The mistaken Differential Adhesion Hypothesis

Albert K Harris
Professor of Biology
UNC-Chapel Hill
Presented in the Embryo Physics Course, January 23, 2013


Malcolm Steinberg invented the “Differential Adhesion Hypothesis” (Steinberg 1963, 1970).  His key idea was that aggregations of cells are thermodynamically equivalent to drops of liquids, and that maximization of adhesions between cells acts like maximization of molecule-to-molecule contact.  He proposed that this is the reason why aggregations of cells behave as if they have contractile surfaces.  Albert Harris (1976) published a critique suggesting an alternative explanation – namely that cell aggregates really do have contractile surfaces, caused by strengthened  acto-myosin contraction of those parts of individual cell surfaces that are not touching other cells.  Steinberg had tested his theory by combining chemically dissociated cells from seven cell types, in pair-wise combinations.  His theory predicted that if cell type A sorted out to a more interior position relative to cell type B, and if cell type B sorted out to an interior position relative to cell type C, then A would always sort out to an interior position relative to C.  He called this “Transitivity”.  The logic was that if A is more adhesive than B, and if B is more adhesive than C, then A would have to be more adhesive than C.  The Harris critique was that transitivity is predicted equally by any hypotheses based on quantitative variables. For example, if A is stronger than B, etc.  In terms of the small field of Mathematics called “relation theory”, the sorting behavior of dissociated cells is a “partial order relation”. Theories based on time sequence would make the same prediction:  If A happens after B, and B happens after C, then A must happen after C.  Therefore Adam Curtis’ “Timing” explanation for sorting (Curtis, 1961) also predicted transitivity.

In collaboration with Herbert Phillips, Steinberg measured the resistance of cell aggregates to flattening by centrifugation and by direct pressure (Phillips et al., 1977; Phillips and Steinberg, 1978).  They proved that the greater the resistance to flattening, the more interior the relative location of sorting out.  Harris argued that the identical prediction is made by explanations of cell rounding by active contraction.  Steinberg was not persuaded, and he and Phillips quantified resistances to flattening as “Reversible Work of Adhesion” and invoked the fearsome and unfathomable mysteries of thermodynamics in opposition to Harris. Unfortunately, that person had taken multiple college courses in thermodynamics (as a way to hang out with his girl-friend) not to mention being a veteran debater in public controversies about when thermodynamics does and doesn’t apply to biological phenomena.

By the early 1980s, many cell-cell adhesion proteins had been discovered, including N-CAM and L-CAM by Edelman et al. (1983) and many different cadherins by Takeichi (1988).  That reduced the need for Steinberg’s explanation by differences in amount of just one kind of adhesion protein.  It also created a paradox: why do relative positions behave as if caused by a differences in an amount of some quantitative variable, given that cell-cell adhesion proteins differ in kind, not just amount?

Harris’ answer was that cadherins, etc. cause like cells to bunch together, and that differences in amount of contractile strength is what produces the inside-versus-outside positions, and likewise that contractile strength is what was measured by the resistance to flattening measurements of Phillips.  Wayne Brodland (Brodland, 2002; used computer simulations to produce a better version of this reasoning.

Steinberg later spent a sabbatical with Takiechi, doing molecular biology (Steinberg and Takeichi, 1994). They used tissue culture  cells from an established mouse line, that happened to lack genes for any cell-cell adhesion proteins.  Into the chromosomes of these non-adhesive cells, they inserted (“transformed”) multiple copies of the a gene for cadherin (some had 2 copies per cell, some had 4 copies, etc.)  They then mixed single-cell suspensions of their transformed lines.  As hoped, those cells with the most copies of the cadherin gene sorted out to the more internal position.  No attempt was made to compare their strengths of contractility.  Prof. Steinberg regarded this as the final, undisputable confirmation of his theory.  C-P Heisenberg (Puech et al., 2005; Krieg et al., 2008) and collaborators used Zebra-fish embryos and Atomic Force Microscopy to compare forces needed  to pull cells apart, versus forces needed to flatten cell aggregates.  P. B. Green’s lucid review of their findings concluded that Brodland and Harris’ interpretation may replace Steinberg’s in textbooks (Green, 2008). This has yet to happen.

