bms mist strategy and experience · guidance on nonclinical safety studies for the conduct of human...

BMS MIST strategy and experience

DV DMDG Rozman Symposium13 June 2017

Griff HumphreysBristol Myers Squibb R&D

Princeton, NJ

Bristol-Myers Squibb R&D1

Bristol-Myers Squibb R&D

Outline

• Intro• BMS strategy to meet MIST• Experiences and learnings• How well has MIST met goals?

2


Safety Testing of Drug Metabolites

• Goal: Proper testing of the safety of a new drug requires that the metabolites that humans are exposed to are also tested in the species used for toxicological evaluation – Obligation to patient safety during clinical

trials and post-approval– While the goal sounds simple in principle, in

practice it can be rather complicated…need proper framework to set expectations

3



•Sometimes the goal of ensuring adequate exposure to human metabolites can be fairly simple

CP

M

Time (min)0

20

40

60

80

100

120

140

160

180

200

0 10 20 30 40 50 60

I h

4 h

6 h

M2

d4T

0

20

40

60

80

100

120

0 10 20 30 40 50 60

4



-5

45

95

145

195

245

295

0 10 20 30 40 50

-5

45

95

145

195

245

0 10 20 30 40 50

-5

95

195

295

395

495

595

695

795

0 10 20 30 40 50

-5

35

75

115

155

195

235

275

315

0 10 20 30 40 50

-5

35

75

115

155

195

0 10 20 30 40 50

Mouse

Rat

Dog

Monkey

Human

Time (min)

CP

M

CP

MC

PM

CP

MC

PM

P

P

P

P

P

4 4

4

5

5

5

5

7575

75

75

3

3

3

3

728180

7113

7281

807113

72

882

7113

88271

1372

21

21

69

69

3

74882

7113

69

15

882

882

1564

64

15

15

6264

59

4417

4417

4417

26

323839

5254

14

585962

434417

39

59

14

5254

3839

43

4446

525414

58 15

596264

3839

5254

14

62

3032

48

5254

1458

8316

7879

7778

777879

7779

16

•Sometimes the goal of ensuring adequate exposure to the complex mixture of human metabolites can be very challenging!

5


Safety Testing of Drug Metabolites• History

– Prior guidance from FDA• ICH S3A, March 1995• Carcinogenicity Study Protocol Submission, since 1997

– PhRMA MIST task force, 1999-2002 (culminated in publication - Baillie et al, TAP, 182, 188-106, 2002)

– PhRMA/FDA Joint workshop, Nov 2000– FDA Draft Guidance, June 2005– PhRMA/FDA/DruSafe Joint workshop, Nov 2005– Safety Testing of Drug Metabolites Guidance for Industry Feb 2008– GUIDANCE ON NONCLINICAL SAFETY STUDIES FOR THE

CONDUCT OF HUMAN CLINICAL TRIALS AND MARKETING AUTHORIZATION FOR PHARMACEUTICALS M3(R2), Dec 2009

– FDA revised guidance, Nov 2016

6

FDA Safety Testing of Drug Metabolites Guidance for Industry

7

“Generally, systemic exposure is assessed by measuring the concentration of the parent drug at steady state, in serum or plasma.”


Other Considerations Relevant for Strategy Around Metabolite Characterization

8

1) On-target pharmacological activity of significant metabolites

2) Off target-related pharmacological activity of significant metabolites

3) Characterization of “other” off-target activity (including CYP and transporter liabilities) of major metabolites (DDI guidances)


Safety Testing of Drug MetabolitesBMS Strategy – Multi-tiered approach• Determine metabolites with high probability to be circulating human

metabolites from in vitro and animal in vivo studies• Analyze metabolite profile present in samples from MAD study. Use

HR-MS with data mining as front line tool to ensure robust metabolite detection

• Prepare samples from satellite animal groups, compare MS response of human metabolites to equivalent animal metabolite

• Employ multiple methods to estimate concentration of metabolites• Characterize major metabolites:

– Determine exposure in human and animals with fit for purpose analytical methodology

– On and Off target activity as needed• Final characterization in human ADME study, typically concurrent

with phase 2Bristol-Myers Squibb R&D

9

General Approach for Characterizing Human Metabolites

Profile plasma metabolites from SAD

by LC/UV/HRMS

Quantitative estimate of major plasma metabolites from MAD by LC/HRMS

Determine definitive structures of major

metabolites in human plasma using NMR

Evaluate exposure coverage of major human

plasma metabolites in Tox species after multiple

doses

What were major plasma metabolites?What were structures of major metabolites?

What are major plasma metabolites at steady state after therapeutic doses?

Do Tox species provideexposure coverage of the major human metabolites at steady state?

