/. Embryo/, exp. Morph. Vol. 35, 3, pp. 445-461, 1976
Printed in Great Britain
445
Clonal development of the retinal epithelium in
mouse chimaeras and X-inactivation mosaics
By J. D. WEST1
From the Department of Genetics, University of Edinburgh, Scotland.
SUMMARY
A comparative study of coherent clones in the retinal epithelium is presented for mouse
aggregation chimaeras and X-inactivation mosaics. There is a basic similarity in the number
of coherent clones and the pattern of clonal development between mosaics and chimaeras.
While this similarity is compatible with the difference in mean patch size reported by other
authors this is at variance with some interpretations of previous work and suggests that no
conclusive evidence on the timing of X-inactivation is provided by comparisons of patch
sizes between chimaeras and mosaics.
The present clonal analysis also suggests that at \2\ days post coitum the cells in the retinal
epithelium are distributed almost randomly while, in the adult, the cells are grouped into
small coherent clones, which comprise an average of five or six nuclei. However, these data
could also be explained by larger, irregular coherent clones.
INTRODUCTION
The pigmented epithelium of the retina was one of the first tissues used to
detect experimental chimaerism in the mouse (Tarkowski, 1964). This tissue
forms a monolayer of cells between the neural retina and the choroid of the eye.
In normal mice each cell of the retinal epithelium produces melanin granules
which remain localized within the cell, whereas certain mutants such as albino
(cjc) or pink-eye (p/p) fail to produce normal pigmentation. The combination
of pigmented and unpigmentated cells in this tissue provides a two-dimensional
system with a convenient cell-autonomous marker for the analysis of clonal
growth by routine histological methods.
Several authors, including Tarkowski (1964), Mystkowska & Tarkowski
(1968), Mintz & Sanyal (1970) and Mintz (1971) have used this system to detect
chimaerism. Deol & Whitten (1972a) have compared patches in chimaeras and
mosaics but few other detailed studies have been reported concerning the pattern
of pigmentation in mouse chimaeras or mosaics.
The theoretical relationship between clone and patch sizes (Roach, 1968;
West, 1975 a), is used here in a comparative investigation of clonal development
in the retinal epithelium of mouse aggregation chimaeras and X-inactivation
1
Present address: Department of Molecular Biology, Roswell Park Memorial Institute,
Buffalo, New York, 14263, U.S.A.
446
J. D. WEST
mosaics. The patch size varies with the proportions of the two cell populations
in the tissue and it is essential to allow for this source of variation between
groups of mice. The analysis used here compares the mean size of the coherent
clones in different groups of chimaeras and mosaics, and compensates for any
differences, in proportions of the cell populations, between the groups.
Using this analysis it has been possible to demonstrate the basic similarity
in the pattern of coherent clones between mouse chimaeras and X-inactivation
mosaics.
MATERIALS AND METHODS
(a) Mice
Eight-cell morulae from pigmented and unpigmented stocks were aggregated
to produce chimaeric mice (McLaren & Bowman, 1969). The pigmented stocks
of mice used included (C57BL/McL£ x C3H/BiMcL^)F!, and pigmented individuals from a closed, random-bred stock of Q-strain mice. Mice with unpigmented eyes were either albino (cjc) members of the Q strain or from a multiple
recessive strain produced by Michie (1955), homozygous for non-agouti (a),
brown (b), dilute (d), pink-eye (p), chinchilla (cch), waved-2 (wa-2), short-ear (se),
vestigial tail (vt), supernatant-NADP isocitrate dehydrogenase type a {Id-la) and
glucose phosphate isomerase type a {Gpi-Id). This multiple recessive stock is
designated 'Recessive' throughout. Chimaerism is indicated by the joining of
the symbols for the two component genotypes or stocks with a double-headed
arrow.
The X-inactivation mosaics used were flecked mice, heterozygous for Cattanach's translocation (Cattanach, 1961). Cattanach's translocation, T{7;X)Ct,
involves the insertion of a large part of chromosome 7 (linkage group 1), carrying the wild-type alleles for albino (c), pink-eye (p), ruby-eye-2 (ru-2) and
shaker-1 (sh-1), into the X-chromosome. The stock of flecked mice used in this
study was derived from six mice obtained in November 1972 from Dr Bruce
Cattanach at Harwell. These mice are of the unbalanced, duplication (type II)
form and have a completely normal set of autosomes with no known deletions
in chromosome 7, and are designated Dp(7; X)Ct. The original six females came
from Dr. Cattanach's 'High line' and had about 75% JU-strain genetic background. The stock was maintained by crossing flecked females to albino males;
to retain fertility JU/Fa and albino Q-strain males were used to sire alternate
generations. The mosaics {Dp (7; X)Ci) used in this study, therefore, have a
complex genetic background which could affect the relative proportions of
pigmented and unpigmented cells in the retinal epithelium.
