|Home | About | Journals | Submit | Contact Us | Français|
Loss-of-function progranulin gene (GRN) mutations have been identified as the major cause of frontotemporal lobar degeneration with transactive response (TAR) DNA-binding protein 43 (TDP-43) pathology (frontotemporal lobar degeneration [FTLD]-TDP); however, little is known about the association between progranulin (PGRN) deficiency and neuronal loss in individuals with FTLD-TDP. Previously we reported enhanced proliferative activity associated with the activation of WNT5A/CDK6/pRb signalling in PGRN-deficient cells. The objective of this work was to elucidate the association between PGRN deficiency, WNT5A signalling and cell proliferation in immortalized lymphoblasts from carriers of the c.709–1G > A GRN mutation (asymptomatic and FTLD-TDP).
We assessed cell proliferation in carriers of the c.709–1G > A GRN gene mutation and controls without GRN mutation and without sign of neurologic degeneration by cell counting or using an MTT assay. We used a luciferase assay to measure the nuclear factor-κ (NF-κ) activity. We evaluated messenger RNA levels using quantitative real-time polymerase chain reaction and protein levels by immunoblotting. Co-immunoprecipitation was used to analyze the interaction between PGRN and its receptors.
We enrolled 19 carriers of the GRN gene mutation and 10 controls in this study. The PGRN-deficient cells showed increased expression of WNT5A due to NF-κB signalling overactivation. We observed a competition between PGRN and tumour necrosis factor-α (TNF-α) for binding both TNF receptors (TNFR) I and II. Blocking NF-κB signalling using wedelolactone or specific antibodies against TNFRs inhibited WNT5A overexpression and proliferation of PGRN-deficient cells. Conversely, the activation of NF-κB signalling by TNF-α increased WNT5A-dependent proliferation of control cells.
All cell lines were derived from individuals harboring the same splicing GRN mutation. Nevertheless, most of the known GRN mutations lead to haploinsufficiency of the protein.
Our results revealed an important role of NF-κB signalling in PGRN-associated FTLD-TDP and confirm that PGRN can bind to TNF-α receptors regulating the expression of WNT5A, suggesting novel targets for treatment of FTLD-TDP linked to GRN mutations.
Frontotemporal lobar degeneration (FTLD) is a clinically, pathologically and genetically heterogeneous disorder resulting from the progressive deterioration of the frontal and temporal lobes of the brain.1,2 It is the second most common form of presenile dementia (after Alzheimer disease), with a prevalence estimated to be 10–30 per 100 000 individuals between the ages of 45 and 65 years.3,4 Depending on the affected regions, patients with FTLD can have dementia, behavioural abnormalities, language impairment or personality changes, among other conditions.5,6 The molecular pathology is heterogeneous and is based on the type of neuronal lesions and abnormal protein aggregates present in neuroectodermic cells. In most cases, the inclusions contain either the microtubule-associated protein tau (FTLD-tau; approximately 40%) or the transactive response DNA-binding protein TDP-43 (FTLD-TDP; approximately 50%), although in a small number of cases pathological inclusions containing the fused in sarcoma (FUS) protein (FTLD-FUS) or ubiquitinated proteins (FTLD-UPS) have been described.7–9
A positive family history is found in 40%–50% of patients with FTLD,10 and 10% of them present an autosomal dominant inheritance.11 Familial forms of FTLD are caused by mutations in 3 common genes: microtubule-associated protein tau (MAPT), granulin (GRN) and chromosome 9 open reading frame 72 (C9ORF72).12–18 Mutations in other genes, such as transactive response (TAR) DNA-binding protein-43 (TARDBP), valosin-containing protein (VCP), chromatin-modifying 2B protein (CHMP2B), sequestome1/p62 (SQSTM1) and ubiquilin-2 (UBQLN2) are rare causes of disease.14,19–23
Mutations in GRN have been identified as a major cause of autosomal-dominant FTLD-TDP, leading to TDP-43 inclusions by a haploinsufficiency mechanism,12,13 but little is known about how progranulin (PGRN) deficiency induces neuronal loss in individuals with FTLD-TDP.24
Previous work from our laboratory showed an inverse association between WNT5A and PGRN abundance. In addition, we found an overactivation of cell proliferation associated with enhanced noncanonical WNT5A/CDK6/pRb signalling in both lymphoblasts from patients with FTLD-TDP carrying a loss-of-function GRN mutation (c.709–1G > A) or GRN knockdown SH-SY5Y neuroblastoma cells.25 It has been suggested that an aberrant activation of the WNT5A cascade could have pathogenic significance in PGRN deficiency–linked FTLD, as it is believed that unscheduled cell cycle entry underlies neuronal loss in patients with neurodegenerative disorders.26 The present work was undertaken to elucidate possible mechanisms involved in the PGRN deficiency–induced increases in WNT5A levels and signalling.
