Modeling the role of AMPA receptor trafficking in the expression of long-term potentiation/depression

Transcription

1 Modeling the role of AMPA receptor trafficking in the expression of long-term potentiation/depression Berton Earnshaw Department of Mathematics Michigan State University October 8, 29 Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, 29 1 / 32

2 Synaptic plasticity Introduction The amazing brain 1 11 neurons 1 1, synapses/neuron regulates body, behavior can learn, remember conscious experience Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, 29 2 / 32

3 Synaptic plasticity Introduction Neurons communicate at synapses Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, 29 3 / 32

4 Synaptic plasticity Communication at a synapse Introduction Kandel, Schwartz & Jessel (2) Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, 29 4 / 32

5 Synaptic plasticity Introduction Synapses can learn synaptic plasticity Collingridge et al., Nat Rev Neurosci (24) Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, 29 5 / 32

6 Synaptic plasticity Introduction Synapses can learn synaptic plasticity Collingridge et al., Nat Rev Neurosci (24) LTP = long-term potentiation (strengthens synapse) LTD = long-term depression (weakens synapse) Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, 29 5 / 32

7 Synaptic plasticity Pre/postsynaptic mechanism How/where does synaptic plasticity occur? Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, 29 6 / 32

8 Synaptic plasticity Pre/postsynaptic mechanism How/where does synaptic plasticity occur? Two major hypotheses: Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, 29 6 / 32

9 Synaptic plasticity Pre/postsynaptic mechanism How/where does synaptic plasticity occur? Two major hypotheses: 1 Presynaptic: change in the number of vesicles/probability of release Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, 29 6 / 32

10 Synaptic plasticity Pre/postsynaptic mechanism How/where does synaptic plasticity occur? Two major hypotheses: 1 Presynaptic: change in the number of vesicles/probability of release 2 Postsynaptic: change in the number/conductance of receptors Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, 29 6 / 32

11 Synaptic plasticity Pre/postsynaptic mechanism Presynaptic vs. postsynaptic mechanisms Depressed LTD Naïve LTP Potentiated POST PRE Depressed LTD Naïve LTP Potentiated Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, 29 7 / 32

12 Synaptic plasticity Presynaptic vs. postsynaptic debate Pre/postsynaptic mechanism...long-term plasticity...is expressed overwhelmingly via presynaptic changes in reliability of transmitter release. Therefore, LTP is the recruitment of new [receptors] to synapses... Enoki et al., Neuron (29) Kerchner & Nicoll, Nat. Rev. Neurosci. (28) Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, 29 8 / 32

13 Can modeling help? Synaptic plasticity Pre/postsynaptic mechanism Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, 29 9 / 32

14 Can modeling help? Synaptic plasticity Pre/postsynaptic mechanism That depends on who you talk to! Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, 29 9 / 32

15 Can modeling help? Synaptic plasticity Pre/postsynaptic mechanism That depends on who you talk to! Question to answer... Can LTP/LTD data be explained by postsynaptic receptor trafficking alone? Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, 29 9 / 32

16 AMPA receptor trafficking Outline Outline for rest of talk Introduce AMPA receptors Describe AMPA receptor trafficking Propose model of AMPA receptor trafficking Present results from model Conclusions & future directions Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, 29 1 / 32

17 AMPA receptor trafficking Introduction AMPA receptors Responsible for excitable synaptic transmission in CNS Formed from four subunits: GluR1 to GluR4 Dominant heteromers: GluR1/2 and GluR2/3 Huganir & Song, Nat. Rev. Neurosci. (22) Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, / 32

18 AMPA receptor trafficking Introduction Excitable synapses located in dendritic spines Matus, Science (2) Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, / 32

19 AMPA receptor trafficking Introduction AMPA receptor trafficking at a spine Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, / 32

20 AMPA receptor trafficking Single-spine model Model of AMPA receptor trafficking at a single spine exocytosis AMPA receptor endocytosis scaffolding protein diffusion Cottrell et al., J. Neurophysiol. (2) Sorra & Harris, Hippocampus (2) Ehlers, Neuron (2) Passafaro et al., Nat. Neurosci. (21) Groc et al., Nat. Neurosci. (24) Time constants Exo/endocytosis: 1-3 min Diffusion: 1 s Surface area of PSD:.1 µm 2 Surface area of spine head: 1 µm 2 Diffusion coefficient:.1-.1 µm 2 /s Binding/unbinding to scaffolding: unknown Production/degradation: unknown Other constants Intracellular AMPAR number: 1-5 AMPAR concentration in dendrite: 1-1/µm 2 Scaffolding concentration: unknown Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, / 32

