Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2023 Dec;143(12):2397-2407.e8.
doi: 10.1016/j.jid.2023.06.192. Epub 2023 Jul 5.

Distinguishing Keratoacanthoma from Well-Differentiated Cutaneous Squamous Cell Carcinoma Using Single-Cell Spatial Pathology

Affiliations

Distinguishing Keratoacanthoma from Well-Differentiated Cutaneous Squamous Cell Carcinoma Using Single-Cell Spatial Pathology

Jesse Veenstra et al. J Invest Dermatol. 2023 Dec.

Abstract

Keratoacanthoma (KA) is a common keratinocyte neoplasm that is regularly classified as a type of cutaneous squamous cell carcinoma (cSCC) despite demonstrating benign behavior. Differentiating KA from well-differentiated cSCC is difficult in many cases due to the substantial overlap of clinical and histological features. Currently, no reliable discriminating markers have been defined, and consequently, KAs are often treated similarly to cSCC, creating unnecessary surgical morbidity and healthcare costs. In this study, we used RNA sequencing to identify key differences in transcriptomes between KA and cSCC, which suggested divergent keratinocyte populations between each tumor. Imaging mass cytometry was then used to identify single-cell tissue characteristics, including cellular phenotype, frequency, topography, functional status, and interactions between KA and well-differentiated cSCC. We found that cSCC had significantly increased proportions of Ki67+ keratinocytes among tumor keratinocytes, which were dispersed significantly throughout non-basal keratinocyte communities. In cSCC, regulatory T-cells were more prevalent and held greater suppressive capacity. Furthermore, cSCC regulatory T-cells, tumor-associated macrophages, and fibroblasts had significant associations with Ki67+ keratinocytes as opposed to avoidances with KA, indicating a more immunosuppressive environment. Our data suggest that multicellular spatial features can serve as a foundation to enhance the histological discrimination of ambiguous KA and cSCC lesions.

PubMed Disclaimer

Conflict of interest statement

Conflict of interest: The authors declare they have no conflicts of interest.

