Keratoacanthoma, committed stem cells and neoplastic aberrant infundibulogenesis integral to formulating a conceptual model for an infundibulocystic pathway to squamous cell carcinoma

Abstract Keratoacanthomas (KAs) are distinctive tumors that are defined by their clinical and histopathological features. Their relationship and distinction from squamous cell carcinoma (SCC), however, remain controversial. All cytogenic and immunohistochemical markers that have been applied in this quest have failed. A close relationship of KAs to hair follicles has been recognized. The descriptive term infundibulocystic or infundibular SCC was introduced to define a more broad‐based pathway encompassing KAs. The follicular infundibulum roles in respect to neoplasia and wound healing are important elements in understanding the pathogenesis of KAs. Mouse models for KA have provided insights into the relationship of KA to follicles and SCCs. These advances and together with the diverse clinical and histopathological aspects of KA have contributed to the formulation of a conceptual pathway. The central element is that ultraviolet (UV)‐mutated or activated committed infundibular stem cells are driven by the combination of a mutated oncogenic RAS pathway linked with the Wnt/beta‐catenin pathway responsible for stem cell maintenance, hair follicle development, wound healing and driving KA proliferation and terminal keratinization. The existence and activation of this mutated pathway may form the basis of the paradoxical emergence of KAs and SCCs in patients receiving BRAF and PD‐1 inhibitor therapy.


| INTRODUCTION
Keratoacanthomas (KAs) are distinctive tumors that have been defined by their clinical and histopathological features, but their relationship to hair follicles and their distinction from squamous cell carcinoma (SCC) remain a controversy. The descriptive term infundibulocystic or infundibular SCC was introduced to define a more broad-based pathway to SCC encompassing KAs and incorporate the follicular infundibulum as an essential element to the process. 1 A broad range of biomarkers, particularly cytogenic markers and immunohistochemistry including CD123 have been unsuccessful in the quest to distinguish KA from SCC. 2 This may reflect the fact that both KAs and SCCs are chronic UV-induced squamoproliferative tumors sharing the same driver RAS mutations for neoplasia. 3 Early carcinogen-induced KA mouse models 4 to show their relationship to follicles have been augmented by genetically engineered models. 5 Significant advances have been made in defining skin stem cells that underpin both the homeostasis and dynamic physiology of follicles and the interfollicular epidermis. [6][7][8] Advances in stem cell biology in humans have been reviewed in detail in respect to epidermal regeneration and carcinogenesis 9 and have provided an opportunity to formulate a conceptual model within the framework of the infundibulocystic pathway to SCC. The central element is the concept of UV-mutated or activated committed infundibular stem cells that are driven by the combination of two major molecular pathways, namely the mutated oncogenic RAS pathway linking KA and squamous cell carcinoma and the Wnt/beta-catenin pathway responsible for skin stem cell maintenance, hair follicle development, wound healing and driving KA proliferation and terminal keratinization. 10 Figure 1A, B). This reflects that for committed infundibular stem cells, squamoproliferation is integral to infundibulogenesis. Complex infundibulogenesis characterizes the early phase and may have distorted infundibular hyperplasia, verrucous, and superficial squamoproliferation with lichenoid inflammation (Figure 2A-C). The broad range of precursors is clearly apparent in eruptive squamoproliferative tumors on the lower limbs that can range from KAs to SCCs that are often not crateriform (Figure 3 A, B).
Large verrucous plaques with papillomatous and infundibulocystic squamoproliferation may complicate skin grafts after KA surgery ( Figure 3C, D). A subset of crateriform SCCs are likely to belong to this pathway 1,12 with the initial infundibulocystic phase rapidly subsumed by SCC producing nonulcerated craters bordered by invasive SCC ( Figure 4A, B). A rarer type of SCC was also defined as a deep and broad-based tumor retaining infundibulocystic differentiation in the deep dermis ( Figure 4C). 1 Well-differentiated infundibulocystic SCCs including KAs dominate the pathway, but with increasing UV-induced mutations, moderate and poorly differentiated SCCs may emerge. To be in the pathway, these tumors can only be identified by the presence of a recognizable infundibular or infundibulocystic precursors representing differentiation allied to KA. The oncogenic burden ultimately signals a departure from the committed stem cell pathway and progressive loss of recognizable committed terminal infundibular differentiation.
The increased mutational burden in the infundibulocystic pathway has been shown in the study by Ko et al. 13 This study included   The authors emphasized that they did not view these SCCs as primary follicular tumors. The importance of this study is that UV-induced F I G U R E 2 A, B, Precursors with convoluted and complex infundibulocystic hyperplasia with prominent keratinization and minimal squamoproliferation or atypia (H&E, x12.5) and, C, broadbased precursor demonstrating prominent irregular papillomatosis with infundibulosquamous proliferation and early invasive squamous cell carcinoma (H&E, x12.5) F I G U R E 4 A, Crateriform nonulcerated squamous cell carcinoma with papillomatous and keratotic cavity without prominent infundibular or infundibulocystic differentiation (H&E, x12.5). B, Detail of base with invasive cords with fullthickness atypia and minimal maturation through enlarged keratinocytes (H&E, x50). C, Broad base deeply and widely invasive welldifferentiated infundibulocystic squamous cell carcinoma with fibrosis and lichenoid inflammation (H&E, x12.5) F I G U R E 5 Comparison of, A, follicular squamous cell carcinoma as a nodular tumor centered on follicles with irregular solid lobules projecting into surrounding dermis (H&E, x12.5) and, B, infundibulocystic follicular squamous cell carcinoma centered on follicle with squamoproliferation and complex irregular anastomosing cords with prominent infundibular differentiation (H&E, x50) The mouse model for carcinogen-induced KAs 4 highlighted the role of pre-existing hair follicles. KA inception in humans may also emerge from the infundibulum of existing hair follicles, but the process is usually more complex in humans. The follicular density in human skin is a fraction seen in mouse skin and the interfollicular epidermis dominates. In humans, epidermal rather than follicular-based infundibulogenesis with squamoproliferation is more frequent for the development of KAs and infundibulocystic SCCs.