Little is known about how WNT5A is regulated; however, there is evidence that nuclear factor (NF)-κB can influence WNT5A expression.27 In this sense, 2 promoter regions have been identified in the human WNT5A gene (promoter A and B), which contains putative NF-κB binding sites.28 NF-κB mediates the expression of many genes involved in inflammation and it is known to be upregulated in some neurodegenerative disorders such as Alzheimer disease,29,30 Parkinson disease31 or amyotrophic lateral sclerosis (ALS).32
Tumour necrosis factor (TNF)-α is a well-known NF-κB inducer.33 There are 2 membrane receptors of TNF-α: TNFRI and TNFRII. TNFRI is widely expressed in most cell lines and primary tissues, although TNFRII is preferentially expressed on cells of the hematopoietic lineage.34 The binding of TNF-α to both TNFRs results in the recruitment of adaptor and interacting proteins that activate the IκB kinase (IKK) complex leading to phosphorylation and degradation of IkBα and activation of NF-κB.35,36 Work from different laboratories supports the hypothesis that PGRN directly binds to TNFRs, blocking their interaction with TNF-α.37–39 Therefore, we hypothesized that PGRN deficiency would eventually facilitate TNF-α/NF-κB signalling, which in turn may control WNT5A expression and signalling.
All components for cell culture were obtained from Invitrogen. We obtained recombinant human progranulin from Enzo and recombinant human TNF-α from R&D Systems. We purchased polyvinylidene fluoride (PVDF) membranes from Merk-Millipore, and the Luciferase Reporter Assay System was obtained from Promega. Other reagents were of molecular biology grade. We obtained antibodies against human NF-κB p50 (H-119) (sc-7178), NF-κB p65 (H-286) (sc-7151), β-actin (sc-81178) and α-tubulin (sc-23948) from Santa Cruz Biotechnologies; antibodies against phospho-IκBα (Ser32) (14D4) (#2859), IκBα (#9242) and NF-κB phospho-p65 (Ser536) (#3031) from Cell Signalling; and antibodies against TNFRI and II from Hycult Biotech. Anti-WNT5A and anti-PGRN (EPR3781) were obtained from Abcam, and mouse anti-human Lamin B1 was from Calbiochem. The kinase inhibitor PD332991 was kindly provided by Pfizer, and we obtained the NF-κB inhibitor wedelolactone (WDL) from Merk-Millipore.
We recruited carriers of the c.709–1G > A GRN gene mutation (presymptomatic and symptomatic) and control individuals without GRN mutation and without sign of neurologic degeneration from 7 different families of Basque descent with a homogeneous genetic background to participate in this study. In all patients, FTD was diagnosed in the Hospital Universitario Donostia using consensus criteria, as published elsewhere.40
All study protocols were approved by the Hospital Universitario Donostia and the Spanish Council of Higher Research Institutional Review Board and are in accordance with National and European Union Guidelines. In all cases, peripheral blood samples to determine the presence of the c.709–1G > A GRN mutation and to establish the lymphoblastoid cell lines were obtained after the patients or their relatives provided written informed consent.
