nephrin expression in adult rodent central nervous system and its … · 2011. 11. 16. · nephrin...

ORIGINAL PAPERJournal of PathologyJ Pathol 2011; 225: 118–128Published online 1 June 2011 in Wiley Online Library(wileyonlinelibrary.com) DOI: 10.1002/path.2923

Nephrin expression in adult rodent central nervous systemand its interaction with glutamate receptors

Min Li,1 Silvia Armelloni,1 Masami Ikehata,1 Alessandro Corbelli,1,2 Marzia Pesaresi,1,3# Novella Calvaresi,1Laura Giardino,1 Deborah Mattinzoli,1 Francesca Nistico,1 Serena Andreoni,1 Aldamaria Puliti,4,5

Roberto Ravazzolo,4,5 Gianluigi Forloni,3 Piergiorgio Messa1 and Maria Pia Rastaldi1*

1 Renal Research Laboratory, Department of Nephrology, Dialysis and Renal Transplantation, Fondazione IRCCS Ca’ Granda Ospedale MaggiorePoliclinico and Fondazione D’Amico per la Ricerca sulle Malattie Renali, Milan, Italy2 MIA Consortium, Universita Milano Bicocca, Monza, Italy3 Department of Neuroscience, Mario Negri Institute, Milan, Italy4 Department of Paediatric Sciences, University of Genova, Italy5 Molecular Genetics and Cytogenetics Unit, Gaslini Insitute, Genova, Italy

*Correspondence to: Maria Pia Rastaldi, Renal Research Laboratory, Fondazione IRCCS Policlinico, Via Pace 9, 20122 Milano, Italy.e-mail: [email protected]

#Present address: Department of Endocrinology, Pathophysiology and Applied Biology (DEFIB), Centre of Excellence on NeurodegenerativeDiseases (CEND), Universita degli Studi di Milano, 20133 Milano, Italy.

AbstractNephrin is an immunoglobulin-like adhesion molecule first discovered as a major component of the podocyteslit diaphragm, where its integrity is essential to the function of the glomerular filtration barrier. Outside thekidney, nephrin has been shown in other restricted locations, most notably in the central nervous system (CNS)of embryonic and newborn rodents. With the aim of better characterizing nephrin expression and its role in theCNS of adult rodents, we studied its expression pattern and possible binding partners in CNS tissues and culturedneuronal cells and compared these data to those obtained in control renal tissues and podocyte cell cultures. Ourresults show that, besides a number of locations already found in embryos and newborns, endogenous nephrinin adult rodent CNS extends to the pons and corpus callosum and is expressed by granule cells and Purkinjecells of the cerebellum, with a characteristic alternating expression pattern. In primary neuronal cells we findnephrin expression close to synaptic proteins and demonstrate that nephrin co-immunoprecipitates with Fynkinase, glutamate receptors and the scaffolding molecule PSD95, an assembly that is reminiscent of those madeby synaptic adhesion molecules. This role seems to be confirmed by our findings of impaired maturation andreduced glutamate exocytosis occurring in Neuro2A cells upon nephrin silencing. Of note, we disclose that thevery same nephrin interactions occur in renal glomeruli and cultured podocytes, supporting our hypothesis thatpodocytes organize and use similar molecular intercellular signalling modules to those used by neuronal cells.Copyright 2011 Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd.

Keywords: nephrin; central nervous system; adhesion molecules; glutamate receptors; neuronal cells; podocytes

Received 21 January 2011; Revised 5 April 2011; Accepted 13 April 2011

No conflicts of interest were declared.

Introduction

In renal glomeruli, podocytes are terminally differen-tiated, highly ramified cells ultimately responsible forglomerular filtration. From each podocyte’s cell body,which bulges into the urinary space, a number of pri-mary processes extend and further divide, giving riseto interdigitating secondary processes that completelyenwrap the glomerular basement membrane.

Starting from these morphological aspects, recentresearch has emphasized biochemical and functionalsimilarities between podocytes and neuronal cells, andour group has contributed to this field of investigationby demonstrating a podocyte-driven neuron-like system

of signalling composed by functional synaptic-likevesicles and cognate neurotransmitter receptors [1,2].Podocytes and neurons also share a similar cytoskele-tal organization, common pathways for process for-mation [3] and several expression-restricted molecules,remarkably among them nephrin [4].

In podocytes, nephrin is a major component ofthe slit diaphragm; the highly specialized adhesionstructure that connects podocyte processes and isactively involved in glomerular filtration [5].

In the developing central nervous system (CNS)nephrin expression was first described in the fourthventricle, spinal cord, hippocampus, olfactory bulb andradial glial cells of the cerebellum [4,6]. Other groupshave described nephrin in the medulla oblongata and

Copyright 2011 Pathological Society of Great Britain and Ireland. J Pathol 2011; 225: 118–128Published by John Wiley & Sons, Ltd. www.pathsoc.org.uk www.thejournalofpathology.com

Nephrin in neurons and podocytes 119

cerebellum [7] and in the cerebral cortex and choroidplexus [8]. Although these studies have been mostlyinformative on nephrin expression in embryos andnewborns, localization of endogenous nephrin in theadult rodent CNS has never been described.

