J. Embryol. exp. Morph. 83, 137-156 (1984)
Printed in Great Britain © The Company of Biologists Limited 1984
The early development of mystacial vibrissae in the
mouse
By JOAN T. WRENN AND NORMAN K. WESSELLS
Department of Biological Sciences, Stanford University, Stanford,
California, 94305, U.S.A.
SUMMARY
The initial generation of the pattern of mystacial vibrissae (whiskers) in the mouse is
described. The maxillary process is present in 10-day embryos but has a relatively flat surface.
Beginning at approximately 11-5 days, thefirstsign of vibrissal development is the formation
of ridges and grooves on the maxillary and lateral nasal processes. Thefirstvibrissal rudiment
to form subsequently appears posterior to the most ventral groove on the maxillary process.
It is the most ventral whisker of the posterior, vertical row. The next few rudiments appear:
(1) dorsal to the first, also in the vertical row; and (2) anterior to the first, on the ventral-most
ridge and in the groove beneath it. Formation of vibrissal rudiments continues in a dorsal and
anterior progression usually by an apparent partitioning of the ridges into vibrissal units.
The hypothesis that this patterning of mystacial vibrissae might be determined by the
pattern of innervation in the early mouse snout was investigated. Nerve trunks and branches
are present in the maxillary process well before any sign of vibrissal formation. Because
innervation is so widespread there appears to be no immediate temporal correlation between
the outgrowth of a nerve branch to a site and the generation of a vibrissa there. Furthermore,
at the time just prior to the formation of the first follicle rudiment, there is little or no nerve
branching to the presumptive site of that first follicle while branches are found more dorsally
where vibrissae will not form until later. Thus, a one-to-one spatial correlation between nerve
and follicle sites also appears to be lacking.
The developmental changes in ultrastructure within the neurites of the trunks and branches
as well as the apparent rearrangements of the nerve trunks suggest that early innervation of
the snout is a labile phenomenon and that the vibrissal pattern begins to be established before
the neural pattern is completely developed. The results indicate that vibrissal pattern formation is likely to be a complex process relying on the interplay of the cells and tissues involved,
rather than on unidirectional instructions from neurons to other cell types.
INTRODUCTION
The development of the mystacial vibrissae (sinus hairs, whiskers) of the
mouse is marked by a high degree of order and pattern in time and space (Danforth, 1925; Gruneberg, 1943; Davidson & Hardy, 1952; Yamakado & Yohro,
1979; Van Exan & Hardy, 1980). The earliest signs of vibrissal formation occur
between 11 and 12 days of gestation as two parallel grooves and two ridges
appear on the maxillary process of the developing snout. In general, the rudiments of the most posterior (caudal) and ventral vibrissae appear first and succeeding vibrissae form in an anterior (rostral) and dorsal progression. By day 14
138
J. T. WRENN AND N. K. WESSELLS
most, if not all, rudiments have formed and the final adult pattern of five
horizontal rows with one vertical row posterior to them is apparent. This pattern
is maintained in most strains of mice studied although there are small variations
in the number of follicles in the horizontal rows (Dun, 1958), and supernumerary
follicles sometimes occur between rows (Yamakado & Yohro, 1979). The
general pattern, however, is quite constant.
The search for the source of the pattern has centred on the fact that vibrissae
are highly innervated sensory organs (Vincent, 1913; Winkelman, 1959) which
are represented in the somatosensory cortex of the brain by their cortical 'barrels' (Woolsey & Van der Loos, 1970). These are discrete groups of neurons, one
unit for each vibrissa, which are arranged in a pattern corresponding to the
pattern of the vibrissae on the snout. This spatial correspondence between the
central nervous system and the periphery suggests that the organization of one
might influence the patterning of the other. Several studies have shown that
nerve trunks are present in the snout both directly beneath the early vibrissal
rudiments and also more anteriorly in the region where rudiments will
subsequently form (Tello, 1923; Wessells & Roessner, 1965). Van Exan & Hardy
(1980) studied the innervation pattern of 12-day embryonic snouts with light
microscopy. They made reconstructions of embryos in which the more caudal
vibrissae were already well established and found that the formation of a nerve
'plexus' preceded the development of the more rostral follicles. These observations suggest the possibility that, since innervation precedes vibrissal formation,
the pattern of that innervation determines the arrangement of vibrissae.
The studies reported here were undertaken to investigate the initial generation
of the vibrissal pattern and its relationship, if any, to the pattern of innervation
at that time.
MATERIALS AND METHODS
Scanning electron microscopy
Mouse embryos were obtained by crossing Balb/c females with C3H male
mice and were staged by day of development with the date of vaginal plug
discovery set at day 0. Embryos of 10,11,11-5,12 and 13 days were taken from
pregnant females which had been killed by cervical dislocation. The uteri were
removed and placed in Hank's balanced salt solution. The heads of the embryos
were removed and immediately put into the primary fixative of 2-5 % glutaraldehyde in 0-07 M-Sorenson's buffer with 0-06 M-sucrose, pH 7-4. There the heads
were cut into right and left halves and trimmed, depending on their age and size,
to leave at least the eye and the area beneath it, the medial and lateral nasal
processes, and the maxillary process. For almost all but the youngest embryos
the right side of the snout was left whole and the left side was cut into two parts
in a plane horizontal through the length of the maxillary process. The tissues
Early development of mystacial vibrissae
139
remained in the primary fix for approximately 1 h before being washed three
times with the same Sorenson's buffer.