{b) Histological methods
Standard histological methods were used throughout this study. Adult eyes
were removed for histological examination, but the eyes of embryos and newborn mice were sectioned in situ. Each adult eye was fixed in Bouin's fluid over-
Chimaeric and Mosaic Retinal Epithelia
447
night, an incision made in the cornea and the lens removed with watchmaker's
forceps, under a dissecting microscope. Embryos and newborn mice were decapitated, the heads fixed overnight in a fixative based on Sanfelice's fluid
(Sanfelice, 1918) and rinsed in running tap water for at least 8h. Specimens from
newborn mice were decalcified in 1 % nitric acid for three or four days to soften
the skull before sectioning. All specimens were dehydrated in graded alcohols,
cleared in toluene, embedded in wax and sectioned at 6 jum using either a Leitz
or Cambridge rotary microtome. The sections were stained Vith Ehrlich's
haematoxylin and eosin and mounted with D.P.X.
If the eye of an adult mouse is compared to a globe with the cornea in the
north pole position, the two planes of section used are equivalent to 'latitudinal'
sections, parallel to the equator, and 'longitudinal' sections, perpendicular to
the equator. Both of the eyes from each mouse were sectioned in the same plane.
Eyes from embryos and newborn mice were all sectioned in a plane para-sagittal,
with respect to the body.
(c) Clonal analysis
The retinal epithelium of each adult eye was examined in three sections: the
mid section, and one section about 500 /im each side of the middle. The two
longitudinal sections either side of the mid-line were arbitrarily designated
mid-500 and mid+ 500, whereas the latitudinal section nearest the cornea was
designated mid + 500 and the section nearest the posterior pole (or 'south pole')
was termed mid - 500. The retinal epithelium of a fixed adult eye normally has a
diameter of about 2-5-3 mm.
The retinal epithelium of chimaeras and mosaics is patchy and appears in
sections as a one-dimensional string of pigmented and unpigmented cells
(Fig. 1). The length of each patch was measured using a microscope fitted with a
calibrated eyepiece-micrometer. From these measurements the relative proportion of pigmented and unpigmented cells in the section was calculated, together
with the mean patch length for each cell type. The mean one-dimensional patch
length for each cell population varies with the relative proportion (p) of that
population in the section, while the number of coherent clones per patch
expected for a random string of coherent clones, can be estimated as 1/(1 — p)
(see Roach, 1968; West, 1975 a).
Our 'coherent clone' is that defined by Nesbitt (1974) as 'a group of clonally
related cells which have remained contiguous throughout the history of the
embryo'. If cell mixing occurs during development these coherent clones will be
smaller than the descendant clones observed by Mintz (1971), which more nearly
represent the clonal pattern in the primordium of the retinal epithelium. Our
'patch' is defined by Nesbitt (1974) as 'a group of cells of like genotype which
are contiguous at the moment of consideration', and may comprise several
coherent clones.
The average coherent clone length for each cell type was calculated as the
IDI
Fig. 1. Histological sections showing the retinal pigment epithelium (PE) (Horizontal
bar represents 25 /on in each case). (A) Adult pigmented-Q <-> unpigmented-Q
chimaera: patches of pigmented and unpigmented cells. (B-D) Retinal epithelium
from mosaic embryo (D), a fully pigmented littermate (B) and an unpigmented
control (C), 12£ days post coitum.
0-75
0-25
G
H
40
1-3
20
20
20
o
OO
OOO
O
OOO
x
XX
XXX
XXX
X
o
o
X
OO
OOO
O
OOO
o
o
o
X
x
o
X
X
o
x
XX
XXX
XXX
X
o
X
x
X
X
o
o
XX XX XX OO
XXX XXX XXX OOO
XXX XXX XXX O
X X X
OOO
x
o
o
x
X
X
X
X
Linear array
o
XX
XXX
XXX
X
o
X
X
o
X
X
o
o
o
o
o
OO OO OO
OOO OOO OOO
O O O
OOO OOO OOO
o
o
X
20
40
60
60
20
1-5
1-5
10
20
30
Mean estimated coherent Actual
clone len th
Observed
8
length of
mean patch [Observed patch length] coherent
length
[Expected patch length] clones
10
20
4-5
11
1-5
1-2
Schematic representation of hypothetical linear arrays of X and O cells, showing independence ofp and the estimated coherent clone length. In the retinas a mean
coherent clone length was derived from estimates of both cell populations whereas, for ease of presentation, in Table 1 the coherent clone length is estimated only
from the population of X cells. The observed patch length is here expressed in terms of numbers of cells per patch whereas in the biological analysis the patch lengths
are first measured in micrometers as described in the Methods section. The expected patch length for a random array of coherent clones, given by 1/(1-/?), predicts
the number of coherent clones per patch when the proportion of each cell population is known. Arrays A-C show increase in patch length with p, and arrays
D - F show increase in patch length with coherent clone length. Arrays G to H show inaccurate estimation of the coherent clone length when the coherent clones
of X and O cells differ in length. This indicates that the estimated coherent clone lengths for the two populations in an array (or tissue) are not independent
measurements.
