons was 20

ACCEPTED MANUSCRIPT observed prior to the DmManflarval lethality. The lethality of DmManflarvae was rescued to adulthood by ubiquitous overexpression of DmManf, demonstrating that the lethality was solely due

T

to the loss of endogenous DmManf. A full rescue of the lethality was observed using a driver line

IP

overexpressing DmManf in the epidermis and CNS, suggesting that the primary reason for

SC R

DmManflethality was defects in the cuticle formation (Palgi et al 2009). Mutant flies lacking both maternal and zygotic DmManf (DmManfmz96) die during

NU

the final stage of embryogenesis. In contrast to DmManf larvae, DmManfmz96 embryos showed dramatic disorganization of the nervous system. Additionally, their cuticle was severely disrupted

MA

probably due to loss of dopamine, which is also needed for cuticle formation (Palgi et al 2009). Ultrastructural studies of secretory cells in DmManfmz96 mutants revealed accumulation of

TE

D

inclusion vesicles, suggesting disturbances in degradation or membrane transport (Palgi et al 2012). Gene expression profiling of DmManf mutants revealed altered expression of several

CE P

genes involved in dopamine uptake, synthesis, and transport, as well as mitochondrial and nuclear genes involved in mitochondrial respiration known to be associated with PD. Expression of genes

AC

involved in ATP-binding, nucleic acid metabolism, transmembrane transport (e.g. sugar transporters), exocytosis and endosomal recycling as well as DNA replication was downregulated. Notably, among the upregulated genes were clusters of immune and defense response genes and genes annotated to UPR, lipid metabolism, oxidoreductases and cell death. In line with this, increased phosphorylation of eIF2α indicating activated UPR in the DmManfmz96 embryos was demonstrated by Western blotting (Palgi et al 2012). Wing-specific knockdown of DmManf dramatically alters the wing morphology (Lindstrom et al 2016). To identify genetic interaction partners of DmManf, transgenic UAS-RNAi lines of selected genes were targeted to the wing disc and adult wings were monitored for distinct phenotypes in wild-type and DmManf-overexpressing backgrounds. Using this method, DmManf 21

ACCEPTED MANUSCRIPT was found to genetically interact with UPR genes homologous to human GRP78, PERK and XBP1 (Lindstrom et al 2016). Taken together, DmManf is important for neuronal development, dopamine

T

neurite survival and maintenance of dopamine levels, cuticle formation as well as for the wing

IP

development in the fly. Similarly to mammalian MANF, DmManf is a cellular survival factor with

SC R

important roles in cellular homeostasis.

Structural features that are important for DmManf protein activity have been

NU

characterized using Drosophila genetics (Lindstrom et al 2013). When the N-terminal signal sequence of DmManf was deleted the expression of this transgenic construct was not able to rescue

MA

the early larval lethality of DmManf mutants. This suggests that DmManf entry into the ER and then into the secretory pathway is needed for its function during development. In line with studies

D

of human MANF, deletion of the putative ER-retention signal RSEL from the C-terminal end led to

TE

increased secretion and decreased ER localization of DmManf. When expressed in a restricted pattern in DmManf mutant background, DmManf-∆RSEL only partially rescued fly viability

CE P

suggesting that increased secretion of DmManf compromised intracellular functions of DmManf. Positively charged residues which potentially interact with lipids were mutated in DmManf and

AC

tested in genetic rescue experiments in DmManf mutant background. These studies indicated that three lysine residues K79, K83, and K86 located in helix α5 of the N-terminal domain are important for the functionality but not for secretion of DmManf. Similarly to human MANF, a CXXC motif in the C-terminal domain is essential for functionality of DmManf. Ubiquitous expression of DmManf in which the CXXC motif was mutated (C129S) failed to rescue DmManf mutant lethality in vivo (Lindstrom et al 2013). Both human MANF and CDNF were able to rescue the larval lethality when ubiquitously overexpressed in the DmManfbackground indicating that they are functionally conserved. However, MANF and CDNF do not functionally fully complement each other since in restricted expression pattern, excluding expression in the muscle, fat body and gastric caeca, only 22

ACCEPTED MANUSCRIPT human MANF but not CDNF was able to rescue the lethality of DmManf mutant larvae to adulthood. Only full-length DmManf, but not the independent N-and C-terminal domains is

T

sufficient to rescue DmManf mutant lethality, indicating that full-length DmManf protein is

SC R

IP

essential for fly viability (Lindstrom et al 2013).

4.8. Subcellular MANF outside the ER?

NU

According to several reports, cellular MANF localizes in the ER (Apostolou et al

MA

2008, Glembotski et al 2012, Matlik et al 2015, Mizobuchi et al 2007, Sun et al 2011). DmManf immunostaining was also detected in the Golgi, in line with the data showing that MANF is secreted (Palgi et al 2012). However, MANF immunoreactivity has been detected outside the

TE

D

secretory pathway as well. In Drosophila, partial co-localization of DmManf with endosomal markers was observed in larval salivary gland cells (Palgi et al 2012), suggesting that MANF is

CE P

transported from the secretory pathway to endosomes or, alternatively, extracellular mature MANF is endocytosed to cells. Indeed, whether extracellularly applied CDNF/MANF is endocytosed

AC

remains to be studied. However, DmManf synthetized in the cytoplasm without a signal peptide failed to rescue DmManf mutant lethality and seemed unstable (Lindstrom et al 2013), suggesting that MANF does not function as a soluble cytoplasmic protein. Still, in contrast, MANF microinjected into the cytoplasm of mouse SCG neurons shows neuroprotective activities (Hellman et al 2011). Future studies will reveal if endogenous cytoplasmic MANF exists in vivo. Nuclear localization of MANF has also been reported. MANF immunoreactivity was detected in the nuclei of rodent glial cells after focal cerebral ischemia (Shen et al 2012). In patients with autoimmune systemic lupus erythematosus and rheumatoid arthritis MANF mRNA was upregulated in peripheral white blood cells. The authors found that inflammation inducing lipopolysaccharide (LPS) and glycosylation-inhibitor tunicamycin caused relocalization of MANF 23

ACCEPTED MANUSCRIPT together with p65, a subunit of transcription factor NF-κB, to the nuclei of rabbit synovial fibroblast-like cells. MANF inhibited DNA binding of p65, thus suppressing inflammatory

T

pathways induced by nuclear NF-κB binding to its target genes, suggesting that MANF could

IP

reduce the inflammatory response and prevent proliferation of inflamed cells (Chen et al 2015). The

SC R

mechanism of MANF relocalization to the nucleus remains to be investigated.