Postfixation was in 1 % osmium tetroxide in 0-028 M-Veronal buffer, pH7-4,
for 30 to 60min, after which the tissues were dehydrated through a series of
acetone and dried with CO2 in a Sorvall critical-point drying system. The tissues
were then mounted on stubs in various orientations with double-sided tape and
sputtered with gold in an argon atmosphere in a Denton vacuum evaporator. The
specimens were examined in a Coates and Welter Model 100-6 field emission
scanning electron microscope.
Light and electron microscopy
Sample preparation for transmission microscopy was the same as for SEM
through postfixation in osmium. After that, specimens for transmission microscopy were dehydrated in a graded series of ethanol which was then replaced by
propylene oxide. The tissues were infiltrated with an Epon 812-propylene oxide
mixture, followed by Epon alone; then they were embedded in fresh Epon and
were cured for one day at 45 °C and two days at 60 °C.
Sections of 1 ,um thickness for light microscopy were cut on an LKB Ultrotome
III and stained with a 1:1 mixture of 1 % azure II in water and 1 % methylene
blue in 1 % sodium borate (Richardson, Jarett & Finke, 1960). Thin sections for
electron microscopy were also cut on the LKB. They were stained with uranyl
acetate and lead citrate before being examined in a Hitachi HU-11E.
Half mounts
Some specimens were treated with acetylthiocholine according to the methods
of Karnovsky & Roots (1964) and Barald & Berg (1979) to detect cholinesterase
activity and thereby aid in viewing nerve cells and their neurites.
RESULTS
Whole mounts
The initial generation of vibrissal rudiments, once started, occurs so rapidly
that embryos must be examined in half-day increments of age in order to
resolve the formation of the early pattern. The search for the order of development is aided, however, by the wide range of developmental characteristics
within the half-day increments. Not even apparently healthy litter mates will
have equally developed facial features at the same hour of gestation
(Gruneberg, 1943). This report will include the range of sizes and features at
each age, although the illustrations will be of specimens selected to show the
progression of development. The elements of the pattern, the rows and the
individual vibrissae, will be labelled according to Fig. 1, with row 1 being the
most ventral row.
140
J. T. WRENN AND N. K. WESSELLS
Fig. 1. A diagrammatic representation of the right side of an embryonic mouse face,
ca. 13 days of gestation. The rows and the individual whiskers are labelled in accord
with their approximate order of development so that row 1 is the most ventral row
and row 5, the most dorsal.
In Figs 2 through 10 of this paper, only the right side of the whole snout will be
shown.
10 days
At 10 days all embryos show the maxillary process diverging from the mandibular process beneath the eye (Fig. 2). The lens of the eye is invaginating and
the olfactory pit in various specimens ranges from broad and relatively shallow,
to narrower and more groove like. The maxillary and lateral nasal processes are
quite short compared to their later outgrowth. The length of the nasolacrimal
groove between them, anterior to the eye, ranges from approximately 100 jm\ to
160[Am (mean = 132/an, S.D. = 21 fm\) in the eight specimens studied. No
vibrissal rudiments are evident at this stage.
11 days
The primary change in the 11-day specimens (Fig. 3) is the forward growth of
the facial processes. The distance from the anterior edge of the eye to the most
anterior point of the maxillary process ranges from approximately 260 jum to
480 fim (mean = 360 fim, S.D. =70 /im in 20 specimens). This wide range reflects
the differences in growth and development in the same-aged embryos reported
by Gruneberg (1943). The nasolacrimal groove provides a deeper separation
between the maxillary and lateral nasal prominences. No vibrissal rudiments are
seen.
11 i days
As the facial processes continue their outgrowth, the nasolacrimal groove
Early development of mystacial vibrissae
141
becomes shallower in its anterior portion. In smaller, less-developed specimens
(Fig. 4) the length of the maxillary process anterior to the eye overlaps with the
length of 11-day samples, beginning at about 320 /im. In more developed 11-5day snouts (Fig. 5), that measurement of the maxillary process ranges up to 680 jUm
in length (mean = 470/im, S.D. = 80]Um in 17 samples).
Most importantly, in these larger, more-developed samples there is an indication of two new grooves appearing on the surface of the maxillary process, more
or less parallel to the length of the nasolacrimal groove. With the two depressions
appear two ridges (the precursors of vibrissal rows 2 and 3), one between the two
new grooves and one between the more dorsal of the depressions and the
nasolacrimal groove. In addition, in some specimens another groove and two
ridges (rows 4 and 5) appear on the surface of the lateral nasal process.
12 days
12-day embryos exhibit a very broad range of facial and vibrissal development
(Figs 6-10). In size, the length of the maxillary process anterior to the eye varies
from approximately 650 fjxn to 1200 /im (mean = 910 /an, S.D. = 150 jum) in the 33
samples studied. The nasolacrimal groove becomes much shallower to form the
groove between rows 3 and 4.
The earliest, shortest samples show that the first vibrissa appears as a hillock
at the posterior end of the ventral-most groove on the maxillary process (Fig. 6).
This is the rudiment of the most ventral whisker in the vertical row, vibrissa a.
