Hair follicles are essentially composed of two tissues. The inner epidermal component, which gives rise to, among other products, the keratinized hair shaft, is confluent with the surface epidermis and is ensheathed by the dermal component which is confluent with the pars papillaris of the dermis.

A specialization of the dermal component is the dermal papilla which, in follicles producing hair, is enclosed by the epidermal matrix of the hair bulb and is connected to the dermal sheath by the papilla stalk.

Many authorities have considered that the dermal papilla is an essential component of the hair follicle (reviews: Cohen, 1965; Oliver, 1969). It has been suggested that the dermal papilla may be involved in both the induction of follicle lengthening and hair growth during the proanagen phase (Chase, 1965) of the hair cycle, a concept now justified by direct experimentation in the vibrissa follicle at least (Oliver, 1967 b), and perhaps also in determining the nature of the hair produced by a follicle.

It also seems likely that the developing dermal papilla is an important, if not the primary, source of induction in follicle development during ontogeny (Jacobson, 1966; Kollar, 1966, 1968). Indeed various authorities have considered that the dermal papilla is capable of inducing follicle formation in the adult after wounding of the skin (Billingham, 1958; Billingham & Silvers, 1963).

In order to investigate these possibilities directly Cohen (1961) devised techniques for isolating dermal papillae from excised bulbs from the comparatively large rodent vibrissa follicles, a possibility suggested by Billingham (1958).

Cohen (1961, 1964, 1965) has reported the results of implanting vibrissa dermal papillae into ear dermis in the adult hooded rat. He has presented persuasive evidence that where the vibrissa papillae contacted ear epidermis, whether of follicular or non-follicular origin, they induced by 14 days and growth of ear-type hairs. These papillae underwent a drastic shrinkage in size to ear follicle papilla proportions, involving a reduction in cell number from hundreds down to tens of cells. However, where papillae did not contact epidermis they remained the same size.

More recently Oliver (1968, 1969) has published preliminary accounts of results of the implantation of vibrissa dermal papillae recombined with epidermis from three different regions of the body into ear dermis. This method was used, in preference to implantation of papillae on their own, in the hope that it would more effectively ensure papilla/epidermal contact and that the influence of papillae on non-follicular epidermis could be evaluated. This report gives full details of the above and further work.

All operations were performed on 2-to 4-month-old animals from an inbred strain of hooded rat. Anaesthesia was induced by the intraperitoneal injection of a 10 × diluted solution of Nembutal (Abbott), 0·65 ml/100 g body weight.

Operative sequence

At each operation thin slices of skin were removed freehand with a scalpel from one of the following sites after bathing the donor area with 70% alcohol: (i) the dorsum of the right ear. This provides ear epidermis which contains follicular epidermis, including sebaceous glands (Fig. 1A); (ii) the anterior region of the stretched scrotum. This provides follicle-free scrotal sac epidermis (Fig. 1B); (iii) the inner medial surface of the lower lip as far as the base of the incisors. This provides keratinizing oral epithelium which is, of course, afollicular (Fig. 1C).

Figure 1

All sections stained with Weigert’s haematoxylin, alcian blue and Curtis’s Ponceau S.

(A) Section of ear epidermis separated from its dermis with trypsin. Note presence of follicular epidermis, including sebaceous gland cells, × 190.

(B) Section of afollicular scrotal sac epidermis separated from its dermis with trypsin, × 160.

(C) Section of oral mucosa epithelium separated from its dermis with trypsin. Note the compact cuticle and slight maceration of the basal cells some of which are detached from the epithelial sheet, × 160.

(D) Section of a vibrissa dermal papilla (arrow) in ear dermis 33 days after implantation of dermal papillae alone. The papilla surface is entirely free of epidermal cells. × 110.

(E-G) Sections of vibrissa dermal papilla/epidermis associations in ear at various times after implantation.

(E) Implant of ear epidermis becoming locally organized around a single vibrissa dermal papilla. 14 days, × 110.

(F) Short, stout hair follicle induced from ear epidermis by a single vibrissa dermal papilla. Note the thick inner root sheath, fine hair shaft and papilla with many cells. 35 days. × 110.

(G) Base of a follicular structure developed from scrotal sac epidermis and incorporating three vibrissa dermal papillae which have fused to form a central ‘pulp’. Note the scalloped inner aspect of the matrix and the lack of recognizable hair follicle epidermal elements. 21 days, ×80.

Figure 1

All sections stained with Weigert’s haematoxylin, alcian blue and Curtis’s Ponceau S.

(A) Section of ear epidermis separated from its dermis with trypsin. Note presence of follicular epidermis, including sebaceous gland cells, × 190.

(B) Section of afollicular scrotal sac epidermis separated from its dermis with trypsin, × 160.

(C) Section of oral mucosa epithelium separated from its dermis with trypsin. Note the compact cuticle and slight maceration of the basal cells some of which are detached from the epithelial sheet, × 160.

