More folksy metaphors, please

Also in the 1937 Journal of Infectious Diseases, a pleasing approach to explaining the mechanism of bacterial clearance by phagocytes.

We don’t see many extended metaphors in scientific papers today. If you want to illustrate how something works, you don’t compare it to some other mechanism, you draw a simplified diagram with a bunch of arrows. But I like the more folksy approach of Arthur Locke of the Western Pennsylvania Hospital.

From Locke A (1937), Lack of Fitness as the Predisposing Factor in Infections of the Type Encountered in Pneumonia and in the Common Cold, J Inf Dis 60(1):106-112 (available from JSTOR, but subscription required).

I also like the use of the word “snuffles”.

Data Update: Pox in rabbits, pox in mice

Here’s a paper by Frank Fenner (1914-2010), which compares quite a lot of different poxviruses to establish basic facts about their biology.

By 1958, when this was published, Fenner was already well known in Australia for his leadership role in releasing extremely deadly myxoma virus among the rabbit population. This was deemed to be worth the risk, as the creatures had for almost a century been locust-like in their consumption of Australia’s crops, only making themselves useful as a source of food during the Depression. In a prelude to fellow Australian Barry Marshall’s auto-experimentation with Helicobacter pylori, Fenner and two other experts infected themselves with myxoma to show that the risk to humanity was negligible.

Fenner F (1958). The biological characters of several strains of vaccinia, cowpox and rabbitpox virus. Virology 5(3):502-529. (abstract here: subscription required for full article)

This paper was an attempt to clear up the categorization of various poxvirus strains, mostly vaccinia virus. Fenner got colleagues around the world to provide samples of Mill Hill (V-MH), Williamsport (V-WILL), Pasteur Institute (V-PI), and other vaccinia isolates, which had been designated as either “dermal” or “neuro-” vaccinia based on whether they had been propagated in rabbit skin or rabbit brain. Were these categories practically useful? What strains of vaccinia, if any, were really typical of the virus?

The vaccinia strains were compared to the “Amsterdam” and “Brighton” variants of cowpox, and the “Utrecht” and “Rockefeller Institute” variants of rabbitpox. Rabbitpox is not the same as myxoma virus, by the way.

* * *

Frank Fenner infecting eggs. Source: Sydney Medical School

Also included in the comparison were “white variants” of certain virus strains. At this time culturing cell lines in vitro was not convenient. Lots of procedures, including growing virus, that we would now do with cellular monolayers in petri dishes were done in a system called the chorioallantoic membrane (CAM). The CAM is the membrane under the shell of an egg. Before doing any experiments on the various viruses, Fenner and associates made sure each virus was pure, by inoculating eggs and then extracting a single “pock” from the CAM for further propagation.

Often they would see that a virus that normally produced red, angry pocks had a few white pocks. Under further study, these “white variants” generally turned out to be milder than their parent strains, as a result of some genetic deletion. The white variants gave me some trouble in putting together the graphs based on Fenner’s data, because they are grouped with non-white strains that are far more virulent.

Sample chorioallantoic membrane pocks. 16a is the “white variant” of 16 (Rabbitpox Utrecht). 15a is the “white variant” of 15 (Cowpox Brighton).

Let’s get to the data. After doing a bunch of CAM experiments, Fenner started infecting animals. He compared the viruses for their ability to kill mice and rabbits after brain infection, and induce skin lesions in rabbits after skin infection.

This table contains an immense amount of data, but it takes up two pages which is never a good thing. I think we can turn it into two good figures, one for mouse infection and one for rabbit infection.

First of all, not all this data needs to be graphed. If you look at the last two columns, every rabbit skin lesion of more than 13 mm in diameter is considered “IPC”, and every smaller lesion is “N”. That’s all we need to know. Also, I don’t think it makes sense to list a mean survival time for experiments where almost every mouse survived. It’s a little sketchy to say that mice that never came close to dying had a “survival time” of 14 days, and then average that arbitrary number with actual survival times of mice that died. So I’ll drop the “mean survival time” data as well.

* * *

My mouse figure has two parts. I didn’t do any statistics on the data.

1A shows the relative virulence, which is a somewhat confusing metric that compares the LD50 (dose of virus at which half of infected animals die) with the virus’s ability to cause pocks on the egg membrane. All of the viruses grow pretty well in eggs, so this shows which viruses are particularly adapted to be dangerous to mice (or rabbits).

