By Abby Van Voorhees, MD, July 01, 2015
In this month’s Acta Eruditorum column, Physician Editor Abby S. Van Voorhees, MD, talks with Scott Binder, MD, and Albert Su, MD, about their recent Modern Pathology article, “Keratoacanthoma and squamous cell carcinoma are distinct from a molecular perspective.”
Dr. Van Voorhees: As you state in your paper keratoacanthomas have been controversial for many years with some believing that they are squamous cell carcinomas and others not. So I thought that many Academy members would find your paper of great interest. Tell us what you did to try to further the understanding of the genetics of KAs.
Drs. Binder and Su: At UCLA we have had success with microarray-based gene expression profiling of various types of tumors including squamous lesions. These comprehensive microarrays allow us to examine the expression of over 47,000 genes. We thought it would be interesting to apply this methodology to examine the genetics of KAs. For this study, we selected classic cases of KA and compared their gene expression profiles to a database of SCC and normal skin that we already had. All the cases of KA had the typical histologic features that we usually associate with KA, such as symmetric crateriform architecture and glassy eosinophilic cytoplasm. We found a relatively large number of genes differentially expressed between KA and SCC, suggesting that they are distinct from a molecular perspective.
Dr. Van Voorhees: Some of your KAs were seen in the setting of patients on BRAF inhibitors. Are there differences between spontaneous KAs and those that develop when on these medications?
Drs. Binder and Su: We have compared KAs occurring in association with BRAF inhibitor treatment to KAs occurring sporadically, and they had similar gene expression profiles. Since the BRAF inhibitor vemurafenib can paradoxically activate the MAP kinase pathway in melanoma cells without the V600E BRAF mutation, we wondered if the KAs that developed with drug treatment occurred due to stimulation of the MAP kinase pathway. We performed quantitative RT-PCR on the various KA samples to amplify several genes within the MAP kinase pathway, but there was no significant difference in expression between drug-induced and sporadic KAs. Basically, we could not find a discernible difference between the two groups of KAs. It’s possible that the BRAF inhibitor treatment may induce a transient activation of the MAP kinase pathway that contributes to KA development but which is no longer present at the time of biopsy. This would be an interesting topic for further research.
Dr. Van Voorhees: What did you learn about the microarray parameters of the KAs, SCCs, and normal skin? How do they compare to each other?
Drs. Binder and Su: We found 1,449 genes differentially expressed between KA and SCC, which is a relatively large number. In an effort to shed light on the pathogenesis of KA, we also compared the gene expression profiles of KA to that of normal skin. There were 2,435 differentially expressed genes, and analysis of the genes showed enrichment of various molecular and cellular functions associated with neoplasia. Interestingly, we also found enrichment of apoptosis/cell death pathways. These results suggest that KA is distinct from SCC and represents an “abortive neoplasm.” In other words, KA might be a squamoproliferative lesion that is held in check by apoptosis.[pagebreak]
Dr. Van Voorhees: How do SCC microarrays compare with what is known about actinic keratoses (AKs)? What about other situations where we see the epidermis go amok like pseudoepitheliomatous hyperplasia?
Drs. Binder and Su: When we studied SCC and AK with the microarrays, we found that they had similar gene expression profiles but there were nine genes that were significantly differentially expressed. These few differentially expressed genes have been implicated in tumorigenesis, and our results are consistent with the hypothesis that AK is a precursor to SCC. We have also compared SCC to pseudoepitheliomatous hyperplasia. Those experiments showed a large number of genes differentially expressed between SCC and pseudoepitheliomatous hyperplasia with enrichment of molecular pathways that have been implicated in malignancy. The distinct gene expression profiles are consistent with the idea that pseudoepitheliomatous hyperplasia and SCC are different entities although they may look similar histologically.
Dr. Van Voorhees: Does your research shed any light on why KAs will spontaneously involute?
Drs. Binder and Su: When comparing KA to normal skin, we found the cell death/apoptotic pathway to be one of the most significantly enriched pathways. There have been a few other studies in the literature that have compared KA to SCC by utilizing immunohistochemical markers for various proteins involved in apoptosis. Those studies have also suggested a role for apoptosis in the spontaneous involution of KAs. It is thought that KAs may be follicle-derived, and thus the involution seen with KAs might be akin to catagen involution of the hair follicle.
Dr. Van Voorhees: How does the microarray data compare with that seen in other cancers?
Drs. Binder and Su: Since there are so many different types of cancers, it is difficult to compare our microarray data from this experiment with that seen in other cancers. It’s probably most useful to make comparisons between similar types of lesions and with a good control. Many of our microarray experiments have focused specifically on squamous lesions with normal skin as a control.
Dr. Van Voorhees: What should the practicing dermatologist learn from this work? Are there take-home messages that we can use to guide our approach to the treatment of these lesions?
Drs. Binder and Su: Our results suggest that KA is distinct from SCC and represents a squamoproliferative lesion that may involute due to upregulation of apoptosis. We think the results are very interesting and give a good explanation of why classic KAs spontaneously involute, but we also want to emphasize that ultimately the diagnosis of KA should be based on a good clinicopathologic correlation. In our study we used classic cases of KA, but we all know that in actual practice the biopsies may not always be so definitive due to factors such as partial sampling, artifact, etc. Although in theory one should be able to treat KAs conservatively, the clinical judgment of the dermatologist is paramount and should guide the management of these squamous proliferations.
Dr. Binder is the director of pathology and lab medicine clinical services and chief of dermatopathology at UCLA Health System.
Dr. Su is senior clinical instructor in UCLA’s department of pathology. Their article appeared in Modern Pathology, 2015 Feb 13. doi: 10.1038/modpathol.2015.5.