Write a critique on the Alzheimer’s disease paper (between 750 to 900 words excluding the title page) I have to read the paper

ARTICLES

APP binds DR6 to trigger axon pruning and neuron death via distinct caspases Anatoly Nikolaev

1 , Todd McLaughlin

2 , Dennis D. M. O’Leary

2 & Marc Tessier-Lavigne

1

Naturally occurring axonal pruning and neuronal cell death help to sculpt neuronal connections during development, but their mechanistic basis remains poorly understood. Here we report that b-amyloid precursor protein (APP) and death receptor 6 (DR6, also known as TNFRSF21) activate a widespread caspase-dependent self-destruction program. DR6 is broadly expressed by developing neurons, and is required for normal cell body death and axonal pruning both in vivo and after trophic-factor deprivation in vitro. Unlike neuronal cell body apoptosis, which requires caspase 3, we show that axonal degeneration requires caspase 6, which is activated in a punctate pattern that parallels the pattern of axonal fragmentation. DR6 is activated locally by an inactive surface ligand(s) that is released in an active form after trophic-factor deprivation, and we identify APP as a DR6 ligand. Trophic-factor deprivation triggers the shedding of surface APP in a b-secretase (BACE)-dependent manner. Loss- and gain-of-function studies support a model in which a cleaved amino-terminal fragment of APP (N-APP) binds DR6 and triggers degeneration. Genetic support is provided by a common neuromuscular junction phenotype in mutant mice. Our results indicate that APP and DR6 are components of a neuronal self-destruction pathway, and suggest that an extracellular fragment of APP, acting via DR6 and caspase 6, contributes to Alzheimer’s disease.

The initial formative phase of nervous system development, invol- ving the generation of neurons and extension of axons, is followed by a regressive phase in which inappropriate axonal branches are pruned to refine connections, and many neurons are culled to match the numbers of neurons and target cells1–3. The loss of neurons and branches also occurs in the adult after injury, and underlies the pathophysiology of many neurodegenerative diseases1,4.

Our understanding of regressive events in development remains fragmentary. Degeneration can result ‘passively’ from the loss of support from trophic factors such as nerve growth factor (NGF)1–3.

There is also evidence for ‘active’ mechanisms in which extrinsic signals trigger degeneration by means of pro-apoptotic receptors, including some members of the tumour necrosis factor (TNF) recep- tor superfamily such as p75NTR (also known as NGFR), Fas and TNFRSF1A (previously known as TNFR1) (Fig. 1a)5. However, the full complement of degeneration triggers remains incompletely understood.

Our understanding of the intracellular mechanisms of neuronal dismantling is also incomplete. It is well documented that devel- opmental neuronal cell body degeneration requires the apoptotic

1Division of Research, Genentech, Inc., 1 DNA Way, South San Francisco, California 94080, USA. 2Molecular Neurobiology Laboratory, The Salk Institute, 10010 North Torrey Pines Road, La Jolla, California 92037, USA.

0%

1%

a E10.5 E11.5 E12.5

P P PVC

S

M M M

V C

S

V

C

S

b

e d 100 80 60 40 20

NGF-deprived

Sensory axons

NGF-deprived + anti-DR6.1

D e g

e n e ra

ti n

g a

xo n b

u n d

le s

(% )

