[Skip to Navigation]
Sign In
February 1999

Architectural Organization of Filiform Papillae in Normal and Black Hairy Tongue Epithelium: Dissection of Differentiation Pathways in a Complex Human Epithelium According to Their Patterns of Keratin Expression

Author Affiliations

From the Departments of Dermatology, Akita University School of Medicine, Akita, Japan (Dr Manabe), and Henry Ford Hospital, Detroit, Mich (Dr Lim); the Dermatohistopathology Laboratory, Westerstede, Germany (Dr Winzer); and The Ronald Perelman Department of Dermatology, New York University School of Medicine, NY (Dr Loomis).

Arch Dermatol. 1999;135(2):177-181. doi:10.1001/archderm.135.2.177

Background  An inadequate understanding of the complex morphologic characteristics of human filiform papillae has hampered the histopathological characterization of disorders affecting tongue keratinization. To better define the 3-dimensional cytoarchitecture of tongue epithelium, we performed detailed immunohistochemical analyses of normal and black hairy tongue tissues using a panel of antikeratin antibodies.

Observations  The dome-shaped base of the human filiform papilla (primary papilla) is surmounted by 3 to 8 elongated structures (secondary papillae). These secondary papillae are composed of a central column of epithelial cells expressing hair-type keratins and an outer rim of cells expressing skin-type keratins. The epithelium overlying the primary papillae and between the individual primary papillae express esophageal-type keratins. In black hairy tongue disease, there is a marked retention of secondary papillary cells expressing hair-type keratins.

Conclusions  Using a panel of antikeratin probes, we define the precise topographical localization of cell populations undergoing 3 distinct differentiation programs in dorsal tongue epithelium. Comparative analyses of black hairy tongue specimens indicate that defective desquamation of the cells in the central column of filiform papillae results in the formation of highly elongated, cornified spines or "hairs"—the hallmark of this disease.

THE DORSAL surface of mammalian tongue is covered by densely packed filiform papillae. The shape and size of these papillae vary markedly from species to species.1-8 In humans, the papillary architecture is more complex, consisting of a central body surrounded by several threadlike cornified projections, often referred to as secondary papillae.5,9 Morphologically, human tongue epithelium seems to be divided into discrete domains that undergo distinct pathways of terminal differentiation, similar to what occurs in rodent and cow tongue epithelium. However, the morphologic compartments of the human tongue have not been as well characterized as those of the rodent because of their greater complexity.

Keratins are a heterogeneous family of polypeptides that form a subclass of intermediate filaments. They serve as excellent markers for various pathways of epithelial differentiation and have been used to define discrete populations of keratinocytes within the tongue epithelium.10-13 The cumulative data indicate that, in addition to the ubiquitous expression of the stratified epithelial-type keratins (K5 and K14), different regions of the tongue epithelium display distinct patterns of keratin expression. The nonkeratinized epithelium of the lateral and ventral surfaces of the tongue, as well as the epithelium covering the lateral aspects of filiform and fungiform papillae and the interpapillary mucosa of the dorsal tongue, express esophageal-type keratins (K4 and K13). In contrast, the orthokeratinized epithelium overlying the tips of filiform papillae makes skin-type keratins (K1 and K10), and the epithelium covering taste buds synthesizes simple epithelial-type keratins (K8, K18, and K19). Finally, we and others14-16have shown that tongue epithelium also produces an acidic hard-type keratin, characteristic of hair- or nail-type differentiation.

In this article, we extend our immunolocalization studies on the structure and compartmentalization of the human filiform papilla and begin to explore the role of such compartments in human disease. We present the characterization of a new antikeratin K1 monoclonal antibody and its use to definitively localize the skin-type compartment in human tongue epithelium. In addition, we provide a more detailed description of the 3-dimensional organization of the separate domains within human filiform papillae. Finally, we demonstrate that the hairlike projections observed in black hairy tongue disease (BHT) are due primarily to the formation of abnormally long extensions of the "hair compartment" of the filiform papillae.

Materials and methods

Human tissue

Normal human tongues were obtained from autopsy within 48 hours of death. The 3 biopsy specimens of BHT were obtained from former patients of Manhattan Veterans Affairs Medical Center in New York as well as Medical University of Lübeck in Germany.