Establishment of lymphoblastic cell lines was performed in our laboratory as previously described41 by infecting peripheral blood lymphocytes with the Epstein–Barr virus (EBV). Cells were grown in suspension in T flasks in an upright position, in approximately 8 mL of RPMI-1640 medium that contained 2 mM of L-glutamine, 100 mg/mL of streptomycin/penicillin and 10% (v/v) fetal bovine serum (FBS). The flasks were maintained in a humidified 5% CO2 incubator at 37°C. Fluid was routinely changed every 3 days by removing the medium above the settled cells and replacing it with an equal volume of fresh medium.
Stable GRN knockdown neuroblastoma SH-SY5Y cells (Clone # 207) was provided by Drs. Joselin and Wu from the Center for Genetic Medicine (Northwestern University, Chicago, Il USA). GRN knockdown was achieved using pSUPERIOR RNAi construct as previously described.42 We designed a sequence of 19 nucleotides targeted against nucleotides 207–226 (#207) of the human GRN messenger RNA (mRNA). The 64 nt short hairpin RNA sense and antisense primer sequences were 5′-GATCCCCGGCCACTCCTGCATCTTTATTCAAGAGATAAAGATGCAGGAGTGGCCTTTTTGGAAA-3′ and 5′-AGCTTTTCCAAAAAGGCCACTCCTGCATCTTTATCTC TTGAATAAAGATGCAGGAGTGGCCGGG-3′.
The sense and antisense primer pairs were annealed and ligated into the pSUPERIOR vector (OligoEngine) according to the manufacturer’s instructions. The vector control was also stably introduced into SH-SY5Y cells to generate the control cell line.
Cell proliferation was determined by total cell counting using a TC10TM Automated Cell Counter from Bio-Rad Laboratories. The EBV-immortalized lymphocytes from controls and GRN mutation carriers were seeded at an initial cell concentration of 1 × 106 cells/mL–1. Cells were enumerated every day thereafter. In some experiments, cell proliferation was assessed using the MTT (3-(4,5-Dimethylthiazol-2-yl)-2,5-Diphenyltetrazolium Bromide) assay.43
To prepare whole cell extracts the lymphoblasts were seeded at an initial density of 1 × 106 cells × mL−1, and 24 hours later cells were harvested, washed in phosphate-buffered saline (PBS) and then lysed in ice-cold lysis buffer (50 mM Tris pH 7.4, 150 mM of NaCl, 50 mM of NaF, 1% Nonidet P-40), containing 1 mM of sodium orthovanadate, 1 mM of phenylmethylsulfonylfluride (PMSF), 1 mM of sodium pyrophosphate and protease inhibitor complete mini mixture (Roche). The protein content of the extracts was determined using a de Pierce BCA Protein Assay kit (Thermo Scientific).
We fractionated 50–100 μg of protein from cell extracts on an SDS polyacrylamide gel and transferred it to PVDF membrane. The membranes were then blocked with 5% BSA (Sigma) and incubated overnight at 4ºC, with primary antibodies in the following concentrations: β-actin (1:1000), α-tubulin (1:1000), antihuman Lamin B (1:500), anti-WNT5A (1:500), anti-PGRN (1:500), NF-κB p50 (1:1000), NF-κB p65 (1:200), NF-κB phosphop65 (Ser536) (1:1000), phospho-IkBα (1:500), IBa (1:1000), anti-TNF-RI (1:500) and TNFR-II (1:500). The specificity of the antibodies used in this work was checked by omitting the primary antibody in the incubation medium. Signals from the primary antibodies were amplified using species-specific antisera conjugated with horseradish peroxidase (Bio-Rad) and detected with a chemiluminiscent substrate detection system. Protein band densities were quantified using Image J software (National Institutes of Health) after scanning the images with a GS-800 densitometer from Bio-Rad.