1B shows how many mice survived brain infection with each virus strain. Nowadays we would show survival curves for this – you know, the lines that always start out horizontal with 100% survival, and as the experiment goes on, the line goes down and down, stepwise, as mice succumb to mortality. But Fenner just shows the percent of mice that died during infection (probably this means the number that were dead after 14 days). I changed this to become a graph that shows the % of mice that survived infection.

Figure 1. Mice are more susceptible to neurovaccinia and rabbitpox than dermal vaccinia. (A) For each strain of virus, mice were infected intracranially with a range of doses, and the LD50 was calculated as the dose which was lethal to 50% of infected mice. LD50 values are normalized to the number of CAM pocks produced by the same dose of virus in eggs. Horizontal line represents the limit of detection. (B) For each strain of virus, mice were infected intracranially with 100,000 infectious particles. Here we show the percent of mice surviving to 14 days post-infection (sample size = 10). Bars with white stars represent “white variants”.

The rabbit figure has three parts.

2A is the relative virulence data again. This time it’s not the LD50, but the ID50. (Even though it says “LD50” in Fenner’s table, it says “ID50” in the legend and in the text.) “LD” means lethal dose, whereas “ID” means infectious dose. So this is the dose of virus at which half of infected animals show signs of infection. “Infection” in this case would be a skin lesion. He doesn’t say exactly how big a lesion needs to be before it’s considered “infection”.

2B is a similar graph showing the diameter of the lesions, in millimeters.

And then there’s a bunch of survival curves. 2D is too busy to be able to decipher each datapoint, with eight overlapping lines in a single graph, but what’s important is that there’s a huge difference between 2C and 2D.

Figure 2. Rabbits are more susceptible to neurovaccinia and rabbitpox than dermal vaccinia. (A) For each strain of virus, rabbits were infected intradermally with a range of doses, and the ID50 was calculated as the dose which produced a lesion in 50% of infected mice. ID50 values are normalized as described in 1A. (B) For each strain of virus, rabbits were infected intradermally with 100,000 infectious particles. After 5 days of infection, the diameter of each skin lesion was measured with a ruler. (C-F) For each strain of virus, rabbits were infected intradermally with 100,000 infectious particles, and survival was monitored daily (sample size = 2-4).

What do you think? Let me know through email (see “About” page). Or through Twitter. Or subscribe via email or RSS (see bottom of this page). Or watch this German rap video.

Rabbit mosaic virus?

This paper (1) seemed weird, at first. With a title like “Immunologic Reactions with Tobacco Mosaic Virus“, and methods that included injecting the virus into rabbits, I wondered if it was an attempt to use all possible means to infect animals with the legendary plant virus. But no, the author of today’s paper points out in her introduction that that work was already done (2) by Maurice Mulvania of the Tennessee Experiment Station, and we now know (as of 1929) that the virus doesn’t reproduce in rabbits.

So what is she doing, looking at the rabbit’s immunologic reactions to tobacco mosaic virus (TMV)?

Well, primarily, trying to back up the evidence that the virus exists at all, as a physical object. The introduction to this paper is particularly interesting in its snapshot of what people thought about viruses, before electron microscopy.

The properties of tobacco virus are frequently linked to those of enzymes. An apparent similarity between them may be due to the fact that they have both been studied in tissue extracts and in an impure state.

In Angela Creager’s 2001 book The Life of a Virus: Tobacco Mosaic Virus as an Experimental Model, 1930-1965, a similar point is made:

Investigations of virus infectivity resembled contemporary biochemical studies of enzyme activity under various conditions. Allard, mentioned above for his insistence on the contagious nature of TMV, assessed the effects of heat, germicidal agents, and chemical reagents on viral infectivity… A few years later, Maurice Mulvania conducted similar studies assessing the effects of light rays, heat, dialysis, and exposure to bacteria on the infectivity of TMV, publishing his work in the other major journal for research on plant diseases, Phytopathology. The susceptibility of TMV to some of these inactivating agents led him to speculate that the virus might be a simple protein or enzyme.

So the word “virus” did not mean “virus” as we now know it, an organism with genetic material that can reproduce when it infects host cells. A “virus” could simply be a toxic substance. After all, “virus” comes from the Latin word for poison.

Helen Purdy, later Helen Purdy Beale, was a legendary figure in plant virology. Here is a 2006 review of her work (3) that mentions, right off the bat, this 1929 paper in which she immunized rabbits against TMV-infected plant saps and uninfected plant saps, and concluded that the virus saps had unique proteins which were not present in the normal plants. In the introduction she also gives a mini-literature review of what people thought TMV might be.