(D) Section of a vibrissa dermal papilla (arrow) in ear dermis 33 days after implantation of dermal papillae alone. The papilla surface is entirely free of epidermal cells. × 110.

(E-G) Sections of vibrissa dermal papilla/epidermis associations in ear at various times after implantation.

(E) Implant of ear epidermis becoming locally organized around a single vibrissa dermal papilla. 14 days, × 110.

(F) Short, stout hair follicle induced from ear epidermis by a single vibrissa dermal papilla. Note the thick inner root sheath, fine hair shaft and papilla with many cells. 35 days. × 110.

(G) Base of a follicular structure developed from scrotal sac epidermis and incorporating three vibrissa dermal papillae which have fused to form a central ‘pulp’. Note the scalloped inner aspect of the matrix and the lack of recognizable hair follicle epidermal elements. 21 days, ×80.

The skin slices were incubated in 0·25% trypsin (Nutritional Biochemicals Corporation 1:300) for 20 × 35 min at 37 °C. Some slices were, in fact, incubated for slightly longer periods which may well have affected the viability of the epidermis, as is discussed below.

While the skin slices were incubating, the vibrissa follicles on the left upper lip were exposed, the bulbar region of follicles removed and placed into Hanks’s solution and the lip stitched back into position as described previously (Cohen, 1961; Oliver, 1966). Dermal papillae were dissected from the bulbs and put into fresh Hanks’s solution, as described by Oliver (1967b). From each rat 1 – 6 dermal papillae were isolated which were considered to be free of contaminating epidermal matrix cells. At this stage the epidermis was cleanly separated from the skin slices in Hanks’s solution using watchmaker’s forceps and cut into rectangular sheets ca. 1 – 3 mm square.

Preparation of host site and implantation procedure

Using a pointed scalpel an L-shaped flap of skin was made then reflected on the dorsum of the left ear. The horizontal plane of the incision was made in the dermis, parallel to and between the skin surface and the ear cartilage.

In some of the earlier work the dermal papillae were wrapped in the epidermal sheets prior to their implantation under the ear flap. However, in the majority of implants the skin flap was reflected, the dermal papillae placed on the exposed dermal bed and the appropriate epidermal sheet gently lowered on to the papillae, basal surface down.

The skin flap was then carefully pulled back into position over the implants and secured either with a single suture through the apex of the flap or with Nobecutane (Duncan, Flockhart and Evans Ltd.).

Design of experiments

Details of animals used, number of papillae implanted and times of biopsy after operation for each of the three types of recombination made are shown in Tables 13. A total of 88 papillae were implanted into 25 rats: 28 papillae in association with ear epidermis, 36 with scrotal sac epidermis and 19 with oral epithelium. Five dermal papillae were implanted alone and, in addition, a limited number of trypsinized but unseparated skin slices and sheets of epidermis alone were also implanted under flaps in ear. All but three of the grafts were autografts.

Table 1

Vibrissa dermal papilla/ear epidermis recombinants in ear skin

Vibrissa dermal papilla/ear epidermis recombinants in ear skin
Vibrissa dermal papilla/ear epidermis recombinants in ear skin

Histology

Implant areas were fixed in buffered formalin or formolsaline fixatives. Serial sections of the entire implant areas were cut at 8 μ and stained in a combination of Weigert’s haematoxylin, alcian blue and Curtis’s Ponceau S.

While the experimental technique employed was essentially very simple, with the object of ensuring papilla/epidermal contact within ear dermis, many of the results cannot be interpreted in terms of this persisting anatomical relationship. Two aspects of the results, summarized in Tables 1, 2 and 3, will therefore be considered before presenting those results in which papillae were found in various associations with epidermis.

Table 2

Vibrissa dermal papilla/scrotal sac epidermis recombinants in ear skin

Vibrissa dermal papilla/scrotal sac epidermis recombinants in ear skin
Vibrissa dermal papilla/scrotal sac epidermis recombinants in ear skin
Table 3

Vibrissa dermalpapilla/oral mucosa epithelium recombinants in ear skin

Vibrissa dermalpapilla/oral mucosa epithelium recombinants in ear skin
Vibrissa dermalpapilla/oral mucosa epithelium recombinants in ear skin

(i) Number of dermal papillae recovered

Of a total of 88 dermal papillae implanted into ear skin, 76 were found with certainty and a possible further three may have been present in the biopsy material. This loss of papillae was not obviously related to duration of implantation since three papillae could not be accounted for in ears biopsied at 7 days (rats I, IX and X), while papillae were often fully accounted for at 35 and 42 days (e.g. rats XVI-XV1I1).

