Ebola 2014 is Mutating as Fast as Seasonal Flu

Ebola 2014 is Mutating as Fast as Seasonal Flu

// alex .at. operonlabs.com //

Background:

The current Ebola 2014 virus is mutating at a similar rate to seasonal flu (Influenza A).  This means the current Ebola outbreak has a very high intrinsic rate of viral mutation.  The bottom line is that the Ebola virus is changing rapidly, and in the intermediate to long term (3 months to 24 months), Ebola has the potential to evolve.  

We cannot predict exactly what the Ebola virus will look like in 24 months.  There is an inherent stochastic randomness to viral evolution which makes predictions on future viral strains difficult, if not impossible.  One basic tenet we can rely on is this: Viruses tend to maximize their infectivity (basic reproduction number) within their biological constraints (Nowak, 2006).  

These evolutionary constraints can be extremely complex, and can include trade-offs between virulence and infectivity, conditions of superinfection, host population dynamics, and even outbreak control measures.

One of the few statements we can make with confidence that the Ebola genome is changing at a specific rate, which is explained below.

Ebola Mutation Rate:

Analysis of the available research suggests that the Ebola 2014 virus is currently mutating at a rate 200% to 300% higher than historically observed (Gire, 2014).  


Ebola Genome Substitution Rates (Gire, 2014)


Furthermore, the Ebola-2014 virus's mutation rate of 2.0 x 10³ subs/site/year is nearly identical to Influenza A's mutation rate of 1.8 x 10³ subs/site/year (Jenkins, 2002).  This means Ebola 2014 is mutating as fast as seasonal flu.

 

Disclaimer: This paper contains no evidence (for or against) alternate modes of transmission for Ebola, nor is this paper postulating that genetic changes have impacted EVD clinical presentation (although evidence for this has started to emerge). This paper is simply demonstrating what appears to be a rapid rate of evolution in the Ebola 2014 Virus. Many recent Ebola viral mutations have been synonymous mutations, some have been in intergenic regions, while others are non-synonymous substitutions in protein-coding regions. All have unknown impact at the present time. Such questions should be the subject of future scientific research. This article simply points out that Ebola in 2014 is undergoing rapid mutation and adaptation.  The future implications of Ebola's rapid evolution are unclear.  

We chose to compare Ebola-2014 to Influenza A (Seasonal Flu) because Influenza is one of the fastest-mutating viruses (Jenkins, 2002).  Unlike chickenpox (VZV), which people usually only contract once per lifetime, Influenza can infect a single individual many times repeatedly over the years.  One of the reasons Influenza is able to re-infect humans each year is because the Influenza's high mutation rate allows the virus to generate 'escape mutants'.  Escape mutants are Influenza viruses which are no longer recognized by human immune systems.  Each winter presents us with a new mutated strain of the Influenza virus. Rapid mutation is beneficial to Influenza genetic fitness (in regards to antigenic regions), because it allows a 'new' Influenza virus to circulate year after year.

The benefit of a high mutation rate in Ebola 2014 is different -- the genetic changes in Ebola-2014 allow for rapid exploration of the entire fitness landscape in a brand new host -- humans. We need to be aware that the Ebola-2014 virus is undergoing rapid adaptation.

Ebola in Zoonotic Reservoir: Viral Genome adapted to Fruit Bats.  (Green)
Ebola in Human Hosts: Viral Genome adapted to Humans.      (Red)
Ebola Genotype will move Green -> Red during serial passage through Humans.

Until the Ebola outbreak is brought under control, the Ebola-2014 virus will continue to seed and adapt in its growing pool of West African human hosts.   We need to consider that as the weeks and months go on, the rapidly-changing Ebola-2014 virus will undergo repeated export from the West African region to countries around the world.

As new Ebola cases grow in West Africa and elsewhere, we are effectively conducting 'serial passage' experiments of Ebola-2014 through human hosts. The repeated passage of Ebola-2014 through humans is exerting selection pressure on the Ebola-2014 virus to adapt to our species (instead of fruit bats).  The introduction of Ebola-2014 into a large pool of West African human hosts (coupled with the complex dynamics of evolutionary selection pressure) may allow the Ebola-2014 virus to become more transmissible as the months go on, particularly in the absence of effective control interventions. 

The high mutation rate we see in Ebola-2014 reflects its ability to rapidly explore the fitness landscape. The ability of Ebola to undergo rapid genome substitutions and SNPs, coupled with genetic recombination, will allow 'survival of the fittest' in Ebola-2014 genetic variants (on both the intra-host and inter-host levels). New Ebola sub-clades are created with each passing month (there are already four sub-clades as of August 2014). New Ebola genetic variants are created with each new infection, though most are selected against. Rapid adaptation emerges from the high intrinsic Ebola-2014 mutation rate, coupled with the virus's ability to undergo RNA recombination during superinfection.

Molecular dating of the Ebola-2014 outbreak (Gire, 2014).
Probability distributions for both 2014 divergence events are overlaid above.

This phylogenetic tree is based on 99 Ebola viral genomes deep-sequenced from 78 distinct patients in Sierra Leone (Gire, 2014). We can see in the figure above that there are at least four Ebola genetic clusters (or sub-clades) based on phylogenetic analysis: These Ebola clusters are called GN, SL1, SL2, and SL3 by Gire et al. The key takeaway is that even prior to July 2014, the current Ebola outbreak had already accumulated significant genetic diversity.  Furthermore, the dominant circulating Ebola variants have changed over time. Up to four different Ebola-2014 viral sub-clades (groups of genetically related Ebola isolates) have circulated between humans since the onset of the 2014 Ebola outbreak.  

As the number of people affected by the 2014 Ebola outbreak has grown, so has the number of Ebola unique viral mutations and unique viral genetic lineages.  We can expect Ebola 2014 viral lineages to grow as some function f(i) proportional to the number of people infected with Ebola.


Ebola-2014: Acquisition of genetic variation over time (Gire, 2014).
Fifty mutational events (short dashes) and 29 new viral
lineages (long dashes) were observed.

 

The diagram above suggests that as the Ebola-infected host pool grows, so does the number of unique Ebola viral lineages (Gire, 2014).  This implies that Ebola acquires genetic diversity as it infects more people, particularly if the virus undergoes recombination during superinfection (Niman, 2007).  The growing number of new Ebola viral lineages will undergo natural selection for some 'optimum' balance of virulence, infectivity, tissue tropism, immune suppression, and other parameters which maximize the reproductive fitness of the Ebola virus in humans.  What that final virus might eventually look like 2 years from now is anyone's guess.  But the explosion of genetic variation suggests that the Ebola virus will become more difficult to contain as time goes on, which is why early action is important.

 

The idea that the Ebola-2014 Virus jumped species, but is now somehow 'static' or 'frozen in time' is a mistake. The Ebola-2014 virus is undergoing a period of rapid adaptation in human hosts, as evidenced by the Ebola RNA sequences deposited in Genbank, and the studies referenced with this article.  Hopefully, interventions (like contact tracing) will be able to stop Ebola-2014 before the virus optimizes its genotype.

 

These are two scenarios to outline what may happen in the future.  The critical variable determining the global outcome of Ebola is the response in West Africa, not the response in the United States.

Best Case Scenario:

WHO immediately deploys contact-tracing teams on the ground in West Africa.  The US Military is deployed as well, and constructs hospitals sufficient to care for the sick. The hospitals are staffed by qualified (read: well trained) caregivers. Teams on the ground  track down and care for Ebola-infected patients across West Africa, distributing self-treatment kits, food, medicine, and expertise.  An effort is made to involve local authorities and community leaders.  These efforts cause measurable reductions in the basic reproduction number of the virus by the end of 2014. 

Within 3 months to 9 months, the outbreak in West Africa peaks, levels-off, and begins to fade.  The Ebola virus never has the opportunity to acquire any significant mutations, due to its limited host pool. Ebola is fully under control by early 2015.  Sporadic cases in other countries are dealt with by treatment and contact tracing.  By Q4 2015, multiple Ebola vaccines and drugs are in the pipeline limiting the overall threat Ebola poses.

Worst Case Scenario:

The international response is perpetually behind the curve. Every response action is 8 to 12 weeks too late.  Statistics from the WHO become volatile and are unreliable as the lack of deployed personnel make hard numbers impossible to pin down. By  2015 the number of infections is in the hundreds of thousands in West Africa. The West African region exports 'asymptomatic infectives' which go undetected by basic screening. These individuals  'seed' outbreaks in other countries.

As more people become infected, a significant mutation arises that allows for a longer asymptomatic but infectious period, increasing the R-0. Globally, cases continue to double every 16 days, contact tracing infrastructure outside the West becomes saturated, and hospitals are overrun. By early-to-mid 2015, the global pool of Ebola-infected patients are in the millions, mainly centered in West Africa and Southeast Asia with multiple strains of varying virulence. A sudden change in the outbreak epidemiology caused by a recombinant Ebola strain causes confusion about how to respond. Efforts at developing treatments/vaccines become logistically complex and ineffective.

 

The implication of the Ebola 2014 mutation rate is this:  A single Ebola mutation doesn't necessarily mean the virus will become 'airborne', or that the virus has altered tissue tropism, or that the virus spreads more easily.  But a high intrinsic rate of Ebola mutation means that such changes may become possible in the future.  If the number of people infected grows into the hundreds of thousands, or even low millions, then the probability of a significant 'constellation' of accumulated Ebola mutations with phenotypic impact becomes more likely.  The problem is that accumulated Ebola mutations will scale with the size of the population infected.  Conversely, in a small population, such Ebola mutations are not likely to have a significant impact.  It's a bit like the virus is buying lottery tickets... The more lottery tickets the Ebola virus 'buys', the more chances it has to 'win'.  

Next Steps:

The general consensus in the scientific and epidemiological community is immediate intervention in West Africa is necessary in order to avoid taking the risky outcomes possible in a 'worst case' scenario.  A suitable response would need to include airlifting self-treatment kits with thermometers, the distribution of life-saving drugs, the construction of Ebola treatment centers, hospital staffing, contact tracing teams, and so forth.  A robust international response must happen soon in order to ensure that the current situation with the Ebola outbreak remains a 'best case' outcome.

