13.
D I F F E R E N T I A T I O N
473
and Wessells: see Grobstein, 1962b) supply another case. Embryonic
pancreatic epithelium cultured for 40 h in association with mesenchyme
will progress to zymogen granule formation even though the mesenchyme be removed. If the 40 h epithelium is cut into fragments, however, progression of the differentiation, as judged by the criteria used,
depends on whether the fragments are cultured en masse or in isolation.
In the former case overt zymogen-granule synthesis occurs; in the latter
case it does not. Once again, disruption of essential intimate relations
among differentiating cells of presumably like type seems the most
direct explanation of the altered differentiative behavior.
These facts suggest that differentiation is, in important aspects,
influenced by factors resulting from group relations among cells. It is
important to note that these group or system-dependent effects may be
either inhibitory or promotional with respect to a given differentiation.
In the experiments with the chick blastoderm, cell groups below the
minimal value for neural differentiation do not remain undifferentiated.
They specialize in an alternative direction, in this instance into cornifying skin-like, or mucoid gut-like cells. In the absence of critical information it is equally valid to speak of the larger masses inhibiting cornification
or as promoting neurogenesis. The point is made even more strikingly
in studies of the differentiative behavior of disaggregated chick retinal
cells in vitro. Moscona (1957) has noted, under particular conditions,
that dissociated neural retina, from 6-7-day chick embryos, reaggregates to form rosettes of sensory cells. When crowded together, however,
clusters of rosettes undergo changes which lead to the formation of
lentoids. On the other hand, Peck (1960) reports that dissociated cells
of 6-14-day chick embryo neural retina, cultured as a dispersed cell
population in a synthetic medium, show a 95-100% transformation
into pigment cells in 10-24 h. It is important to recall that these mass
or density-induced alterations in the differentiative behavior of neural
retina in vitro have a parallel in the interconversions of cell types
experimentally induced in the amphibian eye in vivo (Stone, 1959).
They strongly suggest that homotypic interactions among cells, initially
at least of like type, are important operators in influencing cytodifferentiative mechanisms.
Extrapolation of these facts might suggest that differentiation of a
single cell, or of cells fully isolated from one another, is impossible.
Several considerations argue against the necessity of this conclusion.
First, changes of cell properties very similar to multicellular differentiation occur during the life history of unicellular forms (Weisz, 1954;
Tartar, 1956). Though the control systems in these cell-organisms may
be very different from those operative in metazoa and metaphyta, it is
clear that developmental specialization of cell organization can occur
D I F F E R E N T I A T I O N
473
and Wessells: see Grobstein, 1962b) supply another case. Embryonic
pancreatic epithelium cultured for 40 h in association with mesenchyme
will progress to zymogen granule formation even though the mesenchyme be removed. If the 40 h epithelium is cut into fragments, however, progression of the differentiation, as judged by the criteria used,
depends on whether the fragments are cultured en masse or in isolation.
In the former case overt zymogen-granule synthesis occurs; in the latter
case it does not. Once again, disruption of essential intimate relations
among differentiating cells of presumably like type seems the most
direct explanation of the altered differentiative behavior.
These facts suggest that differentiation is, in important aspects,
influenced by factors resulting from group relations among cells. It is
important to note that these group or system-dependent effects may be
either inhibitory or promotional with respect to a given differentiation.
In the experiments with the chick blastoderm, cell groups below the
minimal value for neural differentiation do not remain undifferentiated.
They specialize in an alternative direction, in this instance into cornifying skin-like, or mucoid gut-like cells. In the absence of critical information it is equally valid to speak of the larger masses inhibiting cornification
or as promoting neurogenesis. The point is made even more strikingly
in studies of the differentiative behavior of disaggregated chick retinal
cells in vitro. Moscona (1957) has noted, under particular conditions,
that dissociated neural retina, from 6-7-day chick embryos, reaggregates to form rosettes of sensory cells. When crowded together, however,
clusters of rosettes undergo changes which lead to the formation of
lentoids. On the other hand, Peck (1960) reports that dissociated cells
of 6-14-day chick embryo neural retina, cultured as a dispersed cell
population in a synthetic medium, show a 95-100% transformation
into pigment cells in 10-24 h. It is important to recall that these mass
or density-induced alterations in the differentiative behavior of neural
retina in vitro have a parallel in the interconversions of cell types
experimentally induced in the amphibian eye in vivo (Stone, 1959).
They strongly suggest that homotypic interactions among cells, initially
at least of like type, are important operators in influencing cytodifferentiative mechanisms.
Extrapolation of these facts might suggest that differentiation of a
single cell, or of cells fully isolated from one another, is impossible.
Several considerations argue against the necessity of this conclusion.
First, changes of cell properties very similar to multicellular differentiation occur during the life history of unicellular forms (Weisz, 1954;
Tartar, 1956). Though the control systems in these cell-organisms may
be very different from those operative in metazoa and metaphyta, it is
clear that developmental specialization of cell organization can occur