RESEARCH ARTICLE

MMP-14 overexpression correlates with the neurodegenerativeprocess in familial amyloidotic polyneuropathyDiana Martins1, Joao Moreira1,2, Nadia Pereira Gonçalves1,2,* and Maria Joao Saraiva1,2,‡

ABSTRACTLevels of matrix metalloproteases (MMPs) can be differentiallyregulated in response to injury or neurological diseases. Forinstance, it is known that selective and short-term inhibition ofMMP-14, a membrane-type 1 MMP, accelerates axon regeneration.Because axon growth and regeneration is impaired in familialamyloidotic polyneuropathy (FAP), a neurodegenerative disordercharacterized by misfolding and deposition of mutant transthyretin(TTR) in the peripheral nervous system (PNS), we presentlyinvestigated the expression levels and the potential role for MMP-14 in this condition. By using cell culture studies, a mouse model ofdisease and human clinical samples, we observed that MMP-14: (i) isoverexpressed in FAP nerves, correlating with TTR deposition; (ii) isupregulated in sciatic nerves from a preclinical transgenic mousemodel, increasing with TTR deposition; (iii) levels in the PNS andplasma are rescued upon treatment of mice with anakinra or TTRsiRNA, drugs acting over the IL-1 signaling pathway or TTR liversynthesis, respectively; (iv) increases in Schwann cells uponincubation with amyloid-like aggregates; and, finally, (v) isincreased in plasma of FAP patients, correlating with diseaseprogression. These results highlight the relevance of MMP-14 in thepathophysiology of FAP, suggesting not only a potential role for thismolecule as a novel biomarker for therapy follow up, but also as a newpotential therapeutic target.

KEY WORDS: MMP-14, Familial amyloidotic polyneuropathy,Neurodegeneration, Neuroinflammation, Peripheral nervous system

INTRODUCTIONTransthyretin (TTR)-related amyloidosis is the most common formof hereditary autosomic dominant systemic amyloidoses; TTR pointmutations lead to the extracellular deposition of amyloidogenicspecies in different tissues (Benson and Kincaid, 2007). Avariety ofTTR mutations are known; however, the most common mutationrelated to peripheral neuropathy is TTRV30M, which represents thereplacement of a methione for a valine at codon 30 (Saraiva et al.,1984). Familial amyloidotic polyneuropathy (FAP) is associated

with deposition of mutant TTR aggregates and amyloid fibrils,particularly in the peripheral nervous system (PNS), accumulatingin the basal lamina of Schwann cells, blood vessels and collagenfibrils, ultimately leading to axonal fiber degeneration and cell death(Coimbra and Andrade, 1971; Planté-Bordeneuve and Said, 2011).Recently, TTR expression was found in Schwann cells of the sciaticnerve (Murakami et al., 2015), but this protein is mainly synthesizedin the liver and choroid plexus of the brain (Soprano et al., 1985),being secreted into the blood stream and cerebrospinal fluid,respectively.

Although the mechanisms leading to TTR deposition comprise analteration in the conformational stability of the homotetrameric form,followed by the deposition of unfolded monomers, aggregates oramyloid fibers (Quintas et al., 2001), the mechanisms mediatingcytotoxicity in FAPare not completely understood. Severalmolecularpathways, such as oxidative stress, apoptosis, the heat-shock responseand extracellular matrix remodeling, have been implicated in thepathogenesis of FAP (Saraiva et al., 2012). Inflammation has animportant role not only as a causing agent but also in TTR depositionand consequently neurodegeneration (Sousa et al., 2001a,b;Buxbaum et al., 2012; Gonçalves et al., 2014a). Also, productionof regeneration mediators, such as chemokines and neurotrophins, bySchwann cells is deficient in FAP nerves (Sousa et al., 2001a,b;Gonçalves et al., 2014b). Therefore, the disclosure of novelmechanisms and disease modifying agents underlying nervedegeneration is essential to understand disease progression. In thisregard, work on high-throughput gene arrays has shown that matrixmetalloproteinase-9 (MMP-9) was increased in FAP and is able toin vitro degrade TTR aggregates and fibrils (Sousa et al., 2005).Moreover, MMP-9 is transcriptionally upregulated by NF-κB, atranscription factor that is activated in FAP nerves.

Given the relationship between inflammation and extracellularmatrix (ECM) remodeling, and the increase of proinflammatorycytokines in FAP, we further search for misexpression of otherMMPs in FAP peripheral nerves. MMPs are zinc- and calcium-dependent endopeptidases, identified as matrix-degrading enzymes(Candelario-Jalil et al., 2009), and are members of an extracellularprotease family of collagenases, gelatinases, stromelysins andmembrane-type MMPs (Liu et al., 2010). The regulation of theECM by proteases is a fundamental biological process for normalgrowth, development and repair (Candelario-Jalil et al., 2009). Innormal conditions, the highly coordinate action of these enzymesremodel the components of the matrix and perform essentialfunctions at the cell surface involved in signaling and cell survival.

The membrane-type 1 matrix metalloproteinase (MMP-14) is apro-tumorigenic factor: its upregulation is associated with gliomaexpansion (Markovic et al., 2009; Langenfurth et al., 2014).Moreover, MMP-14 expression is found in microglia/macrophagesassociated with neurodegenerative and neuroinflammatorypathologies both in mouse models and human biopsies(Langenfurth et al., 2014). In patients with Alzheimer’s diseaseReceived 10 November 2016; Accepted 11 August 2017

1Instituto de Inovaçao e Investigaça o em Saude (I3S), Universidade do Porto,R. Alfredo Allen 208, 4200-135 Porto, Portugal. 2Neurobiologia Molecular - Institutode Biologia Molecular (IBMC), Universidade do Porto, R. Alfredo Allen 208, 4200-135 Porto, Portugal.*Present address: Department of Biomedicine/DANDRITE, Aarhus University,Ole Worms Alle 3, 1171, 8000 Aarhus C, Denmark.

