Transcript Dr. Gelzleichter: Evaluation of Carcinogenic Risk for Biotechnology

Nor Cal SOT
Evaluation of Carcinogenic Risk for
Biotechnology-Derived Therapeutics
Tom Gelzleichter
September 27th, 2012
©2012, Genentech
Topics
1.
2.
3.
4.
5.
Limitations on utility of standard rodent bioassays for biologics
Historical approaches for risk assessment of biologics
2011 revisions to ICH S6
Examples of revised approach
How will these changes impact risk communication?
©2012, Genentech
Main Objective of Carcinogenesis Testing for
Pharmaceuticals
A product-specific assessment of carcinogenic potential is
used to communicate risk and provide input to the risk
management plan along with labeling proposals, clinical
monitoring, post-marketing surveillance
©2012, Genentech
2 year bioassays in general have limited utility for all chemical
classes
Interpretation difficult due to:
– Lack of known negative controls (IARC only classifies one chemical as probably
not carcinogenic in humans)
– Susceptibility determined by genotype, sex and test conditions
• Examples: cigarette smoke, arsenic, benzene were challenging to find
rodent models that gave positive results
• Lack of concordance across sexes, species (rarely are tumors found in all 4
genotypes i,.e., rat/mouse/M/F)
– Inter-rodent predictivity (rat:mouse) 70-75%
• Validation efforts have been heavily skewed towards certain chemical classes
(plus, nearly all have been genotoxicants)
– Poor concordance for immunotoxicants, some hormones
• Poor reproducibility (only 57% concordance when repeated)
• Positivity rate is extremely high
– In NTP studies, 68% of tested chemicals are positive in at least one of
the 4 genotypes
– 40% of marketed pharmaceuticals and food additives are positive
©2012, Genentech
Why are Chronic Rodent Bioassays Still Used?
– Most known human carcinogens are positive in at least one of the 4 genotypes
tested, when evaluated at MTD
– Only 5-10% of positives are strictly rodent carcinogens
• i.e, has some positive predictive value
• However, rate of false positives poorly understood
Limitations of the assay limit the utility
• Typically used to inform label, informed consent
• Rarely will regulatory agencies use this data in isolation for decision-making
©2012, Genentech
What About Biologics?
Rarely are long term studies in rodents feasible for biologics due to
antigenicity concerns or lack of binding
Lack of validation data
– Limited data for nongenotoxic carcinogens
– Virtually no validation efforts with large molecules (e.g., Tg-AC
transgenic model)
– Known lack of concordance for immunosuppressive agents, many
hormones
Surrogate molecules: Discouraged given difficulties in verifying that
surrogate accurately reflects the biology of clinical candidate
Other data:
– Data from in class or related drugs
– Transgenic/ko
– Xenograft studies
©2012, Genentech
Limitations in data
interpretation and lack of
validation limit utility
Challenges with Nongenotoxic Drugs:
What We’ve Learned From Risk Evaluation
of Immunosuppressive Agents
©2012, Genentech
Human Neoplasms Associated with 13 Immunosuppressive Drugs
Type
Drugs
Lymphoma
Dexamethasone, prednisone, busulfan, azathioprine,
methotrexate, mycophenolate, cyclosporin A, tacrolimus
Squamous Cell Carcinoma
Prednisone, busulfan, azathioprine, methotrexate,
mycophenolate, cyclosporin A, tacrolimus
Kaposi Sarcoma
Dexamethasone, prednisone, busulfan, azathioprine,
methotrexate, mycophenolate, cyclosporin A
Urologic
Prednisone, azathioprine, mycophenolate, cyclosporin A,
tacrolimus
Melanoma
mycophenolate
Multiple
cyclophosphamide
Source: Bugelski et al.(2010) Int. J. Toxicol. 29(5) 435-66
©2012, Genentech
Two year bioassay results for immunosuppresive drugs
Drug
Rat 2 yr
Mouse 2 yr
Abatacept
-
Lymphomas and mammary tumors
(MLV/MMT virus)
Dexamethasone
Neg
-
Prednisone
-
Neg
Busulfan
-
Thymic and ovarian
Cyclophosphamide