21 AMPA receptor trafficking Single-spine model State diagram β diffusion Q P h R µ U exocytosis endocytosis PSD α(z-q) σ EXO k Spine Head C AMPA receptor scaffolding protein Variables P = free AMPAR concentration in PSD Q = bound AMPAR concentration in PSD R = AMPAR concentration in spine head Constants C = intracellular AMPAR number U = AMPAR concentration in dendrite Z = scaffolding protein concentration Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, / 32

22 Model equations AMPA receptor trafficking Single-spine model PSD free: PSD bound: Spine head: dp dt = h(r P) α(z Q)P + βq + σexoc a dq = α(z Q)P βq dt dr = µ(u R) h(r P) kr dt diffusion exocytosis endocytosis Q β P h R µ U α(z-q) PSD σ EXO k Spine Head AMPA receptor scaffolding protein C BAE & Bressloff, J. Neurosci. (26) Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, / 32

23 AMPA receptor trafficking Single-spine model Parameter values Known parameter values: Rate of exocytosis, endocytosis: σ EXO,k = 1 3 /s Surface area of PSD, spine head: a =.1 µm 2, A = 1 µm 2 Hopping rate between PSD and spine head: h =.1/s Hopping rate between spine head and dendrite: µ =.5/s Ashby et al., J. Neurosci. (26) Intracellular AMPA receptor number: C = 1 AMPA receptor concentration in dendrite: U = 2/µm 2 Unknown parameter values: Binding/unbinding rates α and β Constitutive recycling 1 3 min α = β = 1 3 /s Scaffolding protein concentration Z Approx. half AMPARs in PSD are bound Z = 2/µm 2 Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, / 32

24 Steady-state AMPA receptor trafficking Results from single-spine model PSD free: PSD bound: Q = P = R + σexo h αp αp + β Z C a 193/µm 2 199/µm2 Spine head: R = µu + σexo C a µ + k 183/µm 2 β Q P h R µ U α(z-q) PSD σ EXO k Spine Head C Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, / 32

25 Block exo/endocytosis AMPA receptor trafficking Results from single-spine model number of receptors in PSD Total Bound Free number of receptors in PSD Total Bound Free time [min] time [min] Luscher et al., Neuron (1999) Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, / 32

26 LTP simulation AMPA receptor trafficking Results from single-spine model GluR1/2 GluR2/3 ΙΙ Activation of GluR1/2 intracellular pool Ι Rapid insertion of receptors into ESM number of receptors in PSD t [min] Total Bound GluR1/2 Free GluR1/2 Bound GluR2/3 Free GluR2/3 Scaffolding O' Connor et al (PNAS 25) Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, 29 2 / 32

27 AMPA receptor trafficking Results from single-spine model LTP simulation with increase in scaffolding GluR1/2 Ι ΙΙ GluR2/3 Activation of GluR1/2 intracellular pool Rapid insertion of receptors into ESM AMPARs transport slot proteins into PSD number of receptors t [min] Total Bound GluR1/2 Free GluR1/2 Bound GluR2/3 Free GluR2/3 Scaffolding O' Connor et al (PNAS 25) Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, / 32

28 LTD simulation AMPA receptor trafficking Results from single-spine model PSD ESM PSD ESM GRIP PICK Switch from AMPA-GRIP to AMPA-PICK receptor-protein complexes Rapid unbinding from PSD and trafficking to ESM followed by endocytosis. number of receptors in PSD Total Scaffolding t [min] Dudek & Bear (PNAS 1993) Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, / 32

29 AMPA receptor trafficking Results from single-spine model LTD simulation with decrease in scaffolding PSD ESM PSD ESM PSD ESM GRIP PICK Switch from AMPA-GRIP to AMPA-PICK receptor-protein complexes Rapid unbinding from PSD and trafficking to ESM followed by endocytosis. Unbound scaffolding proteins are degraded. 7 number of receptors Total Scaffolding Dudek & Bear (1993) t [min] Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, / 32

30 Conclusions AMPA receptor trafficking Conclusions & future directions Can LTP/LTD data be explained by postsynaptic receptor trafficking alone? Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, / 32

31 Conclusions AMPA receptor trafficking Conclusions & future directions Can LTP/LTD data be explained by postsynaptic receptor trafficking alone? LTP/LTD data can be reproduced within our model of AMPA receptor trafficking Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, / 32