Nephrin is a cell adhesion molecule (CAM) of theimmunoglobulin (Ig) superfamily. In addition to its rec-ognized function of maintaining the integrity of theslit diaphragm, nephrin is a powerful signalling pro-tein. In the podocyte, nephrin interacts with the tyrosinekinase Fyn, which phosphorylates tyrosine residues inthe cytoplasmic domain of nephrin [9,10] and regu-lates nephrin signalling. The relevance of this processis demonstrated by Fyn-deficient mice, which showeffacement of podocyte foot processes and developsignificant proteinuria [11]. Interestingly, Fyn mutantsalso display impaired long-term potentiation and spa-tial learning [12], and in neuronal cells Fyn interactswith the Ig-like adhesion molecule NCAM within raft-like microdomains [13]. The cytoplasmic insertion siteof the glomerular slit diaphragm and the postsynapticdensity (PSD) of neurons are both regions of Triton X-100-resistant electron-dense material [14,15] and bearcommon features. In the PSD, multiple neurotransmit-ter receptors and ion channels are physically linkedto synaptic CAMs through a variety of adaptor pro-teins, which in turn are connected to actin filamentsand their associated proteins [16]. Similarly, at the slitdiaphragm, multiple CAMs of the Ig-like and cadherinfamily are connected to actin through a variety of adap-tor proteins [17].

With these premises, we first conducted a detailedanalysis of nephrin expression in the CNS of adultrodents and examined its possible interaction withFyn. Then we compared the expression pattern andinteractions of nephrin in primary neuronal cells andpodocyte cell cultures, to get information on the roleof this molecule in both cell types.

Methods

TissuesThe experimental protocol, conducted according to cur-rent national regulations (D.L.116-27/01/1992), wasapproved by the ethical committee for animal exper-imentation of Milan University. Control CNS and kid-ney tissues were from 3 month-old C57BL/6N miceand Sprague–Dawley rats.

For immunofluorescence, brain and cerebellum weretaken after 4% buffered paraformaldehyde perfusion,and washed in 10% sucrose in phosphate buffer(PBS), pH 7.4, overnight. Kidneys were taken aftercold PBS perfusion. Tissues were then embedded inoptimum cutting temperature cryo-embedding matrix(OCT; Miles Scientific, Naperville, IL, USA), snap-frozen in a mixture of isopentane and dry ice andstored at −80 ◦C. Coronal and sagittal CNS sections(cut as depicted in Figures S1, S2; see Supporting

information) and kidney sections were placed on slidesbefore immunostaining.

Cell culturesPrimary cortical neurons were prepared from 2 day-old C57BL/6N mice and Sprague–Dawley rats. Non-neuronal contamination of the cultures, assessed asdescribed [18], was <3%.

The mouse neuroblastoma cell line Neuro2A (ATCC,cat. no. CCL-131TM), a commonly used model systemfor neuronal differentiation [19], was utilized for trans-fection experiments. The cell line is characterized byrounded cells that become ramified upon incubationin differentiating medium (see Supporting information,Figure S3a, b). Notably for our purposes, nephrin isexpressed by these cells (RT–PCR method describedin Supporting information, Figure S3c).

Primary podocyte cultures were obtained as previ-ously described [1] and the culture methods are detailedin the Supporting information.

ImmunostainingIndirect immunofluorescence was applied to 5 µmthick acetone-fixed kidney sections, 8 µm thick para-formaldehyde-fixed brain sections, acetone-fixed pri-mary podocytes and paraformaldehyde–sucrose-fixedcultured neurons. Details and immunogold methods areprovided in the Supporting information.

Proximity ligation assayProximity ligation assay (PLA) (20) was performedwith Duolink in situ PLA kit (Olink Bioscience,Uppsala, Sweden). Tissues and cells were incubatedwith the first primary antibody [rabbit anti-nephrin(Santa Cruz Biotechnologies); or rabbit monoclonalanti-PSD95 (Epitomics, HistoLine, Milan, Italy)], fol-lowed by the second primary antibody [mouse anti-NMDA1 receptor or mouse anti-PSD95 (both fromAbcam)]. Anti-rabbit IgG and anti-mouse IgG sec-ondary antibodies conjugated with oligonucleotides(PLA probes) were subsequently applied, sequentiallyfollowed by hybridization, ligation, amplification anddetection solution. The advantage of this method isthat it generates red fluorescence signals from oligonu-cleotide pair amplification only when the two PLAprobes are in close proximity (<35–40 nm) [20].

To assess the methodology, positive controls wererun on cultured podocytes and kidney sections usingprimary antibodies directed against synaptopodin andα-actinin (see Supporting information, Figure S4a–f),because their binding is known to occur [21]. Substitu-tion of control immunoglobulins for one of the primaryantibodies served as negative controls (see Supportinginformation, Figure S4g–l).

Membranes detergent extraction,immunoprecipitation and western blottingMembrane separation was conducted by sequentialcentrifugation. Immunoprecipitation and western blots


120 M Li et al

were performed according to established methodologies(details are provided in the Supporting information).

siRNA knockdown of Nphs1 expression in Neuro2AcellsNeuro2A cells were used after 2 days of culture in dif-ferentiating medium. To down-regulate nephrin expres-sion, they were incubated in antibiotic-free medium for24 h, then transfected with 10 nM siRNA duplexes,using Lipofectamine 2000 (Invitrogen) as the trans-fection agent. We used a pool of three commerciallyavailable siRNAs complementary to Nphs1 mRNA(SASI_Mm01_00 152 484, SASI_Mm01_00 152 485and SASI_Mm01_00 152 487). As control, non-targe-ting siRNAs were applied at the same concentration (allfrom Sigma-Aldrich). After 24 h incubation at 37 ◦C,the medium was replaced with fresh medium contain-ing antibiotics.

Transfection efficiency was determined by a siRNAfluorescent-tagged (AlexaFluor 488, Amersham, Per-kin-Elmer, Waltham, MA, USA) and was >70% asestimated from fluorescence distribution. Silencing ofnephrin and its cellular effects were evaluated 72 hafter siRNA transfection in parallel immunofluores-cence and western blot analyses.

