View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector Cell Host & Microbe Previews

Where Does Innate Immunity Stop and Adaptive Immunity Begin?

Junaid Ziauddin1 and David S. Schneider1,* 1Department of Microbiology and Immunology, Stanford University, Stanford, CA 94304, USA *Correspondence: [email protected] http://dx.doi.org/10.1016/j.chom.2012.10.004

The regulation of alternative splicing in the immune effector Dscam reported by Dong et al. (2012) in this issue of Cell Host & Microbe raises important questions about the nature of immune responses. Can we clearly define ‘‘adaptive’’ as being different from ‘‘innate’’ immunity, or is it time for a more flexible description?

Many papers on invertebrate immunity encoded while adaptive immune effectors specific NK cells directed toward fighting contain a statement that invertebrates are not. This implies that an innate system cytomegalovirus (CMV) fits this sort of are ideal for studying innate immunity has a limited capacity to generate specific model (Sun et al., 2009). Finally, there because they lack adaptive immune effectors whereas the adaptive immune are responses that are trained by past responses. In actuality, invertebrates system can create bespoke effectors experience and increase upon a second have spectacularly plastic immune effec- over the course of the host’s life span. exposure. Antibody production following tors that can generate true novelty and This approach doesn’t create two simple a secondary infection is an obvious functional immune response changes in categories; there is a continuum of example of this sort of training. Neuro- relation to past experience. The paper possibilities and no clear boundary that immune reflexes could be placed pro- by Dong et al. (2012) in this issue of Cell can be used to draw a border between vocatively at this end of the spectrum Host & Microbe concerning the mecha- adaptive and innate (Figure 1). We con- as they open the possibility of neural nism behind alternative splicing of an sider five overlapping categories along memory affecting the immune response immune effector, Dscam, adds a mecha- this continuum: first, there are single (Tracey, 2009). There is already evidence nistic dimension to the observed plas- immune effectors, like lysozyme or reac- for this in C. elegans (Zhang et al., 2005). ticity of invertebrate immunity. However, tive oxygen, which work nonspecifically All of this is relevant to the Dscam gene perhaps the most important contribution against a class of microbes. Second, in insects because Dscam fits poorly into of the Dong et al. (2012) study is the there are clouds of immune effectors our classical descriptions of immunity. demonstration of how difficult it is to comprised of many effectors, like antimi- Considerable work has been published divide immune responses into strictly crobial peptides or natural antibodies. on the role of Dscam molecules in wiring ‘‘innate’’ and ‘‘adaptive’’ properties. Third, there are effectors like antibodies the Drosophila nervous system, but Why might authors write that flies, or VLRs that can be produced via recom- Dscam serves a second role as an worms, snails, or sea urchins lack an bination or gene conversion. Fourth, anti- immune effector, and this has been less adaptive immune response? Typically, bodies and snail FREPS can be altered well studied (Dong et al., 2006; Watson the argument takes the following form: B through somatic mutation to further et al., 2005). The Dscam gene in inverte- and T cells generate our adaptive immune increase the range of specificity (Zhang brates is a complex gene that undergoes responses, therefore organisms with B et al., 2004). Finally, there is RNAi, which alternative splicing. Theoretically, the and T cells likely also have an adaptive produces highly specific antiviral effec- Dscam gene can generate up to 38,016 immune response. That is good logic. tors de novo. This is arguably the highest splice variants in dipterans (Dong et al., Organisms lacking B and T cells are often form of specificity found in an immune 2006; Watson et al., 2005). Dscam can said to lack adaptive immunity, but it system because it produces tailor-made be secreted where it can potentially act is a logical fallacy to conclude that just effectors from the pathogen and has no as an opsonin, and in a membrane-bound because an organism lacks these cells, template in the host. This is as far from form it can potentially act as a phagocytic the organism will also lack an adaptive innate as a host can get. receptor. Dscam is reported to be func- immune response. These organisms The memory axis can be divided into tionally important in several invertebrate could simply generate a trained immune four categories that are also drawn as systems including Drosophila and the response in another way. And they do. a continuum in the plot in Figure 1. First, mosquito. In the absence of Dscam, these To describe immunity, we plotted immune effectors can be expressed hosts (or at least their cells) have trouble known immune responses according to constitutively. Second, effectors can be raising an effective immune response their properties on two axes: molecular stereotypically induced upon exposure (Dong et al., 2006; Watson et al., 2005). specificity and memory. A historical to microbes. Third, effectors can persist Ever since Dscam was implicated in method of dividing innate from adaptive following an immune response. This pro- invertebrate immunity, there has been immunity is to discuss the specificity of vides a sort of switch where the chal- some question about how much speci- the immune effectors. One could say lenged animal is different from a naive ficity the insect can squeeze out of these that innate immune effectors are germline animal; for example, amplification of molecules. Do these act like antibodies,

394 Cell Host & Microbe 12, October 18, 2012 ª2012 Elsevier Inc. Cell Host & Microbe Previews

work across the whole memory axis,

Switch and from constitutive expression to altered altered neuro-immune Antibodies second induction. We need to fill in the second signaling Agnathan VLRs induction missing spots for Dscam and all other immune effectors. There is a third axis that we left off