 

References:

 
[1] Genomic surveillance elucidates Ebola virus origin and transmission during the 2014 outbreak. (Gire et al, 2014).
 
[2] Rates of Molecular Evolution in RNA Viruses: A Quantitative Phylogenetic Analysis. (Jenkins et al, 2002).
 
[3] Isolates of Zaire ebolavirus from wild apes reveal genetic lineage and recombinants. (Wittman et al, 2007).
 
[4] Ebola Recombination: Recombinomics Commentary. (Niman, 2007).
 
[5] Evolutionary Dynamics: Exploring the Equations of Life. (Nowak, 2006).
 

Deep Dive: Oct 15

Ebola Deep Dive Discussion: Outline, Figures, and Notes


Outline:

* Background Information and Definitions
* Impact of Ebola Mutations
* Ebola 2014 Mutation Rate: Comparison to previous Ebola outbreaks & Other Viruses
* Superinfection: Mathematical Properties / Evolutionary Dynamics
* Ebola Virulence vs Infectivity: Confounding Variables
* Recombination: Evidence for Horizontal Gene Transfer in Ebola
 
NOTE:  This post represents notes from a 'deep dive' scientific conference call regarding the Ebola 2014 outbreak.  
 
Definitions and Background Information:
Virulence: A virus's adverse impact on  host fitness (ie. host mortality and morbidity)
Infectivity: A virus's inherent ability to spread (ie. reflected in basic reproduction number)
 
There have long been discussions on virulence vs infectivity.  Currently, the consensus is that these two parameters are absolutely related within a virus to some sort of fitness optimum, but the confounding variables present make straightforward analysis impossible.  The important point is to understand that Virulence and Infectivity are linked, and form a sort of 'optimum' for a virus under a given set of conditions.
 
The Ebola Polymerase is an RNA-Dependent RNA Polymerase, which means that it is far more error prone than DNA viruses.  This reduces the reproductive fidelity per cycle and introduces genetic instability in the form in single-nucleotide polymorphisms (SNPs), insertions, and deletions.  Additionally, recent evidence has indicated the Ebola RdRp is capable of engaging in genetic recombination.  This gives Ebola a rich source of genetic change from which to explore it's optimum fitness landscape.
 

Deep Dive Discussion 1:

· The Ebola 2014 viral outbreak is of a has a common genetic,  lineage which goes back to the discovery of the species of the Zaire Ebolavirus in 1976.

· The Ebola 2014 outbreak is considered a 'strain' or 'sub-clade' of Ebola Zaire.

· Ebola Zaire viral species are defined by in reference to a 'consensus strain' -- the Mayinga-76 Ebola Virus (named for Ebola victim Nurse Mayinga N'Seka in 1976).

· The Ebola 2014 outbreak's RNA genome is only 97% similar to the 1976 Ebola Zaire consensus strain.  

· This means out of the 20kb Ebola 2014 genome, approx. 600 nucleotides are different from the 1976 consensus strain.

· The Ebola 2014 outbreak's RNA Genome can be considered to be a separate 'clade' within the species of Ebola Zaire due to these genetic differences.

· The Ebola 2014 outbreak's RNA Genome is most closely related to Ebola strains from 2007-2008 isolated and sequenced in the DRC.

· Computer analysis indicates the Ebola 2014 outbreak and the Ebola DRC 2007-2008 outbreak had a common viral ancestor, perhaps around 2004.

Deep Dive Figure 1:

Ebola 2014 Genetic Lineages
 

· Notice in the above diagram that Guinea and Sierra Leone both have distinct Ebola 2014 'sub-clades'

· Also notice in the above diagram that the current Ebola 2014 Guinea and Sierra Leone strains are most closely related to the DRC outbreak in 2007-2008.

· The current 2014 Ebola outbreak and the 2007-2008 DRC outbreak have an unidentified parent lineage, which ultimately goes back to 1976.

 

Deep Dive Discussion 2:

· Notice the Ebola 2014 outbreak and the DRC 2007-08 outbreak diverged from common ancestor strain in 2004 (Deep Dive Figure 2A).

· With only 97% sequence homology to the Mayinga-76 strain, the current Ebola outbreak could be substantially changed it's reproductive fitness -- but this is unknown.

· Notice on (Figure 2B) we have what looks like 3 or 4 sub-clades ('strains') present in the 2014 Ebola outbreak.

· We can see an Ebola strain in Guinea ("GN") appeared earliest (Feb - March), but then died out by May 2014. (Fig 2B)

· After the Ebola 2014 (GN) strain disappeared, new Ebola (SLx) strains took it's place. (Fig 2B)

· The Ebola strains from Sierra Leone ("SL1, SL2...") appeared after the GN strain, and these continued to spread in May and June 2014. (Fig 2B)

· Within the Ebola 2014 outbreak, we are dealing with multiple genetic sub-clades of Ebola ('sub-strains') which circulate and compete. (Fig 2B)

· The resurgence of the Ebola in May 2014 coincided with the appearance of genetically distinct Ebola viral sub-clades SL2 and SL3. (Fig 2B)

· Deep Dive Figure 2B does not tell us about reproductive fitness, but this is a mystery that must be resolved (do these Ebola genetic changes play any role?).
 

Deep Dive Figure 2:

 
 

Deep Dive Discussion 3:

· Above Deep Dive Figure 3A, we can view the Ebola 2014 Virus Genome, and it's accumulated mutations as of August 28th 2014.

· We can see that circulating Ebola viruses have substantial genetic changes, including non-synonymous mutations (protein changes) in:

NP gene (nucleoprotein)
VP35 gene (L cofactor/immune suppression)
VP40 gene (Ebola Matrix Protein)
GP gene (Ebola Spike Glycoprotein)
VP24 gene (Minor Matrix Protein)
 L gene (Ebola RdRp)

· This means that the Ebola 2014 Virus has protein-changes (red color) to EVERY gene except highly-conserved VP30 -- probably since VP30 is required for transcription activation.

· We can clearly see the Ebola 2014 Virus Genome has accumulated a substantial number of changes, including non-synonymous mutations.

· What is especially curious (h/t to IBM) is the amount of mutations that accumulated in the intragenic region -- the grey lines between VP30 and VP24.

· The implication of these non-synonymous mutations in the 2014 Ebola Virus Genome is unknown at this time.  They could substantially impact viral replication, tissue tropisms, virulence, etc.... Or these mutations could have absolutely no effect.
 

Deep Dive Figure 3:

 

· Notice in Deep Dive Figure 3A we see the Ebola virus genome organized from 5' to 3' end divided into a grid of boxes.  The rows are grouped by 2014 viral sub-clade (or 'sub-strain'). . . "GN", "SL1", "SL2", and "SL3".  The columns represent genetic changes across the Ebola -ssRNA genome.  The Ebola 2014 "SL3" strain can be distinguished by a unique SNP at position 10,218 in the genome.

· Notice in Deep Dive Figure 3B we see that over time, the Ebola 2014 SL1 strain became less and less dominant in the population, and burned itself out by June 2014, meaning that by June 2014, both the Ebola SL1 and GN strains were not actively circulating in humans in West Africa.

· During June 2014, the Ebola SL2 and SL3 strains began to become dominant and co-circulate.  Eventually, both became widespread.

· Deep Dive Figure 3C shows that the iSNP at position 10,218 (associated with SL3) became increasingly frequent over the month of June 2014, indicating reproductive success (for whatever reason) of the Ebola 2014 SL3 strain.

· Current 'deep sequencing' data from the ongoing outbreak as of Oct 16 2014 is not available.  The diagram referenced here stops analysis at August 2014.

 

Deep Dive Discussion 4:

· The phylogenetic tree below contains isolates from patients infected with the Ebola Virus in 2014.

· Their isolates were 'deep-sequenced' and the Ebola RNA sequences were deposited in Genbank.

· Phylogenetic analysis of these Ebola 2014 RNA sequences show (as of August 2014) that there are Four distinct Ebola sub-clades ('strains')

· The earliest strains, named GN and SL1, correlate with Guinea and Cluster1 respectively.

· The circulating strains, named SL2 and SL3, correlate with Cluster2 and Cluster3 respectively.

· It is undetermined if there is any molecular biological or sociological factor which would favor SL2 and SL3 over GN and SL1.

· Observed data regarding changes in Ebola sub-clades 'strains' may simply represent sampling bias of a small number of isolates.

Deep Dive Figure 4:

 

 

Deep Dive Discussion 5:

· This diagram shows the Ebola 2014 mutation rate compared to various parameters

· Mutation Rate and Substitution Rate are not technically the same measurement.  For simplicity, by mutation rate we mean substitution rate.

· Notice in Deep Dive Figure 5F we have a brown and a blue probability distribution.  

· The brown distribution (Figure 5F) shows Ebola mutation rates using 'all prior known human Ebola outbreaks'.  This results in a previous 'all-outbreak' Ebola substitution (mutation) rate average of about 0.9 x 10-3substitutions / base pair / year.

· The blue distribution shows (Figure 5F) Ebola mutation rates using sequences 'only from the 2014 Ebola outbreak'.  This results in a previous 'all-outbreak' Ebola substitution (mutation) rate average of about 2.0 x 10-3 substitutions / base pair / year.  

· The wide probability distribution of Ebola 2014 mutation rate ranges from as low as 1.0 x 10-3 subs/bp/year  to as high as 3.1 x 10-3 subs/bp/year.

· This data indicates that the Ebola 2014 outbreak is undergoing genetic mutation at a rate 220% to 330% faster than previous Ebola outbreaks.

· Part of this may reflect an acceleration of genetic change in order for the virus to be adaptable to human hosts, as it explores the fitness landscape.

Deep Dive Figure 5:



· Notice in Deep Dive Figure 5G that a significant fraction of Ebola 2014 viral mutations were detected and sequenced WITHIN patients (Fig 5G, Within Hosts). 