(ii) Isolated dermal papillae

Of 83 papillae implanted in association with epidermis, 37, perhaps 39, were found isolated in ear dermis, as were the five papillae implanted alone into ear. Sometimes all of the implanted papillae in particular animals were found isolated but in others some of the papillae were isolated while the remainder were either unaccounted for or had formed associations with epidermis. Again isolation of papillae was not related to duration of implantation. All of these isolated papillae were found to be entirely free of adherent epidermal cells (e.g. Fig. 1D). In two rats fusion of papillae had occurred. No obvious reduction in cell number of the papillae was apparent, although some were compacted and showed a loss of cellular cytoplasmic volume. One possible degenerated papilla was found (rat Xlll) but the rest appeared healthy and had acquired a capillary supply and most had taken up the alcian blue stain in varying degrees of intensity. Interestingly three papillae were readily identified at 190 and 207 days after implantation (rats XIX, VIII). These papillae had not stained with alcian blue and collagen fibres were present between the cells.

(iii) Dermal papilla/epidermis associations

A total of 21, possibly 22, papilla/epidermis associations, each containing from one to four papillae, were found (Tables 13).

The distinction has been made between those instances in which papillae had formed associations with epidermis of questionable origin at the ear surface and those which had formed intradermal associations with the epidermis with which they had been implanted.

After implantation with ear epidermis (Table 1)

Both papillae recovered at 14 days were in intradermal contact with the epidermis with which they had been implanted (rat II). However, one of these papillae and its epidermis had degenerated. The other papilla was enclosed by epidermis which was considerably thickened and contained suprabasal cell divisions (Fig. IE). The inner aspect of this ‘matrix’ was irregular. Confluent epidermis lateral to the matrix region was either largely degenerated or completely keratinized.

At 21 days two fused papillae were found in contact with the superficial ear epidermis (rat 111). The epidermis was locally organized as a matrix around the papillae and had an irregular inner aspect.

All four papillae in rat IV, at 28 days, had become fused and had organized epidermis into a large ‘follicular’ structure around them. This structure, which superficially resembled a developing feather follicle, was confluent with the superficial epidermis but projected into the ear. Papillary cells extended throughout the length of this elongated, cylindrical structure and again the inner aspect of the matrix was irregular. The pattern of keratinization was similar to that of superficial epidermis but considerably thicker keratin was being produced. The base of this structure was of greatest interest since it showed, in greatly enlarged form, some of those epidermal elements typical of a hair follicle. Peripherally, vacuolated cells were recognized as outer root sheath bounding an inner root sheath, the latter containing trichohyalin granules which are characteristically found only in hair follicles. Nevertheless, hair keratin was not being produced from the matrix cells, central to the inner root sheath, nor were there any convincing indications of sebaceous gland histogenesis.

At 35 days, in rat V, two follicles producing hair were found, in each of which a single vibrissa dermal papilla had been obviously incorporated. Normal anagen hair bulbs in the mid-dorsal region of the ear, where the implants were made, contain papillae with between ca. 35 and 55 nuclei. Each of the above two follicles had very large bulbs and papillae containing hundreds of nuclei but not, apparently, a vascular supply. One of these follicles was cut transversely and the other vertically (Fig. 1F). This latter follicle was much shorter than the longest ear hair follicles but both were stouter than ear hair follicles. Notwithstanding their comparatively large bulbs they were producing fine non-medullated hairs. Obviously many of the matrix division products were becoming inner root sheath rather than hair shaft since there were six cell layers to the inner root sheath, four contributing to the Huxley layer, in these follicles. Ear hairs generally have three cell layers comprising the inner root sheath and the much larger vibrissa follicles have four layers. Interestingly, the inner root sheath in these follicles extended up to the skin surface rather than disappearing below the level of the sebaceous glands as in normal follicles. Sebaceous glands had not fully developed; sebaceous cells were differentiating within small outgrowths from the neck of each follicle but sebaceous ducts were not present.

At 42 days (rat VI) a follicle with a very large bulb was found but unfortunately the sections were folded in this region and while it seemed likely that it contained a vibrissa papilla it was impossible to determine this with certainty.

With scrotal sac epidermis (Table 2)

A single papilla was found in contact with superficial epidermis at both 7 and 14 days (rats IX, XI). The local epidermis had formed a thickened matrix in both of these associations. Also at 14 days (rat XII) three papillae were intradermally associated with the same implant of scrotal sac epidermis. The papillae were discrete and each had organized a separate matrix.

Rat XIV, at 21 days, showed a structure, confluent with the surface epidermis and containing three fused papillae as an inner dermal ‘pulp’, very similar to that described for rat III. However, there were no indications of hair follicle epidermal elements (Fig. 1G).

At 28 days (rat XV) three separate papilla/scrotal sac epidermis associations were found within the ear dermis, two containing single papillae and one two papillae. A matrix was organized around one of the single dermal papillae. In the other two associations papillary cells were present over the greater surface of the epidermis, which was keratinizing into an internal lumen, with apparent attempts at matrix organization around the focal areas representing the main papillary masses (Fig. 2A). There was no development of recognizable hair epidermal elements in any of these associations.

Figure 2

All sections stained with Weigert’s haematoxylin, alcian blue and Curtis’s Ponceau S.

(A – F) Sections of vibrissa dermal papilla/epidermis associations in ear dermis at various times after implantation.