‡Author for correspondence (mjsaraiv@ibmc.up.pt)

M.J.S., 0000-0002-3360-6899

This is an Open Access article distributed under the terms of the Creative Commons AttributionLicense (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use,distribution and reproduction in any medium provided that the original work is properly attributed.

1253

© 2017. Published by The Company of Biologists Ltd | Disease Models & Mechanisms (2017) 10, 1253-1260 doi:10.1242/dmm.028571

Disea

seModels&Mechan

isms

(AD), MMP-14 was found overexpressed in brain (Yamada et al.,1995), which was later corroborated by studies using a transgenicmouse model where MMP-14 was found in reactive astrocytes, inregions with fibrillar amyloid deposits (Liao and Van Nostrand,2010). More recently, the importance of MMP-14 in the PNS wasnoted to be related to its association with neuronal glial antigen 2(NG2) proteoglycan onmacrophages and further coordination of theresponse to peripheral nerve injury (Nishihara et al., 2015).Based on the above, in the present study, we investigated the

expression levels of MMP-14 in tissues from FAP patients carryingthe TTR V30M mutation and in a disease mouse model tounderstand whether this molecule could function as a new markerfor neurodegeneration in the pathogenesis of FAP.

RESULTSOverexpression of MMP-14 correlates with TTR non-fibrillardeposition in the nerve of a FAP mouse modelWe started by analyzing MMP-14 expression in sciatic nerve from6-month-old Hsf/V30M mice (Hsf-1-deficient mice transgenic forhuman TTR V30M, in the Sv/129 and endogenous Ttr-nullbackground) by both immunohistochemistry (IHC) and quantitativereal-time polymerase chain reaction (qPCR). Significantoverexpression of MMP-14 protein and gene levels were observedwhen compared to age-matched control mouse strains, i.e. non-transgenic wild-type (WT; Sv129 strain) or Hsf-1-deficient mice(Fig. 1A,B). In both control and transgenic animals, MMP-14staining was observed at background levels in other tissues, with theexception of a few dorsal root ganglia (DRGs), and the difference inthe level of staining between control and transgenic animals was notstatistically significant (not shown). MMP-14 immunopositive co-staining in transgenic peripheral nerve was confirmed with doubleimmunofluorescence labeling between this molecule and S100

(Fig. 1C); however, co-staining between MMP-14 and βIII-tubulindid not occur (not shown), indicating that MMP-14 is localized inSchwann cells.

To evaluate whether increased levels of MMP-14 are associatedwith TTR non-fibrillar deposition, we subsequently studied levelsof this MMP protein in 3-, 6- and 22-month-old Hsf/V30M mice.Based on the results, we could perceive that younger animals havethe lowest levels of MMP-14, which increased with age (Fig. S1),accompanying the increased nerve TTR deposition (Santos et al.,2010), thus suggesting an association between MMP-14upregulation with the progress of the neuropathological featuresof FAP disease.

Treatment of transgenic mice with anakinra or TTR siRNArescued MMP-14 nerve protein levelsPrevious studies demonstrated that treatment of Hsf/V30M micewith anakinra or TTR siRNA prevents TTR non-fibrillar depositionin the peripheral nerve due to modulation of the inflammatoryresponse or silencing liver TTR synthesis, respectively (Gonçalveset al., 2014b; Butler et al., 2016). To understand whether removal ofnerve TTR deposits modulates the local synthesis of MMP-14, weperformed IHC and qPCR for this molecule in nerves from treatedand untreated animals. A significant statistical reduction of MMP-14 protein levels was observed, in both the anakinra- and siRNA-treated groups (Fig. 2A), results that were corroborated at thetranscriptional level (Fig. 2B). In the context of being a novelbiomarker for therapy efficacy, determination of MMP-14 levels inplasma are of major importance. Accordingly, measurement ofMMP-14 protein levels in treated versus untreated animals denotea marked reduction of its levels in plasma after anakinra orTTR-siRNA treatment, possibly related to the reduction in TTRnon-fibrillar deposition (Fig. 2C,D).

Fig. 1. Overexpression ofMMP-14 in a naïve FAPmousemodel. (A)MMP-14 protein levels in the peripheral nerve system (n=6) assessed by semi-quantitativeIHC, with respective semi-quantification (right-hand chart) presented as mean±s.e.m. (***P<0.001), demonstrating an increasing expression of MMP-14 in sciaticnerve of Hsf/V30M mice (scale bars: 100 µm). (B) Upregulation of MMP-14 in the sciatic nerve of Hsf/V30M mice (n=6) evaluated by qPCR (***P<0.001),compared with WT (n=6) and Hsf/WT (n=6) mice. Data were analyzed using one-way ANOVA followed by Bonferroni post-test and represented asmean±s.e.m. Normalization was performed against Gapdh mRNA. (C) Confocal microscopy of transgenic peripheral nerve showing immunopositive stainingof MMP-14 and Schwann cells, as highlighted by colocalization with S100 (scale bar: 10 μm).

1254

RESEARCH ARTICLE Disease Models & Mechanisms (2017) 10, 1253-1260 doi:10.1242/dmm.028571

Disea

seModels&Mechan

isms

Transgenic aggregated-TTR species induce expression ofMMP-14 in Schwann cellsBecause MMP-14 expression in Schwann cells was previouslydescribed (Nishihara et al., 2015), we next performed an in vitroassay in order to determine whether aggregated-TTR toxic speciescould induce synthesis of MMP-14 in these glial cells. Therefore,TTR V30M aggregates were produced and incubated with primarymouse Schwann cells that were then lysed for mRNA extraction.IncreasedMMP-14mRNAwas observed after 16 h incubation withTTR V30M oligomers, in contrast with levels in non-treatedSchwann cells or cells incubated with soluble TTR V30M (Fig. 3).

