Molecular Physiology and Pharmacology of ATP1A3 Mutations in AHC Principal Investigator: Kevin C. Ess, M.D., Ph.D., Vanderbilt University Co-Investigator: Alfred L. George, Jr., M.D., Northwestern University

Overview and Progress Report – Vanderbilt

Alternating Hemiplegia of Childhood (AHC) is a perplexing neurodevelopmental disease that usually presents at less than 18 months with weakness of one side of the body (hemiplegia) that lasts for hours

to days. These events are very frightening to families as well as health care professionals. Incorrect diagnoses are then considered including stroke, brain tumors, and usually epilepsy. Over time,

independent alternating involvement of both sides of the body becomes evident, patients can also have episodes of full body weakness. Interestingly, most patients have resolution of their symptoms when they awake from sleep. While most patients are initially misdiagnosed as having seizures or stroke, the pattern of recurrent hemiplegia and recovery in the setting of normal brain MRI scans and lack of any seizure activity eventually leads to a correct diagnosis. This is especially the case as more information and knowledge about AHC is disseminated to the medical community. Later in life, most patients develop developmental delay and intellectual disabilities with variable manifestations ranging from mild to severe.

There is no treatment for AHC that is supported by empirical evidence, most approaches use sleep-inducing agents that can then terminate episodes of hemiplegia. A non-FDA approved drug called flunarizine has been used but there is scant objective evidence supporting its use. While AHC has been recognized as a distinct syndrome for many years, the genetic cause was only discovered in 2013 and is due to mutations ATP1A3. This gene encodes a fundamentally important protein pump used by neurons to control the proper levels of sodium and potassium inside cells. While the definition of the genetic underpinnings of AHC was a tremendous advance that gives great excitement to this field of research, we still have only a basic understanding how ATP1A3 functions, the impact of common and discrete gene mutations, and how new therapies may now be considered based on this information.

Our research has focused on the normal function of ATP1A3 and how common mutations found in patients with AHC cause neuronal dysfunction. Using a novel functional assay, we recently found that the most common gene mutations that cause AHC impair the flux of sodium and potassium at the surface of the cell. Some of these mutations appear to prevent the protein pump from being correctly “tracked” and suitably expressed on the surface of the cell. This is a tremendous insight as drug discovery approaches may work by simply pushing the abnormal form of the protein to the right locartion in the cell.

An additional and equally exciting avenue of our research has been the creation of human stem cells using skin cells taken from patients with AHC. This technology uses what is called induced pluripotent stem cells (iPSC) and was a huge leap of progress as stem cells can now be made from patient skin cells. This completely avoids any use of embryonic stem cells that remain highly controversial given their origin from human fetuses. iPSC work was so ground-breaking that only 6 years after it was first reported, the inventor of this technology, Dr. Shinya Yamanaka, was awarded the Nobel prize in 2012. The use of patient derived stem cells from patients with AHC then allows us to make human neurons in a dish and directly study the impact of ATP1A3 gene mutations on neuronal function. Such approaches are becoming more common in biomedical research and are very complementary to more traditional approaches using mice or rats to model human disease. The drawbacks of using iPSC however include their cost of production and maintenance as they require very specialized growth conditions and daily care by skilled research technicians and scientists.

In the Table below, we summarize our iPSC generation to date and also the production of human neurons derived from them. Ongoing efforts have allowed us to produce these cells from all common mutations that cause AHC. Control lines derived from healthy volunteers are also crucial to define normal cellular behavior and ATP1A3 function. Figures 1, 2, 3, 4, and 5 below show examples of human AHC patient iPSC, validation of their stem cell status, and derivative neurons.