Chemically or enzymatically

synthesize the major metabolites

On / Off target activityTransporter / Cyp

inhibition/ induction

Enable quantitation of metabolites in clinical samples by LC-MS/MSConfirmation of plasma profiles

and determination of major metabolic pathways in definitive

14C ADME studiesBristol-Myers Squibb R&D

10

Metabolite profiling in human plasma sample

5 10 15 20 25 30Time (min)

0

50

1000

50

1000

50

100

31.0030.64

30.77

26.06

30.77

NL: 3.19E6Base Peak F: FTMS - c ESI Full ms [100.00-1000.00] MS

a_2h_bk

NL: 1.56E7Base Peak F: FTMS - c ESI Full

ms [100.00-1000.00] MS

a_2h_30ul

NL: 5.72E5Base Peak F: - c EI Full ms [100.00-1000.00] MS

2h dd 2h slicer1out

P

P

M2M1

M2M1

Placebo plasma

Dosed plasma

Background subtracted,dosed plasma

• Background subtracted LC/HRMS profiles showed that M1 and M2 were significant plasma metabolites


LC/UV profiles of human plasma (MAD)

6 8 10 12 14 16 18 20 22 24 26 28 30 32 34Time (min)

0

2000000

uAU

0

2000000

4000000

uAU

0

2000000

4000000

6000000

uAU

0

500000

1000000

uAU

30.9922.39

22.83

30.43 32.23

30.98

31.8730.97

8.1311.05

31.88

8.017.56 31.43

Day 14, 0 h

M2M1

Day 14, 8 h

Day 14, 4 h

Day 14, 2 h P

P

P

M2

M2

M1

M1

UV

Abso

rptio

n

• At steady state, M2 was the only metabolite that had an exposure greater than 10% of total drug related material in human


How to do initial quantitation from FIH samples?

BMS Strategy for Quantitative Estimate – Use Multiple Approaches

• UV detection, especially if MS points to transformations expected to have a low impact on UV properties

• NMR quantitation of isolated sample, potential to use as a reference standard

• Employ radioactivity based quantitation from an in vitro or in vivo derived samples as a calibrant for MS response


General Approach for Characterizing Human Metabolites


by LC/UV/HRMS

Quantitative estimate of major plasma metabolites from

MAD by LC/HRMS/UV


metabolites in human plasma using NMR

Evaluate exposure coverage of major human plasma metabolites in Tox

species after multiple doses








Enable quantitation of metabolites in clinical samples by LC-MS/MS

Confirmation of plasma profiles and determination of major

metabolic pathways in definitive 14C ADME studies


Comparison of Steady State Metabolite Exposures Without Use of Reference Standards

• Plasma samples from MAD study• Plasma samples from steady multiple-day studies at

NOAEL doses in relevant toxicology species• For each species prepare one representative AUC-

pooled sample• Dilute the AUC-pooled sample from each species with

blank plasma from the other 2 species (matrix match)• Analyze peaks areas of metabolites by LC-HRMS• Metabolite peak area ratio in animal vs human will

represent the AUC ratio between the species

15Bristol-Myers Squibb R&D

Comparison of M1 & M2 exposures in AUC pooled samples from rat, dog and human

Human, Day 14 AUC pool, Dose 1

Human, Day 14 AUC pool, Dose 2

Rat, Day 7 AUC pool

Dog, Day 7 AUC pool

Human placebo

M1M2

5 10 15 20 25 30 35Time (min)

0

50

1000

50

1000

50

1000

50

1000

50

10029.99

24.6433.55

30.14

33.5424.71

29.96

24.10 33.51

29.9324.34

28.75 33.4633.57

17.6816.92 18.7114.3811.069.198.34

M2

M2

M1M2

M1

M1


• LC-HRMS, extracted ion chromatograms• Peak ratios showcoverage for M2 good in both species; coverage For M1 good in dog, challenging in rat

General Approach for Characterizing Huma Metabolites


by LC/UV/HRMS


MAD by LC/HRMS/UV


metabolites in human plasma



doses









Confirmation of plasma profiles and determination of major

metabolic pathways in definitive 14C ADME studies


General Approach for Characterizing Huma Metabolites


by LC/UV/HRMS


MAD by LC/HRMS/UV


metabolites in human plasma



doses






On / Off target activityTransporter / CYP



Confirmation of plasma profiles and determination in definitive

14C ADME studies



Is it Worthy of GLP Bioanalysis?Tiered Approach Reduces Bioanalytical Workload

• At Crystal City May 2006 there was agreement that a tiered approach is possible for metabolite analysis– Not all assays must be GLP– Non-GLP LC/MS assays

• When seen as major metabolites in humans, GLP assays are usually developed, validated and used– Plasma exposures >10-25% threshold– Safety margin < 1– Profiling method may serve as the basis for GLP

assay development

19

20

Plasma Metabolite Profiles from 14C-labeled Study in Human

• Radioprofiles confirm data from exploratory profiling showing M1, M2 were the only prominent metabolites in human after single dose

M1 M2



Safety Testing of Drug Metabolites• How has the strategy worked so far?

– Allowed early determination of major circulating human metabolites and on average both reduced animals 14C-ADME studies and led to delay in 14C-human ADME to later in development

– Only one instance of a late “surprise” in 14C-study in ca. 10 years of HRMS based screening• Learning: In case above, new metabolite would have been

detected using negative ion electrospray. All FIH samples analysis now done on instrument with positive/negative switching

21


Safety Testing of Drug Metabolites –coming up on 10 year anniversary!

• Has MIST (and ICH) lived up to its goals?