MMP-14 increases with the ongoing neurodegenerativeprocessBased on the nerve fiber density, a scoring system has been used forthe classification of patient material and stage of disease. In thisway, patients classified as FAP0 are asymptomatic, FAP1 subjectspresent a discrete decrease in nerve fiber density and fibrillar TTRdepositions start to appear in the peripheral nerve, whereas

individuals with FAP2 or 3 have evident or severe fiber reduction,respectively, with extensive amyloid deposition (Sousa et al.,2001a,b). To validate the in vitro and preclinical data describedabove, we next investigated MMP-14 expression in sural nervebiopsies from FAP patients in different stages of disease ascompared with normal control subjects. IHC analyses demonstrateda significant increase in MMP-14 expression in nerve in advancedstages of disease, with non-fibrillar TTR and amyloid depositsbeing simultaneously present in the tissue (Fig. 4A,B).Asymptomatic carriers (FAP0) had no amyloid deposits in thesural nerve and low MMP-14 protein expression was observed. Incontrast, at stage 3 (FAP3), amyloid deposition was heavilydistributed throughout the tissue with massive neurodegeneration.Interestingly, MMP-14 seems to colocalize with amyloid deposits(Fig. 4C).

Additionally, we analyzed MMP-14 concentrations in plasmasamples from patients in different stages of disease (as per the abovecriteria), asymptomatic carriers and normal individuals. We did notnotice a difference in plasma levels between normal, asymptomatic

Fig. 2. Treatments with anakinra or TTR siRNA reduce MMP-14 expression in Hsf/V30M animals. (A) Representative images obtained with IHC for MMP-14using the Hsf/V30M mouse model, showing reduced MMP-14 expression in peripheral nerves of animals treated with anakinra or TTR siRNA. Charts representquantification of the immunohistochemical images (scale bars: 100 µm) and data were analyzed using one-way ANOVA followed by Bonferroni post-test andrepresented as mean±s.e.m. (***P<0.001). (B) Downregulation of MMP-14 in the sciatic nerve of Hsf/V30M mice (n=6/group) evaluated by qPCR (**P<0.01),compared with mice treated with anakinra (n=6). Data were analyzed using one-way ANOVA followed by Bonferroni post-test and represented as mean±s.e.m.Normalization was performed against Gapdh mRNA. (C) Reduced levels of MMP-14, determined by ELISA, in sciatic nerves of mice treated with anakinra (n=6),compared with control mice (n=6). (D) Reduced levels of MMP-14, determined by ELISA, in sciatic nerves of mice treated with TTR siRNA (n=6), compared withcontrol mice (n=6). Data were analyzed using one-way ANOVA followed by Bonferroni post-test and represented as mean±s.e.m. (**P<0.01; ***P<0.001).

1255

RESEARCH ARTICLE Disease Models & Mechanisms (2017) 10, 1253-1260 doi:10.1242/dmm.028571

Disea

seModels&Mechan

isms

and early stages of disease; however, a significant difference inMMP-14 levels was noted in advanced stages (Fig. 5A). Finally, wemeasuredMMP-14 in plasma samples from individuals who receiveda liver transplantation from a donor with FAP (a FAP liver) as a formof treatment for severe liver conditions. We measured concentrationsbefore transplantation, and at day 10 and 120 after domino livertransplantation (DLT). As illustrated in Fig. 5B, significant changes inMMP-14 plasma levels occurred between 10 and 120 days, ascompared with pre-transplant, indicating that systemic molecularchanges are taking place in liver recipient patients.

Taken together, these results suggest that MMP-14 might be apotential disease and therapeutic biomarker in TTR polyneuropathy.

DISCUSSIONIn the present study, we describe for the first time a significantassociation between MMP-14 and FAP, a degenerative peripheralneuropathy. We observed that MMP-14 was significantlyoverexpressed in degenerated FAP nerves presenting massiveamyloid deposits and was also upregulated in sciatic nerves froma preclinical transgenic mouse model. Regulation of ECM is afundamental biological process for normal growth, developmentand repair of the nervous system (Candelario-Jalil et al., 2009),enhancing the rate of sensory axonal regeneration by promotingSchwann cell mitosis (Liu et al., 2010), limiting the extent ofmyelin and axonal damage (Nishihara et al., 2015). Results from thepresent work suggest that major alterations in ECM-related genesoccur in Schwann cells during FAP disease, such as MMP-14misexpression. Our data also demonstrated that treatment with aTTR siRNA or anakinra significantly reduced MMP-14 expressionin Hsf/V30M animals, both in peripheral nerves and plasmasamples, pointing towards a role of both deposit clearance andinflammation in the modulation of MMP-14 expression. In fact,MMP-14 overexpression was also noticed in disorders with a strongneuroinflammatory component in their pathogenesis, such asmultiple sclerosis or stroke, further pointing to a link betweenregulation of MMP-14 and neuroinflammation (Bar-Or et al., 2003;Langenfurth et al., 2014). Furthermore, increased levels ofMMP-14were also found in other amyloid-like neurodegenerative disorderssuch as Alzheimer’s disease (AD). In AD patients, MMP-14 wasfound overexpressed in brain (Yamada et al., 1995), which was latercorroborated by studies using a transgenic mouse model where

Fig. 3. MMP-14 expression is induced by aggregated-TTR species.Upregulation of MMP-14 expression after incubation with Thioflavin-positivemisfolded V30M aggregates (referred to in the chart as TTR oligomers) inmouse Schwann cells evaluated by qPCR, compared with non-treated cellsor cells incubated with soluble TTR. Data were analyzed using one-wayANOVA followed by Bonferroni post-test and represented as mean±s.e.m.(**P<0.01; ***P<0.001).

Fig. 4. MMP-14 expression in FAP patients at different stages of disease. (A) Representative images obtained with IHC for MMP-14 in human sural nerve,showing an increased expression of MMP-14 in FAP patients in different stages of the disease, compared with asymptomatic and control individuals (n=6/group).(B) Chart represents quantification of the immunohistochemical images. Data were analyzed using one-way ANOVA followed by Bonferroni post-test andrepresented as mean±s.e.m. (**P<0.01; ***P<0.001). (C) Representative IHC images of MMP-14 expression and Congo Red (CR) staining showing associationbetween MMP-14 expression and amyloid deposition (n=5/group). Scale bars: 100 µm.

