Fleshing out Yersinia pestis

Up until a few months ago there were a few representative samples of the Yersinia pestis genome. Important windows into its secrets, but windows none the less. In January a Chinese group remedied this situation by expanding the number of fully sequenced genomes from 15 to 133 (Cui et al, 2013).  China supplied 107 genomes selected from over 900 genotyped specimens collected since 1955 to represent bacterial and host diversity. To these, 11 additional isolates from Mongolia, Myanmar (Burma), the former Soviet Union, and Madagascar were fully sequenced. For the analysis, the previously sequenced 15 genomes were added bringing the total up to 133 including the ancient specimens from 14th century London.

The Core-Genome and the Pan-Genome

Even for a bacterium like Yersinia pestis that is considered to have little genetic diversity, its genome is more elastic than any eukaryote (everything but bacteria). The bacterial genome can be divided into its core genome, found in all members of the species, and the accessory genome, sequences found only in some strains. Plasmids are part of the accessory genome but not all of it. Extra genes are also found on the bacterial chromosome. The core genome is 3.53 Mb long with 3450 genes; the accessory genome has 1.92 Mb with 1249 genes (including 451 on the six known plasmids) (Cui et al, 2013, Table S1). So the accessory genome contains 26% of genes found in the species. This may seem like a lot, but more promiscuous species like Escherishia coli (E. coli) have many more accessory genes than core genes. With E. coli the more specimens that are sequenced, the larger the accessory genome gets with no end in sight.

Combining all of the genes found in Yersinia pestis (core and accessory genome), we have the pan-genome. The pan-genome is 5.46 Mb with 4699 genes (Cui et al, 2013).  No one strain has all of these genes. So different strains do have significant differences in their functions but, as far as I know, there are no significant differences in human prognosis. Hopefully, there will be more study in the future that cross-references strain type  or particular genes with human prognosis, transmission routes (% bubonic vs pneumonic), hosts etc.

Branching Out

Using known and new SNPs, the phylogenetic tree has finally been fleshed out into a healthy looking tree . We couldn’t keep the sickly looking Charlie Brown tree of the past forever! Even so, the tree below represents only the main branches.

Click to enlarge, (Cui et al, PNAS, 2013)

To my mind, the most important aspect of the new tree is that nodes of increased diversity are much more apparent. The authors are the most excited by node 7 where there is a four-way branch, adding two new branches  (3.ANT1 and 4.ANT1) to the main stem of the tree. They refer to this diversity point as the ‘big bang’. This node gains the most attention because the 14th century London genomes are just one step off of node 7 down the 1.ANT1 branch. So it stands the reason that node 7 represents a period of diversity that produced the second pandemic. Yet, looking at their diagram, other locations like node 12 have greater diversity. The 1.IN strains are intermediary on the same lineage between the second the third pandemic. Node 14 is the initial diversity that produced the third pandemic.  Calling node 7 a ‘big bang’ seems to me to have more to do with it producing the second pandemic rather than the diversity at the node itself. The new third and fourth branch (3.ANT and 4.ANT) are concentrated in Mongolia, putting emphasis on the importance of doing such deep sequencing in other Central Asian regions. It is impossible to tell which host species these bursts of diversity occurred within, almost certainly not humans. It’s not that diversity can’t be generated in humans especially during a pneumonic plague, but since it is not endemic in humans,  it must make it back to a reservoir to be preserved anywhere other than in ancient DNA.

Biogeography shows clustering of related strains in regions as would be expected, though they are fairly well mixed within the circled zone in the map above. Samples seem to follow ancient roads, although keep in mind all of these strains have been isolated within the last 60 years.   I do wonder why they were not able to identify a route for the eastern branch two isolates. All of the branch two isolates appear to be running along a fairly straight line from southwest to northeast China (extending trade route III to Manchuria). The 107 Chinese specimens were chosen from > 900 strains identified from 5000 isolates for their diversity revealed by genotyping, host diversity and geography (Cui et al, 2013). It would have been interesting to see a map with all 5000 on it as a measure of abundance (with or without typing).

The oldest strain 0.PE7 is found only on the Qinghai-Tibet plateau in China, an area framed by the ancient trade routes along which most of the western strains are found. This has led Cui et al, 2013 to postulate that the  Qinghai-Tibet plateau as the origin of  Yersinia pestis.