Glutamate release assayGlutamate release was detected by an enzymatic assay[22] based on the following reaction that occurs in thepresence of glutamate dehydrogenase:

Glutamate− + NAD+ + H2O ↔ ketoglutarate2−

+NADH + NAD4+ + H+

After thorough washes with medium without L-glutamine, cells were incubated for 5 min in mediumsupplemented with glutamate dehydrogenase (60 U/ml)and NAD+ (1 mM). Then, medium (to evaluate spon-taneous exocytosis) or 2.5 nM α-latrotoxin (LTX; todetect regulated exocytosis) were added and the spec-trophotometric increase of optical density (OD) dueto increase of NADH was monitored at 340 nm every5 min. These reagents were from Sigma-Aldrich.

Results

Endogenous nephrin expression in adult rodent CNSWe first verified mRNA expression of nephrin byRT–PCR on total cDNA from adult rodent brain,obtaining bands of the expected size (see Supportinginformation, Figure S5). Sequencing of PCR productsrevealed 100% identity with rat nephrin sequence.

Immunofluorescence was then conducted on tissuesections. Renal control tissue was used to demon-strate antibody specificity, with selective stainingof the glomerular tuft (see Supporting information,Figure S6a) and renal nerves (Figure S6b). In the

brain, nephrin-positive cells were diffusely identi-fied in the motor cortex, whereas the somatosensorycortex appeared completely negative. Sparse positivecells were present in the corpus callosum and a dif-fuse expression was found in the choroid plexus (seeSupporting information, Figure S7). Stronger nephrinstaining was detected in the pons, medulla oblon-gata and olfactory bulb (Figures 1a,5a; see also Sup-porting information, Figure S8). Expression was alsoobserved in the dorsal striatum (caudate nucleus andputamen) and thalamus (Figure 1b, c). Within thehippocampus, nephrin was expressed by some pyra-midal neurons of the CA3 region (Figure 5d; seealso Supporting information, Figure S9). A mild scat-tered immunostaining was found in the CA1 region,whereas the ‘ilo’ was completely negative (see Sup-porting information, Figure S9). In the cerebellum,nephrin was expressed by Purkinje cells and by granulecells of the nuclear layer (Figure 5g; see also Sup-porting information, Figure S9). Notably, a peculiaralternating nephrin pattern was observed in Purkinjecells, with positive cells alternating with negative ones(Figure 2).

To check for different nephrin expression by spe-cific neuronal cell types, we performed a doublestaining with the following: anti-ChAT for choliner-gic neurons, anti-tyrosine hydroxylase for dopamin-ergic neurons and anti-tryptophan hydroxylase forserotoninergic neurons. No prevalent or exclusiveassociation was found between nephrin and theanalysed neuronal populations (Figure 3) and a scat-tered double-labelling was observed for all three celltypes.

Nephrin and the kinase Fyn

Fyn kinase is known to be expressed at the podocyteslit diaphragm. Specificity of the antibody was there-fore tested on renal sections by immunofluorescenceand immunogold electron microscopy (see Supportinginformation, Figure S10).

A western blot analysis was conducted on membranefractions from different brain regions, and from renalglomeruli as positive control. Western blotting ofTX-insoluble pellet and supernatant was performed(Figure 4a) and nephrin was detected in all insolublefractions, not in the supernatant.

Western blot analysis of Fyn was then performedon CNS and kidney membrane fractions, showinga corresponding band of 59 kDa in all samples(Figure 4b).

Immunoprecipitation confirmed that Fyn co-im-munoprecipitates with nephrin, not only in the renalglomerulus but also in CNS protein extracts (Figure 4c,d). Co-localization of nephrin and Fyn in the CNS wasobserved by double staining (Figure 5; see also Sup-porting information, Figure S11a–c) and confirmed byPLA (see Supporting information, Figure S11d, e).



(a)

(d)

(e)

(b)

(c)

Figure 1. Nephrin expression in the brain cortex. Nephrin-positive cells are diffusely detected in (a) pons, (b) caudate nucleus–putamenand (c) thalamus. (d) Negative control, performed by applying control Ig instead of the primary antibody. (e) Positive control conducted byapplying an anti-Grm1 antibody, disclosing positive cells in the substantia nigra. Indirect immunofluorescence. Magnifications: (a, d, e)×200; (b, c) ×400. Small icons next to each figure show the areas of staining (empty circles).

Figure 2. Nephrin pattern in Purkinje cells. Cell bodies of Purkinjecells and axons of granule cells are labelled in the molecular layer ofthe cerebellum (×100). Note that not all Purkinje cell bodies appearnephrin-positive, and positive cell bodies periodically alternate withnephrin-negative cells. The small icon indicates the area of staining.

Nephrin expression in cultured neuronal cellsand podocytesIn mouse cortical neuronal cultures both nephrin andFyn showed a homogeneous distribution within the cellbody and a discontinued punctuated staining along cellprocesses (Figure 6a, b).

Double labelling of neuronal cells with nephrin andsynapsin-1 or synaptophysin revealed these two synap-tic proteins closely connected, but not co-localizingwith nephrin (Figure 6c–f). Interestingly, this occurrednot only in neuronal cells but also in podocytes(Figure 6g).

Immunoprecipitation was then conducted withnephrin, the two glutamate receptors NMDAR1 andGrm1, and the scaffolding molecule PSD95. Our anal-ysis disclosed that in brain protein extracts nephrin

co-immunoprecipitates with glutamate receptors andPSD95 (Figure 7a–c). Co-localization was confirmedby double staining immunofluorescence and PLA(Figure 7d–f).

The same investigations were then conducted onpodocyte cell cultures, glomerular protein extracts andhuman kidney sections, revealing that also in podocytesnephrin co-immunoprecipitates and co-localizes withglutamate receptors and PSD95 (Figure 8; see alsoSupporting information, Figure S12).