· What this means if someone is infected with a single copy of Ebola, by the time they are sick, they may actually posses MULTIPLE Ebola viral-substrains within themselves. (Fig 5G)

· In other words, an original Ebola genome might be in a host liver cell, but a mutated Ebola genetic copy might be in the same host's spleen cell.  This is how fast the virus is changing. (Fig 5G)

· Another important point from Figure 5G is that a substantial fraction of intra-host Ebola genetic mutation involves non-synonymous mutations, which can result in changes to amino acid residues which comprise the Ebola proteins.  Amino acid substitutions can have no effect, they can be beneficial for the virus, or they can be detrimental to the virus.

· This is how the virus explores the 'fitness landscape'.

· Lastly, in Figure 5H, notice that from May to June 2014, the virus acquired significant genetic variation which seemed to correlate with the number of hosts it infected.  As of June 16th, the Ebola virus had acquired 29 new viral lineages, which seemed to scale very closely with the number of new Ebola patients. (Fig 5H)

· Bottom Line:  The larger the pool of individuals sick with Ebola 2014, the more opportunities the virus will have to adapt for better genetic fitness (better transmissibility, etc)

 

 

Deep Dive Figure 6:

 

Deep Dive Discussion 7:

· This diagram shows how mutation rate and substitution rates relate to Viral classifications.

· Notice that the viruses with the highest substitution rates include are ssRNA viruses  (which includes Ebola).

Deep Dive Figure 7:

 

Deep Dive Discussion 8:

· Deep Dive Figure 8 shows the average mutation rate of the Influenza A virus NS genes.

· The average mutation rate for Influenza A is 2.6 x 10-3 subs / bp / year.  (Influenza A / Seasonal Flu)

· The average mutation rate for Influenza B is 0.5 x 10-3 subs / bp / year.

· The average mutation rate for Ebola 2014 is 2.0 x 10-3 subs / bp / year.  (Ebola 2014 Outbreak)

· Thus, the mutation rate for the Ebola 2014 outbreak is comparable to that of seasonal flu  (Influenza A is one of the fastest changing viruses known).

· The future mutations in Ebola will be impossible to predict. This why it is critical to get the outbreak contained.
 

Deep Dive Figure 8:

 

Deep Dive Discussion 9:

· The average mutation rate for Ebola 2014 is 2.0 x 10-3 subs / bp / year.  (Ebola 2014 Outbreak)

· The average mutation rate for all previous Ebola Outbreaks is under 0.9 x 10-3 subs / bp / year.  (Previous Ebola outbreaks)

· The average mutation rate for Influenza A NS genes is 2.6 x 10-3 subs / bp / year.  (Seasonal Influenza)

· In Deep Dive Figure 9, we can compare the Ebola 2014 viral mutation rate to 50 other common RNA Viruses.

· In Deep Dive Figure 9, we see that Ebola 2014's viral mutation rate is among the highest mutation rates in the literature for RNA Viruses.

· In Deep Dive Figure 9, any virus with a mutation rate comparable to Ebola 2014 Virus (2.0 x 10-3 subs / bp / year ) is marked in red.

· The Ebola 2014 Virus is currently mutating at a very fast pace by any metric, so much so that it is changing as fast as Seasonal Flu.
 

Deep Dive Figure 9:

 

Deep Dive Discussion 10:

· The pages below are extracts from the book Evolutionary Dynamics by Martin Nowak.

· These diagrams show how it is mathematically possible for a pathogen NOT maximize it's R0 under conditions of Superinfection.

· These diagrams also illustrate how it is possible for Nature to permit the evolution of Virulence in pathogens, when it seems such virulence is not in the pathogen's self-interest.
 

Deep Dive Figure 10:

 
 
 
 

Deep Dive Figure 11:

 
 


Deep Dive Discussion 12:

· Contrary to previous thinking, negative-sense ssRNA viruses can undergo genetic recombination.

· Recombination is where the viral RdRp essentially 'swaps' reading from one RNA strand to another, creating a chimeric viral RNA.

· Evidence has emerged to show Ebola Virus has undergone recombination events, event within the last 15 to 20 years.

· This evidence should seriously question our perspective and approach to how viruses evolve, especially if they can 'swap' genetic material so easily in a non-segmented genome.
 

Deep Dive Figure 12:

 
 

source: Isolates of Zaire ebolavirus from wild apes reveal genetic lineage and recombinants
 


Deep Dive Bibliography:

[1] Genomic surveillance elucidates Ebola virus origin and transmission during the 2014 outbreak. (Gire et al, 2014).
[2] Comparison of the Mutation Rates of Human Influenza A and B Viruses. (Nobusawa et al, 2006).
[3] Rates of Molecular Evolution in RNA Viruses: A Quantitative Phylogenetic Analysis. (Jenkins et al, 2002).
[4] Rates of evolutionary change in viruses: patterns and determinants. (Duffy, 2008).
[5] Viral Mutation Rates. (Sanjuan et al, 2010).
[6] Evolutionary Dynamics: Exploring the Equations of Life. (book), (Nowak, 2006).
[7] Isolates of Zaire ebolavirus from wild apes reveal genetic lineage and recombinants. (Wittman et al, 2007).

 
 

US Ebola Death: Antivirals

The first U.S. Ebola patient, Thomas Eric Duncan, has died despite being administered the experimental nucleoside-type antiviral pro-drug Brincidofovir.  How is this possible?  Well , the first problem is that Duncan was not treated until it was far too late for antiviral therapy to work.  

Let's take this opportunity to look at some nucleoside antivirals, incuding the drug that was administered to Mr. Duncan in Dallas, TX (Brincidofovir).

The first thing to know is remember Brincidofovir and Cidofovir were developed to work against dsDNA Viruses, though Ebola is a negative-sense ssRNA virus.  The analogy is similar to the previous Operon Labs post on the potential action of HIV RT inhibitor Lamivudine and our Ebola Virology Overview.

Broncidofovir, Cidofovir, Lamivudine:  Background

350
Cidofovir Brincidofovir (Cidofovir pro-drug)
Cidofovir, is a nucleoside-derived antiviral that is active against a variety of DNA Viruses including Cytomegalovirus (CMV), Adenovirus, HPV, Herpes Virus, and Orthopox viruses. Brincidofovir is a highly-fat soluble orally active form of Cidofovir.  Brincidofovir is orally absorbed and circulates as the lipid-conjugate above; it enters cells through diffusion, and is cleaved intracellularly into Cidofovir by Phospholipase enzymes.

In both examples above, intracellular Cidofovir is converted to it's active form , Cidofovir Diphosphate , by intracellular protein kinases.  These enzymes phosphorylate Cidofovir, resulting in its active DP form. 

*NOTE:  Oral Brincidofovir is useful because it can result in Cidofovir-DP intracellular concentrations that are approximately 10 to 100 times higher than can intravenous Cidofovir, with far fewer side effects.

Brincidofovir, the drug to the right, is orally active and was administered to Thomas Eric Duncan as an anti-Ebola therapy.  Cidofovir , the drug to the left , is not orally active.

Brincidofovir (right molecule) is eventually converted into Cidofovir (left molecule) inside human cells by enzymes.  

Cidofovir's phosphonate group limits its bioavailability, due to the negative charge of the phosphonate group at physiological pH, which prevent G.I absorption.  Thus, Cidofovir can only be administered via intravenous infusion.   Conversely, Brincidofovir is orally absorbed from the G.I. tract into the lymphatic system, due to it's long hydrophobic chain,

source: http://www.chimerix.com/c/discovery-clinical-trials/technology-discovery.php

The active antiviral compound produced in either case -- Cidofovir-Diphosphate -- is the same for both the above drugs.

Cidofovir DP -- Intracellular Antiviral active form of both Brincidofovir and Cidofovir.
 

 

Broncidofovir, Cidofovir, Lamivudine:  Mononegavirales RnRp Inhibition
 

Now what's interesting is that Cidofovir has zero published test data showing that it has high affinity for RdRp.  Here is an example on a study where Cidofovir was tested against Measles Virus -- a good virus to compare to as an Ebola reference.  Measles, like Ebola, is a negative-sense single-stranded virus in the same Genus as Ebola -- Mononegavirales.

In the Measles assay, Cidofovir has very low affinity for Measles RNA Dependent Rna Polymerase (RdRp) -- Cidofovir has zero anti-Measles activity up to 70uM (meaning it's IC50: far above >100 uM) ... This is in the same range as HIV RT drug Lamivudine (IC50: 180 uM for Lamivudine-TP against Measles RdRp).  Lamivudine actually might be more potent than Cidofovir against Measles RdRp.

Perhaps both these drugs work against Ebola for another reason not identified... Perhaps the accumulation of high levels of the active drug in PBMCs during conditions of high cellular stress?  Perhaps Ebola RnRp is particularly susceptible due to it's high polymerase error rate , or it's biomechanical structure?  Or perhaps one, both, or neither of these drugs work against RNA viruses like Ebola.  In terms of Cidofovir (the ultimate active drug administered to Thomas Duncan)... Let's see how it does against Measles, a virus comparable to Ebola...

"The [antivirals] were tested versus wild type measles virus in B95a cells (Figure 1) [...] Cidofovir, a DNA polymerase inhibitor, shows no effect in this assay [against Measles negative-sense ssRNA Virus Polymerase]." (McGuigan et al, 2013).   This is suggesting that Cidofovir (the active component in Broncidofovir) has very poor activity against Measles RnRp.
 
Now, granted, perhaps the doses weren't high enough.  Or perhaps Ebola RnRp is more susceptible.  Perhaps something else is in play.  But regardless, this suggests that we absolutely need more published scientific data from Chimerix (manufacturer of Broncidofovir) before we conclude that Brincidofovir (or even Lamivudine for that matter) is the latest and greatest Ebola antiviral -- since Cidofovir does not even inhibit Measles at 70 uM.   Keep in mind that Measles is susceptible (IC50) to Lamivudine at approx 180 uM.  Results for Cidofovir against Measles are relatively poor.
 