multiple
multiple
Azathioprine
Lymphoma and squamous
Lymphoma,
hemangioendothelioma
Leflunomide
Neg
Lymphoma and lung
Methotrexate
Neg
Neg
Mycophenolate
Neg
Neg
Cyclosporine
Neg
Neg
Tacrolimus
Neg
Neg
sirolimus
Neg
Lymphoma and liver
everolimus
Neg
Neg
• Of the 5 positives, 4 are known genotoxicants
• Poor concordance with known human risks
• Only 2 correctly predict specific tumor risks
©2012, Genentech
ICH Guidance for Biologics (Original ICH S6, 1997)
Standard carc bioassays are generally inappropriate for biotech drugs
When there is a concern, “a variety of approaches may be considered to
evaluate risk”
In case where product is biologically active and nonimmunogenic and
other studies have not provided sufficient information to assess risk,
then consider a singe rodent bioassay
©2012, Genentech
For Products that Support or Induce Proliferation (ICH S6, 1997)
Evaluate/review receptor expression in malignant and normal cells
Is there evidence that can stimulate growth of normal or malignant
cells?
Cause for concern?
Yes
• Further studies in relevant model
• Incorporate sensitive indices of
proliferation into chronic studies
• If biologically active and
nonimmunogenic consider long
term assay
©2012, Genentech
No
• In vitro evaluations may be
sufficient
Question: Given the limitations of chronic bioassays and ICH
guidance, what type of carc studies have been conducted for biologics?
Answer: In reality, very few chronic studies have been conducted that
have actually impacted product labels
©2012, Genentech
32 FDA-Licensed MAb’s to date: Two sponsors have
conducted preclinical studies that impacted label
Generic Name
Muromonab-CD3
Abcixumab
Daclizumab
Rituximab
Basiliximab
Infliximab
Palivizumab
Trastuzumab
Gemtuzumab
Alemtuzumab
Adalimumab
Efalizumab
Ibritumomab
Tositumomab
Bevacizumab
Cetuximab
Omalizumab
Natalizumab
Panitumumab
Ranibizumab
Eculizumab
Certolizumab pegol
Canakinumab
Golimumab
Year Licensed by
FDA
1986
1994
1997
1997
1998
1998
1998
1998
2000
2001
2002
2002
2002
2003
2004
2004
2004
2006
2006
2006
2007
2008
2009
2009
Target
T cell CD3 Receptor
inhibition of glycoprotein IIb/IIIa
IL-2Rα receptor (CD25)
CD20
IL-2Rα receptor (CD25)
inhibition of TNF-α signaling
an epitope of the RSV F protein
ErbB2
CD33
CD52
inhibition of TNF-α signaling
CD11a
CD20
CD20
VEGF
epidermal growth factor receptor
immunoglobulin E (IgE)
CD20
RANK Ligand inhibitor
LT Carc studies on label?
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
xenograft models of a4+ cell
lines: no drug related impact
no
no
no
no
no
no
Decreased host defense to
tumors with surrogate; knockout
mice susceptible to tumors
[Data not from Sponsor-run
trials]
7 month cyno tumorigenicity
data included in label - no drugrelated effects noted
no
Anti- IL-6R
inihibition of B- cell activating factor
CD30
no
no
no
blocks CTLA-4
no
alpha-4 (α4) integrin,
epidermal growth factor receptor
VEGF-A
Complement system protein C5
inhibition of TNF-α signaling
IL-1β
TNF-alpha inihibitor
2009
Ustekinumab
anti-IL12/IL23
2009
Ofatumumab
Denosumab
Tocilizumab
(Atlizumab )
Belimumab
Brentuximab vedotin
Ipilimumab ( MDX101 )
©2012, Genentech
2010
2010
2011
2011
2011
• Two Sponsors
conducted trials that
impacted label
• One label impacted
by published
literature reports
(ustekinumab)
Label Claims for Non-MAb Therapeutics
Label Claim
2 yr studies in one or two species
glargine (insulin analog), ocreotide (somatostatin analog), teriparatide
(parathyroid hormone analog), IGF-1, gonadotropin releasing hormone,
exenatide, liraglutide (incretin mimetic), pulmozyme (rhDNAse), abatacept
Two year carc studies have not been
performed…
But tumors ID’ed in chronic tox
studies
aspart, glusine, and lispro (insulin analogs), calcitonin, pamlinitide (amylin
hormone analog),
Hypothical risk stated in label
asparaginase (alkylating agent)
Two year carc studies in animals have
not been performed….