0-5
0-5
0-5
D
E
F
A
B
C
Expected
mean patch
length for
random array
of coherent
Proportion
clones
of X 0 0
1/(1-/0 <
0-5
20
X
0-75
40
X
1-3
0-25
o
Table 1
S"
S*
a
450
J. D. WEST
observed mean patch length divided by the expected number of coherent clones
per patch in a linear array which is given by the formula 1/(1 -p) (see Table 1).
The average coherent clone area was taken to be the square of the coherent
clone length, and the number of cells per coherent clone was estimated from the
calculated cell area. The cell area was calculated from measurements of mean
length of unpigmented tissue per nucleus, and nuclear diameter, taking into
account the section thickness (6/tm) and using a slight modification of the
formula derived by Abercrombie (1946), described in the Appendix. This area
was used for both cell types.
The mean coherent clone size, estimated in this way, represents the mean size
of groups of cells which would produce the observed mean patch size by a
random distribution. Our coherent clone is therefore a statistical estimate which
in an ideal tissue represents a group of cells descended from a common ancestral
cell which have remained adjacent throughout development. The accuracy of this
statistical estimate will depend on the variation of coherent clone size and shape
in the retinal epithelium. If the coherent clones are large and very irregular the
analysis is likely to underestimate their mean size, but this limitation can be
ignored for comparisons between different groups of mice.
RESULTS
(a) Preliminary analysis
Figure 1 shows patches of pigmented and unpigmented retinal-epithelium in
a pigmented-Q <-> unpigmented-Q chimaera. The aabbcohcchddpp genotype of the
Recessive stock also results in an unpigmented retinal epithelium and provides
good contrast to the pigmented strains.
Three sections were examined in each plane (latitudinal and longitudinal) for
each adult eye, as explained in the Materials and Methods section. Of these, the
latitudinal section nearest the cornea was commonly entirely iris, so only data
from the remaining two latitudinal sections are considered.
Data were collected from mosaics and several groups of adult chimaeras.
Statistical analyses of the three longitudinal sections by analysis of variance,
and the two latitudinal sections using Student's f-test, showed no significant
difference in the mean coherent clone lengths, between sections, for any of the
four groups of mice. The mid sections are chosen as representative samples of
the eye for subsequent clonal analysis.
The number of cells per coherent clone in two dimensions (coherent clone
size) is estimated as
/mean coherent clone length\ 2
\
mean cell length
/ '
The estimation of the mean coherent length for each cell population is explained
in the Materials and Methods section, and these estimates are averaged to
Chimaeric and Mosaic Retinal Epithelia
451
Table 2. Number of cells per two-dimensional coherent clone in the retinal
epithelium {mean ± standard error of mean) for mosaic and pigmented-Q <-> unpigmented-Q chimaeric eyes at various stages of development
Age
(days post coitum)
Pigmented-Q *-> unpigmented-Q
Chimaeras
Mosaics
12*
13*
14*
15|
18*
Birth
1-56 ±0-21 (n = 10)
2-24 + 018 (#i = 10)
2-18 ±0-23 (#i = 10)
1-33 ±0-18 (/i = 10)
—
1-83 ±0-14 (/i = 2)
4-72 ± 0-35 (n = 2)
—
20i
2-79 ± 010 (n = 10)
4-85 ± 0-27 (n = 20)
4-36 ± 0-56 (n = 8)
605 + 0-44 {n = 1 1 )
Adult
1-20 ± 0 - 0 9 ( I I =
1-82 ± 0 - 2 6 ( I I =
12)
10)
provide one estimate of coherent clone length per eye. The mean coherent clone
lengths for the two cell populations are not independent observations, and the
hypothetical arrays G and H in Table 1 show that even if the two cell populations
had widely differing mean coherent clone lengths, this would not be detected.
The estimation of mean cell length is discussed in the Appendix.
Comparison between longitudinal and latitudinal mid-sections using Student's
Mest, shows no significant difference in the mean number of cells per coherent
clone. This suggests that, on average, the coherent clones are symmetrical, so the
data from the two planes are pooled. Data from left and right eyes are also
pooled. Although there is a significant negative correlation for clone size between left and right eyes for the small group of adult pigmented-Q <-> unpigmented-Q chimaeras (r = -0-90; P < 0-01), the biological basis for this negative correlation in one group is unclear, and it is probably an artifact resulting
from the small sample size.