1) Make drugs safer, both during clinical trails and post-marketing2) Give sponsors and regulatory agencies a common framework and set of expectations for evaluation of overall safety testing of a drug and its metabolites

22

Pre-MIST experience…Metabolic pathways of apixaban circa 2005

• M1 accounts for approximately 20% of the circulating radioactivity in human subjects, 1-5% in animals

NN

N

N

ONH2

O

OCH3

O

BMS-562247

NN

N

N

ONH2

O

OH

O

NN

N

N

ONH2

O

OH

O

NN

N

N

ONH2

O

OCH3

O

M2

M7 or M4 M10

NN

N

N

ONH2

O

OSO3H

O

M1

SO3H

OH*

Urine: 21.5%Feces: 34.0%

Urine: 1.58%Feces: 1.16%

M7: Urine: 1.46%, Feces: 3.70M4: Urine: ND, Feces: 0.37

Urine: NDFeces: 12.2%


Recovery (% dose):Urine: 24.5%Feces: 56.0%

6 subjects following 20 mg oral dose

SULT1A1 &1A2

CYP3A4

CYP3A4

NN

N

N

ONH2

O

OH

O

M13

O


OH


Characterization of M1Mean Cmax and AUC in toxicological species

and humans at relevant doses

Species Dose Cmax (ng/mL)

AUC (ng/mL*h)

Rat (7 day) 600 mg/kg/day 45 197

Dog (7 day) 100 mg/kg/day 66 1204

Human (single dose) 20 mg (solution) 79 1829

• Initial HA interaction indicated a need to study M1 in a dedicated rat study

Characterization of M1• Potential for chemical reactivity

– Sulfate conjugates with well characterized potential for chemical reactivity:• Conjugates of benzylic and allylic alcohols; aryl amines,

hydroxylamines, and hydroxamic acids; secondary nitroalkanes – Phenolic sulfates, such as M1, are not known to be

chemically reactive• No association of carcinogenicity/mutagenicity with phenolic

sulfates– Drugs, endogenous substances and food constituents

• Thermodynamically stable in aqueous solution• Chemical reactions limited to hydrolysis of conjugate to yield

starting phenol – Quantum chemical calculations demonstrate that the

energetics of formation of cationic reactive species from phenolic sulfates is highly unfavorable


M1 Issue Resolution• M1 was not a chemically reactive and did not inhibit

factor Xa• The enzymes predicted to be responsible for

formation of M1 were not subject to significant polymorphic distribution

• Cmax exposure values of M1 in humans at clinical doses are similar to those found in rats and dogs at toxicological doses

• AUC exposure values of M1 in humans at clinical doses are similar to those found in dogs at toxicological doses

• Eventually the request to do direct animal studies with M1 was dropped and study not conducted…Safety Testing of Drug Metabolites Guidance likely would have led to a more rapid resolution of this request



What biological activity will a metabolite likely possess?

• Typical metabolites are very closely related structurally to the parent, thus would be expected to have some degree of activity against the target receptor, i.e., it is not surprising that they would follow the same SAR as parent (Humphreys and Unger, CRT, 19, 1564-1569)

• Conversely, these closely related species would not necessarily be expected to gain new activities not seen with parent– Does this hypothesis hold true? Seems to be from the

literature as there are few examples of where a stable metabolite produced toxicity through a pathway not impacted as well by the parent.

27


What biological activity will a metabolite likely possess?

• How often do parent-metabolite pairs display significantly different off-target pharmacology?– Pairs such as terfenadine-fexofenadine are examples of

parent-metabolite with significant off-target effects (HERG binding), however, the off-target activity resides with the parent

– Metabolites can have significantly different physicochemical properties which may make them prone to physical deposition during elimination leading to toxicity (eg., guaifenesin induced urolithiasis)

28


So why characterize stable metabolites?• There is a reasonable chance they will contribute to the

pharmacological activity, and may even drive a significant portion of the observed efficacy

• While examples of stable metabolites mediating off-target toxicity are relatively rare there are some notable recent examples:– An IQ working group confirmed that DDIs mediated by

metabolites were rare (Yu H, et al. DMD (2015) 43:620-30), however, a Vertex clinical candidate demonstrated a very significant metabolite mediated CYP3A DDI (Zetterberg C, et al. DMD (2016) 44:1286-95)

– A clinical candidate from Incyte was halted due to unexpected renal toxicity thought to be driven by the local crystallization of a human selective AO metabolite (Diamond S, et al. DMD (2010) 38:1277-85)

29


Safety Testing of Drug Metabolites –coming up on 10 year anniversary!

• Has MIST (and ICH) lived up to its goals?

1) Make drugs safer, both during clinical trails and post-marketing2) Give sponsors and regulatory agencies a common framework and set of expectations for evaluation of overall safety testing of a drug and its metabolites

30


AcknowledgementsMingshe Zhu Donglu ZhangLi MaWeiping ZhaoWenying LiRama IyerLisa ChristopherJinping GanHaiying ZhangMAP/Biotransformation

Group

31

bms mist strategy and experience · guidance on nonclinical safety studies for the conduct of human...

Documents