Roger W. Hendrix, 74

Virologist Roger W. Hendrix died on 15 August 2017. I only met Roger once, at the 2011 ASM meeting in New Orleans where we recorded an episode of This Week in Virology. The video of that episode is below, starting at my conversation with Roger at 30:34. Harmit Malik and Rachel Katzenellenbogen were my other guests on TWiV 135.

Thirty-five years later

Thirty-five years ago this month, in September 1982, I arrived at Columbia University College of Physicians & Surgeons to open my virology laboratory. I brought with me an infectious DNA copy of the poliovirus RNA genome, the first of its kind, and a lot of enthusiasm. Over the years we used this infectious DNA to study poliovirus neurovirulence, pathogenesis, and translation, among other topics; I wrote grant applications, published papers, and trained new scientists. In short, I was a typical academic scientist.

My career forked in 2000 with the publication by the American Society for Microbiology of the textbook Principles of Virology. Because this book was written by process, not by virus, each of the authors learned far more virology than ever before. As a consequence of writing this book, I became interested in disseminating virology to the public. Beginning with virology blog in 2004, I began to use social media to communicate science. This interest has lead to a collection of blogs, podcasts, lectures, and videos, in addition to four editions of Principles of Virology.

Recently virologist Islam Hussein, founder of Virolvlog and an avid science communicator, decided to summarize my modest scicomm career with an infographic. I’m grateful to Islam for this lovely chart, which was produced by Mohamed Gaawan. Here’s to the next 35 years.


TWiV 457: The Red Queen meets the White Rabbit

Brianne returns to the TWiV Gang to discuss the distribution of proteins on the influenza viral genome, and the evolution of myxoma virus that was released in Australia to control the rabbit population.

Click arrow to play
Download TWiV 457 (74 MB .mp3, 123 min)
Subscribe (free): iTunesRSSemail

Become a patron of TWiV!

Show notes at

A plasmid on the road to becoming a virus

Origin of virusesPlasmids have been discovered that can move from cell to cell within membrane vesicles in a species of Archaea (link to paper). They provide clues about the origin of virus particles.

Electron microscope analysis of the culture medium from Halobrum lacusprofundi R1S1, an Archaeal strain from Antarctica, revealed spherical particles which were subsequently shown to contain a 50,000 base pair circular double-stranded DNA molecule. When added to H. lacusprofundi, the purified membrane vesicles entered the cells and the DNA replicated.

Nucleotide sequence analysis of the plasmid within the membrane vesicles revealed 48 potential protein coding regions and an origin of DNA replication. None of these proteins showed any similarity to viral stuctural proteins, leading the authors to conclude that these particles are not viruses.

Many of the proteins encoded in the plasmid DNA were found in the membrane vesicles. Some of these are similar to cell proteins known to be involved in the generation of membrane vesicles. However no DNA polymerase-like proteins are encoded in the plasmid. These data suggest that the plasmid encodes proteins that generate, from the membranes of the cell, the vesicles needed for their transport to other cells. However, replication of the plasmid is carried out by cellular DNA polymerases.

It is likely that the plasmid-containing membrane vesicles are precursors of what we know today as virus particles. It is thought that viruses originated from selfish genetic elements such as plasmids and transposons when these nucleic acids acquired structural proteins (pictured; image credit). Phylogenetic analyses of the structural proteins of many enveloped and naked viruses reveal that they likely originated from cell proteins on multiple occasions (link to paper).

The membrane-encased Archaeal plasmid seems well on its way to becoming a virus, pending acquisition of viral structural proteins. Such an early precursor of virus particles has never been seen before, emphasizing that science should not be conducted only under the streetlight.

TWiV 456: Be careful of canons

Brianne joins the TWiVMasters to explain how mutations in genes encoding RNA polymerase III predispose children to severe varicella, and detection of an RNA virus by a DNA sensor.


Click arrow to play
Download TWiV 456 (75 MB .mp3, 124 min)
Subscribe (free): iTunesRSSemail

Become a patron of TWiV!

Show notes at

Phage synergy with the immune system

bacteriophage modelNot long after their discovery, viruses that infect bacteria – bacteriophages – were considered as therapeutic agents for treating infections. Despite many years of research on so-called phage therapy, clinical trials have produced conflicting results. They might be explained in part by the results of a new study which show that the host innate immune system is crucial for the efficacy of phage therapy.