Steinberg and Harris remained lifelong friends: agreeing to disagree. They both believed that cell sorting is caused by the same forces that cause gastrulation, neurulation, and other normal rearrangements embryonic movement.  They also agreed that many cancer cells differ from normal equivalents in the same properties that cause sorting out to external positions (Foty and Steinberg, 2004; Danowski and Harris, 1988).  Too bad they couldn’t reach agreement, and that thermodynamics gets a reputation as a mystery-weapon for repulsion of logic.  It’s a fallacy that only conservative (“reversible) forces  (i.e. that don’t continually expend energy) can cause cells to gravitate to consistent end results, by multiple series of intermediate geometries.  Body cells crawl actively, somewhat  like amoebae; they pull themselves toward and past adhesions; they are not  passively pulled around by forces of attraction toward glass, collagen etc.  Passive pulling was advocated  vigorously by Paul Weiss (1961)     and Stephen Carter (1967,1970).  Proponents of “Differential Adhesion” don’t realize that their calculations of “Reversible Work of Adhesion” are not valid except if cells are passively pulled by enlargement of cell-cell adhesions.

This misbegotten enterprise is being carried forward by Professor Ramsey Foty (Foty and Steinberg, 2005).  Both sides ought to give more consideration to the possibility that the other has some truth on their side, lest they give stubbornness a bad name.

From a larger perspective, I believe that Steinberg’s theory, although misguided in details, was ahead of its time, aimed in the right direction, and closer to the truth than most of his critics: 

(One) Steinberg was right to believe that cell sorting is caused by combinations of physical properties of individual cells. 

(Two) Steinberg and Holtfreter were both right that these same properties of cells also cause embryos to gastrulate, to neurulate, and to rearrange themselves to form many organs of the body, possibly all organs.

(Three) Steinberg and Phillips were correct that one particular quantitative variable causes the inside/outside location of cell types, when they sort out, and also exerts the driving force that causes aggregations of cells to round up, and to resist flattening.   

Never mind that acto-myosin driven contractility is this key variable, not amounts of cell-cell adhesiveness.

(Four) Even more profoundly, Steinberg and Phillips were also on the right track to envision that anatomical shapes and arrangements are caused by counterbalances of forces.

Regarding their misunderstanding of thermodynamics, especially their belief that cellular forces correspond to “reversible works of adhesion”: They are right that counterbalance is what matters;  But active forces are just as much able to be counterbalanced, either against one another or against “conservative” (equals “reversible”) forces, like elastic stretching.  If they exist, works of adhesion could be reversible, and measurable. Thermodynamics is like some electric saw: when used correctly; very powerful; but if you don’t read the instructions carefully enough, look out!

Stephen Levin in the US (Levin, 2002;, and Lev Beloussov in Moscow (Beloussov, 2012) have come closest to the truth in regarding mechanical stabilities of organs in terms of what Fuller and Ingber (Ingber, 1997) called tensegrity structures.  To achieve geometric stability, it is sufficient (and, I believe, necessary) to have pairs of opposed forces that are counterbalanced (equal in strength: opposite in direction) only when a certain shape exists.  One or both of the counterbalanced forces need to vary as some mathematical function of the currently existing shape.  All imbalances need to be back toward stable balance. This is an example of what Otter and colleagues (Andersen et al., 2009)  call “Shape Homeostasis”.  Active feedback controls are prone to over-correct, which I suggest is at least part of the Hyper-Restoration phenomenon, discovered by Lev Beloussov, who showed that it causes important embryonic cell rearrangements.

Therefore, I regard Steinberg, Phillips and Foty as having been 90% correct, analogous to following instructions to someone’s house, getting 9 out of 10 turns correct, but unfortunately turning left instead of right at one of the early turns. We need to sort this out.

Andersen, T., Newman, R., and Otter, T. (2009).  Shape homeostasis in virtual embryos.  Artif. Lif 15, 161-183.

Beloussov, L.V. (2012).  Morphogensis as a macroscopic self-organizing process.  BioSystems 109, 262-279.

