478
CLIFFORD GROBSTEIN
dependent upon the influence of another is to be expected from the
phenomenon of embryonic induction. It is now well established in
normal development in vivo that interactions between tissue components
of developing systems play a vital role, and that the effects of these
interactions include the origin of new cell-types. It is less well recognized
that interactive processes extend beyond the initiation of differentiation, that they are important as well for the progression and, in many
instances, for the maintenance of differentiated states. The overt
dedifferentiation of cultured cell populations, observed very early in the
effort to obtain "pure" cultures, now seems to be one demonstration of
the need for continuing tissue interaction in support of differentiation.
This was illustrated in classical experiments of Drew (1923) and others,
who showed that "pure", dedifferentiated, epithelially-derived strains
of cells would, on recombination with fibroblasts, show tubuloid
organization and other evidences of more specialized behavior.
Recent years have brought confirmation and expansion of these
observations, based largely on exploitation and gradual modification
of organ-culture techniques and the enzymatic separation of tissues. In
particular, the relations between epithelium and mesenchyme have
been explored in a number of rudiments of mouse and chick, and the
following general statements may now be made. The epithelial
component of such early rudiments as salivary gland (Grobstein,
1953a); kidney (Grobstein, 1955b); pituitary (Sobel, 1958); lung
(Dameron, 1961); thymus (Auerbach, 1960); thyroid (Hilfer, 1962);
uropygial gland (Gomot, 1958); pancreas (Golosow and Grobstein,
1962); skin (Sengel, 1958; McLoughlin, 1961a, b ; Wessells, 1962);
mammary gland (Lasfargues, 1957), etc. does not continue its
characteristic differentiation when isolated in vitro. On recombination
with mesenchyme from its own rudiment type (homogenous) the
epithelium renews its morphogenetic and differentiative behavior,
providing it has not been isolated for more than a few hours. The
epithelium at the time of isolation is relatively undifferentiated overtly,
but cannot be said to be completely undifferentiated since epithelium
from different rudiments shows somewhat different morphology and,
in favorable cases, characteristically different developmental behavior
when recombined with the same mesenchyme. This can be demonstrated in those instances, e.g. thymus (Auerbach, 1960), pancreas
(Golosow and Grobstein, 1962), where the specificity ofthe epithelial
requirement is low and mesenchyme from a number of sources will
suffice. Salivary mesenchyme, for example, will promote morphogenesis of both salivary and pancreatic epithelia but the differentiation
is ofthe expected epithelial type. In other instances, e.g. salivary gland,
kidney, the epithelial requirement is more specific and only the
CLIFFORD GROBSTEIN
dependent upon the influence of another is to be expected from the
phenomenon of embryonic induction. It is now well established in
normal development in vivo that interactions between tissue components
of developing systems play a vital role, and that the effects of these
interactions include the origin of new cell-types. It is less well recognized
that interactive processes extend beyond the initiation of differentiation, that they are important as well for the progression and, in many
instances, for the maintenance of differentiated states. The overt
dedifferentiation of cultured cell populations, observed very early in the
effort to obtain "pure" cultures, now seems to be one demonstration of
the need for continuing tissue interaction in support of differentiation.
This was illustrated in classical experiments of Drew (1923) and others,
who showed that "pure", dedifferentiated, epithelially-derived strains
of cells would, on recombination with fibroblasts, show tubuloid
organization and other evidences of more specialized behavior.
Recent years have brought confirmation and expansion of these
observations, based largely on exploitation and gradual modification
of organ-culture techniques and the enzymatic separation of tissues. In
particular, the relations between epithelium and mesenchyme have
been explored in a number of rudiments of mouse and chick, and the
following general statements may now be made. The epithelial
component of such early rudiments as salivary gland (Grobstein,
1953a); kidney (Grobstein, 1955b); pituitary (Sobel, 1958); lung
(Dameron, 1961); thymus (Auerbach, 1960); thyroid (Hilfer, 1962);
uropygial gland (Gomot, 1958); pancreas (Golosow and Grobstein,
1962); skin (Sengel, 1958; McLoughlin, 1961a, b ; Wessells, 1962);
mammary gland (Lasfargues, 1957), etc. does not continue its
characteristic differentiation when isolated in vitro. On recombination
with mesenchyme from its own rudiment type (homogenous) the
epithelium renews its morphogenetic and differentiative behavior,
providing it has not been isolated for more than a few hours. The
epithelium at the time of isolation is relatively undifferentiated overtly,
but cannot be said to be completely undifferentiated since epithelium
from different rudiments shows somewhat different morphology and,
in favorable cases, characteristically different developmental behavior
when recombined with the same mesenchyme. This can be demonstrated in those instances, e.g. thymus (Auerbach, 1960), pancreas
(Golosow and Grobstein, 1962), where the specificity ofthe epithelial
requirement is low and mesenchyme from a number of sources will
suffice. Salivary mesenchyme, for example, will promote morphogenesis of both salivary and pancreatic epithelia but the differentiation
is ofthe expected epithelial type. In other instances, e.g. salivary gland,
kidney, the epithelial requirement is more specific and only the