Nephrin silencing in Neuro2A cellsThe silencing procedure did not affect cell health, asdemonstrated by the morphology and number of cells72 h after transfection (see Supporting information,Figure S13a, b). Nephrin expression was unaffectedin control cells, but almost completely abolished innephrin siRNA-transfected cells (Figure 9a, b; seealso Supporting information, Figure S13c–e). As aconsequence of nephrin silencing, we observed amarked reduction in the length and complexity of cellramifications, as was clearly apparent by morphologyand actin staining (Figure 9c, d). Similarly, NMDARand PSD95 expression were limited to the cell body(Figure 9e–h). These changes were parallelled byreduction of both spontaneous and LTX-stimulatedglutamate release (Figure 9i, j).

Discussion

Evidence for nephrin expression in the rodent centralnervous system (CNS) has been gathered from investi-gation in embryo and newborn rodents, using reporter


122 M Li et al

(a) (b) (c)

(d) (e) (f)

Figure 3. Nephrin labelling of specific neuron populations. Nephrin does not show a preferential staining for single neuron populations, asidentified by antibodies directed against tyrosine hydroxylase (dopaminergic), ChAT (colinergic) and tryptophan hydroxylase (serotoninergicneurons): double labelling is detectable on a few cell bodies or is represented by small dots along cell processes (arrows). (a–c) Pons:(a) nephrin (green), tyrosine hydroxylase (red), ×200; (b) nephrin (red), ChAT (green), ×400; (c) nephrin (red), tryptophan hydroxylase(green), ×200; (d) substantia nigra: nephrin (red), tryptophan hydroxylase (green), ×400; (e, f) substantia reticularis: (e) nephrin (red),ChAT (green) ×200; (f) nephrin (red), tryptophan hydroxylase (green) (×400).

Triton X100 insoluble fraction

Triton X100 insoluble fraction

IP nephrin - glomeruli IP nephrin - brain

Triton X100insol. fract.

supernatant

WB Nephrin

WB Fyn

WB Fyn

WB Fyn

(a)

(b)

(c) (d)

OB OB

OB

K

K

Cb CbCx

Cx CbP

CxMO

MO

MO

180kDa

60kDa

60kDa

60kDa

MWM MWM1 12 23

Figure 4. Nephrin and Fyn; western blot and co-immunoprecipitation from brain and kidney. (a) Western blot analysis conducted onmembrane extracts and supernatants from CNS areas and on membrane extracts from kidney. Nephrin is present only in material frommembrane-associated fractions. (b) Western blot conducted on CNS and kidney membrane extracts shows positivity for the kinase Fynin all analysed fractions. OB, olfactory bulb; Cb, cerebellum; MO, medulla oblongata; Cx, cortex; P, pons; K, kidney. (c) After nephrinimmunoprecipitation from isolated glomeruli, western blot was conducted with anti-Fyn antibody: lane 1, starting material from the50 µm mesh filter; lane 2, material obtained from the 36 µm mesh filter; lane 3, negative control, performed by loading the bead pelletmaterial obtained after addition of rabbit IgG instead of rabbit anti-nephrin antibody. (d) After immunoprecipitation of nephrin from brainprotein extracts (lane 1), WB was performed with a Fyn antibody: lane 2, negative control (see above). MWM, molecular weight marker.



(a) (b) (c)

(d) (e) (f)

(g) (h) (i)

Figure 5. Nephrin and Fyn immuno-co-localization in the CNS. Double staining of nephrin (green) and Fyn (red) in mouse brain tissueshows (a–c) co-localization (yellow) of the two molecules in the brainstem, (d–f) CA3 region of hippocampus, and (g–i) cerebellum.Magnifications: (a–c) ×200; (d–i) ×400.

(a) (b)

(c) (d)

(e) (f) (g)

Figure 6. Primary neuronal cells and podocyte cultures. The immunostaining clearly displays similar labelling for nephrin (a, ×200) andFyn (b, ×200), both present in the cell body as well as along neuronal cell processes, with a characteristic punctuate pattern of positivity.The double staining demonstrates that synaptophysin (red, c, d) and synapsin I (red, e, f) do not co-localize with nephrin (green, c–f), butperiodically surround nephrin around its positive dots. The same pattern of positivity can be observed in a primary cell process (g) andsecondary processes (g, inset) of a podocyte, where nephrin (green) and synaptophysin (red) closely associate but do not co-localize.Magnifications: (c) ×400; (d) ×600; (e) ×200; (f) ×630; (g) ×1000.

genes under the control of NPHS1 promoters [4,6–8].Although the precise function of nephrin in the CNSremains to be completely clarified, a neurological phe-notype was recently reported for nephrin-KO mice afterselective rescue of nephrin in the kidney [23], support-ing the idea that this Ig-like CAM, at least in rodents,has an important role in the CNS. This is confirmedby the expression in the CNS of other nephrin fam-ily members, viz Neph1, Neph2 and Neph3 [24,25],which have been shown to interact with nephrin bycis- or trans-heterophilic interactions [25].

Our present results add further information, bydescribing the expression and identifying some molec-ular partners of nephrin in adult rodent CNS.

