Pandora's box: how GM mosquitos could have caused Brazil's microcephaly disaster

Since August 2015, a large number of babies in Northeast Brazil have been born with very small heads, a condition known as microcephaly, and with other serious malformations. 4,180 suspected cases have been reported.

Epidemiologists have found a convincing correlation between the incidence of the natal deformities and maternal infections with the Zika virus, first discovered in Uganda's Zika Valley in 1947, which normally produces non-serious illness.

The correlation has been evidenced through the geographical distrubution of Zika infections and the wave of deformities. Zika virus has also been detected in the amniotic fluids and other tissues of the affected babies and their mothers.

Please also see Notes 1, 2 and 3 at the foot of this article.

This latter finding was recently reported by A S Oliveira Melo et al in a scientific paper published in the journal Ultrasound in Obstetrics & Gynecology, which noted evidence of intra-uterine infection. They also warn:

"As with other intrauterine infections, it is possible that the reported cases of microcephaly represent only the more severely affected children and that newborns with less severe disease, affecting not only the brain but also other organs, have not yet been diagnosed."

The Brazilian Health Minister, Marcelo Castro, says he has "100% certainty" that there is a link between Zika and microcephaly. His view is supported by the medical community worldwide, including by the US Center for Disease Control.

Oliveira Melo et al draw attention to a mystery that lies at the heart of the affair: "It is difficult to explain why there have been no fetal cases of Zika virus infection reported until now but this may be due to the underreporting of cases, possible early acquisition of immunity in endemic areas or due to the rarity of the disease until now.

"As genomic changes in the virus have been reported, the possibility of a new, more virulent, strain needs to be considered. Until more cases are diagnosed and histopathological proof is obtained, the possibility of other etiologies cannot be ruled out."

And this is the key question: how - if indeed Zika really is the problem, as appears likely - did this relatively innocuous virus acquire the ability to produce these terrible malformations in unborn human babies?

Oxitec's GM mosquitoes

An excellent article by Claire Bernish published last week on AntiMedia draws attention to an interesting aspect of the matter which has escaped mainstream media attention: the correlation between the incidence of Zika and the area of release of genetically modified Aedes aegypti mosquitos engineered for male insterility (see maps, above right).

The purpose of the release was to see if it controlled population of the mosquitos, which are the vector of Dengue fever, a potentially lethal disease. The same species also transmits the Zika virus.

The releases took in 2011 and 2012 in the Itaberaba suburb of the city of Juazeiro, Bahia, Northeast Brazil, about 500 km west of the coastal city of Recife. The experiment was written up in July 2015 in the journal PLOS Neglected Tropical Diseases in a paper titled 'Suppression of a Field Population of Aedes aegypti in Brazil by Sustained Release of Transgenic Male Mosquitoes' by Danilo O. Carvalho et al.

An initial 'rangefinder of 30,000 GM mosquitos per week took place between 19th May and 29th June 2011, followed by a much larger release of 540,000 per week in early 2012, ending on 11th February.

At the end of it the scientists claimed "effective control of a wild population of Ae. aegypti by sustained releases of OX513A male Ae. aegypti. We diminished Ae. aegypti population by 95% (95% CI: 92.2%-97.5%) based on adult trap data and 78% (95% CI: 70.5%-84.8%) based on ovitrap indices compared to the adjacent no-release control area."

So what's to worry about?

The idea of the Oxitec mosquitoes is simple enough: the males produce non-viable offspring which all die. So the GM mosqitoes are 'self-extinguishing' and the altered genes cannot survive in the wild population. All very clever, and nothing to worry about!

But in fact, it's not so simple. In 2010 geneticist Ricarda Steinbrecher wrote to the biosafety regulator in Malaysia - also considering a release of the Oxitec mosquitoes - with a number of safety concerns, pointing out the 2007 finding by Phuc et al that 3-4% of the first generation mosquitos actually survive.

The genetic engineering method employed by Oxitec allows the popular antibiotic tetracycline to be used to repress the lethality during breeding. But as a side-effect, the lethality is also reduced by the presence of tetracycline in the environment; and as Bernish points out, Brazil is among the world's biggest users of anti-microbials including tetracycline in its commercial farming sector:

"As a study by the American Society of Agronomy, et. al., explained, 'It is estimated that approximately 75% of antibiotics are not absorbed by animals and are excreted in waste.' One of the antibiotics (or antimicrobials) specifically named in that report for its environmental persistence is tetracycline.

In fact, as a confidential internal Oxitec document divulged in 2012, that survival rate could be as high as 15% - even with low levels of tetracycline present. 'Even small amounts of tetracycline can repress' the engineered lethality. Indeed, that 15% survival rate was described by Oxitec."