A second hillock, vibrissa /3, appears dorsal to the first (Fig. 7), at the posterior
end of the more dorsal groove of the maxillary process. The next rudiments
appear at or about the same time anterior to the first two (Figs 7, 8), one on the
ventral ridge (horizontal row 2) and one in the groove beneath the ridge
(presumptive row 1).
More advanced 12-day specimens show vibrissal formation proceeding dorsally and rostrally as two more hillocks appear in the vertical row posterior to the
ridges, new vibrissae form as the ridges themselves are blocked off into new
hillocks, and more rudiments form in the ventral-most groove (Figs 9,10).
Late 12-day and 13-day embryos continue to form new whiskers in the
established pattern from ventral to dorsal and posterior to anterior. The whisker
pad becomes elevated above the level of the surrounding head tissue and the
epithelium of the earliest rudiments begins its downward growth into the mesenchyme to form whisker follicles. These later stages have been thoroughly
described by Hardy (1951), Davidson & Hardy (1952), and Yamakado & Yohro
(1979).
Half mounts and sectioned snouts
Large nerve trunks are present in the maxillary process very early as it is
growing out, before any signs of vibrissal formation. Fig. 11 shows an 11-5-day
maxillary process which has been favourably bisected frontally for scanning
142
J. T. WRENN AND N. K. WESSELLS
electron microscopy. Lengths of nerve trunk course through the mesenchyme
after leaving the trigeminal ganglion which is the source of sensory innervation
to the vibrissae (Vincent, 1913).
Thiocholine staining of snouts, bisected as above, permits the visualization of
nerves on and beneath the surface of the cut so that their branching patterns are
more easily viewed. Although it is not known how deeply the stain penetrates in
these half mounts, it is clear that vibrissal rudiments in 13-day snouts (Fig. 12)
are associated with well-defined nerve branches coming to them and that other
nerve branches course more rostrally where vibrissal rudiments are not yet apparent.
The information provided by examining bisected snouts is, however, severely
limited. Only those portions of nerves in or near the plane of the cut can be seen.
Furthermore, it is impossible to distinguish between small nerve fibres or terminals and the surrounding mesenchymal cells.
To gain more information about the distribution of neurites relative to the
positions of developing vibrissae it was necessary to section the snouts in various
planes for light and transmission electron microscopy. In general, 1/im-thick
plastic sections were taken first and examined for areas of special interest requiring higher resolution. When such areas were seen, sections 50 nm to 70 nm thick
were taken of the adjacent tissue and examined in the TEM. In this way very
small neural elements could be identified, their ultrastructure observed, and
Fig. 2. The right half of the facial region of a 10-day mouse embryo. The lens (/) of
the eye is in an early stage of invagination. The lateral (In) and medial (mn) nasal
processes surround the olfactory groove. The maxillary process (mx) diverges from
the mandibular process (md) beneath the developing eye. The nasolacrimal groove
(ng) separates the maxillary and lateral nasal processes. Bar = 100pm.
Fig. 3. An 11-day facial region. The lens (/) is completing its invagination; only a
small opening remains in the surface epithelium of the eye. The maxillary (mx) and
lateral nasal (In) processes have both grown out anterior to the eye and the
nasolacrimal groove between them has deepened. Bar = 100 fun.
Fig. 4. An early 11-5-day mouse snout. The surface epithelium (e) of the eye is now
continuous over the lens. The nasolacrimal groove (ng) is quite deep between the
maxillary (mx) and lateral nasal (In) processes which have continued their outgrowth. Bar= 100 /an.
Fig. 5. A more advanced 11-5-day facial region. Note the two depressions (arrows)
which appear on the maxillary process and the single slight groove (arrow) on the
lateral nasal process. The raised ridges between the grooves are the precursors of
horizontal rows 2 through 5 of vibrissae. Bar = 100/zm.
Fig. 6. An early 12-day snout. At the posterior end of the most ventral groove is a
slight hillock, the first sign of the first vibrissal rudiment (a) to form. The
nasolacrimal groove (ng) has become much shallower than it was at 11 days.
Bar = 100/im.
Fig. 7. A slightly later stage of a 12-day mouse embryo. The first vibrissa (a) of the
vertical row, is more apparent, and the second (fi) has begun to form. Likewise, the
initial signs of vibrissa 2a appear on its ridge. Bar = 100/an.
*~Tfc^
ss^afe:^1
8
10
Early development of mystacial vibrissae
143
their connections followed to larger nerve bundles, thus providing maps of the
innervation pattern at the time of initial vibrissal formation.
Frontal sections of 11-day snouts, parallel to the length of the maxillary
process, show that nerve trunks are present in the process approximately one day
before any sign of the first vibrissal rudiment. There is limited branching of the
trunks toward the periphery (Fig. 13). Light micrographs show that mesenchymal cells of the maxillary process are very densely packed at 11 days, especially in the areas adjacent to the nerve trunks. Away from the nerve tracts, toward
the epithelium, the mesenchyme is slightly less densely packed but, compared to
the later stages, there is little extracellular matrix. A network of capillaries lies
in the mesenchyme between the nerves and the epithelium. At this stage the
capillaries appear to remain at least 6 or 7 cell diameters away from the
epithelium.
Ultrastructurally, neurites in the large nerve trunks appear somewhat
'immature' (Fig. 14) compared to those seen at later stages. Some of the individual neurites of a bundle contain the organelles typical of axons: numerous
neurotubules, neurofilaments, and mitochondria in a ribosome-free cytoplasm.