Statistical analysis by Student's /-test shows a significant difference in the
proportion of pigmentation between longitudinal and latitudinal sections for
mosaics (t = 5-25; P < 0-01) but not in any of the three chimaeric groups. This
difference between the two planes of section in mosaics is clearly shown in Fig. 2
and could be due to sampling error or reflect a non-random distribution of pigmented and unpigmented coherent clones in mosaic eyes. No significant difference in proportion of pigmentation, between the two planes, was found in a
further sample of eight mosaic eyes, each of which was sectioned in both planes.
(These eyes were sectioned to the equator and turned perpendicularly and sectioned again.) This suggests that the difference observed in the first group of
mosaic eyes is not biologically significant and can be ignored.
u
C
8O
1
ao '
2 6-
•
o
•
c<u
_
& 5-
un
•
34-
• •
•
•
^3-
§2-
•
1-
00-2
0-4
0-6
0-8
1 0
0-2
Fig. 2. Estimated two-dimensional coherent clone sizes in the retinal epithelium
from mid-sections of mosaic and chimaeric eyes, showing independence of coherent
clone size and proportion of pigmented cells (/?). Both left and right eyes are
shown in each graph. (A) Eyes from mosaic embryos, 12^ days post coitum. (B)
Eyes from pigmented-Q +-> unpigmented-Q chimaeras, \2\ days post coitum. (C)
Eyes from adult mosaics. (D) Eyes from adult pigmented-Q <-> unpigmented-Q
chimaeras.
In (C) and (D) # represents longitudinal mid-sections and • represents latitudinal mid-sections. The proportion of pigmented cells (p) is estimated from the
sum of the patch lengths for each population in the section. The coherent clone size
for each eye is an average of the coherent clone sizes estimated from both pigmented and unpigmented populations.
(b) Estimation of coherent clone size in the developing retinal epithelium
Figure 1 shows that patches in the retinal epithelium of mosaic eyes can be
detected as early as \2\ days/7051/ coitum, and mosaics can be clearly distinguished
from fully pigmented or unpigmented individuals of the same age. The mean
coherent clone size was estimated in mosaics and pigmented-Q «-• unpigmented-Q
chimaeras at various stages of development, and the results are summarized in
Table 2 and Fig. 2. The results suggest that at 12^ days the mean coherent clone
size in both the mosaic and chimaeric samples is close to one cell, so the cells of
the two populations are distributed almost randomly in the retinal epithelium.
At this stage the mean coherent clone size in the chimaeric group does not
significantly differ from that in the mosaics, (t = 0-70; P > 0-05). The marked
predominance of the pigmented population in the 12^ day chimaeras is most
probably an artifact due to the small number of animals studied, and is considered
more fully in the Discussion section below. The coherent clone size increases
during development to an average of just under three cells in mosaics one day
Chimaeric and Mosaic Retinal Epithelia
453
Table 3. Approximate dimensions of cells and coherent clones of the retinal
epithelium in adult mice and comparison of growth between mosaic and chimaeric
retinal epithelia
Growth: 12^ days to aduli
Adult retinal epithelium
Mosaics
Cells per coherent
clone
Coherent clones per
retinal epithelium*
Cells per retinal
epithelium*
Cell area
Area of retinal
epithelium*
A
Mosaics
Pig. Q <-*
unpig. Q
chimaeras
A
c
Pig. Q <-> unpig. Q
chimaeras
4-85
605
40 x
4-5 x
8,700
9,100
3-1 x
3-2 x
42,000
207 /tm2
55,000
208 /*m2
12-4 x
2-6 x
14-3 x
2-3 x
8-7 x 106 /*rn2
ll-4xlO 6 /tm 2
33-5 x
32-2 x
* Based on very crude estimates of the area of the retinal epithelium.
after birth, and nearly five cells in the mature adult. Chimaeric coherent clones
are significantly larger both at 20^ days post-coitum {t = 3-12, P < 0-01) and in
the adult {t = 2-92; P < 0-01), although in each case this difference is only
about one cell. It is unlikely that any difference in the age structure between the
two groups of adults influences the coherent clone size, as there is no correlation
between coherent clone size and the age of the adult for either mosaics or
pigmented-Q <-> unpigmented-Q chimaeras.
