DOI: 10.1126/science.1213798, 771 (2011);334 Science

et al.Stephanie JamesMosquito Trials

This copy is for your personal, non-commercial use only.

clicking here.colleagues, clients, or customers by , you can order high-quality copies for yourIf you wish to distribute this article to others

  here.following the guidelines

can be obtained byPermission to republish or repurpose articles or portions of articles

  ): April 25, 2014 www.sciencemag.org (this information is current as of

The following resources related to this article are available online at

http://www.sciencemag.org/content/334/6057/771.full.htmlversion of this article at:

including high-resolution figures, can be found in the onlineUpdated information and services,

http://www.sciencemag.org/content/334/6057/771.full.html#ref-list-1, 1 of which can be accessed free:cites 6 articlesThis article

http://www.sciencemag.org/content/334/6057/771.full.html#related-urls2 articles hosted by HighWire Press; see:cited by This article has been

http://www.sciencemag.org/cgi/collection/ecologyEcology

subject collections:This article appears in the following

registered trademark of AAAS. is aScience2011 by the American Association for the Advancement of Science; all rights reserved. The title

CopyrightAmerican Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by theScience

on

Apr

il 25

, 201

4w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

o

n A

pril

25, 2

014

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

on

Apr

il 25

, 201

4w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

www.sciencemag.org SCIENCE VOL 334 11 NOVEMBER 2011 771

PERSPECTIVES

lated by phosphorylation ( 8). Szijgyarto et

al. show that IP7 may transfer its pyrophos-

phate to key phosphorylated serine residues

in GCR1. The pyrophosphorylated GCR1

is thereby unable to interact with GCR2,

resulting in decreased transcription of the

genes for the glycolytic enzymes (see the

second fi gure). Pyrophosphorylation of ser-

ine residues by IP7 is a controversial topic; it

occurs in vitro ( 9), but there is no consensus

as to whether it can happen in vivo. Pyro-

phosphorylation of a clathrin coat adaptor

protein, AP3β1 ( 10), was recently demon-

strated with an elegant “back-pyrophospho-

rylation” assay that used radioactive IP7 to

show differences in the pyrophosphorylated

state of the relevant AP3β1 peptide after

expressing it in yeast strains with differing

amounts of IP7. The fi ndings were consis-

tent with stoichiometric pyrophosphoryla-

tion of that peptide in vivo.

By using as a control a mutant GCR1

with all four sequential serine residues that

are candidates to be pyrophosphorylated ( 9,

10) mutated to alanines, Szijgyarto et al.

show that one or more of these serines can

be pyrophosphorylated by IP7 in vitro (after

prior “primer” phosphorylation with the

protein kinase CK2), and that this construct

can act as a “dominant negative” in kcs1

cells. This and other evidence in their study

imply that in IP7-defi cient cells, GCR1 may

be less pyrophosphorylated, thus increasing

its interactions with GCR2. However, direct

evidence for GCR1 pyrophosphorylation by

IP7 in vivo is lacking.

Two further questions arise from the

study of Szijgyarto et al. Is this an evo-

lutionarily conserved mechanism? The

authors studied embryonic fi broblasts from

mice lacking IP6K1, which have low IP7

concentrations. These cells showed some

similarities to the kcs1 yeast strains in terms

of increased ATP content and compromised

mitochondrial function. This suggests that

the role of IP7 in regulating energy-related

metabolism may extend to all eukaryotes,

although this idea is not yet proven, not

least because mammals have three distinct

IP6K enzymes ( 2). What is the relationship

between the up-regulation of glycolysis and

the other major changes seen in yeast lack-

ing IP7 (including loss of mitochondrial

function and the massive increase in ATP

content)? It is not clear whether these other

events are regulated independently by IP7 or

whether they also initiated directly or indi-

rectly by pyrophosphorylation of GCR1.

Overall, the study of Szijgyarto et al. indi-

cates that IP7 may play an important role in

coordinating energy-related metabolism in

cells, presumably alongside other known

mechanisms such as those involving AMP

kinase ( 11) and, in higher organisms, cal-

cium ions ( 12).

References

1. Z. Szijgyarto, A. Garedew, C. Azevedo, A. Saiardi, Science

334, 802 (2011).

2. A. Chakraborty, S. Kim, S. H. Snyder, Sci. Signal. 4, re1

(2011).

3. S. Mulugu et al., Science 316, 106 (2007).

4. H. Lin et al., J. Biol. Chem. 284, 1863 (2009).

5. M. C. Glennon, S. B. Shears, Biochem. J. 293, 583

(1993).

6. F. S. Menniti, R. N. Miller, J. W. Putney Jr., S. B. Shears,

J. Biol. Chem. 268, 3850 (1993).

7. S. B. Shears, Mol. Pharmacol. 76, 236 (2009).

8. T. Mizuno et al., Yeast 21, 851 (2004).

9. R. Bhandari et al., Proc. Natl. Acad. Sci. U.S.A. 104,

15305 (2007).