Tissue Culture Infective Dose:  Lower value means more potent Measles antiviral
 
So suffice to say, just looking at the public data, it does not look promising that Cidofovir or Brincidofovir would be active against Ebola.  Pharmacology studies indicate Brincidofovir has a Cmax blood concentration of 623.1084  nanomolar, or 0.623 micromolar.  If blood Cidofovir-DP levels are normalized to Cmax, this results in intracellular concentrations of CDV-DP of 3.4 uM.  To normalize for the 10x to 100x greater bioavailability of Brincidofovir, we multiply  to yield a Cidofovir-DP effective intracellular concentration range of 34 uM to 340 uM of in human PBMCs treated with Brincidofovir.  The long story short is that Cidofovir is not a particularly potent drug to begin with, with only 1/100 of the extracellular drug being converted to the active form intracellular.
 
BUT:  With Brincidofovir having great bioavailablity and pharmacokinetics, we reach Cidofovir-DP intracellular concentrations from 34uM to 340uM. . . The active level of Cidofovir-DP in the Measles assay is at minimum 0.7uM.   So Brincidofovir has potential.
 
But Brincidofovir has not been tested on RNA Viruses.  There is no published research that I can find besides press releases from Chimerix.  Hopefully, further data is forthcoming.
 
Any nucleoside-type antivirals likely must be administered soon after onset of symptoms, or they probably will not be effective in containing disease mortality.  Most all nucleoside antiviral must be activated by intracellular host kinases to their DP or TP (Triphosphate) Form, which can take 24h to 48h to distribute into tissue compartments, and for the drug to encounter the rate-limiting enzymatic conversion steps.  This is particularly true for pro-drugs like Brincidofovir and T-705.
 
Here is an example of how Brincidofovir is activated by host cellular kinases.  This process is very similar for most all nucleoside related antiviral drugs.
 
 

Favipiravir (T-705):  The Best Ebola Antiviral currently Approved

 
A better antiviral Ebola therapy that inhibits most all ssRNA Viruses is an RdRp (RNA-Dependent RNA Polymerase) Inhibitor called T-705 -- (similar in mechanism of action to Brincidofovir) -- T-705 only works in mice and other small animals if administered within the first 6 to 8 days Post-Infection (P.I.).  The same seems true in humans, based on it's short clinical testing in Europe against Ebola infected individuals.  
 
 

T-705 (aka Famipiravir) has been tested in murine models of Ebola, and has been shown to have potent anti-Ebola activity when administered at 150mg/kg B.I.D.  This drug has been used to save several patients' lives in Europe.  The European treatment protocols are not yet published, but using standard FDA Mouse/Human dose scaling,  a 65kg human probably needs to be administered (at minimum) 750 mg of Famipiravir twice per day to treat Ebola Virus Disease (EVD).

Fortunately, Famipiravir has the benefit of being tested in animal models against Ebola, and was designed and intended to substitute as a purine nucleotide in viral RdRp.  Famipiravir causes chain-termination of the enzyme after two anti-viral are incorporated sequentially.  We have strong data to show Famipiravir will treat Ebola in humans. This is not the case for Brincidofovir (yet).   The experimental drug Brincidofovir was designed as a Cidofovir pro-drug, which is ultimately active against some dsDNA viruses.  There are no in vivo tests (even in animals) to show Brincidofovir is active against Ebola.  

Thus, it is probably too early to evaluate Brincidofovir as an Ebola therapy -- whether positively or negatively.  It does have some promise, and should be investigated.  However, given that other existing therapies tested in animals are promising (Famipiravir, BCX-4430, Clomiphene, etc), these should be the first-line therapies against Ebola -- not Brincidofovir, which has yet to be tested in animal models.  Famipiravir production should be immediately scaled up, and selected as front-line treatment against Ebola.     BCX-4430 should immediately enter 'right to try' Phase I trials.  Clomiphene (already FDA approved) should immediately be tested on patients willing to give Informed Consent based on Clomiphene's 90% protection rate in murine models.

 

Brincidofovir and the first U.S. Ebola Fatality:  Too Late for Lethal Host Genes

Most likely Duncan received Brincidofovir too late, even if the drug is effective in vivo.  Duncan should have been administered therapy sooner by the medical authorities.  By the time he received any antiviral, most likely the vascular and organ dysfunction (Severe Sepsis / DIC) was just too severe. . . Even if the viral replication were slowed or stopped, there is still circulatory system and organ damage to deal with -- not to mention being over 14 days into a severe Ebola infection.  Why did Duncan take so long to be identified, receive treatment?  If this was done sooner, he may have survived.  By the time Brincidofovir was started, he was probably in DIC and viral sepsis.

Part of Ebola's lethality is likely due to a DIC/Septic Shock like syndrome that many patients progress to during the terminal phases of the disease.  Fatal strains of Ebola up-regulate a whole constellation of host genes.  One of the most damaging of these are the MMP class of genes -- specifically MMP-3 (Matrix Metalloproteinase 3).

"Proteins of the matrix metalloproteinase (MMP) family are involved in the breakdown of extracellular matrix and during tissue remodeling"

http://en.wikipedia.org/wiki/MMP3

Host MMP-3 upregulation has been studied and identified (among other genes) to correlate with the activity of lethal Ebola strains.  Most likely, MMP-3 is involved in the destruction of surrounding collagen networks around Ebola-infected Fibroblasts.  Interestingly (and of potential theraputic note) is that Doxycycline is one of the few 'broad-spectrum' human MMP inhibitors that is FDA-approved.

"Of particular relevance is our finding that metalloproteinase genes were upregulated only in lethal [Ebola] infections and not in response to the nonlethal wild-type or single-mutant viruses. Furthermore, the metalloproteinase gene MMP3 was one of 7 genes associated with lethality that was induced by the double mutant but not by the virus carrying the VP24 mutation. Metalloproteinases control chemokine activity (14, 34, 37, 43) and can regulate inflammation by controlling the activity of chemokines (13, 28), and MMP3-deficient mice are prone to severe inflammation "

Molecular signature associated with [Ebola] lethality. Biological network analysis determined by Ingenuity Pathways Analysis to identify functional relationships between the differentially regulated inflammatory genes associated with lethality. This analysis highlights five different subsets of genes that showed a direct functional relationship. These genes are involved in chemotaxis (green), apoptosis (yellow), acute-phase signaling (orange), cytotoxicity of leukocytes (blue), and leukocyte extravasation (purple). All these genes were strongly upregulated at 72 h postinfection as a consequence of lethal Ebola virus infection, with the exception of four chemotaxis genes (CCL21, CD209, CLEC4M, and FCER2) that were downregulated.

http://jvi.asm.org/content/85/17/9060.full

 

Ebola Lands in US

First U.S. Case of Ebola Virus Detected in Dallas, TX


source: http://toprightnews.com/?p=6257

 

The first case of Ebola Virus Disease (EVD) has been confirmed in a Dallas, TX man who recently returned from Liberia.  Thus, the first undetected case of Ebola Virus has now arrived within the borders of the United States.  This article will focus on the epidemiological aspects of this case, as the major media outlets are providing more-or-less non-stop coverage of other details.

One important aspect that must be discussed in regard to this first initial case is disease surveillance.  The fact that a passenger from Liberia was able to fly from Liberia through Belgium, Virginia, and land in Texas is a good opportunity to re-evaluate disease surveillance efforts.

The patient, Thomas Eric Duncan, was asymptomatic at the time of his flight.  He also filled out an exit form in Liberia indicating he did not have contact with any sick people.  The combination of these two factors (1. Lack of EVD Symptoms, and 2. Incorrect Declaration on Health Screening Form) allowed Duncan to exit Liberia and travel across the world.

The case of Thomas Eric Duncan suggests that we must seriously consider the role of asymptomatic carriers ("asymptomatics") in terms of 2014 Ebola disease epidemiology -- We must consider the role of asymptomatics even if the idea that "asymptomatics are not transmissible" turns out to be correct (a claim which should be scientifically re-examined).

Here we have a case where an individual with EVD traveled while asymptomatic.  Asymptomatics may represent a major 'loophole' in which EBOV (Ebola Virus) may exit West Africa and travel internationally.  We should consider efforts which may be put in place to screen for asymptomatic carriers, including perhaps the use of salivary, buccal, or blood-drop screening (rapid antigen technology similar to the OraQuick rapid home HIV test).  For Ebola, a blood-drop test would probably be most suitable.

Dr. Robert Garry of Tulane University has worked with other scientists to develop a field test for Lassa Fever which works on a single drop of blood.  Currently, he and his colleagues are working to adapt and field-test this device for Ebola Virus Disease (EVD).  These sort of tests would be a crucial method to stop asymptomatic carriers (Ebola viral embers) from exiting one country and entering another.  They could also prove highly-effective in other of the outbreak, removing the need for 'bandwidth-limited' RT-PCR testing during a situation where large numbers of patients could present with Ebola-like symptoms.

 

A second epidemiological issue regarding Thomas Eric Duncan is it shows the amount of rapid-response and manpower which is required to do isolation and contact-tracing on Ebola patients.  A single case of Ebola in Dallas, TX has resulted in well over 100 people under surveillance... this number could easily grow to 200 in the next few days as contact tracing continues.  The question that remains is one of manpower.  

While it may be improbable that the case of Thomas Duncan results in a sustained chain of transmission in the United States, this event underscores the fact that for each case of Ebola that comes to the U.S. from West Africa, there will be several hundred people that need to be monitored.  This is a ratio of 100:1 to 200:1 contacts per confirmed Ebola patient.  The problem is that if the outbreak continues to grow in West Africa, asymptomatics will continue to enter the U.S., and will require rapid detection, isolation, and treatment by the CDC and our health infrastructure.  

Eventually, however (if the outbreak in Africa is not brought under control), the United States will experience saturation of its contact-tracing capabilities.  Incoming cases will saturate our response, and Ebola cases will eventually fly 'under the radar' for one to three incubation periods until the index case can no longer be identified through contact tracing.  At this point, if the virus's effective reproduction number is sufficiently high (>1), the Ebola virus will be very difficult to stop within the U.S.  

From an epidemiological perspective, we cannot simply rely on our ability to contact trace asymptomatic Ebola patients after they walk off the tarmac.  This is not a sustainable strategy. 