Table 3 gives a crude comparison of the various parameters in the adult
pigmented epithelium between mosaics and the chimaeras, and shows approximate increases in size between 12^ days and maturity. It seems likely that the
area of the retinal epithelium increases by both cell expansion and cell multiplication. The increase in coherent clone size does not keep pace with the increase
in cell number and there is an increase in the estimated number of coherent
clones indicating cell mixing during development.
(c) Comparison of adult coherent clone size between different groups
The mean coherent clone size estimated from adult chimaeras of various
strain combinations are shown in Table 4. Statistical analysis of the four groups
of animals, using Student's /-test, showed no significant difference in coherent
clone size between left and right eyes within each group, but a significantly
larger coherent clone size for (C57BL x C3H)F! <-> Recessive chimaeras than for
mosaics and possible other groups (Table 5). The mean coherent clone sizes
estimated for the left and right eyes of this group are 8-37 and 12-65 respectively.
The mean for the right eyes includes an estimate of 42-79 cells per coherent clone
29
EMB 35
454
J. D. W E S T
Table 4. Mean number of cells per two-dimensional coherent clone (mean ± s.E.)
for adult mosaic and chimaeric retinal epithelia
Mosaics
Cells per coherent clone = 4-85±027 (« = 20 eyes)
(Range = 2-25-7-67)
Chimaeras
Unpigmented component
Pigmented component
Pigmented-Q
Unpigmented-Q
Recessive
605±0-44 (n = 11)
(Range 2-54-7-88)
5-70±0-76 (n = 16)
(Range 1 -84-11 -35)
10-52 ± 1 -85 (n = 20)
(Range 3-46^2-79)
882 ± 0-77 (« = 19)
(Range 3-46-14-77)
(C57BL x C3H)Fi
(C57BL x
excluding X18
* Mean coherent clone size for right eye of XI8 = 42-79 cells; proportion of pigmentation
(p) = 0 0 1 . (The left eye of X18 was not examined.)
Table 5. Comparison of number of cells per two-dimensional coherent clone of the
retinal epithelium, between unrelated groups of chimaeras and mosaics, showing
t-values
Mosaic
Pig.-Q < unpig.-Q
Pig.-Q < Recessive
Pig.Q <-> unpig.-Q
Pig.Q <-> Recessive
2-92**
1-44
0-35
F1 <-> Recessive
5-22***
2-57*
2-38*
* P < 002; ** P < 001 ; * * * / > < 0001; others P > 005. F x = (C57BLx C3H)Fi
Fig. 3. (A) Tangential section of chimaeric eye showing two cell populations and
both binucleate and uninucleate cells in the retinal epithelium. (The horizontal
bar represents 50 /*m). (B-F) Low power photomicrographs of unstained mosaic
and chimaeric eyes showing variegation in the pigmentation of both the dense
choroid layer and the underlying retinal epithelium. (Actual diameter of eyes is
about 3 mm).
No obvious stripes are seen in the retinal epithelium of the chimaeric eye (B) or
mosaic eye (C). Striping is clear in mosaics (D) and (E). Comparison of equatorial
(E) and polar (F) views of the same eye suggests that striping may be restricted to
the equatorial region. The four radiating stripes of dense pigmentation seen in (F)
are due to pigmentation in the overlying choroid. (t marks a region of the retinal epithelium which was torn during the removal of the muscle from the eye.)
Chimaeric and Mosaic Retinal Epithelia
455
B
29-2
456
J. D. WEST
for chimaera XI8, which is almost three times the next largest for the group. The
proportion of pigmentation in the retinal epithelium estimated from the midsection of the right eye of XI8, is only 0-01. The mean coherent clone length
estimate is, therefore, based on a very small number of pigmented patches and
so is more prone to inaccuracies from sampling error. A two-dimensional
reconstruction of part of the retinal epithelium, from serial sections either side
of the mid-section, revealed two pigmented patches. One patch was about the
size of eleven cells and was assumed to be a single coherent clone, whereas the
other was equivalent to about seventy cells and may have represented a patch of
six or seven coherent clones. A coherent clone size of eleven cells agrees more
closely with the other estimates from right eyes in this group. It is assumed
that the original high estimate is due to the small number of patches sampled
in one dimension, and data for this eye is omitted from the statistical analyses
shown in Table 5.
DISCUSSION
The comparison of clone size estimated in two planes suggests that the clones
of the retinal epithelium are, on average, symmetrical. Figure 3 (B-F) shows
eyes from one chimaera and three mosaics. Longitudinal stripes, as described by
Mintz (1971), can clearly be seen in some of the eyes, although, when present,
these are normally restricted to the equatorial region. Figure 3 (E and F) shows
two views of the left eye from one mosaic, which illustrate prominent striping
near the equator but no evidence of any stripes over most of the eye. (The four
prominent stripes radiating from the posterior pole in Fig. 3F, are regions of
pigmentation in the overlying choroid.) The failure of the clonal analysis to
detect any pattern of stripes from sections of the retinal epithelium probably
reflects the apparent restriction of any stripes to a relatively small part of the
retina.