NU

5. MANF and CDNF in disease models

MA

5.1. Effects of CDNF and MANF proteins in animal models of Parkinson’s disease PD is a progressive neurodegenerative disorder associated with motor and non-motor

D

dysfunctions (Bartels & Leenders 2009). The precise relationship among most of these dysfunctions

TE

and the underlying pathology is unfortunately not clear. However, many of the main motor symptoms of the disease appear to result from the loss of dopamine neurons projecting from the SN

CE P

to the caudate putamen (Dauer & Przedborski 2003). Commissiong’s group purified MANF and demonstrated that it can support the

AC

survival of embryonic dopamine neurons in culture (Petrova et al 2003). In further studies, MANF was shown to enhance GABAergic inhibition to dopamine neurons which was suggested to mediate the protective activity of MANF to dopamine neurons (Zhou et al 2006a). Since MANF and CDNF are both expressed in the striatum i.e., in the target field of the nigral dopamine neurons (Lindholm et al 2008, Lindholm et al 2007), and DmManf-deficient flies had defects specifically in dopamine neurons and also dramatically decreased levels of dopamine (Palgi et al 2009) it was of great interest to study the effects of CDNF and MANF in animal models of PD. The two most studied growth factors in PD, GDNF and neurturin (NRTN), efficiently protect and repair dopamine neurons in animal models of PD, but have given only modest clinical benefit in phase 2 clinical

24

ACCEPTED MANUSCRIPT trials (Kordower & Bjorklund 2013). Thus, there is a need to find new dopaminotrophic factors for the development of disease modifying drugs for PD.

T

In the first study CDNF was delivered 6 hours after neurotoxin and it dose-

IP

dependently protected dopamine neurons in a rat 6-hydroxydopamine (6-OHDA) model of PD.

SC R

Most importantly, CDNF was also able to protect and restore the function of nigral dopamine neurons even when given one month after the 6-OHDA injection i.e., during progressive

NU

degeneration of dopamine nerve endings from the 6-OHDA injection site towards the SN (Lindholm et al 2007). In these experiments CDNF was as efficient as GDNF and a single injection

MA

of CDNF produced a long-lasting beneficial effect on rat motor behavior. Similarly to CDNF intrastriatally injected MANF also protected nigrostriatal dopamine neurons from 6-OHDA-induced

TE

striatum (Voutilainen et al 2009).

D

degeneration when administered either 6 hours before or 4 weeks after 6-OHDA injection in the

CE P

The effects of two-week continuous infusions of CDNF, MANF and GDNF were studied and compared in rats using a severe 6-OHDA hemiparkinsonian model. The amphetamineinduced rotational behavior was gradually normalized in rats treated with CDNF for two weeks

AC

following the intrastriatal 6-OHDA injection. CDNF was also able to inhibit 6-OHDA-induced loss of dopamine neurons in the SN and restore TH-positive fibers in the striatum, whereas MANF and GDNF at similar doses had no statistically significant effect in any of the measured parameters (Voutilainen et al 2011).The low benefit of GDNF and NRTN in phase 2 clinical trials on PD patients can possibly be partially ascribed to their poor diffusion in the brain parenchyma (Domanskyi et al 2015, Hamilton et al 2001). Indeed, the calculated (http://web.expasy.org/cgibin/compute_pi/pi_tool) pI of monomeric fully processed GDNF (UniProt P39905) is 9.44 and that of NRTN (Q99748) is 9.01, and they both bind with high affinity to heparan sulfate proteoglycans (Bespalov et al 2011, Hamilton et al 2001). Using 125I-labelled growth factors the diffusion of the CDNF and MANF was compared to that of GDNF. The distribution volume of MANF and CDNF 25

ACCEPTED MANUSCRIPT in the striatum was significantly larger than that of GDNF after 24-hour and 3-day infusions suggesting that differently from GDNF and NRTN, CDNF and MANF do not bind to heparan

T

sulfates in the extracellular matrix. The molecular weight of CDNF/MANF is lower than that of

IP

homodimeric GDNF and NRTN [i.e., 32-42 kDa (Lin et al 1993) and 25 kDa (Kotzbauer et al

SC R

1996), respectively] possibly facilitating their diffusion in the brain. Although the pI 8.6 of MANF is higher than that of CDNF (pI 7.7) MANF diffused slightly better than CDNF in the brain tissue (Voutilainen et al 2011). The reason for this difference is unclear. Both 125I-CDNF and 125I-MANF

NU

were retrogradely transported from the striatum to other brain regions. The CDNF transportation profile seems to follow that of GDNF, whereas the mechanism of MANF action apparently differs

MA

from that of GDNF and CDNF since intrastriatally injected 125I-MANF was transported to the frontal cortex, whereas 125I-labelled GDNF and CDNF were transported mainly to the SN. These

TE

D

data suggest that CDNF and MANF can efficiently protect and rescue dopamine neurons in rat 6OHDA PD models, and in severe rat 6-OHDA neurorestoration models of PD CDNF seems to be

CE P

the most efficient. Both CDNF and MANF diffuse better than GDNF in the brain and are retrogradely transported from the striatum to SN, but the mechanism of transportation for MANF

AC

seems to be different (Voutilainen et al 2011). CDNF protein has also been tested in a mouse 1-methyl-4-phenyl-1,2,3,6tetrahydropyridine (MPTP) model of PD. MPTP is a neurotoxin, which produces parkinsonian symptoms in humans, monkeys and in C57/Bl6 mice. To study whether CDNF protein has neuroprotective and neurorestorative properties for the nigrostriatal dopamine system, C57/Bl6 mice received MPTP systemically, and CDNF protein was delivered intrastriatally. CDNF delivery 20 hours before MPTP improved horizontal and vertical motor behavior, increased TH immunoreactivity in the striatum, as well as the number of dopamine neurons in the SN. In the neurorestoration paradigm CDNF was delivered 1 week after MPTP injections and it increased horizontal and vertical motor behavior of mice, as well as dopamine fiber densities in the striatum 26

ACCEPTED MANUSCRIPT and the number of TH-positive cells in SN (Airavaara et al 2012). Unexpectedly, CDNF did not alter any of the analyzed dopaminergic biomarkers or locomotor behavior in MPTP-untreated mice,

T

whereas GDNF is known to induce axonal sprouting and also enhance dopamine uptake in healthy

IP

animals. In line with this observation in naïve mice, similar results with CDNF were also obtained

SC R

in naïve rats (Voutilainen et al 2011). These data show that CDNF is both neuroprotective and neurorestorative for the TH-positive cells in the nigrostriatal dopamine system in the mouse MPTP model, and indicate that differently from GDNF, CDNF has minimal or no effects on naïve nigral

NU

dopamine neurons. Effects of CDNF are currently studied and compared with the effects of GDNF

MA

in a rhesus monkey MPTP model of PD (Saarma, Cameron, et al. unpublished). Recently, toxicology studies on non-human primates using good manufacturing practice human CDNF were conducted and demonstrated the safety of CDNF (Saarma, personal communication). These data

TE

D

suggest that CDNF holds promise as a clinical therapy for PD patients. Translational potential of convection-enhanced delivery (CED) of MANF for the

CE P

treatment of PD was recently assessed by studying its distribution in porcine brain. MANF was coinfused with gadolinium-DTPA into putamen and SN using an implantable CED catheter system to

AC

allow real-time magnetic resonance imaging (MRI) of infusate distribution (Barua et al 2015). The distribution of gadolinium-DTPA on MRI correlated well with immunohistochemical analysis of MANF distribution and confirmed the translational potential of CED of MANF as a novel treatment strategy in PD. Characteristic neuropathological feature of PD is the accumulation of Lewy bodies which are mainly composed of fibrillary aggregated alpha-synuclein (Goedert et al 2013). As activated UPR has been associated with alpha-synuclein aggregation in the brain of PD patients (Hoozemans et al 2007) and in mouse PD models (Mercado et al 2013), investigating the role of MANF and CDNF in alpha-synuclein-related UPR would be of great interest.

