MiRNA-10b marks aggressive squamous cell carcinomas, and confers a cancer stem cell-like phenotype

Background Cutaneous squamous cell carcinomas (cSCC) are the primary cause of premature deaths in patients suffering from the rare skin-fragility disorder recessive dystrophic epidermolysis bullosa, which is in marked contrast to the rarely metastasizing nature of these carcinomas in the general population. This remarkable difference is attributed to the frequent development of chronic wounds caused by an impaired skin integrity. However, the specific molecular and cellular changes to malignancy, and whether there are common players in different types of aggressive cSCCs, remain relatively undefined. Methods MiRNA expression profiling was performed across various cell types isolated from skin and cSCCs. Microarray results were confirmed by qPCR and by an optimized in situ hybridization protocol. Functional impact of overexpression of a dysregulated miRNA was assessed in migration and 3D spheroid assays. Sample-matched transcriptome data was generated to support the identification of disease relevant miRNA targets. Results Several miRNAs were identified as dysregulated in cSCCs as compared to controls. These included the metastasis-linked miR-10b, which was significantly upregulated in primary cell cultures and in archival biopsies. At the functional level, overexpression of miR-10b conferred the stem cell-characteristic of 3D-spheroid formation capacity to keratinocytes, and impaired their mobility. Analysis of miR-10b downstream effects identified a novel putative target of miR-10b, the actin- and tubulin cytoskeleton-associated protein DIAPH2. Conclusion The discovery that miR-10b confers an aspect of cancer stemness – that of enhanced tumor cell adhesion, known to facilitate metastatic colonization - provides an important avenue for future development of novel therapies targeting this metastasis-linked miRNA.

3 Background 57 While the incidence of cutaneous squamous cell carcinomas (cSCC) in Western industrialized societies has 58 increased in recent years, it has remained relatively easy to treat, especially at early stages of the disease 59 (1). However, in some cases cSCCs metastasize, worsening the prognosis of patients. A special case of 60 cSCC occurs in patients suffering from the rare genodermatosis recessive dystrophic epidermolysis bullosa 61 (RDEB). These patients are at high risk of developing a particularly aggressive type of cSCC with a high 62 metastatic potential that is linked to changes in the extracellular matrix (ECM), caused by loss-of-function 63 mutations in the COL7A1 gene (2). A lack of functional type-VII collagen (C7) at the dermal-epidermal

74
We focused on post-transcriptional regulatory processes in aggressive cSCCs, in particular on micro-RNAs 75 (miRNAs). miRNAs are short (20 -25 nucleotide) RNA molecules, which are key regulators of normal cell 76 functions. In a healthy system, miRNAs are predicted to mediate the post-transcriptional control of up to 77 60% of all expressed genes (10). Their dysregulation is associated with several pathologic states, including 78 cancer, heart disease, and obesity, and they are attributed a promising potential for therapeutic 79 developments (11,12). In recent years, both, oncogenic miRNAs (onco-miRs) and tumor suppressive 80 miRNAs, have been identified as playing important roles in cancer progression. In addition, a class of 81 miRNAs have been shown to have specific pro-metastatic properties. A key metasta-miR, miR-10b, has 82 been attributed with tumor promoting properties, as well as the growth of metastatic foci in breast cancer in 83 4 various landmark studies (13-15). MiR-10b is encoded by a highly conserved genomic region, which is 84 located near the homeobox D (HOXD) cluster on chromosome 2 (16). It has been linked to a range of 85 functions, including regulation of angiogenesis and promotion of cell invasion (17)(18)(19)(20). In addition, increased 86 serum levels of miR-10b are associated with poor prognosis in melanoma and breast cancer (21,22), and 87 intravenous injections of miR-10b inhibitors in tumor-bearing mice dramatically reduced breast cancer cell 88 metastasis (13,15,23). In addition, a meta-analysis of miR-10b levels and clinical outcomes in various 89 cancers, demonstrated that overexpression was associated with poor overall survival, indicating that miR-90 10b might be a promising prognostic biomarker (24).

91
In this study, we substantially add to our understanding of the role of miR-10b, by reporting for the first time 92 on the dysregulation of this miRNA in aggressive cSCC. Both, RDEB-and otherwise healthy donor-related 93 cSCCs (HC-cSCCs) were classified and included in this study according to their potential to metastasize.

94
We further show that miR-10b expression is linked to a cancer stem cell-like phenotype in a 3D organotypic 95 model. Taken as a whole, this work provides a new explanation of maligancy in cSCCs, and a novel target 96 for further development of markers and therapies to treat cSCCs.