+ NGF

100 80 60 40 20

TF-deprived

Motor axons

TF-deprived + anti-DR6.1

1.5%

+ TFs

100 80 60 40 20

48 h, control

Commissural axons

48 h, + anti-DR6.1

24 h, control

+ N

G F

N

G F

-d e p

ri ve

d

Sensory, TuJ1

+ A

n ti -D

R 6 .1

+

I g

G

+ T

F s

+ I g

G

+ A

n ti -D

R 6 .1

Motor, p75NTR

T F

-d e p

ri ve

d

f

2 4 h

GFP TUNEL

4 8 h

+

A n ti -D

R 6 .1

+

I g

G

+ I g

G

Commissural

c

? Neuro- trophins FasL

Fasp75NTRDR6

TNF

TNFRSF1A

0

0

0

Figure 1 | DR6 regulates degeneration of several neuronal classes. a, Diagram of several TNF receptor superfamily members possessing death domains. b, DR6 mRNA is expressed by differentiating spinal neurons (including motor (M) and commissural (C)), and by sensory (S) neurons in DRG at E10.5–E12.5. Expression is low in neuronal progenitors (P) in the ventricular zone (V). c–f, Anti-DR6.1 (50 mg ml

21 ) reduces

degeneration in vitro. c, Anti-DR6.1 inhibits commissural axon degeneration (visualized with green fluorescent protein (GFP), right) and cell body death (TUNEL labelling, left; dots indicate explants) seen after 48 h in dorsal spinal cord cultures. Red arrow indicates degenerating axon. d, e, Effect on degeneration of sensory (d) or motor (e) axons triggered by trophic deprivation. TFs, trophic factors (BDNF and NT3). Axons were visualized by immunostaining for tubulin (TuJ1; sensory) or p75NTR (motor). f, The percentage of degenerating axon bundles in c–e (mean and s.e.m., n 5 3 replicates). Scale bars, 110 mm (b, c) and 50 mm (d, e).

Vol 457 | 19 February 2009 | doi:10.1038/nature07767

981 Macmillan Publishers Limited. All rights reserved©2009

effectors BAX and caspase 3 (refs 6–8); pruning of a particular dend- rite in Drosophila is also caspase-dependent9,10. Developmental axo- nal degeneration similarly has many hallmarks of apoptosis— including blebbing, fragmentation, and phagocytic clearing of debris by neighbouring cells2,4. However, it has been argued that axonal degeneration is caspase-independent, because caspase 3 inhibitors block cell body but not axonal degeneration8 (reflecting higher activation of caspase 3 in cell bodies compared to axons11), and because genetic manipulations to inhibit apoptosis did not block axonal degeneration in some models12,13. These results indicated the existence of a caspase-independent program of axonal degenera- tion1,2,4, but its molecular nature has remained elusive.

While studying the expression of all TNF receptor superfamily members14, we found that DR6—one of eight members possessing a cytoplasmic death domain (Fig. 1a)—is widely expressed by neurons as they differentiate and enter a pro-apoptotic state. DR6 is an orphan receptor15. In transfected cells, it triggers cell death in a Jun N-terminal kinase-dependent manner16. In vivo, it regulates lymphocyte develop- ment17,18, but its involvement in neural development is unknown.

Here we show that DR6 links passive and active degeneration mechanisms. After trophic deprivation, DR6 triggers neuronal cell body and axon degeneration. Because DR6 signals via BAX and cas- pase 3 in cell bodies, we revisited the involvement of caspases in axonal degeneration, and found that axonal degeneration indeed requires both BAX and a distinct effector, caspase 6. Our results also indicated that DR6 is activated by a prodegenerative ligand(s) that is surface-tethered but released in an active form after trophic depriva- tion. In searching for candidate ligands with these properties, we considered APP, a transmembrane protein that undergoes regulated shedding and is causally implicated in Alzheimer’s disease19–22, because we had previously found it to be highly expressed by devel- oping neurons and especially axons (see later). Because Alzheimer’s disease is marked by neuronal and axonal degeneration, we had long wondered whether APP participates in developmental degeneration. We show that an extracellular fragment of APP is indeed a ligand for DR6—as is a fragment of its close relative APLP2—that triggers degeneration of cell bodies via caspase 3 and axons via caspase 6, and we propose that this developmental mechanism is hijacked in Alzheimer’s disease.

