Left-right asymmetry in embryonic development: a comprehensive reviewIntroductionPre-molecular dataLeft-right asymmetry meets molecular biologyInvertebratesFishAmphibiaChickMammalsTwinning and asymmetryLaterality and brain asymmetryConservation of mechanismsOpen questionsConclusionAcknowledgementsSupplementary materialReferencesReviewLeft–right asymmetry in embryonic development:a comprehensive reviewMichael Levin*Cytokine Biology Department, The Forsyth Institute, and Department of Oral and Dev. Biol., Harvard School of Dental Medicine,140 The Fenway, Boston, MA 02115, USAReceived 18 July 2004; received in revised form 22 August 2004; accepted 23 August 2004Available online 11 September 2004AbstractEmbryonic morphogenesis occurs along three orthogonal axes. While the patterning of the anterior–posterior and dorsal–ventral axes hasbeen increasingly well characterized, the left–right (LR) axis has only recently begun to be understood at the molecular level. Themechanisms which ensure invariant LR asymmetry of the heart, viscera, and brain represent a thread connecting biomolecular chirality tohuman cognition, along the way involving fundamental aspects of cell biology, biophysics, and evolutionary biology. An understanding ofLR asymmetry is important not only for basic science, but also for the biomedicine of a wide range of birth defects and human geneticsyndromes. This review summarizes the current knowledge regarding LR patterning in a number of vertebrate and invertebrate species,discusses several poorly understood but important phenomena, and highlights some important open questions about the evolutionary originand conservation of mechanisms underlying embryonic asymmetry.q 2004 Elsevier Ireland Ltd. All rights reserved.Keywords: Embryogenesis; Left–right asymmetry; Chirality1. IntroductionThe geometrical invariance known as symmetry is aprominent aspect of developmental morphology duringembryogenesis. Animal body-plans occur in a wide varietyof symmetries: spherical (e.g. volvox), radial (e.g. seaanemone), chiral (e.g. snails, ciliates), bilateral (e.g.planaria) and pseudo-bilateral (e.g. man). Vertebrates havea generally bilaterally symmetrical body-plan, but thissymmetry is broken by the consistently asymmetricplacement of various internal organs such as the heart,liver, spleen, and gut, or the asymmetric development ofpaired organs (such as brain hemispheres and lungs).Symmetries are repeatedly broken during development.For example, the radial symmetry of the early chickblastoderm is broken into a bilateral symmetry by theappearance of Ko¨hler’s sickle and then the primitive streak.This is further broken into a definitive pseudo-symmetry bythe right-sided looping of the heart tube. A fascinating atlasof morphological asymmetries throughout the animal king-dom is given in Neville (1976).Developmental noise often results in pseudo-randomcharacteristics and minor stochastic deviations known asfluctuating asymmetry; however, the most interestingphenomenon is invariant (i.e. consistently biased among allnormal individuals of a given type) differences between theleft and right sides. For reasons of space as well as becausethese are likely to be secondary to embryonic asymmetries,this review largely neglects behavioral/sensory asymmetries(such as lobster claw morphology which is determined byneurological activity). A huge literature on brain lateraliza-tion phenomena in human beings exists as well (Harnad,1977), but many of these asymmetries are secondary andarise as a result of cultural environmental biasing factors.The establishment of left–right (LR) asymmetry raises anumber of fascinating biological questions. Why doesasymmetry exist at all? What are the implications ofasymmetry for the normal structure and physiology of theheart, gut, and brain? Why are all normal individuals not only0925-4773/$ - see front matter q 2004 Elsevier Ireland Ltd. All rights reserved.doi:10.1016/j.mod.2004.08.006Mechanisms of Development 122 (2005) 3–25www.elsevier.com/locate/modo* Tel.: C1 617 892 8403; fax: C1 617 892 8423.E-mail address: [email protected], but asymmetric to the same direction (i.e. why aconsistent bias and not a 50/50% racemic population, giventhat individuals with full inversion are not phenotypicallyimpaired)? When, during evolution, did handed asymmetryappear, and were there true bilaterally symmetrical orga-nisms prior to the invention of oriented asymmetry (Cooke,2004)? Is it connected to chirality in lower forms (such assnail shell coiling and chirality in some plants)? At whatdevelopmental stages is asymmetry initiated in vertebrateembryos? How conserved are the molecular mechanismsestablishing correct asymmetry in animals with drasticallydifferent modes of gastrulation? And, how can the LR axis beconsistently oriented with respect to the anterior–posterior(AP) and dorso-ventral (DV) axes in the absence of anymacroscopic feature of chemistry or physics which dis-tinguishes left from right? Answers to these questions requirea detailed understanding, at the molecular, genetic, andbiochemical levels, of the formation of biased asymmetry inembryos.The LR axis itself follows automatically from thedefinition of the AP and DV axes, as it is perpendicular toboth; however, consistently imposed asymmetry across it isfundamentally different from patterning along the other twoaxes. Firstly, while the AP and DV axes can be set byexogenous cues such as gravity, or sperm entry point, there isno independent way to pick out the left (or right) direction,since no known macroscopic aspect of nature differentiatesleft from right. One possible way to use a fundamental forceto orient the LR axis relative to the other two axes wassuggested by Huxley and deBeer (1963). They proposed thatLR asymmetry was oriented during embryonic developmentby an electric current running down the length ofthe notochord, which would generate a magnetic field vectorpointing R or L, if measured at the dorsal or ventral sides.Although a correlation between the Earth’s geomagneticfield reversals and shell chirality has been observed (Harrisonand Funnel, 1964), the nature of a causal relationship (if any)is unknown, and there is no evidence to date of a magneticfield being utilized during LR patterning in any species.While in most species all normal individuals areasymmetrical in the same direction, animals with completemirror reversal of internal organs can arise (situs