The investigation of the mean coherent clone size during development suggests
that at 12^- days post coitum the cells of the two populations are distributed almost
randomly in both mosaics and pigmented-Q <-> unpigmented-Q chimaeras. This
indicates that cell movement mixes the two cell populations sufficiently to
prevent significant coherent clonal growth at this stage, and further suggests
that Nesbitt's assumption of 'limited coherent clonal growth in the developing
mouse embryo' (Nesbitt, 1974) may not be true for the retinal epithelium, at
this stage of development. The analysis also predicts that limited clonal growth
occurs after \2\ days. The estimate of mean coherent clone size is really an
estimate of the mean number of nuclei per coherent clone. Ts'o & Friedman
(1967) observed a high frequency of binucleate cells in flat preparations of rat
and rabbit retinal epithelium, and inspection of tangential sections of the retinal
epithelium from adult mosaics and chimaeras shows that these contain both
binucleate and unicleate cells (see Fig. 3A), whereas there is no clear evidence of
binucleate cells in similar sections from newborn mosaics and chimaeras. Part
Chimaeric and Mosaic Retinal Epithelia
457
of the increase in the mean number of nuclei per coherent clone after birth may
be due to formation of binucleate cells.
Comparison between the coherent clone sizes for developing mosaic and
pigmented-Q <-> unpigmented-Q chimaeras show a marked similarity at 12\ days,
and only a small difference after birth. The high proportion of pigmented cells,
shown in Fig. 2B, for the 12^-day chimaeric group might suggest that the eyes
are non-chimaeric and the development of pigment is incomplete. Although
Tarkowski (1964) found unpigmented cells in some of his control stocks up to
13 days, this seems an unlikely explanation here. The chimaeric eyes were at
least as well developed as the mosaics shown in Fig. 2D, and the chimaeras
were readily distinguishable from fully pigmented embryos of the same age. Of
the six 12^-day chimaeric embryos used, five were littermates and it is possible
that either genetic or developmental differences between the pigmented and
unpigmented embryos used as aggragants are responsible for the high proportion of pigmentation. Even if the comparison between the 12^-day embryos is
ignored, the results from chimaeras at later stages, which show no marked bias
in proportion of pigmentation, suggest similar patterns of development for
mosaics and this chimaeric group. While the present analysis clearly indicates
a basic similarity in clonal development between chimaeras and mosaics, the
mean size of the coherent clones may be consistently underestimated if they are
very irregular in shape.
Comparisons between adult chimaeras of different strain combinations
suggest that (C57BL x C3H)FX <-> Recessive chimaeras have more nuclei per
coherent clone that the other groups considered, although crude estimates
indicate that the number of nuclei per retinal epithelium is very similar in all
chimaeric groups. Presumably coherent clonal growth in this strain combination
has been less disrupted by cell mixing than in the other groups studied. One
possibility is that the association between cells of unlike genotype is less stable
than associations of like genotype in this particular strain combination.
The smaller differences between other groups may also be due to differences
in cell interactions. The two cell populations in a mosaic will only differ genetically by the activity of X-linked genes, and so differences in interactions between
cells of the two populations are less likely in mosaics than in chimaeras. The
smaller mean coherent clone size shown by mosaics than by any chimaeric
group is consistent with a slightly greater tendency of cells of like genotype to
stay together in chimaeras. However, the difference may be simply an artifact of
the method of analysis. It has already been noted that some chimaeras have very
unequal proportions of the two cell populations, and in these cases the coherent
clone size estimation is based on fewer patches, and so is less reliable than usual.
Mosaics, as Deol & Whitten (1972 a) have also noted, tend to have more equal
proportions of pigmented and unpigmented cells.
Deol & Whitten's one dimensional analysis of histological sections of the retinal
epithelium also revealed that the mean number of patches in mosaics was three
458
J. D. WEST
times that in C57BL/10Wt <-> SJL/Wt chimaeras (Deol & Whitten, 1972a). If
this reflects a similar difference in coherent clone number, the results suggest a
far greater difference between mosaics and chimaeric coherent clones than any
seen in the present study. Interpretations of these results have been made to suggest either late X-chromosome inactivation (Deol & Whitten, 1972a), or coherent
clonal growth, in both developing chimaeric and mosaic embryos, following
X-inactivation earlier in development (Nesbitt, 1974).