DR6 regulates neuronal death

To explore the involvement of the TNF receptor superfamily in neural development, we screened its 28 members by in situ hybrid- ization in midgestation mouse embryos. We came to focus on DR6 (Fig. 1a), because its messenger RNA is expressed at low levels in proliferating progenitors in the spinal cord, but is highly expressed by differentiating neurons within the spinal cord and adjacent dorsal root ganglia (DRG) (Fig. 1b).

Because DR6-expressing neurons are becoming dependent for sur- vival on trophic support at these stages, we examined whether DR6 regulates neuronal death after trophic-factor deprivation in vitro, focusing on three sets of spinal neurons: commissural, motor and sensory (Supplementary Fig. 1a). Initially, we found that short inter- fering RNA (siRNA) knockdown of DR6 protected commissural neurons from degeneration (Supplementary Fig. 2). This prompted us to screen monoclonal antibodies to DR6 for their ability to mimic this protection; we selected antibody 3F4 (anti-DR6.1). When embryonic day (E)13.5 rat dorsal spinal cord explants are cultured for 24 h, commissural cell bodies and axons are healthy, but they degenerate if cultured for 24 h longer 23; anti-DR6.1 inhibited this degeneration (Fig. 1c, f), mimicking DR6 knockdown. Anti-DR6.1 also protected sensory neurons from E12.5 mouse DRGs cultured for 48 h with NGF, and motor neurons from E12.5 mouse ventral spinal cord explants cultured for 24 h with brain-derived neurotrophic fac- tor (BDNF) and neurotrophin 3 (NTF3, also known as NT3): when these cultures were deprived of trophic factor and cultured for 24 h longer, they showed massive cell death and axonal degeneration,

which were largely inhibited by anti-DR6.1 (Fig. 1d–f and Supplementary Fig. 1b). Similar protection was observed when DRGs or ventral explants from a DR6 null mutant17 were deprived in the absence of anti-DR6.1 (Supplementary Fig. 4b and data not shown), confirming that anti-DR6.1 is function-blocking. DR6 inhibition (by antibody, siRNA or genetic deletion) caused a delay rather than a complete block, because more degeneration was observed in each case 24–48 h later (Fig. 2b, Supplementary Fig. 4b and data not shown). Consistent with a delay, there was a higher motor-neuron number at E14.5 in the DR6 mutant, but this returned to the wild-type level by E18 (Supplementary Fig. 3), after the cell death period. Thus, antagonizing DR6 delays the death of several neuronal populations in vitro and in vivo.

DR6 regulates axonal pruning

DR6 protein is expressed not just by cell bodies (data not shown) but also by axons (Supplementary Fig. 4a). Protection of axons by DR6

a

b + NGF 48 h

C o

n tr

o l I

g G

A

n ti -D

R 6 .1

NGF-deprived

12 h 24 h

1%

D e g

e n

e ra

ti n

g a

xo n

b u

n d

le s

(% )

NGF-deprived + control IgG NGF-deprived + anti-DR6.1

+ NGF

48 h24 h

100

80

60

40

20

c

d

e

f

Scratch in substratum Teflon divider

Neurites NeuritesCell bodies

WT

WT

DR6–/–

DR6–/–

0

Figure 2 | DR6 regulates axon pruning in vitro and in vivo. a, Diagram of Campenot chamber (adapted from ref. 24). b, Images show the local degeneration of sensory axons (TuJ1 immunostain) in Campenot chambers after NGF deprivation from the axonal compartment (top) was delayed by anti-DR6.1 (50 mg ml

21 ) added at the time of deprivation (bottom). The

graph shows the percentage of degenerating bundles at 24 and 48 h (mean and s.e.m., n 5 3 replicates). c–f, Compromised pruning of retinal axons in DR62/2 mice.Dorsalviewof (c, e),andvibratomesectionsthrough(d, f),the superior colliculus of wild-type (WT; c, d) or DR62/2 (e, f) mice at P6. In wild-type mice (c, d), DiI-labelled temporal RGC axons form a dense termination zone (TZ) in anterior superior colliculus (arrowheads denote the anterior border). Few are outside the immediate termination zone area (arrows). In DR6

2/2 mice (e, f), temporal RGC axons and arbors are present

in areas far from the termination zone (inset, magnified in e, right) and well posterior to it (arrows). L, lateral; M, medial; P, posterior. Scale bars, 100 mm (b), 400 mm (c, e, left), 170 mm (e, right) and 250 mm (d, f).