After cell extracts preparation, 1 mg of protein was subjected to immunoprecipitation overnight at 4ºC with anti-PGRN antibody or an irrelevant Ig. Samples were incubated with protein A Sepharose (GE Healthcare Bio-Sciences), and the resultant immunoprecipitates were washed 3 times in ice-cold lysis buffer. The samples were then treated with protein sample buffer and boiled before immunoblotting. Western blotting was performed using the TNFRI and TNFRII antibodies. Membranes were incubated with anti-PGRN antibody as positive control of the experiment.
Lymphoblasts from controls and carriers of the GRN mutation (2 × 106 cells) were transiently transfected with 2 μg of NF-κB-luciferase reporter plasmid (3EnhConA) using the Amaxa nucleofector II device (Lonza) and the P-16 program. To evaluate transfection efficiency and cell viability, cells were co-transfected with the GFP-expressing construct pmaxGFP (Amaxa Biosystems). Twenty-four hours after nucleofection, 3 × 105 cells were staining with propidium iodide and analyzed by flow cytometry. The transfection efficiency was approximately 17%. Luciferase activity measurement took place 24 hours after nucleofection using the Luciferase Assay System (Promega) according to the manufacturer’s instructions.
Total RNA was extracted from cell cultures using Trizol reagent (Invitrogen). We quantified RNA yields spectrophotometrically, and RNA quality was checked using the A260:A280 ratio and on a 1.2% agarose gel to determine the integrity of 18S and 28S rRNA. RNA was then treated with DNase I Amplification grade (Invitrogen). One microgram was reverse transcribed with the Superscript III Reverse Transcriptase kit (Invitrogen). Quantitative real-time polymerase chain reaction (PCR) was performed in triplicate using Taq-Man Universal PCR MasterMix No Amperase UNG (Applied Biosystems) reagent according to the manufacturer’s protocol. Primers were designed using the Universal Probe Library Set, Human (Roche Applied Science) and used at a final concentration of 20 mM. The following sequences of forward and reverse primers were used: 5′-ATTGTACTGCAGGTGTACCTTAAAAC-3′ and 5′-ACCCCCTTATAAATGCAACTGTTC-3′ for WNT5A and 5′-CCAACCGCGAGAAGATGA-3′ and 5′-CCAGAGGCGTACAGGGATAG-3′ for β-actin. Quantitative real-time PCR was performed in the Bio-Rad iQ5 system using a thermal profile of an initial 5-min melting step at 95 ºC followed by 40 cycles at 95 ºC for 10 s and 60 ºC for 60 s. Relative mRNA levels of the genes of interest were normalized to β-actin expression using the simplified comparative threshold cycle delta-delta Ct method (2−[DCT WNT5A-DCT Actin]).
Statistical analyses were performed using Graph Pad Prism software version 6. Data are presented as means ± standard errors of the mean (SEM). Normality was checked using the Shapiro–Wilk test. Parametric tests were therefore used in the statistical analysis. Based on the expertise achieved in previous works25,44,45 we expected that with the sample size that we used and the significance level that we established the variability within groups would be low enough and the power to detect differences between groups would be high enough to ensure a statistical power above 0.9. Statistical significance was estimated using the Student t test or, when appropriate, using 1-way and 2-way analyses of variance (ANOVA) followed by the Bonferroni test for multiple comparisons. We considered results to be significant at p < 0.05.
A total of 29 individuals from 7 families were enrolled in this study: 19 carriers of the GRN gene mutation (12 presymptomatic and 7 symptomatic) and 10 controls. Table 1 summarizes the demographic characteristics and the plasma levels of PGRN of all participants enrolled in this study.