She then quotes the leaked Oxitec paper: "After a lot of testing and comparing experimental design, it was found that [researchers] had used a cat food to feed the [OX513A] larvae and this cat food contained chicken. It is known that tetracycline is routinely used to prevent infections in chickens, especially in the cheap, mass produced, chicken used for animal food. The chicken is heat-treated before being used, but this does not remove all the tetracycline. This meant that a small amount of tetracycline was being added from the food to the larvae and repressing the [designed] lethal system."

So in other words, there is every possibility for Oxitec's modified genes to persist in wild populations of Aedes aegypti mosquitos, especially in the environmental presence of tetracycline which is widely present in sewage, septic tanks, contaminated water sources and farm runoff.

'Promiscuous' jumping genes

On the face of it, there is no obvious way in which the spread of Oxitec's GM mosquitos into the wild could have anything to do with Brazil's wave of microcephaly. Is there?

Actually, yes. The problem may arise from the use of the 'transposon' ('jumping' sequence of DNA used in the genetic engineering process to introduce the new genes into the target organism). There are several such DNA sequences in use, and one of the most popular is known as known as piggyBac.

As a 2001 review article by Dr Mae Wan Ho shows, piggyBac is notoriously active, inserting itself into genes way beyond its intended target: "These 'promiscuous' transposons have found special favour with genetic engineers, whose goal is to create 'universal' systems for transferring genes into any and every species on earth. Almost none of the geneticists has considered the hazards involved ...

"It would seem obvious that integrated transposon vectors may easily jump out again, to another site in the same genome, or to the genome of unrelated species. There are already signs of that in the transposon, piggyBac, used in the GM bollworms to be released by the USDA this summer.

The piggyBac transposon was discovered in cell cultures of the moth Trichopulsia, the cabbage looper, where it caused high rates of mutations in the baculovirus infecting the cells by jumping into its genes ... This transposon was later found to be active in a wide range of species, including the fruitfly Drosophila, the mosquito transmitting yellow fever, Aedes aegypti, the medfly, Ceratitis capitata, and the original host, the cabbage looper.

"The piggyBac vector gave high frequencies of transpositions, 37 times higher than mariner and nearly four times higher than Hirmar."

In a later 2014 report Dr Mae Wan Ho returned to the theme with additional detail and fresh scientific evidence (please refer to her original article for references): "The piggyBac transposon was discovered in cell cultures of the moth Trichopulsia, the cabbage looper, where it caused high rates of mutations in the baculovirus infecting the cells by jumping into its genes ...

"There is also evidence that the disabled piggyBac vector carrying the transgene, even when stripped down to the bare minimum of the border repeats, was nevertheless able to replicate and spread, because the transposase enzyme enabling the piggyBac inserts to move can be provided by transposons present in all genomes.

"The main reason initially for using transposons as vectors in insect control was precisely because they can spread the transgenes rapidly by 'non-Mendelian' means within a population, i.e., by replicating copies and jumping into genomes, thereby 'driving' the trait through the insect population. However, the scientists involved neglected the fact that the transposons could also jump into the genomes of the mammalian hosts including human beings ...

"In spite of instability and resulting genotoxicity, the piggyBac transposon has been used extensively also in human gene therapy. Several human cell lines have been transformed, even primary human T cells using piggyBac. These findings leave us little doubt that the transposon-borne transgenes in the transgenic mosquito can transfer horizontally to human cells. The piggyBac transposon was found to induce genome wide insertion mutations disrupting many gene functions."

Has the GM nightmare finally come true?

So down to the key question: was the Oxitec's GM Aedes aegypti male-sterile mosquito released in Juazeiro engineered with the piggyBac transposon? Yes, it was. And that creates a highly significant possibility: that Oxitec's release of its GM mosquitos led directly to the development of Brazil's microcephaly epidemic through the following mechanism:

1. Many of the millions of Oxitec GM mosquitos released in Juazeiro in 2011/2012 survive, assisted, but not dependent on, the presence of tetracycline in the environment.

2. These mosquitos interbreed with with the wild population and their novel genes become widespread.

3. The promiscuous piggyBac transposon now present in the local Aedes aegypti population takes the opportunity to jump into the Zika virus, probably on numerous occasions.

4. In the process certain mutated strains of Zika acquire a selective advantage, making them more virulent and giving them an enhanced ability to enter and disrupt human DNA.

5. One way in which this manifests is by disrupting a key stage in the development of human embryos in the womb, causing microcephaly and the other reported deformations. Note that as Melo Oliveira et al warn, there are almost certainly other manifestations that have not yet been detected.

6. It may be that the piggyBac transposon has itself entered the DNA of babies exposed in utero to the modified Zika virus. Indeed, this may form part of the mechanism by which embryonic development is disrupted.