153
Imaging was performed on a LSM800 Airyscan confocal microscope (Zeiss). Analysis was conducted in

204
In parallel to every migration assay, proliferation of each cell type was measured in 24-well plates, to exclude 205 varying proliferation rates as confounding factor.      Figure 1A). Even though the miR-10 family was found 265 to be overall upregulated in cSCCs, in-depth analysis of major contributors driving the unsupervised cluster 266 separation highlighted miR-10b, which was 2.2-fold (p < 0.05) upregulated in RDEB-SCCs and 2.2-fold (p 267 < 0.05) in HC-cSCCs, respectively, compared to their non-malignant controls ( Figure 1B,C). As opposed to 268 miR-10b, miR-10a was only found to be slightly upregulated by qPCR. As miR-10a appeared to be the most 269 deregulated miRNA in RDEB-SCCs in microarray analysis but not in subsequent qPCR, we assume that 270 high microarray scores most likely derived from a certain hybridization error rate (miRs-10a and -10b differ 271 in only one nucleotide) ( Figure 1E, Supplementary Fig. S1C-E). In addition, increased miR-10b (9-fold) , but 272 not miR-10a, levels in RDEB-SCC were found in previously generated RNA sequencing (RNA-seq) data,

273
where immortalized HC-KC lines were used as controls (Supplementary Fig. S1F,G). Thus, we focused on 274 miR-10b, and dropped miR-10a from further experiments. Predicted miR-10b targets were further tested for 275 gene set enrichment in cancer hallmarks (Molecular Signature Database v6.2), and showed significant 276 association with metastatic processes like epithelial to mesenchymal transition (EMT) ( Figure 1D).

11
In summary, the results of a microarray-based miRNA expression profiling revealed a deregulated 278 miRNome able to distinguish cSCC from keratinocytes. In particular, miR-10b was significantly upregulated 279 in cSCC.

293
To investigate whether miR-10b expression was correlated to differentiation, cells were incubated in the 294 presence of calcium and serum in order to induce differentiation, and respective marker gene expression 295 was analyzed in parallel to miR-10b expression levels. However, no overall correlation between 296 differentiation and miR-10b expression was observed (data not shown).

297
Taken together, results of the IHC-ISH support a potential link between miR-10b expression levels and 298 aggressive RDEB-SCC.  at distant niches (31). In order to investigate the biological role of miR-10b, we analyzed its expression 304 levels in experimental cell lines using IHC-ISH, in parallel to 3D tumor spheroid formation assays. As the 12 cytoplasm is considered as the site of miRNA maturation and action, fluorescence intensity, excluding the 306 nuclear region, was assessed at a single cell resolution. MiR-10b was confirmed to be highly expressed in 307 RDEB-SCC cell lines using ISH probes specific for mature miR-10b, and in two out of three HC-cSCC 308 cultures, as compared to primary HC-KCs (Figure 2A,B, Supplementary Fig. S3,4). Notably, we observed 309 that miR-10b levels in cSCCs varied between individual cell lines, however, in correspondance to phenotypic 310 features linked to increased malignancy, like an EMT-associated spindle-like phenotype.

321
In order to further substantiate the link between miR-10b overexpression and enhanced spheroid formation 322 capacities, we next knocked-out the MIR10B gene locus in RDEB-SCC1 cells (RDEB-SCC1 MIR10B-/-) using 323 the CRISPR-Cas9 technology, and performed minimal dilution experiments to potentially obtain single 324 clones ( Supplementary Fig. S5J,K). In spheroid formation assays we observed, that miR-10b knock-out 325 reduced the stability of aggregates and resulted in an increased number of single cells and fragmented 326 aggregates ( Figure 3A-C). While PCR-mediated confirmation of MIR10B knock-out showed only bands 327 corresponding to successful deletion, we found that over time single cells that had escaped knock-out and 328 subsequent clearance by minimal dilution returned to dominance, independent of a potential proliferative 329 advantage ( Figure 3D,E). This was observed in several clones and over several cultivation passages,

334
Spheroids adhered to dishes, and circularly outgrowing cells became visible after 24 hrs in RDEB-SCC1 335 derived aggregates, and to a much lower extent in respective miR-10b knock-out cells. Again, this was 336 reversed in mixed culture experiments ( Figure 3F). This outgrowth pattern was also observed in two out of 337 three HC-cSCC derived spheroid assays ( Figure 3G).

338
As spheroid formation is a key attribute of cancer stem cells (CSCs), and points towards the presence of

344
We next examined whether the above observed properties were attributed to mobility-, or adhesion-345 associated mechanisms (32). Therefore, we conducted wound closure assays using miR-10b

352
Taken together, our data suggest for the first time that CSC-like properties can be conferred by miR-10b, a 353 heretofore unknown aspect of miR-10b driven malignancy in cSCCs.