Much of the difference between patch length in mosaics and the C57BL/
lOWt <-> SJL/Wt chimaeras is probably due to the observed differences in the
proportions of the two cell-populations. More mosaics have nearly equal proportions which, in a one-dimensional analysis, will result in a larger number of
patches even if the coherent clone sizes are equal to the chimaeric coherent
clones. Strain specific interactions between C57BL/10Wt and SJL/Wt might also
occur and reduce the degree of cell mixing in these chimaeras, and account for
part of the difference between the results from Deol & Whitten's analysis and the
present study. A combination of strain-specific interactions between C57BL/
lOWt and SJL/Wt cells and a difference in proportions of the two cell-populations between the mosaics and chimaeras, would probably reconcile Deol &
Whitten's results with early X-inactivation without the need to postulate limited
coherent clonal growth.
The present observations agree with Deol & Whitten's finding that chimaeric
retinal epithelia tend to have less equal proportions of the two cell-populations.
The more equal proportions seen in the mosaic pigmented epithelia may partly
reflect more equal proportions of the two cell-populations in the whole body
(see Nesbitt, 1971), which might be expected if X-inactivation occurred fairly
soon after the formation of the inner cell mass, as is widely believed (Lyon,
1972).
Several authors have compared chimaeric and mosaic phenotypes in order to
investigate the validity and timing of X-chromosome inactivation in the mosaics.
In a number of cases larger patch sizes have been claimed for chimaeras. These
differences can only be interpreted on the basis of late X-inactivation if the larger
chimaeric patches are known not to result from differences in the proportions
of the two populations, or from a tendency for cells of like genotype to remain
together in the chimaeric group. These possibilities were not excluded in the
studies of the retinal epithelium (Deol & Whitten, 1972a), migratory melanocytes of the eye and inner ear (Deol & Whitten, \912b) and tail banding patterns
caused by the tabby gene (Ta) (McLaren, Gauld & Bowman, 1973).
In conclusion, the retinal epithelium of X-inactivation mosaics and chimaeras
show a broad phenotypic similarity and it is suggested that there is no basic
difference in the coherent clone size or the pattern of clonal development
between mosaics and chimaeras. This similarity may be masked by cellular interactions in chimaeras of some strain combinations and observed differences
can be attributed either to cell selection or to early sampling processes (such as
Chimaeric and Mosaic Retinal Epithelia
459
the formation of the inner cell mass) which occur after the two chimaeric
populations are aggregated but before X-inactivation. Comparisons between
mosaics and chimaeras provide no conclusive evidence on the timing of Xchromosome inactivation. The present analysis suggests that at \2\-&a.ys postcoitum cell mixing in the retinal epithelium is sufficient to disrupt any clonal
growth but after this time limited coherent clonal growth occurs in both mosaics
and chimaeras until, in the adult, coherent clones comprise an average of five or
six nuclei.
I wish to thank Dr Anne McLaren for generous help, supervision and encouragement given
to me throughout this study. I am also grateful to Drs Anne McLaren, Patricia Bowman and
Mrs Janet Carter for providing the chimaeras, to Mr B. Doyle for reliable technical help, and
to Dr W. K. Whitten of the Jackson Laboratory, Bar Harbor, Maine, for reading the manuscript and making many helpful suggestions. This study was supported by a Science Research
Council Studentship, and by the Ford Foundation.
REFERENCES
ABERCROMBIE, M. (1946). Estimation of nuclear population from microtome sections. Anat.
Rec. 94, 239-247.
CATTANACH, B. M. (1961). A chemically-induced variegated type position effect in the
mouse. Z. VererbLehre 92, 165-182.
DEOL, M. S. & WHITTEN, W. K. (1972a). Time of X chromosome inactivation in retinal
melanocytes of the mouse. Nature New Biol., Lond. 238, 159-160.
DEOL, M. S. & WHITTEN, W. K. (19726). X-chromosome inactivation: does it occur at the
same time in all cells of the embryo? Nature New Biol., Lond. 240, 277-279.
LYON, M. F. (1972). X-chromosome inactivation and developmental patterns in mammals.
Biol. Rev. 47, 1-35.
MCLAREN, A., GAULD, I. K. & BOWMAN, P. (1973). A comparison between mice chimaeric &
heterozygous for the X-linked gene tabby. Nature, Lond. 241, 180-183.
MCLAREN, A. & BOWMAN, P. (1969). Mouse chimaeras derived from fusion of embryos
differing by nine genetic factors. Nature, Lond. 224, 238-240.
MICHIE, D. (1955). Genetical studies with 'vestigial tail' mice. III. New independence data.
/. Genet. 53, 285-294.
MINTZ, B. (1971). The clonal basis of mammalian differentiation. In Control Mechanisms of
Differentiation and Growth (eds. D. D. Davies and M. Balls), pp. 345-370. Symp. Soc. Exp.