27

ACCEPTED MANUSCRIPT

5.2. CDNF and MANF gene therapy in animals models of PD

IP

T

Gene therapy strategies with GDNF and NRTN in rodents and non-human primate models of PD have produced promising results providing structural and functional recovery

SC R

(Kordower & Bjorklund 2013). GDNF protein delivery and NRTN gene therapy have been successfully translated to clinical trials with success in open-label clinical trials, although this has

NU

not yet produced significant clinical benefit in phase 2 double-blinded studies (Domanskyi et al 2015).

MA

AAV has been used as an effective and safe delivery vehicle for the NRTN clinical trial showing efficacy especially in the early stages of the disease (Bartus & Johnson 2016).

D

Collectively, the results of NTF gene therapy studies for the treatment of PD will require further

TE

development. Gene therapy, using regulated viral vectors and novel dopaminotrophic factors has the

CE P

potential to be a safe and effective new therapeutic option for PD. CDNF gene therapy has been tested in three independent studies. In the first two

AC

studies CDNF was delivered using AAV serotype 2 and in the third study lentiviral delivery of CDNF was used. In the first study AAV2-CDNF was injected into striatum in a rat 6-OHDA model of PD 2 weeks before the neurotoxin and the effects were compared with those of AAV2-GDNF. Treatment with AAV2-CDNF resulted in a marked decrease in amphetamine-induced ipsilateral rotations while it provided only partial protection of dopamine neurons and TH-reactive fibers in the striatum (Back et al 2013) In the second study, AAV2-CDNF was delivered to the striatum 6 weeks after 6-OHDA. Striatal delivery of AAV2-CDNF was able to recover 6-OHDA-induced behavior deficits and resulted in a significant restoration of dopamine neurons in SN and THpositive fiber density in the striatum. In addition to neurohistochemical analysis, positron emission tomography (PET) analyses were used with [11C]-2β-carbomethoxy-3β-(4-fluorophenyl)-tropane

28

ACCEPTED MANUSCRIPT ([11C] β-CFT) probes and the results strongly suggested functional recovery of dopamine system (Ren et al 2013).

T

Intranigral or intrastriatal lentiviral vector-mediated expression of CDNF and MANF

IP

was used to evaluate their efficacy in the 6-OHDA model of PD in rats (Cordero-Llana et al 2015).

SC R

Surprisingly, no beneficial effects were found by striatal overexpression of either protein, whereas GDNF, used as the positive control, was neuroprotective. However, nigral overexpression of CDNF

NU

decreased amphetamine-induced rotations and increased TH-positive striatal fiber density but had no effect on numbers of TH-positive cells in the SN. On the contrary, nigral MANF overexpression

MA

had no effect on amphetamine-induced rotations or TH-positive striatal fiber density but resulted in a significant protection of dopamine neurons. Combined nigral overexpression of both factors led to

D

a robust reduction in amphetamine-induced rotations, greater increase in striatal TH-positive fiber

TE

density and significant protection of dopamine neurons in the SN. This interesting study, however,

CE P

did not report which cell types overexpressed CDNF and MANF and at what intracellular locations.

AC

5.3. CDNF and MANF effects on other neurological diseases In addition to studies in PD models, effects of CDNF and MANF have been tested in animal models of AD, brain ataxia and also in peripheral nerve injury. Here we very briefly summarize the main observations of these studies. In an APP/PS1 mouse model of AD, intrahippocampal injections of CDNF protein or CDNF transgene in an AAV2 viral vector was used. APP/PS1 mice co-express chimeric mouse/human amyloid precursor protein (APP) harboring mutations linked with Swedish familial AD pedigrees and mutant human presenilin 1 (PS1) in neurons and astrocytes in the CNS, which leads to accumulation of amyloid deposits in the brain (Jankowsky et al 2004). Quite unexpectedly, CDNF-therapy improved long-term memory in both 1-year-old APP/PS1 mice and wild-type 29

ACCEPTED MANUSCRIPT controls, but was not associated with decreased brain amyloid load or enhanced hippocampal neurogenesis. On the other hand intrahippocampal CDNF did not affect spontaneous exploration,

T

object neophobia or early stages of spatial learning (Kemppainen et al 2015). Although the

IP

intracranial CDNF treatment has beneficial effects on long-term memory its mechanism of action

SC R

on memory warrants further studies.

In AD models, PERK phosphorylation has been linked to activation of Aβ-producing β-secretase (BACE1), tau hyperphosphorylation, and memory impairment and neuronal loss (Devi

NU

& Ohno 2014, van der Harg et al 2014). Decreased insulin-signaling (T1D) and insulin-resistance

MA

(T2D) are factors contributing to higher risk for the development of PD, cognitive decline and agerelated memory impairment in humans (Blazquez et al 2014, Kleinridders et al 2014). As blood

D

insulin-levels are reduced with age in conventional MANF knockout mice, it would be of great

TE

interest to dissect the effect of MANF-removal both in conventional diabetic mice and in nonfl/fl

mice considering PD and AD pathologies.

CE P

diabetic brain-specific Nestin-Cre::Manf

Changes in the expression levels of Manf mRNA in the hippocampus and cerebral cortex were reported after experimentally induced global forebrain ischemia and epileptic insult in

AC

mice (Lindholm et al 2008). Thereafter AAV7-MANF protective effects were studied in a rat model of stroke in vivo. The right middle cerebral artery was ligated for 1 hour one week after the vector injections. Interestingly, AAV7-MANF expression reduced the volume of cerebral infarction and improved behavioral recovery of the rats (Airavaara et al 2010). In accordance, in a rat model of focal cerebral ischemia, MANF immunoreactivity was upregulated in neurons in the peri-infarct regions (Yu et al 2010). Thus, MANF might be involved in the protection of existing neurons and in the regeneration and migration of newly formed neurons to the injury site in brain ischemia. A spinocerebellar ataxia 17 (SCA17) knock-in mouse model has been established where induction of expression of a mutant TATA box binding protein (TBP) leads to TBP accumulation in Purkinje cells (Yang et al 2014). TBP accumulation in older mice correlates with 30

ACCEPTED MANUSCRIPT age-related decreases in chaperone HSC70 activity and more severe Purkinje cell degeneration. Surprisingly, mutant TBP shows decreased association with XBP1s, resulting in the reduced

T

transcription of MANF, particularly in Purkinje cells. Overexpression of MANF reduces mutant

IP

TBP-mediated Purkinje cell degeneration suggesting an important role of MANF in Purkinje cell

SC R

maintenance. Further studies on MANF knockout mice are required to reveal the role of MANF in Purkinje cell development and survival. Since CDNF is also expressed in Purkinje cells at high

NU

levels it would be interesting to test its activity in spinocerebellar ataxia models. Peripheral nerve injury (PNI) often results in axonal degeneration and the loss of

MA

neurons, leading to limited nerve regeneration and severe functional impairment. Currently, there are no effective treatments for PNI. Effect of lentiviral-mediated transfer of CDNF on regeneration

D

of the rat sciatic nerve was tested in a transection model in vivo (Cheng et al 2013). CDNF

TE

expression resulted in significant improvement of axonal and Schwann cell regeneration and an increase in the thickness of the myelination around the axons. In an independent study

CE P

mesenchymal stem cells (MSCs) were transduced with lentiviral CDNF and placed into collagen tubes to investigate their regenerative effects on the same model (Liu et al 2014). Again CDNF-

AC

MSCs treatment showed significant enhancement of axonal and Schwann cell regeneration, increased myelination thickness, axon diameter and the axon-to-fiber diameter ratio. Thus, CDNF promotes peripheral nerve regeneration and functional recovery, but further studies are required to dissect the molecular and cellular effects of CDNF in this PNI model. CDNF therapeutic effects have recently been tested in traumatic spinal cord injury (SCI) where effective therapies are lacking (Zhao et al 2016). Bone marrow-derived MSCs overexpressing CDNF were developed which after transplantation suppressed neuroinflammation, decreased the production of pro-inflammatory cytokines after SCI, and promoted the locomotor function and nerve regeneration of the injured spinal cord. This study presents a novel promising strategy for the treatment of SCI. 31

ACCEPTED MANUSCRIPT Taken together, CDNF and MANF have mostly been tested in animal models of PD and the first studies in animal models of other neurological diseases have been promising.

