FIFRA Scientific Advisory
Panel to Consider and Review
RNA Interference:
Problem Formulation for
Human Health and Ecological
Risk Assessment
Monsanto Oral Public Comments
Monsanto Company
Creve Coeur, MO
January 28, 2014
Docket Number: EPA-HQ-OPP-2013-0485
Food and Feed
Safety of dsRNA
Jay S. Petrick, Ph.D., DABT
Senior Toxicologist,
Associate Fellow
Toxicology and Nutrition Center
Monsanto Company
Creve Coeur, MO
January 28, 2014
Biological Barriers Protect Against Ingested RNA
Systemic Barriers
Saliva
Stomach acid
Pancreatic nucleases
• High capacity
• Multi-enzyme
Intestinal epithelium
Vascular endothelium
Blood/systemic nucleases
Cellular Barriers
Plasma membrane barrier
Endosomal sequestration
Lysosomal degradation
3
Safety of Exogenous dsRNA in Higher Organisms
dsRNA has been
consumed safely
over millennia

After decades of
research, oral
RNA/DNA
therapeutics
remain an elusive
goal

Small RNAs and long dsRNAs with identity to human (and
animal) transcripts are safely consumed in staple crops1,2
 RNAi is not new to agriculture: underlies domesticated crop
phenotypes and traits in approved biotech crops
Very low oral bioavailability for oligo therapeutics (<1%)3
 Direct injection, formulation, and stabilizing modifications
needed for systemic activity3
 siRNA drugs are extensively metabolized; half-life of ~5 mins;
cleared by kidney within mins of i.v. dose4,5.
 RNA drugs have been safely administered at doses of up to
200 mg/kg i.v. in rats5
Zhang et al (2012): novel report of activity and significant
uptake of ingested miRNAs in mammals6
 Dickinson et al (2013): Zhang’s findings resulted from
nutritional differences, not ingested miRNAs7
 Witwer et al (2013), Snow et al (2013) and Dickinson et al
(2013) together show negligible small RNA uptake from the
diet in rodents, bees, primates, and humans7,8,9

Dietary RNAs:
Not active and
negligible uptake
4
Bioinformatics: Not a Safety Assessment Tool
Bioinformatics assessment of dsRNA vs human transcriptome
is not a safety assessment, only a predictor of putative
sequence matches
Matches themselves are not a safety hazard (safely consumed)1,2;
1000’s of matches in rice to human genes1; very low exposures
 Bioinformatics is over-predictive; 100% ID does not guarantee
effectiveness in vitro
• Number of “matches” identified is a function of search criteria
• Context of matches is key10: 1) where match occurs; 2) target
accessibility; 3) thermodynamic criteria for hybridization;
4) multiple matches needed for off-target effects11

Off-target suppression has been characterized in transfected
cells at high doses; potency of suppression is orders of
magnitude lower than for “on-target” matches12
5
The Existing Regulatory Paradigm
is Appropriate for RNAi
The existing paradigm for biotech crops is robust and utilizes a number of
endpoints to evaluate product safety:

Compositional assessment and animal performance assessment
 Potential for toxicity and allergenicity of the inserted protein
• Nucleic acids are not orally toxic nor allergenic
 Agronomic, phenotypic, and environmental assessments
Digestive barriers to exogenous RNA make in vitro digestibility assessments
unnecessary for RNAi based products

Assays place undue weight on stomach and intestine, only two of
many barriers to exogenous dsRNA
Nucleic acids are considered GRAS by the US FDA (FDA, 1992):
“Any safety considerations would focus on the intended effects of the
anti-sense RNA.”
As stated by FSANZa: “The current case-by-case approach to GM food safety
assessment is sufficiently broad and flexible to addresses the safety of GM foods
developed using gene silencing techniques.”
ahttp://www.foodstandards.gov.au/consumer/gmfood/Pages/
Response-to-Heinemann-et-al-on-the-regulation-of-GM-crops-and-foods-developed-using-gene-silencing.aspx
6
Conclusions
There are many effective barriers to exogenous dsRNA
exposures
Exogenous dsRNA has a history of safe exposure in
higher organisms
Bioinformatics is not useful to inform safety assessment of
dsRNA in higher organisms
The existing regulatory paradigm is robust and
appropriate for evaluating safety of RNAi based products;
potential adverse effects should be evaluated by
traditional measures
7
References
1.
Ivashuta, et al., (2009) Food Chem. Tox. 47, 353-360.
2.
Jensen, et al., (2013) GM Crops and Food, 4, 1-8.
3.
Reviewed by Petrick et al., (2013) Reg. Tox. Pharm., 66, 167-176.
4.
Christensen, et al.,(2013) Drug Metab. Distr. 412, 1211-1219.
5.
Thompson et al., (2012) Nucleic Acid Ther. 22, 255-264.
6.
Zhang, et al., (2012) Cell Research, 22: 107-126.
7.
Dickinson, et al., (2013) Nature Biotech 31, 965-967.
8.
Witwer, et al., (2013) RNA Biology 10: 7, 1-7.
9.
Snow, et al., (2013) RNA Biology 10: 6, 1-10.
10. Reviewed in Liu et al., (2013) Genomics, 102, 215-222.
11. Broderick et al., (2011) RNA, 17, 1858-1869.
12. Vaishnaw et al., (2010) Silence, 1: 14.
8
Environmental
Exposure
Assessment for
RNA-based
Agricultural Products
David Carson, Ph.D.
Biotech Environmental Fate &
Microbiology Lead
Monsanto Company
Creve Coeur, MO
January 28, 2014
Outline
Present results from laboratory soil degradation
studies indicating that dsRNAs are unlikely to persist
or accumulate in the environment
Current tiered-testing approach used by EPA is
appropriate for assessing environmental exposure for
dsRNA PIPs and exogenously applied dsRNAs
10
Laboratory Studies Assess Soil Degradation
to Estimate Environmental Exposure
• RNA degradation or loss of functional bioactivity,
and degradation rate estimates
• Measure the length of the exposure period
(e.g., DT50 or DT90)
QuantiGene®
Environmental
Fate
Assessment
Insect Bioassay
(pesticidal dsRNAs)
11
Degradation Evaluated using Diverse and
Representative Agricultural Soils
Source
USDA
Textural
Class
% Sand/Silt/Clay
pH
% Organic
Matter
Missouri
Loamy
Sand
85/8/7
5.5
3.0
19/55/26
5.9
2.6
26/38/36
7.1
5.1
Illinois
North
Dakota

Silt
Loam
Clay
Loam
Soil - key receiving environment for dsRNAs from
PIPs and exogenously applied dsRNAs
12
Degradation of a Corn Rootworm (CRW)-active
dsRNA from Different Agricultural Soils
140
120
120
120
100
100
100
80
60
40
% of initial
140
80
60
40
80
60
40
20
20
20
0
0
0
60
140
120
100
80
60
40
20
0
0
0
12 24 36 48 60
Incubation Time (h)
Dubelman et al. 2014 (submitted for publication)
12
24
36
48
Incubation Time (h)
60
0
12
24
36
48
Incubation Time (h)
60
180
140
120
100
80
60
40
20
0
150
% of Initial
12
24
36
48
Incubation Time (h)
% of initial
% of initial
Clay Loam (ND)
140
0
SOIL+TISSUE
Silt Loam (IL)
% of initial
% of initial
SOIL
Sandy Loam (MO)
120
90
60
30
0
0
12 24 36 48 60
Incubation Time (h)
0
12 24 36 48 60
Incubation Time (h)
13
Degradation of a CRW-active dsRNA
Correlates with a Loss of Functional Activity
Percent of Initial
QuantiGene
Silt Loam (IL)
QuantiGene
100
120
Sandy Loam (MO)
90
60
30
0
0
SCR Bioassay
75
12 24 36 48
Incubation Time (h)
50
QuantiGene
25
160
0
0
12
24
36
Incubation Time (h)
Dubelman et al. 2014 (submitted for publication)
48
Percent of Initial
Percent of Initial
125
SCR Bioassay
SCR Bioassay
Clay Loam (ND)
120
80
40
0
0
12 24 36 48
Incubation Time (h)
14
Degradation Kinetics of a CRW-active dsRNA
are Largely Independent of Dose
Dubelman et al. 2014 (submitted for publication)
15
Summary of Results from Laboratory Studies
Results from a soil degradation study indicate that a CRWactive dsRNA degraded completely within ~2 days in soil
Degradation kinetics estimated for the 3 soils:

DT50 <30 hrs, DT90 <35 hrs
Degradation was largely independent of dose
Preliminary results for other dsRNA molecules support a
lack of persistence in soil, independent of MW (24bp,
160bp, 240bp, 968nt) and sequence
16
Assessing Environmental Exposure for
Exogenously Applied dsRNAs
We recommend EPA employ a tiered-testing approach, utilizing
information on key properties and environmental persistence of
the active ingredient
EPA should not presume de facto that Tier II testing be required
for exogenously applied dsRNAs
We recommend a tiered risk assessment approach, for example:
Use existing biopesticide or Bt soil degradation protocols
Evaluate the Technical Grade Active Ingredient (TGAI)
If formulation stabilized, test formulation and compare to TGAI
Evaluate need for Tier II testing
17
Conclusions
Results to date indicate PIP or exogenously applied
dsRNA is unlikely to persist in the environment, minimizing
potential exposure to non-target organisms
Environmental exposure for dsRNA can be adequately
assessed by generating soil degradation data utilizing
protocols developed for Bt and microbial pesticides
The existing tiered-testing approach employed by EPA is
appropriate for assessing environmental exposure for
dsRNA PIPs and exogenously applied dsRNAs
18
Assessing Ecological
Risk for RNA-based
Agricultural Products
Steven L. Levine, Ph.D.
Senior Science Fellow
Ecotoxicology & Risk Assessment
Monsanto Company
Creve Coeur, MO
January 28, 2014
RNAi Technology Has Potential for High
Taxonomic Specificity
The sequence specificity of RNAi allows:
 Targeted
suppression of essential gene(s) in pests
 Development of highly efficacious and highly selective
products that have a low likelihood to adversely impact NTOs
Whyard et al. (2009) showed:
 Using
the vATPase target that ingested dsRNAs can act as
selective insecticides
 Using the tubulin target and four Drosophila species, that
single species specificity of transfected dsRNA can be
achieved when targeting the 3’ UTR where no 19 to 21 nt
sequence length was shared among the species
Whyard et al., (2009) Insect Biochemistry and Molecular Biology 39:824-832
20
Taxonomic Specificity Informs Non-target
Species Selection and the ERA
Work done at Monsanto confirms the conclusion of Whyard et al.
(2009) that insecticidal dsRNAs can achieve high taxonomic
specificity
We used a hypothesis-based taxonomic approach to establish
the relationship between biological activity and taxonomic
relatedness with a CRW active RNA in diet feeding assays
(Bachman et al. 2013)
Results from testing 18 species from 10 families and 4 orders in
sub-chronic or chronic bioassays shows that activity with a CRW
active RNA is only evident at the subfamily level (next slide)
Characterization of the spectrum of activity informs the
assessment plan by aiding non-target species selection and can
narrow the scope of hazard testing
Bachman et al., (2013) Transgenic Research 22:1207-22
21
A CRW Active dsRNA is Highly Specific
Order
Duration
Endpoint
Diabrotica virgifera virgifera
12 d
S, G
Southern Corn Rootworm
Diabrotica undecimpunctata howardi
12 d
S, G
Colorado Potato Beetle
Leptinotarsa decemlineata
▄▄
12 d
S, G
Red Flour Beetle
Tribolium castaneum
▄▄
12 d
S, G
Pink-spotted Lady Beetle
Coleomegilla maculata
▄▄
21 d
S,G, E
Mexican Bean Beetle
Epilachna varivestis
▄▄
28 d
S,G,E
Ground Beetle
Poecilus chalcites
▄▄
35 d
S,G,E
European Corn Borer
Ostrinia nubilalis
▄▄
12 d
S,G
Fall armyworm
Spodoptera frugiperda
▄▄
8d
S,G
Corn earworm
Helicoverpa zea
▄▄
12 d
S,G
Silkworm
Bombyx mori
▄▄
14 d
S,G
Jewel Wasp
Nasonia vitripennis
▄▄
20 d
S
Eulophid Wasp
Pediobius foveolatus
▄▄
21 d
S
Honey Bee adult
Apis mellifera
▄▄
14 d
S, B
Honey Bee larvae
Apis mellifera
▄▄
17 d
S, G, E
Hemiptera
Insidious Flower Bug
Orius insidiosus
▄▄
20 d
S, G
Collembola
Springtail
Folsomia candida
▄▄
28 d
S, R
Haplotaxida
Earthworm
Eisenia andrei
▄▄
14 d
S, G
Hymenoptera Lepidoptera
Coleoptera
Common Name
Species
Western Corn Rootworm
Activity
S= Survival; G= Growth; E = Emergence; B= Behavior; R = Reproduction
22
Factors to Consider when Assessing Risk to
NTOs from dsRNA during Problem Formulation
Exposure
Relative
Sensitivity

Barriers to exposure such as endonucleases in saliva (Lygus, aphids), midgut
fluids, and the hemolymph

Magnitude, duration and temporal nature of exposure after application

For PIPs, tissue-specific expression that eliminates routes of exposure

Not all taxa possess RNAi machinery and demonstrate quantifiable responses
to environmental exposure to dsRNA and there is a large range of sensitivities
across taxa

Sequence information can ONLY indicate whether an insect is potentially
sensitive but could be used to exclude organisms from testing

Bioinformatics can be used to address specific questions that reduce
uncertainty in the hazard assessment, but its application is limited due to
barriers, sensitivity and exposure; therefore sequence cannot be used as a
standalone to predict hazard to NTOs

Our research has demonstrated that ≥21 nt contiguous sequence embedded
in a ≥60 nt dsRNA is required for oral activity in a highly sensitive insect
species (CRW)
Sequence
Match
Bolognesi et al., (2012) PlosOne 7(10): e47534
23
Off-Target Effect Concerns from in vitro
Pharmaceutical Research Have Low Potential for
Adverse Effects with Environmental Exposure
Pharma literature using extremely high concentrations and invasive
techniques (injection, transfection) have created artifacts such as
immune stimulation and saturation of RNAi machinery
These effects are extremely unlikely to result from the low
concentrations of RNA applied for agricultural uses; organisms have a
history of safe exposure due to the ubiquitous presence of RNA in
natural food sources
For transitivity to occur the organism MUST have the following:
• RNA dependent RNA polymerase (not documented in insects)
• Sequence identity to an RNA
 Therefore, transitivity is highly unlikely to be a significant risk to
NTOs when not observed in target organism
24
EPA’s Tiered Testing and Assessment
Framework Provides Sufficient Data for ERA
The design, measurement endpoints and selection of Tier 1
hazard studies:

Should take into consideration the Mode of Action, taxonomic
specificity, routes and levels of exposure

Can be tailored in Problem Formulation to provide reliable and
relevant information for PIP and exogenously applied RNA
assessments
Tier 1 studies not only assess mortality but typically include
sublethal endpoints (e.g., growth and development)
Currently, EPA PIP requirements include developmental
milestones for NTOs and study designs for spray applications
could be further addressed on a case-by-case basis
25
Conclusions
RNA-based technology has high potential for taxonomic specificity and
low likelihood of adverse effects to NTOs
Key factors that drive potential ecological risk are exposure, relative
sensitivity, and sequence
Bioinformatics cannot be used as standalone information for an
ecological assessment
There is extremely low potential for off-target effects from saturation of
RNAi machinery, immuno and suppression transitivity in NTOs
Ecological testing for RNA-based products can be adequately
assessed within the existing framework using properly designed Tier 1
assays to evaluate endpoints relevant to EPA’s protection goals
26