356
To analyze the impact of miR-10b on its targetome, we first screened the scientific literature in an automated 357 text-mining approach for miR-10b in cancer, which highlighted the previously reported, direct miR-10b target 358 transcription factor HOXD10 (14, Supplementary Fig. S7A). We observed significantly reduced levels of 359 HOXD10 protein in three RDEB-SCC cell lines compared to HC-KC, and also in an RDEB-SCC tissue 360 section, compared to HC-and RDEB-skin (34, Supplementary Fig. S7B-D).

14
To next identify novel downstream targets of miR-10b by the example of RDEB, data driven miR-10b target 362 identification was implemented based on transcriptome data generated from miRNA microarray-matched 363 RDEB-SCC and RDEB-KC samples. Of the 576 differentially expressed genes identified (≥ 2-fold ↓↑) in 364 RDEB-SCCs, 114 were reported in merged repository data (n = 3,923) of validated miR-10 targets 365 (miRTarbase v6.1), as well as computationally predicted targets by seed sequence and evolutionary 366 conservation (TargetScan v7.2). Dysregulated putative miR-10 interaction partners were further prioritized 367 by strength of their inverse correlation of expression with miR-10 signal (Supplementary Fig. S7E). To 368 nominate a disease relevant miR-10b target, we analyzed publicly available survival data from metastatic 369 stage IV head and neck squamous cell carcinoma (HNSCC) (n = 86 patients, The Cancer Genome Atlas / 370 TCGA), as this cancer type was previously described to have high genetic similarities to RDEB-SCC (8).

371
The top 20 candidate miR-10b targets with highest inverse correlation were then used to stratify HNSCC 372 patients ( Supplementary Fig. S8A and Supplementary Table S6). We nominated diaphanous related formin 373 2 (DIAPH2) for further evaluation based on significant differences in Kaplan Meier survival curves (p < 0.05, 374 log-rank test, Figure 5A), together with the fact that it was listed as a putative target of miR-10b. Its potential 375 disease-relevance was substantiated by its recent association with colon carcinoma, and its potential role

383
To confirm DIAPH2 as a direct target of miR-10b, a dual luciferase reporter assay was established by cloning 384 the 3'UTRs of DIAPH2 and HOXD10 as a control, respectively, downstream of a firefly luciferase reporter 385 gene. Constructs were then co-transfected with a miR-10b mimic into HC-KCs, which express only low 386 levels of endogenous miR-10b. In the presence of miR-10b mimic, luciferase signal was significantly 387 reduced (p-value DIAPH2 = 0.015; p-value HOXD10 < 0.01) compared to scrambled (SCR) control (Supplementary 388 Fig. S7F,G).

15
We next assessed DIAPH2 expression in cultured RDEB-and control KC. Both, at the mRNA and the 390 protein level, DIAPH2 was downregulated in RDEB-SCC cells (Supplementary Fig. S7I, Figure 5B,C).

394
To examine if a loss of DIAPH2 in keratinocytes phenocopies the observed migratory behavior in KC miR-10 ,

436
Alternatively, changes in the actin-and / or microtubule cytoskeleton, which impact cell motility (42), could 437 account for slower wound closure rates and increased stickiness in RDEB skin. In this context, the putative 438 miR-10b target DIAPH2, a protein belonging to the formin homology family, which has previously been 439 associated with actin filament assembly, microtubule formation and vesicle shuttling, might be a modulator 440 of cell motility in RDEB-SCCs (43,44). In this study, DIAPH2 was predicted to be a target of miR-10b in 441 silico, which we additionally confirmed using a luciferase-reporter assays. Notably, while miR-10b

459
In summary, overexpression of miR-10b impacts cellular processes at various levels. The extent of the 460 response is dependent on the cellular context, the presence of certain target mRNAs, and/or mutations in 461 those target mRNAs (49). In the context of cSCCs, we showed that miR-10b confers anchorage-462 independent spheroid formation capacities, indicating a phenotypic shift towards stem cell-like properties.

463
We hypothesize, that these findings might be involved in tumor progression at the stages of extravasation 464 and metastatic colonization, processes in which both anchorage-free survival, as well as cell surface 465 interaction, are important pre-requisites.

468
Overall, our results demonstrate for the first time the upregulation of miR-10b in aggressive RDEB-SCCs.

469
While new, these findings provide a potential so far unreported pro-metastatic function of miR-10b, as this