When mice are infected intranasally with Pseudomonas aeruginosa (which causes pneumonia in patients with weak immune systems), the bacterium multiplies in the lungs and kills the animals in less than two days. When a P. aeruginosa lytic phage (i.e. that kills the bacteria) is instilled in the nose of the mice two hours after bacterial infection, all the mice survive and there are no detectable bacteria in the lungs. The phage can even be used prophylactically: it can prevent pneumonia when given up to four days before bacterial challenge.

The ability of phage to clear P. aeruginosa infection in the mouse lungs depends on the innate immune response. When bacteria infect a host, they are rapidly detected by pattern recognition receptors such as toll-like receptors. These receptors detect pathogen-specific molecular patterns and initiate a signaling cascade that leads to the production of cytokines, which may stop the infection. Phage cannot clear P. aeruginosa infection in mice lacking the myd88 gene, which is central to the activity of toll like receptors. This result shows that the innate immune response is crucial for the ability of phages to clear bacterial infections. In contrast, neither T cells, B cells, or innate lymphoid cells such as NK cells are needed for phage therapy to work.

The neutrophil is a cell of the immune system that is important in curtailing bacterial infections. Phage therapy does not work in mice depleted of neutrophils. This result suggests that humans with neutropenia, or low neutrophil counts, might not respond well to phage therapy.

A concern with phage therapy is that bacterial mutants resistant to infection might arise, leading to treatment failure. In silico modeling indicated that phage-resistant bacteria are eliminated by the innate immune response. In contrast, phage resistant bacteria dominate the population in mice lacking the myd88 gene.

These results demonstrate that in mice, successful phage therapy depends on a both the innate immune response of the host, which the authors call ‘immunophage synergy’. Whether such synergy also occurs in humans is not known, but should be studied. Even if observed in humans, immunophage synergy might not be a feature of infections in other anatomical locations, or those caused by other bacteria. Nevertheless, should immunophage synergy occur in people, then clearly only those with appropriate host immunity – which needs to be defined – should be given phage therapy.

TWiV 455: Pork and genes

Erin Garcia joins the TWiVirions to discuss a computer exploit encoded in DNA, creation of pigs free of endogenous retroviruses, and mutations in the gene encoding an innate sensor of RNA in children with severe viral respiratory disease.


Click arrow to play
Download TWiV 455 (64 MB .mp3, 105 min)
Subscribe (free): iTunesRSSemail

Become a patron of TWiV!

Show notes at

Purging the PERVs

pigThere aren’t enough human organs to meet the needs for transplantation, so we have turned to pigs. Unfortunately pig cells contain porcine endogenous retroviruses, PERVS, which could infect the transplant recipient, leading to tumor formation. But why worry? Just use CRISPR to purge the PERVs.

The genomes of many species on Earth are littered with endogenous retroviruses. These are DNA copies of retroviral genomes from previous infections that are integrated into germ line DNA and passed from parent to offspring. About 8% of the human genome consists of ERVs. The pig genome is no different – it contains PERVs (an acronym made to play with). The genome of an immortalized pig cell line called PK15 contains 62 PERVs. Human cells become infected with porcine retroviruses when they are co-cultured with PK15 cells.

The presence of PERVS is an obvious problem for using pig organs for transplantation into humans – a process called xenotransplantation. The retroviruses produced by pig cells might infect human cells, leading to problems such as immunosuppression and tumor formation. No PERV has ever been shown to be transmitted to a human, but the possibility remains, especially with   the transplantation of increasing numbers of pig organs into humans.

The development of CRISPR/Cas9 gene editing technology made it possible to remove PERVs from pigs, potentially easing the fears of xenotransplantation. This technology was first used to remove all 62 copies of PERVS from the PK15 cell line. But having PERV-free pig cells doesn’t help humans in need of pig organs – for that you need pigs.

To make pigs without PERVs, CRISPR/Cas9 was used to remove the PERVs from primary (that is, not immortal) pig cells in culture. Next, the nuclei of these PERV-less cells was used to replace the nucleus of a pig egg cell. After implantation into a female, these cells gave rise to piglets lacking PERVs.

In theory such PERV-less piglets can be used to supply organs for human transplantation, eliminating the worrying about infecting humans with pig retroviruses. But first we have to make sure that the PERV-free pigs, and their organs, are healthy. The more we study ERVs, the more we learn that they supply important functions for the host. For example, the protein syncytin, needed to form the placenta, is a retroviral gene, and the regulatory sequences of interferon genes come from retroviruses. There are likely to be many more examples of essential functions provided by ERVs. It would not be a good idea to have transplanted pig organs fail because they lack an essential PERV!