T

Nevertheless additional information about the mechanism of action, site of delivery, optimal doses

SC R

IP

and delivery regimens for CDNF and MANF are required.

5.4. MANF and CDNF in neuroinflammation

NU

Microglia function in the innate immune system in the brain and are important for neuronal plasticity, neurogenesis and tissue homeostasis. In several neurodegenerative diseases such

MA

as AD, PD, ALS and ischemia, microglia are activated by misfolded proteins leading to release of pro-inflammatory and cytotoxic factors that contribute to neuronal damage (Heneka et al 2014)

D

Whereas microglia represent the major types of immune cells in the CNS, astrocytes are the main

TE

supportive cells of neurons in the brain. In response to acute injury they can stimulate the secretion

CE P

of chemokines and pro-inflammatory cytokines, enhancing the inflammatory response. Conversely, astrocytes as well as microglia may contribute to the neuronal growth and repair by secreting

AC

neurotrophic factors (Amor et al 2010). CDNF may function as a regulator of inflammation as it was found upregulated in cultured rat primary microglia activated with LPS. Pretreatment with CDNF attenuated LPS induced secretion of pro-inflammatory cytokines and inhibited LPS induced cytotoxicity, possibly by suppressing the activation of JNK pathway in the microglia (Zhao et al 2014). Additionally, CDNF transfected into SN using a nanoparticle carrier neurotensin-polyplex was shown to reduce glial markers and interleukin-6 levels in 6-OHDA-induced parkinsonian rats (Nadella et al 2014). MANF expression was upregulated in activated microglia and also in astrocytes in rat ischemic cortex after focal cerebral ischemia (Shen et al 2012). In an in vitro model of ischemia, pretreatment of rat primary astrocytes with MANF inhibited oxygen-glucose deprivation-induced cell damage

32

ACCEPTED MANUSCRIPT and decreased the secretion of pro-inflammatory cytokines by suppressing ER stress (Zhao et al 2013). Thus, CDNF and MANF may function as neuroprotective factors by dampening the

IP

T

inflammatory response in the CNS.

SC R

6. Challenges and future prospects

Although MANF was discovered in 2003 and CDNF in 2007 our knowledge about the

NU

basic biology, mode(s) of action and therapeutic potential of these very unique and exciting proteins is still limited. The major challenge is to understand the mechanism of action of these proteins.

MA

Compared to other NTFs this is very challenging since CDNF and MANF act inside the cells and most obviously also as extracellular proteins. Although the analysis of MANF-deficient mice has

D

revealed its important role in the development and maintenance of pancreatic beta cells and raised

TE

hopes that it can be developed as a novel drug for the treatment of diabetes, several challenges still

CE P

remain. It is of great interest to understand what is the in vivo role of MANF for neurons, especially for midbrain dopamine neurons, Purkinje cells, and for cortical and hippocampal neurons. Its strong

AC

expression in several peripheral secretory organs predicts a role for MANF there also. Likewise, it is of great interest to understand the role of CDNF both in the CNS and PNS, and particularly in the dopamine neurons.

The question whether MANF and CDNF are involved in the regulation of ER stress and UPR is particularly interesting. Are they directly involved in the regulation of UPR and thus playing a key role in protein aggregation? Recent data suggest that CDNF regulates alpha-synuclein aggregation in neurons (Latge et al 2015). Thus it is of great interest to study the mechanism of CDNF and MANF in the protein folding and regulation of UPR. Experiments on animal models of PD give hope that CDNF and MANF can soon enter clinical trials and hopefully offer an alternative and efficient strategy for the development of novel 33

ACCEPTED MANUSCRIPT disease modifying therapies. Although MANF and CDNF have been tested in a limited number of disease models, the data obtained are very promising. Initial studies also encourage further studies

T

of these factors in animal models of AD, stroke and spinal cord injury. In the coming years genome

IP

wide association studies and other genomic approaches will be utilized to investigate changes in the

SC R

expression levels of CDNF and MANF in different diseases and in the search for disease associated mutations in CDNF and MANF. Furthermore, evidence of increased MANF levels in the urine of mouse models of ER stress-related kidney disease (Kim et al 2016) and in the sera of newly

NU

diagnosed T1D patients (Galli, et al. in press) suggest that MANF might have a strong potential as a

MA

noninvasive biomarker in ER stress-related diseases.

D

Acknowledgements

TE

We acknowledge the support from Jane and Aatos Erkko Foundation, Juvenile Diabetes Research

CE P

Foundation (grants 17-2013-410, 1-INO-2014-162-A-V), Michal J. Fox Foundation for Parkinson’s Research, and the Academy of Finland (grant 268662). We are grateful to Dr. Mikko Airavaara, Dr.

AC

Urmas Arumäe, and Dr. Riitta Lindström for critical reading of the manuscript, and to Katrina Albert and Ave Eesmaa for language editing.

References Airaksinen MS, Saarma M. 2002. The GDNF family: signalling, biological functions and therapeutic value. Nature reviews. Neuroscience 3: 383-94 Airavaara M, Chiocco MJ, Howard DB, Zuchowski KL, Peranen J, et al. 2010. Widespread cortical expression of MANF by AAV serotype 7: localization and protection against ischemic brain injury. Experimental neurology 225: 104-13 Airavaara M, Harvey BK, Voutilainen MH, Shen H, Chou J, et al. 2012. CDNF protects the nigrostriatal dopamine system and promotes recovery after MPTP treatment in mice. Cell transplantation 21: 1213-23 Ameri K, Harris AL. 2008. Activating transcription factor 4. The international journal of biochemistry & cell biology 40: 14-21 34