Unsteady Molecular Clocks

Estimating ages from genetics can be a very risky business. To estimate years since the last common ancestor, it requires a steady molecular clock , measured in base changes per unit of time. In theory all of the genes from the core genome should have changed to the same degree from the common ancestor, but that is not the case at all. The number of SNPs in the Yersinia pestis core genome varies greatly. Even excluding the most divergent Angola (0.PE3) strain, there is “a nearly 40 fold difference between the slowest and the fastest evolving branches” (Cui et al, 2013). An unsteady molecular clock was also suggested by previous data from Madagascar, though the discussion was buried in the supplementary material (Morelli et al, 2010, p. S10-s18). Mutator phenotypes do occur (Rajanna et al, 2013), though Cui et al, 2013 assure us that none of these strains are mutators.  On the other hand, a Georgian group suggest that the mutator phenotype, a single point mutation, could naturally reverse (back mutate) altering the predictability of the lineage age (Rajanna et al, 2013). The Chinese group concluded that the faster clock rates for some branches are due to a higher reproduction rate, probably due to more or larger epidemics in the lineage (Cui et al, 2013). The types of genetic changes (SNPs) indicate neutral selection, so the increased reproduction rate is not due to the genetic changes.

While I understand that calculating divergence dates an important exercise to people who focus on phylogenetics, for the understanding of historical plague it is not useful. It is not solid or specific enough to base historical events upon alone. Predictions are just that; all of these groups have been proven wrong, sometimes later by themselves, too often.  Most importantly, it appears that it will eventually be trumped by ancient DNA analysis with an archaeological and/or documentary context. As far as I’m concerned, the Angola strain is a genetic and geographic outlier of uncertain provenance. We don’t know important factors like how long it was kept in active culture before it was made into a stock or the conditions of storage. Both of these can effect mutation rates and the molecular clock (Rajanna et al, 2013).  I’m sure the Angola strain’s story is interesting but unlikely to be useful for understanding the whole species unless it turns up in ancient DNA.

Gaining and Loosing Diversity

Returning to these starburst points on the tree, called polytomys, where multiple lineages share the same ancestor, we have some of the most valuable information in the new phylogenetic tree. Epidemics (and presumably epizootics) are believed to have an increased reproduction rate over enzootic plague. Since the mutation rate is directly tied to the reproduction rate, increased reproduction rates predict an increased mutation rate and, therefore, production of genetic diversity.  The team predicts that “higher clock rates are an indicator of epidemic disease, even in the absence of historical evidence” (Cui et al, 2013). It is unclear how an epidemic can be differentiated from an epizootic by genetics alone. We know from modern observations that not all epizootics spill over into the human population. Yet, major polytomys can at least be used to estimate how many bursts of growth the bacterium has gone through in China. We should see other polytomys with increased sequencing of other Central Asian regions.

While these polytomys show a starburst of new lineages, there is also a loss of diversity during every epidemic. Most of the new lineages produced during an epidemic (or epizootic) will die out (become extinct) when the epidemic ends. If the changes are truly neutral, then which lineage survives to endure in the reservoir will be completely random (as will be the number of surviving lineages). We should also remember that clinical isolates  during an epidemic and ancient DNA can preserve lineages that become extinct (and this is normal). In the four individuals they sequenced from 14th century East Smithfield, they found two different clones, with the second being derivative of the first. Both of these clones may only be found in ancient DNA, not in any living specimen. The more time that passes the greater the likelihood that the minor lineages will become extinct. This tends to make the earlier sections of the pylogenetic tree look cleaner by stripping off side branches.

Another recent study by Vogler et al (2013), supports their scenario on a finer scale during the 9 year epidemic in a port town of Mahajana,  Madagascar from 1991 to 1999. Over a decade we can compare the incidence of plague vs. the genetic diversity. Yersinia pestis evolution can be plotted with great precision. In the lower diagram, clones are color coded to the year of isolation. From 1995 to 1999 it is possible to see the next year’s primary clone emerge in the previous year’s epidemic, which supports local cycling within the city. At the same time, most of the diversity generated is not represented in later outbreaks.

Vogler et al, 2013

Vogler et al, 2013

Host Diversity

Host genus vs Y. pestis strain collected (Cui et al, 2013).

The hosts of these 107 strains give us a glimpse into the host diversity for Yersinia pestis within China (Cui et al, 2013). The figure to the right gives an indication of strain diversity within each host but does not tell us abundance or location within China. What jumps out at me, is that humans and marmots have the most strain diversity. The high strain diversity in humans including 0.PE7, the strain closest to the most recent common ancestor, suggests to the Chinese team that Yersinia pestis has been pathogenic to humans since it evolved (Cui et al, 2013). Thus, at no point in its evolution did it gain the ability to infect humans. The few strains that can not infect humans are hypothesized to have lost their ability to infect humans possibly as a function of purifying selection for voles as hosts. It is interesting that the 1.ORI strains of the third pandemic are only found in humans, rats and mice.  We have to be careful about taking this figure to represent abundance or importance of a particular host. The great gerbil, Rhombomys opimus, is a primary host throughout central Asia is is represented by only one strain in this figure.