These organelles are often loosely arrayed, however, in neurites whose profiles
may measure up to 4/im or 5 /im in diameter, larger than the 1 /im-wide profiles
which, as we shall see, are typical of axons at later stages. Other neurites in a
bundle contain few, if any, of these organelles. These neurites appear to be more
like growth cones with much fine filamentous material, dense-core granules, and
Fig. 8. A later stage of facial development at 12 days. Vibrissa la is forming in the
groove beneath 2a. There are hints of hillock development at 3a and in the vertical
row at positions y and A. The remnant of the nasolacrimal groove (ng) separating
rows 3 and 4 appears to be similar in depth to the more recently formed grooves on
the maxillary process. Bar = 100jum.
Fig. 9. Another 12-day specimen. All four rudiments in the vertical row are present
as are more rudiments in the horizontal rows, including a hint of vibrissa 4a above
the nasolacrimal groove (ng). Bar = 100 jUm.
Fig. 10. A late 12-day embryo. At least 12, and possibly 14, rudiments have formed
and development is proceeding in ventral-(V)-to-dorsal (D) and posterior-(P)-toanterior (A) directions. Bar = 100 ^m.
Fig. 11. The bottom half of an 11-5-day maxillary process and adjacent tissue which
has been bisected frontally, i.e., in a horizontal plane with the anterior (A), nasal
portion of the process at the bottom and the posterior (P) portion toward the top.
The posterior boundary of the process proper is approximately at the level of the P.
Vibrissae will form on the outer surface to the right. At the top lies the trigeminal
ganglion (tg). Nerves (n) are seen leaving the ganglion in a large trunk and smaller
sections of nerve tracts (arrows) are present in much of the length of the process.
Bar =100jmi.
Fig. 12. The lower half of a frontally bisected 13-day maxillary process stained with
acetylthiocholine to demonstrate a portion of the innervation pattern. Developing
vibrissae (v) appear as small, rounded contours on the epithelial surface and some
of the nerve branches (n) terminate near them. The trigeminal ganglion (tg) appears
at the upper right of the photo. Bar = 100 jum.
144
m
J. T. WRENN AND N. K. WESSELLS
Early development of mystacial vibrissae
145
membranous structures, some vesicular and some reminiscent of smooth
endoplasmic reticulum, in a homogeneous, ribosome-free cytoplasm. Finally,
some neurites have an intermediate appearance with both groups of organelles
present. Occasionally small filopodia with few organelles are seen extending
from the larger profiles. In general, the appearance of the large nerve trunks at
11 days suggests that most of the neurites have just recently grown out or are in
the process of growing out even as the maxillary process itself is enlarging.
By 12 days of development ridges are present on the maxillary process and
vibrissae have begun to form (Figs 6-10). Frontal sections parallel to the ridges
(the same orientation as in the 11-day sections described above) show the
epithelial hillocks and their underlying dermal condensations. In Fig. 15 the
vibrissa marked j8 on the right is a member of the vertical row (Fig. 1). The plane
of section passes through one side of the structure. Vibrissa a is the first to
develop in its horizontal row, and, to its left, b is just beginning to form. Within
the more advanced condensations as in j3, the mesenchymal cells are closely
apposed. Outside of the vibrissal dermal condensations the mesenchymal cells
are much less densely packed than they were at 11 days. The growth of the
maxillary process mentioned above appears to be at least partly attributable to
the increase in extracellular matrix within the process.
The nerves of the 12-day snout are more highly branched than at 11 days. Fig.
16 shows several main nerve trunks near the bottom of the picture and portions
of nerve tracts nearer the periphery. On the right a large nerve trunk splays out
beneath the dermal condensation of vibrissa /3 (the same vibrissa as seen in Fig.
15 but sectioned more centrally). Transmission EM of the same nerve bundle in
a section approximately 5 /im away from that in Fig. 16 shows that beneath the
condensation the neurites have a typical growth-cone appearance (Fig. 17). At
the end of that portion of nerve trunk away from the condensation several
neurites resemble axons with arrays of neurotubules and neurofilaments (Fig.
18), but most of the neurites at this level have an intermediate appearance similar
to that described for the neurites of 11-day nerve trunks.
This hint of a proximodistal sequence of neurite ultrastructure is reinforced by
Fig. 13. A light micrograph of an 11-day maxillary process sectioned frontally with
the anterior (A) end of the process to the right and the posterior (P) end to the left.
Large nerve trunks (n) with little branching are present in the cell-dense mesenchyme (m). A network of capillaries (c) filled with red blood cells lies about six or
seven cell diameters from the epithelial surface (e). Bar = 50/xm.
Fig. 14. A transmission EM view of a region of nerve tract (11 days) similar to those
seen in Fig. 13. The nerve is composed of individual neurites of varying size and
contents. Some neurites (ax) have regions containing the organelles typical of mature
axons, i.e., neurotubules and neurofilaments. Other neurites look more like growth
cones (gc) with dense-core granules and membranous structures in a rather
homogeneous background cytoplasm. Many neurites are intermediate (i) in appearance with some features of both types. Note how closely the mesenchymal cells
(m) at the bottom are apposed to the neurons. Bar =
J. T. WRENN AND N. K. WESSELLS
V * -
-# .V.r:
Figs 15 and 16. Light micrographs of sections of 12-day maxillary process taken
approximately 50/um apart. The anterior (A) of the snout is to the left and the
posterior (P) is to theright.The vibrissa marked /?in eachfigureis the same structure
cut at different levels. It is a member of the posterior, vertical row of follicles.