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

MA

NU

SC R

IP

T

Amor S, Puentes F, Baker D, van der Valk P. 2010. Inflammation in neurodegenerative diseases. Immunology 129: 154-69 Apostolou A, Shen Y, Liang Y, Luo J, Fang S. 2008. Armet, a UPR-upregulated protein, inhibits cell proliferation and ER stress-induced cell death. Experimental cell research 314: 2454-67 Aravind L, Koonin EV. 2000. SAP - a putative DNA-binding motif involved in chromosomal organization. Trends in biochemical sciences 25: 112-4 Back S, Peranen J, Galli E, Pulkkila P, Lonka-Nevalaita L, et al. 2013. Gene therapy with AAV2CDNF provides functional benefits in a rat model of Parkinson's disease. Brain and behavior 3: 75-88 Bartels AL, Leenders KL. 2009. Parkinson's disease: the syndrome, the pathogenesis and pathophysiology. Cortex; a journal devoted to the study of the nervous system and behavior 45: 915-21 Bartus RT, Johnson EM, Jr. 2016. Clinical tests of neurotrophic factors for human neurodegenerative diseases, part 2: Where do we stand and where must we go next? Neurobiology of disease Barua NU, Bienemann AS, Woolley M, Wyatt MJ, Johnson D, et al. 2015. Convection-enhanced delivery of MANF--volume of distribution analysis in porcine putamen and substantia nigra. Journal of the neurological sciences 357: 264-9 Bertolotti A, Zhang Y, Hendershot LM, Harding HP, Ron D. 2000. Dynamic interaction of BiP and ER stress transducers in the unfolded-protein response. Nature cell biology 2: 326-32 Bespalov MM, Sidorova YA, Tumova S, Ahonen-Bishopp A, Magalhaes AC, et al. 2011. Heparan sulfate proteoglycan syndecan-3 is a novel receptor for GDNF, neurturin, and artemin. The Journal of cell biology 192: 153-69 Blazquez E, Velazquez E, Hurtado-Carneiro V, Ruiz-Albusac JM. 2014. Insulin in the brain: its pathophysiological implications for States related with central insulin resistance, type 2 diabetes and Alzheimer's disease. Frontiers in endocrinology 5: 161 Blin M, Norton W, Bally-Cuif L, Vernier P. 2008. NR4A2 controls the differentiation of selective dopaminergic nuclei in the zebrafish brain. Molecular and cellular neurosciences 39: 592604 Bourougaa K, Naski N, Boularan C, Mlynarczyk C, Candeias MM, et al. 2010. Endoplasmic reticulum stress induces G2 cell-cycle arrest via mRNA translation of the p53 isoform p53/47. Molecular cell 38: 78-88 Brewer JW, Diehl JA. 2000. PERK mediates cell-cycle exit during the mammalian unfolded protein response. Proceedings of the National Academy of Sciences of the United States of America 97: 12625-30 Bruhn H. 2005. A short guided tour through functional and structural features of saposin-like proteins. The Biochemical journal 389: 249-57 Chen L, Feng L, Wang X, Du J, Chen Y, et al. 2015. Mesencephalic astrocyte-derived neurotrophic factor is involved in inflammation by negatively regulating the NF-kappaB pathway. Scientific reports 5: 8133 Chen YC, Sundvik M, Rozov S, Priyadarshini M, Panula P. 2012. MANF regulates dopaminergic neuron development in larval zebrafish. Developmental biology 370: 237-49 Cheng L, Liu Y, Zhao H, Zhang W, Guo YJ, Nie L. 2013. Lentiviral-mediated transfer of CDNF promotes nerve regeneration and functional recovery after sciatic nerve injury in adult rats. Biochemical and biophysical research communications 440: 330-5 Cnop M, Foufelle F, Velloso LA. 2012. Endoplasmic reticulum stress, obesity and diabetes. Trends in molecular medicine 18: 59-68 Cordero-Llana O, Houghton BC, Rinaldi F, Taylor H, Yanez-Munoz RJ, et al. 2015. Enhanced efficacy of the CDNF/MANF family by combined intranigral overexpression in the 6-

35

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

MA

NU

SC R

IP

T

OHDA rat model of Parkinson's disease. Molecular therapy : the journal of the American Society of Gene Therapy 23: 244-54 Cox JS, Shamu CE, Walter P. 1993. Transcriptional induction of genes encoding endoplasmic reticulum resident proteins requires a transmembrane protein kinase. Cell 73: 1197-206 Dauer W, Przedborski S. 2003. Parkinson's disease: mechanisms and models. Neuron 39: 889-909 Delepine M, Nicolino M, Barrett T, Golamaully M, Lathrop GM, Julier C. 2000. EIF2AK3, encoding translation initiation factor 2-alpha kinase 3, is mutated in patients with WolcottRallison syndrome. Nature genetics 25: 406-9 Devi L, Ohno M. 2014. PERK mediates eIF2alpha phosphorylation responsible for BACE1 elevation, CREB dysfunction and neurodegeneration in a mouse model of Alzheimer's disease. Neurobiology of aging 35: 2272-81 Domanskyi A, Saarma M, Airavaara M. 2015. Prospects of Neurotrophic Factors for Parkinson's Disease: Comparison of Protein and Gene Therapy. Human gene therapy 26: 550-9 Dooley J, Tian L, Schonefeldt S, Delghingaro-Augusto V, Garcia-Perez JE, et al. 2016. Genetic predisposition for beta cell fragility underlies type 1 and type 2 diabetes. Nature genetics 48: 519-27 Dudek J, Benedix J, Cappel S, Greiner M, Jalal C, et al. 2009. Functions and pathologies of BiP and its interaction partners. Cellular and molecular life sciences : CMLS 66: 1556-69 Eizirik DL, Cardozo AK, Cnop M. 2008. The role for endoplasmic reticulum stress in diabetes mellitus. Endocrine reviews 29: 42-61 Eizirik DL, Miani M, Cardozo AK. 2013. Signalling danger: endoplasmic reticulum stress and the unfolded protein response in pancreatic islet inflammation. Diabetologia 56: 234-41 Fonseca SG, Gromada J, Urano F. 2011. Endoplasmic reticulum stress and pancreatic beta-cell death. Trends in endocrinology and metabolism: TEM 22: 266-74 Glembotski CC. 2011. Functions for the cardiomyokine, MANF, in cardioprotection, hypertrophy and heart failure. Journal of molecular and cellular cardiology 51: 512-7 Glembotski CC, Thuerauf DJ, Huang C, Vekich JA, Gottlieb RA, Doroudgar S. 2012. Mesencephalic astrocyte-derived neurotrophic factor protects the heart from ischemic damage and is selectively secreted upon sarco/endoplasmic reticulum calcium depletion. The Journal of biological chemistry 287: 25893-904 Goedert M, Spillantini MG, Del Tredici K, Braak H. 2013. 100 years of Lewy pathology. Nature reviews. Neurology 9: 13-24 Hamilton JF, Morrison PF, Chen MY, Harvey-White J, Pernaute RS, et al. 2001. Heparin coinfusion during convection-enhanced delivery (CED) increases the distribution of the glial-derived neurotrophic factor (GDNF) ligand family in rat striatum and enhances the pharmacological activity of neurturin. Experimental neurology 168: 155-61 Harding HP, Ron D. 2002. Endoplasmic reticulum stress and the development of diabetes: a review. Diabetes 51 Suppl 3: S455-61 Harding HP, Zeng H, Zhang Y, Jungries R, Chung P, et al. 2001. Diabetes mellitus and exocrine pancreatic dysfunction in perk-/- mice reveals a role for translational control in secretory cell survival. Molecular cell 7: 1153-63 Harding HP, Zhang Y, Ron D. 1999. Protein translation and folding are coupled by an endoplasmicreticulum-resident kinase. Nature 397: 271-4 Hartley CL, Edwards S, Mullan L, Bell PA, Fresquet M, et al. 2013. Armet/Manf and Creld2 are components of a specialized ER stress response provoked by inappropriate formation of disulphide bonds: implications for genetic skeletal diseases. Human molecular genetics 22: 5262-75 Haze K, Yoshida H, Yanagi H, Yura T, Mori K. 1999. Mammalian transcription factor ATF6 is synthesized as a transmembrane protein and activated by proteolysis in response to endoplasmic reticulum stress. Molecular biology of the cell 10: 3787-99 36