Studies published this winter have moved us significantly down the road to fleshing out Yersinia pestis. The genetic survey of Y. pestis in China provides a firm foundation to build on as more ancient DNA becomes available and extensive sequencing is done in other regions. Madagascar continues to be the best laboratory for plague ecology and epidemiology, while the Georgian study begins to address unintended intra-laboratory evolution that may shed light on Y. pestis in the wild. I’ll return to these papers again soon as I continue to examine Y. pestis from different perspectives and ruminate on answers to other questions.

References:

Cui, Y., Yu, C., Yan, Y., Li, D., Li, Y., Jombart, T., Weinert, L., Wang, Z., Guo, Z., Xu, L., Zhang, Y., Zheng, H., Qin, N., Xiao, X., Wu, M., Wang, X., Zhou, D., Qi, Z., Du, Z., Wu, H., Yang, X., Cao, H., Wang, H., Wang, J., Yao, S., Rakin, A., Li, Y., Falush, D., Balloux, F., Achtman, M., Song, Y., Wang, J., & Yang, R. (2013). Historical variations in mutation rate in an epidemic pathogen, Yersinia pestis Proceedings of the National Academy of Sciences, 110 (2), 577-582 DOI: 10.1073/pnas.1205750110

Morelli G, Song Y, Mazzoni CJ, Eppinger M, Roumagnac P, Wagner DM, Feldkamp M, Kusecek B, Vogler AJ, Li Y, Cui Y, Thomson NR, Jombart T, Leblois R, Lichtner P, Rahalison L, Petersen JM, Balloux F, Keim P, Wirth T, Ravel J, Yang R, Carniel E, & Achtman M (2010). Yersinia pestis genome sequencing identifies patterns of global phylogenetic diversity. Nature genetics, 42 (12), 1140-3 PMID: 21037571

Rajanna C, Ouellette G, Rashid M, Zemla A, Karavis M, Zhou C, Revazishvili T, Redmond B, McNew L, Bakanidze L, Imnadze P, Rivers B, Skowronski EW, O’Connell KP, Sulakvelidze A, & Gibbons HS (2013). A Strain of Yersinia pestis With a Mutator Phenotype from the Republic of Georgia. FEMS microbiology letters PMID: 23521061

Vogler, A., Chan, F., Nottingham, R., Andersen, G., Drees, K., Beckstrom-Sternberg, S., Wagner, D., Chanteau, S., & Keim, P. (2013). A Decade of Plague in Mahajanga, Madagascar: Insights into the Global Maritime Spread of Pandemic Plague mBio, 4 (1) DOI: 10.1128/mBio.00623-12

Reactivation of Ancient Plague Foci in Libya, 2009

Landscape around Oran, Algeria (2003), and Tobruk, Lybia (2009) that produced plague cases. (Cabanel et al, 2013)

Plague has been called a re-emerging disease primarily because cases have begun to appear in areas where plague has been absent for decades. Two recent surprising outbreaks occurred in Algeria, where plague had been absent for over 50 years, and in Libya after a 25 year absence. A team led by the Institut Pasteur explored possible relationships between the recent Libyan outbreak and the Algerian outbreaks. All of the information in this post comes from their report to be published in the February issue of Emerging Infectious Diseases (citation and link below).

The outbreaks under consideration were just south of Oran, Algeria in 2003, at Lanhouat, Algeria in 2008 and near Tobruk near the Libyan-Egyptian border in 2009. Another possible outbreak of plague occurred at Tobruk during the Libyan revolution in 2011.  Political unrest prevented a complete disease investigation of the 2011 Libyan epidemic. Past Libyan plague outbreaks have occurred from 1913-1920, 1972, 1976, 1977, and 1984. The largest outbreak in 1917 is credited with 1,449 deaths.

The 2009 Libyan index cases consisted of three children from one nomad family; one child died after two days of intensive care and the other two eventually recovered. Only one child had a tender cervical node. The other two, including the child who died, had signs of a severe infection but no visible buboes. The father reported having axillary lymphadenitis and a couple of sudden deaths in the region in the previous two months. A week after admission Libyan authorities reported 13 possible cases to the World Health Organization and requested assistance. The WHO-Libyan team identified two more women with painful inguinal nodes and “infectious syndrome”, but also concluded the initial estimate overstated the number of cases. There are five confirmed cases. The cases were spread 30-60 km from the index family’s home in Eltarsha, 30 km south of Toburk. Regional response included antibiotic treatment of contact persons, and insect and rodent control measures. No further cases were reported.