1256

RESEARCH ARTICLE Disease Models & Mechanisms (2017) 10, 1253-1260 doi:10.1242/dmm.028571

Disea

seModels&Mechan

isms

MMP-14 was found in reactive astrocytes, in regions with fibrillaramyloid deposits (Liao and Van Nostrand, 2010). A commonphysiological mechanism involved in amyloidogenic disorders suchas AD or FAP is the activation of inflammatory cascades (Sousaet al., 2001a,b). In FAP, the main inflammatory mechanism relieson, among others, the binding of amyloidogenic TTR species to thereceptor for advanced glycation end products (RAGE) and, as aresult, on activation of extracellular signal-regulated kinases 1/2 andNF-κB, with consequent transcription of pro-inflammatoryproteins, such as IL-1β and TNF-α (tumor necrosis factor-α)(Monteiro et al., 2006; Sousa et al., 2000). Thus, it is possible thatthe increased levels of MMP-14 observed in neurological tissuesresult from the inflammatory process itself, also substantiated by therescue upon treatment with anakinra. Further corroborating theinterplay between inflammation and MMP-14 are recent data fromGao and colleagues (Gao et al., 2015), where an inhibition ofMMPs, including MMP-14, accelerates wound healing in diabeticanimals by decreasing inflammation.MMP-14 is also responsible for pericellular remodeling events

mediated by NG2 in the nervous system. Nishihara and colleaguesshowed that NG2 is expressed by macrophages and progenitor glia,and that NG2 shedding and axonal growth depends on MMP-14,highlighting the relevance of this MMP for axon regeneration, andpinpointing MMP-14 as a novel target in PNS injury andneuropathic pain (Nishihara et al., 2015). In preliminaryexperiments using macrophages or sciatic nerve from FAPtransgenic mice, we did not confirm NG2 shedding in our model(data not shown). However, MMP-14 expression by macrophagesmight explain the entering of this molecule into the circulation andthe augmented levels found in FAP plasma samples. Anotherpossibility is that the increased MMP-14 produced by the Schwanncells could reach the blood stream through the blood-nerve barrier.Further corroborating this last hypothesis is the fact that MMP-14was found increased in the plasma of patients even after livertransplantation, a likely consequence of nerve molecular changes.MMP-14 can cleave other molecules of the ECM, such as

fibronectin, laminin, cell adhesion molecules and even cytokinesand growth factors (Ohuchi et al., 1997; Koshikawa et al., 2000;Kajita et al., 2001; Tam et al., 2004). Additionally, MMP-14 alsoenhances the proteolytic capacity through the activation of otherMMPs, interfering directly with MMP-2 (Ishigaki et al., 1999;Langenfurth et al., 2014). In fact, it was previously described that

MMP-14 is upregulated in glioma-associated microglia/macrophages through pro-MMP-2, suggesting MMP-14 as aprotumorigenic factor (Markovic et al., 2009; Langerfurth et al.,2014). Interestingly, a previous study by our group has alreadyshown an association between MMP-2 and ECM remodeling inFAP. In particular, the inhibition of TTR expression by an siRNAprevented and reversed TTR deposition, with a concomitantdecrease on MMP-2 protein levels in DRG of a transgenic mouseline (Gonçalves et al., 2016). Based on our present results, it ispossible to speculate that MMP-14 overproduced by FAP Schwanncells might consequently be transported to sensory neurons,contributing to the activation of MMP-2.

The cleavage of ECM components byMMPs, includingMMP-14,can also originate damage-associated molecular patterns (DAMPs),which can interact with DAMP receptors, such as Toll-like receptors,amplifying the immune response to injury (Schaefer, 2014) andcreating a feed-forward mechanism of injury in the course of FAPdisease progression. Overall, we hypothesized that overexpression ofMMP-14 in FAP might be associated with the inflammatory processand can also contribute to further remodeling of the ECM, forexample by MMP-2 activation. Nevertheless, the specific role ofMMP-14 in the pathophysiology of FAP remains unclear andwarrants further investigation. So far, the demonstration of increasedMMP-14 in the periphery, together with its augmented levels inplasma from FAP patients, especially in advanced stages of disease,denote a correlation between this MMP and neurodegeneration,highlighting its potential role as a novel disease biomarker or even asa promising future therapeutic target.

MATERIALS AND METHODSHuman samplesAll human samples were collected from donors after written informedconsent for the use of the material and clinical information for researchpurposes.

Formalin-fixed, paraffin-embedded sural nerve biopsy slides wereobtained from FAP patients (n=3 for each disease stage), asymptomaticcarriers (n=3) and matched disease controls (n=4) that were near-relatives ofFAP patients who ultimately turned out not to have mutations in TTR. Thismaterial was kindly provided by the Hospital Geral de Santo António, Porto,Portugal and obtained as part of the clinical diagnosis, before the current useof less invasive methods, as previously described (Sousa et al., 2001a,b).The material was previously characterized, after informed consent, for TTRreactivity and amyloid deposition by histological analyses. Additionally,

Fig. 5. MMP-14 levels, determined by ELISA, on plasma from human patients in different stages of disease. (A) Increased levels of MMP-14 in FAP2 andFAP3 plasma samples, as compared to controls (n=5/group). (B) Increased plasma concentration levels of MMP-14 in patients transplanted with a FAPliver (10 and 120 days post-transplant), as compared with before transplant. Data were analyzed using one-way ANOVA followed by Bonferroni post-test andrepresented as mean±s.e.m. (*P<0.05; **P<0.01; ***P<0.001).