Figures

Figure 1.
Figure 1.. KA and cSCC transcriptomic profiles.
(a) Heatmap of significant DEGs identified from bulk RNAseq of KA and cSCC (n=10/group). (b) Principal component analysis with VST normalization. Top 10 predicted biofunctions for (c) cSCC and (d) KA using all significant genes. (e) Top predicted biofunctions from cSCC and KA combined with fold change >1 (positive z-score representative of cSCC, negative z-score representative of KA). Predicted biofunctions assessed through Qiagen Ingenuity Pathway Analysis.
Figure 2.
Figure 2.. Imaging Mass Cytometry of KA and cSCC and cluster identification.
Representative H&E (a) and pseudocolored IMC images from KA (left) and cSCC (right) demonstrating keratinocyte morphology and T-cell infiltrates. (b) aSMA – white, Ki67 - red, K14 – cyan , K10 – yellow, and Collagen I – blue. (c) Pan keratin – white, CD3 – red, CD8 – green, CD4 – blue, and Foxp3 – cyan. (d) ECAD – white, TP63 – red, EGFR – blue. IMC images visualized in MCD Viewer with pseudocoloration (Standard BioTools); scale bar = 300 μm. (e) Heatmap of median epitope expression (x-axis) and 17 identified cellular clusters (y-axis) based on hierarchal grouping. (f) Proportion of identified cellular clusters within each sample. (g) Mean cluster proportion calculated based on total cells per sample. Welch t-test with Holm-Sidak correction for multiple comparisons. *P ≤ 0.05. DCs, dendritic cells; Eccrine, Eccrine glands and ducts; K10, Keratin 10; LCs, Langerhans cells; NOC, not otherwise classified; PMN, Polymorphonuclear neutrophils.
Figure 3.
Figure 3.. Keratinocyte phenotypes.
(a) Keratinocyte cluster proportion based on total population of keratinocytes. Representative pseudocolored IMC images from (b) KA, (c) cSCC, and (d) normal skin demonstrating K10 (cyan) and K14 (red) expression. Images visualized in MCD Viewer with pseudocoloration (Standard BioTools). Scale bar = 300 µm. (e) Relative expression of p53 and TP63 among keratinocyte clusters (Ki67+, Basal (Ki67−), K10 high, and NOC keratinocyte clusters shown, left to right). (f) Relative expression of targets in Ki67+ keratinocyte cluster. Keratinocyte classification by CIBERSORTx analysis from bulk RNAseq using reference single-cell RNAseq libraries from (g) healthy skin (Reynolds et al., 2021) and (h) cSCC (Ji et al., 2020); Proportion of total cells/sample. Welch t-test with Holm-Sidak correction for multiple comparisons. *P ≤ 0.05, **P ≤ 0.01. K10, keratin 10; NOC, not otherwise classified; TSK, Tumor-specific keratinocyte.
Figure 4.
Figure 4.. Ki67+ keratinocyte topography.
Representative pseudocolored IMC images from KA (a-b) and cSCC (c-d) demonstrating proliferating Ki67+ keratinocytes. Pan-keratin (red) and Ki-67 (green); scale bar = 300 μm. (e) Voronoi plot of representative keratinocyte and stromal communities; from cSCC. (f) Distribution (left) and spatial clustering (right) of proliferating Ki67+ keratinocytes within basal and non-basal keratinocyte communities; spatial clustering assessed by Clark-Evans aggregation index (R), where R=1 is random, R<1 indicates greater clustering, and R>1 indicates greater dispersion. (g) Validation cohort of KA and cSCC (n=20/group) using whole tissue sections with Ki67 IHC; Ki67+ spatial index (% Ki67+ cells+R). Welch t-test with Holm-Sidak correction for multiple comparisons. *P ≤ 0.05, **P ≤ 0.01, ****P ≤ 0.0001.
Figure 5.
Figure 5.. Immune phenotypes in tumor microenvironment.
(a) Cluster proportion based on total population of lymphoid, myeloid, and stroma lineages, respectively. (b) Cluster proportion located in tumor and stroma compartments based on total sample cluster population. Relative expression of functional markers among (c) T-cell and (d) myeloid clusters. (e) Macrophage MMP10 expression levels by tumor or stromal compartment. Welch t-test with Holm-Sidak correction for multiple comparisons. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001.
Figure 6.
Figure 6.. Spatial analysis of KA and cSCC TME.
(a) Heatmap of cellular associations and avoidances between clusters (from cluster, x-axis; to cluster, y-axis). Outlined boxes indicate proliferating keratinocyte cluster interaction with fibroblast (red), Treg (blue), macrophage (green), and PMN (orange) clusters. Interaction testing performed with 1,000 iterations and a significance threshold of 0.01. Only significant values were allowed to contribute, which averaged associations (red, +1) and avoidances (blue, −1) across samples per group. (b) Interactions from Ki67+ keratinocytes to clusters of interest; values of 0 were nonsignificant. Welch t-test with Holm-Sidak correction for multiple comparisons. *P ≤ 0.05, **P ≤ 0.01, ****P ≤ 0.0001. DCs, dendritic cells; GrB, granzyme B; LCs, Langerhans cells; PMN, Polymorphonuclear neutrophils.

Similar articles

Cited by

References

    1. Amarnath S, Mangus CW, Wang JC, Wei F, He A, Kapoor V, et al. The PDL1-PD1 axis converts human TH1 cells into regulatory T cells. Sci Transl Med 2011;3(111):111ra20. - PMC - PubMed
    1. Ambur A, Clark A, Nathoo R. An Updated Review of the Therapeutic Management of Keratoacanthomas. J Clin Aesthet Dermatol 2022;15(12 Suppl 1):S16–S22. - PMC - PubMed
    1. Bankhead P, Loughrey MB, Fernandez JA, Dombrowski Y, McArt DG, Dunne PD, et al. QuPath: Open source software for digital pathology image analysis. Sci Rep 2017;7(1):16878. - PMC - PubMed
    1. Bedir R, Gucer H, Sehitoglu I, Yurdakul C, Bagci P, Ustuner P. The Role of p16, p21, p27, p53 and Ki-67 Expression in the Differential Diagnosis of Cutaneous Squamous Cell Carcinomas and Keratoacanthomas: An Immunohistochemical Study. Balkan Med J 2016;33(2):121–7. - PMC - PubMed
    1. Boyd S, Virolainen S, Parssinen J, Skoog T, van Hogerlinden M, Latonen L, et al. MMP-10 (Stromelysin-2) and MMP-21 in human and murine squamous cell cancer. Exp Dermatol 2009;18(12):1044–52. - PubMed

Publication types