Thus, the most important steps to take immediately are: (1) Immediate mobilization of resources to West Africa to stop the outbreak at the source, and (2) Re-assessment and expansion of containment strategies within the United States, to include rapid-antigen tests, better surveillance, and scientifically-sound coordinated policy response across Local, State, and Federal levels of government.

HIV drug may stop Ebola

Ebola Breakthrough? 

HIV drug shows promise

Here's how it might work.

Ebola has no approved therapies, but a few repurposed FDA-approved drugs show promise.   

A recent report out of West Africa indicates that a doctor gave HIV antivirals to Ebola patients out of despiration, and has apparently achieved success.  The numbers are small, and the study was not controlled -- indicating the results have low predictive value.  Nonetheless, this may represent a breakthrough in Ebola treatment if the reports hold up to scrutiny.  

Dr. Gobee Logan in Tubmanberg, Liberia treated 15 Ebola patients with the anti-HIV drug Lamivudine -- an anti-HIV reverse transcriptase inhibitor.  So far, 13/15 patients of his pateints have survived -- a 7% mortality rate, compared to the average Ebola 2014 mortality rate, which is closer to 71%.  The ten-fold reduction in mortality deserves further rapid and comprehensive investigation. 

The mechanism of action of Lamivudine (and perhaps other anti-HIV drugs) is plausible. Let's evaluate why.

Retroviruses such as HIV-1 must carry with them an enzyme called Reverse-Transcriptase, which is only found in viruses.  Reverse-Transcriptase is an enzyme is responsible for multiple functions in the HIV-1 viral replication cycle.  One of the first steps of HIV-1 genetic replication is to convert HIV's dsRNA genome to a dsDNA intermediate.  

Reverse Transcriptase Enzymatic Activity (HIV-1) NNS RNA Replicase Activity (Influenza, Ebola) 

 

In contrast, Ebola's RNA Replicase has three simpler functions -- Ebola's RNA Replicase (1) converts Ebola's negative-sense ssRNA into capped viral mRNAs (for protein syntesis), (2) synthesizes Ebola positive-sense ssRNA (a full length Ebola genome which acts as a replication template), and (3) synthesizes Ebola negative-sense ssRNA (full length Ebola genome for progeny viruses) .  

The overall similarity is that both the HIV-1 and Ebola viruses must bring with them an enzyme to transcribe and replicate their RNA genomes.  

HIV Reverse Transcriptase Activities Ebola RNA Polymerase Activities
RNA-Dependent DNA Polymerase RNA-Dependent RNA Polymerase
Ribonuclease H     A. RNA Transcriptase
DNA-Dependent DNA Polymerase     B. RNA Replicase

Lamivudine is a Nucleoside analog reverse-transcriptase inhibitor (NARTI).   Lamivudine is a small-molecule analog of the nucleoside molecule Cytidine. Lamivudine's mechanism of action in HIV-1 is to act as a chain-terminator in HIV-Reverse Transcriptase at an early phase of HIV-1 replication.  When HIV-1 attempts to copy its dsRNA into dsDNA using HIV RT's RNA-Dependent DNA Polymerase , it fails due to the induction of chain termination by Lamivudine.  

The following diagram shows how Lamivudine's chemical structure is related to it's natural nucleoside analog, Cytidine.

Lamivudine Chemical Structure Cytidine Chemical Structure

Lamivudine has only two structural differences from Cytidine.  Thus, Lamivudine is converted by host Nucleoside-Kinase , then host Nucleotide-Kinase to it's corresponding tri-phosphoryl analog. The HIV-1 RT enzyme recognizes and incorporates Lamivudine Triphosphate in the place of Cytidine Triphosphate.  However, after this step HIV-1, cannot elongate its dsDNA, because of Lamivudine Triphosphate's absense of a 3'-Hydroxy (3'-HO) group.  Lamivudine Triphosphate's lack of the 3'-HO group means the HIV-1 RT enzyme cannot attach next nucleotide, due the inability to form a 5' to 3' phosphodiester linkage.   There is a the lack of an attachment point at the 3'-position in Lamivudine.  Thus, Lamivudine is a DNA chain terminator in HIV-1 RT.

Now here's where it gets interesting... Cytidine (C) is present in both DNA chains (base pairs GCAT) and RNA chains (GCAU).  Since Lamivudine acts at the 'Replicase' portion of the HIV-1 RT enzyme, it's quite conceivable that Lamivudine has affinity for active sites in other viral enzymes.  In particular, NNS viruses like Ebola must carry their own viral Replicase enzyme.   Furthermore, since HIV-1 and Ebola target a similar part of the body (White Blood Cells like Macrophages), the pharmacokinetics and pharmacodynamics for Lamivudine 'overlap'.  We already know that Lamivudine will be present in sufficient quantities and accumulate in White Blood Cells, of which Monocytes, Macrophages, and Dendritic Cells are a primary (and preferential) cellular target of Ebola.

What may happen in Ebola is that Lamivudine competes with Cytidine as a nucleic acid base for Ebola RNA Replicase in human macrophages.  This could cause Ebola's RNA replicase to fail during one or all of these three steps: (1) Ebola RNA Replicase could fail to produce 'read-through' full-length +ssRNA template from -ssRNA genome due to chain-termination by Lamivudine, (2) Ebola RNA Replicase could fail to produce 'read-through' full-length -ssRNA genome from +ssRNA template due to chain-termination by Lamivudine, and/or (3) Ebola RNA Replicase could fail to produce viral mRNA (vRNAs) for ribosomal protein synthesis, due to incorporation into Ebola viral mRNAs , resulting in truncated protein products with probable loss-of-function.

A quick review of the literature shows that there has been minimal testing of HIV-1 RT nucleoside or nucleotide analogs against Filoviruses.  In fact, a quick review did not even produce evaluation of these drugs for activity against RNA Replicase in NNS viruses in general (those with a -ssRNA genome).  This must be explored.

Following the logic above, HIV-1 Reverse Transcriptase inhibitors that are nucleoside analogs (and possibly nucleotide analogs), which are related to Cytidine (preferably) , or possibly Guanine or Adenine should quickly be evaluated for in vitro and in murine models for chain-termination activity against Ebola RNA Replicase.  If results are promising, immediate trials in NHP models are warranted.

In the meantime, Lamivudine and other HIV-1 RT inhibitors should be added to 'right to try ' drugs for those sufferring from the 2014 Ebola outbreak.

Other anti-HIV drugs in addition to Lamivudine that should be immedialy explored for anti-Ebola activity include Emtricitabine, as well as perhaps newer nucleoside/nucleotide anti-RnRp HBC drugs.

A Review of Ebola Virology

(*PAPER IN PROGRESS -- Estimated Completion , Tues. Sep 30th, 2014)

Introduction

The Ebola 2014 outbreak has been nothing short of unprecedented.  


STEM Simulation of Ebola Virus 2014 West Africa Outbreak (Operon Labs, 2014).

Many medical doctors, virologists, epidemiologists, public health specialists, researchers, bioinformaticians, and other scientific experts are closely following ongoing events.   

The recent epidemiological simulations from the CDC suggest Ebola cases have begun geometric expansion.  Effective intervention is required should these assumptions prove correct.

The CDC shows Ebola virus infections above 1 million if interventions are not successful by Jan 2015.  A reduction of 50% or more would be necessary to ensure the current outbreak does not become a global problem.   The current response will be discussed in future updates.

With that said, the purpose of this post is to provide a background on EVD (Ebola Virus Disease) to interested scientists, researchers, or members of the public.  The content is primarily geared towards scientists with an interest in Ebola, who may not specialize in filoviruses.


Ebola Paleovirology

Ebola is an ancient disease.  Evidence continues to mount which supports this claim.  In fact,  with the rise of bioinformatics, an entire discipline (Paleovirology) has emerged to evaluate such ideas.

Our problem is that Viruses generally do not make good fossils.    Luckily for us, there are exceptions.

Sometimes, Viral genetic material can become incorporated into eukaryotic genomes through retrotransposons (L1 in humans).  The resultant genetic 'trash' from viral infections is passed on to subsequent generations of a species, resulting in what are called EVEs -- Endogenous Viral Elements .   EVEs are found in animals, plants, and fungi as 'relics' of viral infections. These viral fragments entered the host genome through long-deceased ancestors.  


Viral replication strategies, endogenous viral elements, and the genomic fossil record.

 

A 2010 PLOS Genetic Analysis indeed found watermarks of viral genetic data in vertebrate retrotransposons -- to include humans and other mammals.  To everyone's surprise, what they found were not the usual suspects (Rhinovirus, Coronavirus, etc).

This PLOS study ran data mining on 5666 viral genes (from all known non-retroviral ssRNA viruses at the time) and compared them to the complete genomes of 48 vertebrate species.  The strongest statistical associations (right at the top of the list) were Ebolavirus and Bornavirus. 

The result was so unexpected, the title of the paper started with "Unexpected Inheritance".  In other words, no one thought Ebola would show up.

 

Phylogenetic tree of vertebrates that encode Bornavirus- and Filovirus- like proteins in their genomes.
Bornaviruses-related sequences are denoted by icosahedrons and Filoviruses-related sequences by triangles. Times of the viral gene integrations are approximate. (Belyi et al, 2010).

This suggests that Ebola (or an Ebola-like disease) has been an enemy of mankind (or at least fruit bats) for quite a long time.  

So exactly how long have Ebola-like viruses been with us?   Oh, at least 35 million years.  As a lower estimate.  This suggest Ebola-like viruses are much older than our frozen ancestor Lucy, and shows Ebola predates many species.  Filovirus genetic fossils have been confirmed in many different animals, including primates.  The chart below implies that human and macaque Ebola genetic fossils had a common ancestor over 40 million years ago.  Note the blue column for filoviruses.

Timescaled phylogenetic tree of mammals screened in this study showing the known distribution of EVEs and of exogenous Borna-, Filo-, Circo-, and Parvoviruses. (Katzourakis et al, 2010)

To be fair, the studies quoted here did not find Filovirus 'fossils' in human genomes.  But ancient Ebola virus fragments were found in microbat, wallaby, guinea pig, shrew, opossum, and tarsier.   This suggests Filoviridae have been ubiquitous for millions of years, and are probably one of the oldest viruses conclusively known.  The finding of Ebola-specific gene fragments in tarsier is highly significant, as the tarsier is a non-human primate (NHP) -- a common route from the probably Ebola reservoir species (bat) to human. 