(A) Implant of scrotal sac epidermis keratinizing into a central lumen with localized hyperplasia and attempts at matrix formation, with columnar basal cells, in association with vibrissa papillary cells (left and especially right). 28 days, × 80.

(B) Huge, irregular hair bulb which has developed from scrotal sac epidermis and three vibrissa dermal papillae, showing vacuolated outer root sheath cells (arrow) and epidermal matrix and its differentiation products around the fused papillary tissue. Note that scrotal sac epidermis unorganized by the papillae has formed a thinwalled cyst at left. 35 days, × 80.

(C) Oblique section through a large hair bulb which has developed from scrotal sac epidermis and two vibrissa dermal papillae. Note the vacuolated outer root sheath (top arrow), inner root sheath (middle arrow) and keratinizing hair (lower arrow). 35 days, × 80.

(D) Implant of oral mucosa epithelium which has thickened and is keratinizing into a central lumen. Note localized attempts at matrix formation, with columnar basal cells, in relation to vibrissa papillary tissue (at bottom and especially top). Suprabasal cell morphology and compact cuticle resemble normal mucosa epithelium. 24 days, × 80.

(E) Implant of oral mucosa epithelium surrounded by vibrissa papillary tissue. The columnar basal cells are separated, at right and left, from an epithelial stellate reticulum by a distinct intermediate cell layer. Note the two nidi of parakeratotic cells confluent with the inner keratinized layer. 35 days, × 80.

(F) Implant of oral mucosa epithelium which has been organized by vibrissa papillary tissue at the top and has formed a confluent thin-walled epithelial cyst below. At top are seen the typical hair follicle epidermal elements, outer root sheath (left arrow) and inner root sheath (middle arrow), which enclose not developing hair shaft, but a parakeratotic layer surrounding an inner epithelial stellate reticulum (right arrow). The stellate reticulum contacts papillary tissue via a columnar basal cell layer in other sections. 35 days, × 80.

Figure 2

All sections stained with Weigert’s haematoxylin, alcian blue and Curtis’s Ponceau S.

(A – F) Sections of vibrissa dermal papilla/epidermis associations in ear dermis at various times after implantation.

(A) Implant of scrotal sac epidermis keratinizing into a central lumen with localized hyperplasia and attempts at matrix formation, with columnar basal cells, in association with vibrissa papillary cells (left and especially right). 28 days, × 80.

(B) Huge, irregular hair bulb which has developed from scrotal sac epidermis and three vibrissa dermal papillae, showing vacuolated outer root sheath cells (arrow) and epidermal matrix and its differentiation products around the fused papillary tissue. Note that scrotal sac epidermis unorganized by the papillae has formed a thinwalled cyst at left. 35 days, × 80.

(C) Oblique section through a large hair bulb which has developed from scrotal sac epidermis and two vibrissa dermal papillae. Note the vacuolated outer root sheath (top arrow), inner root sheath (middle arrow) and keratinizing hair (lower arrow). 35 days, × 80.

(D) Implant of oral mucosa epithelium which has thickened and is keratinizing into a central lumen. Note localized attempts at matrix formation, with columnar basal cells, in relation to vibrissa papillary tissue (at bottom and especially top). Suprabasal cell morphology and compact cuticle resemble normal mucosa epithelium. 24 days, × 80.

(E) Implant of oral mucosa epithelium surrounded by vibrissa papillary tissue. The columnar basal cells are separated, at right and left, from an epithelial stellate reticulum by a distinct intermediate cell layer. Note the two nidi of parakeratotic cells confluent with the inner keratinized layer. 35 days, × 80.

(F) Implant of oral mucosa epithelium which has been organized by vibrissa papillary tissue at the top and has formed a confluent thin-walled epithelial cyst below. At top are seen the typical hair follicle epidermal elements, outer root sheath (left arrow) and inner root sheath (middle arrow), which enclose not developing hair shaft, but a parakeratotic layer surrounding an inner epithelial stellate reticulum (right arrow). The stellate reticulum contacts papillary tissue via a columnar basal cell layer in other sections. 35 days, × 80.

However, at 35 days (rat XVI), two intradermal structures were found which were instantly recognizable as extraordinarily large although not perfectly organized hair bulbs (Fig. 2B, C). One of these follicles contained two and the largest three dermal papillae. In each follicle a large amount of epidermal matrix was present around the papillae and giving rise to inner root sheath, hair cuticle and hair keratin bounded by a layer of outer root sheath cells. The hair keratin did not display the regular organization presented by naturally occuring hair follicles. Neither of these two follicles showed the development of sebaceous glands.

With oral epithelium (Table 3)

At 24 days (rat XX) two papillae were in intradermal association with oral epithelium. As seen in rat XV the papillary cells enclosed most of the surface of the thickened epithelium which was keratinizing into a central lumen. The outer surface of the epithelium was locally lobulated and arranged as matrices with a columnar basal layer around the two main papillary masses (Fig. 2D).

