3,000

Proprietary Co-crystal Structures

90

0x >120x

6MM

Selective Tyk2 Lead Series

Compound

Ideas

Milestone Achieved

Compounds Synthesized

Selectivity over JAK2

Tyk2 Potency 2.5 nM

0

50

100

150

% A

ctiv

ity

JAK2 Tyk2

Cellular Selectivity

31,979

400

>500x

0.3 nM

6 10 0

Months

12,400

260

>250x

0.5 nM

14

In vivo PoM

GMCSF/pSTAT5

IL-12/pSTAT4

0

60

80

100

% A

ctiv

ity

20

40

NDI-031232 NDI-031301

Biochemical Tyk2 Kinase Assay Tyk2 Ki (nM) 0.14 0.53

JAK Family Kinase Biochemical Selectivity (Fold over Ki)

Fold Selectivity over JAK2 210x 85x

Fold Selectivity over JAK1 93x 107x

Fold Selectivity over JAK3 24x 15x

Plasma Protein Binding Human PPB (%bound) 83 74

Human in vitro metabolism Clint (mL/min/kg)

Microsomes 5 3

Hepatocytes 2 9

Mouse PK

IV 3 mg/kg Clobs (mL/min/kg) 35 39

PO 30 mg/kg

T1/2 (h) 1.6 4.8

Cmax (µM) 15 13

F (%) 85 100

AUC (µM*hr) 29 38

Physical Properties MW (Da); solubility (µM) 400-450; 10 400-450; 302

Vehicle - IL-12

Vehicle + IL-12

10 mg/kgNDI-031232

+ IL-12

30 mg/kgTofacitinib

+ IL-12

0

50

100

150

JAK2

pSTAT5pSTAT4

Tyk2/JAK2

GMCSF IL-12

Ki (nM)0.1 to 0.5

5 to 10

10 to 150

Tyk2

JAK3

• IL-12 induced pSTAT4 IC50 = 17 nM • GMCSF induced pSTAT5 IC50 ~ 1.2 µM • Selectivity Tyk2/JAK2 = 70x

0

25

50

75

100

125

Act

ivity

(%)

NDI-031232 (µM)1000.001 0.01 0.1 1 10

IL-12/pSTAT4GMCSF/pSTAT5

IFN

γ (p

g/m

l)

pY pY

IL-12/32

JAK2Tyk2

Tyk2 JAK3

Ki (nM)0.1 to 0.5

5 to 10

10 to 150

• IL-12 induced pSTAT4 IC50 = 74 nM • GMCSF induced pSTAT5 IC50 ~ 4.5 µM • Selectivity Tyk2/JAK2 = 61x

0

25

50

75

100

Act

ivity

(%)

NDI-031301 (µM)1000.0001 0.01 0.1 1 10

IL-12/pSTAT4GMCSF/pSTAT5

0.001

IC50 = 0.4 µM

-1000

0

200

300

400

IFN

γ (p

g/m

l)

NDI-031301 (µM)1000.001 0.01 0.1 1 10

100

IL-12

p35 subunit

p40 subunit

IL-23

INF /

p40 subunit

p19 subunit

Tyk2

Jak2

Jak2

Tyk2 Jak1

Tyk2

STAT 1-5

P

via Th17 Th1

600

JAK

2 Se

lect

ivity

(Fol

d O

ver K

i)

~560x

3503002500 50 100 150 2000

500

400

300

200

100

Number of Compounds Synthesized

Use of FEP for Tyk2 Program

Additional ProprietaryCo-crystal Structures

New sub-seriesthat takesadvantage ofadditionalinteractions withnon-conservedresidues