In the above annotated image, ancient Ebola was detected in non-human primates well over 35 million years ago -- probably much longer.

The study by Katzourakis et al had this to say about Filoviruses like Ebola in Paleovirology:

Filoviruses EVEs were identified not only in North American bats (M. lucifugus) and Asian primates (tarsier), but also in insectivores, rodents, and in both South American and Australian mammals (Figure 6). In concordance with the recent identification of Ebola Reston in swine [45], this unexpected result indicates that the distribution of filoviruses is likely much broader than has previously been recognized (Katzourakis et al, 2010). 

 

Ebola is very old.   So far we have prevailed, which is the good news.

 


Ebola Virus Structure: Overview

The following image (an NSF contest runner-up) visualizes the Ebola Virus Structural Biology:

Image Credit: Ivan Konstantinov, Yury Stefanov, Alexander Kovalevsky, Anastasya Bakulina, Visual Science

The above excellent artistic image above shows how an infectious Ebola viral proteins are assembled into a 3D model.  The current 2014 outbreak possess changes in multiple genes, both ncRNA and coding regions.  The 2014 genetic changes will be the topic of a follow up discussion.  For the moment, let's continue with our Ebola review...

The above artistic image shows how the proteins in the Ebola virus are assembled (along with it's genetic material, -ssRNA) to make a complete viral particle.  The Ebola virus is made out of  seven distinct proteins (nine proteins if we include secreted products).  The seven structural proteins are formed from seven transcriptional units on the RNA genome.   

The Ebola secreted proteins are an additional two soluble products which can be made from Ebola's seven transcriptional-unit mRNA genome.  The remaining two proteins (sGP and ssGP) are secreted into the extracellular fluid, and are not virus structural components.  

The purpose of Ebola's secreted soluble proteins sGP and ssGP are unclear at this time, although they appear to play a role in acting as 'decoys' to subvert the host immune system, by absorbing anti-GP antibodies (Mohan, 2012).  We will not discuss the Ebola secreted proteins in further detail, as they are still poorly understood, and are not critical to viral replication.

Regarding the Ebola virus gene products themselves, they are characterized as follows:

Ebola Virus Genome
GENE FULL NAME PRIMARY FUNCTIONS SECONDARY FUNCTIONS
NP Nucleoprotein -Essential Component of Nucleocapsid. 
-Transcription & Assembly. 
-ssRNA Packaging.  
-Spontanously Self-Assembles 
into non-functional helical structures (~20nm)
VP35 Virion Protein 35

-Essential Component of Nucleocapsid.
-L Polymerase Cofactor

-ssRNA Packaging.  
-Immune dysregulation
-Interferon suppression
-Silences host RIG-1 dsRNA binding
-Inhibits host IRF-3 phosphorylation. 

 

-Targets Innate immune response.
-
Interacts with cellular kinases IKKε & TBK-1.
-Can disrupt 
IKKε and IPS-1 interactions.

VP40 Virion Protein 40
/ "Matrix Protein"
-Maintains virion structural integrity
-Required for viral cellular egress
-Associated with late endosomes
-Can form 'budding' viral-like particles
in absence of all other Ebola proteins
-VP40 + GP VLPs can enter cells

-May exploit COPII transfer system to
reach 
host cell internal membrane for egress
GP Glycoprotein -'Spikes' on virion outer membrane
-Required for viral cellular entry
-Broad tissue tropism
-GP is Ebola's 'cell fusion' protein
-GP1/GP2 complex is heterodimer
-Outer GP1/2 is a trimer of heterodimers
-GP trimer forms 10nm long spikes
-GP is Post-Tr Cleaved into GP1/GP2
-GP1/GP2 linked via S-S bond
-GP is RNA-edited and Golgi-proc'd

-Dimeric sGP product is secreted
-sGP assists in viral cloaking ('decoy')
VP30 Virion Protein 30 -Essential Component of Nucleocapsid. 
-Transcription Activation & Reinitiation

-ssRNA Packaging.  
-Transcription will not proceed without VP30
-Contains Zn-binding Cys-His Motif
-Overall, poorly understood at present.
VP24 Virion Protein 24 /
"2nd Matrix Protein"
-Required for fully-functional Nucleocapsid
-Associated with Matrix Protein (VP40)
-Blocks IFN-α/β + IFN-γ signaling
-Immune Dysregulation

 
-Located between capsid & envelope
-Possibly associated with capsid assembly
-Competes with STAT1 for karyopherin
-Viral transcription inhibitor
-May play regulatory role in switch from
transcription <--> translation
L RNA Replicase -Essential Component of Nucleocapsid. 
-Genome Transcription & Translation
-Copies Viral -ssRNA into segmented
and capped viral mRNAs for translation.
-Copies -ssRNA to +ssRNA template
-Positive ssRNA template used to make
new neg ssRNAs for packaging into viral progeny

-Ebola RNA Replicase is complex
-Can engage in RNA editing (many ORFs)
-Has interactions in relation to ORFs
-Upstream repeat ORFs can regulate L
-Cellular stress can regulate L

 

The above table should give a high-level overview of the various genes in the Ebola virus, and their functions within both EBOV structural and functional biology.  The life-cycle of the Ebola virus is complex, but closely resembles other filoviruses as well as well-studied NNS viruses such as VSV.

As described above, there are seven main Ebola structural proteins, as in the 3D model presented above.  Let's review the virus diagram from the outside inward.  
 



Ebola Virus Structure: Cellular Entry via GP

A notable feature of Ebola is that underneath the GP spikes, there is a lipid-bilayer (grey in above diagram), making Ebola an enveloped virus. The lipid bilayer is derived from infected cell membranes as the virus buds from the cell.  Thus, Ebola membranes contain not only 'normal' lipid bilayer, they also contain a fair amount of other proteins and lipid rafts which were derived from the virus's previous host cell.  A viral envelope (as seen in Ebola) can help many viruses evade the immune system, because most of the virus's antigenic components besides GP are not exposed to scrutiny  (they are 'cloaked' by a real human-cell derived lipid bilayer).   The video to the right gives a good background on encapsulated viral fusion and entry.

 

 

Notice that the Ebola virus outer viral 'fusion proteins' (GP) spikes resemble that of influenza.  This is expected as both are believed to be Class I fusion proteins...Both are also homotrimers and undergo pH-dependent conformational change in the late endosome.  The interaction of fusion proteins with the endosome is how these viruses 'trick' their way out of the late endosome, which in practice means these viruses pop out of the cellular 'trash can' and into the cytoplasm. (SIB, 2014)

 

 

 

The surface proteins on the Ebola virus bind to a target cell (mostly monocytes, fibroblasts, and endothelial cells) and triggers a process called macropinocytosis. The cell unwittingly brings the virus inside, enclosed in a digestive and sorting compartment called the endosome.

The endosome is processed , and eventually becomes a 'late endosome' where acidification of its viral payload occurs. This triggers the activity of endosome enzymes Cathepsin B and L.  

The process is: (1) An interaction between Ebola GP1,2 and cell receptors trigger macropinocytosis into an endosome, (2) The endosome pH drops, activating cathepsin proteases, (3) Endosomal Cathepsin B and L sequentially cleave the Ebola GP1,2 into a 'primed' form, ready for fusion, (4) The primed Ebola GP1,2 is reduced and interacts with NPC1 receptor, (5) An unidentified event occurs, triggering primed GP1,2 to fuse with NPC1 and the endosomal membrane, liberating the Ebola virus contents into the host cell cytoplasm. (Hoffman-Winkler, 2012)

 

 

This is a detailed view of the Ebola virus fusion process in the late endosome.  First, Cathepsin L cleaves GP1 into a smaller 20kDa GP1 subunit.  Next, Cathepsin B cleaves the 20kDa GP1 into a 19kDa form that is now 'primed'.  Finally, an additional unidentified event occurs (possibly involving reduction and NPC1 interaction), triggering fusion with the endosome compartment and viral escape into the cell cytoplasm. (White, 2012)

 

 

"All enveloped viruses penetrate into host cells using a viral membrane fusion protein. Class I fusion proteins are trimers of three identical units (see the figure; the initial and final protein depictions are based on X‑ray structures of several class I fusion proteins). For most of these proteins, including Ebola virus (EBOV) glycoprotein, influenza virus haemagglutinin and retroviral Env proteins, each monomeric unit consists of a receptor-binding (rb) and a fusion (f) subunit, which are initially present in a single polypeptide chain. Priming by proteolytic cleavage (generally between the receptor-binding and fusion subunits) converts the protein from a fusion-incompetent to a fusion-competent (metastable) form that can respond to a fusion trigger. Triggering exposes and repositions the previously hidden (or tacked-down) fusion loop (or fusion peptide), which then binds hydrophobically to the target membrane." (White, 2012)

 

 


Ebola Virus Structure: Internal Proteins & Nucleocapsid

Beneath Ebola's GP spikes and lipid bilayer exists the Ebola virus matrix protein (VP40 aka "major matrix protein", and VP24 aka "minor matrix protein).  Both of these proteins have a structural role to maintain the shape of the viral particle, but they also have additional functions such as suppressing the host immune response, regulating cellular trafficking to the benefit of the virus, etc.

Beneath the major and minor matrix proteins is what is called the Nucleocapsid.  The Ebola virus nucleocapsid is a complex of four proteins (NP, VP35, VP30, and L) and the viral genome (ssRNA).  The NP protein forms a sort of helix which allows the Ebola ssRNA to be wound and packed tightly.   RNA is negatively charged (thus self-repels, like two North poles) so it is necessary for the virus to overcome (or at least minimize) the repulsive energy of its RNA.   So the Ebola RNA is wound onto the protein complex of NP (nucleoprotein), VP35, VP30, and L.