Vibrissa a in Fig. 15 is the first to form in its horizontal row, and the beginning of
vibrissa b is suggested by the slight thickening of the epithelium seen in this section.
In both figures numerous small sections of nerve trunks are scattered through the
mesenchyme (arrows); some are far anterior to the youngest developing vibrissa.
Fig. 16 shows a longitudinal section through a major nerve (n) with a small branch
coming from it (br). Other nerve trunks (nt) lie deep in the mesenchyme.
Bars = 100 jum.
further study of nerve trunks at various distances from the epithelium. In general
most of the individual neurites of the main nerve trunks deep in the mesenchyme
are axon like with neurotubule and neurofilament arrays in neuritic cell profiles
of 0-5 /an to 1 jum in diameter (Fig. 19). This gives the trunks an appearance of
Early development of mystacial vibrissae
147
'maturity' compared to those at 11 days. The population is mixed, however, and
neurites with an intermediate appearance (described above) and growth cones
are also present at this level. The closer the nerves get to the periphery, the
greater the proportion of their neurites which look like growth cones. In all cases
these remain at least four or five mesenchymal cell diameters away from the
epithelium. No nerve branches or endings have been observed to contact either
the basal lamina or the epithelium.
Figs 15 and 16 show a number of nerve bundles across the width of their
sections. These are typical in that virtually any single 1 /im-thick, frontal section
will have nerve branches both in the vicinity of obvious dermal condensations
and also further nasally where there is, as yet, no sign of vibrissal development.
In an attempt to represent the distribution of nerves in one portion of serially
sectioned 12-day maxillary process, drawings were made on clear plastic of
approximately every second 1 jUm-thick section (electron microscopy was used
for nerve identification when necessary). The drawings were mounted with air
space between them to approximate the depth of the intervening sections. The
three-dimensional representations thus obtained (Fig. 20) show that neural elements are distributed through the length of the maxillary process including the
rostral portion some distance beyond the last vibrissal site. If one were to divide
the area between the site of the youngest recognizable rudiment and the site of
the most nasal nerve endings into units approximately the same size as the
already developing vibrissae, one could say that nerves are present at least two
or three units in advance of the next site to develop. In truth, however, the great
forward growth of the maxillary process during early vibrissal formation makes
such a division into units meaningless in terms of predicting future follicle sites.
The best one can say is that nerves are present at sites in the mesenchyme well
in advance of the time that dermal condensations would be forming near those
sites.
To investigate the possibility that the nerve pattern may play a role at an
earlier stage, before the first vibrissa begins to form, sections were taken of 11-5day embryonic snout in a plane perpendicular to the length of the ridges on the
maxillary process (Figs 21-24). The sectioning was done in a posterior-toanterior direction through the trigeminal ganglion beneath the eye and continuing well into the maxillary and nasal processes. In this manner one can trace the
smaller branches of the superior maxillary branch of the trigeminal ganglion and
can see the position of those branches relative to the ridges and to the sites of
initial vibrissal formation.
Serial 1/im-thick sections at 11-5 days reveal that the maxillary branch is
composed of approximately 20 to 25 discrete nerve trunks interspersed with
mesenchymal cells and small capillaries as it leaves the trigeminal ganglion. The
trunks remain together as a loosely associated group until they enter the region
of the maxillary process proper, i.e., in sections beyond the eye where the
nasolacrimal groove separates the maxillary and nasal processes. Then the
148
^ £^
J. T. WRENN AND N . K. WESSELLS
'•:<i1A>
;^^w;>^^::*;iv--
..;.r:. *. ;i> ; ^.. :. -r : .• • . V f ^ g •'•*•' -'r-*^
'>^R1. i > - 1
fc^;
^
i*v rf. \ ^ - ^ ? ^ W ^ '• **•
Early development of mystacial vibrissae
149
Fig. 20. Three reconstructions from different angles of the same portion of 12-day
maxillary process in which a few vibrissae have begun to form. In each view solid lines
indicate nerves on the surface toward the viewer. Nerve trunks and branches are
distributed throughout the tissue, including in the anterior region (A) well beyond
the last vibrissal site. Posterior = P.
Fig. 17. The distal end of the nerve tract directly beneath the dermal condensation
(dc) of vibrissa j3 in Fig. 16 (arrow 1). Most of the neurite profiles are enlarged and
contain the organelles such as dense-core granules (dcg) and smooth endoplasmic
reticulum (ser) which are typical of growth cones. Bar = 1 jum.
Fig. 18. The proximal end of the portion of nerve tract beneath vibrissa /?, i.e., the
end away from the dermal condensation (arrow 2 in Fig. 16). At this level of the nerve
tract, the population of neurites is mixed in appearance: some resemble growth cones
(gc), some resemble axons (ax) with arrays of neurotubules (ntu), and some are
intermediate (i), containing the organelles of both types. The general appearance of
the nerve is similar to that seen in nerve trunks at 11 days (Fig. 14). Bar =
Fig. 19. A portion of the major nerve trunk (nt) seen deep in the mesenchyme of Fig.