Antikeratin antibodies

Detailed characterizations of the monoclonal antibody AE 8, which reacts specifically with the esophageal-type keratin K13, and the antibody AE 13, which reacts with 44- to 46-kd acidic hair keratins, are described elsewhere.17,18 The AE 20 antibody was generated against trypsinized human skin cells according to the methods described previously.19 The antibody to K6 was generously provided by Dennis Roop, PhD, Baylor University, Houston, Tex.

Keratin extraction, gel electrophoresis, and immunoblotting

Normal human epidermis and cultured epithelial cells from various tissues were first extracted with 25-mmol/L Tris-hydrochloride, 0.6-mol/L potassium chloride, 1% Triton X-100 supplemented with 5 protease inhibitors.20 The residual aqueous-insoluble cytoskeletal preparation, containing mainly keratin proteins, was then solubilized with 2% sodium dodecyl sulfate in 25-mmol/L Tris-hydrochloride (pH 7.4) as described.20One-dimensional sodium dodecyl sulfate–polyacrylamide gel electrophoresis and subsequent immunoblotting were performed according to the methods described previously.21

Immunohistochemical staining

The biopsy specimens were embedded in OCT compound (Ty Miles Inc, Westchester, Ill), snap frozen in liquid nitrogen, and cut into 6-µm cryostat sections. Sections were stained by the indirect immunofluorescent and immunoperoxidase technique.19,22


Specificity of the ae 20 antibody for human epidermal keratin

In our prior study of human tongue epithelium,7indirect immunofluorescent staining was performed using the monoclonal antibody AE 2, which recognizes markers of skin-type differentiation, K1 and K10.19 Unfortunately, AE 2 cross-reacts with filaggrin, a major component of keratohyalin granules (data not shown), making it difficult to localize the skin-type keratins. To obtain more definitive immunolocalization data, we generated a monoclonal antibody, AE 20, that reacts specifically with the basic 67-kd human epidermal keratin, K1, as assessed by immunoblotting (data not shown). No AE 20 immunoreactivity was detected in keratin extracts from a variety of other tissues and cultured cells tested (data not shown).

Several distinct differentiation domains of crown-shaped human filiform papilla

The human filiform papilla consists of an elevated connective tissue core covered by a partially keratinized stratified epithelium. Small connective tissue protrusions emanate from the top of this central core and are surmounted by a column of epithelial cells undergoing cornification. These elongated cornified spines are referred to as secondary papillae and characteristically tilt toward the pharynx. In contrast to the papillary epithelium, the interpapillary surface is nonkeratinized.

To facilitate the topological analysis of this complex structure, we used a panel of monoclonal antibodies recognizing keratins that are characteristic of different types of stratified epithelia. The AE 8 monoclonal antibody specifically reacts with the K4 keratin, which is expressed in nonkeratinized, stratified epithelia such as in the esophagus. The AE 13 antibody recognizes acidic hair keratins, which are expressed in hair cortex, nail plate, and mouse tongue filiform papillae.14,17 The AE 20 antibody, described above, recognizes K1, which is characteristic of keratinized epithelia such as skin. The antibody to K6 recognizes a keratin that is normally expressed at a few restricted sites of the body, including volar skin, foreskin, oral mucosa, and outer root sheath cells of the hair follicle, as well as by all stratified epithelial cells undergoing hyperproliferation.23-25

Using this panel of antikeratin antibodies, we confirmed that the epithelium covering filiform papillae is subdivided into at least 3 distinct cell domains (data not shown). As reported previously, the antiesophageal keratin antibody, AE 8, reacted with the nonkeratinizing epithelia between the filiform papillae as well as between the cornified hairlike projections. The AE 20 immunoreactivity was confined to cells at the bases of the cornified spines, indicating that the skin-type keratin K1 is restricted to this cell population. The AE 13 antihair keratin antibody reacted only with those cells directly beneath the cornified spines, adjacent to the AE 20–positive cells. Immunostaining with the antibody to K6 was uniformly positive in the suprabasal cells of the dorsal tongue, similar to other oral epithelium.13

