Contribution
A framework that defines cancer by what it breaks rather than by its human clinical form, so that it can be looked for in any multicellular lineage, and a survey of the literature using it. It also proposes that dysregulated differentiation may be a missing hallmark of cancer.
Key insight
Being multicellular is not only cooperation among cells but continuous suppression of the cells that would gain by defecting. Cancer is what that suppression failing looks like, and because complex multicellularity evolved several times, so did the suppression.
Assumptions
- Cell-level and organism-level fitness can conflict, and multicellularity requires the first to be suppressed in favour of the second.
- “Cheating” is functional: “We do not imply that cheating is a pre-adapted or evolved ‘strategy’” (Introduction).
- A broad notion of cancer, including neoplastic growth with abnormal proliferation or differentiation, is the right unit for comparison (Survey).
Key results
- Five foundations and the cheating phenotypes that break them (Figure 1 caption).
- Survey: bilaterian animals all or nearly all susceptible; cnidarians show it, so it was likely a problem for early metazoans; galls and calluses in plants and in red and brown algae; masses of undifferentiated, sometimes invasive hyphae in fungi; cheating in Volvox mutants; cooperation and cheating in aggregative slime moulds and bacterial biofilms.
- Generalising (Discussion): division-of-labour cheating “was also present in all forms of multicellularity”; resource and environment cheating documented only in animals and plants; metastasis, rare plant cases aside, only in animals.
- Two kinds of cheating: demographic (proliferation, cell death) and economic (resource, labour, environment). Both occurred in every report reviewed, and “cheating of both types may be necessary”.
- Suppression: cell-autonomous, architectural and systemic mechanisms; clonal development reduces conflict but is “not strictly necessary”; proliferative units may break a large suppression problem into “more manageable and modular sub-problems”.
- Peto’s paradox: large organisms likely need more cheater detection and suppression across all five foundations.
- Within tumours: cooperation may re-emerge, with neoplasms recapitulating the foundations for the benefit of the cancer cells.
Claims
| id | claim | strength | support |
|---|---|---|---|
| C1 | Multicellularity requires suppressing cell-level fitness, and cancer is a breakdown of that | moderate | framing drawn from cooperation theory and cited work |
| C2 | Cancer-like phenomena occur across the lineages of complex multicellularity | moderate | literature survey; absences flagged as possibly lack of study |
| C3 | Division-of-labour cheating occurs in every form of multicellularity | moderate | survey, with the definitional breadth that makes it possible |
| C4 | Animals are more susceptible than other branches | weak | survey, with a sampling-bias caveat stated by the authors |
| C5 | Complex multicellularity is a cheater detection and suppression system | moderate | an interpretation of the survey and of known suppression mechanisms |
| C6 | Proliferative units are an intermediate level of selection | weak | proposed, with “may” throughout |
Concepts
- cheating — “Breaking of shared rules, including genetically encoded phenotypes or behaviours, that leads to a fitness advantage for the cheater” (Box 1).
- cooperation — “Transmission of benefits and/or coordination of actions that augments fitness of the larger ensemble” (Box 1).
- five foundations of multicellularity — the five forms of cellular cooperation above.
- demographic and economic cheating — the authors’ two categories.
- complex multicellularity — after Knoll: several cell types, intercellular communication, three-dimensional organization, resource transport (Endnote).
Connections
It draws on cooperation theory and the major-transitions literature. Okasha’s overview (LIT-749, read in NOTE-586) states the general principle this paper applies, that a collective is stable only if mechanisms such as policing align its units’ interests. It also gives a caution this paper partly answers: terms like “cheating” have been misapplied where no collective action problem was shown. Aktipis et al. define cheating functionally, not as a strategy, which meets the caution in part, but the paper does not show that each reported growth is a conflict of fitness interests; for slime moulds it concedes a cheater “might be considered a parasite rather than a neoplasm”. The levels of selection literature (LIT-160) is where its proliferative-unit proposal would be tested.
Bearing on the record
Source of THEORY-135. For the owner’s essay:
- §6.3 (capture from below). This is the biological model: units gaining at the collective’s expense, and a collective that survives only by detecting and suppressing them. The paper adds that the environment matters as much as the units: “Cancer is not just a matter of cells that are inherently cheaters, but also a matter of being in an environment that has poor cheater detection and suppression”.
- A positive feedback. Cheating degrades the local environment, which selects for cells that move, which may favour more cheating. The record notes the parallel with dysregulation from within (§6.3), without claiming it.
- Modularity. Proliferative units that contain a defector’s spread are a structural defence, a candidate parallel for organizational subdivision. The paper does not consider organizations.
Limitations
A review: no model and no new data. The survey depends on what has been studied, and the authors flag sampling bias and lack of study more than once. The broad definition of cancer makes the cross-lineage comparison possible and also makes cheating easier to find in every branch. Multilevel selection is invoked by citation, not modelled.
Open questions
Which forms of cheating are necessary and sufficient for cancer-like growth, which the authors pose; whether proliferative units are a level of selection; whether bacterial proliferation inhibition serves the aggregate or the individual cell.