At 35 days (rat XXII) a follicle opening at the ear surface was seen which had obviously incorporated a single vibrissa dermal papilla. In all essential respects it was identical to the induced follicles described in rat V, and shown in Fig. 1 F.

Also at 35 days (rat XXI11) were found two intradermal papilla/epithelial associations, each probably containing two papillae. In one of these the papillary tissue almost completely enveloped the considerably thickened epithelial mass which, over its greater part, had a highly columnar outer basal layer. In regions immediately central to the columnar layer, was a layer of cells arranged parallel to the prominent basement membrane and central to this layer epithelial cells were present as a stellate reticulum. Two foci of parakeratotic cells were also present and confluent with a layer of keratinized cells bounding a keratin-filled central lumen (Fig. 2E). Near to this implant were two separate areas of intramembranous ossification.

In the second and larger of these two associations was a more familiar matrix region, irregularly enveloping papillary tissue and giving rise to inner root sheath which was bounded by a layer of typically vacuolated outer root sheath cells. Intriguingly, the central core of this structure, capping the papillary tissue, was present as an epithelial stellate reticulum which was separated from the inner root sheath by a layer of parakeratotic cells (Fig. 2F). Hair shaft could not be identified and there were no indications of sebaceous gland histogenesis.

(iv) Subsidiary experiments

Single implants of trypsinized but unseparated ear skin, scrotal sac skin and oral mucosa were recovered at 42, 35 and 44 days respectively. The ear skin had encysted but otherwise had a normal epidermis and follicles in catagen and telogen. The scrotal sac implant, which had contacted the surface epidermis, had an apparently normal non-pigmented epidermis without follicles, but had a smoother than normal dermal/epidermal interface. The oral mucosa implant had encysted, forming a distended sac packed with keratin. The thin, flattened epithelium was non-pigmented and afollicular.

Implants were also made of ear epidermis alone (recovered at 42 days), scrotal sac epidermis (recovered at 28, 35 and 42 days) and oral epithelium (recovered at 24 days). Most of these presented a picture of disorganized intradermal cell masses, with no organization of a basal cell layer, although attempts at keratinization were seen in one of the scrotal sac epidermis implants. Occasionally tongues of viable epidermal cells, confluent with the surface epidermis, were seen protruding into the dermis but it could not be ascertained whether they derived from native ear epidermis or from implanted epidermis.

Generally speaking the epidermis implanted with papillae, and not locally organized by the papillae, also showed degenerative changes. However, cysts, keratinizing internally and containing neither follicles nor sebaceous glands, were seen formed from ear epidermis (rat VII), scrotal sac epidermis (rat XVI; Fig. 2B) and oral mucosa epithelium (rat XXIII; Fig. 2F).

(v) Other comments

No special study was made of the stage of the growth cycle of the ear hair follicles at the time of operation. However, it is probable that most, if not all, of the implant regions contained follicles in the resting phase since anagen follicles are rarely seen in undisturbed ear skin taken from rats of the same age as those used for operations. Apparently the implantation procedure coincided with or, much more likely (Argyris &Trimble, 1964), precipitated a new hair growth cycle in the implant areas for, at 7 days, follicles in the biopsied regions were in proanagen and producing hair by 14 days. Local ear hair follicles were at their maximum length at 21 and 28 days but most had entered catagen by 35 days and, with some exceptions, were in telogen by 42 days.

Unfortunately little information was obtained of the wound healing process which follows the implantation procedure used. Sections through the scab present in rat IX showed that much of the original ear skin flap had been undergrown by epidermis from the wound edges and was being exfoliated. A similar process may have occurred in other animals, especially where the epidermis covering the implant area was devoid of hair follicles, and this is the only reasonable explanation for the presence of dermal papillae immediately under the surface epidermis.

Rigorous examination of all sections showed no indications of neoformation of hair follicles except, of course, those follicles which had developed in association with the implanted vibrissa dermal papillae.

Fulfilment of experimental requirements

A proper investigation of the influence of vibrissa dermal papillae on epidermis essentially requires that viable papillae, known to be entirely free of contaminating epidermal cells, are brought into uninterrupted contact with viable epidermis in a site conducive for the growth of hair.

Ear dermis has been shown to support the growth and regeneration of vibrissae, albeit of a stunted nature, from grafts of the bulbar and follicle wall regions of vibrissa follicles (Cohen, 1961; Oliver, 1967 a).

The majority of the papilla implants were obviously viable since only one, possibly two, out of 76, perhaps 79, recovered 7 – 207 days after implantation showed degenerative changes.

There can also be little doubt that most, if not all, of the vibrissa dermal papillae were uncontaminated with vibrissa epidermal cells. Oliver (1967 b) has already observed that serial sections of vibrissa dermal papillae, dissected from vibrissa bulbs by the same method used in this work, show the papillae to be free of epidermal cells. Corroborative evidence is also convincingly available from the present results. Of a total of 88 papillae implanted into ear skin, including the five papillae implanted alone, 42, perhaps 44, were found isolated in the dermis and all were entirely free of adherent epidermal cells. It would appear unlikely that the 34 or 35 papillae found in association with epidermis, often obviously graft epidermis, also happened to be papillae contaminated with epidermal cells of vibrissal origin.

