Authors: Goulart E, Carlos de Caires Jnr. L, Kaid C, Passos-Bueno MR, Melo US, Okamoto OK & Zatz M (University of São Paulo, Brazil)
This is the story on how Zika virus, a terrible threat to pregnant women, turned out to be hopefully a ‘friend’. It all began in 2015 when the media started to report an increase in the incidence of babies born with microcephaly in Northeast Brazil, mainly in Recife, Pernambuco. When the first cases came out nobody suspected that they had been caused by a Zika virus epidemic until Dr. Vanessa van den Linden , a pediatric neurologist in Recife, suspected an association between gestational infection with Zika virus, recently introduced in Brazil, and the birth of affected babies.
Soon, there were thousands of reported cases, mainly in states of Pernambuco, Bahia, Paraiba and Rio Grande do Norte. Many newborns had also other associated malformations in addition to microcephaly, such as hearing and vision loss as well as arthrogryposis. The syndrome was named Zika congenital syndrome (ZCS).
However, although each case represents a tragedy for the family, we were intrigued by the fact that most women who had been infected by Zika virus during pregnancy delivered non-affected babies. The incidence of ZCS varies from 6–13%, in some countries even less. Could this be explained by an increased genetic susceptibility in affected babies or a protective mechanism in those who were born normal?
One way to address this question would be to compare groups of non-affected newborns with groups of affected babies, all confirmed to be exposed to Zika virus during pregnancy. However, in practice, it would be extremely difficult to perform such a study. First, because it would be difficult to pinpoint exactly when the infection occurred and also because we would have to compare mothers with different ethnic backgrounds and exposed to different environmental conditions. We therefore decided that the best approach would be to focus on twins.
Interestingly, the first reported twin case of ZCS was in Recife, and the outcome was discordant (one affected and one normal). We hypothesized that if there were a higher concordance among identical twins than among fraternal ones this would indicate a strong genetic component. With this in mind we searched for reports on twins and we identified nine pairs where at least one had ZCS: two were identical (monozygotic) and seven fraternal (dizygotic). The two identical twins were concordant (both affected) but six out of the seven dizygotic twins were discordant for ZCS (one normal and one affected), which supported our hypothesis of a genetic susceptibility component.
Next, we obtained induced pluripotent stem cell (IPS)-derived neuroprogenitor cells (NPCs) from three pairs of discordant twins and infected them ‘in vitro’ with Zika virus in the same laboratory conditions. We observed that the virus destroyed significantly more NPCs from affected babies – replicating what had happened in nature. Moreover, Zika replicated significantly more in NPCs derived from affected newborns than from the non-affected newborns . We then compared gene expression (through RNAseq) and confirmed that there were 64 genes differentially expressed in affected as compared with non-affected twins. Three genes in particular, DDIT4L, FOXG1 and LHX2 were significantly downregulated, supporting an increased genetic susceptibility in twins who developed ZCS. These results support the role of WNT-mTOR signaling in determining NPC predisposition to Zika infection. Overall, our study indicates that ZCS is not a stochastic event and depends on NPC intrinsic susceptibility, possibly related to oligogenic and/or epigenetic mechanisms.
Zika preferential tropism
We interviewed 91 mothers who had affected babies (unpublished data) and were intrigued by the fact that the majority reported no symptoms or very mild ones following Zika infection (such as a skin rash or some fever). It was clear that the virus was not very harmful to the mothers but had a strong tropism towards fetuses’ developing brains.
Could Zika virus destroy brain tumors?
This was our next question. While we were investigating the role of Zika virus in baby-derived NPCs, researchers from our group were studying pediatric brain tumors comprised by cells with neural stem-like features, also known as cancer stem cells (CSCs) . These aggressive central nervous system (CNS) embryonal tumors with high incidence in infants and resistance to classic cancer therapies are originated from NPC aberrations, affecting key cell signaling pathways that regulate normal neurogenesis . Since Zika infects preferentially NPCs, we hypothesized whether Zika could act as an oncolytic agent (molecular entity capable of induce cancer cells lysis/death) against such aggressive and metastatic human CNS embryonal tumors.