ARTICLES NATURE | Vol 457 | 19 February 2009

982 Macmillan Publishers Limited. All rights reserved©2009

inhibition might therefore reflect a direct role for DR6 in axons. To explore this, we used compartmented (‘Campenot’) chambers24

(Fig. 2a). Sensory neurons are placed in a central chamber containing NGF; their axons grow under a partition into NGF-containing side- chambers. Fluid exchange between the chambers is limited, so NGF deprivation in a side-chamber elicits local axon degeneration while sparing cell bodies24. Locally deprived axons degenerate in a stereo- typical manner with initial signs by 6 h and extensive degeneration by 12–24 h, but when anti-DR6.1 was added to the deprived side- chamber, degeneration was blocked at 24 h and still largely impaired at 48 h (Fig. 2b). A similar delay was observed when axons of neurons from DR6 knockout mice were locally deprived, but in the absence of anti-DR6.1 (Supplementary Fig. 4b, c). Thus, DR6 function is required in axons for degeneration.

To determine whether DR6 functions in axonal pruning in vivo, we studied the well-characterized retino-collicular projection, which develops from an initially exuberant projection of retinal ganglion cell (RGC) axons to a focused termination zone in the superior colliculus. All temporal RGC axons initially extend into posterior superior colliculus, well past their future termination zone in anterior superior colliculus (Supplementary Fig. 5a). This diffuse projection is then refined by selective degeneration of inappropriate axon seg- ments2, such that by postnatal day (P)6 in wild-type mice few axon segments persist in areas well beyond the termination zone, as revealed by focal injection of the lipophilic dye DiI into temporal retina (Fig. 2c, d and Supplementary Fig. 5a, b). In contrast, in P6 DR6 mutant mice, many more RGC axons and arbors are present in areas far from the termination zone (Fig. 2e, f and Supplementary Fig. 5c, d): we found an 83% increase in axon-positive domains more than 400 mm from the termination zone (Supplementary Fig. 5f) in DR6

2/2 (n 5 7) compared to wild-type mice (n 5 7; P , 0.05,

Student’s t-test). The defect at P6 represents a delay in pruning, not a complete block, as assessed by examining labelled axons at P4, P5, P6 and P9: at each age, the mutant has more extraneous axons than the wild-type, and fewer are observed in both wild-type and mutant at each age compared to earlier time points, but by P9 the mutant and wild-type projections are indistinguishable (data not shown). Thus, blocking DR6 function delays pruning of sensory axons in vitro and retinocollicular axons in vivo.

Caspase 6 regulates axonal degeneration

Because DR6 regulates both cell body apoptosis and axonal degen- eration, we revisited whether an apoptotic pathway is also involved in axons. In support, we found that BAX, an effector in the intrinsic apoptotic pathway, is required in axons, because local sensory axon degeneration in Campenot chambers was blocked by the genetic deletion of Bax (Fig. 3a) or by local addition of a BAX inhibitor (for example, Supplementary Fig. 10b). Consistent with evidence that caspase 3 mediates cell body but not axon degeneration8,11, we found that procaspase 3 is highly enriched in cell bodies, and that zDEVD-fmk, an inhibitor of effector caspases 3 and 7, blocked cell body but not axon degeneration (Fig. 3b, c and Supplementary Fig. 6a–c). There is, however, a third effector caspase, caspase 6. We found that procaspase 6 is expressed in both cell bodies and axons, and that the caspase 6 inhibitor zVEID-fmk blocked degeneration of sensory, motor and commissural axons (Fig. 3b, c and Supplementary Fig. 6a–c), suggesting that caspase 6 regulates axonal degeneration. We verified these results using RNA interference in sensory and commis- sural neurons: Casp3 knockdown protected cell bodies significantly but had only a minor protective effect on axons, whereas Casp6 knockdown protected axons significantly with only minor effect on cell bodies (Fig. 3d, e). Thus, distinct caspases mediate cell body and axon degeneration.