Diagnosis was confirmed by standard bacteriological assays and molecular characterization. All five confirmed cases were positive with the F1 antigen dipstick.  Yersinia pestis cultures were isolated from three patients,  all phenotyped to the Medievalis biovar by metabolic assays. Molecular characterization confirmed that all are the same Medievalis strain. Hybridization analysis indicates that it is most closely related to, but distinct from, strains isolated from Iranian Kurdistan in 1947 – 1951.

Using the same methods, the 2003 Algerian isolates were phenotyped to the Orientalis biovar. Molecular characterization confirmed that they are all related but not identical Orientalis strains. Activation of multiple related strains from an ancient foci in the same year suggests an environmental trigger. Comparing the 2003 strains to those isolated in 1944 and 1945 illustrate the complexities of plague foci. The 1944 isolate is a Orientalis strain that belongs to the same cluster of strains as the 2003 isolates and other strains from Morocco and Senegal.  The 1945 strain matched a molecular characterization of  Orientalis isolates from Saigon, Vietnam and is believed to have been transmitted by military transports during World War II.  Cabanel et al conclude that the 2003 Algerian outbreaks were caused by local Yersinia pestis strains. It should be noted that the third pandemic from the turn of the 20th century was a Orientalis biovar (1.Ori1).

Cabanel et al. note this is the only instance they could find of a Medievalis strain in Africa. The spread of cases over a 30-60 km region and isolation of related but different strains support the reactivation of an ancient plague focus. Unfortunately they did not have access to isolates from previous 20th century Libyan outbreaks (if they exist) that could have provided more certainty.

Reactivation of plague foci around the Mediterranean has been associated with climate change. They note that an unusually humid winter and good crops in Libya in 2009 favored rodent and flea abundance. Long dormancies may be part of Yersinia pestis’ natural history particularly in resource limited environments. This possibility will be one of the topics of my next post.

Cabanel et al. note that camel meat and livers have been associated with human plague cases in Libya (1976), Saudi Arabia (1994), Jordan (1997), and Afghanistan (2007). Additional local evidence suggested that the highly susceptible camels contracted the plague from local foci in these instances. Although camels do not survive plague long enough to transmit it very far, camel caravan routes may still have played a role in transmission if only by the other organisms also along the camel caravan route. Camels would have provided an abundant host to amplify the organism along the route. Camel fleas could have been carried among the cargo not unlike rat fleas in ship cargoes. Camel caravans would provide an ancient route for a Medievalis strain to reach Libya from the central Asia.

Reference

Cabanel, N., Leclercq, A., Chenal-Francisque, V., Annajar, B., Rajerison, M., Bekkhoucha, S., Bertherat, E., & Carniel, E. (2013). Plague Outbreak in Libya, 2009, Unrelated to Plague in Algeria Emerging Infectious Diseases, 19 (2), 230-236 DOI: 10.3201/eid1902.121031

Toward a Molecular History of Yersinia pestis (AHA)

This post a resource for the presentation I gave at the AHA meeting in New Orleans on January 5, 2013. A color handout of the slides can be downloaded here.

This map will be continually updated as new finds are published. Some of the balloons mark sites with multiple studies. Click on the balloons for citations.

View Larger Map

References:

Achtman, M. (2012). Insights from genomic comparisons of genetically monomorphic bacterial pathogens. Philosophical Transactions of the Royal Society B: Biological Sciences, 367(1590), 860–867. doi:10.1098/rstb.2011.0303

Bos, K. I., Schuenemann, V. J., Golding, G. B., Burbano, H. A., Waglechner, N., Coombes, B. K., et al. (2011). A draft genome of Yersinia pestis from victims of the Black Death. Nature, 1–5. doi:10.1038/nature10549

Bos, K. I., Stevens, P., Nieselt, K., Hendrik N Poinar, DeWitte, S. N., & Krause, J. (2012). Yersinia pestis: New Evidence for an Old Infection. PLoS ONE, 7(11), e49803.