Detailed analysis of horizontal sections of human filiform papillae allowed us to further refine our understanding of the 3-dimensional architecture of the filiform papillae and the topological locales of different epithelial domains (data not shown). Within the large central area surrounded by a ring of 5 to 12 cornified spines are AE 8–positive cells, which synthesize esophageal-type keratins. The interpapillary epithelial cells likewise produce esophageal-type keratins, but they appear to constitute an esophageal-type compartment distinct from the AE 8–positive papillary epithelium. A thin band of AE 20–positive cells that undergo skin-type differentiation are distributed at the base of each of the cornified spines, encircling the cluster of AE 13–positive cells. Clusters of AE 13–positive cells, as well as their overlying cornified spines, are distributed in a ring at the periphery of the filiform papillae. These cells synthesize a "hair" keratin and undergo hair-type differentiation. Thus, the human filiform papilla is essentially crown shaped, with the outer rim being formed by the ring of hairlike spines. A scheme of the histological architecture of human filiform papillae is shown in Figure 1.

Figure 1. 
A schematic summary of filiform papillae in normal and black hairy tongue epithelium. Horizontal (A) and longitudinal (B) sections of normal dorsal tongue epithelium, and a longitudinal section (C) of black hairy tongue. Areas expressing hair- (h), skin- (s), and esophageal- (e) type keratins are the hatched, filled, and dotted areas, respectively.

A schematic summary of filiform papillae in normal and black hairy tongue epithelium. Horizontal (A) and longitudinal (B) sections of normal dorsal tongue epithelium, and a longitudinal section (C) of black hairy tongue. Areas expressing hair- (h), skin- (s), and esophageal- (e) type keratins are the hatched, filled, and dotted areas, respectively.

Aberrant elongation of the hairlike domain in bht tongue epithelium

Since the histological architecture of normal filiform papillae has never been described precisely, many common denominators make it difficult to differentiate between normal and BHT epithelium. Furthermore, the most striking finding on routine histological examination of BHT is the presence of numerous small fragments of cornified cells, which results from tangential sectioning. To avoid this problem, we made an effort to prepare well oriented longitudinal sections of the BHT biopsy specimens. Immunofluorescent staining using our panel of monoclonal antibodies demonstrated that the overall pattern of keratin expression in BHT is similar to that of normal tongue epithelium. The cells in the center of the filiform papillae as well as those in the interpapillary region are AE 8–positive, indicating the presence of esophageal-type keratins (Figure 2, A). The column of cells directly beneath the cornified spines are AE 20–positive, reflecting the presence of skin-type keratins ( Figure 2, B). The cells adjacent to the AE 20–positive cells are stained with AE 13 and thus undergo hair- and nail-type differentiation ( Figure 2, C and D). Our studies could not exclude, however, the possibility that the entire filiform papilla is involved, resulting in mild hyperkeratosis in both the skin-type and central intraspinous esophageal-type domains. Interestingly, the interpapillary esophageal domain (the dorsal tongue epithelium between the individual primary filiform papillae) shows no alterations in its differentiation, based on both morphological and immunolocalization criteria.

Figure 2. 
Immunofluorescent staining of black hairy tongue epithelium. Vertical sections are stained with AE 8 (A), AE 20 (B), or AE 13 (C). Panel D is a phase contrast image of panel C. P1 indicates primary filiform papillae; P2, secondary filiform papillae; and IP, interpapillary epithelium.

Immunofluorescent staining of black hairy tongue epithelium. Vertical sections are stained with AE 8 (A), AE 20 (B), or AE 13 (C). Panel D is a phase contrast image of panel C. P1 indicates primary filiform papillae; P2, secondary filiform papillae; and IP, interpapillary epithelium.