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

MA

NU

SC R

IP

T

Hellman M, Arumae U, Yu LY, Lindholm P, Peranen J, et al. 2011. Mesencephalic astrocytederived neurotrophic factor (MANF) has a unique mechanism to rescue apoptotic neurons. The Journal of biological chemistry 286: 2675-80 Henderson MJ, Richie CT, Airavaara M, Wang Y, Harvey BK. 2013. Mesencephalic astrocytederived neurotrophic factor (MANF) secretion and cell surface binding are modulated by KDEL receptors. The Journal of biological chemistry 288: 4209-25 Henderson MJ, Wires ES, Trychta KA, Richie CT, Harvey BK. 2014. SERCaMP: a carboxyterminal protein modification that enables monitoring of ER calcium homeostasis. Molecular biology of the cell 25: 2828-39 Heneka MT, Kummer MP, Latz E. 2014. Innate immune activation in neurodegenerative disease. Nature reviews. Immunology 14: 463-77 Hetz C. 2012. The unfolded protein response: controlling cell fate decisions under ER stress and beyond. Nature reviews. Molecular cell biology 13: 89-102 Hoozemans JJ, van Haastert ES, Eikelenboom P, de Vos RA, Rozemuller JM, Scheper W. 2007. Activation of the unfolded protein response in Parkinson's disease. Biochemical and biophysical research communications 354: 707-11 Hoseki J, Sasakawa H, Yamaguchi Y, Maeda M, Kubota H, et al. 2010. Solution structure and dynamics of mouse ARMET. FEBS letters 584: 1536-42 Inagi R, Ishimoto Y, Nangaku M. 2014. Proteostasis in endoplasmic reticulum--new mechanisms in kidney disease. Nature reviews. Nephrology 10: 369-78 Inoue H, Tanizawa Y, Wasson J, Behn P, Kalidas K, et al. 1998. A gene encoding a transmembrane protein is mutated in patients with diabetes mellitus and optic atrophy (Wolfram syndrome). Nature genetics 20: 143-8 Jankowsky JL, Fadale DJ, Anderson J, Xu GM, Gonzales V, et al. 2004. Mutant presenilins specifically elevate the levels of the 42 residue beta-amyloid peptide in vivo: evidence for augmentation of a 42-specific gamma secretase. Human molecular genetics 13: 159-70 Kaufman RJ, Back SH, Song B, Han J, Hassler J. 2010. The unfolded protein response is required to maintain the integrity of the endoplasmic reticulum, prevent oxidative stress and preserve differentiation in beta-cells. Diabetes, obesity & metabolism 12 Suppl 2: 99-107 Kemppainen S, Lindholm P, Galli E, Lahtinen HM, Koivisto H, et al. 2015. Cerebral dopamine neurotrophic factor improves long-term memory in APP/PS1 transgenic mice modeling Alzheimer's disease as well as in wild-type mice. Behavioural brain research 291: 1-11 Kim Y, Lee H, Manson SR, Lindahl M, Evans B, et al. 2016. Mesencephalic Astrocyte-Derived Neurotrophic Factor as a Urine Biomarker for Endoplasmic Reticulum Stress-Related Kidney Diseases. Journal of the American Society of Nephrology : JASN Kleinridders A, Ferris HA, Cai W, Kahn CR. 2014. Insulin action in brain regulates systemic metabolism and brain function. Diabetes 63: 2232-43 Kordower JH, Bjorklund A. 2013. Trophic factor gene therapy for Parkinson's disease. Movement disorders : official journal of the Movement Disorder Society 28: 96-109 Kotzbauer PT, Lampe PA, Heuckeroth RO, Golden JP, Creedon DJ, et al. 1996. Neurturin, a relative of glial-cell-line-derived neurotrophic factor. Nature 384: 467-70 Latge C, Cabral KM, Almeida MS, Foguel D. 2013. (1)H-, (13)C- and (15)N-NMR assignment of the N-terminal domain of human cerebral dopamine neurotrophic factor (CDNF). Biomolecular NMR assignments 7: 101-3 Latge C, Cabral KM, de Oliveira GA, Raymundo DP, Freitas JA, et al. 2015. The Solution Structure and Dynamics of Full-length Human Cerebral Dopamine Neurotrophic Factor and Its Neuroprotective Role against alpha-Synuclein Oligomers. The Journal of biological chemistry 290: 20527-40

37

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

MA

NU

SC R

IP

T

Lee AH, Iwakoshi NN, Glimcher LH. 2003. XBP-1 regulates a subset of endoplasmic reticulum resident chaperone genes in the unfolded protein response. Molecular and cellular biology 23: 7448-59 Li Y, Iida K, O'Neil J, Zhang P, Li S, et al. 2003. PERK eIF2alpha kinase regulates neonatal growth by controlling the expression of circulating insulin-like growth factor-I derived from the liver. Endocrinology 144: 3505-13 Lin LF, Doherty DH, Lile JD, Bektesh S, Collins F. 1993. GDNF: a glial cell line-derived neurotrophic factor for midbrain dopaminergic neurons. Science 260: 1130-2 Lindahl M, Danilova T, Palm E, Lindholm P, Voikar V, et al. 2014. MANF is indispensable for the proliferation and survival of pancreatic beta cells. Cell Rep 7: 366-75 Lindholm D, Wootz H, Korhonen L. 2006. ER stress and neurodegenerative diseases. Cell death and differentiation 13: 385-92 Lindholm P, Peranen J, Andressoo JO, Kalkkinen N, Kokaia Z, et al. 2008. MANF is widely expressed in mammalian tissues and differently regulated after ischemic and epileptic insults in rodent brain. Molecular and cellular neurosciences 39: 356-71 Lindholm P, Voutilainen MH, Lauren J, Peranen J, Leppanen VM, et al. 2007. Novel neurotrophic factor CDNF protects and rescues midbrain dopamine neurons in vivo. Nature 448: 73-7 Lindstrom R, Lindholm P, Kallijarvi J, Palgi M, Saarma M, Heino TI. 2016. Exploring the Conserved Role of MANF in the Unfolded Protein Response in Drosophila melanogaster. PloS one 11: e0151550 Lindstrom R, Lindholm P, Kallijarvi J, Yu LY, Piepponen TP, et al. 2013. Characterization of the structural and functional determinants of MANF/CDNF in Drosophila in vivo model. PloS one 8: e73928 Liu Y, Nie L, Zhao H, Zhang W, Zhang YQ, et al. 2014. Conserved dopamine neurotrophic factortransduced mesenchymal stem cells promote axon regeneration and functional recovery of injured sciatic nerve. PloS one 9: e110993 Luo GR, Chen Y, Li XP, Liu TX, Le WD. 2008. Nr4a2 is essential for the differentiation of dopaminergic neurons during zebrafish embryogenesis. Molecular and cellular neurosciences 39: 202-10 Matlik K, Yu LY, Eesmaa A, Hellman M, Lindholm P, et al. 2015. Role of two sequence motifs of mesencephalic astrocyte-derived neurotrophic factor in its survival-promoting activity. Cell death & disease 6: e2032 Mercado G, Valdes P, Hetz C. 2013. An ERcentric view of Parkinson's disease. Trends in molecular medicine 19: 165-75 Mizobuchi N, Hoseki J, Kubota H, Toyokuni S, Nozaki J, et al. 2007. ARMET is a soluble ER protein induced by the unfolded protein response via ERSE-II element. Cell structure and function 32: 41-50 Mori K, Ma W, Gething MJ, Sambrook J. 1993. A transmembrane protein with a cdc2+/CDC28related kinase activity is required for signaling from the ER to the nucleus. Cell 74: 743-56 Nadella R, Voutilainen MH, Saarma M, Gonzalez-Barrios JA, Leon-Chavez BA, et al. 2014. Transient transfection of human CDNF gene reduces the 6-hydroxydopamine-induced neuroinflammation in the rat substantia nigra. Journal of neuroinflammation 11: 209 Norisada J, Hirata Y, Amaya F, Kiuchi K, Oh-hashi K. 2016. A Comparative Analysis of the Molecular Features of MANF and CDNF. PloS one 11: e0146923 Oh-Hashi K, Hirata Y, Kiuchi K. 2013. Transcriptional regulation of mouse mesencephalic astrocyte-derived neurotrophic factor in Neuro2a cells. Cellular & molecular biology letters 18: 398-415 Oh-Hashi K, Tanaka K, Koga H, Hirata Y, Kiuchi K. 2012. Intracellular trafficking and secretion of mouse mesencephalic astrocyte-derived neurotrophic factor. Molecular and cellular biochemistry 363: 35-41 38