Drancourt, M., & Raoult, D. (2005). Palaeomicrobiology: current issues and perspectives. Nature Reviews Microbiology, 3(1), 23–35. doi:10.1038/nrmicro1063

Drancourt, M., Houhamdi, L., & Raoult, D. (2006). Yersinia pestis as a telluric, human ectoparasite-borne organism. The Lancet Infectious Diseases, 6(4), 234–241. doi:10.1016/S1473-3099(06)70438-8

Haensch, S., Bianucci, R., Signoli, M., Rajerison, M., Schultz, M., Kacki, S., et al. (2010). Distinct Clones of Yersinia pestis Caused the Black Death. (N. J. Besansky, Ed.)PLoS Pathogens, 6(10), e1001134. doi:10.1371/journal.ppat.1001134.t001

Houhamdi, L., Lepidi, H., Drancourt, M., & Raoult, D. (2006). Experimental model to evaluate the human body louse as a vector of plague. The Journal of Infectious Diseases, 194(11), 1589–1596. doi:10.1086/508995

Little, L. K. (2011). Plague Historians in Lab Coats*. Past & Present, 213(1), 267–290. doi:10.1093/pastj/gtr014

Malou, N., Tran, T.-N.-N., Nappez, C., Signoli, M., Le Forestier, C., Castex, D., et al. (2012). Immuno-PCR – A New Tool for Paleomicrobiology: The Plague Paradigm. (S. Bereswill, Ed.)PLoS ONE, 7(2), e31744. doi:10.1371/journal.pone.0031744.g006

Morelli, G., Song, Y., Mazzoni, C. J., Eppinger, M., Roumagnac, P., Wagner, D. M., et al. (2010). Yersinia pestis genome sequencing identifies patterns of global phylogenetic diversity. Nature Genetics. doi:10.1038/ng.705

Nguyen-Hieu, T., Aboudharam, G., Signoli, M., Rigeade, C., Drancourt, M., & Raoult, D. (2010). Evidence of a Louse-Borne Outbreak Involving Typhus in Douai, 1710-1712 during the War of Spanish Succession. PLoS ONE, 5(10), e15405. doi:10.1371/journal.pone.0015405

Parkhill, J., Wren, B. W., Thomson, N. R., Titball, R. W., Holden, M. T., Prentice, M. B., et al. (2001). Genome sequence of Yersinia pestis, the causative agent of plague. Nature, 413(6855), 523–527. doi:10.1038/35097083

Pusch, C. M., Rahalison, L., Blin, N., Nicholson, G. J., & Czarnetzki, A. (2004). Yersinial F1 antigen and the cause of Black Death. The Lancet Infectious Diseases, 4(8), 484–485. doi:10.1016/S1473-3099(04)01099-0

Raoult, D., Dutour, O., Houhamdi, L., Jankauskas, R., Fournier, P.-E., Ardagna, Y., et al. (2006). Evidence for louse-transmitted diseases in soldiers of Napoleon’s Grand Army in Vilnius. The Journal of Infectious Diseases, 193(1), 112–120. doi:10.1086/498534

Schuenemann, V. J., Bos, K., Dewitte, S., Schmedes, S., Jamieson, J., Mittnik, A., et al. (2011). PNAS Plus: Targeted enrichment of ancient pathogens yielding the pPCP1 plasmid of Yersinia pestis from victims of the Black Death. Proceedings of the National Academy of Sciences, 1–22. doi:10.1073/pnas.1105107108

Tran, T., Forestier, C., & Drancourt, M. (n.d.). Brief communication: Co‐detection of Bartonella quintana and Yersinia pestis in an 11th–15th burial site in Bondy, France. American Journal of ….

Tran, T.-N.-N., Signoli, M., Fozzati, L., Aboudharam, G., Raoult, D., & Drancourt, M. (2011). High throughput, multiplexed pathogen detection authenticates plague waves in medieval venice, Italy. PLoS ONE, 6(3), e16735. doi:10.1371/journal.pone.0016735

Wiechmann, I., & Grupe, G. (2004). Detection ofYersinia pestis DNA in two early medieval skeletal finds from Aschheim (Upper Bavaria, 6th century A.D.). American Journal of Physical Anthropology, 126(1), 48–55. doi:10.1002/ajpa.10276

Wiechmann, I., Harbeck, M., & Grupe, G. (2010). Yersinia pestis DNA Sequences in Late Medieval Skeletal Finds, Bavaria. Emerging Infectious Diseases, 16(11), 1806–1807.

AHA 2013: The Power of Cartography: Remapping the Black Death in the Age of Genomics and GIS

This coming year’s American Historical Association meeting will be held in New Orleans, Jan 2 – Jan 6, 2013. The schedule went online within about the last day. I’ll be at the AHA meeting for the first time this year at this session. The links below are to the session and individual speaker abstracts. I think it will be a great session and hopefully something more permanent will come of it.