In the latter case, one implication is that the action of the gene could be blocked by giving pregnant women tetracycline in order to block its activity. The chances of success are probably low, but it has to be worth trying.

No further releases of GM insects!

While I am certainly not claiming that this is what actually took place, it is at least a credible hypothesis, and moreover a highly testable one. Nothing would be easier for genetic engineers than to test amniotic fluids, babies' blood, wild Aedes mosquitos and the Zika virus itself for the presence of the piggyBac transposon, using well established and highly sensitive PCR (polymerase chain reaction) techniques.

[See Notes 1 & 2, below. Recent Zika genome snapshots examined as a result of this article are, in fact, free of the piggyBac transposon.]

If this proves to be the case, those urging caution on the release of GMOs generally, and transgenic insects bearing promiscuous transposons in particular, will have been proved right on all counts.

But most important, such experiments, and any deployment of similar GM insects, must be immediately halted until the possibilities outlined above can be safely ruled out. There are plans, for example, to release similarly modified Anopheles mosquitos as an anti-malarial measure.

There are also calls for even more of the Oxitec Aedes aegypti mosquitos to be released in order to halt the transmission of the Zika virus. If that were to take place, it could give rise to numerous new mutations of the virus with the potential to cause even more damage to the human genome, that we can, at this stage, only guess at.

Note 1: A reader of this article, David Murphy, reports on Facebook (see thread here) as follows:

Zika, 4 strains, 2016 from the Center for Technological Innovation, Brazil

http://www.ncbi.nlm.nih.gov/nuccore/KU365777.1
http://www.ncbi.nlm.nih.gov/nuccore/KU365778.1
http://www.ncbi.nlm.nih.gov/nuccore/KU365779.1
http://www.ncbi.nlm.nih.gov/nuccore/KU365780.1

The releases of GM mosquitos took place in 2011 and 2012. Fortunately people sequenced zika before that.

Zika 01-AUG-2006

http://www.ncbi.nlm.nih.gov/nuccore/GU937109.1
id:AY632535.2

So lets compare. This isn't anything amazing, the viruses are tiny, this is rare in biology but we can actually eyeball the data since you could fit a whole viral genome on a single A4 sheet of paper.

View the alignment online:
http://web.archive.org/web/20160204113740/http://www.ebi.ac.uk/Tools/services/web/toolresult.ebi?jobId=mview-I20160203-123326-0036-25505577-pg

http://toolkit.tuebingen.mpg.de/alnviz/results/8144137

If you take a peek at the last 2 links you can see that while they're not identical the differences are almost all small changes with no big new chunks of code added. 

From a quick scan through eyeballing it, the only chunk of new bases was a 15 base sequence that doesn't show up in any piggyBac.

Zika is small, only 10000 bases long and it's hard to hide anything big in it. 

piggyBac is not tiny, ( http://www.ncbi.nlm.nih.gov/nuccore/?term=piggybac ), you wouldn't have to do any amazing analysis to see that it's not been added in there. Zika virus strain BeH818995 polyprotein gene, complete cds - Nucleotide - NCBI National Center for Biotechnology Information, U.S. National Library of Medicine 8600 Rockville Pike, Bethesda MD, 20894 USANCBI.NLM.NIH.GOV

Note 2: I'm grateful to David Murphy for carrying out this work, and to James Babcock for drawing it to my attention. It appears that the hypothesis set out above is probably incorrect, and this must be a matter of considerable relief to all concerned. However it remains my opinion that considerable caution should be exercised with releases of GE insects containing 'promiscuous' DNA sequences such as piggyBac.

The Zika virus, as a pure RNA virus, is resistant to the uptake of DNA of any kind (more on this in Note 3), and being very small, any such uptake would be rapidly detectable. However there are many other viruses, bacteria, disease vectors, etc, for which this remains a possiblity.

Another check that should also be made is to test for the presence of piggyBac in wild Aedes mosquitos around the release sites to see if, in fact, these 'programmed to die out' sequences are indeed as evanescent as claimed.

Note 3: I have received this comment by Dr Michael Antoniou, respected researcher in this field, originally posted on a GM news list, and reproduce it here in full.

"I urge great caution in viewing and using the message of this article in the Ecologist as it contains a number of technical errors with what it says especially about the piggyBac transposon system used to generate the GM mosquitos.

"Native transposons by their very nature do move around the genome of the host into which they have inserted and do so by the action of an enzyme called a 'transposase' for which they carry the gene.