Biol. 25. Cambridge University Press.
MINTZ, B. & SANYAL, S. (1970). Clonal origin of mouse visual retina mapped from genetically
mosaic eyes. Genetics, Princeton (Suppl.), 64, S43-S44.
MYSTKOWSKA, E. T. & TARKOWSKI, A. K. (1968). Observations on CBA-p/CBA-T6T6 mouse
chimaeras. /. Embryol. exp. Morph. 20, 33-52.
NESBITT, M. N. (1971). X-chromosome inactivation mosaicism in the mouse. Devi Biol.
26, 252-263.
NESBITT, M. N. (1974). Chimaeras vs X-inactivation mosaics: Significance of differences in
pigment distribution. Devi Biol. 38, 202-207.
ROACH, S. A. (1968). The Theory of Random Clumping. London: Methuen.
SANFELICE, F. (1918). Recherches sur la genese des corpuscules du Molluscum contagiosum.
Annls Inst. Pasteur, Paris, 32, 363-371.
FARKOWSKI, A. K. (1964). Patterns of pigmentation in experimentally produced mouse
chimaeras. /. Embryol. exp. Morph. 12, 575-585.
Ts'o, M. O. M. & FRIEDMAN, E. (1967). The retinal pigment epithelium. I. Comparative
histology. Archs Ophthal., N.Y. 78, 641-649.
460
J. D. WEST
J. D. (1975a). A theoretical approach to the relation between patch size and clone size
in chimaeric tissue. J. theor. Biol. 50, 153-160.
WEST, J. D. (19756). Cell populations in mouse chimaeras and mosaics. Ph.D. Thesis, Edinburgh University.
WEST,
{Received 18 July 1975, revised 23 December 1975)
APPENDIX
Estimation of mean cell area in retinal epithelium
Abercrombie (1946) showed that the estimation of the nuclear population
density from histological sections is only possible if the mean nuclear size is
taken into account. Some of the nuclei visible in a section are whole nuclei,
whereas some are nuclear fragments, so extrapolation of crude counts per area
to apparent number of nuclei per volume will result in over-estimation. This is
relevant to the estimation of the mean cell area over the surface of the retinal
epithelium, where the extrapolation is one of counts per length to counts per
area. Abercrombie derived the following equation to correct for the exaggeration of nuclear counts:
where P is the average number of ' nuclear points' per section, A is the crude
count of visible nuclei per section, M is the section thickness (in /m\) and L is
the average length of the nuclei (in /*m) perpendicular to the plane of the section.
A 'nuclear point' is any geometrical point of the same relative position in all
nuclei and cannot overlap two adjacent sections. The function M\{L + M) is the
proportion of visible nuclei whose 'nuclear points' lie within the section.
In chimaeric or mosaic retinal epithelia the nuclei can only be clearly seen in
unpigmented patches, so the estimation is based entirely on one population of
cells. The length of the nuclei was not measured in a plane perpendicular to the
section as it was shown that the nuclei were symmetrical. The mean nuclear
length, based on twenty nuclei, for each of ten mosaic eyes, sectioned in the
'latitudinal' plane (6-31 ±0-11 fim) did not differ significantly from the equivalent mean nuclear length derived from sections in the 'longitudinal' plane
(5-96 ±0-16 /an). (Values from a Student's f-test: t = 1-94; P > 0-05.)
In tangential sections, cells of the retinal epithelium appear roughly symmetrical, and this assumption is supported by the similarity of estimates of
cell length from longitudinal and latitudinal sections of the same eyes (West
1975 b). If the estimated mean cell length (parallel to the epithelial surface) is
termed C the cell area is C2. The length of unpigmented retinal epithelium
considered will be called R. The cell area to be considered is the surface area,
perpendicular to the plane of the section, and is equal to the area of retinal
epithelium (RxM) divided by the number of nuclear points (P), or:
r2_RM
L
"
P '
As
Chimaeric and Mosaic Retinal Epithelia
M 1
P=
„
cell area
_
i?M
C2 = —7- x
A
A
cell length
C
461
M
R(L + M)
= J(«±±^),
where L is the mean nuclear length, visible in the section and parallel to the
surface of the epithelium, Mis the section thickness (6 /im), and A is the number
of nuclei counted in the length of unpigmented retinal epithelium, R. This estimate of cell area (C2) is a measure of the area per nucleus and is equivalent to
the cell area only if each cell is uninucleate.
REFERENCES
M. (1946). Estimation of nuclear population from microtome sections.
Anat. Rec. 94, 239-247.
WEST, J. D. (19756). Cell populations in mouse chimaeras and mosaics. Ph.D. Thesis. Edinburgh University.
ABERCROMBIE,