1257

RESEARCH ARTICLE Disease Models & Mechanisms (2017) 10, 1253-1260 doi:10.1242/dmm.028571

Disea

seModels&Mechan

isms

human plasma samples from control subjects, asymptomatic carriers andFAP patients in different stages of disease (n=5 for each group) were alsoavailable after informed consent. Plasma samples from individuals who hadundergone DLT, hosting the liver of a FAP V30M patient, were availablebefore DLT (n=5) and subsequently collected at different times after DLT(10 and 120 days; n=5 and n=2, respectively) after approval by the EthicsCommittee of the Transplantation Department, University of Coimbra.

AnimalsAll animal experiments were carried out in accordance with National andEuropean Union guidelines for the care and handling of laboratory animals,and were performed in compliance with the institutional guidelinesand recommendations of the Federation for Laboratory Animal ScienceAssociation (FELASA) and approved by the National Authority for AnimalHealth (DGAV; Lisbon, Portugal).

In the present study, transgenic mice for human TTRV30M, in the Sv/129and endogenous Ttr-null background, with heterozygous deletion of the geneencoding transcription factor Hsf-1 (labeled as Hsf/V30M) (Santos et al.,2010) were analyzed at 3, 6 and 22 months of age (n=6 of each age).WTmiceand WT animals heterozygous for Hsf-1 deletion, both in the Sv/129background, were used as controls (n=6 of each strain). Mice were housed inpathogen-free conditions in a temperature-controlled room and weremaintained on a 12-h light/dark cycle, with water and food ad libitum. Foranimal sacrifice, a lethal injection of a premixed solution containing ketamineplus medetomidine was used. Plasma, sciatic nerve, DRG and other tissuestypical of TTR deposition in FAP, such as the gastrointestinal tract, kidney andheart, were collected, as well as brain, liver and pancreas, representing tissuesof TTR synthesis. They were frozen at −80°C, or fixed in formalin.

Drug design and animal treatmentFor anakinra (Kineret®, Biovitrum) treatment, Hsf/V30M mice (4.5 monthsof age) were daily injected subcutaneously with 25 mg anakinra/kg bodyweight over 6 weeks. Age-matched untreated control mice were injectedwith phosphate buffered saline (PBS). Because mice respond to anakinra ina dose range of 1 to 100 mg/kg (Abbate et al., 2008; Salloum et al., 2009), inthis work we chose the intermediate dose of 25 mg/kg, as previouslydescribed (Gonçalves et al., 2014a).

For TTR siRNA treatment, Hsf/V30M mice (5 months) were injected inthe tail vein with a human TTR siRNA (Semple et al., 2010), at aconcentration of 1 mg/kg over 4 weeks. Untreated age-matched controlsreceived vehicle only (n=4). In both groups, one intravenous injection wasperformed per week and animals were euthanized 48 h after the lastinjection.

Semi-quantitative immunohistochemistryParaffin sections of human nerve and mouse sciatic nerve and sensoryneurons were deparaffinized in Histoclear™ (National diagnostics, Atlanta,GA, catalog #HS-200) and hydrated. Endogenous peroxidase activity wasblocked with 3% H2O2 in methanol for 30 min and sections were pretreatedwith blocking solution (10% fetal bovine serum and 0.5% Triton X-100 inPBS) to avoid undesired background staining. Sections were incubatedovernight at 4°C with rabbit anti-MMP-14 (1:75, Millipore, Darmstadt,Germany, catalog #AB8345) in blocking buffer and, after washing, sectionswere incubated with the secondary biotinylated antibody horse anti-rabbitIgG (1:200, Vector Laboratories, Burlingame, CA, catalog #BA-1100).Avidin-biotin peroxidase complex (ABC Elite, Vector Laboratories,catalogb#PK-6100) labeling was performed with 3,3′-diaminobenzidine(Sigma, Darmstadt, Germany, catalog #D5637-1G) for antigenvisualization. Sections were counterstained with hematoxylin (Millipore,catalog #1051752500) for 20 s, mounted using Entelan (Millipore,catalog#1079610500) and coverslipped for microscope visualization.

Semi-quantitative analysis of immunostaining was performed in 5different areas at 20× magnification, representative of the whole tissue,using the Image pro-plus 5.1 software (Media Cybernetics, Rockville, MD).Quantification was blind. The results shown represent the percentage ofoccupied area by pixels corresponding to the substrate reaction colornormalized relative to the total area occupied by the tissue, with thecorresponding standard error of the mean (s.e.m.).

Congo Red stainingThe presence of amyloid deposits was investigated in human samples byCongo Red staining (Puchtler and Sweat, 1965). Briefly, deparaffinizedtissue sections were incubated for 20 min with 0.01%NaOH in 80% ethanolsaturated with NaCl followed by staining with 0.5% Congo Red in theprevious solution. The slides were stained with hematoxylin and analyzedunder polarized light. Amyloid was identified by the characteristic greenbirefringence.

Confocal microscopyDouble immunofluorescence was performed to colocalize MMP-14 withSchwann cells (axons). After deparaffinazation, sections were hydrated andpermeabilized with 0.5% Triton X-100 for 10 min and then blocked withphosphate buffer supplemented with 4% fetal bovine serum and 1% bovineserum albumin for 1 h at room temperature. After blocking, tissues wereincubated overnight at 4°C with the primary antibodies: mouse monoclonalanti-βIII-tubulin (1:500, Promega), mouse monoclonal anti-S100 (1:100,Abcam), rabbit polyclonal anti-MMP-14 (1:75, Millipore). The followingday, sections were incubated with secondary antibodies: Alexa-Fluor-488-conjugated goat anti-mouse IgG and goat Alexa-Fluor-568-conjugated anti-rabbit IgG (1:1000, Molecular Probes) for 2 h at room temperature. Finally,sections were mounted using Vectashield with 4′,6-diamino-2-phenylindole(DAPI; Vector Laboratories) for nuclei visualization. Image stacks wereobtained using the laser scanning confocal microscope Leica TCS SP5 II(Leica Microsystems).