humulin, novolin, lente, ultralente, Exubera, detemir, genotropin,
humantrope, norditropin, norIVitropin, nutropin, omnitrope, protropin, siazen,
serostim, valtropin, iplex, bioclate, helixate, kogenate, recombinate, reFacto,
BeneFIX, ceprotin, aldurazyme, elaprase, naglazyme, fabrazyme, aralast,
prolastin, lactaid, arco-lase, cotazym, creon, donnazyme, pancrease,
viokase, zymase, adagen octagam, albumarc, albumin, albuminar, albuRx,
albutein, flexbumin, buminate, plasbumin, neupogen, neulasta, leukine,
neumega, Gonal-F, Follistim, ovidrel, luveris, infergen, roferon-A, Pegasys,
Intron A, Peg-Intron, Alferon N, Avonex, rebif, betaseron, actimmune,
proleukin, activase, retavase, TNKase, abbokinase, NovoSeven, Xigris,
kepivance, regranex, granulex, natrecor, botox, myoblock collagenase,
xiaflex, santyl, amphadase, hydase, vitrase, hylenex, oncaspar, elitek,
refludan, angiomax, streptase, eminase, antril, kinaret, thioglobulin, fuzeon,
somavert, crofab, digifab, ontak,
©2012, Genentech
New ICH S6 Revision (ICH S6 R1, June 2011)
To better inform risk, a new paradigm has been
implemented by ICH
When an assessment is warranted (i.e., chronic dosing,
appropriate patient population, etc.) a weight of
evidence approach is now advocated
More emphasis on post-marketing surveillance
©2012, Genentech
What Can this Include?
Assessment of risk based on published literature and internal data
Clinical
–
–
–
–
–
Market surveillance
Human epidemiology
Genetic diseases
Polymorphisms
Class effects
Mechanistic data
• Is there impact on pathways known to be associated with malignancy risk
• Immunosuppression, chronic inflammation
• Downstream signaling through pathways associated with risk
Transgenic models
KO models
Animal disease models
Xenograft models
In vitro data
Chronic tox data
Alternative data (lifetime phenotyping, labeling for proliferation)
©2012, Genentech
Recommendations per ICH S6 R1, 2011
Outcome of Weight Based Assessment
Cause for concern
• Hazard best
addressed by product
labeling and risk
management practices
• Sponsor can propose
additional studies to
mitigate concern
©2012, Genentech
Risk unclear
• Consider studies as
discussed in ICH S6,
1997)
Risk considered low
• Additional rodent
bioassays not
warranted
Example: PCSK9 Inhibitor Class
FDA has provided guidance to all sponsors that are targeting PCSK9
(LDL-c lowering therapies)
Recommends a “Thorough Carcinogenicity Assessment” as
described in ICH S6 (R1)
– Requests that it is submitted early in development program (e.g., EOP2)
– Includes formal evaluation of immunosuppressive potential
(recommends 12 week study in cynos in combo w/ statin)
– Specific interest in NK cell activity, CD8+ T cell cytolytic activity
– Includes evaluation of impact on bile acid synthesis
– “…evidence of immunosuppression and/or a sustained increase in bile
acid secretion and/or intestinal bile acid load would be disclosed in the
label as potential cancer risks”
©2012, Genentech
Example: Studies to Mitigate Cause for Concern
GLP-1 analogs and C cell tumors
• Rodent bioassays identified increases in C-cell tumors
• Follow-up in vitro studies evaluated GLP-1 expression in rodent, monkey and human C
cells
• GLP-1 expression was much lower in humans, monkeys relative to rodents
• GLP-1 agonists stimulated measurable C-cell calcitonin release in rodents but not
human or monkey cells
• Calcitonin levels evaluated in 5000 patients treated for up to two years with no
evidence of increase