10. C. Azevedo, A. Burton, E. Ruiz-Mateos, M. Marsh, A.

Saiardi, Proc. Natl. Acad. Sci. U.S.A. 106, 21161 (2009).

11. D. G. Hardie, Genes Dev. 25, 1895 (2011).

12. R. M. Denton, Biochim. Biophys. Acta 1787, 1309

(2009).

10.1126/science.1214727

Mosquito TrialsECOLOGY

Stephanie James 1, Cameron P. Simmons 2, Anthony A. James 3

Field trials of modifi ed mosquitoes present

complex but manageable challenges.

After two decades of research at the

bench, strategies based on biologic

or genetic modifi cation of mosqui-

toes to control vector-borne diseases are

now advancing to fi eld testing. Such strate-

gies seek either to reduce the overall number

of target mosquitoes to levels unable to sup-

port pathogen transmission (population sup-

pression) or to introduce into the local mos-

quito population a genetic modifi cation that

renders them unable to transmit the patho-

gen (population replacement). Dengue rep-

resents a good target for such interventions.

The disease is caused by a fl avivirus transmit-

ted primarily by the mosquito Aedes aegypti

(see the fi gure) and is a major problem in

tropical and subtropical regions. Conven-

tional methods (insecticide fogging, larva-

ciding, manual elimination of breeding sites)

are diffi cult to sustain at effective levels and

largely have failed to control dengue. Small

experimental trials of genetically engineered

( 1) or Wolbachia-infected ( 2) Ae. aegypti are

showing promising results in meeting their

respective entomological goals of popula-

tion suppression and population replacement

(infection with Wolbachia bacteria inhibits

growth of dengue virus in the mosquitoes).

What indications of success are required for

these technologies to be accepted as public

health tools? A recent meeting to consider

this question brought together vector biolo-

gists, epidemiologists, infectious disease and

clinical trial experts, and others interested in

dengue control ( 3).

Most vector biologists agree that success

is refl ected ultimately in reduced morbidity

and mortality. Some consider entomological

endpoints, such as local elimination of the

principal vector species or complete intro-

gression of a gene or symbiotic species that

causes pathogen refractoriness, as surrogate

markers for impact on infection and disease.

However, those familiar with trials of con-

ventional interventions (vaccines, drugs, and

insecticides) maintain that sustained epide-

miological and clinical impact should be the

primary effi cacy endpoint. Observations that

dengue transmission can sometimes continue

even with low mosquito population densities

( 4) are cited as a reason for vector biologists

to conduct trials to measure the incidence of

infection and/or disease.

Although cluster-randomized trials have

been adopted for studying the impact of

insecticide-treated nets on the vector-borne

disease malaria ( 5), similar trials of modi-

fi ed mosquitoes for dengue control present a

number of complications. Dengue transmis-

sion can be endemic, epidemic, multiyear

episodic, and unpredictable, so that trials may

have to continue for years. Trials also must

encompass large geographic areas to ensure

that there is sufficient human infection to

1Foundation for NIH, Bethesda, MD 20814, USA. 2Oxford University Clinical Research Unit, Hospital for Tropical Dis-eases, Wellcome Trust Major Overseas Program, Ho Chi Minh City, Vietnam. 3Departments of Microbiology and Molecular Genetics and Molecular Biology and Biochem-istry, University of California, Irvine, CA 92697–3900, USA. E-mail: sjames@fnih.org, csimmons@oucru.org, aajames@uci.edu

Published by AAAS

11 NOVEMBER 2011 VOL 334 SCIENCE www.sciencemag.org 772

PERSPECTIVES

detect differences between

control and treated popu-

lations. For a drug or vac-

cine, it is possible to ran-

domize individuals with a

similar risk of infection into

control or treatment groups

and follow them over time

to measure the direct effect

of the intervention. How-

ever, mosquito trials involve

area-wide, rather than indi-

vidual, intervention. Thus,

the effect must be measured

at community (cluster) lev-

els. Clusters must be large

enough and suffi ciently rep-

licated to detect an indirect

effect and overcome con-

founding factors such as

access to a health facility,

propensity for travel into

and out of the cluster, level

of local vector control efforts, or presence of

insecticide resistance. In urban environments,

there may be many such confounders. Fur-

thermore, should the modifi ed mosquitoes

prove effi cacious within treatment clusters, it

is likely that new dengue outbreaks will occur

disproportionately in control clusters because

transmission would be higher. If local gov-

ernment policy is to react to these outbreaks

by insecticide application, this would occur

more frequently in control clusters and poten-

tially bias results.