Enzyme-linked immunosorbent assayThe levels of MMP-14 in human plasma of non-V30M-carrier controls(n=5), asymptomatic subjects (n=5) and FAP patients in different stages ofdisease (FAP1, FAP2 and FAP3) (n=5 per group) were quantitativelydetermined by enzyme-linked immunosorbent assay (ELISA), according tothe manufacturer′s instructions (R&D Systems, Minneapolis, MN, catalog#DY918-05).

The levels of MMP-14 in plasma from Hsf/V30M mice before and aftertreatment with anakinra and TTR siRNA were also quantitativelydetermined, according to the manufacturers’ instructions.

RNA extraction, cDNA synthesis and qPCRRNA from sciatic nerves was isolated with RNeasy Mini columns, followingthe manufacturer′s instructions (Qiagen, Hilden, Germany, catalog #74804).cDNA synthesis was performed using the SuperScript II reverse transcriptase(Invitrogen, Waltham, MA, catalog #18064014). The primers to target MMP-14 were designed with the Beacon Designer 8 software (Premier BiosoftInternational) and the MMP-14 mouse forward 5′-TTACAAGTGACAGG-CAAGG-3′ and mouse MMP-14 reverse 5′-GCTTCCTCCGAACATTGG-3′. A standard calibration curve was performed by 10-fold serial cDNAdilutions, to assess primer efficiency. qPCR was performed in duplicatesusing an IQ5multi-color real-time detection systemwith IQSybrGreen SuperMix and gene expression analysis performed with Bio-Rad iQ5 software(Bio-Rad Laboratories, Hercules, CA, catalog #1708880). Melting curveswere produced in order to detect non-specific amplification products. Fivebiological replicates were used per group. Differential expression wasdetermined by the 2−ΔΔCT method, using the Gapdh gene for normalization.

Preparation of TTR V30M aggregatesRecombinant TTRV30Mwas produced inEscherichia coliBL21 and purified(Furuya et al., 1991). Soluble TTRV30Mwas filtered through 0.2 μmAnotopsyringe filters (Whatman, England) and TTR aggregates generated byincubating the protein (2 mg/ml), with stirring, at room temperature for7 days (Teixeira et al., 2006). TTR preparations were then tested by ThioflavinT spectrofluorometric assay to search for their amyloidogenic potential.Dynamic light scattering (DLS) at 25°C in a Malvern Zetasizer Nano ZS(Malvern,Worcestershire, UK)was also performed to confirmTTRpathogenicaggregation and specimen size (Ferreira et al., 2011).

Schwann cell primary cultureMouse Schwann cell primary cultures were obtained from ScienCellResearch Laboratories™ (San Diego, CA, catalog #M1700) and were

1258

RESEARCH ARTICLE Disease Models & Mechanisms (2017) 10, 1253-1260 doi:10.1242/dmm.028571

Disea

seModels&Mechan

isms

cultured according to the manufacturer’s instructions. Briefly, cells werecultured in Schwann cell medium, supplemented with 5% fetal bovineserum, 1% Schwann cell growth supplement and 1% penicillin/streptomycin (all from ScienCell Research Laboratories). After monolayerpropagation in T-75 flasks at 37°C in a 5% CO2 humidified chamber, 5×105

cells were seeded in 12-well plates and stimulated with TTR aggregates orsoluble protein, both at 2.5 μM in duplicates. Twenty-four hours afterstimulation, cell lysates were prepared in trizol (Invitrogen, Waltham, MA,catalog #15596026) and assayed for the expression of MMP-14 mRNA byqPCR. Unstimulated cells were also used as controls.

Statistical analysisStatistical comparison of datawas performed using Student’s t-test or one-wayanalysis of variance (ANOVA) followed by Bonferroni post-test, withGraphPad Prism software. Quantitative data are expressed as mean±s.e.m.Statistical significancewas established for *P<0.05, **P<0.01, ***P <0.001.

AcknowledgementsWe thank Paula Gonçalves, Paul Moreira and Sofia Martins from IBMC for paraffintissue processing, protein production and confocal microscopy, respectively.

Competing interestsThe authors declare no competing or financial interests.

Author contributionsConceptualization: M.J.S.; Methodology: D.M., J.M., N.P.G.; Investigation: D.M.,J.M., M.J.S.; Writing - original draft: D.M.; Writing - review & editing: N.P.G., M.J.S.;Supervision: M.J.S.; Project administration: M.J.S.; Funding acquisition: M.J.S.

FundingThis work was financed by FEDER (Fundo Europeu de Desenvolvimento Regional)funds (European Regional Development Fund) through the COMPETE 2020 –

Operational Programme for Competitiveness and Internationalisation (POCI), Portugal2020, and by Portuguese funds through FCT (Fundaçao para aCiência e a Tecnologia/Ministerio daCiência, Tecnologia e Inovaçao in the frameworkof the project ‘Institute forResearch and Innovation in Health Sciences’ (POCI-01-0145-FEDER-007274). Thiswork was also the result of funding from project NORTE-01-0145-FEDER-000008,supported by Norte Portugal Regional Operational Programme (NORTE 2020), underthe PORTUGAL 2020 Partnership Agreement, through the European RegionalDevelopment Fund (ERDF). A contribution fromCordex, Portugal (www.cordex.com) isalso acknowledged. A Ph.D. fellowship to N.P.G. (SFRH/BD/74304/2010) and apostdoctoral fellowship to D.M. (ON2-201302-ND-III), supported by FEDER fundsthrough the Operational Competitiveness Program – COMPETE, by national fundingfrom the FCT, are acknowledged.

Supplementary informationSupplementary information available online athttp://dmm.biologists.org/lookup/doi/10.1242/dmm.028571.supplemental

ReferencesAbbate, A., Salloum, F. N., Vecile, E., Das, A., Hoke, N. N., Straino, S., Biondi-Zoccai,G.G. L., Houser, J.-E.,Qureshi, I. Z., Ownby, E.D. et al. (2008). Anakinra,a recombinant human interleukin-1 receptor antagonist, inhibits apoptosis inexperimental acute myocardial infarction. Circulation 117, 2670-2683.