To visualize caspase activation, we first used the fluorescent repor- ters FAM-DEVD-fmk (for caspase 3 and 7) and FAM-VEID-fmk (for caspase 6), which bind covalently to activated target caspases. In NGF-deprived sensory neurons, the caspase 3/7 reporter labelled cell

bodies but not axons, consistent with a previous study11; in contrast, caspase 6 reporter labelling was observed in both cell bodies and axons, and axonal labelling occurred in regularly spaced ‘puncta’, giving a beads-on-a-string appearance (Supplementary Fig. 6f). To control for reporter specificity, we used a selective antibody to cleaved caspase 6 and observed a similar punctate pattern in axons (Fig. 3f, g), whereas antibodies to cleaved caspase 3 only label cell bodies11. Caspase 6 activation was confirmed biochemically (Supplementary Fig. 6e). Interestingly, caspase 6 activation appeared at sites of microtubule fragmentation (assessed by the loss of tubulin immunoreactivity) (Fig. 3f, g), suggesting that caspase 6 activation drives microtubule destabilization. Punctate caspase 6 activation was markedly reduced by anti-DR6.1 (Fig. 3f) and abolished in Bax

2/2

neurons (not shown), suggesting that caspase 6 acts downstream of BAX in the pathway triggered by DR6. However, the possibility of feedback loops in apoptotic pathways makes this interpretation tent- ative.

Regulated shedding of a DR6 ligand(s)

As DR6 is a receptor-like protein, we addressed whether it is activated by a ligand(s). If so, the DR6 ectodomain might be capable of binding the ligand(s) and blocking its action (Fig. 4a). Consistent with this, the DR6 ectodomain fused to human Fc (DR6–Fc) mimicked anti- DR6.1 in delaying degeneration (Fig. 4a–c and Supplementary Figs 7a and 13). To search for DR6 binding sites on axons and in conditioned medium, we used the DR6 ectodomain fused to alkaline phosphatase (DR6–AP). Purple alkaline phosphatase reaction product was observed on sensory and motor axons cultured with trophic factors when they were pre-incubated with DR6–AP but not with alkaline phosphatase alone, but binding was markedly reduced after trophic deprivation (Supplementary Fig. 7b, c). To control for the loss of axonal membrane, we blocked degeneration using a BAX inhibitor (data not shown) or using neurons from Bax

2/2 mice (Fig. 4d) and

observed an even greater reduction in DR6–AP binding (residual binding seen without BAX inhibition might reflect nonspecific bind- ing to degenerating axons). To determine whether DR6-binding sites were shed, we collected medium conditioned by sensory axons (in Campenot chambers) or motor neurons (in explant culture) (a BAX inhibitor was added to prevent nonspecific release resulting from degeneration). Proteins were separated on non-reducing gels, blotted to nitrocellulose, and probed with DR6–AP. Little signal was seen in medium conditioned by either neuronal type in the presence of trophic factor. However, 48 h after trophic deprivation, DR6–AP bound a prominent band around ,35 kDa and a minor band around ,100 kDa in both cultures (Fig. 4e). Together, these results support a ‘ligand activation’ model in which a prodegenerative DR6 ligand(s) is present on the neuronal surface and inactive, but is shed into med- ium in an active form after trophic deprivation (Fig. 4f), allowing it to bind and activate DR6.