The fact that only 41% of the papillae were found in association with epidermis raises another requirement necessary to validate the experiment: uninterrupted papilla contact with viable epidermis. It is highly possible that in some instances papillae became displaced from under their epidermal sheets on repositioning of the skin flap at operation. An alternative and perhaps additive explanation is that the viability of some of the epidermal sheets may have been impaired by too long exposure to the effects of trypsin. Billingham & Silvers (1967) emphasize that trypsinization at 37 °C may result in variable degrees of epidermal maceration, and thus the loss of germinal basal cells, as was occasionally seen in the present work. Such an effect may explain the degenerated appearance of many of the intradermally sited implants of epidermis, whether implanted alone or with papillae. This explanation gains credence from those examples of epidermis which did manage to maintain some organization. Presumably here epidermal viability was less impaired by the trypsin although to what extent epidermal organization is dependent on a suitable substrate for the basal layer cannot here be assessed (cf. Van Scott & Flaxman, 1968).

Twelve, but possibly only nine papillae (10 – 14%), were not identified in the biopsy material which may be accounted for by one or more of the following reasons: non-inclusion in the biopsy area (although great care was taken to avoid this eventuality), exfoliation from the skin with the skin flaps by epidermal undergrowth during the wound healing process, degeneration and resorption of papillae or incorporation of papillae into new ear-type follicles indistinguishable from native follicles. The fact that three papillae were not identified at 7 days strongly suggests that these at least were lost in the scab. Cohen (1964) could only account for 9 out of 57 papillae implanted alone into ear and thought that while some at least were exfoliated, others probably became incorporated into new ear-type follicles.

Influence of vibrissa dermal papillae on epidermis

It has already been demonstrated that vibrissa dermal papillae, implanted into the bases of the superficial halves of vibrissa follicles, induce lengthening of the follicle and the growth of whiskers (Oliver, 1967b). In the present studies none of the implants of epidermal sheets alone showed any indication of follicle organization, rather most showed degenerative changes. This emphasizes both the profound influence of vibrissa papillae on epidermis and the pluripotential characteristics of epidermis since not only can vibrissa papillae induce follicle formation from ear epidermis but also from afollicular scrotal sac epidermis and from oral epithelium.

A variety of papilla/epidermis associations were seen, some containing single and others fused papillae. The development of follicles seemed to involve the following incomplete sequence of events. The epidermis enveloped one or more papillae to become locally organized into a matrix containing suprabasal mitoses (7 – 14 days). The development of follicular structures, without elements characteristic of hair follicles, then occurred (21 days), and by 28 days the development of an outer root sheath separated from the matrix by an inner root sheath was observed, but without the growth of hair. However, at 35 days, examples were found of follicles producing either hair, or at least structures resembling it, some with developing but non-functioning sebaceous glands. Obviously time is an important factor in these events and in a previous publication (Oliver, 1968) I erroneously suggested that vibrissa papillae were unable to induce follicle formation since there were no indications that this was occurring up to 3 weeks after implantation.

The papillae in those follicles which incorporated single vibrissa papillae, although obviously much larger than local ear follicle papillae, were smaller than when implanted. This may have been due to the fact that they now did not contain capillaries; the papillae may also have contributed to the new dermal sheaths as seems to occur after implantation into vibrissa follicles.

These follicles also differed from ear follicles in that they were shorter and stouter with comparatively large bulbs, yet producing fine, non-medullated hairs. The first hairs produced during ontogenetic development may also be non-medullated (Davidson &Hardy, 1952; Montagna, 1962). The presence of the inner root sheath extending as far as the skin surface, associated with the presence of non-functional sebaceous glands, would seem to relate to the suggestion that sebum may contain enzymes which normally effect the dissolution or fragmentation of the inner root sheath at this level (Straile,1965).

Comparatively huge intradermal follicles had developed from scrotal sac epidermis and oral epithelium where they had incorporated two or more papillae.

In other associations papillary tissue was seen to invest, rather than be invested by, epidermis which had encysted. Even in these instances the papillae promoted epidermal hyperplasia. The development of an epithelial stellate reticulum adjacent to a highly columnar basal layer in two of the oral epithelia implants associated with papillary tissue is most intriguing, since this configuration is presented by the enamel organ of developing teeth. It would appear that here papillary cells are promoting the expression of a potential in the mucosal epithelium to form the epithelial contribution which is required for the development of teeth.

It is worth noting that isolated dermal papillae can persist for up to 207 days in ear dermis; obviously their integrity is not dependent on contact with epidermal cells.

Tract specificity

So far the results have only been considered in terms of the dermal papilla and epidermis. However, it has been demonstrated that many characteristics of the superficial epidermis in adult mammals, but not necessarily mucosal epithelia, are determined and maintained by specific and persistent stimuli from the underlying dermis (Billingham &Silvers, 1965, 1967, 1968).