The tongue is covered by a complex epithelium composed of several functionally distinct cell populations. Our data have enabled us to definitively localize at least 3 unique domains within the tongue epithelium: (1) cells in the secondary filiform papillae that express acidic hair-type keratins, (2) a ring of cells surrounding this hair compartment that express skin-type keratins, and (3) cells overlying the central mound of the primary papillae as well as between the papillae that express esophageal-type keratins. It is fascinating that filiform papillae express both soft (epithelial) keratins and hard (trichocytic) keratins. It has been proposed that the coexistence of these different programs of keratin expression reflects the dual requirements of tongue epithelium to be both rigid and flexible to resist friction and expansion accompanying tongue movements during food handling and grooming.14

While the importance of the above observation remains to be determined, the existence of 3 distinct domains in filiform papillae raises an interesting question with respect to their cell origin. Do cells in the hair-, skin-, and esophageal-type domains originate from a common pluripotent stem cell, or does each arise from unique stem cells located at the base of its domain? Recent results26-29 suggest that epithelial stem cells, as defined by their kinetic properties, reside in specific sites within each epithelial tissue. Since stem cells are normally slow cycling, they can be identified as label-retaining cells after the labeling of all cycling cells by a continuous administration of isotope for a prolonged period. Using this approach, the stem cells of epidermis, intestinal epithelium, corneal epithelium, and hair follicles were shown to be located at the bottom of deep rete ridges, the crypts, the limbal region, and the bulge area, respectively.26,28-30Interestingly, these findings were further supported by the studies of graft-vs-host disease, in which the sites of involvement were thought to correlate precisely in the localization of the stem cells. In graft-vs-host disease, it was shown that cytotoxic lymphocytes infiltrate preferentially around the bottom of the epidermal rete ridges, the follicular bulge cells, and the anterior and posterior shoulders of filiform papillae.31-33 In this regard, our finding that filiform papillae can be divided into several differentiation compartments is noteworthy. It may be important to distinguish the stem cells of these domains because different stem cells may give rise to different neoplasms of the tongue epithelium.

The distinctive spatial arrangement of the hair domains of filiform papillae is of particular interest, given that hairs in skin also develop with a patterned distribution rather than at random.34-37 For example, a common grouping of follicles is the "trio" in which the primary hair follicle or tylotrich first erupts from the germ, and then 2 secondary hair follicles develop on both sides of the primary one, resulting in a central large tylotrich hair follicle flanked by 2 smaller hair follicles. Alternatively, the primary tylotrich follicle can be surrounded by a cluster of up to a dozen smaller follicles. If one assumes that the topological arrangement of the hair-type compartment (secondary papillae) at the periphery of the primary papillae in the tongue is analogous to the distribution of secondary hairs around a central tylotrich follicle in the skin, then the molecular mechanisms responsible for regulating these embryonic patterning events may be similar.

Black hairy tongue is an acquired disorder seen most frequently in chronic smokers.38 Clinically, the middorsal tongue is covered with long hairlike structures, hence the name. Although BHT has been histopathologically characterized extensively by Winzer et al,39 most of their descriptions also apply to normal histological characteristics of the tongue. For example, they noted that pathological changes associated with BHT include a markedly digitated surface, pointed rete ridges, focal parakeratosis, focal preservation of the granular zone, ballooned cells with pale cytoplasm in the spinous zone, and the presence of both keratohyalin- and trichohyalin-like granules. However, since normal tongue epithelium consists of distinct but closely juxtaposed differentiation domains, the above "pathological" findings are also observed in normal tongue. Without a clear understanding of the precise histological topography of the tongue and its specialized epithelial domains, one cannot meaningfully analyze the pathological processes in tongue diseases. Our present study shows that the designation of black hairy tongue is by chance consistent with the specific keratin expression in cells that cause the characteristic morphologic features. Cells underneath the strikingly elongated cornified spines are indeed derived from the hair domain of filiform papillae. However, it remains unclear why the hair-type compartment shows a prominent retention of cornified spines, and how the pathogenic factors implicated in promoting BHT, such as smoking, oxidizing agents, and antibiotics, inhibit hair compartment desquamation. Further studies are necessary to elucidate the relevant mechanisms. In this regard, it might be interesting to investigate whether medicines that influence hair growth may also cause morphologic alterations of the filiform papillae.

Accepted for publication June 22, 1998.

Research for this article has been supported in part by a Dermatology Foundation Career Development Award (Dr Loomis).

We thank Tung-Tien Sun, PhD (New York University, New York, NY), and Helmut H. Wolff, MD (Medical University of Lübeck in Germany), for their useful discussions.

Reprints: Motomu Manabe, MD, PhD, Department of Dermatology, Juntendo University School of Medicine, Hondo 1-1-1, Akita 113-8421, Japan (e-mail: manabe@doc.med.akita-u.ac.jp).