16. At this level of the nerve most of its neurites are axon-like (ax) with smaller
profiles and neurotubules and neurofilaments. However, there are some larger
profiles which resemble growth cones (gc). This suggests that later neurites continue
to grow out along the established nerve tract. Bar =
150
J. T. WRENN AND N . K. WESSELLS
D
Early development of mystacial vibrissae
151
trunks begin to splay out in a dorsoventral direction across the base of the
process. Continued sectioning passes through the posterior end of the dorsal
groove on the maxillary process as the trunks continue to spread dorsoventrally.
Within the next 10 to 20 sections (Figs 21-22) the more dorsal nerve trunks send
several branches into the mesenchyme beneath the dorsal ridge (the area between the dorsal groove and the nasolacrimal groove, presumptive row 3 of
vibrissae). Approximately 40 sections further on, the ventral groove begins to
appear. It is within the area of these sections, just before the groove, that the first
vibrissa would have begun to form if development had continued for another half
day. But within these sections there is no sign of the more ventral nerve trunks
branching toward the epithelium. A small branch does appear approximately
20 /im more anteriorly, but it is not until another 40 or so sections are taken that
the ventral nerve trunks begin to branch (Figs 23,24) as profusely as those under
presumptive row 3. Those more dorsal nerve bundles have continued their branching throughout these sections into regions where vibrissae would not be
developing until after they form in the ventral region. Thus, at 11-5 days the site
of the first vibrissal rudiment is not as highly innervated as areas that would
develop vibrissae later. It is also important to note that at this stage the many
nerve trunks are splayed out dorsoventrally through the mesenchyme with no
clear segmentation or other patterning corresponding to the positions of the
ridges and grooves above. Therefore, at the time just prior to the development
of the earliest vibrissa, there appears no obvious spatial nor temporal one-to-one
correlation between the innervation pattern and the positions of the ridges, the
grooves, or the first whisker.
Figs 21 to 24. These illustrate the distribution of nerve trunks and branches relative
to the ridges and presumptive sites of vibrissal formation in an 11-5-day maxillary
process. Sections were taken perpendicularly to the length of the process (they are,
consequently, perpendicular to the plane of the previous figures) beginning at the
trigeminal ganglion and continuing anteriorly (see diagram inset, Fig. 21, for locations of these sections). The ventral edge (V) is to the left, and the dorsal (D) is to
the right.
Fig. 21. The first sign of branching (arrows) from the major trunks (n) occurs in
the dorsal region of the process at the level where the dorsal groove (dg) and ridge
(dr) are just beginning to appear. This region will eventually become horizontal row
3 of follicles but not until after vibrissae begin to develop more ventrally (V) where,
as yet, no sign of branching occurs. Bar = 100 /an.
Fig. 22. A higher power view of the region in the box in Fig. 21. The smaller nerve
branches (arrows) are more easily seen. Bar = 10/an.
Fig. 23. No branching of the ventral nerve trunks (in box) occurs until this level
of section where the ventral groove (vg) is already well established. The site where
the first vibrissa would have formed is in previous sections, at the posterior end of
the ventral groove where no ventral branching was seen. Branching continues more
dorsally (arrows) under presumptive row 3. Bar = 100/an.
Fig. 24. A higher power view of the region in the box in Fig. 23. A number of small
nerve branches (arrows) are seen in the area under the ventral groove. Bar = 10 jum.
152
J. T. WRENN AND N. K. WESSELLS
DISCUSSION
The variation in the rate of facial development in mouse embryos, even among
litter-mates, facilitates constructing what we interpret to be a precise sequence
of vibrissal rudiment formation. This study, in establishing the locations of the
first three or four rudiments, demonstrates that the common method of labelling
the sinus hairs from dorsal to ventral (Danforth, 1925; Davidson & Hardy, 1952;
Van der Loos & Woolsey, 1973, are examples) is developmentally inaccurate
and confusing in the description of early pattern formation. Fig. 1 presents an
alternative nomenclature which reflects the ventral-to-dorsal developmental pattern.
The early appearance of the ridges and the linearity of the horizontal rows
suggest that ridge formation might be an essential precondition for whisker
formation. In considering this possibility, one must remember that there are two
distinct facets of vibrissal development: (1) the generation of the overall whisker
pattern, and (2) the generation of a single vibrissa. It seems quite likely that the
ridges play a role in the first case, establishing the overall pattern, since the
positions of all five horizontal rows and the vertical row are closely correlated
with the positions of the four original ridges and the grooves. Indeed, the linearity of horizontal rows 2 through 5 arises directly from the partitioning of the
ridges. Ridge formation is a feature also found in mammary gland and tooth
development and may be a common mechanism for linear alignment of organs
as discussed by Van Exan & Hardy (1980).
In the second case, however, this study shows that it is not necessary for a
particular site to be on a ridge for that site to develop into a vibrissa. The vertical
row does not arise directly from a ridge nor do the first few vibrissae to form in
row 1. Therefore, while it may be true that ridge formation provides a general
plan for vibrissal pattern, it probably does not provide a specific and essential
morphogenetic signal to position a vibrissa in a specific spot.
Another conspicuous feature of facial development concurrent with early
vibrissal formation is the great amount of growth and remodelling of the maxillary and nasal processes. It seems likely that the forward growth of the two
processes provides new tissue to participate in vibrissal formation anterior to
rudiments which have already begun to form. To the authors' knowledge, no
studies on mice have yet been published, but Minkoff (1980) has shown that in
the chick embryo there are higher labelling indices at the tip of the maxillary
process and at the boundaries of the maxillary and nasal processes than in the
centres of the processes. If the same differential could be shown in the mouse
snout, it would help to explain the posterior-to-anterior gradient of development.