Cohen (1965, 1969 a) has extended such findings to account for the determination of tract specificity in skin appendages (reviews: Cohen, 1969b,Oliver, 1969). As mentioned previously Cohen considers that those vibrissa dermal papillae implanted into ear skin which contact ear epidermis induce the formation of ear-type follicles by 14 days. He visualizes the induction of follicle formation by the papillae and the determination of follicle type by the local dermis. The present results showed no evidence of the development of ear-type hairs nor of the shrinkage of vibrissa papillae to ear hair proportions where they contacted epidermis. The characteristics of the follicles formed did not readily implicate the local dermis or the epidermis which invested the papillae as determinants of follicle specificity. Rather the size of the matrix of these follicles, but not necessarily the size of the hair shaft produced, seemed to be related to the amount of papillary tissue present. It is conceivable that in Cohen’s work cellular viability in the papillae was in some way reduced, before or at implantation, so that follicles were formed with small papillae. However, it may be relevant that at 14 days, when Cohen considered such follicles to be forming, the local ear hair follicles, sometimes disorientated by the implantation procedure, are in various stages of early anagen when they could perhaps be interpreted as newly forming follicles.

Papilla/epidermis, and particularly oral epithelium, recombinants will need to be left in ear dermis for much longer periods than reported here. Induced follicles might then assume recognizable tract characteristics. It would also be of great interest to discover whether the formation of an epithelial stellate reticulum in oral epithelium/papilla recombinants is a reproducible phenomenon and if so whether it is a terminal or an intermediary product. The behaviour of recombinants could also be evaluated in non-skin environments.

Refinement of present techniques could lead to more closely controlled follicle formation in the adult. This would permit a more detailed causal analysis of the factors involved in follicle formation and such problems as the determination of tract specificity than has been previously possible.

  1. The influence of vibrissa dermal papillae on epidermis of non-vibrissal origin has been investigated in the adult hooded rat.

  2. A method is described for preparing and implanting recombinants of vibrissa dermal papillae and ear epidermis, afollicular scrotal sac epidermis and oral epithelium into ear dermis.

  3. A variety of dermal papilla/epidermis associations were found demonstrating a profound localized influence by dermal papillae on epidermis, including formation of new hair follicles by 35 days.

  4. It was thus conclusively shown that vibrissa dermal papillae can induce the formation of hair follicles and hair growth, even from previously afollicular epidermis.

  5. Possible factors involved in determining the characteristics of these follicles are discussed.

  6. Dermal papillae isolated in ear dermis maintained their integrity for up to 207 days after implantation.

  7. Epidermis from ear and scrotal sac skin and oral epithelium implanted alone into ear generally appeared disorganized or degenerated, with no regeneration of hair follicles.

Induction de la formation de follicules pileux chez le rat mantelé adulte par les papilles dermiques des vibrisses

  1. L’influence des papilles dermiques des vibrisses sur de l’épiderme d’autre origine a été étudiée chez le rat mantelé adulte.

  2. On décrit une méthode de préparation et d’implantation, dans le derme de l’oreille, de réassociations de papilles dermiques de vibrisses avec de l’épiderme de l’oreille, ou de l’épiderme afolliculaire du sac scrotal, ou encore de l’épithélium buccal.

  3. On constate que diverses associations papilles dermiques/épiderme révèlent une profonde influence localisée des papilles dermiques sur l’épiderme, aboutissant, entre autres, à la formation de nouveaux follicules pileux au bout de 35 jours.

  4. On a donc montré de manière concluante que les papilles dermiques des vibrisses peuvent induire la formation de follicules pileux et la croissance des poils, même à partir d’épiderme d’origine afolliculaire.

  5. Les facteurs éventuels impliqués dans la détermination des caractéristiques de ces follicules sont discutés.

  6. Des papilles dermiques isolées et implantées dans le derme de l’oreille, gardent leur intégrité jusqu’à 207 jours après l’implantation.

  7. L’épiderme de l’oreille et de la peau du sac scrotal ainsi que l’épithélium buccal, implantés seuls dans l’oreille, apparaissent en général désorganisés ou en dégénérescence, ne montrant aucune régénération des follicules pileux.

I would like to express my thanks to Dr C. N. D. Cruickshank, Director of the M.R.C. Unit for Research on the Experimental Pathology of the Skin, The University of Birmingham, for his valuable suggestions and Mr J. Cooper and Mr N. Newbold for their technical assistance.