• Longitudinal studies have not identified causal association between GLP-1 analogs
and C cell pathology
• However, FDA AERS database supports a potential risk of thyroid cancer with
exenatide
Current label (liraglutide):
 In mice … a dose-related increase in benign thyroid C-cell adenomas was seen…
 In rats … a treatment-related increase in benign thyroid C-cell adenomas was seen…
 Human relevance of thyroid C-cell tumors in mice and rats is unknown and could not
be determined by clinical studies or nonclinical studies
©2012, Genentech
Example: Class Effect Labeling
13.1 Carcinogenesis, Mutagenesis, Impairment of Fertility
A carcinogenicity study was not conducted with belatacept. However, a murine carcinogenicity study was
conducted with abatacept (a more active analog in rodents) to determine the carcinogenic potential of CD28
blockade. Weekly subcutaneous injections of 20, 65, or 200 mg per kg of abatacept were associated with
©2012, Genentech
increases
in the incidence of malignant lymphomas (all doses) and mammary gland tumors….
Utility of General Tox Study Results
Predictivity of 6- or 12 month general tox studies for 2 yr bioassay (rats)
– Histology evaluation (+ = increase in hyperplasia, hypertrophy, and atypical cellular
foci (e.g., multinucleated cells, dysplasia, etc.)
– 2 yr rat bioassay (+ = increase in significant increase in tumors)
– 80 pharmacuticals evaluated (all FDA approved, sufficient rat data available for eval)
– 30 rat carcinogens, 50 noncarcinogens
Chronic Tox (rat)
Positive predictivity: 63%
Negative predictivity: 88%
False negatives: 6%
2 yr
Bioassay
+
-
+
25
5
-
15
35
Reddy, deGeorge, et al., 2010. Vet. Path. 47(4) 614-629
©2012, Genentech
Where is carcinogenic risk communicated currently in label?
Boxed warning
Section 5: Warnings and Precautions
– “Immunosuppression” or “Malignancies”
Section 6: Adverse Events
– “Malignancies”
Section13: Nonclinical Toxicology
13.1 …carcinogenesis
“…must state whether long term studies in animals have been performed to evaluate the
carcinogenic potential and, if so, the species and results”
“…any precautionary statement on these topics must include practical, relevant advice
to the prescriber on the significance of these animal findings.
Human data suggesting that the drug may be carcinogenic … as described in the
‘Warning and Precautions’ section, must not be included in this subsection of the
labeling.”
Section17: Patient Counseling Information
Source: Dan Mellon, FDA SOT 2012 Presentation.
©2012, Genentech
Posited Strategy for Labeling Revisions (Proposed for SOT
Discussion Only: Not Formal FDA Position)
©2012, Genentech
©2012, Genentech
©2012, Genentech
Bottom Line: Changes in risk communication in
not only product labels but informed consent
documents, investigator brochures, etc. are
anticipated but regulatory agencies have yet to
address what these changes will look like
©2012, Genentech
Summary
• Historically, classical lifetime rodent bioassays have had limited utility
for malignancy risk assessment for biologics and have had little
impact on informing product labels
• ICH has implemented new paradigm: Weight based assessment
that incorporates clinical, preclinical and mechanistic data
• It remains to be seen how these risk assessments will be
communicated in product labels
©2012, Genentech
Page 28
Thank You
Slide Credits:
Dan Mellon(FDA)
Heather Taylor (Genentech)
© 2009, Genentech / Proprietary information – Please do not copy, distribute or use without prior written consent.
©2012, Genentech