The nucleocapsid proteins alone are sufficient for self-assembling Viral-like Particles (VLPs), which spontaneously form structures almost identical in size and shape to Ebola viral particles... The nucleocapsid complex (when wound with proper viral ssRNA) is , quite remarkably, sufficient for transcription and replication.  To be specific, the VP30 protein is the necessary factor for transcription initiation -- and so can be thought of as a protein critical in transcription initiation in the L polymerase.  

However, the four protein nucleocapsid complex alone , as described above, is not sufficient for a fully-functional virus -- viral functions such as cellular ingress and egress are impaired to non-existent.  Recombinant non-pathogenic Ebola viruses (VLPs) often consist of Ebola Matrix Protein (VP40) with other Nucleocapsid genes , and are often used for study of viral behavior under non-BSL4 conditions.

 

 



Ebola Virus Structure: Immune Suppression

An additional interesting property of the Ebola nucleocapsid complex (NP, VP35, VP30, and L) is it also has 'multi-functional' roles for the proteins involved.  For example, the VP35 protein has a secondary function whereby it is partially responsible for 'shutting off' the host cell's immune response.  

 

The Ebola VP35 protein suppresses the activation of a host enzyme RIG-I.  Ebola binds to RIG-I, and prevents induction of downstream Interferon genes.

When a cell is infected by an RNA virus like Ebola, the presence of short ssRNA or dsRNA in the cytoplasm alerts the cell's pattern-recognition enzymes that something is amiss.
 

dsRNA is not normally present in large amounts in eukaryotic cytoplasm, so when it is detected, an cellular enzyme called RIG-I binds to the dsRNA and induces a signal cascade.

Essentially, RIG-I 'tattles' to the nucleus that a virus might be in the cell.  In response to detection of a virus (for example, a signal from RIG-I), the cell upregulates a constellation of antiviral genes, the most important of which are Type I Interferons.  

Ebola's VP35 gene prevents this process from happening.

 

 

The VP35 protein has the dual-role with the ability to bind to RIG-I, thus preventing the host cell from detecting the Ebola dsRNA that's amplifying in the cytoplasm.  An analogy here is that the VP35 protein is a cloaking mechanism that makes Ebola 'invisible' to RIG-I... What happens is that RIG-I becomes 'clogged up' with Ebola VP35 protein instead of viral ssRNA.  The end result is that no antiviral signaling occurs via the RIG-I pathway.  This prevents induction of cellular Interferon.

Interferon is categorically one of the most important 'master' antiviral genes that exists.  Interferon is the 'switch' for a whole constellation of gene products (called Interferon Stimulated Genes), most all of which cause the cell to 'double check' and 'inspect' everything, making it nearly impossible for a virus to properly replicate.   The power of Interferon is why Ebola (and other dangerous viruses) target it so ruthlessly.  Ebola must absolutely destroy and disable the host Interferon response in order to be able to succeed in it's prime directive -- replicate at all costs.

The Type I Interferons put the cell into a highly antiviral state -- I like to think of the cellular result of Interferon is a sort of 'stasis' where non-essential cell machinery gets shut down, and all non-essential staff go home.  The call calls in the engineers, and the engineers go back and 'double-checks' the cell's machinery to ensure that everything is okay.  If it's a false alarm, the cell returns to business as usual after about 6h to 48h.   However, if a virus is indeed present, Interferon stimulated genes will take action.  Nearby cells to alerted to 'be on the lookout' for a virus. If the cell cannot 'clear' the virus, it will actually undergo programmed cell death (often with help from immune cells) to destroy the virus within it.

 

In addition to inhibiting RIG-I, Ebola VP35 also targets IKKε, TBK1, IPS-1, and possibly TRAF3 interactions. All these host proteins are downstream signals of RIG-I. Thus, Ebola VP35 targets enzymes throughout the entire host antiviral signaling pathway.
 
Ebola is so intent on shutting off all host antiviral signalling, that targeting four to five proteins is not enough. 
 
VP35 protein actually has a final target for Interferon suppression: VP35 inhibits the phosphorylation of IRF-3. Phosphorylation of IRF-3 is the last step prior to antiviral signals reaching the cell's nucleus (for production of Interferon Beta).  
 
A second Ebola protein (VP24) has even more anti-interferon activity. Ebola's VP24 protein inhibits the nuclear import of phosphorylated STAT1 through competitive binding for karyopherin (KPNA5) (Xu et al, 2014) .
 
To summarize, using the diagram on the right: VP35 targets RIG-I, IKKε, IPS-1, TRAF3, TBK1, and IRF-3.  VP24 targets nuclear importation of STAT1.
 
In summary, Ebola inhibits multiple antiviral pathways as shown in this diagram, by acting through synergystic antiviral signal blockades.

 

It is quite remarkable that a Ebola virus protein (VP35) has (at least) four secondary and distinct anti-Interferon functions... All of VP35's targets involve the production of cellular Interferon or ISG.  The Ebola virus as a whole actually targets both the first (RIG-I) and final steps (IRF-3 / STAT1) of the host anti-viral signal cascade.  The specificity is actually remarkable, if it weren't so deadly and hostile.

The end result is that Ebola has multiple ways to 'cloak' itself from immune responses. . . Ebola thus relies on suppressing host antiviral signalling to allow early-phase replication to proceed in the host undetected by the innate immune response (NK-Cells, Interferon, ISG, etc).  By the time the host realizes that it's been tricked, the Ebola virus has already undergone extreme amplification undetected, and has claimed the Liver, Spleen, and Lymphatic system as it's turf.   



Ebola Virus Genome

Filoviridae like Ebola have the longest genomes of all the Mononegaviruses, clocking in at almost 20 kbp.

Filoviridae are also curious in that they have long non-translated sequences that flank the coding regions (after the stop signals).  These regions are of unknown purpose, but may be involved in auto-control of viral gene expression.

There are seven genes in the Ebola virus; The following comprise their order within the negative-sense (-) ssRNA non-segmented virion.  



 

The overall Ebola viral gene products have decreasing expression (as read from 3' to 5') , as in other NNS (nonsegmented negative-strand) viruses, due to the activity of the L-polymerase (via polymerase entry, initiation, and termination).   Put another way... Ebola has only a single promoter at the 3' end of it's genome for the RNA Polymerase to bind.   The RNA Polymerase can only load at one site.  As the RNA Polymerase moves down the genome, it tends to 'fall off' (and start over at 3') as it hits the 'bumps' of translation start and stop signals of the seven individual viral genes.  Thus, genes located near the promoter (3' end) are expressed at much higher levels than genes toward the 5' end.

Another consequence of this is that it makes things complex in terms of making full-length positive-sense RNA.  The polymerase needs to (sooner or later) override or ignore the start and stop signals it encounters in order to make a full-length positive sense template (polymerase read-through) .  The resulting complete 19kb +ssRNA serves as a master copy for production of full-length 19kb negative-sense RNAs.  The replicated -ssRNAs are copied to be incorporated into progeny viral genomes (after success of +ssRNA/-ssRNA read-through translation, mRNA gene transcription, and ribosomal/Golgi production of viral gene products).  The result of this process leads accumulation of new full-genome +/-ssRNAs in the cytoplasm, the self-assembly of the nucleocapsid complex with 19kb -ssRNAs, full virion assembly, and ultimately the budding functional new viral particles.
 

A diagram of NNS (ie Ebola virus) replication (transcription/translation) scheme (source)

 

 

 

 

Another visual example of the Ebola virus -ssRNA (NNS) genome 

 

All NNS viruses use a genetic regulation mechanism analogous to Ebola, which is best characterized as follows:

"The levels of gene expression [in NNS viruses] are primarily regulated by their position on the genome. The promoter proximal gene is transcribed in greatest abundance and each successive downstream gene is synthesized in progressively lower amounts due to attenuation of transcription at each successive gene junction"  (Barr JN, Whelan SP, Wertz GW., 2002)

Again, to put it another way, Ebola virus gene expression is regulated by a transcriptional gradient from the negative-sense 3' to 5' end, with gene products towards the 3' end (NP mRNA) transcribed at much higher concentrations than those at the 5' end (L mRNA). (Shabman, et al, 2013).    The Ebola mRNA levels are roughly characterized as follows:  NP > VP35 > VP40 > GP > VP30 > VP24 > L, but are time-dependent.  This would be characterized as the study of Ebola proteomics.

The following diagram shows the Ebola virus transcriptional gradient -- shown are mRNAs at 6h, 12h, and 24h post infection (Shabman, et al, 2013).

"Representative mRNA levels for each EBOV mRNA [...] at 6, 12, and 24 hpi. Each bar corresponds to a different EBOV mRNA."
source: 
An Upstream Open Reading Frame Modulates Ebola Virus Polymerase Translation and Virus Replication (Shabman, et al, 2013)
 

Notice that NP (nucleoprotein) is expressed at the highest levels, with VP40 (matrix protein) at third-highest levels, while L (RNA Polymerase) is usually expressed at the lowest levels.



Ebola Virus: Life Cycle Overview

The main process could be summarized as follows:

 


(White & Schornberg, 2012).
1.
Ebola virus (EBOV) binds to attachment factors and receptors on the cell surface through the viral spike protein, glycoprotein (GP) (step 1) 
2. The virus is then internalized into a macropinosome (step 2).
3. The virus is trafficked to an endosomal compartment containing the cysteine proteases cathepsin B (CatB) and CatL. These proteases digest GP to a 19 kDa form. (step 3)
4. Primed GP Initiates fusion between the viral and endosomal membranes (step 4). 
5. The viral nucleocapsid is released into the cytoplasm, where the genome is replicated (step 5).
6. The viral genome is transcribed with the aid of the viral proteins VP35, VP30 and L  (step 6).
7. Viral mRNAs are then translated (step 7).
8. mRNAs encoding GP are brought to the endoplasmic reticulum (ER), where GP is synthesized, modified with N-linked sugars and trimerized (step 8)
9. GP is further modified in the Golgi and delivered to the plasma membrane in secretory vesicles (step 9).
10. At the plasma membrane the ribonucleoprotein complex (RNA plus nucleoprotein (NP)) and associated viral proteins assemble with the membrane-associated proteins (matrix proteins VP24 and VP40, and GP), and the resultant virions bud from the cell surface (step 10).
11. Non-structural forms of GP, including soluble GP (sGP), are also secreted (step 11).


