Fig 1.
Study design and tumor development.
A) Mastomys coucha as a model for cutaneous papillomavirus infection. In the study, naturally MnPV-infected animals (MnPV+) as well as virus-free control animals (MnPV-) were irradiated three times per week with UVB. The starting dose of 150 mJ/cm2 was increased weekly by 50 mJ/cm2 until the desired final dose was reached (450, 600 or 800 mJ/cm2, respectively). Black arrows indicate an increase of the dose, gray arrows the subsequent application of this dose. The irradiation was continued until the animals were sacrificed or died. B) Kaplan-Meier curves demonstrating the percentage of irradiated virus-infected (MnPV+, UV+), virus-free (MnPV-, UV+) and unirradiated virus-infected (MnPV+, UV-) tumor-bearing animals. C) Two examples of spontaneous skin lesions arising in naturally infected animals. D) Examples of UV-induced keratinizing SCCs (KSCC) with similarities to human keratoacanthomas. E) Examples of UV-induced non-keratinizing SCCs (nKSCC) (C, D and E: scale bars: 10 mm). F) Number of KSCCs and nKSCCs in correlation with the final UV doses. Note that KSCCs preferentially appeared at the lowest dose, nKSCCs preferentially at higher doses (Mean ± SEM; animal numbers: see Table 1; av: average number of tumors).
Table 1.
Summary of the absolute numbers of tumors in the different groups, percentage of tumor-bearing animals and median time of tumor development.
Fig 2.
Histological analyses of a non-UV tumor and UV-induced tumors.
A) Tumors from unirradiated sites show papilloma-like growth of well-differentiated neoplastic squamous cells (H&E). Especially the basal layers are hyperproliferative as indicated by strong Ki-67 staining. Throughout all layers of the lesion, neoplastic cells strongly express cytokeratins (pan-Cytokeratin). B) A UV-induced KSCC with well-differentiated exoendophytic proliferations of squamous cells expressing Ki-67 throughout all neoplastic squamous layers. C and D) In some cases well-differentiated KSCCs (C) further developed into more aggressive poorly differentiated nKSCCs (D). Proliferating altered squamous cells thereby invaded deeper layers and often changed to a spindle-like phenotype (H&E). The Ki-67 staining becomes diffuse in this process (compare insets) and cytokeratin expression is reduced. (d: dermis; e: epidermis; f: fat; k: keratin; m: muscle; u: ulceration; t: tumor. Scale bars: macroscopic: 10 mm, overviews: 1 mm, insets 100 μm).
Fig 3.
Collagen IV staining on tissue sections reveals invasion of keratinocytes through the basal membrane (BM).
The BM was stained against collagen IV (green). Nuclei were counterstained with DAPI (blue). Consecutive sections stained for pan-cytokeratin are shown in comparison. A) In normal skin, the BM (white arrows) marks the barrier between epidermis and dermis. B) