ArticlePDF Available

Pathway to Deployment of Gene Drive Mosquitoes as a Potential Biocontrol Tool for Elimination of Malaria in Sub-Saharan Africa: Recommendations of a Scientific Working Group †


Abstract and Figures

Gene drive technology offers the promise for a high-impact, cost-effective, and durable method to control malaria transmission that would make a significant contribution to elimination. Gene drive systems, such as those based on clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR associated protein, have the potential to spread beneficial traits through interbreeding populations of malaria mosquitoes. However, the characteristics of this technology have raised concerns that necessitate careful consideration of the product development pathway. A multidisciplinary working group considered the implications of low-threshold gene drive systems on the development pathway described in the World Health Organization Guidance Framework for testing genetically modified (GM) mosquitoes, focusing on reduction of malaria transmission by Anopheles gambiae s.l. mosquitoes in Africa as a case study. The group developed recommendations for the safe and ethical testing of gene drive mosquitoes, drawing on prior experience with other vector control tools, GM organisms, and biocontrol agents. These recommendations are organized according to a testing plan that seeks to maximize safety by incrementally increasing the degree of human and environmental exposure to the investigational product. As with biocontrol agents, emphasis is placed on safety evaluation at the end of physically confined laboratory testing as a major decision point for whether to enter field testing. Progression through the testing pathway is based on fulfillment of safety and efficacy criteria, and is subject to regulatory and ethical approvals, as well as social acceptance. The working group identified several resources that were considered important to support responsible field testing of gene drive mosquitoes.
Content may be subject to copyright.
Am. J. Trop. Med. Hyg., 98(Suppl 6), 2018, pp. 149
Copyright © 2018 by The American Society of Tropical Medicine and Hygiene
Pathway to Deployment of Gene Drive Mosquitoes as a Potential Biocontrol Tool for Elimination of
Malaria in Sub-Saharan Africa: Recommendations of a Scientic Working Group
Stephanie James,
* Frank H. Collins,
Philip A. Welkhoff,
Claudia Emerson,
H. Charles J. Godfray,
Michael Gottlieb,
Brian Greenwood,
Steve W. Lindsay,
Charles M. Mbogo,
Fredros O. Okumu,
Hector Quemada,
Moussa Savadogo,
Jerome A. Singh,
Karen H. Tountas,
and Yeya T. Tour ´e
Foundation for the National Institutes of Health, Bethesda, Maryland;
University of Notre Dame, Notre Dame, Indiana;
Institute for Disease
Modeling, Bellevue, Washington;
McMaster University, Hamilton, Canada;
Oxford University, Oxford, United Kingdom;
London School of
Hygiene & Tropical Medicine, London, United Kingdom;
Durham University, Durham, United Kingdom;
Kenya Medical Research Institute,
Nairobi, Kenya;
Ifakara Health Institute, Ifakara, Tanzania;
University of Glasgow, Glasgow, Scotland;
University of the Witwatersrand,
Johannesburg, South Africa;
Donald Danforth Plant Science Center, Saint Louis, Missouri;
New Partnership for Africas Development,
Ouagadougou, Burkina Faso;
Centre for the AIDS Programme of Research in South Africa, Durban, KwaZulu-Natal, South Africa;
University of
Sciences, Techniques and Technologies of Bamako, Bamako, Mali
Abstract. Gene drive technology offers the promise for a high-impact, cost-effective, and durable method to control
malaria transmission that would make a signicant contribution to elimination. Gene drive systems, such as those based
on clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR associated protein, have the potential to
spread benecial traits through interbreeding populations of malaria mosquitoes. However, the characteristics of this
technology have raised concerns that necessitate careful consideration of the product development pathway. A multi-
disciplinary working group considered the implications of low-threshold gene drive systems on the development pathway
described in the World Health Organization Guidance Framework for testing genetically modied (GM) mosquitoes,
focusing on reduction of malaria transmission by Anopheles gambiae s.l. mosquitoes in Africa as a case study. The group
developed recommendations for the safe and ethical testing of gene drive mosquitoes, drawing on prior experience with
other vector control tools, GM organisms, and biocontrol agents. These recommendations are organized according to a
testing plan that seeks to maximize safety by incrementally increasing the degree of human and environmental exposure
to the investigational product. As with biocontrol agents, emphasis is placed on safety evaluation at the end of physically
conned laboratory testing as a major decision point for whether to enter eld testing. Progression through the testing
pathway is based on fulllment of safety and efcacy criteria, and is subject to regulatory and ethical approvals, as well as
social acceptance. The working group identied several resources that were considered important to support responsible
eld testing of gene drive mosquitoes.
Mosquitoes modied with gene drive systems are being
proposed as new tools that will complement current prac-
tices aimed at reducing or preventing transmission of
vector-borne diseases such as malaria. Gene drive systems
have the potential to spread new genetic traits through in-
terbreeding populations of malaria mosquitoes from low
initial introductions (Figure 1), and the transgenic construct
could persist in those mosquitoes indenitely or until the
target mosquito population is locally eliminated. Having
observed naturally occurring drive mechanisms in insects
and other organisms, scientists speculated for decades
about how these mechanisms could be harnessed to in-
sert benecial traits into a population of vector mosquitoes
to create a high-impact, low-cost, sustainable tool for
controlling disease transmission.
With the advent of new
molecular tools for modifying mosquitoes,
was envisioned to use synthetic genes with the capability
of spreading in populations, even if they confer a tness
cost (driving transgenes). The envisioned goal for apply-
ing this technology is to reduce or eliminate vector mos-
quito populations or, alternatively, to render them less
competent to transmit pathogens. Either of these out-
comes should contribute to disease reduction. However,
the characteristics that make gene drive technology so
attractive as a cost-effective and durable vector control
tool raise questions about possible adverse effects on
human or animal health or the environment that must be
seriously considered in product development.
Several mechanisms are being examined to achieve gene
Until recently, the attempted methods either did not
work in mosquitoes or were difcult to engineer
; however,
discovery of the clustered regularly interspaced short palin-
dromic repeats (CRISPR)/CRISPR associated protein 9 (Cas9)
system for gene editing has provided a widely accessible and
versatile molecular tool for creating driving transgenes.
use of CRISPR/Cas9 in mosquitoes follows from an idea,
conceived in 2003, that naturally occurring genes producing
homing endonuclease enzymes that target and cut specic
deoxyribonucleic acid (DNA) sequences could be used to
create gene drive.
Conceptually, constructs incorporating
the CRISPR/Cas9 system spread in the same way as these
natural endonuclease genes, but the easy manipulation of the
guide ribonucleic acid (RNA) that specically selects the site
where the chromosome cut occurs allows for the targeting of a
* Address correspondence to Stephanie James, Foundation for the
National Institutes of Health, 11400 Rockville Pike, Suite 600, North
Bethesda, MD 20852. E-mail:
These recommendations represent the collective effort of working
group members. They reect a consensus opinion, acknowledging
that individual participants had differing views on some issues. The
issues addressed here are complex and cover a wide range of
potential gene drive technologies. Therefore, these recommendations
will need to be interpreted in a case-by-case manner depending on the
nature of the investigational gene drive products and the environment
where they are to be tested. The working group members agree that
these recommendations provide important context and direction for
further planning.
greater range of gene sequences. Computational model-
ing based on other gene drive systems suggests that the type
of drive that can be achieved with the CRISPR/Cas9 system
can be so effective that release of low numbers of modied
mosquitoes into the environment could result in establish-
ment of the genetic modication in the natural interbreeding
population (P. A. Welkhoff, personal communication).
though still in the process of being optimized, such mosqui-
toes have already been developed in the laboratory with the
ultimate intent of testing in the eld.
Computer simulations
and population genetic analyses suggest that gene drive
strategies for reducing or modifying the population of vector
mosquitoes both have the potential to provide a transformative
new tool for conquering malaria and to make a valuable con-
tribution toward the elimination, and ultimate eradication, of this
In 2014, the World Health Organization (WHO) released the
Guidance Framework for testing genetically modied (GM)
mosquitoes (WHO Guidance Framework) that describes a
phased testing pathway and best practices for evaluating GM
mosquitoes (GMM) intended as public health tools.
proposed product development pathway moves from physi-
cally conned (also termed contained) studies in the labora-
tory and insectary (Phase 1) to small-scale physically and/or
ecologically conned eld-testing (Phase 2). Early small-scale
releases in Phase 2 are intended to allow observation of the
behavior of GMM in natural environments, and thus assess-
ment of entomological endpoints of efcacy, under conditions
that would minimize risk to the environment and/or human
health. Contingent on satisfactory results from conned
testing, the WHO Guidance Framework advocates pro-
ceeding to a series of staged open release trials of increasing
size, duration, and complexity (Phase 3).
These trials assess
performance under various conditions, such as different levels
of pathogen transmission, seasonal variations in mosquito
density, or presence of other disease vectors in the region.
Larger trials in this phase allow measurement of the impact of
GMM on infection and/or disease in human populations, in
addition to entomological endpoints. Following successful
completion of Phase 3, national authorities will determine
whether to move a specic GMM product into application as a
malaria control tool (Phase 4), which would include ongoing
surveillance of the effectiveness of the product under opera-
tional conditions, accompanied by monitoring of safety under
diverse conditions of use.
New low-thresholdgene drive technologies, such as those
using the CRISPR/Cas9 system, have broad implications
at multiple phases of the recommended WHO Guidance
Framework testing pathway because of the potential to be self-
that is, to spread a genetic modication through the
local mosquito population, and for that modication to become
established and to persist. Recognizing both the benets and
risks accompanying this new technology, there have been calls
for additional guidance and oversight before any eld-testing
The recommendations provided here represent the
response of a multidisciplinary working group that comprised
international experts in mosquito research (including, but not
limited to, molecular entomologists and individuals with eld ex-
perience in vector ecology and control), as well as experts in
containment/quarantine of exotic arthropods, mathematical
modeling, epidemiology, clinical trial design, statistics, ethics,
regulatory science, and policy (Box 1). Working group mem-
bers considered whether mosquitoes modied with low-
threshold gene drive could be developed appropriately and
used against malaria, and, if so, the resources and activities
needed to ensure their safe and efcient eld-testing and
implementation. These consensus recommendations build
primarily on existing guidance provided by the WHO Guidance
but also take into account recommendations
from the report of the National Academies of Science, Engi-
neering, and Medicine (NASEM) Gene Drives on the Horizon:
Advancing Science, Navigating Uncertainty, and Aligning
Research with Public Values (NASEM report),
which con-
sidered the broader public health, conservation, and agricul-
tural potential of gene drive technology, as well as widely
accepted guiding principles for sponsors and supporters of
gene drive research.
The recommendations presented here
attempt to envision the entire development pathway for gene
drive mosquitoes, from discovery research to implementation,
to provide a basis for establishment of standards of best
practice before the initiation of any eld trials. Like the WHO
Guidance Framework
and NASEM report,
they are inten-
ded to inform decision-making by researchers, funders, reg-
ulators, and policy-makers. It is anticipated that these
recommendations will be revised and rened as more expe-
rience with gene drive technologies is accumulated.
The investigational product for these recommendations is
considered to be the transgenic mosquito carrying a low-
threshold gene drive system (for convenience, herein referred to
as a gene drive mosquito). Over the course of three face-to-face
meetings, with ongoing discussions between each meeting, the
working group systematically examined how utilizing low-
threshold gene drive might inuence the planning and conduct
of each testing phase described in the WHO Guidance Frame-
This report does not attempt to summarize the detailed
FIGURE 1. Spread of novel traits by gene drive vs. Mendelian
The term thresholdrefers to the proportion of modied mosquitoes
with respect to the total mosquito population that will reliably initiate
spread of the modication to high levels within the local mosquito
population by mating. The goal of gene drive is to rapidly increase the
proportion of vector mosquitoes carrying the benecial modication.
Low-threshold gene drives are dened here to include those that are
predicted to spread from a rare introduction (zero threshold) or low
initial release frequency.
information contained within the WHO Guidance Framework,
which was accepted by the working group as the foundation for
the additional considerations related here. Readers are encour-
aged to consult the WHO Guidance Framework for underlying
information on efcacy and biosafety testing, ethics, public en-
gagement, and regulatory issues relevant to GMM.
working group envisioned that these recommendations will be
used as a companion to the earlier WHO guidance.
Scope and rationale. To focus the discussions, the work-
ing group concentrated on the example of malaria trans-
mission in Africa by mosquitoes of the Anopheles gambiae
(see Box 2).
It was assumed that the transformation event would be
performed in An. gambiae s.s. and later transferred to sibling
species by introgression in the laboratory or by natural hy-
bridization in the eld.
Anopheles gambiae s.l. mosquitoes are reported only on the
African continent.
This geographic limitation is an important
consideration in evaluating the potential spread of gene drive
approaches targeting these mosquitoes. Although An. gam-
biae s.s. and sibling species An. coluzzi and An. arabiensis are
major malaria vectors in sub-Saharan Africa, the working
group recognized that other Anopheles species (notably
Anopheles funestus) also transmit malaria, and may, in certain
situations, contribute a signicant proportion of the residual
Products directed at these mosquitoes also
will be required for malaria elimination.
Because of massive deployment of currently available
malaria control tools, Plasmodium falciparum infection prev-
alence in endemic Africa halved and the incidence of clinical
disease fell by 40% between 2000 and 2015.
Yet, residual
levels of transmission still persist even in places where cov-
erage with existing interventions is already very high.
According to the most recent World Malaria Report 2017,
despite best control efforts undertaken to date, there were 216
(95% condence interval = 196263) million cases of malaria
and an estimated 445,000 deaths from malaria in 2016, with
90% of cases and deaths occurring in sub-Saharan Africa and
with a leveling off in the recent decline in malaria mortality. In
2015, malaria killed an estimated 303,000 children under the
age of 5 years globally, and 96% of these deaths occurred in
the African region.
Although the African region has shown
considerable recent progress, malaria remains stubbornly
persistent in some areas and is increasing in others
; the
substantial progress that has been made is fragile and is
threatened by insecticide resistance,
changes in vector
behavior, resistance to antimalarial therapeutics, and high
ongoing costs of malaria control (estimated at over $6 billion
per year to meet the 2020 target for reduction in malaria
with over half the costs going toward
vector control).
Thus, control of malaria in Africa is arguably where the use of
self-sustaining gene drive mosquitoes could yield the greatest
public health benet, and, therefore, where their initial use
would be most justied. Although there are still many issues to
be resolved, initial indications are that low-threshold gene
drive technology, if optimized, has the potential to be readily
deployable across diverse geographical and socioeconomic
areas, including low-income communities and those with poor
access to health care, thus protecting millions of people and
achieving extremely high impact over relatively short periods
of time.
Because of these potential benets, NASEM and a
Working group composition
Core Working Group Members: participated in all working group activities and authored the recommendations
Frank H. Collins, University of Notre Dame; Philip A. Welkhoff,Institute for Disease Modeling; Claudia Emerson, McMaster University; H. Charles
J. Godfray, Oxford University; Brian Greenwood, London School of Hygiene & Tropical Medicine; Steve W. Lindsay, Durham University;Charles
M. Mbogo, Kenya Medical Research Institute; Fredros O. Okumu, Ifakara Health Institute, University of Glasgow, University of the
Witwatersrand; Hector Quemada, Donald Danforth Plant Science Center; Moussa Savadogo, New Partnership for Africas Development
(NEPAD); Jerome A. Singh, Center for the AIDS Program of Research in South Africa; Yeya T. Tour´e, University of Sciences, Techniques and
Technologies of Bamako
Ad Hoc Working Group Participants: attended specic working group meetings as appropriate to areas of expertise and provided comments
Aggrey Ambali, NEPAD; Mark Benedict, Foundation for the Centers for Disease Control and Prevention; Christophe Boete,* Institut pour
Recherche pour le D ´eveloppement; Catherine Bourgouin,* Institut Pasteur; Paul De Barro, The Commonwealth Scientic and Industrial
Research Organisation; Abdoulaye Diabate, Institut de Recherche en Science de la Sant ´e/Center Muraz; Azra Ghani, Imperial College London;
Fred Gould, North Carolina State University; Lee Hall, US National Institute of Allergy and Infectious Diseases; Steve Higgs, Kansas State
University; Immo Kleinschmidt, London School of Hygiene & Tropical Medicine; Greg Lanzaro, University of California, Davis; Christian
Lengeler, Swiss Tropical and Public Health Institute; Jo Lines, London School of Hygiene & Tropical Medicine; David Malone, Innovative Vector
Control Consortium; Kevin Marsh, University of Oxford; Leonard Mboera, National Institute for Medical Research Tanzania; Abraham Mnzava,
African Leaders Malaria Alliance; Scott ONeill, Monash University; Seth Owusu-Agyei, University of Health & Allied Sciences; Malla Rao, US
National Institute of Allergy and Infectious Diseases; Larry Slutsker, Program for Appropriate Technology in Health; Willy Tonui, National
Biosafety Authority (NBA) Kenya; Kenneth Vernick, Institut Pasteur
Contributors: provided written or verbal comments or information for working group consideration
Adam Bennett, University of California, San Francisco; Austin Burt, Imperial College, London; Nora Besansky, Notre Dame University; Lorna
Clark, Imperial College London; George Christophides, Imperial College London; Andrea Crisanti, Imperial College London; Anthony James,
University of California, Irvine; John Marshall, University of California, Berkeley; Tony Nolan, Imperial College, London; Nikolai Windbichler,
Imperial College, London
Observers: attended one of more working group meetings
Anne Cheever, Booz Allen Hamilton and Contractor support to the Defense Advanced Research Projects Agency; Adriana Costero-Saint Denis,
US National Institute of Allergy and Infectious Diseases; Anna Drexler, World Health Organization (WHO), Florence Fouque, WHO; Fil Randazzo,
Bill & Melinda Gates Foundation; Mike Reddy, Bill & Melinda Gates Foundation; Emmanuel Temu, WHO; Raman Velayudhan, WHO; Renee
Wegrzyn, Defense Advanced Research Projects Agency
* Invited at the recommendation of WHO observers
WHO expert advisory committee have encouraged continued
research on gene drive mosquitoes as a new tool to work
synergistically with other malaria interventions.
Although this working group considered only malaria
transmission by An. gambiae s.l. in Africa, it is expected that
the recommendations related here will have relevance to similar
research on other mosquito vectors of malaria, including those
prevalent in other regions, and on other disease vectors.
However, the testing pathway will need to be reconsidered
according to the specics of these other cases.
As dened in the WHO Guidance Framework, gene drive
approaches that are self-sustaining(sometimes termed
self-propagating) are intended to spread through the target
mosquito population.
The drive mechanism must be capa-
ble of overcoming any tness costs and capable of increasing
in frequency from low initial levels to xation, or near xation, in
the population into which it was introduced within a time frame
that will be meaningful for malaria elimination. Although other,
more limited, approaches are now being considered (see Self-
limiting alternatives), this denition remains valid for low-
threshold gene drive strategies that are the subject of these
There are two major categories of gene drive strategies
population suppression and population replacement§ (Figure 2).
Computer simulation indicates that both have the potential to
interrupt malaria transmission by the targeted mosquito spe-
cies even in the most challenging settings, and these studies
provide insight into how deployment methods and spatio-
temporal extents can be tailored to local conditions to over-
come obstacles such as extreme seasonality.
Population suppression strategies are intended to reduce
the size of the vector population to such an extent that it will
not be able to sustain malaria transmission. This is an exten-
sion of the goal of all current vector-control products and does
not require driving a population to extinction. Population
suppression strategies are based on inactivation, or knock-
out, of genes involved in the target mosquitos survival or re-
production (e.g., reducing fertility or production of female
progeny), and/or bias of the sex ratio toward males. These may
be termed loss of functiontechniques.
Population replacement strategies are intended to reduce the
inherent ability of individual mosquitoes to transmit the malaria
pathogen. These strategies may be built around inactivation of a
gene or genes that facilitate parasite survival in the mosquito
vector or that are required for the mosquito to transmit malaria,
such as a tendency to feed on humans. Population replacement
strategies based on inactivation of genes directly involved in
vectorial capacity are also termed loss of functiontech-
niques. Other population replacement strategies involve the
introduction of a new gene or genes, such as those that
produce effector molecules that will kill the malaria parasite
in the mosquito. To perform successfully, such introduced
genes must be carried into the mosquito genome in tight
linkage with the gene drive mechanism.
Mosquitoes modied with low-threshold gene drive con-
structs are expected to persist in the environment. Those
strategies aiming for population replacement require the
modication to persist at high levels for as long as malaria
continues to be transmitted to achieve their objective. For
those strategies aiming for population suppression, modied
mosquitoes are expected to decrease to low numbers over the
period of a few years as the overall population of target
mosquitoes is reduced. Phenotypic traits identied as relevant
to efcacy and/or safety should be observed in the laboratory
over multiple generations to obtain information on their sta-
bility. Because the anticipated mechanism of action and pe-
riod of environmental exposure will differ among various gene
drive strategies, researchers will be responsible for proposing
an adequate plan for demonstrating the durability of efcacy in
their regulatory applications. Modeling will provide a critical
The An. gambiae complex
The An. gambiae complex (also known as An. gambiae sensu lato [s.l., in the broad sense]), which includes some of the most important and
efcient vectors of malaria in sub-Saharan Africa, consists of eight named siblingspecies that are difcult to distinguish morphologically but can
be identied using molecular methods:
An. gambiae sensu stricto (s.s. in the strict sense)
Anopheles amharicus
Anopheles arabiensis
Anopheles bwambae
Anopheles coluzzii
Anopheles melas
Anopheles merus
Anopheles quadriannulatus
The individual species exhibit distinct behavioral and ecological preferences. As examples, An. gambiae s.s. and An. coluzzii, which are closely
related, feed almost exclusively on humans (anthropophilic), whereas An. quadriannulatus takes its blood meal from animals (zoophilic).
Anopheles melas and An. merus can breed in salt water, whereas An. gambiae and the other species breed in fresh water. Anopheles
quadriannulatus is not considered to be a malaria vector. Although these species are considered to be reproductively isolated, there is evidence
of interbreeding between some of them. Patterns of introgression between An. arabiensis and An. gambiae/An. coluzzii,and between An. merus
and An. quadriannulatus, are similar across their geographic range.
Consideration must, therefore, be given to the diversity of members of this complex present at eld testing sites and to whether the gene
sequence targeted by gene drive constructs is present in more than one species. Nontarget Anopheles species should be examined for the
extent of gene ow between sibling species and the potential effects of any genetic transfer events.
§ Population suppression is sometimes called population reduction.
Population replacement is sometimes termed population modication,
population alteration, population transformation, or population
conversion. The present article retains the terminology that was used
in the WHO Guidance Framework.
tool for determining the number of generations over which key
stability, efcacy, and safety characteristics must be moni-
tored in the laboratory to provide sound justication for pro-
ceeding to eld-testing.
Self-limiting alternatives. There are circumstances in
which consideration might be given to the testing of a self-
limiting approach.
Self-limiting constructs constitute a form
of biological or molecular connement, which would supple-
ment physical and ecological connement. The genetic sterile
insect technique is the most extreme self-limiting technology,
and several fertile but self-limiting or self-exhausting ap-
proaches also are possible. One such approach would use a
closely related but nondriving version of the proposed self-
sustaining construct, which is expected to be passed on in
diminishing proportion through subsequent generations
according to normal Mendelian inheritance until eventually
becoming highly diluted in the population through outcrossing
and lost if it conveys any tness cost. Other proposed ap-
proaches include genetic manipulations aiming to purposely
limit the time period or geographic range over which gene
drive is expected to remain functional.
Additional alter-
natives likely will continue to be conceived.
The focus of these recommendations is on developing gene
drive mosquitoes to contribute to elimination/eradication of
malaria across sub-Saharan Africa, a vast continent
malaria is largely present in rural regions and endemic in most
countries. To date, there are no modeling data to support the
possibility that any of the currently contemplated self-limiting
approaches might achieve an appreciable reduction of
malaria transmission across sub-Saharan Africa.kHowever,
the working group recognized three circumstances in which
testing of a self-limiting intermediate before moving to the eld
with a self-sustaining driving construct could be particularly
pertinent to the development pathway for gene drive mos-
quitoes to control malaria in Africa.
1. To provide additional data regarding the interaction of the
genetic construct with the environment for a rst-in-class
gene drive strategy, if deemed necessary to answer
questions raised in the risk assessment or to build con-
dence with regulators, communities, and other stake-
holders. For example, this might be useful to gain
multinational acceptance of a new technology. In this case,
the composition of the self-limiting construct with respect
to promoter, effector, and marker genes should be as
similar as possible to that of the self-sustaining construct to
maximize the relevance of information to be gained from
this approach.
2. To conduct eld testing for efcacy of population re-
placement strategies under conditions of lower risk. Initial
efcacy testing may be performed using a laboratory strain
of P. falciparum, and gametocyte-producing laboratory
strains of the parasite are available. However, before pro-
gressing to eld-testing of gene drive mosquitoes, it will be
important to determine whether the construct exhibits
predicted activity against locally transmitted parasite
strains. Because malaria gametocytes only remain viable
for a short period, such testing with local strains must be
performed in a malaria endemic region. Vector competence
for local parasites could initially be performed in-country
using a self-limiting intermediate.
3. To provide additional training and capability strengthening
for an unproven containment facility and/or inexperienced
staff before initiating work with a low-threshold self-
sustaining form of gene drive. Combined experience sug-
gests that many breaches of containment are associated
with human error due to failure to follow established pro-
cedures, emphasizing the importance of training and ex-
perience for physically conned studies. Initial work with a
self-limiting strain will provide an opportunity for re-
searchers to evaluate system capacity and compliance
with standard operating procedures (SOPs) and un-
derstand regulatory requirements, under conditions of
decreased risk.
Researchers should consider the following factors when
making the decision on whether to include a self-limiting step
in the development pathwayfor a self-sustaininginvestigational
product. First, under most conditions found in sub-Saharan
Africa, it is questionable whether a self-limiting version could be
effective over a sufcient area and time frame to provide a cost-
effective and sustainable reduction of malaria transmission.
Thus, although short-term population reduction or replacement
effects may be measurable, the feasibility of obtaining mean-
ingful and cost-effective epidemiological efcacy, especially in
those regions most in need of additional controltools to achieve
malaria elimination, should be examined. Second, no matter
how similar the self-limiting intermediate isto the intended self-
sustaining version it cannot be expected to have exactly the
same environmental interactions and implications because of
the intentionally limited level and period of exposure. Finally,
eld-testing of a self-limiting intermediate will require a
sizable expenditure of time and resources, which may result
in an appreciable delay in the availability of a self-sustaining
FIGURE 2. Comparison of population replacement (A) and pop-
ulation suppression (B) strategies.
kThe working group acknowledged that self-limiting or self-
exhausting drive approaches might be useful in restricted locations,
for other diseases or under other transmission conditions, and
recognized the importance of modeling for making this determination.
gene drive mosquito product for use against malaria. On the
positive side, although a self-limiting strain may not in itself
be an effective tool against malaria transmission in sub-
Saharan Africa, it may enable researchers to gain useful
experience and information that will increase the likelihood
of success when the self-sustaining version is released.
Testing of a self-limiting strain can also help to build relevant
regulatory experience, allowing in-country regulators an
opportunity to consider country-specic risk questions and
Careful consideration must be given on a case-by-case
basis to determine what information can reasonably be ac-
quired from testing a self-limiting intermediate, how vital this
information is to decision-making, and how extensive such
testing must be to obtain the necessary information. It is
possible that regulators and policy makers may desire an in-
termediate step in the testing pathway. Researchers must be
prepared to explain the advantages and disadvantages of
such a step.
Self-limiting alternatives will be subject to relevant risk as-
sessment and regulatory requirements for importation and
use of GM organisms (GMO). Beyond that, because of the
diversity of potential self-limiting approaches, it is possible
that some connement and release requirements described
here for low-threshold gene drive may not be applicable.
These should be determined on a case-by-case basis
according to the nature of the construct.
Follow-on products. Given the current state of develop-
ment, the recommendations provided here focus primarily on
rst-in-class applications of low-threshold gene drive tech-
nology. However, it is anticipated that this research will not
end after eld testing of the rst gene drive mosquito product.
For purposes of these recommendations, the working
group assumed that initial products will be targeted at An.
gambiae s.l. It is expected that the transgenic construct can be
transferred to other major vector species within this complex
through interbreeding in the laboratory. There will be a need,
however, to move the technology into other malaria vector
species, notably to the An. funestus complex which is also
important for malaria transmission in large regions of Africa
and is developing resistance to common insecticides.
probably will require a new transgenic event and, thus, likely
will be considered an independent investigational product.
Moreover, as has been observed repeatedly with insecti-
cides and antimalarial drugs, it can be expected that re-
sistance eventually will develop to rst generation products
through selection or evolution of variations in the targeted
genetic sequence in the mosquito or parasite.
nisms are being explored to delay the expression of resistance
(see Resistance). However, unless malaria has been eradicated
before this occurs, resistance is likely to generate a need to
identify effector mechanisms for population suppression or
replacement that target different gene sequences.
Finally, gene editing is a rapidly evolving eld of research,
and new advances may result in second generation products
with improved efcacy or other desirable features. Recog-
nizing the likelihood of ongoing advances and new ap-
proaches in gene drive technology, this working group
endeavored to avoid being too prescriptive in its recommen-
dations, aiming to provide advice that will be relevant to cur-
rent and future efforts to develop gene drive technology for
use in mosquitoes. Nonetheless, it should be understood that
some of the stated recommendations and requirements de-
scribed here might be changed for follow-on products after
uncertainties have decreased.
Many of the issues that must be considered for eld-testing
an investigational product will be common to all low-threshold
gene drive strategies and all along the continuum of the de-
velopment pathway. Product development will be more ef-
cient if planning for these issues begins early in the project (see
Box 3).
Product characteristics. The WHO Guidance Framework
identied two major issues to be addressed in the critical
path for development of GMM as public health tools: 1) proof
of efcacy, determined through testing for entomological
and epidemiological impact; and 2) evidence of acceptabil-
ity, determined through biosafety, ethics and engagement
activities, and compliance with regulatory requirements.
Similarly, these two issues are priorities for gene drive mos-
quitoes. Development of a target product prole (TPP) will
help researchers identify their specic goals in each of these
areas and facilitate decision-making about when an in-
vestigational product is ready to move further along the
testing pathway. The TPP is a tool that aids investigators to
begin their work with the ultimate goal in mind, focusing on
the specic claims of the envisioned product, and is fre-
quently recommended for product development.
product, including the advantages it will provide beyond
existing tools. Engagement activities should include con-
sulting the potentially exposed community and relevant
government authorities early in the process of TPP formu-
lation to understand what characteristics would make the
product attractive from their perspective. For example, this
engagement could begin during the period in which baseline
eld data are collected. The TPP will include parameters such
as efcacy, safety (including ecosystem impact), stability/
durability, and production and release characteristics
(discussed further in the following text). Establishment of
TPP criteria should be informed by modeling. The TPP can be
rened as additional eld experience and data are obtained.
However, even early in development, researchers should
think about such practical issues as how the potential gene
drive mosquito will be manufactured, distributed, and mon-
itored,asthismayinuence fundamental decisions, in-
cluding construct composition and whether to protect
intellectual property (see Implementation as a public health
tool). A cost analysis likely will be required before decision-
making on implementation, so researchers should consider
their goals for cost of eventual deployment.
Development pathway. The predicted properties of rapid
spread and persistence that make low-threshold gene drive
such an attractive tool for controlling malaria transmission
also complicate the ability to remove the modication once it
has become established in the local mosquito population
should demonstrable harms be observed. Unlike GM crops for
agriculture where the modication is not intended to spread,
low-threshold gene drive modications purposely are
designed to continue spreading after releases of the gene
drive mosquitoes are halted.
Similarly to the pathway described in the WHO Guidance
testing of gene drive mosquitoes is expected
to proceed through multiple phases or stages, each incre-
mentally increasing the degree of human and environmental
exposure to the investigational product. The transition along
the pathway will be subject to fulllment of efcacy and
safety criteria as dened in the TPP and evaluated in the
context of specically designed eld trials, as well as regu-
latory and ethical approvals, and social acceptance. How-
ever, the characteristics of gene drive may make it difcult or
even undesirable to delineate distinct cutoffs between pha-
ses in the testing pathway beyond initial studies under
physically conned laboratory and insectary conditions.
Thus, to avoid confusion, these working group recommen-
dations consider the goals and requirements of successive
major stages of eld testing as a continuum of expanding
releases, rather than as distinct or discontinuous phases
(Figure 3).
As is the case with biocontrol agents, which also are
expected to spread and persist in the environment and whose
releases may be difcult to reverse, emphasis is placed on
safety evaluation at the end of physically conned laboratory
testing (including testing in large indoor cages that simulate
the natural environment, if applicable) as a major decision
point for whether to enter eld testing. Initial transition to the
eld may begin with testing in a large outdoor cage (semi-eld
testing), although this was not considered an essential re-
quirement. The rst small eld release should strive for geo-
graphic isolation to limit environmental exposure to the extent
practicable as safety observation continues. Increasingly
larger scale open releases will allow for assessment of rst
entomological and then epidemiological efcacy of the in-
vestigational gene drive product. Acceptance as a public
health tool would initiate more systematic scale-up releases
and initiation of post-implementation surveillance for ongoing
efcacy and safety. Requirements for each of these phases
are described in detail in the following text.
In determining the minimal requirements that an in-
vestigational mosquito product should meet to justify moving
beyond the laboratory and along the development pathway,
safety concerns and potential benets to be gained from
further testing both must be weighed. Benet should be
considered not only from the perspective of the nal product,
which aims to provide cost-effective control of malaria
transmission, but also at each level of testing before expan-
sion of releases is allowed. Researchers must be prepared to
justify each study or trial to regulators and other decision-
makers in terms of how the information to be gained will con-
tribute to decision-making. Although safety will be assessed
carefully before proceeding to eld testing, ongoing consider-
ation of safety is necessary as a matter of due diligence.
Risk assessment. The predicted ease of spread of gene
drive mosquitoes calls for extremely thorough evaluation un-
der careful connement before release into a hospitable en-
vironment (i.e., conditions that could support the survival of
the mosquito). However, as emphasized in the WHO Guidance
it is important that safety expectations should
be proportionate to those for other vector control tools and
should take into account the risks associated with maintaining
the status quo.
Risk assessment will provide guidance on decision-making
for the project team, including information for preparation of
regulatory applications and development of risk mitigation
plans. This can identify additional questions that need further
research to fully assess risk. The WHO Guidance Framework
discussed risk assessment and risk management consider-
ations at each phase of testing.
Quantitative ecological risk
assessment was endorsed in the NASEM report as especially
useful for estimating the probability of specied outcomes.
In their consideration of synthetic gene drives in Australia, the
FIGURE 3. Pathway to deployment of gene drive mosquitoes.
Planning considerations for testing along the continuum of the development pathway
Establish a team with appropriate expertise and experience
Develop/rene the target product prole (TPP) and criteria for advancement
Conduct modeling to inform experimental study design and understand potential benets
Develop processes for information and data sharing to promote transparency
Establish partnerships as necessary
Characterize the eld testing site (ecology, vector, and clinical)
Undertake environmental risk and biosafety assessments
Address ethical issues
Plan for and conduct stakeholder engagement at multiple levels
Plan for and meet regulatory requirements
Design remediation/mitigation plans
Australian Academy of Science recommended that any
decision to release a synthetic gene drive be made on a case-
by-case basis following a comprehensive environmental risk
assessment which includes ecological and evolutionary
Others have recommended an integrated ap-
proach to risk assessment of gene drive technologies that in-
cludes the participation of ethicists
and biosafety professionals.
The risk assessment should be grounded in the protection
goals established by the countries that would host the testing
and/or use the technology.
However, it should cover not
only environmental and health risks, but also social and eco-
nomic risks.
There are challenges associated with weighing
risks that the research team identies as most signicant
against those of greatest concern to the lay public in risk as-
which will be especially true for gene drive, a
technology that is expected to cross national borders. Thus,
there will need to be a plan for how public input on hazards is
solicited and integrated. Principles for both environmental and
social impact assessment have been proposed.
The risk
assessment conducted for testing Wolbachia-infected mos-
quitoes for controlling transmission of other vector-borne
diseases provides an example,
but this would have to be
adapted to the context of gene drive applications in Africa.
An external risk assessment, conducted by qualied indi-
viduals with no vested interest in the success of the product,
can be valuable for building community, stakeholder, and
public condence. This will be particularly important for rst-
in-class gene drive strategies. The working group recom-
mended that researchers and/or funders commission an
external all-hazards risk assessment, to be conducted by
experts that are unafliated with the research project, and
that results of the risk assessment be made publicly avail-
able. Funders should be prepared to support the costs for
risk assessment as an integral part of the overall research
The risk assessment must be reexamined and updated
before moving forward along the testing pathway, to take into
account any changes in human or environmental exposure,
additional data, and any further public concerns.
researchers are encouraged to make such external assess-
ments available to regulators and the public, it should be
understood that these would not supersede the risk as-
sessments performed by the regulators in connection with
evaluating applications, in compliance with national regulatory
requirements and guidelines.
Decision-making. As described in the WHO Guidance
Framework, it is expected that decision-makers will take into
consideration criteria of both safety (risk) and efcacy (benet)
for a products intended use.
Benet will be perceived rel-
ative to the particular context, which in this case is the need for
malaria control in Africa.
Early on in the development of individual gene drive mos-
quito products, funders will need to make important deci-
sions about their commitment to move an investigational
product to the eld. This decision will be facilitated by
mathematical modeling of the predicted effects under re-
alistic transmission conditions. As the science continues to
evolve, there always will be the possibility of new products on
the horizon; however, anticipation of a better product needs
to be balanced with the potential life-saving benet(s) of a
current investigational product if support is provided to
complete the development and testing process. It can be
expected that eld-testing of gene drive mosquitoes will be
rigorous and expensive. The effort necessary to move an
investigational product forward through eld testing in de-
veloping countries in a responsible manner represents a
major commitment, as described in the following text. If an
investigational product meets mutually agreed TPP criteria of
safety and efcacy during contained laboratory testing, in-
dicating that it could have a signicant impact in reducing
malaria transmission in the setting in which release is con-
templated, a decision is made to move forward to eld testing
and releases begin; funders must be prepared to commit
sufcient resources to meet long-term obligations to the
researchers and to the countries where the testing will take
place to which they committed in the research plan. It is
advised that all involved institutions (including funders) de-
velop a joint research collaboration agreement in advance,
which makes each institutions obligations clear.
Modeling. Just as pharmacokinetic/pharmacodynamic
modeling is an essential component of developing and test-
ing a new drug, mathematical modeling has an important role
to play in each step of the testing pathway for gene drive
mosquitoes. However, the validity of the modeling will be
inuenced by the strength of the dataset it utilizes, including its
relevance to conditions at the trial site, and this underscores
the need to collect relevant baseline eld data and to make
it widely available to the research community (see Data
Even before developing constructs in the laboratory,
mathematical modeling can guide the specication of re-
quired properties of the construct, such as homing/drive rate,
effector strength, frequency of development of resistance,
and more. Mathematical modeling can also help to identify
baseline data required from potential eld sites before the rst
eld trials, which will assist in determining whether the sites
under consideration will be sufciently informative for the
proposed trial objectives.
Computational modeling can use performance character-
istics measured during early development (laboratory and
large cage testing) to predict possible outcomes in openeld-
testing for the investigational product before any actual eld
releases are performed. These types of model-based infer-
ences provide an important contribution to decision-making
about whether eld releases are justied.
Data collected from small and large-scale eld testing can
be used in computational models to help plan a resource-
optimized robust release strategy for wide-scale imple-
mentation that achieves the goals of disease control and
elimination efforts before scale-up begins. Risks of resistance
also can be explored in mathematical models to develop
sampling schemes to identify any occurrences in the course of
implementation with substantial potential effects on the dis-
ease. Modeling also may provide insights into the effective-
ness of proposed remediation strategies.
Transparency. The working group members noted that
development of gene drive technology carries an obligation for
transparency and accountability. This is important for earning
public condence, ensuring that the product meets stake-
holder needs, encouraging inter-project coordination neces-
sary for responsible eld testing, and minimizing any risks to
human health and/or the environment. Gene drive researchers
should commit to being appropriately transparent about their
With respect to public engagement, failure to be transparent
about data can heighten anxiety by creating the impression
that scientists know things they are not willing to reveal, and
this may fuel distrust. From the perspective of product de-
velopment, inappropriately conducted eld trials have the
potential to negatively impact the future success of other gene
drive products; to undermine community, stakeholder, and/or
public condence in the technology; and to contaminate the
regulatory and funding environment. Also, even though the
release of an ineffective gene drive construct in the context of
an efcacy trial may not appreciably alter mosquito function or
result in any direct biosafety threat, such a release might
create subtle genetic changes in the target mosquito pop-
ulation that could impact the effectiveness of subsequent in-
vestigational products and inuence the use of subsequent
new gene drive products at the study site. This could result in
loss of time and resources spent developing other gene drive
products and preparing eld sites, and possibly prevent
the sites from beneting from future products. At worst, ill-
conceived eld trials might cause damage to human or animal
health, or the environment. Thus, transparency should in-
clude, but is not necessarily limited to, keeping open and ac-
cessible records of any (accidental or intended) releases,
containing a full description of the investigational product.
Policies and mechanisms for inter-project coordination and
broader data and information sharing are a necessity. This
level of cooperation is best driven by research funders, as
exemplied by prior data sharing agreements.
nizing the importance of transparency for public condence
and future development of gene drive technology, the working
group recommended that funders work cooperatively on the
early establishment of policies for appropriate sharing of data
from gene drive research.
Researchers, funders, policy makers, and government au-
thorities will need to consider whether currently available sites
for publicly disclosing relevant information (e.g., the Biosafety
Clearing House of the Convention on Biological Diversity
(CBD), various clinical trial and nucleic acid databases, and
national regulatory agency websites) are sufcient for gene
drive technology or whether additional reporting mechanisms
are necessary.
Coordination. The working group encouraged funders to
support efforts to establish mechanisms for coordination
across projects and programs on gene drive technology.
Formation of networks among gene drive funders, re-
searchers, and regulators and policy makers, could encour-
age information sharing and cooperation in areas of mutual
interest and overall importance to the eld. For example, co-
ordination of communication strategies among teams working
on similar technologies, different approaches, and/or in the
same region is desirable and would contribute to research
advancement through enabling better community, stake-
holder, and public understanding. Such coordination should
be encouraged by those who are aware of various projects
within a region, such as academic institutions, regulators,
ethics committees, and funders. A forum for researchers in-
terested or involved in gene drive research would be especially
useful to promote evidence-based self-regulation, sharing
information on best practices, and supporting appropriate
management of eld trials.
Development of gene drive technology from initial research
through eld-testing and deployment will require complex
interactions among researchers, funders, and national and
international authorities at multiple levels, including broad
alignment of public engagement efforts, biosafety, and ethical
standards. The working group recommended the establish-
ment of a neutral body empowered to manage high-level co-
ordination among the various stakeholders and to organize
centralized responses to the diverse challenges that will arise
in the development pathway for gene drive mosquitoes as
public health tools.
Data systems. Researchers are strongly encouraged to
share eld data openly and collaboratively for the greater
benet of the malaria research and control communities. Ad-
equate database platforms for data gathering and storage for
evaluation/analysis, therefore, should be available. It is rec-
ommended that the data be archived in centralized, widely
accessible data repositories with the aim of having common
data formats. VectorBase
and PlasmoDB
are examples of
databases established for the purposes of such research.
Mosquito data should not only include sequence information,
but also extensive meta-data describing the type of mosquito
(gene drive or wild), source of collection, and experimental
study design. As data systems are being designed for eld
trials, it is recommended that they be developed following
Clinical Data Interchange Standards Consortium (CDISC)
guidelines. Clinical Data Interchange Standards Consortium is
anonprot standards-developing organization and has de-
veloped some standards for data instruments for malaria
Investigators could engage CDISC to develop
the data ontology relevant to mosquito vectors and related
information and expand this suite of standards for gene drive
Ethical obligations. The development and deployment of
gene drive mosquitoes for control of vector-borne diseases
will involve interaction with a diverse spectrum of groups, as
recognized by both the WHO Guidance Framework and the
NASEM report.
The WHO Guidance Framework distin-
guishes between communitiesthat live at the trial sites and
third partiesthat also have interest in the research but do not
live at the eld trial site.
The NASEM report denes com-
munitiesas those who live in or near sites where gene drive
organisms will be used and further distinguishes stake-
holdersas those who have direct professional or personal
interest in gene drive and publicsas those who lack a direct
connection but have interests or concerns that may contribute
to decision-making.
The composition and extent of these
groups likely will change with each successive phase of
testing. It could be argued that because gene drive constructs
theoretically could spread across large regions of Africa, most
of the African population legitimately falls in the category of
stakeholder regardless of where the trials begin. This points to
the importance of engaging with regional and multinational
bodies with authority to represent transnational sets of
As described in the WHO Guidance Framework, obligations
to these different communities, stakeholders, and publics will
vary in their ethical signicance and may be addressed
through a range of activities.
Researchers are responsible
for obtaining fair and legitimate authorization for eld-testing
gene drive mosquitoes.
At the highest level, safety is a
paramount public interest that is addressed through the reg-
ulatory mechanisms put in place by governments. Moreover, it
is a standard requirement to obtain ethical clearance for any
research involving human participation. The process for doing
so may differ among countries, and in some may require
considerable lead time.
At any point along the continuum of an investigational
products testing, individual informed consent is required from
those who meet the internationally accepted criteria of re-
search subjects (examples of requirements may be found at
the websites of the WHO
and U.S. Department of Health &
Human Services, Ofce for Human Research Protections
such as those who provide clinical specimens or identiable
information at the individual or household level.
simply living near a vector release site does not qualify
someone as a research subject.
Nonetheless, researchers
are obligated to respect the interests of those within the
community(ies) hosting trials of gene drive mosquitoes who,
although not research subjects, may be associated with and/
or affected by the research in a meaningful way. As discussed
in the following text under community engagement, this re-
quires practices undertaken to inform such persons about the
project, and to understand, respond to, and learn from their
perceptions and reactions in a way that makes it clear their
opinions have inuence.
Once the decision to eld-test a particular investigational
product has been made, researchers and funders incur a re-
sponsibility for the safety of the host community. Prematurely
discontinuing eld-testing and/or monitoring for lack of
funding could be considered irresponsible. Funders must be
prepared to commit to continued support for trial and post-
trial activities as long as is required by regulators and by ethical
obligations to the community hosting the eld testing. Like-
wise, researchers should not initiate eld releases until ade-
quate funds are secured to carry out their regulatory and
ethical obligations.
Given the complex ethical and community engagement is-
sues accompanying gene drive technology, an ethics advisory
group comprising experts external to the project would be an
important mechanism to supplement the input from commu-
nity advisory boards or other community engagement activi-
ties, providing additional and broader perspectives. This
group would be distinct from the institutional or national ethics
committee to which researchers must submit their proposed
activities for review and approval, and would advise the re-
searchers on ethical issues related to the project. This advice
could be especially helpful in determining how to anticipate
and address controversial or sensitive issues. Mechanisms
should be established to allow this group to obtain relevant
information on issues such as risk assessment, policy, en-
gagement activities, and trial status from the project and other
advisors. The working group strongly recommended that re-
searchers establish an independent group of ethics experts
that is external to the project team and includes in-country
experts and those from involved communities, to advise their
projects throughout the research and eld testing trajectory.
Engagement. Appropriate engagement will be crucial to
the success of the research on a number of levels. Therefore,
funders must be prepared to provide support for ongoing
engagement activities as an integral component of the re-
search plan. Acceptability of the research project, and of the
ultimate gene drive mosquito product, is fundamental to its
success. Engagement is essential to meeting ethical obliga-
tions of informed consent, building trust, and gaining accep-
tance of the research. When conducted through an open
exchange of ideas, engagement can also support knowledge
sharing that leads to development of a better and more ac-
ceptable product. Engagement will be an iterative process
that continues throughout the development pathway, un-
derstanding that opinions can change over time. Consider-
ation must be given, however, to mechanisms to monitor for
and avoid stakeholder fatigue over the course of lengthy trials.
Before releases begin, researchers, in collaboration with
government authorities of countries hosting the trial, funders,
or other advisors should create a plan for achieving effective
engagement with communities and other stakeholders,
thereby providing for opinions of various groups to be con-
sidered in the decision-making process over the course of a
project. For this, it will be important to conduct a systematic
analysis of inuential stakeholders at different levels.
At the
early stages of research, in addition to in-country members of
the project team and community members, researchers
should seek to learn from other in-country and/or regional
experts and organizations familiar with the local political, re-
ligious, social, and cultural structure to establish an appro-
priate engagement strategy. It is important to understand the
different levels of government when planning the engagement
approach and respect the requirements at each level. Re-
searchers should engage early with relevant ethics commit-
tees (e.g., institutional or national) for eld sites to determine
the extent of public engagement required in preparing for and
conducting eld studies, and guidance in identifying local
leaders and key in