Free Energy Perturbation (FEP) Calculates Differences in Binding Energy Between Two Structures

via Intermediary States

P=10-25 in >30,000 case-control

samples

P=10-18 in ~15,000

Ass

ocia

tion

(-log

P) 2

2

41 Type 1 diabetes2 Type 1 diabetic ketoacidosis3 Insulin pump user

4 Psoriasis vulgaris 5 Rheumatod arthritis& related disorders6 Rheumatod_arthritis

5

6

4321

OR<

1O

R>1

0

GAA 0.008

CGA 0.038

GGC 0.086

GGA 0.871 2 3 F

0.4 0.6 0.8 1 1.2 0.4 0.6 0.8 1 1.20.2OROR

ref

GGAA 0.15

GAAA 0.006

CGAA 0.036

GGAC 0.711 2 3 F

GGCA 0.083

4ref

case-control samples

1200

1000

800

600

400

200

0Foot

pad

Swel

ling

(µm

)

Tyk2 +/+ Tyk2 +/- Tyk2 -/-

***

*

*

#

###

SalinemBSA

Infectious and ParasiticNeoplasmsEndocrine, nutritional and metabolic; immunity disordersBlood and blood-forming organsMental disordersNervous system and sense organsCirculatory systemRespiratory systemDigestive systemGenitourinary systemSkin and subcutaneous tissueMusculoskeletal system and connective tissue

PheWAS phenotypes (ICD9 Classification)

0.0

0.2

0.4

0.6

0.8

Hours After mBSA Challenge

Vehicle

Paw

Sw

ellin

g (m

m)

-1 23

Dexamethasone 3 mg/kgND-031301 10 mg/kgND-031301 30 mg/kgND-031301 100 mg/kg

**** p = 1x10-8

*** p = 2x10-6

** p = 2x10-4

* p = 1x10-3****

***

***

Identification of Highly Potent and Selective Tyk2 Inhibitors for the Treatment of Autoimmune Diseases Through Structure-based Drug DesignCraig E. Masse1, Wenyan Miao1, Jeremy Greenwood2, Mee Shelley2, Joshua Kennedy-Smith2, Rosana Kapeller1

Nimbus Therapeutics, Inc., Cambridge, MA, USA1; Schrödinger, Inc., New York, NY, USA2

25 First St #404, Cambridge, MA 02141, USA • www.nimbustx.com • Tel: 857.999.2009 • info@nimbustx.com

ABSTRACT

CONCLUSIONS• Tyk2 is a sought after target for the treatment of autoimmune dis-

eases with compelling human genetic data from GWAS/PheWAS studies

• Given the high degree of structural homology amongst the JAK ki-nase family members, designing potent and selective inhibitors has remained a considerable challenge

• Using a physics-based computational approach, Nimbus has uncov-ered previously unexploited drivers of potency and JAK family se-lectivity

• Potent inhibition of IL-12-induced STAT4 phosphorylation and cyto-kine production was observed in human PBMC and whole blood and the Nimbus compounds maintain high levels of functional selectivity over JAK2

• In vivo proof of mechanism was demonstrated in a mouse delayed type hypersensitivity model

• Nimbus compounds have good drug-like properties and are candi-dates for further development for inflammatory diseases

The JAK family kinase Tyk2 is essential for IL-12 and IL-23 signaling, which are associated with Th1 and Th17 cell differentiation and activa-tion. The Th1 and Th17 pathways have been implicated in the patho-genesis of psoriasis and inflammatory bowel diseases (IBD) thereby making Tyk2 a highly attractive target for the treatment of these disor-ders. However, given the high degree of sequence homology between the JAK family kinases, designing potent and selective Tyk2 inhibitors remains a challenge. Using an innovative structure-based approach, we have designed, synthesized and characterized small molecule in-hibitors optimized for JAK family selectivity using computational free energy perturbation (FEP) methods and medicinal chemistry SAR. We have identified selective Tyk2 inhibitors with pM activity against Tyk2 (Ki=140-520 pM) and >100 fold selectivity over JAK2 and JAK1 with more moderate selectivity over JAK3. These analogs are orally bio-available (85%F) with suitable drug-like properties. NDI-031232 was determined to be highly selective across 359 kinases, and is a potent inhibitor of IL-12 induced pSTAT4 in hPBMCs (IC50=17 nM) and IL-12 induced IFNg in human whole blood (IC50=520 nM). NDI-031232 also demonstrated robust efficacy in blocking IL-12-mediated IFNg pro-duction in an ex vivo mouse model. Therefore, selective inhibitors of Tyk2 retain the anti-inflammatory activity while reducing potential for dose-limiting side effects observed with non-selective JAK inhibitors.