In order to address this question, we decided to test first different cancer cell lines ‘in vitro’. While we were performing these experiments, a paper came out showing that Zika significantly reduced mouse glioblastoma growth . However, these authors tested human glioblastoma only in vitro and at a fixed, very high, single MOI (viral dose) of 5. Regarding safety assessment in possible future clinical trials, our study tested different virus MOI at different time points (kinetics), concluding that Zika has strong oncolytic effects at very low MOI (0.01 to 2). We also compared Zika oncolytic effect in six cell lines from different human tumors (breast, prostate, colon rectal and three different CNS embryonal tumors), normal human NPC cells and neurons (6). With this approach, we were able to show that Zika oncolytic effects are not generalized to all kinds of cancer, but more specific to CNS tumors.
The greatest novelty of our study was the Zika oncolytic properties tested for the first time in pre-clinical experiments with human tumors. This was only possible by generating human tumor orthotopic xenografts in immunocompromised athymic mice. With this experimental model we were able to show that Zika not only induced complete tumor remission but also effectively inhibited metastatic spread of CNS tumor cells. Interestingly, we also observed that following Zika virus infection, the overall survival of tumor-bearing mice is significantly higher than that of sham animals. We also carried out additional cross-infection experiments indicating that the higher overall survival of tumor-bearing mice involves low production of functional viral particles by tumor cells. This new knowledge has important implications to future clinical trials.
Furthermore, additional functional experiments indicate the importance of the WNT signaling pathway to ZCS host predisposition and oncolytic effects of Zika virus. We observed that activation of this pathway increases Zika-induced tumor cell death. These findings provide novel mechanistic insights for the oncolytic effects of Zika and are of particular clinical relevance since activation of WNT signaling is a hallmark of CNS embryonal tumors. However, the oncolytic effects of Zika and its selectivity in CNS tumors are probably more complex and it is of imperative interest to determine the optimal conditions for highly specific cancer cells death through virus infection aiming successful Zika virus translation in future clinical trials.
In order to address the safety of a possible Zika tumor therapy, we also speculated if the persistence of virus in peripheral blood, brain, spinal cord and spleen in the animal model treated with one injection could spread Zika virus infection. In a cross-infection experiment, we observed that NPCs were not infected with Zika obtained from supernatants of embryonal CNS tumor cells treated with Zika. These experiments suggest that, in a possible clinical setting, the increased sensitivity of CNS tumor cells to Zika and subsequent low production of functional viral particles due to death of infected cells would favor a more efficient and specific destruction of CNS tumor cells over normal NPCs. Furthermore, the fact that the great majority of individuals infected by Zika virus during the epidemic reported very mild or no symptoms at all is very reassuring.
In short our results showing how an enemy can become an important ally are very exciting and we hope that we will be able to witness soon the effectiveness of Zika in treating brain tumors with safety in humans.
We are very grateful to all our colleagues that participated in this study and particularly to the patients and their families. This work was supported with grants from FAPESP ( CEPID-2013/08028-1), INCT, CNPq
Ernesto Goulart* ,Luiz Carlos de Caires Júnior*, Carolini Kaid*, Maria Rita Passos-Bueno, Uirá Souto Melo, Oswaldo Keith Okamoto&, Mayana Zatz&
Human Genome and stem-cell Center, Institute of Biosciences, University of São Paulo, São Paulo, Brazil
*These three authors contributed equally to this work
You might also like:
- Linden VV, Linden HV Jr., Leal MC et al. Discordant clinical outcomes of congenital Zika virus infection in twin pregnancies. Neuropsiquiatr. 75(6) 381–386 (2017).
- Caires-Júnior LC, Goulart E, Melo US et al. Discordant congenital Zika syndrome twins show differential in vitro viral susceptibility of neural progenitor cells. Commun. 9(1) 475 (2018).
- Kaid C, Silva PBG, Cortez BA, Rodini CO, Semedo-Kuriki P, & Okamoto OK. miR-367 promotes proliferation and stem-like traits in medulloblastoma cells. Cancer Sci. doi:10.1111/cas.12733 (2015).
- Xie Z. Brain tumor stem cells. Res. 34(12), 2055–2066 (2009).
- Zhu Z, Gorman MJ, McKenzie LD et al. Zika virus has oncolytic activity against glioblastoma stem cells. Exp. Med. 214(10) 2843–2857. (2017).
- Kaid C, Goulart E, Caires-Júnior LC et al. Zika virus selectively kills aggressive human embryonal CNS tumor cells in vitro and in vivo. Cancer Res. doi:10.1158/0008-5472.CAN-17-3201 (2018).