The Power of Cartography: Remapping the Black Death in the Age of Genomics and GIS

AHA Session 143
Medieval Academy of America 4

Saturday, January 5, 2013: 9:00 AM-11:00 AM

Bayside Ballroom A (Sheraton New Orleans)

Chair: Nukhet Varlik, Rutgers University–Newark

Papers:

Remapping the Black Death
David Mengel, Xavier University

Toward a Molecular History of Yersinia pestis
Michelle Ziegler, Saint Louis University

In Search of the Black Death in Central Eurasia
Uli Schamiloglu, University of Wisconsin–Madison

Plague Patterns in Fifteenth Century Milan
Ann Carmichael, Indiana University Bloomington

The Public Practice of History in and for a Digital Age (session theme tag)

Dr Seuss Does Malaria

NEWSMAP, United States Military, 8 Nov. 1943

This Malaria map was illustrated by Theodor Seuss Geisel, better known as Dr. Seuss, during World War II to educate young GIs. According to the Naval Department Library, this map was printed on the back of a Newsmap (two sided poster) that showed the five war fronts in 1943: Russia, Italy, “air offensive”, southwest Pacific and Burma.

The text as transcribed by the Navy Department Library reads:

THIS IS ANN…..she drinks blood!

Her full name is Anopheles Mosquito and she’s dying to meet you. Her trade is dishing out MALARIA! If you’ll take a look at the map below you can see where she hangs out.

She can knock you flat so you’re no good to your country, your outfit or yourself. You’ve got the dope, the nets and stuff to lick her if you will USE IT.

Use a little horse sense and you can lick Ann. Get sloppy and careless about her and she’ll bat you down just as surely as a bomb, a bullet or a shell.

This text is taken from a booklet done by Theodor Geisel to train soldiers or sailors during the war. Cartoon educational materials were probably fairly effective. Newspaper cartoons were very popular at the time and the average GI was very young. Many (if not most) of these soldiers/sailors dropped out of school, often in grade school, to work during the Great Depression so educational materials had to be targeted at a lower reading level that modern military materials.

Since this is now a 60+ year old government publication, I will assume that its public domain. I found this at the Young Dipterists website. “What to do about Ann” was apparently a header on each page after the first page. I’ve tried to reproduce it as well as I could. The site appears to be missing the last page(s) since the last page has a “turn the page”.

When Yellow Fever Came to the Americas

“Yellow Jack”, Cornhill Mag., 1892

In the early Americas, nothing scared people more than when Yellow Jack came knocking at the door of their city. Yellow Jack, or as we know it better today Yellow Fever, has rightly been called the plague of the Americas.

It has long been assumed that yellow fever came to the Americas with its vector, Aedes aegypti, in the hold of slave ships. These ships would have been an irresistible feast to the mosquito. Yet, little was known about the origin, locations, and dates of transmission to South America. Juliet Bryant, Edwarld Holmes and Alan Barrett (2007) looked to DNA analysis of yellow fever virus (YFV) strains from 22 countries ( 14 African and 8 South American) to resolve and date the phylogentic tree for YFV. They analyzed 133 isolates from humans and animal hosts collected over a 75 year period.

Bryant, Holmes and Barrett (2007: e75) made four clear observations.

1. The American strains represent a single clade (monophyletic).
2. There are two distinct sub-clades in east and west South America respectively.
3. The South American clade is most similar to the West African isolates.
4. The East African clade is the most distinctive.

These observations support an east or central African origin for the Yellow Fever Virus dominated by enzootic transmission. Its development parallels the transmission of its vector Aedes aegypti.

The split between the east and west African clades has been calculated to an average distance of 723 years (roughly 1284 AD). The West African isolates are the most diverse in Senegal, suggesting this was an early focus for West African YSF. From West Africa Yellow Fever was transmitted to Brazil a calculated average of 470 years ago (roughly 1537 AD). Early Portuguese seamen frequented this part of Africa and Brazil was their largest colony, founded in 1500. This suggests that Yellow Fever was transmitted to Brazil virtually from the beginning of the Portuguese colony. It is possible that Yellow Fever was one of the imported diseases brought by the Portuguese that decimated native Brazilians before large-scale importation of Africa slaves. The South American clade split into eastern and western populations when it was transmitted to Peru a calculated average of 306 years ago (roughly 1700). There is no evidence of transmission back to Africa or other areas where Aedes aegypti have spread in Asia. Byrant, Holmes and Barrett (2007) argue that sylvatic transmission is the primary means of maintaining YSF in South America. They note that there hasn’t been an urban epidemic of YSF in South America since 1928, unlike the annual urban outbreaks in West Africa.