32 AMPA receptor trafficking Conclusions & future directions Conclusions Can LTP/LTD data be explained by postsynaptic receptor trafficking alone? LTP/LTD data can be reproduced within our model of AMPA receptor trafficking LTP requires increase in scaffolding (Shi et al., Cell 21) LTD requires decrease in scaffolding (Colledge et al., Neuron 23) Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, / 32

33 AMPA receptor trafficking Conclusions & future directions Future directions AMPAR trafficking along dendrite receptor scaffolding protein PSD EXO END J soma DEL DEG x = spine U t = D 2 U η(x)µ(x)(u R) x2 D U U x = J soma, x= x =. x=l x = L Bressloff, BAE, Ward, SIAM J Appl Math (28) BAE & Bressloff, J Comput Neurosci (28) Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, / 32

34 AMPA receptor trafficking Conclusions & future directions Future directions AMPAR trafficking along dendrite 1-fold reduction in rate of exocytosis in gray region Consequences of diffusive coupling 1-fold increase in rate of endocytosis in gray region PSD receptors PSD receptors before perturbation after perturbation distance from soma [µm] distance from soma [µm] Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, / 32

35 AMPA receptor trafficking Conclusions & future directions Future directions AMPAR trafficking along dendrite Fast or slow recycling? Passafaro et al., 21 Adesnik et al., 25 recovery [% baseline] µm 3 µm time [hr] Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, / 32

36 AMPA receptor trafficking Future directions stochastic models dp = α(z q)p + βq µp + σ dt dq = α(z q)p βq dt DEG Conclusions & future directions A PSD EXO END PROD ESM B AMPA receptor scaffolding protein β µ q p α(z - q) σ PSD ESM Bressloff & BAE Biophys J (29) Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, / 32

37 AMPA receptor trafficking Future directions stochastic models dp = α(z q)p + βq µp + σ dt dq = α(z q)p βq dt Conclusions & future directions P n,m (t) = Prob{n unbound,m bound at time t} A ESM EXO PROD PSD END DEG B AMPA receptor scaffolding protein β µ q p α(z - q) σ PSD ESM Bressloff & BAE Biophys J (29) Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, / 32

38 AMPA receptor trafficking Future directions stochastic models dp = α(z q)p + βq µp + σ dt dq = α(z q)p βq dt Conclusions & future directions P n,m (t) = Prob{n unbound,m bound at time t} dp n,m dt = σp n 1,m + µ(n + 1)P n+1,m + α(n + 1)[Z (m 1)]P n+1,m 1 + β(m + 1)P n 1,m+1 [σ + µn + αn(z m) + βm]p n,m A B ESM ESM EXO PROD PSD PSD END DEG AMPA receptor scaffolding protein β µ q p α(z - q) σ Bressloff & BAE Biophys J (29) Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, / 32

39 AMPA receptor trafficking Future directions stochastic models dp = α(z q)p + βq µp + σ dt dq = α(z q)p βq dt Conclusions & future directions P n,m (t) = Prob{n unbound,m bound at time t} dp n,m dt = σp n 1,m + µ(n + 1)P n+1,m + α(n + 1)[Z (m 1)]P n+1,m 1 + β(m + 1)P n 1,m+1 [σ + µn + αn(z m) + βm]p n,m stochastic gate : < µ open γ γ + µ closed = σ(t) = Cµ(t) (C bath conc.) A B ESM ESM EXO PROD PSD PSD END DEG AMPA receptor scaffolding protein β µ q p α(z - q) σ Bressloff & BAE Biophys J (29) Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, / 32

40 AMPA receptor trafficking Future directions stochastic models Conclusions & future directions A variance D variance dynamic static dynamic 2 static 1.6 mean time [s] B E time [s] C F time [s] Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, / 32

41 AMPA receptor trafficking Conclusions & future directions Future directions trafficking of other proteins Rose et al., Neuron (29) Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, 29 3 / 32

42 AMPA receptor trafficking Conclusions & future directions Future directions trafficking of other proteins p t = D 2 p x 2 kap a t = D 2 a + kap ha x2 s t = ha BAE & Bressloff, J. Comput. Neurosci. CaMKII [norm. conc.] CaMKII [norm. conc.] total in spines total in dendrite primed in dendrite activated in dendrite x [µm] 1 sec 65 sec 14 sec 348 sec x [µm] Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, / 32

43 AMPA receptor trafficking The end Thank you! Thanks to Paul Bressloff (Oxford) Michael Ward (UBC) National Science Foundation Earnshaw (MSU) AMPAR trafficking and LTP/LTD October 8, / 32