TWiV 454: FGCU, Zika

Sharon Isern and Scott Michael return to TWiV for a Zika virus update, including their work on viral evolution and spread, and whether pre-existing immunity to dengue virus enhances pathogenesis.


Click arrow to play
Download TWiV 454 (65 MB .mp3, 108 min)
Subscribe (free): iTunesRSSemail

Become a patron of TWiV!

Show notes at

Does prior dengue virus infection exacerbate Zika virus disease?

Antibody dependent enhancementThe short answer to the question posed in the title of this blog is: we don’t know.

Why would we even consider that a prior dengue virus infection would increase the severity of a Zika virus infection? The first time you are infected with dengue virus, you are likely to have a mild disease involving fever and joint pain, from which you recover and develop immunity to the virus. However, there are four serotypes dengue virus, and infection with one serotype does not provide protection against infection with the other three. If you are later infected with a different dengue virus serotype, you may even experience more severe dengue disease involving hemorrhagic fever and shock syndrome.

The exacerbation of dengue virus disease has been documented in people. Upon infection with a different serotype, antibodies are produced against the previous dengue virus encountered. These antibodies bind the new dengue virus but cannot block infection. Dengue virus then enters and replicates in cells that it does not normally infect, such as macrophages. Entry occurs when Fc receptors on the cell surface bind antibody that is attached to virus particles (illustrated). The result is higher levels of virus replication and more severe disease. This phenomenon is called antibody-dependent enhancement, or ADE.

When Zika virus emerged in epidemic form, it was associated with microcephaly and Guillain-Barré syndrome, diseases that had not been previously known to be caused by infection with this virus. As Zika virus and dengue virus are closely related, because ADE was known to occur with dengue virus, and both viruses often co-circulated, it was proposed that antibodies to dengue virus might exacerbate Zika virus disease.

It has been clearly shown by several groups that antibodes to dengue virus can enhance Zika virus infection of cells in culture. Specifically, adding dengue virus antibodies to Zika virus allows it to infect cells that bear receptors for antibodies – called Fc receptors. Without Fc receptors, the Zika virus plus dengue antibodies cannot infect these cells. ADE in cultured cells has been reported by a number of groups; the first was discussed here when it appeared on bioRxiv.

The important question is whether antibodies to dengue virus enhance Zika virus disease in animals, and there the results are mixed. In one experiment, mice were injected with serum from people who had recovered from dengue virus infection, followed by challenge with Zika virus. These sera, which cause ADE of Zika virus in cultured cells, led to increased fever, viral loads, and death of mice.

These finding were not replicated in two independent studies conducted in rhesus macaques (paper one, paper two). In these experiments, the macaques were first infected with dengue virus, and shown to mount an antibody response to that virus. Over one year later the animals were infected with Zika virus (the long time interval was used because in humans dengue ADE is observed mainly with second infections 12 months or more after a primary infection). Both groups concluded that prior dengue virus immunity did not lead to more severe Zika virus disease.

Which animals are giving us the right answer, mice or monkeys? It should be noted that the mouse study utilized an immunodeficient strain lacking a key component of innate immunity. As the authors of paper one concluded, it’s probably not a good idea to use immune deficient mice to understand the pathogenesis of Zika virus infection of people.

When it comes to viral pathogenesis, we know that mice lie; but we also realize that monkeys exaggerate. Therefore we should be cautious in concluding from the studies on nonhuman primates that dengue virus antibodies do not enhance Zika virus pathogenesis.

The answer to the question of whether dengue antibodies cause Zika virus ADE will no doubt come from carefully designed epidemiological studies to determine if Zika virus pathogenesis differs depending on whether the host has been previously infected with dengue virus. Such studies have not yet been done*.

You might wonder about the significance of dengue virus antibodies enhancing infection of cells in culture with Zika virus. An answer is provided by the authors of paper one:

In vitro ADE assays using laboratory cell lines are notoriously promiscuoius and demonstrate no correlation with disease risk. For example, DENV-immune sera will enhance even the homotypic serotype responsible for a past infection in the serum is diluted to sub-neutralizing concentrations.

The conundrum of whether ADE is a contributor to Zika virus pathogeneis is an example of putting the cart before the horse. For dengue virus, we obtained clear evidence of ADE in people before experiments were done in animals. For Zika virus, we don’t have the epidemiological evidence in humans, and therefore interpreting the animals results are problematic.

*Update 8/12/17: A study has been published on Zika viremia and cytokine levels in patients previously infected with dengue virus. The authors find no evidence of ADE in patients with acute Zika virus infection who had previously been exposed to dengue virus. However the study might not have been sufficiently powered to detect ADE.