Compared to the expression observed during devel-opment and at birth, in adult CNS nephrin extends tothe pons, with evident and diffuse positivity in thesubstantia nigra and corpus callosum. The protein isinstead reduced in the hippocampus, mostly limited tothe CA3 area. Adult rodents also display a diffuse pres-ence of nephrin in basal ganglia and motor cortex butcomplete negativity of the sensory cortex, suggesting


124 M Li et al

Brain – IP nephrin

MWM 1 2MWM

MWM

1

1

2

2

180kDa

100kDa

80kDa

WB nephrin

WB NMDAR1

WB PSD95

(a)

(d) (e) (f)

(b) (c)

Nephrin-NMDAR Nephrin-PSD95 PSD95-NMDAR

Bra

in ti

ssue

sec

tions

Figure 7. Immunoprecipitation and PLA; brain. After nephrin immunoprecipitation (a–c, lanes 1), western blot was conducted with(a) nephrin, (b) the NMDA1 receptor, and (c) PSD95, disclosing bands of corresponding molecular weight for all three molecules. Lanes 2,negative controls. PLA conducted on brain tissue sections (d, e, f) confirms the co-localization in the motor cortex of the three molecules(IF, ×200).

Glomeruli – IP nephrin

MWM 1 2 3

MWM 1 2 3

MWM 1 2 3

MWM 1 2 3

180kDa

80kDa

100kDa

140kDa

(a) (b)

(c) (d)

(e) (f) (g)

WB nephrin

WB PSD95 WB Grm1

WB NMDAR1

Nephrin-NMDAR PSD95-NMDARNephrin-PSD95

Cul

ture

d po

docy

tes

Figure 8. Immunoprecipitation and PLA—glomeruli and cultured podocytes. After nephrin immunoprecipitation (a–d, lanes 1, 2), westernblot was conducted on glomerular protein extracts with (a) nephrin, (b) the NMDA1 receptor, (c) PSD95, and (d) metabotropic glutamatereceptor 1 (Grm1), disclosing bands of corresponding molecular weight for all molecules: lanes 1, starting material from the 50 µm meshfilter; lanes 2, material obtained from the 36 µm mesh filter; lanes 3, negative controls. (e–g) Co-localization of nephrin, NMDAR1 andPSD95, as demonstrated by PLA on cultured primary podocytes (IF, ×630).

the involvement of nephrin in distinct brain networksrelated to movement, not to sensory processing.

The association of nephrin with movement activitiesis further confirmed by its presence in the cerebellum,and helps to explain the symptoms of nephrin-deficientmice [23], which consist of decreased spontaneousactivity, reduced locomotor activity in novel environ-ments and impaired coordination in the hind legs andtail. In these mice morphological changes of Purkinjecells are also observed, differently from nephrin-KOmice described by Putaala et al [4], displaying normalcerebellar morphology.

In humans, most patients with nephrin mutationsseem not to have neurological symptoms [26], althoughmuscular dystonia and athetosis were reported by

Laakkonen et al [27]. Apart from obvious speciesdifferences, the lack of symptoms in the majority ofhuman subjects can have several explanations, firstof all the possibility that nephrin changes can berescued in the CNS by other CAMs. The same is likelyoccurring in the heart, where nephrin plays a role invessel formation [28]. Actually, it is not uncommon forthe same protein to play non-redundant or redundantroles in different tissues, generating diverse tissue-specific phenotypes.

Interestingly, in the cerebellum we found a char-acteristic pattern of nephrin-positive Purkinje cellsalternating with nephrin-negative Purkinje cells, iden-tical to the expression of aldolase C [29] and 5′-nucleotidase [30], and called ‘stripe-like distribution’.



(a) (b)

(c) (d)

(e) (f)

(i)

(j)(g) (h)

800

750

700

650

600

550

500

OD

340

nm

800

750

700

650

600

550

500

OD

340

nm

T0/LTX T5 T10 15 20 25 30min

T0/LTX T5 T10 15 20 25 30min

Figure 9. Nephrin silencing in Neuro2A cells. Images are obtained from Neuro2A cells transfected with (left panels, a, c, e, g) controlsiRNA and (right panels, b, d, f, h) nephrin siRNA. (a) Nephrin, (c) F-actin, (e) NMDAR1 and (g) PSD95 are positive on the cell body andalong cell processes in control cells. (b) Nephrin is completely negative and the other markers (d, f, h) are positive in the cell body ofnephrin-silenced cells. Magnification, ×630. Spontaneous (i) and α-latrotoxin-stimulated (j) glutamate exocytosis is reduced at all timepoints in nephrin-silenced cells (black triangle) as compared to cells transfected with control siRNA (white rhombi). ∗Significant values, asobtained by t-test (∗p < 0.03; ∗∗p < 0.003).

This longitudinal stripe-shaped expression pattern hasbeen used to better describe the longitudinal compart-mentalization of the cerebellar cortex [31–33] beyondthe conventional division in vermis, pars intermediaand hemisphere. Although the functional significanceof this specific pattern remains to be investigated, it islikely related to the existence of different populationsof Purkinje cells.

Similarly to glomerular nephrin [34], neuronalnephrin is recovered in membrane Triton X-100-insoluble fractions, or lipid rafts, which are special-ized membrane domains enriched in sphingolipids,cholesterol and proteins involved in signal transduction[35]. Previous experiments conducted on glomeruli andpodocytes have clearly demonstrated the involvementof the tyrosine kinase Fyn in nephrin phosphoryla-tion and raft-mediated internalization [36]. Therefore,our co-localization and co-immunoprecipitation experi-ments, by disclosing the same nephrin–Fyn interactionin the CNS, are suggestive of similar functions.

Furthermore, our data show that neuronal nephrinco-localizes and co-immunoprecipitates with glutamatereceptors and with the scaffolding molecule postsynap-tic density protein 95 (PSD-95).

In the CNS, presynaptic and postsynaptic membranesare held together by trans-synaptic interactions betweenCAMs. Among major groups of synaptic CAMs,such as neurexins and neuroligins, ephrins and ephrin

receptors, cadherins and nectin-like molecules, thereare also members of the Ig superfamily, to whichnephrin belongs. CAMs serve to facilitate organizationand adhesion of the synapse, and help to recruit andorganize key components, such as synaptic vesicles atthe presynaptic terminal and neurotransmitter receptorsat the PSD [37]. At the presynaptic terminal, CAMsare in close association, but do not co-localize withsynaptic proteins [38], which is in line with ourfindings.