N-APP is a regulated DR6 ligand

Several properties of APP made it a candidate for a DR6 ligand: (1) it is highly expressed by developing spinal and sensory neurons and their axons (Fig. 4g), (2) its ectodomain can be shed in a regulated fashion19,20, and (3) it is tied to degeneration through its links to Alzheimer’s disease19–22. In an initial experiment, we found that DR6–AP bound APP expressed in COS-1 cells (Supplementary Fig. 8a). This prompted us to test whether the bands detected by DR6–AP in conditioned medium (Fig. 4e) represent APP ectodomain frag- ments. APP is cleaved by a- or b-secretases (including, in neurons, BACE1; ref. 25) at distinct sites in its juxtamembrane region (Fig. 4h) to release ,100-kDa ectodomain fragments termed sAPPa or sAPPb, respectively19,20. We probed conditioned medium with a polyclonal antibody to the APP N terminus (anti-N-APP(poly), which also binds the APP relative APLP2; see later) and an antibody selective for the carboxy-terminal epitope of sAPPb exposed by BACE cleav- age (anti-sAPPb) (Fig. 4h). Notably, anti-N-APP(poly) detected

NATURE | Vol 457 | 19 February 2009 ARTICLES

983 Macmillan Publishers Limited. All rights reserved©2009

similar bands to those detected by DR6–AP: a major band at ,35 kDa and a minor band at ,100 kDa, both highly enriched after trophic deprivation (Fig. 4i); anti-sAPPb detected a minor ,100-kDa band and a major ,55-kDa band (Fig. 4i), also both enriched after trophic deprivation. These results indicate that trophic deprivation triggers BACE cleavage of APP to yield the ,100-kDa sAPPb (detected by both antibodies), which undergoes a further cleavage(s) to yield a ,55-kDa C-terminal fragment (detected by anti-sAPPb) and a ,35- kDa N-terminal fragment (detected by anti-N-APP(poly)) that we term N-APP. The site of additional cleavage(s) is unknown, but on the basis of fragment sizes it is expected to be around the junction between the APP ‘acidic’ and ‘E2’ domains (amino acid 286); indeed, recombinant APP(1–286) ran at ,35 kDa and was detected with anti-N-APP(poly) (Fig. 4j), similar to N-APP.

Supporting cleavage of APP by BACE, we found that APP express- ion on the surface of cultured sensory and motor axons, as assessed with anti-N-APP(poly) and with antibody 4G8 to the APP juxta- membrane region (Fig. 4h), is high in the presence of trophic factor

but lost after trophic deprivation; the surface loss was blocked by three structurally divergent BACE inhibitors—OM99-2, BACE inhibitor IV, and the highly selective AZ29 (ref. 26) but not the a-secretase inhibitor TAPI (Fig. 4k, Supplementary Figs 9a–c and 10a, and data not shown). Interestingly, 4G8 partially inhibited sur- face loss (Supplementary Fig. 9d), presumably through the steric hindrance of BACE. Loss of surface APP occurred progressively and in ‘patches’, with little lost at 3 h, more at 6–12 h, and most lost by 24 h (Fig. 4k, Supplementary Fig. 10b and data not shown). Total APP visualized after permeabilization did not change detectably (Supplementary Fig. 10b). Surface loss was not affected by BAX or caspase 6 inhibitors, or in neurons from Bax

2/2 mice (Fig. 4k and

Supplementary Figs 9c and 10c). The marked similarly of bands detected by anti-N-APP(poly) and