Patient

AHC-4

AHC-8

AHC-12

AHC-14

AHC- 2

AHC 95

CA

CC

CD

t ID

(AHC 35189

(AHC 91759

2 (AHC 4193

4 (AHC 4633

20 (AHC 421

5454

ATP1A

9) D801N

9) E815K

35) G947R

37) G947R

03) G947R

Splice c

None (c

None (c

None (c

A3 Mutation

c.2542+2A

control line)

control line)

control line)

iPSC

yes

yes

yes

yes

yes

G yes

yes

yes

yes

C

Neur

yes

yes

yes

yes

yes

In pro

yes

yes

yes

on

ogress

P S

WinchFmFmEeainwmcofeamETin

WstexAaspsoof

C AATDbaindochte

Progress Re

Screening of

We have conn the presenhemical librurther, we h

manner. TherDA-approve

most commoE815K, G947

ach drug ncubation, awere very femutant ATP1ompound (cemale sex hll three muta

most robust. E815K, we h

herefore, wen our future e

We also scretructures thaxposure par

ATP1A3 cell ll 71 chemicpecific to ATomething enf different ce

Computationa

As an alternaATP1A3 mut

hese studieDiscovery at

ased on pren the vicinityocking was hemical struesting of thes

eport – North

FDA-approv

tinued to usnce of test crary of all Fhad not perforefore, we p

ed drugs ann AHC-asso7R) against under twond after a sew drugs tA3 activity chlormadinoormone proations (Fig. In our prior

had also obse will be puexperiments

eened a seat are sharradigm. The line as well cals activate

TP1A3. Manndogenous tell types and

al Screening

ative strategations, we p

es were perNorthweste

eviously detey of several performed

uctures werese compoun

hwestern

ved drugs th

e a high-throcompounds FDA-approveormed a comurchased a nd re-screenociated ATPeach drug in

o conditionsshort (20 mthat exhibiteby either ex

one acetateogesterone, 6) with effecscreen for c

served an ersing the me

s (Specific A

econd chemred by drugresults incluas the origined somethiny of the 71 cto HEK-293 d to target a

g for ATP1A3

gy to find smperformed arformed in

ern Universitermined mol

known ATPagainst a lib

e identified tnds will be pe

hat restore m

oughput ass(drugs and

ed drugs, wmplete screenew chemicned each o1A3 mutation the librarys: followinginute) expoed positive xposure para), a derivashowed actcts on D801compounds

effect of chloechanism fo

Aim 2).

ical compous that targe

uded 71 activnal HEK-293ng in the HEcompounds cells. Furthemore chemi

3 Docking C

mall chemicaa computer-bcollaboratioty. The workecular strucP1A3 mutatbrary of 10 that met criterformed in

mutant ATP1

say system ( drug-like c

we had conen of the thr

cal library ofof the threeons (D801N,y. We testedg overnightsure. There

effects onadigm. One

ative of theivity againstN being thethat rescue

ormadinone.or this effect

und library wet ion transve compoun

3 cell line (toEK-293 cellhad a share

er assay devcally diverse

Compounds

al compoundbased screen with the k began by

ctures (Fig. 7tions includimillion smaeria for docSpecific Ai

A3 activity

(thallium uptchemicals). Ancerns that ree most cofe ,

d t

e n e e t

e e . t

with more tporting prot

nds that wereo exclude nos suggestin

ed chemical velopment ee library.

ds that mayen for compCenter for constructin

7a). A potenng D801N (

all chemical king (exampm 2.

take) to meaAlthough wethese resul

ommon muta

than >6,000teins. This e then assay

on-specific efng that none

structure thefforts are on

y correct thepounds that

Molecular g a comput

ntial binding (Fig. 7b). Nstructures.

ples illustrat

asure ATP1Ae previouslyts were no

ations in a s

0 entities enscreen usedyed again agffects). Unfoe of the effeat might be ngoing to ex

e functional ‘dock’ to thInnovation ater model ofpocket was

Next, compuFrom this sed in Fig. 8

A3 activity y tested a t optimal.

systematic

nriched in d a short gainst the

ortunately, ects were activating

xplore use

effects of e protein. and Drug f ATP1A3 identified

uter-based creen, 62

8). Further