Bar-Or, A., Nuttall, R. K., Duddy, M., Alter, A., Kim, H. J., Ifergan, I., Pennington,C. J., Bourgoin, P., Edwards, D. R. and Yong, V. W. (2003). Analyses of allmatrix metalloproteinase members in leukocytes emphasize monocytes as majorinflammatory mediators in multiple sclerosis. Brain 126, 2738-2749.

Benson, M. D. and Kincaid, J. C. (2007). The molecular biology and clinicalfeatures of amyloid neuropathy. Muscle Nerve 36, 411-423.

Butler, J. S., Chan, A., Costelha, S., Fishman, S., Willoughby, J. L. S., Borland,T. D., Milstein, S., Foster, D. J., Gonçalves, P., Chen, Q. et al. (2016). Preclinicalevaluation of RNAi as a treatment for transthyretin-mediated amyloidosis.Amyloid23, 109-118.

Buxbaum, J. N., Tagoe, C., Gallo, G., Walker, J. R., Kurian, S. and Salomon,D. R. (2012). Why are some amyloidoses systemic? Does hepatic “chaperoning ata distance” prevent cardiac deposition in a transgenic model of human senilesystemic (transthyretin) amyloidosis? FASEB J. 26, 2283-2293.

Candelario-Jalil, E., Yang, Y. and Rosenberg, G. A. (2009). Diverse Roles ofMatrix Metalloproteinases and Tissue Inhibitors of Metalloproteinases inNeuroinflammation and cerebral Ischemia. Neuroscience 158, 983-994.

Coimbra, A. and Andrade, C. (1971). Familial amyloid polyneuropathy: an electronmicroscope study of the peripheral nerve in five cases. II. Nerve fibre changes.Brain 94, 207-212.

Ferreira, N., Saraiva, M. J. and Almeida, M. R. (2011). Natural polyphenols inhibitdifferent steps of the process of transthyretin (TTR) amyloid fibril formation. FEBSLett. 585, 2424-2430.

Furuya, H., Saraiva, M. J. M., Gawinowicz, M. A., Alves, I. L., Costa, P. P., Sasaki,H., Goto, I. and Sakaki, Y. (1991). Production of recombinant human transthyretinwith biological activities toward the understanding of themolecular basis of familialamyloidotic polyneuropathy (FAP). Biochemistry 30, 2415-2421.

Gao, M., Nguyen, T. T., Suckow, M. A., Wolter, W. R., Gooyit, M., Mobashery, S.and Chang, M. (2015). Acceleration of diabetic wound healing using a novelprotease-anti-protease combination therapy. Proc. Natl. Acad. Sci. USA 112,15226-15231.

Gonçalves, N. P., Vieira, P. and Saraiva, M. J. (2014a). Interleukin-1 signalingpathway as a therapeutic target in transthyretin amyloidosis. Amyloid 21, 175-184.

Gonçalves, N. P., Teixeira-Coelho, M. and Saraiva, M. J. (2014b). Theinflammatory response to sciatic nerve injury in a familial amyloidoticpolyneuropathy mouse model. Exp. Neurol. 257, 76-87.

Gonçalves, N. P., Gonçalves, P., Magalhaes, J., Ventosa, M., Coelho, A. V. andSaraiva, M. J. (2016). Tissue remodeling after interference RNA mediatedknockdown of transthyretin in a familial amyloidotic polyneuropathymousemodel.Neurobiol. Aging 47, 91-101.

Ishigaki, S., Toi, M., Ueno, T., Matsumoto, H., Muta, M., Koike, M. and Seiki, M.(1999). Significance of membrane type 1 matrix metalloproteinase expression inbreast cancer. Jpn J. Cancer Res. 90, 516-522.

Kajita, M., Itoh, Y., Chiba, T., Mori, H., Okada, A., Kinoh, H. and Seiki, M. (2001).Membrane-type 1 matrix metalloproteinase cleaves CD44 and promotes cellmigration. J. Cell Biol. 153, 893-904.

Koshikawa, N., Giannelli, G., Cirulli, V., Miyazaki, K. and Quaranta, V. (2000).Role of cell surface metalloprotease MT1-MMP in epithelial cell migration overlaminin-5. J. Cell Biol. 148, 615-624.

Langenfurth, A., Rinnenthal, J. L., Vinnakota, K., Prinz, V., Carlo, A.-S.,Stadelmann, C., Siffrin, V., Peaschke, S., Endres, M., Heppner, F. et al.(2014). Membrane-type 1 metalloproteinase is upregulated in microglia/brainmacrophages in neurodegenerative and neuroinflammatory diseases. J. Neurosci.Res. 92, 275-286.

Liao, M. C. and Van Nostrand, W. E. (2010). Degradation of soluble and fibrillaramyloid beta-protein by matrix metalloproteinase (MT1-MMP) in vitro.Biochemistry 49, 1127-1136.

Liu, H., Kim, Y., Chattopadhyay, S., Shubayev, I., Dolkas, J. and Shubayev, V. I.(2010). Matrix metalloproteinase inhibition enhances the rate of nerveregeneration in vivo by promoting de-differentiation and mitosis of supportingSchwann cells. J. Neuropathol. 69, 386-395.

Markovic, D. S., Vinnakota, K., Chirasani, S., Synowitz, M., Raguet, H., Stock,K., Sliwa, M., Lehmann, S., Kalin, R., van Rooijen, N. et al. (2009). Gliomasinduce and exploit microglial MT1-MMP expression for tumor expansion. Proc.Natl. Acad. Sci. USA 106, 12530-12535.

Monteiro, F. A., Sousa, M. M., Cardoso, I., do Amaral, J. B., Guimaraes, A. andSaraiva, M. J. (2006). Activation of ERK1/2 MAP kinases in familial amyloidoticpolyneuropathy. J. Neurochem. 97, 151-161.