"However, when adapted for use as a gene transfer vehicle ('vector') as was used to generate the GM mosquitos they have been stripped of their own genes including the one encoding the transposase, which as I mentioned is responsible for the insertion and mobility of the DNA transposon element. Thus following insertion of the GM transposon vector into the target cell genome by the transient provision (over a small time scale) of the transposase via a separate gene construct, the inserted GM transposon has no mechanism for subsequent mobilisation from its initial site of genome integration. Also, please note that transposases are specific for a given transposon. I worked with a transposon vector called Sleeping Beauty; the Sleeping Beauty transposase can only interact and move around its cognate transposon and not that of another type such as piggyBack, which has its own specific transposase.

"Furthermore, the suggestion that the transposon vector inserted into the GM mosquitos can be mobilised and insert itself into the Zika virus genome that may be carried by these GM mosquitos is even more farfetched. The Zika virus genome is made of RNA and resides/replicates after infection completely in the part of the cell known as the cytoplasm, which surrounds the cell nucleus, which is where the cell DNA with its inserted transposon are located. Not only are the Zika virus RNA in the cell cytoplasm and transposon DNA in the cell nucleus separated spatially and thus incapable of physically interacting but the piggyBack transposase enzyme even if somehow present is totally incapable of inserted mobilised transposon DNA into the Zika virus RNA genome. 

"Overall, I believe the link between the release the GM mosquitos and increased incidence of microcephaly is at best highly tenuous. I draw your attention to what some other individuals have highlighted as a far more likely cause of the increase in microcephaly; that is, escalation and spread in the use of toxic larvicides used to combat mosquito reproduction."

Dr Mae Wan Ho responds:

"Michael Antoniou claimed and argued with me for a long time that it was impossible for the CaMV 35S promoter to move into genomes. This has been proven to happen.

"I think the whole issue of horizontal gene transfer, not just of piggyBac, but the other bits of the construct cannot be ignored.

"Also, interaction between construct and virus could simply be mutagenic, as you suggested originally. And of course, there is now a suggestion that the 'killing gene product' from the transgenic mosquito may have been transferred into pregnant women, thereby causing microcephaly.

"If the female mosquito can transfer protein and viruses, why not nucleic acids? Both RNA (transcripts) and DNA can undergo horizontal gene transfer and recombination, and are hence also mutagenic."

I also add a highly relevant comment received in an earlier email from Dr Mae Wan Ho which addresses the same points as those raised by Dr Antoniou, though in another context:

"The point in the follow-up email is the most serious objection. piggbac is a DNA transposon that has no RNA intermediate, and zika virus is a flavivirus that has no DNA intermediate, therefore on the face of it, recombination between them is impossible.

"However, this has recently been proven wrong. See Stedman KM. Deep Recombination: RNA and ssDNA Virus Genes in DNA Virus and Host Genomes. Annual Review of Virology Vol. 2: 203-217 (Volume publication date November 2015).

"It states: 'genetic exchange across viral kingdoms, for instance between nonretroviral RNA viruses or ssDNA viruses and host genomes or between RNA and DNA viruses, was previously thought to be practically nonexistent. However, there is now growing evidence for both RNA and ssDNA viruses recombining with host dsDNA genomes and, more surprisingly, RNA virus genes recombining with ssDNA virus genomes. Mechanisms are still unclear, but this deep recombination greatly expands the breadth of virus evolution and confounds virus taxonomy.'

"The contention that disabled piggyBac cannot transpose is known to be false from studies cited in articles submitted by Cummins & Ho to the USDA in 2001, see here, for example: http://www.i-sis.org.uk/terminsects-pr.php. That is because transposases are abundant in genomes.

"A more recent article confirming this is Aziz RK, Breitbart M and Edwards RA. Transposases are the most abundant, most ubiquitous genes in nature. Nucleic Acid Res 2010, http://nar.oxfordjournals.org/content/early/2010/03/09/nar.gkq140.full "

My conclusion: there are obstructions to the onward transfer of genetic material from the transposon into the Zika virus, as suggested in my article. But it would be wrong to declare it impossible, as a number of critics have done in firm and absolute terms. These are obstructions, not hard barriers, and similar obstructions are proven to have been breached. It appears (from examination of the Zika genome) not to have taken place in the manner proposed on this occasion, however there is no guarantee that something similar may not happen in the future.

Dr Mae Wan Ho also refers to two additional possibilities as deserving of consideration, first, that "interaction between construct and virus could simply be mutagenic", and second, "that the 'killing gene product' from the transgenic mosquito may have been transferred into pregnant women, thereby causing microcephaly".

There remains every reason to take a highly precautionary approach to the release of genetically modified organisms and especially those that contain promiscuous transposons whose DNA may, as in the case of mosquitos, enter into people with potentially dangerous impacts.

Oliver Tickell edits The Ecologist.