Farbman  AI The dual pattern of keratinization in filiform papillae on rat tongue.  J Anat. 1970;106233- 242Google Scholar
Steflik  DESingh  BBMckinney  RV  Jr  et al.  Correlated TEM, SEM, and histological observations of filiform papillae of the cow tongue.  Acta Anat (Basel). 1983;11721- 30Google ScholarCrossref
Iida  MYoshikawa  IMuto  H Three dimensional and surface structures of rat filiform papillae.  Acta Anat (Basel). 1985;121237- 144Google ScholarCrossref
Iwasaki  SMiyata  KKobayashi  K Comparative studies of the dorsal surface of the tongue in three mammalian species by scanning electron microscopy.  Acta Anat (Basel). 1987;128140- 146Google ScholarCrossref
Mikkonen  AHynynen  MHynynen  P Filiform papillae of human, rat and swine tongue.  Acta Anat (Basel). 1987;130280- 284Google ScholarCrossref
de Paz Cabello  PChamorro  CASandoval  J  et al.  Comparative scanning electron microscopic study of the lingual papillae in two species of domestic mammals (Equus caballus and Bos taurus), II: mechanical papillae.  Acta Anat (Basel). 1988;132120- 123Google ScholarCrossref
Hume  WJPotten  CS The ordered columnar structure of mouse filiform papillae.  J Cell Sci. 1976;22149- 160Google Scholar
Mackenzie  ICBickenback  JR Identification and localization of label-retaining cells in hamster epithelia.  J Invest Dermatol. 1984;82618- 622Google ScholarCrossref
Mikkonen  ASorvari  TE A scanning electron microscopic study of the dorsal surface of the human tongue.  Acta Anat (Basel). 1985;123114- 120Google ScholarCrossref
Galvin  SLoomis  CManabe  M  et al.  The major pathways of keratinocyte differentiation as defined by keratin expression: an overview.  Adv Dermatol. 1988;4277- 300Google Scholar
Weidauer  HNagle  RBMoll  R  et al.  Patterns of cytokeratin expression in normal and malignant epithelium of the tongue.  Laryngol Rhinol Otol (Stuttg). 1986;65164- 168Google ScholarCrossref
Rentrop  MBKnapp  HWinter  H  et al.  Differential localization of distinct keratin mRNA-species in mouse tongue epithelium by in situ hybridization with specific cDNA probes.  J Cell Biol. 1986;1032583- 2591Google ScholarCrossref
Sawaf  MHOuhayoun  JPShabana  AHM  et al.  Cytokeratin expression in human tongue epithelium.  Am J Anat. 1990;189155- 166Google ScholarCrossref
Dhouailly  DXu  CManabe  M  et al.  Expression of hair-related keratins in a soft epithelium: subpopulations of human and mouse dorsal tongue keratinocytes express keratin markers for hair-, skin-, and esophageal-types of differentiation.  Exp Cell Res. 1989;181141- 158Google ScholarCrossref
Heid  HWMoll  IFranke  WW Patterns of expression of trichocytic and epithelial cytokeratins in mammalian tissues, II: concomitant and mutually exclusive synthesis of trichocytic and epithelial cytokeratins in diverse human and bovine tissues (hair follicle, nail bed and matrix, lingual papilla, thymic reticulum).  Differentiation. 1990;37215- 230Google ScholarCrossref
Tobiasch  ESchweizer  JWinter  H Structure and site of expression of a murine type II hair keratin.  Mol Biol Rep. 1992;1639- 47Google ScholarCrossref
Lynch  MHO'Guin  WMHardy  C  et al.  Acidic and basic hair/nail ("hard") keratins: their colocalization in upper cortical and cuticle cells of the human hair follicle and their relationship to "soft" keratins.  J Cell Biol. 1986;1032593- 2606Google ScholarCrossref
Pang  YYSchermer  AYu  J  et al.  Suprabasal change and subsequent formation of disulfide-stabilized homo- and hetero-dimers of keratins during esophageal epithelial differentiation.  J Cell Sci. 1993;104727- 740Google Scholar