Data in Fig. 1 confirm and extend previous results from our laboratory by showing that lower cellular and secreted levels of PGRN found in lymphoblasts harboring the c.709–1G > A GRN mutation are associated with a significant increase in both cellular content and secreted WNT5A protein (Fig. 1A and B). In consonance, PGRN-deficient lymphoblasts, either from asymptomatic individuals or patients with FTLD-TDP, showed higher WNT5A mRNA expression levels (Fig. 1C). The increased WNT5A mRNA levels were blunted by the addition of 100 ng/mL of rhPGRN (Fig. 1C), resulting in the normalization of the enhanced proliferative activity of lymphoblasts harboring the GRN mutation (Fig. 1D). This dose of rhPGRN falls within the plasma concentration of PGRN found in control individuals25,46 (Table 1).
Because it has been reported that WNT5A expression could be regulated by NF-κB,27 we were interested in assessing whether PGRN deficiency influences NF-κB activity. Fig. 2A showed a clear increase of NF-κB activation in PGRN-deficient lymphoblasts, as examined using a NF-κB luciferase reporter assay. As NF-κB activation occurs via phosphorylation of IkBα followed by proteasome-mediated degradation of this protein, resulting in the release and nuclear translocation of active p50 and p65 NF-κB subunits,47,48 we determined that the intracellular levels of total and phosphorylated IκBα in controls and carriers of GRN-mutated lymphoblasts. We also performed a fractionation of cell extracts to look for nuclear p50 and p65 NF-κB subunit enrichment in PGRN-deficient cells. Fig. 2B shows increased levels of phosphorylated IkBα protein along with a diminution of total IκBα protein in cellular extracts from PGRN-deficient lymphoblasts, thereby facilitating the translocation of NF-κB subunits to the nucleus. As shown in Fig. 2C and 2D, the levels of p50 and p65 NF-κB subunits are increased in the nucleus of cells carrying the GRN mutation. In addition, we observed a significant decrease of these proteins in the cytosolic compartment of PGRN-deficient cells. Taken together these results indicate that PGRN deficiency is associated with activation of NF-κB signalling.
We evaluated the impact of experimentally modulating NF-κB activity using WDL, a selective inhibitor of IκB kinase,49 on the proliferative activity and cellular content of WNT5A in controls and PGRN-deficient lymphoblasts.
Treatment with escalating concentrations of WDL (0–5 μM) suppressed the cell proliferation, which we assessed using the MTT reduction assay, in a dose-dependent manner (Fig. 3A). Maximal effects were found at the concentration of 2.5 μM. The treatment with this concentration of WDL for 72 hours was effective in decreasing the total number of PGRN-deficient lymphoblasts with no effect in control cells (Fig. 3B). As shown in Fig. 3C, WDL was able to decrease the levels of both WNT5A mRNA and WNT5A protein in PGRN-deficient lymphoblasts without affecting control cells. The effectiveness of WDL blocking NF-κB activation in PGRN-deficient lymphoblasts is shown in Fig. 3D. The addition of WDL (2.5 μM) was sufficient to inhibit the enhanced phosphorylation of IκBα protein and thus to increase the levels of total IκBα in the lymphoblasts of the GRN-mutation carriers. Taken together these results demonstrated that the PGRN haploinsufficiency–induced increase in levels of WNT5A and the enhanced proliferation of lymphoblasts harboring the GRN mutation depends on the activation of NF-κB.
It has been reported that PGRN interacts with the TNF-α receptors, TNFRI and II,37 raising the possibility that disruption of PGRN–TNFRI/II interactions may underlie FTLD pathogenesis. On these grounds, we considered the possibility that PGRN haploinsufficiency might induce overactivation of TNF-α/NF-κB signalling, which results in increased levels of WNT5A and cell proliferation. First, we checked if PGRN effectively binds to both TNFRs by immunoprecipitating PGRN with an anti-PGRN antibody, and then we probed for TNFRs with antibodies against TNFRI and II (Fig. 4A, left panel). A reverse experiment immunoprecipitating with TNFRI and II antibodies is shown in the right panel of Fig. 4A. We found that PGRN binds to both TNFRI and II. Then, we treated lymphoblasts from controls and carriers of the GRN mutation with anti-TNFR antibodies or with soluble TNF-α and determined the proliferative activity, NF-κB activation and levels of WNT5A (Fig. 4 and Fig. 5).