Auguste et al (2010) confirmed the overall structure of the YSF phylogenetic tree in the Americas, including its Brazilian origin in the Americas. Their analysis of strains collected over the last decade also confirm that Brazil is the reservoir and origin for most strains in the Americas today with the Peruvian strains remaining primarily localized in Peru and neighboring Bolivia. The analysis of Auguste et al (2010) also supports enzootic maintenance and local evolution in areas of spread from Brazil such as Trinidad and Columbia.

What I find most surprising about the YSF tree is its relative youth. This all suggests that Yellow Fever originated in the Middle Ages and probably did not circulate outside of local areas of central Africa until the late medieval period. We still have a lot of learn about the landscape epidemiology of yellow fever including possible vertical transmission among mosquitoes and the importance of difference primate species as reservoirs. Although we have had an effective vaccine for decades, yellow fever is still a very clear and present danger in both the Americas and Africa.

References:

J E Bryant, E C Holmes, & A D T Barrett (2007). Out of Africa: A Molecular Perspective on the Introduction of Yellow Fever Virus into the Americas PLOS Pathogens, 3 (5) : doi:10.1371/journal.ppat.0030075

Auguste, A.J., Lemey, P., Pybus, O.G., Suchard, M.A., Salas, R.A., Adesiyun, A.A., Barrett, A.D., Tesh, R.B., Weaver, S.C. & Carrington, C.V.F. (2010). Yellow Fever Virus Maintenance in Trinidad and Its Dispersal throughout the Americas, Journal of Virology, 84 (19) 9977. DOI: 10.1128/JVI.00588-10

The Landscape of Super-Spreading

Super-spreading individuals and disease hot spots have been known for over a century, but rarely have they been considered together. Sara Paull and colleagues [1] have pulled together all of the recent work the ecology of disease hot spots and transmission heterogeneity (super spreading) to explore the continuum between individual transmission heterogeneity and the landscape of disease, otherwise known as disease ecology. Hawley and Altizer [2] have developed the interrelationships further to the cellular level by linking disease ecology directly to ecological immunology, examining the effects of the environment on the immune system of  host populations. Taken together, these fields can be developed into a continuum of macro to micro landscapes  with transmission heterogeneity, better known as  super-spreading, at its center.

Adapted from Hawley and Altizer (2011). (Click to enlarge)

Disease Ecology

Disease ecology is the study of pathogens in the context of their dynamic environment. Unlike most medical experiments conducted in vitro, removing or stabilizing as many variables as possible, disease ecology seeks to understand the place of the pathogen in an an open, responsive environment with an unknown number of variables [3]. In disease ecology, multiple hosts and vectors of zoonotic diseases are considered together as a complex system. It is normal for the system to respond in unpredictable ways and for multiple stable states to be possible; the human gut microbiome consists of up to a thousand bacterial species that can come to at least three stable states (enterotypes) unrelated to easily identified variables (age, gender, ethnicity, etc.) [3].

Some of the basic areas of disease ecology include understanding the multiple hosts of a pathogen including factors effecting the lifecycle of the hosts, ecological drivers, endemic/enzootic and epidemic/epizootic dynamics,  characterization of the rural or urban landscape, and transmission dynamics. The US Environmental Protection Agency (EPA) defines an ecological driver as any biotic or abiotic factor that causes change to the ecological landscape. Ecological drivers can be as diverse as invasive species, climate change or changes to a man-made landscape. Transmission dynamics is also a basic factor in disease ecology.

Transmission Heterogeneity

Transmission between hosts is rarely, if ever, a uniform process. Extreme cases of transmission heterogeneity, where a small number of individuals are responsible for most successful transmissions, are called super-spreaders. These super-spreaders are major factors in epidemic dynamics of an increasing number of human pathogens. Drilling down into why transmission dynamics are heterogeneous yields a variety of causes.

Species level transmission heterogeneity has led to the recognition of ‘amplifying species’ that act as super-spreaders within the landscape.  American Robins as preferential hosts and transmitters of West Nile Virus are a classic example of an amplifying species [1]. Presumably changes in the landscape could alter the degree to which each possible host species contributes to the overall transmission within the environment and which host becomes the amplifying species. It would naturally follow that not all possible vectors are equally responsible for disease transmission. For example, we know that rat fleas, cat fleas, human fleas and possibly human lice can all transmit Yersinia pestis, but they are not equally responsible for transmission.