Argyris
,
T. S.
&
Trimble
,
M. E.
(
1964
).
On the mechanism of hair growth stimulation in wound healing
.
Devi Biol.
9
,
230
54
.
Billingham
,
R. E.
(
1958
).
A reconsideration of the phenomenon of hair neogenesis, with particular reference to the healing of cutaneous wounds in adult mammals. In The Biology of Hair Growth. Ed. W. Montagna and R. A. Ellis, pp. 451 –68
.
New York
:
Academic Press
.
Billingham
,
R. E.
&
Silvers
,
W. K.
(
1963
).
The origin and conservation of epidermal specificities
.
New Engl. J. Med.
268
,
477
–80, 539 –45.
Billingham
,
R. E.
&
Silvers
,
W. K.
(
1965
).
Some unsolved problems in the biology of skin. In Biology of the Skin and Hair Growth.
Ed.
A. G.
Lyne
and
B. F.
Short
, pp.
1
24
.
Sydney
:
Angus and Robertson
.
Billingham
,
R. E.
&
Silvers
,
W. K.
(
1967
).
Studies on the conservation of epidermal specificities of skin and certain mucosas in adult mammals
.
J. exp. Med.
125
,
429
46
.
Billingham
,
R. E.
&
Silvers
,
W. K.
(
1968
).
Dermoepidermal interactions and epithelial specificity. In Epithelial-Mesenchymal Interactions.
Ed.
R.
Fleischmajer
and
R. E.
Billingham
, pp.
252
66
.
Baltimore
:
Williams and Wilkins
.
Chase
,
H. B.
(
1965
).
Cycles and waves of hair growth
. In
Biology of the Skin and Hair Growth.
Ed.
A. G.
Lyne
and
B. F.
Short
, pp.
461
65
.
Sydney
:
Angus and Robertson
.
Cohen
,
J.
(
1961
).
The transplantation of individual rat and guinea-pig whisker papillae
.
J. Embryol. exp. Morph.
9
,
117
27
.
Cohen
,
J.
(
1964
).
Transplantation of hair papillae
.
Symp. zool. Soc. Lond.
12
,
83
96
.
Cohen
,
J.
(
1965
).
The dermal papilla
. In
Biology of the Skin and Hair Growth
Ed.
A. G.
Lyne
and
B. F.
Short
, pp.
183
99
.
Sydney
:
Angus and Robertson
.
Cohen
,
J.
(
1969a
).
Dermis, epidermis and dermal papilla interacting
. In
Hair Growth.
Ed.
W.
Montagna
and
R. L.
Dobson
, pp.
1
18
.
London
:
Pergamon Press
.
Cohen
,
J.
(
1969b
).
Interaction in the skin
.
Br. J. Derm.
81
,
Suppl. 3
,
46
54
.
Davidson
,
P.
&
Hardy
,
M. H.
(
1952
).
The development of mouse vibrissae in vivo and in vitro. J. Anat.
86
,
342
56
.
Jacobson
,
C. M.
(
1966
).
A comparative study of the mechanisms by which X-irradiation and genetic mutation cause loss of vibrissae in embryo mice
.
J. Embryol. exp. Morph.
16
,
369
79
.
Kollar
,
E. J.
(
1966
).
An in vitro study of hair and vibrissae development in embryonic mouse skin
.
J. invest. Derm.
46
,
254
62
.
Kollar
,
E. J.
(
1968
).
The inhibition of vibrissa development in vitro by beta-2-thienylalanine
.
J. invest. Derm.
50
,
319
22
.
Montagna
,
W.
(
1962
).
The Structure and Function of Skin.
New York
:
Academic Press
.
Oliver
,
R. F.
(
1966
).
Whisker growth after removal of the dermal papilla and lengths of follicle in the hooded rat
.
J. Embryol. exp. Morph.
15
,
331
47
.
Oliver
,
R. F.
(
1967a
).
Ectopic regeneration of whiskers in the hooded rat from implanted lengths of vibrissa follicle wall
.
J. Embryol. exp. Morph. Yl,
27
34
.
Oliver
,
R. F.
(
1967b
).
The experimental induction of whisker growth in the hooded rat by implantation of dermal papillae
.
J. Embryol. exp. Morph.
18
,
43
51
.
Oliver
,
R. F.
(
1968
).
The regeneration of vibrissae—a model for the study of dermal-epidermal interactions. In Epithelial-Mesenchymal Interactions.
Ed.
R.
Fleischmajer
and
R. E.
Billingham
, pp.
267
79
.
Baltimore
:
Williams and Wilkins
.
Oliver
,
R. F.
(
1969
).
The vibrissa dermal papilla and its influence on epidermal tissues
.
Br. J. Derm.
81
,
Suppl. 3
,
55
65
.
Straile
,
W. E.
(
1965
).
Root sheath—dermal papilla relationships and the control of hair growth. In Biology of the Skin and Hair Growth.
Ed.
A. G.
Lyne
and
B. F.
Short
, pp.
35
57
.
Sydney
:
Angus and Robertson
.
Van Scott
,
E. J.
&
Flaxman
,
B. A.
(
1968
).
Environmental control of epithelial cells in vivoand in vitro.
In
Epithelial-Mesenchymal Interactions.
Ed.
R.
Fleischmajer
and
R. E.
Billingham
, pp.
280
94
.
Baltimore
:
Williams and Wilkins
.