Treatment of controls and PGRN-deficient lymphoblasts with anti-TNFRI or anti-TNFRII antibodies normalized the proliferative activity of cells harboring the GRN mutation without significantly affecting the proliferation of control cells (Fig. 4B). Blocking just 1 of the TNFRs was enough to blunt the enhanced proliferative response of PGRN-deficient lymphoblasts, as we didn’t observe additive effects in cell proliferation after treatment with both anti-TNFR antibodies (Fig. 4B). By contrast, as shown in Fig. 4C, the addition of soluble TNF-α increased proliferation of control cells, but not that of PGRN-deficient lymphoblasts. This effect of TNF-α on control cells was prevented in the presence of recombinant human PGRN (rhPGRN). In addition, TNF-α partially blunted the effect of exogenous rhPGRN, decreasing the total cell number in PGRN-deficient lymphoblasts (Fig. 4C).
We then evaluated the effect of anti-TNFRs and exogenous TNF-α treatments on NF-κB activity and WNT5A levels. As expected, both anti-TNFRs decreased NF-κB overactivation in asymptomatic patients and patients with FTLD, which was indicated by the inhibition of the phosphorylation of IκBα (Fig. 5A). In agreement with the idea that NF-κB is involved in the increased levels of WNT5A in PGRN-deficient cells, Fig. 5B shows that the blockade of TNFRs decreased the cellular content of WNT5A in lymphoblast-carriers of the GRN mutation to values close to those of control cells (Fig. 5B). Conversely, the addition of TNF-α–activated NF-κB signalling (Fig. 5C) increased WNT5A protein levels in control cells to values close to those observed in lymphoblasts harboring the GRN mutation (Fig. 5D).
Previous work from our laboratory demonstrated that the PGRN haploinsufficiency–induced stimulation of proliferation was associated with the activation of CDK6/pRb cascade.45 Since it has been reported that CDK6 may affect the transcriptional activity of NF-κB by its ability to phosphorylate the p65 subunit at serine 536,50,51 we considered it interesting to evaluate whether this kinase could influence total levels of WNT5A. To address this issue we treated the cells with a known CDK6 activity inhibitor (PD332991). As expected, carriers of the GRN mutation presented higher levels of pSer536p65 NF-κB subunit than controls. The inhibition of CDK6 activity by PD332991 selectively decreased the overphosphorylation of p65 NF-κB subunit in PGRN-deficient cells without affecting the pSer536p65 levels of control cells (Fig. 6A). Under these experimental conditions, blocking CDK6-associated kinase activity leads to the decrease of cell proliferation (Fig. 6B) of PGRN-deficient lymphoblasts to values close to those of control cells. Moreover, the addition of PD332991 blocked the enhanced expression of WNT5A observed in lymphoblasts from both presymptomatic and symptomatic carriers of the c.709–1G > A GRN mutation.
Previous work form our laboratory has shown that PGRN-deficient neuroblastoma SH-SY5Y cells (GRN KD) presented enhanced WNT5A protein levels and signalling.25 Here we addressed the issue as to whether these changes were accompanied by enhanced NF-κB activity. First, in agreement with our previous observations, we detected higher WNT5A mRNA expression levels in GRN KD cells than in control cells (Fig. 7A). Second, as shown in Fig. 7B, GRN KD cells displayed higher levels of pIkBα (left panel) together with a decrease in total IkBα content (right panel). Moreover a significant increase in the nuclear content of p50 and p65 NF-κB subunits was found in these PGRN-deficient neuronal cells (Fig. 7C and D). Taken together these results indicate that PGRN deficiency stimulates NF-κB activity in this neuronal model of FTLD-TDP, as it does in peripheral cells from patients with FTLD-TDP.