Verdict of history:

Light and electron microscopy of cauliflower mosaic virus in cells, from Martelli and Castellano, 1971 (5). “Ib” means inclusion bodies. Click for higher resolution.

• Egiz: Wrong!
• Hunger: Wrong! That sounds like a prion disease.
• Beijerinck: Right about the “contagium vivum” part, wrong about “fluidum”. Beijerinck is generally credited as the first scientist to demonstrate that an infection could be transmitted by something smaller than a bacterium.
• Heintzel/Woods/Chapman: Wrong!
• Iwanowski: Right about being unculturable, wrong about being bacterial.
• Allard: Right!
• Palm: Right! Those inclusions are made up of solid particles of the infectious virus.
• Olitsky: Wrong!

Even Palm didn’t contemplate a new form of “parasite” other than extra-tiny bacteria. His paper (4) was written in Dutch and published in the now-elusive Bull. Deliproefsta. te Medan-Sumatra. This publication is so elusive that I can’t figure out what “Deliproefsta” stands for. Among others interested in Palm’s findings, Dr. L.O. Kunkel provided a synopsis at the end of an article in the 1921 Bulletin of the Experiment Station of the Hawaiian Sugar Planters’ Association. (This is what came up first from a web search for “tobacco” and “deliproefsta”.

In a recent paper Palm has described certain “amoebiform corpuscles” which he finds in mosaic tobacco leaves. He apparently considers these bodies to be analogous to those of corn mosaic. In addition to the “larger foreign corpuscles,” which he says “lie either in intimate contact with the nucleus or more or less in its vicinity”, he finds numerous extraordinarily small granules which, he thinks, may represent an orgnaism. On the assumption that the granules are alive and that they are the same bodies which Iwanowski described as bacteria, he gives them the name Strongloplasma iwanowskii.

…

In considering the intracellular bodies as a possible cause of mosaic disease, the writer meets the same difficulty which prevented Iwanowski from looking on the “Plasmaanhaufungen” as causal organisms. They are too large to pass through bacterial filters. Iwanowski did not know, however, that these bodies are associated with mosiac disease in several different plants. Perhaps this further evidence may justify an attempt to harmonize statements which at first appear to be contradictory. It may be that the amoeboid bodies… represent only one stage in the life of a causal organism. At another stage they may be so small and plastic that they can pass through the fine pores of a filter and escape detection under the microscope. They probably become visible only after a certain period of growth within the host cell.

Bingo!

* * *

The 1929 Purdy paper has a remarkably coherent set of conclusions that agree with what we now know. Purdy did a large number of experiments that mostly involved immunizing rabbits with “virus-saps” and “normal saps” from tobacco, tomato, pepper and petunia plants. She then took the immune sera from these rabbits and saw if they could block the activity of the various saps. There’s a whole page of “summary” and conclusions. In brief:

• Tobacco virus-saps contain all the antigens that normal tobacco saps contain, plus a few others.
• Virus-saps from tobacco, tomato, pepper and petunia plants all have some antigens in common, which are probably not found in normal sap.
• Normal rabbit or guinea pig serum cannot block TMV from infecting plants. Neither can serum from animals immunized with normal plant sap. These sera do not contain the proper antibodies. But serum from animals immunized with virus-saps can block TMV from causing plant disease.

Well, that’s pretty conclusive. But these plant saps are mostly just soluble proteins. So we still don’t know if the virus is made of of particles, or if it is liquid in nature. Some sort of extra-sensitive microscope would be very helpful.

* * *

1. Purdy HA (1929). Immunologic reactions with tobacco mosaic virus. J Exp Med 49(6): 919-935.

2. Mulvania M (1926). Studies on the nature of the virus of tobacco mosaic. Phytopathology XVI: 853-872.
(note: I can’t find any archive that includes Phytopathology any earlier than 1970, which is a shame since a lot of papers and indices refer to Mulvania’s work).

3. Scholthof KBG, Peterson PD (2006). The role of Helen Purdy Beale in the early development of plant serology and virology. Adv Applied Microbiol 59: 221-241.

4. Palm BT (1922). Die Mosaikziekte vac de Tabakeen chlamydozoonose? Bull Deliproefsta te Medan-Sumatra xv: 7-10.

5. Martinelli GT, Castellano MA (1971). Light and electron microscopy of the intracellular inclusions of cauliflower mosaic virus. J Gen Virol 13: 133-140.