While ridge formation and growth of the snout may have roles in establishing
the general vibrissal pattern, neither could explain the precise temporal and
spatial ordering of the vibrissae in their vertical and horizontal rows. Certainly
Early development of mystacial vibrissae
153
a likely patterning agent would seem to be the distribution of nerves within the
snout. The results of Van Exan & Hardy (1980) support this idea. Those investigators examined follicle formation at stages we believe correspond to our late
12-day or early 13-day embryos, at times when many vibrissae had already begun
to form and few sites remained to begin development. Their reported correlation
between innervation pattern and vibrissal location appears to be similar to the
patterns seen in the 13-day bisected snouts stained with an acetylthiocholine
procedure to identify nerves. Indeed, nerve branches do seem to be intimately
associated with the positions of vibrissae at this stage. Furthermore, with so few
vibrissae left to form in each row it would be possible to visualize their future sites
and see their associated neural elements as Van Exan & Hardy have done.
The results reported here, however, show that at earlier times, as the arrangement of whiskers is just beginning to be established, innervation is already
widespread within the mesenchyme. This profusion of nerve trunks, of neurites
and growth cones within the maxillary process complicates any attempt to show
a one-to-one correlation between a nerve ending (termed a 'plexus' by Van Exan
& Hardy) and the later development of a vibrissa. In the first place, there appears
to be no immediate temporal correlation between many nerve branches at a
certain site and the beginning of a dermal condensation at that place. Secondly,
any precise spatial correlation would be difficult, if not impossible, to show
because of the interplay of two factors: (1) the close proximity of the early
rudiments (see Fig. 15), and (2) the prodigious growth of the maxillary process
as the rudiments are forming (from approximately 650 /im to 1200 jum long in the
12-day samples studied). With so little space between the rudiments (each
measures ca. 100/im in diameter) one would have to postulate a very precisely
located set of growth cones corresponding to a specific site of a dermal condensation with no growth cones between them. In fact, the nerve tracts in this study
do not appear so localized. Neurites do appear between the developing condensations. Furthermore, even if the growth cones were localized in the early 12day snout, one could not be confident that the sites above them would be future
centres of dermal condensation because of the unknown effects of growth within
the process on the spatial relationships between the elements of the process.
It appears, then, that at the time the follicular pattern is beginning to be
established, the overall pattern of innervation is too complex to account for a
simple one-to-one correlation. Furthermore, any influence the nerves may have
on vibrissal formation would have to be mediated through some action on the
mesenchyme. Nerve endings are never closer than approximately five mesenchymal cell diameters from the epithelium. As the dermal cells condense, the
growth cones remain on the outside, forming shallow, cup-like figures beneath
the condensations at the latest stages studied. Therefore, whatever influence
neurons may have on vibrissal development at this time must not be exerted
directly on the epithelium.
The results of this study suggest that early innervation of the snout may be a
154
J. T. WRENN AND N. K. WESSELLS
labile phenomenon. Most interesting is the large number of nerve trunks splayed
out in the mesenchyme of thell-5-day maxillary process as ridges are forming.
Van Exan & Hardy (1980) reported that five major branches of the maxillary
branch underlie the five ridges across the entire whisker pad in later embryos.
Three of these large branches would be expected under the area of the three
ridges derived from the maxillary process. The many trunks seen at 11-5 days in
perpendicular sections must be rearranged at a later time to form three major
branches. If this is so, the positioning of the ridges precedes the positioning of
the nerve trunks beneath them.
In addition, the nerve tracts themselves are in a process of development during
the time studied. Neurites in the trunks at 12 days look more 'mature' than those
at 11 days because most of the 12-day neurites have a more axon-like appearance. But even at 12 days growth cones are present in the main trunks and
become progressively more prevalent as the nerve tracts are traced toward the
periphery. The ultrastructure of the neurons is similar to that found during the
outgrowth of the interosseus nerve of the chick wing by Al-Ghaith & Lewis
(1982). These authors suggested that 'pioneer' growth cones were the first to
advance through the wing mesenchyme and that later growth cones followed
their tracts. Whether or not the most distal growth cones in the maxillary process
are true 'pioneers' in the same sense as those first suggested by Harrison (1910),
found by Bate (1976) in invertebrates, and studied by others (Keshishian, 1980,
and Edwards, Chen & Berns, 1981, are examples), it appears that vibrissae can
begin to form well before the time that the innervation of the maxillary process
is completely developed.
One further comment deserves mention. The high degree of spatial correspondence between the arrangements of the vibrissae on the snout and of their barrels
in the sensory cortex was mentioned earlier as being suggestive of the possibility
that the central nervous system may direct the patterning of vibrissae. But the
correspondence might equally suggest an influence acting in the opposite direction. Indeed, while the follicular pattern is established at 12 days of embryonic
life, the barrel pattern is not morphologically apparent until about 5 days after
birth (Rice & Van der Loos, 1977), about the same time that barrel development
can no longer be altered by lesions of the vibrissae (Van der Loos & Woolsey,
1973; Weller.& Johnson, 1975). Weller & Johnson concluded that 'the periphery
can at least be strongly suspected of exerting an organizational influence on the
cortex'. Similarly, after a study of supernumerary whiskers and barrels in mice,
Van der Loos & Dorfl (1978) concluded that the vibrissal pattern is responsible
for the patterning of the cortical barrels.