It has to be remembered that most studies on synapticadhesion focus on the PSD, whereas the precise molec-ular events regulating vesicle positioning by CAMsremain relatively unknown. Absence of N-cadherin isresponsible for impaired clustering of synaptic vesi-cles [39], but additional studies are certainly needed toclarify the signalling pathways regulating presynapticorganization.

Further support for the hypothesis that nephrinbehaves as a synaptic CAM is provided by oursilencing experiments, where nephrin down-regulationimpairs the development of neuronal projections [40].This role for nephrin is not unexpected, becausenephrin orthologues SYG-2 in Caenorhabditis elegansand C-roughest in Drosophila melanogaster are crucialplayers in synapse targeting and positioning [41,42].

The presence of Fyn fits nicely into this scheme;Fyn phosphorylation of the NMDA receptor facilitates


126 M Li et al

its interaction with PSD-95 [43,44] and Fyn-mediatedNMDAR phosphorylation is elevated upon LTP induc-tion, highlighting the importance of Fyn in synapticplasticity and memory processes [45].

PSD-95, a scaffold molecule involved in the assem-bly and organization of signal transduction com-plexes at the PSD, belongs to the membrane-associatedguanylate-kinase family of proteins (MAGUK), whichare modular synaptic adaptors [46]. Like other mem-bers of this family, PSD-95 contains three consecutivePDZ domains, relevant to clustering of ion channelsand synaptic CAMs [46,47].

In the CNS, the narrow extracellular space com-monly referred to as the synaptic cleft is character-ized by electron-dense material that tightly links thepresynaptic and postsynaptic membranes. It is widelyaccepted that the cleft material consists of extracellularmatrix components together with extracellular domainsof presynaptic and postsynaptic CAMs engaged inhomophilic or heterophilic interactions [46]. Notably,these features exactly correspond to the description ofthe slit diaphragm in renal glomeruli, where extra-cellular domains of cadherin and Ig-like proteins areengaged in homophilic and heterophilic interactions[48].

We, and others, have emphasized the strong molec-ular [1,17] and functional [1,2] similarities betweenthe slit diaphragm in the glomerulus and the synapticjunction of neurons. Recent reports have documentedthe expression at the slit diaphragm of the synapticCAMs ephrin B1 [49] and neurexin 1 [50] and haveidentified the podocyte molecule podocalyxin as a neu-ronal CAM [51]. Our present data further strengthenthese analogies by showing that glomerular nephrinco-immunoprecipitates and co-localizes with glutamatereceptors and PSD95.

In summary, our results point to a common neuronaland podocyte function for nephrin, which clustersglutamate receptors, recruits synaptic vesicles and, byinteracting with the scaffolding molecule PSD95, linksglutamate receptors to downstream signalling proteins,such as the kinase Fyn.

It becomes more and more evident that, outsidethe nervous system, glutamate signalling is utilizednot only by podocytes but also by other cells, suchas lymphocytes, testis cells and pancreatic β cells,all of them requiring precise modulation of specificfunctions. Not surprisingly, nephrin expression hasbeen selectively found in the very same locations[52–55], and it will be interesting to investigate itsmolecular interactions also in these systems.

Acknowledgment

The authors wish to thank Dr Pia Irene Anna Rossifor technical support, Mr Guido Brusini for animalcare and Professor Harry Holthofer for providing theanti-nephrin antibody. Funding for this research wasprovided by Fondazione ‘La Nuova Speranza’ Lotta

alla Glomerulosclerosi Focale, Rho, Milan, and Asso-ciazione per il Bambino Nefropatico (ABN ONLUS),Milan.

Author contributions

LM, MP and RMP conceived and designed the study;IM, CA and CN did the tissue immunostaining andanalysed the results; PM, GL, NF and PA carriedout experiments on neuronal cells; LM, AS, MD andAS performed the experiments on podocytes; and RRand GF contributed to the critical revision of themanuscript. All authors were involved in writing thepaper and approved the submitted version.

References

1. Rastaldi MP, Armelloni S, Berra S, et al . Glomerular podocytescontain neuron-like functional synaptic vesicles. FASEB J 2006;20: 976–978.

2. Giardino L, Armelloni S, Corbelli A, et al . Podocyte glutamater-gic signaling contributes to the function of the glomerular filtrationbarrier. J Am Soc Nephrol 2009; 20: 1929–1940.

3. Kobayashi N, Gao SY, Chen J, et al . Process formation of therenal glomerular podocyte: is there common molecular machineryfor processes of podocytes and neurons? Anat Sci Int 2004; 79:1–10.

4. Putaala H, Soininen R, Kilpelainen P, et al . The murine nephringene is specifically expressed in kidney, brain and pancreas:inactivation of the gene leads to massive proteinuria and neonataldeath. Hum Mol Genet 2001; 10: 1–8.

5. Welsh GI, Saleem MA. Nephrin-signature molecule of theglomerular podocyte? J Pathol 2010; 220: 328–337.

6. Putaala H, Sainio K, Sariola H, et al . Primary structure of mouseand rat nephrin cDNA and structure and expression of the mousegene. J Am Soc Nephrol 2000; 11: 991–1001.

7. Moeller MJ, Kovari IA, Holzman LB. Evaluation of a new toolfor exploring podocyte biology: mouse Nphs1 5′ flanking regiondrives LacZ expression in podocytes. J Am Soc Nephrol 2000; 11:2306–2314.