DR6–AP suggested that DR6 binds N-APP. Indeed, depletion of conditioned medium with anti-N-APP(poly) eliminated DR6–AP binding sites (Fig. 4i), and purified DR6–Fc bound to purified recombinant APP(1–286) in pull-down (Fig. 4j) and enzyme-linked

a

Cleaved caspase 6 TuJ1/Cleaved caspase 6

+ N

G F

N

G F

-d e p

ri ve

d +

c tr

l I g

G

N G

F -d

e p

ri ve

d +

a n ti -D

R 6 .1

NGF-deprived, 24 h

+ Caspase 6 inhibitor + Caspase 3 inhibitor

T u J1

d

Procaspase 6 TuJ1 Procaspase 3 TuJ1

T u J1

/T U

N E

L

+ NGF

+ Casp6 siRNA1 + Casp3 siRNA1

NGF-deprived NGF-deprived

D e g

e n e ra

ti n g

a xo

n b

u n d

le s

(% )

+ NGF

+ anti-NGF

+ CASP3_si1

+ CASP3_si2

+ CASP3_si3

+ CASP6_si1

+ CASP6_si2

+ CASP6_si3

Casp3 siRNAs

Casp6 siRNAs

Casp3 siRNAs

Casp6 siRNAs

+ NGF + Anti-NGF + Casp3 siRNA1 + Casp3 siRNA2 + Casp3 siRNA3 + Casp6 siRNA1 + Casp6 siRNA2 + Casp6 siRNA3

S u rv

iv in

g

D R

G n

e u ro

n s

(% )

NGF-deprived, 48 h

B ax

+ /+

B ax

–/ –

+ NGF b

c

f

ge

NGF-deprived

NGF-deprived, + caspase 3i

+ NGF

NGF-deprived, + caspase 6i

1%

D e g

e n

e ra

ti n

g a

xo n

b u

n d

le s

(% )

100

80

60

40

20

0

NGF-deprived, 12 h, confocal

T u J1

/c le

a ve

d c

a sp

a se

6

0 10 20 30 40 50 60 70 80 90

100

0 10 20 30 40 50 60 70 80 90

100

Figure 3 | BAX and caspase 6 regulate axonal degeneration. a, Local sensory axon degeneration (TuJ1 immunostain) 48 h after NGF deprivation in Campenot chambers was blocked in neurons from Bax

2/2 mice.

b, Dissociated sensory neurons double-labelled for procaspase 3 and TuJ1 (left), or procaspase 6 and TuJ1 (right). Caspase 3 is detected in cell bodies (arrowheads), whereas caspase 6 is seen in both cell bodies and axons. c, Local degeneration of sensory axons in Campenot chambers deprived of NGF for 24 h is inhibited by a caspase 6 inhibitor (zVEID-FMK; caspase 6i), but not by a caspase 3/7 inhibitor (zDEVD-FMK; caspase 3i). Quantification is shown to the right (mean and s.e.m., n 5 3 replicates). d, In dissociated sensory neuron cultures deprived of NGF for 24 h, siRNA knockdown of Casp3 primarily rescues cell body death (TUNEL label), whereas Casp6 knockdown primarily rescues axonal degeneration. e, Quantification of data

from d. Shown are the percentage of degenerating axon bundles (that is, still visible bundles that show breakdown) and the percentage of surviving neurons (that is, TUNEL

2 , TuJ1

1 ) (mean 6 s.e.m., n 5 3 replicates). Note

that surviving axons may have TUNEL 1 cell bodies. The extent of inhibition

by individual siRNAs correlates with the degree of target knockdown (Supplementary Fig. 6d). f, Detection of caspase 6 activation in sensory axons with a cleaved caspase-6-specific antibody (left; TuJ1 double-label on right). Punctate activation of caspase 6 after NGF deprivation (16 h, middle panel) was reduced by anti-DR6.1 (bottom panel). g, Confocal section of a field from f shows that active caspase 6 puncta correspond to sites of tubulin loss (fraction non-overlapping: 82 6 3.5%; mean 6 s.e.m., n 5 8 fields). Scale bars, 75 mm (a, d), 100 mm (b), 50 mm (c, f) and 25 mm (g).