Mounting evidence has suggested that some biochemical dysfunctions affecting the brains of patients with neurodegenerative diseases, including FTD, may also be traced outside the central nervous system (CNS).46,52–54 For this reason, peripheral cells, such as fibroblasts or blood lymphocytes, have been used extensively in the search for useful biomarkers that may correlate with expression and/or progression of disease.55–58 In this regard, it has been shown that peripheral cells from patients with neurodegenerative disorders display cell cycle disturbances,45,59–62 reflecting the unscheduled cell cycle re-entry of postmitotic neurons and leading to apoptotic cell death.63–65 The present work was undertaken to elucidate possible mechanisms involved in the PGRN deficiency–induced WNT5A expression and signalling in PGRN-deficient lymphoblasts from patients with FTLD-TDP that were considered responsible for the increased activation of the CDK6/pRb-mediated cell proliferation.25 The main finding of our study is that PGRN deficiency facilitates TNF-α/NF-κB signalling to regulate WNT5A expression.
We investigated whether WNT5A expression and signalling could be regulated by NF-κB activity based on 2 pieces of information: first, NF-κB putative binding sites have been identified in the 2 alternative promoters A and B of the human WNT5A gene,28,66 and second, we considered that PGRN deficiency could favour TNF-α-induced NF-κB activation as TNFRs have been proposed to serve as the binding receptor for PGRN.37
Several lines of evidence support the idea that WNT5A expression is regulated through NF-κB signalling in carriers of the c.709–1G > A GRN mutation. First, PGRN-deficient lymphoblasts showed higher NF-κB activity than control cells. Second, inhibition of NF-κB activation by WDL resulted in the decrease of WNT5A expression together with the normalization of the proliferative activity of mutant lymphoblasts. Third, the blockade of the NF-κB activation by using antibodies to both TNFRs (I and II) was also effective in decreasing WNT5A levels and proliferation of PGRN-deficient cells. Finally, the addition of soluble TNF-α induced the activation of NF-κB and increased WNT5A expression and proliferation of control cells. Although all the cell lines used were derived from individuals harboring the same splicing GRN mutation, the demonstration that rhPGRN rescued the phenotype of diseased lymphoblasts as well as the finding that NF-κB is also overactivated in GRN KD SH-SY5Y neuroblastoma cells allows the conclusion that PGRN deficiency induces the expression of WNT5A through the NF-κB pathway.
It has previously been reported that PGRN binds to TNFRs with high affinity.37 Other authors did not find evidence for a direct interaction between PGRN and TNFRs;67 however, their inability to demonstrate the direct binding of PGRN to the TNFRs could have been due to improper folding of the rhPGRN used or to inappropriate chip selection for the surface plasmon resonance. Since then, several groups independently confirmed the PGRN–TNFR interaction and the inhibitory effect of this binding on TNF-α-induced effects.39,68,69 Our results suggest the existence of a physical interaction between PGRN and both TNFRs (I and II), as we found that PGRN co-immunoprecipitates with either TNFRI or II. Moreover a functional interaction between PGRN and TNFRs is likely, as we found opposite effects after treatment with anti-TNFR antibodies or exogenous TNF-α on NF-κB signalling, WNT5A expression and proliferative activity of human lymphoblasts. Together these results suggest the cross-talk of TNF-α and PGRN signalling pathways to regulate the NF-κB-mediated expression of WNT5A, which in turn determines the proliferative activity of human lymphoblasts. It has been reported that Grn−/− mice developed a more severe collagen-induced inflammatory arthritis than their control littermates as the result of the competition between TNF-α and PGRN for signalling through TNFR.37 Moreover, previous work showed that PGRN is a therapeutic target for rheumatoid arthritis.70 On these grounds it would be interesting to evaluate if individuals harboring loss-of-function GRN mutations are more susceptible to inflammatory pathologies. A deficiency in PGRN could eventually disinhibit TNF-α-mediated signalling in neurons and/or microglia, leading to a persistent proinflammatory state over the lifetime that ultimately could contribute to neurodegeneration.