Murakami, T., Sango, K., Watabe, K., Niimi, N., Takaku, S., Li, Z., Yamamura,K.-i. and Sunada, Y. (2015). Schwann cells contribute to neurodegeneration intransthyretin amyloidosis. J. Neurochem. 134, 66-74.

Nishihara, T., Remacle, A. G., Angert, M., Shubayev, I., Shiryaev, S. A., Liu, H.,Dolkas, J., Chernov, A. V., Strongin, A. Y. and Shubayev, V. I. (2015). Matrixmetalloproteinase-14 both sheds cell surface neuronal glial antigen 2 (NG2)proteoglycan on macrophages and governs the response to peripheral nerveinjury. J. Biol. Chem. 290, 3693-3707.

Ohuchi, E., Imai, K., Fujii, Y., Sato, H., Seiki, M. and Okada, Y. (1997). Membranetype 1 matrix metalloproteinase digests interstitial collagen.s and otherextracellular matrix macromolecules. J. Biol. Chem. 272, 2446-2451.

Plante-Bordeneuve, V. and Said, G. (2011). Familial amyloid polyneuropathy.Lancet Neurol. 10, 1086-1097.

Puchtler, H. and Sweat, F. (1965). Congo red as a stain for fluorescencemicroscopy of amyloid. J. Histochem. Cytochem. 13, 693-694.

Quintas, A., Vaz, D. C., Cardoso, I., Saraiva, M. J. M. and Brito, R. M. M. (2001).Tetramer dissociation and monomer partial unfolding precedes protofibrilformation in amyloidogenic transthyretin variants. J. Biol. Chem. 276,27207-27213.

Salloum, F. N., Chau, V., Varma, A., Hoke, N. N., Toldo, S., Biondi-Zoccai, G. G.,Crea, F., Vetrovec, G. W. and Abbate, A. (2009). Anakinra in experimental acutemyocardial infarction-does dosage or duration of treatment matter? Cardiovasc.Drugs Ther. 23, 129-135.

Santos, S. D., Fernandes, R. and Saraiva, M. J. (2010). The heat shock responsemodulates transthyretin deposition in the peripheral and autonomic nervoussystems. Neurobiol. Aging 31, 280-289.

Saraiva, M. J., Birken, S., Costa, P. P. and Goodman, D. S. (1984). Amyloid fibrilprotein in familial amyloidotic polyneuropathy, Portuguese type. Definition ofmolecular abnormality in transthyretin (prealbumin). J. Clin. Invest. 74, 104-119.

Saraiva, M. J., Magalhaes, J., Ferreira, N. and Almeida, M. R. (2012).Transthyretin deposition in familial amyloidotic polyneuropathy. Curr. Med.Chem. 19, 2304-2311.

1259

RESEARCH ARTICLE Disease Models & Mechanisms (2017) 10, 1253-1260 doi:10.1242/dmm.028571

Disea

seModels&Mechan

isms

Schaefer, L. (2014). Complexity of danger: the diverse nature of damage-associated molecular patterns. J. Biol. Chem. 289, 35237-35245.

Semple, S. C., Akinc, A., Chen, J., Sandhu, A. P., Mui, B. L., Cho, C. K., Sah,D. W. Y., Stebbing, D., Crosley, E. J., Yaworski, E. et al. (2010). Rational designof cationic lipids for siRNA delivery. Nat. Biotechnol. 28, 172-176.

Soprano, D. R., Herbert, J., Soprano, K. J., Schon, E. A. and Goodman, D. S.(1985). Demonstration of transthyretin mRNA in the brain and other extrahepatictissues in the rat. J. Biol. Chem. 260, 11793-11798.

Sousa, M. M., Yan, S. D., Stern, D. and Saraiva, M. J. (2000). Interaction of thereceptor for advanced glycation end products (RAGE) with transthyretin triggersnuclear transcription factor kB (NF-kB) activation. Lab. Invest. 80, 1101-1110.

Sousa, M. M., Cardoso, I., Fernandes, R., Guimaraes, A. and Saraiva, M. J.(2001a). Deposition of transthyretin in early stages of familial amyloidoticpolyneuropathy: evidence for toxicity of nonfibrillar aggregates. Am. J. Pathol.159, 1993-2000.

Sousa, M. M., Du Yan, S., Fernandes, R., Guimaraes, A., Stern, D. and Saraiva,M. J. (2001b). Familial amyloid polyneuropathy: receptor for advanced glycation

end products-dependent triggering of neuronal inflammatory and apoptoticpathways. J. Neurosci. 21, 7576-7586.

Sousa, M. M., do Amaral, J. B., Guimaraes, A. and Saraiva, M. J. (2005). Up-regulation of the extracellular matrix remodeling genes, biglycan, neutrophilgelatinase-associated lipocalin, andmatrix metalloproteinase-9 in familial amyloidpolyneuropathy. FASEB J. 19, 124-126.

Tam, E. M., Morrison, C. J., Wu, Y. I., Stack, M. S. and Overall, C. M. (2004).Membrane protease proteomics: Isotope-coded affinity tag MS identification ofundescribedMT1-matrix metalloproteinase substrates.Proc. Natl. Acad. Sci. USA101, 6917-6922.

Teixeira, P. F., Cerca, F., Santos, S. D. and Saraiva, M. J. (2006). Endoplasmicreticulum stress associated with extracellular aggregates. Evidence fromtransthyretin deposition in familial amyloid polyneuropathy. J. Biol. Chem. 281,21998-22003.

Yamada, T., Yoshiyama, Y., Sato, H., Seiki, M., Shinagawa, A. and Takahashi, M.(1995). White matter microglia produce membrane-type matrix metalloprotease,an activator of gelatinase A, in human brain tissues. Acta Neuropathol. 90,421-424.

1260

RESEARCH ARTICLE Disease Models & Mechanisms (2017) 10, 1253-1260 doi:10.1242/dmm.028571

Disea

seModels&Mechan

isms