Woodcock-Mitchell  JEichner  RNelson  WG  et al.  Immunolocalization of keratin polypeptides in human epidermis using monoclonal antibodies.  J Cell Biol. 1982;95580- 588Google ScholarCrossref
Schermer  AGalvin  SSun  T-T Differentiation-related expression of a major 64K corneal keratin in vitro and in culture suggests limbal location of corneal epithelial stem cells.  J Cell Biol. 1986;10349- 62Google ScholarCrossref
Eichner  RBonitz  PSun  T-T Classification of epidermal keratins according to their immunoreactivity, isoelectric point, and mode of expression.  J Cell Biol. 1984;981388- 1396Google ScholarCrossref
Manabe  MO'Guin  WM Existence of trichohyalin-keratohyalin hybrid granules in non-follicular epithelia: the co-localization of two major intermediate filament-associated proteins.  Differentiation. 1994;5865- 67Google ScholarCrossref
Weiss  RAEichner  RSun  T-T Monoclonal antibody analysis of keratin expression in epidermal diseases: a 48- and 56-k dalton keratin as molecular markers for hyperproliferative keratinocytes.  J Cell Biol. 1984;981397- 1406Google ScholarCrossref
Moll  RFranke  WWSchiller  DLGeiger  BKrepler  R The catalog of human cytokeratins: patterns of expression in normal epithelia, tumors and cultured cells.  Cell. 1982;3111- 24Google ScholarCrossref
Cooper  DSchermer  APruss  RSun  T-T Classification of human epithelia and their neoplasms using monoclonal antibodies to keratins: strategies, applications, and limitations.  Lab Invest. 1985;52243- 256Google Scholar
Leblond  CP The life history of cells in renewing systems.  Am J Anat. 1981;160114- 158Google ScholarCrossref
Bickenbach  JR Identification and behavior of label-retaining cells in oral mucosa and skin.  J Dent Res. 1981;601611- 1620Google ScholarCrossref
Lavker  RMSun  T-T Heterogeneity in epidermal basal keratinocytes: morphological and functional correlations.  Science. 1982;2151239- 1241Google ScholarCrossref
Cotsarelis  GCheng  S-ZDong  G  et al.  Existence of slow-cycling limbal epithelial basal cells that can be preferentially stimulated to proliferate: implications on epithelial stem cells.  Cell. 1989;57201- 209Google ScholarCrossref
Cotsarelis  GSun  T-TLavker  RM Label-retaining cells reside in the bulge area of pilosebaceous unit: implications for follicular stem cells, hair cycle, and skin carcinogenesis.  Cell. 1990;611329- 1337Google ScholarCrossref
Sale  GEShulman  HMGallucci  BB  et al.  Young rete ridge keratinocytes are preferred targets in cutaneous graft-versus-host diseases.  Am J Pathol. 1985;118278- 287Google Scholar
Sale  GEBeauchamp  MDAkiyama  M Parafollicular bulge, but not hair bulb keratinocytes, are attacked in graft-versus-host disease of human skin.  Bone Marrow Transplant. 1994;14411- 413Google Scholar
Sale  GERaff  RFStorb  R Stem cell regions in filiform papillae of tongue as targets of graft-versus-host disease.  Transplantation. 1994;581273- 1287Google ScholarCrossref
Noback  CR Morphology and phylogeny of hair.  Ann N Y Acad Sci. 1951;53477- 492Google ScholarCrossref
Hussein  MAF The overall pattern of hair follicle arrangement in the rat and mouse.  J Anat. 1971;109307- 316Google Scholar
Straile  WE Sensory hair follicles in mammalian skin: the tylotrich follicle.  Am J Anat. 1960;106133- 141Google ScholarCrossref
Zhou  PByrne  CJacobs  JFuchs  E Lymphoid enhancer factor 1 directs hair follicle patterning and epithelial cell fate.  Genes Dev. 1995;9700- 713Google ScholarCrossref
Sarti  GMHaddy  RISchaffer  D  et al.  Black hairy tongue.  Am Fam Physician. 1990;411751- 1755Google Scholar
Winzer  MGilliar  UAckerman  AB Hairy lesions of the oral cavity: clinical and histopathologic differentiation of hairy leukoplakia from hairy tongue.  Am J Dermatopathol. 1988;10155- 159Google ScholarCrossref