ARTICLES NATURE | Vol 457 | 19 February 2009

984 Macmillan Publishers Limited. All rights reserved©2009

immunosorbent assay (ELISA) (Supplementary Fig. 8c) assays. The interaction detected by ELISA is of high affinity (effector concentra- tion for half-maximum response (EC50) 5 ,4.6 nM). The inter- action of DR6–AP with full-length APP expressed in COS cells was also of high affinity (half maximal saturation 5 ,1.3 nM) (Supplementary Fig. 8a, b). This binding was blocked by anti-N- APP(poly) (data not shown) and anti-DR6.1 (Supplementary Fig. 8a), consistent with APP being a functional DR6 ligand.

Antibodies 4G8 and anti-sAPPb used earlier are highly specific for APP. However, like other antibodies to the N terminus of APP27,

anti-APP(poly) also binds the close APP relative APLP2 (data not shown). We found that a recombinant N-terminal fragment of APLP2 also binds DR6 (Supplementary Fig. 11a). Thus, APLP2 might contribute with APP to the bands detected on western by DR6–AP. Indeed, an antibody selective for the APLP2 N terminus detected a shed fragment in conditioned medium after trophic deprivation (Supplementary Fig. 11b). The relative contribution of APP and APLP2 fragments to DR6–AP binding sites remains to be determined.

To evaluate receptor specificity, we examined by pull-down the binding of APP(1–286) to ectodomains of the seven other

98

62

38

28

16 14

98

62

38

28

16 14

+ – + – TFs SA

98

62

38

28

16

DR6–AP VSC

e f

c d b

T u J1

NGF-deprived

Control Fc DR6–Fc Control Fc

100

80

60

40

20

DR6–Fc

Sensory axons

D e g

e n e ra

ti n g

a xo

n b

u n d

le s

(% )

+ NGF NGF-deprived

B ax

-/ –

S e n so

ry a

xo n s

DR6–AP a

A P

P

E11.5 E12.5 DREZV

C

S

VLF

S

VLF

V DREZ

SN

g

k

S u rf

a c e A

P P

Control + BACE inhibitor + TAPI

NGF-deprived + BAX inhibitor (12 h) + NGF

h

22C11

Anti-N-APP(poly)

Anti-sAPPβ

4G8 Anti-Aβ (33-42)

SA CM

NGF

DR6–AP

98

62

38

28

16 14

Anti-APP depletion Mock depletion

i

IP: Fc (protein A/G)

APP(1–286)–His

Inputs: N-APP

APP(1–286) (~35 kDa)

DR6–Fc

Heparin

Anti-APP

APP(1–286) (~35 kDa) Anti-APP

DR6–Fc

Fc control

Anti-Fc

j

98

62

38

28

16 14

N-APP

NGF + – + –

sAPPβ

98

62

38

28

16 14

NGF

Ligand?

DR6–Fc

DR6

+ TFs – TFs

DR6

Released ligand, active

Surface ligand, inactive DR6

? ?

– – –+ – – +– + – ––

– – +– – + –– – – ++ + + ++

CuBD

KPI OX2

TM

Aβ

CytoE2/carbohydrate binding

β-secretase γ-secretase

Acidic

APP(1–286) ~35 kDa

Growth factor-like

N C

0

Figure 4 | The N terminus of APP is a regulated DR6 ligand. a, Diagram of the hypothesis: if DR6 is ligand-activated, then DR6–Fc might sequester

Collepals.com Plagiarism Free Papers

Are you looking for custom essay writing service or even dissertation writing services? Just request for our write my paper service, and we'll match you with the best essay writer in your subject! With an exceptional team of professional academic experts in a wide range of subjects, we can guarantee you an unrivaled quality of custom-written papers.

Get ZERO PLAGIARISM, HUMAN WRITTEN ESSAYS

Why Hire Collepals.com writers to do your paper?

Quality- We are experienced and have access to ample research materials.

We write plagiarism Free Content

Confidential- We never share or sell your personal information to third parties.

Support-Chat with us today! We are always waiting to answer all your questions.