Brodland, G.W. (2002).  The differential interfacial tension hypothesis (DITH): A comprehensive theory for the self-rearrangement of embryonic cells and tissues.  J. Biomech. Eng. 124, 188-197.

Carter, S.B. (1967). Haptotaxis and the mechanism of cell motility.  Nature 213, 256-260.

Carter, S.B. (1970). Cell movement and cell spreading: a passive or an active process?  Nature 225, 858-859.

Curtis, A.S.G. (1961).  Timing mechanisms in the specific adhesion of cells.  Exp. Cell Res. 8(suppl), 107-122.

Danowski, B.A., and Harris, A.K. (1988).  Changes in fibroblast contractility, morphology, and adhesion in response to a phorbol ester tumor promoter.  Exp. Cell Res. 177, 47-59.

Edelman, G.M., Gallin, W.J., Delouvée, A., Cunningham, B.A., and Thierry, J.-P. (1963).  Early epochal maps of two different cell adhesion molecules.  Proc. Natl. Acad. Sci. USA 80, 4384-4388.

Foty, R.A., and Steinberg, M.S. (2004).  Cadherin-mediated cell-cell adhesion and tissue segregation in relation to malignancy.  Int. J. Dev. Biol. 48, 397-409.

Foty, R.A., and Steinberg, M.S. (2005).  The differential adhesion hypothesis: a direct evaluation.  Dev. Biol. 278, 255-263.

Green, J.B.A. (2008).  Sophistications of cell sorting.  Nature Cell Biol. 10, 375-377.

Harris, A.K. (1976).  Is cell sorting caused by differences in the work of intercellular adhesion? A critique of the Steinberg hypothesis.  J. Theor. Biol. 61, 267-285.

Ingber, D.E. (1997)  Tensegrity: the architectural basis of cellular mecahnotransduction.  Annu. Reb. Physiol. 59, 575-599

Krieg, M., Arboleda-Estudillo, Y., Puech, P.H., Käfer, J., Graner, F., Müller, D.J., and Heisenberg, C.-P. (2008).  Tensile forces govern germ-layer organization in zebrafish.  Nat. Cell Biol. 10, 429-436.

Levin, S.M. (2002).  The tensegrity-truss as a model for spine mechanics: Biotensegrity.  J. Mech. Med. Biol. 2, 3750 .

Phillips, H.M., and Steinberg, M.S. (1978). Embryonic tissues as elasticoviscous liquids.  I. Rapid and slow shape changes in centrifuged cell aggregates.  J. Cell Sci. 30, 1-20.

Phillips, H.M., Steinberg, M.S., and Lipton, B.H. (1977). Embryonic tissues as elasticoviscous liquids.  II. Direct evidence for cell slippage in centrifuged aggregates.  Dev. Biol. 59, 124-134.

Puech, P.-H., Taubenberger, A., Ulrich, F., Krieg, M., Muller, D.J., and Heisenberg, C.-P. (2005).  Measuring cell adhesion forces of primary gastrulating cells from zebrafish using atomic force microscopy.  J. Cell Sci. 118, 4199-4206.

Steinberg, M.S. (1963). Reconstruction of tissues by dissociated cells. Some morphogenetic tissue movements and the sorting out of embryonic cells may have a common explanation.  Science 141, 401-408.

Steinberg, M.S. (1970). Does differential adhesion govern self-assembly processes in histogenesis? Equilibrium configurations and the emergence of a hierarchy among populations of embryonic cells. J. Exp. Zool. 173, 395-433.

Steinberg, M.S., and Takeichi, M. (1994).  Experimental specification of cell sorting, tissue spreading, and specific spatial patterning by quantitative differences in cadherin expression.  Proc. Natl. Acad. Sci. USA 91, 206-209.

Takeichi, M. (1988).  The cadherins: cell-cell adhesion molecules controlling animal morphogenesis.  Development 102, 639-655.

Weiss, P. (1961). Guiding principles in cell locomotion and cell aggregation.  Exp. Cell Res. 8(suppl.), 260-281.






Comments are closed.