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

MA

NU

SC R

IP

T

Palgi M, Greco D, Lindstrom R, Auvinen P, Heino TI. 2012. Gene expression analysis of Drosophilaa Manf mutants reveals perturbations in membrane traffic and major metabolic changes. BMC genomics 13: 134 Palgi M, Lindstrom R, Peranen J, Piepponen TP, Saarma M, Heino TI. 2009. Evidence that DmMANF is an invertebrate neurotrophic factor supporting dopaminergic neurons. Proceedings of the National Academy of Sciences of the United States of America 106: 2429-34 Paratcha G, Ledda F. 2008. GDNF and GFRalpha: a versatile molecular complex for developing neurons. Trends in neurosciences 31: 384-91 Parkash V, Lindholm P, Peranen J, Kalkkinen N, Oksanen E, et al. 2009. The structure of the conserved neurotrophic factors MANF and CDNF explains why they are bifunctional. Protein engineering, design & selection : PEDS 22: 233-41 Petrova P, Raibekas A, Pevsner J, Vigo N, Anafi M, et al. 2003. MANF: a new mesencephalic, astrocyte-derived neurotrophic factor with selectivity for dopaminergic neurons. Journal of molecular neuroscience : MN 20: 173-88 Raykhel I, Alanen H, Salo K, Jurvansuu J, Nguyen VD, et al. 2007. A molecular specificity code for the three mammalian KDEL receptors. The Journal of cell biology 179: 1193-204 Ren X, Zhang T, Gong X, Hu G, Ding W, Wang X. 2013. AAV2-mediated striatum delivery of human CDNF prevents the deterioration of midbrain dopamine neurons in a 6hydroxydopamine induced parkinsonian rat model. Experimental neurology 248: 148-56 Sawada M, Sun W, Hayes P, Leskov K, Boothman DA, Matsuyama S. 2003. Ku70 suppresses the apoptotic translocation of Bax to mitochondria. Nature cell biology 5: 320-9 Scheuner D, Kaufman RJ. 2008. The unfolded protein response: a pathway that links insulin demand with beta-cell failure and diabetes. Endocrine reviews 29: 317-33 Senee V, Chelala C, Duchatelet S, Feng D, Blanc H, et al. 2006. Mutations in GLIS3 are responsible for a rare syndrome with neonatal diabetes mellitus and congenital hypothyroidism. Nature genetics 38: 682-7 Shen Y, Sun A, Wang Y, Cha D, Wang H, et al. 2012. Upregulation of mesencephalic astrocytederived neurotrophic factor in glial cells is associated with ischemia-induced glial activation. Journal of neuroinflammation 9: 254 Shridhar R, Shridhar V, Rivard S, Siegfried JM, Pietraszkiewicz H, et al. 1996. Mutations in the arginine-rich protein gene, in lung, breast, and prostate cancers, and in squamous cell carcinoma of the head and neck. Cancer research 56: 5576-8 Simon HH, Bhatt L, Gherbassi D, Sgado P, Alberi L. 2003. Midbrain dopaminergic neurons: determination of their developmental fate by transcription factors. Annals of the New York Academy of Sciences 991: 36-47 Stratoulias V, Heino TI. 2015. Analysis of the conserved neurotrophic factor MANF in the Drosophila adult brain. Gene expression patterns : GEP 18: 8-15 Sun ZP, Gong L, Huang SH, Geng Z, Cheng L, Chen ZY. 2011. Intracellular trafficking and secretion of cerebral dopamine neurotrophic factor in neurosecretory cells. Journal of neurochemistry 117: 121-32 Szegezdi E, Logue SE, Gorman AM, Samali A. 2006. Mediators of endoplasmic reticulum stressinduced apoptosis. EMBO reports 7: 880-5 Tadimalla A, Belmont PJ, Thuerauf DJ, Glassy MS, Martindale JJ, et al. 2008. Mesencephalic astrocyte-derived neurotrophic factor is an ischemia-inducible secreted endoplasmic reticulum stress response protein in the heart. Circulation research 103: 1249-58 van der Harg JM, Nolle A, Zwart R, Boerema AS, van Haastert ES, et al. 2014. The unfolded protein response mediates reversible tau phosphorylation induced by metabolic stress. Cell death & disease 5: e1393