Landscape contributes to transmission heterogeneity as well. Studies have shown that oak tree vulnerability to sudden oak death syndrome varied with tree genotype but in field tests temperature and rainfall differences  mattered more than the genetics of individual trees [1]. These landscape characteristics go beyond the classical definitions of disease ‘hot spots’. Typical hot spots are defined as areas where there is higher pathogen levels or more amplification hosts [1].  For example, buffalo wallows and sites of buffalo deaths can become anthrax hot spots. Human have created animal disease hot spots; chronic wasting disease is increased among Colorado mule deer near human settlements rather than more remote areas [1]. In Belize, agricultural run-off alters the aquatic plant communities increasing breeding efficiency of the malaria vector Anopheles vestitipennis [1]. Some environmental hot spots can be so strong that they are “analogous to superspreaders” [1].

Which comes first, the hot spot or the super-spreader? Paull et al  note that “logically, superspreaders create hotspots of transmission around them, and hosts in a disease hotspot, by definition, experience an increased infection pressure as compared with others in the population” [1]. Ultimately what really matters is understanding why there is increased transmission. Does the ‘hot spot’ move with the hosts or is it tied to a locality or architectural structure? Some ‘hot spots’, like ships, are capable of moving although the increased transmission is tied to a piece of architecture. Likewise, some types of air conditioning systems can become ‘hot spots’ for Legionellosis, but redesign of the air conditioning system eliminates the ‘hot spot’.

Ecological Immunity

Ecological immunity takes factors of disease ecology and heterogeneous transmission particularly to the cellular and molecular level. For example, high testosterone levels suppress the male immune system (in birds and rodents) while triggering behavior that increases their territory and contact rates making them behavior super-spreaders [1,2]. It has been suggested that testosterone levels could be “a common driver of superspreading phenomena, broadly linking within- and among-host processes across vertebrate host taxa” [2].

The role of the environment and immunologic costs intersect when we look at seasonal effects and nutritional needs of the immune system. Malnutrition depresses the immune system directly and indirectly through hormones like leptin. The cost of immunity comes home when we realize that leptin intentionally cripples the immune system as part of the starvation response. Seasonal effects are easier to understand in animals that have a more defined seasonal pattern of migration, or a defined mating period [2]. Yet human infectious disease fluctuates with the season; influenza being a prime example of a disease with seasonal fluctuations.

Coinfections are also best understood for their impact at the cellular level of the immune system. The cost of the immune response makes immune trade-offs necessary during coinfections.  Focus has zeroed in on a classic immune trade-off between production of T-helper 1  and T-helper 2 cells as at least a partial explanation of poor prognosis for  helminth (worm)  and an intracellular pathogen coinfections [2].  Several studies have now shown that HIV infections progress faster if there is a helminth coinfection [2]. Likewise, in African buffalo helminth infections correlate with tuberculosis vulnerability and rapid progression; “Joelles et al (2008) found evidence for a direct trade-off between circulating levels of IFNγ (Th1 response) and eosinophils counts, their measure for Th2-type response in buffalo” [2]. To decrease vulnerability and improve prognosis for HIV and tuberculosis, treatment for helminths should become a priority.

What constitutes immune fitness must also be reevaluated in light of ecological immunity findings. The high cost of immune defenses are such that unnecessary defenses for the current environment are generally not maintained [2]. Likewise, costs and trade-offs may make it necessary for the immune system to develop tolerance to some pathogens deemed low risk to fight other greater risk challenges. Hawley and Altizer  note that “a ‘competent’ immune system may employ tolerance strategies, making immune ‘competence’ difficult to assess via infection outcome alone” [2].

Today we are only at the very beginning of understanding complex disease ecologies. For zoonotic diseases the impact of ecological drivers and enzootic dynamics all around the world make the identification of the best predictors an imperative. Yet, outbreak prediction will remain an imprecise science, not unlike the prediction of hurricanes or volcanic eruptions.  Increased surveillance will remain key to a rapid response and epidemic mitigation.

References:

[1] Paull, S., Song, S., McClure, K., Sackett, L., Kilpatrick, A., & Johnson, P. (2012). From superspreaders to disease hotspots: linking transmission across hosts and space Frontiers in Ecology and the Environment, 10 (2), 75-82 DOI: 10.1890/110111

[2] Hawley, D., & Altizer, S. (2011). Disease ecology meets ecological immunology: understanding the links between organismal immunity and infection dynamics in natural populations Functional Ecology, 25 (1), 48-60 DOI: 10.1111/j.1365-2435.2010.01753.x

[3] Pepper, J.W. & Rosenfeld, S. (2012). The emerging medical ecology of the human gut microbiome, Trends in Ecology & Evolution, 27 (7) 384. DOI: 10.1016/j.tree.2012.03.002