This study makes no attempt to investigate the hypothesis of peripheral-tocentral patterning. Rather, by focusing on early innervation and the lack of clear
spatial or temporal correlations between it and initial vibrissal formation, the
study suggests that patterning is not unidirectional but is a complex process
requiring mutual interactions between the cells and tissues involved.
Early development of mystacial vibrissae
155
The authors wish to thank Heather Hindman for her patience and careful work in drawing
the reconstructions. This project was supported by NIH grant No. HD 04708.
REFERENCES
L. K. & LEWIS, J. H. (1982). Pioneer growth cones in virgin mesenchyme: an
electron-microscope study in the developing chick wing. J. Embryol. exp. Morph. 68,
149-160.
BARALD, K. F. & BERG, D. W. (1979). Ciliary ganglion neurons in cell culture: High affinity
choline uptake and autoradiographic choline labeling. Devi Biol. 72, 15-23.
BATE, C. M. (1976). Pioneer neurons in an insect embryo. Nature 26, 54-56.
DANFORTH, C. H. (1925). Hair in its relation to questions of homology and phylogeny. Am.
J. Anat. 36, 47-68.
DAVIDSON, P. & HARDY, M. H. (1952). The development of mouse vibrissae in vivo and in
vitro. J. Anat. 86, 342-356.
DUN, R. B. (1958). Growth of the mouse coat. VI. Distribution and number of vibrissae in
the house mouse. Aust. J. biol. Sci. 2, 95-105.
EDWARDS, J. S., CHEN, S. & BERNS, M. W. (1981). Cereal sensory development following
laser microlesions of embryonic apical cells in Acheta domesticus. J. Neurosci. 1, 250-258.
GRUNEBERG, H. (1943). The development of some external features in mouse embryos. J.
Hered. 34, 89-92.
HARDY, M. H. (1951). The development of pelage hairs and vibrissae from skin in tissue
culture. Ann. N.Y. Acad. Sci. 53, 546-561.
HARRISON, R. G. (1910). The outgrowth of the nerve fiber as a mode of protoplasmic movement. /. exp. Zool. 9, 787-845.
KARNOVSKY, M. J. & ROOTS, L. (1964). A "direct-coloring" thiocholine method for cholinesterases. /. Histochem. Cytochem. 12, 219-221.
KESHISHIAN, H. (1980). The origin and morphogenesis of pioneer neurons in the grasshopper
metathoracic leg. Devi Biol. 80, 388-397.
MINKOFF, R. (1980). Regional variation of cell proliferation within the facial processes of the
Chick embryo: a Study Of the role of "merging" during development. /. Embryol. exp.
Morph. 57, 37-49.
RICE, F. L. & VAN DER LOOS, H. (1977). Development of the barrels and barrel fields in the
somatosensory cortex of the mouse. J. comp. Neurol. 171, 545-560.
RICHARDSON, K., JARETT, L. & FINKE, E. (1960). Embedding in epoxy resins for ultra-thin
sectioning in electron microscopy. Stain Technol. 35, 313-323.
TELLO, J. F. (1923). Genese des terminaisons motrices et sensitives. II. Terminaisons dans les
poils de la souris blanche. Traveaux du Laboratoire de recherches biologiques de I'University
de Madrid 21, 257-384.
VAN DER LOOS, H. & DORFL, J. (1978). Does the skin tell the somatosensory cortex how to
construct a map of the periphery? Neurosci. Lett. 7, 23-30.
VAN DER LOOS, H. & WOOLSEY, T. A. (1973). Somatosensory cortex: Structural alterations
following early injury to sense organs. Science 179,-395-398.
VAN EXAN, R. J. & HARDY, M. H. (1980). A spatial relationship between innervation and the
early differentiation of vibrissa follicles in the embryonic mouse. /. Anat. 131, 643-656.
VINCENT, S. B. (1913). The tactile hair of the white rat. /. comp. Neurol. 23, 1-38.
WELLER, W. L. & JOHNSON, J. I. (1975). Barrels in cerebral cortex altered by receptor disruption in newborn, but not in five-day-old mice (Cricetidae and Muridae). Brain Res. 83,
504-508.
WESSELLS, N. K. & ROESSNER, K. D. (1965). Nonproliferation in dermal condensations of
mouse vibrissae and pelage hairs. Devi Biol. 12, 419-433.
WINKELMANN, R. K. (1959). The innervation of a hair follicle. Ann. N.Y. Acad. Sci. 83,
401-407.
WOOLSEY, T. A. & VAN DER LOOS, H. (1970). The structural organization of layer IV in the
somatosensory region (SI) of mouse cerebral cortex. Brain Res. 17, 205-242.
AL-GHAITH,
156
J. T. WRENN AND N . K. WESSELLS
M. & YOHRO, T. (1979). Subdivision of mouse vibrissae on an embryological
basis, with descriptions of variations in the number and arrangement of sinus hairs and
cortical barrels in Balb/c (nu/+; nude, nu/nu) and hairless (hr/hr) strains. Am. J. Anat.
155, 153-174.
YAMAKADO,
{Accepted 15 May 1984)