| INFUNDIBULOGENESIS AS AN EXPRESSION OF COMMITTED INFUNDIBULAR STEM CELLS
Epidermal infundibulogenesis has to be distinguished from the complex process of trichogenesis. Pluripotential stem cells within the follicular bulge and rarely from the interfollicular dermis in adults need dermal mesenchyme signaling for trichogenesis. The process is complex and is dependent on assembling a highly integrated set of differing committed stem cells that form a follicle. The identity of committed stem cells is declared through terminal differentiation. The infundibulum is one such component. RAS-mutated infundibular neoplasia is not dependent on mesenchyme stroma associated with the mutated sonic hedgehog (Shh) pathway driving basal cell carcinomas. 9 Misago et al 16  SCCs that can be eruptive. The relationship of stem cells, carcinogenesis, and the wound pathway is complex. 9,20 Particularly relevant is the Wnt/beta-catenin pathway, which plays a critical role in skin stem cell maintenance, hair follicle development, wound healing and driving KA proliferation and subsequent terminal keratinization. 10,11 In hair-bearing skin, wound repair is accelerated by infundibular keratinocytes driven by committed infundibular stem cells. CK15 (clones KRT15, C8/1448) was the first molecular marker for bulgelocated stem cells. This label primarily localized to the bulge has been shown to be absent in KAs 16 and to only persist temporarily within the interfollicular epidermis following wound healing. 18 In contrast, CK15 clone LHK 15 was shown by Misago et al 16

| RETINOID THERAPY IN KA AND INFUNDIBULOCYSTIC SCC
Retinoids have been used to treat and slow down the appearance of multiple KAs and SCCs particularly those that are not amenable for surgery. This modality has been investigated in a pivotal study using a mouse-based KA model. 11 The investigators found that self-regressing tumors shift into terminal differentiation during regression. The Wnt follicular signaling pathway sustains the growth phase of KAs and retinoids induced regression by slowing proliferation and promoting differentiation and terminal keratinization. Importantly, the retinoids were found to also induce regression in associated SCCs. This has implications for the nonsurgical treatment of the wider group of infundibular and infundibulocystic SCCs that may not evolve from KAs but arise within the wider spectrum of broadly based infundibulocystic and verrucous precursors. SCCs. 30,31 Significantly mutations could also be found in in a subset of early precursors that were infundibulocystic supporting an oncogenic pathway with increased mutational burden. In general, these papular keratotic tumors appear rapidly but are often indolent, have not been destructive or metastasized and can regress. The combining of a MEK inhibitor with the BRAF inhibitor has been shown to prevent the emergence of these tumors. 32 Topical BRAF inhibitors in wound healing mice models have been shown to accelerate wound healing through similar paradoxical MAPK activation. 33  The infundibulocystic pathway to KAs and SCCs can also be activated by programmed cell death 1 (PD-1) inhibitors. 35   A comparative study of KA and SCC utilizing comparative DNA microarray analysis revealed that KAs preferentially expressed genes that upregulate the cell death/apoptosis pathway in comparison to SCCs. 36 This confirms that KA's have distinct features but does not indicate that they are not SCC's within the infundibulocystic pathway.
The reason for the emergence of the controversies in the relationship of KA and SCC is probably to be directly related to the complex nature of the conceptual pathway combining committed infundibular stem cells, UV-induced RAS-mutated KA and SCC and the Wnt/beta-catenin pathway. This is not a linear pathway as the burden of UV-induced RAS mutations may at any phase overcome the involutional capacity linked to infundibulogenesis.
The focus on committed skin stem cells was initially triggered by a rare case presentation of KA centrifugum marginatum, dominated by verrucous and acanthotic gross keratoses without infundibulogenesis, limited to a finger. 37 The complex presentation was a challenge to analyze without formulating a stem cell-based model.
The importance of this case is that the early precursor stage of the pathway that is verrucous, squamoproliferative and keratotic without infundibulogenesis was sufficient to present clinically as KA centrifugum marginatum appearing as benign nodules with gross keratinization that did not readily involute.
The separation of the infundibulocystic stem cell pathway to SCC from other committed epidermal derived stem cell pathways to SCC may not be simple and needs further exploration. It is possible that the infundibulocystic pathway may also be secondarily activated as an innate response to SCCs emerging from these alternative mutated pathways and is not always a primarily mutated infundibulocystic event.
This conceptual pathway potentially provides further areas to explore the many facets linked to the complex nature of KA particularly as a longstanding member of a broad group of tumors including SCC.

CONFLICT OF INTEREST
The author declares no conflicts of interest.

DATA AVAILABILITY STATEMENT
No new data was generated in this review leading to a proposed conceptual infundibulocystic pathway as this was based on published studies and the author's analysis of the controversies linked to keratoacanthoma.