Generally, NF-κB activation depends on the phosphorylation and subsequent degradation of IkBα, an inhibitory NF-κB protein that forms a complex with p50 and p65 NF-κB subunits. Once released, NF-κB heterodimer is translocated into the nucleus, where it activates the transcription of its target genes.71 However, it has been shown that NF-κB activity can also be regulated by phosphorylation of the p50 and p65 NF-κB subunits.72,73 Of particular interest for our work is a report indicating that the p65 subunit of NF-κB is phosphorylated at Ser536 by CDK6.50 As already mentioned, PGRN-deficient lymphoblasts display increased levels and kinase activity of CDK6 protein.44,45 On these grounds we considered the possibility that the enhanced CDK6 levels could contribute to increased NF-κB activity in these cell lines. Our results show that the treatment of PGRN-deficient cells with PD332991, a specific CDK6 kinase activity inhibitor, inhibited p65 NF-κB subunit phosphorylation, decreased WNT5A levels and blunted the enhanced proliferative response of the PGRN-deficient lymphoblasts.
Taken together, these findings suggest the existence of a positive feedback loop where the overactivation of WNT5A/CDK6 signalling induced by PGRN deficiency could contribute to maintaining higher levels of WNT5A protein by favouring the phosphorylation of the p65 subunit and the nuclear transcriptional activity of NF-κB. On the other hand, this connection between NF-κB and CDK6 suggests a role of this kinase in inflammation and indicates that CDK6 could contribute to increase the proliferative activity by means other than pRb activation.
A limitation of the present work is that all cell lines used were derived from individuals harboring the same splicing GRN mutation. Nevertheless, since most of the known GRN mutations lead to haploinsufficiency of the protein, one can assume that PGRN deficiency mediates the stimulation of WNT5A signalling.
We provide evidence that PGRN deficiency favours TNF-α/NF-κB signalling, which then results in the increase of WNT5A expression and signalling. Moreover, the WNT5A-induced stimulation of CDK6 activity, in addition to being responsible for the enhanced proliferation of PGRN-deficient lymphoblasts, amplifies the NF-κB-mediated activation of WNT5A transcription. The proposed scenario is represented schematically in Fig. 8.
In agreement with previous work, our results show no differences in the proliferative activity, NF-κB signalling and WNT5A levels among lymphoblasts derived from carriers of the c.709–1G > A GRN mutation who are asymptomatic or who have a clinical diagnosis of FTD.25,45 Because most of the asymptomatic carriers are younger than the patients with FTD, it is suggested that these features are probably early etiologically relevant events during the development of this disorder. These presymptomatic GRN mutation carriers showed no symptoms of disease, but previous studies have described poorer neuropsychological performance and distinctive age-related cortical thinning in asymptomatic individuals compared with controls74,75 that may reflect a presymptomatic phase of the disease.
Changes detected in peripheral cells from patients with FTLD-TDP might not fully reflect those in the FTLD brain; however, mounting evidence suggest that, in addition to neuronal damage, there are also peripheral aspects of the disease.44,46,54,76–79 Altered activation of both TNF-α and WNT5A signalling has been found in the brains of patients with several types of neurodegenerative disorders.80–82 Assuming that the TNF-α and WNT5A-mediated cell cycle dysfunction reported here could be peripheral sign of FTLD associated with PGRN deficiency, we think that targeting TNF-α and/or WNT5A signalling could have potential implications for designing novel therapeutic strategies.
The authors thank Drs. F. Aillet and M.S. Rodríguez for helpful criticism and for their generous gift of the NF-κB luciferase reporter plasmid.
Competing interests: None declared.
Contributors: C. Alquézar and Á. Martín-Requero designed the study. C. Alquézar and A. de la Encarnación acquired the data, which all authors analyzed. All authors wrote and reviewed the article and approved the final version for publication.
Funding: This work has been supported by grants from Ministerio de Economía y Competitividad (SAF2011-28603) and Fundación Ramón Areces to Á. Martín-Requero. A. López de Munain received research support from the Ilundain Foundation.