39

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

MA

NU

SC R

IP

T

Wang H, Ke Z, Alimov A, Xu M, Frank JA, et al. 2014. Spatiotemporal expression of MANF in the developing rat brain. PloS one 9: e90433 Wang W, Dai H, Zhang Y, Chandrasekar R, Luo L, et al. 2015. Armet is an effector protein mediating aphid-plant interactions. FASEB journal : official publication of the Federation of American Societies for Experimental Biology 29: 2032-45 Wei J, Sheng X, Feng D, McGrath B, Cavener DR. 2008. PERK is essential for neonatal skeletal development to regulate osteoblast proliferation and differentiation. Journal of cellular physiology 217: 693-707 Voutilainen MH, Back S, Peranen J, Lindholm P, Raasmaja A, et al. 2011. Chronic infusion of CDNF prevents 6-OHDA-induced deficits in a rat model of Parkinson's disease. Experimental neurology 228: 99-108 Voutilainen MH, Back S, Porsti E, Toppinen L, Lindgren L, et al. 2009. Mesencephalic astrocytederived neurotrophic factor is neurorestorative in rat model of Parkinson's disease. The Journal of neuroscience : the official journal of the Society for Neuroscience 29: 9651-9 Yang S, Huang S, Gaertig MA, Li XJ, Li S. 2014. Age-dependent decrease in chaperone activity impairs MANF expression, leading to Purkinje cell degeneration in inducible SCA17 mice. Neuron 81: 349-65 Yavarna T, Al-Dewik N, Al-Mureikhi M, Ali R, Al-Mesaifri F, et al. 2015. High diagnostic yield of clinical exome sequencing in Middle Eastern patients with Mendelian disorders. Human genetics 134: 967-80 Yoshida H, Matsui T, Yamamoto A, Okada T, Mori K. 2001. XBP1 mRNA is induced by ATF6 and spliced by IRE1 in response to ER stress to produce a highly active transcription factor. Cell 107: 881-91 Yoshida H, Okada T, Haze K, Yanagi H, Yura T, et al. 2000. ATF6 activated by proteolysis binds in the presence of NF-Y (CBF) directly to the cis-acting element responsible for the mammalian unfolded protein response. Molecular and cellular biology 20: 6755-67 Yu YQ, Liu LC, Wang FC, Liang Y, Cha DQ, et al. 2010. Induction profile of MANF/ARMET by cerebral ischemia and its implication for neuron protection. Journal of cerebral blood flow and metabolism : official journal of the International Society of Cerebral Blood Flow and Metabolism 30: 79-91 Zhao H, Cheng L, Du X, Hou Y, Liu Y, et al. 2016. Transplantation of Cerebral Dopamine Neurotrophic Factor Transducted BMSCs in Contusion Spinal Cord Injury of Rats: Promotion of Nerve Regeneration by Alleviating Neuroinflammation. Molecular neurobiology 53: 187-99 Zhao H, Cheng L, Liu Y, Zhang W, Maharjan S, et al. 2014. Mechanisms of anti-inflammatory property of conserved dopamine neurotrophic factor: inhibition of JNK signaling in lipopolysaccharide-induced microglia. Journal of molecular neuroscience : MN 52: 186-92 Zhao H, Liu Y, Cheng L, Liu B, Zhang W, et al. 2013. Mesencephalic astrocyte-derived neurotrophic factor inhibits oxygen-glucose deprivation-induced cell damage and inflammation by suppressing endoplasmic reticulum stress in rat primary astrocytes. Journal of molecular neuroscience : MN 51: 671-8 Zhou C, Xiao C, Commissiong JW, Krnjevic K, Ye JH. 2006a. Mesencephalic astrocyte-derived neurotrophic factor enhances nigral gamma-aminobutyric acid release. Neuroreport 17: 2937 Zhou J, Liu CY, Back SH, Clark RL, Peisach D, et al. 2006b. The crystal structure of human IRE1 luminal domain reveals a conserved dimerization interface required for activation of the unfolded protein response. Proceedings of the National Academy of Sciences of the United States of America 103: 14343-8

40

ACCEPTED MANUSCRIPT Figure legends Figure 1. (A) Schematic primary structure of human MANF and CDNF in comparison to that of

T

GDNF. The saposin-like domain of MANF and CDNF is indicated in blue, and the SAP-domain in

IP

orange. Cysteine residues are marked as yellow bars and numbered according to the sequence of

SC R

mature protein. Disulphide bridges are indicated by black connecting lines according to (Parkash et al 2009) and suggested disulphide bridges (Mizobuchi et al 2007, Wang et al 2015) in grey below

NU

and above the MANF scheme, respectively. GDNF contains a pro-region which is enzymatically cleaved. Mature GDNF is indicated in green and disulphide bonds forming a cysteine knot motif are

MA

indicated with black lines. Two GDNF molecules form a homodimer via a disulphide bridge and a cysteine residue involved is indicated by an arrow. H.s.; Homo sapiens. (B) NMR solution

D

structure of human MANF consisting of an N-terminal saposin-like domain (residues 1-95) and a C-

TE

terminal SAP-domain (residues 96-158) according to (Hellman et al 2011). The linker region (residues 96-103), CXXC motif and ER retention signal are indicated. N, amino-terminus; C,

AC

CE P

carboxy-terminus. The image is modified from (Lindstrom et al 2013).

Figure 2. UPR and hypothetic involvement of MANF in it. (A) In the absence of ER stress, binding of GRP78 to the ER luminal domain of UPR sensors ATF6, IRE1 and PERK keeps them inactive. MANF is retained in the ER by KDELR via the C-terminal RTDL sequence, and by calcium dependent interaction with GRP78 (Glembotski et al 2012, Henderson et al 2013). KDELR may retrieve MANF from the Golgi back to the ER. Hypothetically (question marks) MANF may bind GRP78, which is interacting with the UPR sensors, helping to keep them inactive. (B) Activation of UPR in ER stress. Binding of GRP78 to accumulating unfolded proteins leads to its dissociation from ATF6, IRE1 and PERK, resulting in their activation. ATF6 translocates to the Golgi, where it is cleaved by site 1 and site 2 proteases. Active ATF6 functions as a transcription factor and induces genes involved in protein folding, lipid biosynthesis and the expression of GRP78, MANF, 41

ACCEPTED MANUSCRIPT proapoptotic transcription factor C/EBP homologous protein CHOP, and Xbp1 mRNA (Yoshida et al 2001, Yoshida et al 2000). It also directly controls genes encoding for components of ERAD.

T

IRE1 is activated by dimerization and trans-autophosphorylation. Endoribonuclease activity of the

IP

cytoplasmic domain of active IRE1 removes an intron from XBP1, generating spliced XBP1

SC R

(XBP1s) transcription factor. XBP1s upregulates genes for chaperons, ERAD machinery, phospholipid synthesis and Manf. IRE1 can degrade ER associated mRNAs such as that of insulin, thus reducing translational workload in pancreatic beta cells (Han et al. 2009; Pirot et al. 2007).

NU

IRE1 phosphorylates Jun N terminal kinase (JNK) and activates nuclear translocation of NF-κB,

MA

leading to apoptosis and regulation of autophagy genes in beta cells. Dimerized and activated PERK inhibits general protein translation by phosphorylating translation initiation factor eIF2α subunit, while favoring the translation of ATF4 transcription factor. ATF4 induces pro-survival genes of

TE

D

redox-balance, amino acid metabolism and protein folding (Ameri & Harris 2008). Marked translational inhibition may lead to decreased levels of inhibitory nuclear kappa B (IkB), leading to

CE P

nuclear translocation of NF-κB and increased transcription of inflammatory response genes, and apoptotic cell death. ATF4 induces the expression of genes involved in restoring ER homeostasis,

AC

CHOP, and protein phosphatase 1 regulatory subunit 15A (PPP1R15A) which in complex with protein phosphatase 1 (PP1) can relieve translation arrest by dephosphorylating eIF2α. Phosphorylation of transcription factor NRF2 by PERK results in upregulation of antioxidant response genes (Cnop et al 2012). In ER stress, Ca2+ is depleted from the ER which may lead to dissociation of MANF from GRP78 and its secretion. MANF retention by KDELR in the ER may remain intact (Glembotski et al 2012). Alternatively, MANF interaction with KDELR may regulate MANF secretion to the extracellular space during ER stress (Henderson et al 2013). UP, unfolded protein.

42

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

MA

NU

SC R

IP

T

Figure 1

43

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

MA

NU

SC R

IP

T

Figure 2

44