8. Beltcheva O, Kontusaari S, Fetissov S, et al . Alternatively usedpromoters and distinct elements direct tissue specific expression ofnephrin. J Am Soc Nephrol 2003; 14: 352–358.

9. Huber TB, Benzing T. The slit diaphragm: a signaling platform toregulate podocyte function. Curr Opin Nephrol Hypertens 2005;14: 211–216.

10. Harita Y, Kurihara H, Kosako H, et al . Phosphorylation of nephrintriggers Ca2+ signaling by recruitment and activation of phospho-lipase C-γ1. J Biol Chem 2009; 284: 8951–8962.

11. Verma R, Wharram B, Kovari I, et al . Fyn binds to and phospho-rylates the kidney slit diaphragm component nephrin. J Biol Chem

2003; 278: 20716–20723.12. Grant SG, O’Dell TJ, Karl KA, et al . Impaired long-term potenti-

ation, spatial learning, and hippocampal development in fyn mutantmice. Science 1992; 258: 1903–1910.

13. Kramer EM, Klein C, Koch T, et al . Compartmentation of Fynkinase with glycosylphosphatidylinositol-anchored molecules inoligodendrocytes facilitates kinase activation during myelination.J Biol Chem 1999; 274: 29042–29049.

14. Mundel P, Kriz W. Structure and function of podocytes: an update.Anat Embryol 1995; 192: 385–397.



15. Kennedy MB. The postsynaptic density. Curr Opin Neurobiol

1993; 3: 732–737.16. Kim E, Sheng M. PDZ domain proteins of synapses. Nat Rev

Neurosci 2004; 5: 771–781.17. Faul C, Asanuma K, Yanagida-Asanuma E, et al . Actin up: reg-

ulation of podocyte structure and function by components of theactin cytoskeleton. Trends Cell Biol 2007; 17: 428–437.

18. Miller TM, Johnson EM Jr. Metabolic and genetic analyses ofapoptosis in potassium/serum-deprived rat cerebellar granule cells.J Neurosci 1996; 16: 7487–7495.

19. Ma’ayan A, Jenkins SL, Barash A, et al . Neuro2A differentiationby Gαi/o pathway. Sci Signal 2009; 2: cm1.

20. Soderberg O, Leuchowius KJ, Gullberg M, et al . Characterizingproteins and their interactions in cells and tissues using the in situproximity ligation assay. Methods 2008; 45: 227–232.

21. Kremerskothen J, Plaas C, Kindler S, et al . Synaptopodin, amolecule involved in the formation of the dendritic spine appa-ratus, is a dual actin/α-actinin binding protein. J Neurochem 2005;92: 597–606.

22. Bezzi P, Carmignoto G, Pasti L, et al . Prostaglandins stimulatecalcium-dependent glutamate release in astrocytes. Nature 1998;391: 281–285.

23. Juhila J, Lassila M, Roozendaal R, et al . Inducible nephrin trans-gene expression in podocytes rescues nephrin-deficient mice fromperinatal death. Am J Pathol 2010; 176: 51–63.

24. Gerke P, Benzing T, Hohne M, et al . Neuronal expression andinteraction with the synaptic protein CASK suggest a role forNeph1 and Neph2 in synaptogenesis. J Comp Neurol 2006; 498:466–475.

25. Nishida K, Hoshino M, Kawaguchi Y, et al . Ptf1a directly controlsexpression of immunoglobulin superfamily molecules Nephrin andNeph3 in the developing central nervous system. J Biol Chem 2010;285: 373–380.

26. Patrakka J, Kestila M, Wartiovaara J, et al . Congenital nephroticsyndrome (NPHS1): features resulting from different mutations inFinnish patients. Kidney Int 2000; 58: 972–980.

27. Laakkonen H, Lonnqvist T, Uusimaa J, et al . Muscular dystoniaand athetosis in six patients with congenital nephrotic syndrome ofthe Finnish type (NPHS1). Pediatr Nephrol 2006; 21: 182–189.

28. Wagner N, Morrison H, Pagnotta S, et al . The podocyte proteinnephrin is required for cardiac vessel formation. Hum Mol Genet

2011; 20: 2182–2194.29. Anh AH, Dziennis S, Hawkes R, et al . The cloning of zebrin

II reveals its identity with aldolase C. Development 1994; 120:2081–2090.

30. Eisenman LM, Hawkes R. 5′-Nucleotidase and the mabQ113 anti-gen share a common distribution in the cerebellar cortex of themouse. Neuroscience 1989; 31: 231–235.

31. Hawkes R, Leclerc N. Antigenic map of the rat cerebellar cortex:the distribution of parasagittal bands as revealed by monoclonalanti-Purkinje cell antibody mabQ113. J Comp Neurol 1987; 256:29–41.

32. Brochu G, Maler L, Hawkes R. Zebrin II: a polypeptide antigenexpressed selectively by Purkinje cells reveals compartments in ratand fish cerebellum. J Comp Neurol 1990; 291: 538–552.

33. Voogd J, Pardoe J, Ruigrok TJH, et al . The distribution of climb-ing and mossy fiber collateral branches from the copula pyramidisand the paramedian lobule: congruence of climbing fiber corticalzones and the pattern of zebrin banding within the rat cerebellum.J Neurosci 2003; 23: 4645–4656.

34. Simons M, Schwarz K, Kriz W, et al . Involvement of lipid raftsin nephrin phosphorylation and organization of the glomerular slitdiaphragm. Am J Pathol 2001; 159: 1069–1077.

35. Lingwood D, Simons K. Lipid rafts as a membrane-organizingprinciple. Science 2010; 327: 46–50.

36. Qin XS, Tsukaguchi H, Shono A, et al . Phosphorylation ofnephrin triggers its internalization by raft-mediated endocytosis.J Am Soc Nephrol 2009; 20: 2534–2545.