All kinase assays were performed with radiolabeled peptides. PK study was conducted in C57BL/6 mice. The ve-hicle for NDI-031232 was 20% DMSO/60% PEG400 in saline, and the vehicle for NDI-031302 was 20% HPbCD in saline.

1. Tyk2: Key Mediator of Th17 and Th1 Pathogenesis

2. Free Energy Perturbation (FEP) Used to Develop Quantitative Selectivity Model to Drive SAR

3. Highly Selective Lead Series Identified in ~10 Months

4. Potency, Selectivity, and Drug-Like Properties of Tyk2 Lead Candidates

5. NDI-031232 Is A Potent and Selective Tyk2 Inhibitor

Kinome selectivity is based on radio-peptide kinase assays. For cell-based assays, human PBMC were pre-in-cubated with NDI-031232 for one hour and stimulated with 1.67 ng/ml of IL-12 for 30 minutes or 10 ng/ml of GMCSF for 10 minutes. Cell lysates were prepared and phospho and total STAT protein was measured by MSD. The ratio of pSTAT/total STAT was used for IC50 calculation. For the ex vivo study, C57BL/6 mice were dosed orally with vehicle (20% DMSO/60% PEG400 in saline), NDI-031232, or Tofacitinib. Whole blood was collected 30 minutes post dose and stimulated with 2 ng/ml IL-12 on anti-CD3 antibody coated plate for 24 hours. At the end of the incubation, IFNg level in the supernatant was quantified by MSD.

6. NDI-031301 Is a Potent and Selective Tyk2 Inhibitor

7. NDI-031301 Reduces Inflammation in the mBSA-induced Delayed Type Hypersensitivity Model

Kinome selectivity is based on radio-peptide kinase assays. Cell selectivity assays were performed with human PBMC as above. Human whole blood assay was performed by incubating NDI-031301 with human whole blood for one hour followed by 0.5 ng/ml of IL-12 stimulation for 24 hours on anti-CD3 antibody coated plate. At the end of the incubation, IFNg level in the supernatant was quantified by ELISA.

C57BL/6 mice were immunized with 0.11 mg of methylated BSA/CFA emulsion at lower back on day 0. Mice were challenged on day 5 by injecting 0.1 mg mBSA/PBS solution into one hind paw and PBS into the other hind paw. Vehicle (20% HPbCD in saline) or NDI-031301 were dosed twice daily orally and Dexamethasone was dosed once daily IP on day 0 through day 5. Paw thickness was measured one day post the challenge (day 6) and paw swelling was calculated by subtracting the thickness of the PBS paw from the mBSA paw of the same mouse. Mice received a final dose on day 6 and plasma was collected one hour post the dose and analyzed for NDI-031301 by LC/M/MS. p values are student t-test vs. the vehicle.

Similarity of Tofacitinib Co-Crystal Structure with JAK1, 2, 3 and Tyk2

FEP-enabled Selectivity Breakthrough (Tyk2 vs JAK2)

Psoriasis and Crohn’s Pathology Lupus Pathology

Multiple Alleles Protect from RA Multiple Alleles Protect from SLE

No Obvious Adverse Events in ~30K Patients

Diogo D, et al. (2015) PLoS ONE 10: e0122271

• Active sites virtually identical

• Gatekeeper residue (methionine) conserved across JAK family

See details in Wang, L., et al. (2015) J. Am. Chem. Soc.,137: 2695

Kinome Selectivity

Cell-Based Assay

Cellular Selectivity

Blocks ex vivo IL-12-induced INFa in the Whole Blood of NDI-031232 Dosed Mice

Kinome Selectivity Cellular Selectivity

Human Whole Blood Activity

mBSA-induced DTH Model

Ishizaki M, et al. (2011) J. Immunol.,187: 181