Immune Systems and Ecosystems

Simon A. Levin

• Introduction
• Do Ecosystems Have Immune Systems? If Not, Why Not?
• Should the Development of Immune Systems for Ecosystems be a Goal?
• Lessons Learned
• Responses to this Article
• Acknowledgments
• Literature Cited

KEY WORDS: complex adaptive systems, ecosystem management, immune systems, normative behavior.

Published: June 18, 2001

The parallels between naturally evolved immune systems and emergent or designed reactive response schemes for systems of importance to humans are intriguing both intellectually and practically. In the computer world, the analogies are especially striking; hence our concern with computer viruses, worms, and other bugs, all of which represent terminology lifted from biology that needs no explanation to even a casual computer user. Computers are like organisms in many functional ways, and exchange useful as well as harmful data by horizontal transfer. Why not develop immune systems for computers, or even for integrated computer networks (Hofmeyr and Forrest 2000)?

Indeed, why stop at computer networks or the Internet? Don't we face similar problems in ecosystems, economic systems, and social systems? Such systems all provide services to humans that allow us to live comfortably. Could imitation of the immune system not help us sustain those systems and those services? In a fascinating paper in this issue, Janssen (2001) explores these tantalizing questions with regard to ecosystems, providing much food for thought.

How well does the parallel hold up? After all, both ecosystems and individuals are complex adaptive systems (CASs), in which macroscopic properties emerge from local interactions and selection among diverse components. In fact, for individual vertebrates, the immune system is a prototypical CAS, relying on evolutionary mechanisms over rapid time scales to adapt to novel threats. As Janssen emphasizes, however, the analogy is imperfect.

Ecosystems are also complex adaptive systems, but with features quite different from those of organisms. First of all, ecosystems are open systems, so much so that they are better thought of as idealizations rather than real entities. An organism has well-defined boundaries and composition. Although it must deal with invaders, there is a sharp and clear distinction between self and nonself, reinforced to a large extent by genetic signatures. In contrast, what we call an ecosystem is an operational convenience, a loosely defined and ephemeral assemblage of co-occurring organisms of very mixed genetic backgrounds. The boundaries are defined by the investigator for convenience of description. In the case of agroecosystems, or of bodies such as lakes with well-defined boundaries, the natural limits may be clear, but in most situations there is considerable arbitrariness in setting boundaries. Ecosystems do not evolve as units, but rather self-assemble (or are assembled, in the case of some managed systems) from components that have evolved relatively independently over much larger spatial scales. For example, plants develop chemical defenses to herbivores and pathogens, and these in turn develop detoxification mechanisms. However, such developments arise largely from diffuse co-evolutionary feedbacks involving generalized responses to suites of enemies. The vertebrate immune system is a model of the fruits of diffuse co-evolution, relying on reactive adaptive responses to deal with unpredictable challenges.

The weakness of the metaphor is the same failing that undercuts all efforts to measure ecosystem "health" or "integrity": the ecosystem is not an organism and has not been shaped by evolution to perform particular functions. Hence, no natural immune responses have evolved in the same way as they have for vertebrates. Indeed, through whatever perceptive filters we choose to impose, ecosystems do have characteristic features, such as those involving patterns of species abundance, food web relations, and element cycling, that exhibit stability and homeostasis over time. However, to the extent that such regularities exist, they have largely emerged from selective forces acting at much lower levels of organization.

Given that we depend upon ecosystems for our life-support systems, we obviously have an interest in maintaining their functioning in accordance with some norms. A principal difficulty, however, is defining what normative behavior is desirable. Organisms have well-defined attributes associated with proper functioning, although even here it is impossible to ignore societal influences in defining what is normative (too thin, too fat, too moody, too flighty); ecosystems do not. Hence, management of an ecosystem in accordance with some defined normative behavior rests on judgments as to what is important in those systems. In this enterprise, reasonable people will differ. Furthermore, the open nature of ecosystems means that they will undergo continual change in composition. How would an ecosystem-level immune system, designed to deal with invaders, decide whether a particular set of invaders should be permitted to join the system? And finally, change and renewal characterize the dynamics of ecosystems, so regulation to constancy can be a prescription for loss of services, and hence of the very features we treasure.

What then should be the lessons learned from contemplating the analogies between the immune system of vertebrates and the intelligent management of ecosystems? The complex adaptive nature of ecosystems means that evolutionary forces are strongest at lower levels of organization; we have learned that the hard way in our continual battles with the evolution of resistance to pesticides and antibiotics, and the unwillingness of microorganisms to take a reasonable approach to making things easy for us. Buzz Holling has in many venues emphasized the importance of surprise and argued forcefully for the need to develop adaptive approaches to management. To a large extent, this does, in fact, mean maintaining the features that allow adaptation, such as heterogeneity and diversity or modularity and redundancy, while tightening feedback loops (for example, through enhancing property rights) to shorten the time scales of response (Levin 1999). In many ways, this imitates features of the immune system, but with a fundamental difference: overall sustainability of the biosphere must emerge from the democracy of distributed responses and competitive renewal, and not from system-level regulation at constant normative levels. Rather than focusing on the development of immune systems to sustain particular functions at predetermined levels, we must emphasize resiliency, and find ways to encourage the adaptability of the mechanisms that sustain critical services.

Responses to this article are invited. If accepted for publication, your response will be hyperlinked to the article. To submit a comment, follow this link. To read comments already accepted, follow this link.

S. A. Levin acknowledges the support of The David and Lucile Packard Foundation through awards to Johns Hopkins University/Princeton University and to the Santa Fe Institute.

Hofmeyr, S. A., and S. Forrest. 2000. Architecture for an artificial immune system. Evolutionary Computational Journal 7: 45-68.

Janssen, M. A. 2001. An immune system perspective on ecosystem management. Conservation Ecology 5(1): 13 [online] URL: http://www.consecol.org/vol5/iss1/art13.

Levin, S. A. 1999. Fragile dominion: complexity and the commons. Perseus Books, Reading, Massachusetts, USA.