Blind treatments could be important com-

ponents in cluster-randomized trials. The

release of mosquitoes at some sites but not

others is likely to be obvious to the specifi c

communities, and human behavioral changes

may complicate trials. For example, people

living in treatment clusters may become more

conscientious about removing mosquito

breeding sites around their homes because of

regular visits from the trial team, which could

decrease the incidence of infection. Partici-

pants in control sites may perceive that they

are not receiving the same level of care and

decide to withdraw from the trial. Release of

some type of “placebo,” most likely unaltered

male mosquitoes (males do not bite humans

and therefore are harmless), could improve

the robustness of trials but would increase

considerably their complexity and expense.

Lengthy trials of modified mosquitoes

may increase the possibility of “contamina-

tion” of the treatment or control clusters as

people and mosquitoes migrate into and out of

the selected areas. Moreover, some of the new

technologies are designed to spread modifi ca-

tions into the native mosquito populations by

mating, so as to introduce long-lasting pro-

tection against transmission ( 6). These modi-

fi cations may be disseminated into mosquito

populations in the control clusters. Methods

must be included to detect these events so that

adjustment can be made in the trial analysis.

One way to avoid this contamination is for the

control and trial sites to be separated widely,

but geographic distance may hamper the abil-

ity to equally match control and experimental

sites for other factors that infl uence transmis-

sion, such as proximity to mosquito breeding

sites or health care.

There currently is no pharmaceutical or

agrochemical industry investment for modi-

fi ed mosquitoes, as new biological products

presumably have low profi t potential. Philan-

thropies or small biotech companies may be

unable to make the major fi nancial commit-

ments required for large trials or may not be

willing to fund such efforts in countries that

could support them on their own. Large-scale

trials of these technologies will be costly, and

thus the type of evidence required to prove

their utility should be carefully considered in

the context of potential availability of support.

Showing that mosquito-based technologies

produce the desired biological effect on local

mosquitoes under various conditions will be a

crucial fi rst step. In 2009, a transgenic strain

of Ae. aegypti that does not produce fertile

offspring was released on a small scale in the

Cayman Islands and demonstrated that trans-

genic males could survive and mate with

wild females (generating sterile larvae) ( 1). A

larger-scale trial was subsequently launched

there and has reportedly achieved an 80%

reduction in mosquito numbers ( 7). Moreover,

Wolbachia has been shown

to spread into local mosquito

populations (2). But is this

confi rmation enough?

Demonstrating decreased

incidence of infection in a

cluster-randomized trial,

based on seroconversion

(the production of antibodies

against the fl avivirus) within

a subset of individuals resid-

ing in treatment clusters, may

provide a relatively feasible

epidemiologic endpoint with

respect to trial size and cost.

Must we assume a need to

prove a reduction in actual

cases of dengue, which will

require very large trials? This

ultimately will be determined

by policy-makers, who will

decide whether these tech-

nologies are taken up as pub-

lic health interventions. Much will depend on

estimated cost-effectiveness and whether the

threat posed by dengue continues to expand.

The recent history of cluster-randomized tri-

als of insecticide-treated nets for malaria pre-

vention, where broad uptake followed con-

vincing proof of effi cacy, demonstrates the

value of generating evidence to inform policy

and attract donor funding for roll-out of the

intervention.

The challenges raised by large fi eld tri-

als of modifi ed mosquitoes are manageable,

and some precedents exist from trials of other

types of interventions. However, research-

ers must understand the level of evidence

required by policy-makers, which will allow

the appropriate studies to be designed and the

challenges to be addressed systematically. In

the meantime, additional smaller-scale pilot

studies will help to elucidate the potential

value of modifi ed mosquito technologies.

References and Notes

1. A. F. Harris et al., Nat. Biotechnol. 10.1038/nbt.2019 (2011).

2. A. A. Hoffmann et al., Nature 476, 454 (2011). 3. Designing fi eld trials of new vector-based dengue con-

trol strategies, 11 to 12 May 2011, Seattle, WA (www.fnih.org/sites/all/fi les/documents/Agenda%20for%20May%2011-12%20dengue%20meeting.pdf).

4. S. B. Halstead, Annu. Rev. Entomol. 53, 273 (2008). 5. P. A. Phillips-Howard et al., Am. J. Trop. Med. Hyg. 68

(suppl.), 23 (2003). 6. S. P. Sinkins, F. Gould, Nat. Rev. Genet. 7, 427 (2006). 7. N. Subbaraman, Nat. Biotechnol. 29, 9 (2011). 8. We thank T. W. Scott, J. Farrar, and S. O’Neill for helpful

comments and the meeting attendees (3) for discus-sions. Support was provided by the Foundation for the NIH through the Vector-Based Control of Transmission: Discovery Research Program of the Grand Challenges in Global Health Initiative.

10.1126/science.1213798 CR

ED

IT: O

. M

AR

INO

TT

I A

ND

P. J. B

RYA

NT

Modifi ed for release. Aedes aegypti (Higgs strain shown) can be modifi ed genetically so that they are unable to support transmission of dengue viruses. Their potential value as a means to control disease spread requires fi eld trials with appropriate indicators of success.

Published by AAAS