37. Yamagata M, Sanes JR, Weiner JA. Synaptic adhesion molecules.Curr Opin Cell Biol 2003; 15: 621–632.

38. Song JY, Ichtchenko K, Sudhof TC, et al . Neuroligin 1 is a post-synaptic adhesion molecule of excitatory synapses. Proc Natl Acad

Sci USA 1999; 96: 1100–1105.39. Stan A, Pielarski KN, Brigadski T, et al . Essential cooperation of

N-cadherin and neuroligin-1 in the transsynaptic control of vesicleaccumulation. Proc Natl Acad Sci USA 2010; 107: 11116–11121.

40. Kiryushko D, Berezin V, Bock E. Regulators of neurite outgrowth:role of cell adhesion molecules. Ann NY Acad Sci 2004; 1014:140–154.

41. Shen K, Fetter RD, Bargmann CI. Synaptic specificity is generatedby the synaptic guidepost protein SYG-2 and its receptor, SYG-1.Cell 2004; 116: 869–881.

42. Ramos RG, Igloi GL, Lichte B, et al . The irregular chiasm C-roughest locus of Drosophila, which affects axonal projectionsand programmed cell death, encodes a novel immunoglobulin-likeprotein. Genes Dev 1993; 7: 2533–2547.

43. Nakazawa T, Komai S, Tezuka T, et al . Characterization of Fyn-mediated tyrosine phosphorylation sites on GluR ε2 (NR2B)subunit of the N-methyl-D-aspartate receptor. J Biol Chem 2001;276: 693–699.

44. Tezuka T, Umemori H, Akiyama T, et al . PSD-95 promotes Fyn-mediated tyrosine phosphorylation of the N-methyl-D-aspartatereceptor subunit NR2A. Proc Natl Acad Sci USA 1999; 96:435–440.

45. Isosaka T, Hattori K, Kida S, et al . Activation of Fyn tyrosinekinase in the mouse dorsal hippocampus is essential for contextualfear conditioning. Eur J Neurosci 2008; 28: 973–981.

46. Han K, Kim E. Synaptic adhesion molecules and PSD-95. Progr

Neurobiol 2008; 84: 263–283.47. Kornau HC, Schenker LT, Kennedy MB, et al . Domain interaction

between NMDA receptor subunits and the postsynaptic densityprotein PSD-95. Science 1995; 269: 1737–1740.

48. Holthofer H. Molecular architecture of the glomerular slitdiaphragm: lessons learnt for a better understanding of diseasepathogenesis. Nephrol Dial Transpl 2007; 22: 2124–2128.

49. Hashimoto T, Karasawa T, Saito A, et al . Ephrin-B1 localizes atthe slit diaphragm of the glomerular podocyte. Kidney Int 2007;72: 954–964.

50. Saito A, Miyauchi N, Hashimoto T, et al . Neurexin 1, a presy-naptic adhesion molecule, localizes at the slit diaphragm of theglomerular podocytes in kidneys. Am J Physiol Regul Integr CompPhysiol 2011; 300: R340–R348.

51. Vitureira N, Andres R, Perez-Martınez E, et al . Podocalyxin is anovel polysialylated neural adhesion protein with multiple roles inneural development and synapse formation. PLoS One 2010; 5:e12003.

52. Astrom E, Rinta-Valkama J, Gylling M, et al . Nephrin in humanlymphoid tissues. Cell Mol Life Sci 2006; 63: 498–504.

53. Liu L, Aya K, Tanaka H, et al . Nephrin is an important componentof the barrier system in the testis. Acta Med Okayama 2001; 55:161–165.

54. Palmen T, Ahola H, Palgi J, et al . Nephrin is expressed in thepancreatic β cells. Diabetologia 2001; 44: 1274–1280.

55. Fornoni A, Jeon J, Varona Santos J, et al . Nephrin is expressedon the surface of insulin vesicles and facilitates glucose-stimulatedinsulin release. Diabetes 2010; 59: 190–199.


128 M Li et al

SUPPORTING INFORMATION ON THE INTERNET

The following supporting information may be found in the online version of this article:

Detailed methods

Table S1. Primary antibodies used for the study.

Figure S1. Schematic representations and corresponding PAS staining of CNS areas in a sagittal section corresponding to some of thesections of rodent tissue used along this study.

Figure S2. Images representing some of the sections of rodent tissue used along this study.

Figure S3. Brightfield images of Neuro2A cells taken after 1 week incubation in propagating medium and 1 week incubation indifferentiating medium.

Figure S4. Proximity ligation assay (PLA)-positive and -negative controls.

Figure S5. RT–PCR conducted on 3 month-old rat kidney and brain mRNA, demonstrating the presence in both organs of the Ig-likedomain and the fibronectin domain of nephrin.

Figure S6. Immunostaining of rat renal sections with nephrin antibody shows positivity selectively present on the glomerular tuft and ona renal nerve.

Figure S7. Nephrin expression in the brain cortex.

Figure S8. Nephrin expression in the olfactory bulb, pons and medulla oblongata.

Figure S9. Nephrin expression in the hippocampus and cerebellum.

Figure S10. Immunogold electron microscopy conducted on a rat glomerular section, proving the specificity of Fyn antibody.

Figure S11. Nephrin and Fyn co-localization in the CNS.

Figure S12. Double staining conducted on primary podocyte cell cultures human control glomeruli, showing co-localization of nephrin andthe NMDA1 receptor.

Figure S13. Efficiency of transfection conducted on Neuro2A cells, as demonstrated by cell health and decrease of nephrin expression.


nephrin expression in adult rodent central nervous system and its … · 2011. 11. 16. · nephrin...

Documents