Memory Switch NK cell mediated Figure 1: function. Some of the effectors CMV immunity placed on the plot, like snail FREPS or Drosophila Antimicrobial sea urchin 333/118 molecules, have inter- Peptides esting molecular specificity properties, Inducible Reactive Mosquito Snail Freps RNAi but these have not been shown to func- Oxygen Dscam tionally affect memory. In contrast, there Urchin 185/333 many examples of functionally more effective immune responses in inverte- Constitutive Eye Lysozyme brates where we do not have enough mechanistic information to place them Immune challenges on the memory and specificity axes Single effectors Effector Recombinant/spliced Effector True Cloud effectors Mutation Novelty (Pham et al., 2007; Rodrigues et al., 2010). This figure will become more Molecular Specificity complete as we resolve how each of these systems work. Studies such as the one from Dong Figure 1. A Map of Immune Effectors in Memory by Specificity Space Immune responses are plotted qualitatively against two axes, memory and specificity. The memory et al. (2012) should provoke immunolo- axis imagines induction profiles for two successive immune responses and categorizes the responses gists to reassess their definition of ‘‘adap- with regard to how the host responds the second time it encounters a pathogen. The specificity axis tation’’ and to look for signs of trained reports how many specific molecules can be produced as immune effectors and whether they are immunity (Netea et al., 2011) in systems germline encoded, partially encoded, or generated de novo upon an immune response. A discussion of snail FREPS and sea urchin 185/333 molecules can be found in Buckley et al. (2008) and Zhang et al. that were originally considered innate (2004). See text for more details. and stereotypical.

where the most specific antibody is ex- and resembles antibodies in this respect; REFERENCES pressed in response to stimulation by an however, Dscam is not expressed as antigen? Or is this just potential specificity a single highly specific molecule, but Buckley, K.M., Terwilliger, D.P., and Smith, L.C. (2008). Journal of Immunology 181, 8585–8594. that is never realized because Dscam is rather it is produced as a cloud of effec- just expressed in the same cloud upon tors with different specificities. Dong, Y., Taylor, H.E., and Dimopoulos, G. (2006). every exposure? Dong et al. demonstrate The paper by Dong and colleagues PLoS Biol. 4, e229. that the answer lies in between these two places Dscam near the middle of Fig- Dong, Y., Cirimotich, C.M., Pike, A., Chandra, R., extremes. ure 1. The memory axis was not studied and Dimopoulos, G. (2012). Cell Host Microbe 12, The authors of Dong et al. (2012) find in this paper; the authors report only this issue, 521–530. that different immune elicitors generate what happens during a primary immune Netea, M.G., Quintin, J., and van der Meer, J.W.M. different splicing patterns of Dscam, response. It would be interesting to (2011). Cell Host Microbe 9, 355–361. tracing this through the pattern recogni- know how Dscam affects a second Pham, L.N., Dionne, M.S., Shirasu-Hiza, M.M., and tion pathways to the splicing factors immune response. Does the mosquito Schneider, D.S. (2007). PLoS Pathog. 3, e26. involved. Induction through the insect do better upon a second challenge of immune deficiency (IMD) pathway regu- the same pathogen because there is ex- Rodrigues, J., Brayner, F.A., Alves, L.C., Dixit, R., and Barillas-Mury, C. (2010). Science 329, 1353– lates splicing of Dscam through the isting Dscam protein from the previous 1355. splicing factors Caper and/or IRSF1. The challenge? Is the correct cloud of Dscam authors demonstrated that Dscam alter- expressed more intensely upon re- Sun, J.C., Beilke, J.N., and Lanier, L.L. (2009). Nature 457, 557–561. native splicing is important for protec- exposure? Does the mosquito do worse tion of the host from infection. If the host upon a challenge with a second different Tracey, K.J. (2009). Nat. Rev. Immunol. 9, is unable to splice appropriately for a pathogen because the splicing is now 418–428. given microbe, then the host is more directed toward another pathogen? Watson, F.L., Pu¨ ttmann-Holgado, R., Thomas, F., susceptible. The authors present a model The best-studied effectors in Figure 1 Lamar, D.L., Hughes, M., Kondo, M., Rebel, V.I., whereby the nature of pattern recognition cover large swaths of the specificity by and Schmucker, D. (2005). Science 309, 1874– 1878. signaling leads to the expression of memory space. For example, Drosophila different clouds of alternatively spliced antimicrobial peptides (AMPs) are ex- Zhang, S.M., Adema, C.M., Kepler, T.B., and Dscam molecules. Therefore, Dscam pressed as a cloud at the constitutive, Loker, E.S. (2004). Science 305, 251–254. defies simple descriptions. Dscam is a inducible, and switch levels. Antibodies, Zhang, Y., Lu, H., and Bargmann, C.I. (2005). non-germline-encoded immune effector considered as a class of molecules, Nature 438, 179–184.

Cell Host & Microbe 12, October 18, 2012 ª2012 Elsevier Inc. 395