 

Ebola Virus: Tissue Tropism

The Ebola virus first enters a human or NHP host through mucus membranes, small skin abrasions,  or via parenteral administration.   Historically, EVD exposure is generally through direct contact or contaminated needles.  

Infectious Ebola virus particles in sick individuals are found on skin, mucus membranes, in bodily fluids, and nasal secretions in non-human primates (NHP).  Exposure generally occurs through direct contact with an infected individual, and may also occur through viral shedding and transmission via fomites.   Oral consumption of infectious particles can also result in disease.  Aerosol transmission of infectious particles have been demonstrated in animal models in a laboratory setting, though this finding remains controversial due to methodology.   Aerosol transmission may need to be re-evaluated in the context of the 2014 outbreak, especially considering the high numbers of medical personnel that have become infected.

Route of infection affects clinical course and disease outcome.  IV or IP administration of Ebola is almost invariably fatal (CFR=100%).  Direct Contact transmission results in lower fatality rates (80%).  Furthermore, IV/IP infection of Ebola was calculated to have an incubation period of 6.3 days, while contact exposures had a significantly longer incubation period of 9.5 days.   This fact may also be of relevance to the 2014 outbreak, perhaps if there are longer-than-expected generation times or incubation periods (i.e. oral vs nasopharyngeal infection).

An infectious Ebola virus particle can result in ultimate organ titers as high as 10 million to 100 million (10^7 to 10^8) PFU / g, which correlates with very high viral amplification in hosts.  Viral load has also been demonstrated to correlate with mortality.

Ebola infection will be divided into 'conceptual' phases based on available information.  These are provided for purposes of a conceptual framework, rather than the ultimate word on disease progression.  The following diagram shows the overall process, with all disease 'phases' superimposed.

 

NOTE:  THE FOLLOWING SECTION IS IN PROGRESS --  TO BE COMPLETED , Tue Sep 30th, 2014

A 'first-draft' conceptual model is presented below is provided as a repository of the author's knowledge to be refined and updated.  This model is not an experimentally verified framework , but simply a conceptual way of organizing information.  It will be refined and updated as time permits.

Ebola Infection: Phase 1:

Ebola infection begins when an individual comes into contact with the viral particles at an endothelial or epithelial site.  These can include mucus membranes like the nose, mouth, or eyes, or can include the nasopharynx or gastrointestinal tract.   Regardless, a substantial amount of infectious viral particles have been deposited on an endothelial surface.  The virus infects local cells immediately -- preferentially, it infects monocyte-derived white blood cells (WBCs) -- including monocytes, macrophages, and dendritic cells (DCs).  If such cells are not readily available, the virus will infect endothelial cells or fibroblastic cells at the site of infection.  

The only cells the Ebola virus cannot infect other WBCs like NK Cells, T-cells or B-cells; a substantial subset of non-monocytic WBCs are innately immune to Ebola infection.  Ebola has broad tissue tropism otherwise.  But Ebola can generally only infect three main types of White Blood Cell -- Monocyte, Macrophage, and Dendritic Cell (DC).

Transient Local inflammation results from Ebola primary replication and progeny release at site of initial infection.  This results in a muted and understated immune response and suppression of most all Interferon and ISGs.  Cytokines may be induced, but Interferon is switched off.  Monocyte-derived WBC arrive at the infection site , and are themselves quickly infected by the Ebola virus.  The virus continues to infect monocyte-derived cells in the vicinity.  Some of these infected monocyte-derived WBCs migrate to local lymph nodes, carrying the Ebola virus as a payload.  Once in a local lymph node, the Ebola virus rapidly infects more and more monocytes, macrophages, and dendritic cells.   A secondary target is Fibroblastic Reticular Cells of the FRC Conduit.  A tertiary target is endothelial cells.

At a certain point, Ebola infected monocyte-derived WBCs probabilistically cross over into blood circulation in a concentraiton sufficient to infect the Liver and Spleen.  At this point, this virus enters Phase 2.

Ebola Infection: Phase 2:

Once Ebola begins to appear in the bloodstream , originating from the site of entry and nearby lymph nodes, the virus specifically exhibits tissue tropism for the Liver and Spleen (as well as Adrenal Glands).  Blood containing Ebola-infected WBCs is filtered by the liver and spleen, where infected Ebola cells and/or viral particles are removed and deposited in these organs.  

<REMAINING SECTIONS IN PROGRESS>  

 

 

 

Death occurs due to hypovolemic shock or multiple organ failure.

"Lymph node of an infected African Green Monkey. The location of the Ebola Zaire antigen is indicated by the red stain. The large ovoid structure at the center of the picture is a high endothelial venule (HEV) that is infected with Ebola. The viral replication in the fibroblastic cells that control the HEV's structure has almost totally destroyed the HEV" (Dr. Tom Geisbert) "Photomicrograph of the lip of an African green monkey. Ebola virus (red) is penetrating between the epithelial cells of the lip overlying a lymphoid aggregate in the submucosa. This immunohistochemistry preparation was done by Keith Steele." (Dr. Tom Geisbert)

 

 

This process is under study and review.

Ebola 2014 Spread SEIR West Africa (initial run)

 

SEIR+D Diagram for Ebola 2014 Strain, using Liberian strain parameters from:

[1] http://currents.plos.org/outbreaks/article/estimating-the-reproduction-n...

Outbreak simulation is fixed with 10 people in New York on October 1st.

Incubation Period: 1/σ = 5.3 days [1]

Infectivity / Recovery Period: 1/γ = 5.61 days [1]

Case Fatality Rate: 71% [1] (Liberian Strain)

Comments: This simulation uses the precise parameters referenced in the above paper. The generation time G seems to be closer to 10 days than the observed average generation time of 24 days. Thus, further modelling will be required for accurate results. This graph is likely a very aggressive estimate, both in terms of clinical attack rate and rate-of-spread. Further research forthcoming.

 

 

 

 

 

 

 

 

 

 

 

Some Notes on STEM (Spatiotemporal Epidemiological Modeler)



STEM is an excellent piece of epidemiology software. STEM supports disease compartment models such as SIR, SIRS and SEIR, and here at Operon Labs we have successfully built some preliminary models to trace the 2014 Ebola outbreak. There are, however, some important details of which to be aware when using this software.

Normally , the parameter β (or β(t) when dynamic) represents the basic transmission rate of a disease, and is directly related to the parameter R0. This parameter is often found in published literature, and one would expect this parameter could be directly plugged into the STEM compartment models.

Unfortunately, reality is not this simple.

When using STEM to model a disease using an SEIR compartment model, it is important to remember that STEM treats the R compartment as 'Recovered' -- where R does not include disease-induced deaths -- while in much scientific literature the R compartment is known as 'Removed' and represents patients both recovered and dead. In other words, when STEM is used to model a disease with any significant mortality (like Ebola) , then the SEIR compartment model is actually more accurately termed SEIRD (Susceptible / Exposed / Infectious / Recovered / Dead).

The reason this is important is because when modelling a disease with significant mortality (such as Ebola) using STEM, the input parameters must be corrected to account for the fact that STEM represents the disease model differently.

One would expect the model to behave as follows... With Recovered/Removed containing both Recovered cases as well as Disease-Induced Deaths. STEM does not work like this.

Unfortunately, this is not the case. STEM's documentation is not quite accurate either. Here is STEM's (rather inaccurate) picture of its SEIR model...


β: Transmission Rate
ε or σ: Incubation Rate (per unit time)
γ: Recovery Rate (per unit time)
1 / σ: Average Incubation Period
1 / γ: Average Recovery Period
µ*: Population Birth Rate
µ: Population Death Rate
α: Immunity Loss Rate
δ: Infectious Mortality Rate (not shown in STEM compartment docs)

All of these are fine, and for simplified models, we can simply set the Population Birth and Death Rates equal to one another (or to zero), and we can safely ignore the µ* and µ parameters. The problem is when we want to add the parameter δ to represent the infectious mortality rate.

Notice STEM does not include δ in their compartment diagram! Read on...

Here at Operon Labs, we had noticed our Ebola simulations would grind to a halt upon addition of any significant mortality rate parameter, δ. After much analysis, we realized the reason for this is the addition of this parameter reduced R0 below 1. But why should this be so?

The reason is because STEM models disease mortality as a separate compartment, so this must be accounted for by changing either β or γ to increase R0 above 1 (to correspond with literature values in the 2014 Ebola outbreak). Since γ (and σ as well) are often derived from scientific literature in models such as ours, and since their reciprocals represent the duration of the Incubation and Infectious periods, respectively, these parameters cannot be changed without fundamentally changing the underlying disease model behavior!

This leaves us with no choice but modification of β (Transmission Rate). Luckily for us, if we do this properly, we get the correct values for our simulation. . . Fortunately, STEM calculates the basic reproductive rate R0 from a slightly different equation than in the published SEIR literature. In STEM's SEIR+D model (without birth/death dynamics), the equation looks as follows...

Neglecting Birth/Death population dynamics... We'd expect a time-invariant SEIR equation to be ...
R0 = β / γ

However, in STEM, this is actually calculated *with the mortality rate component as a rate term*.

R0 = β / (γ + δ)

Thus, when adding data from scientific literature to STEM for modelling in an SEIR model, a new βδ parameter must be recalculated to account for disease mortality , as follows...
βSTEM = R0 * (γ + δ).

Notice in the above image (a more accurate diagram of the STEM SEIR+D compartment model), we actually have a separate compartment called D, which represents Deaths from disease. This compartment is SEPARATE and DISTINCT from the R compartment . . . Thus in STEM, R means recovered, and only includes non-fatalities from disease.

To solve our problem, when taking a value of β from known literature to plug into STEM, β must be scaled appropriately to fit the STEM compartment model to include a Death compartment. Let the literature value of β be βL , and let the STEM value of β to be called βSTEM.

R0 = βL / (γ + δ) (literature)
βSTEM = R0 * (γ + δ). (STEM value of βδ )
R0-STEM = βSTEM / (γ + δ) (STEM true value of time-invariant R0)

More to come soon.

Subscribe to Operon Labs RSS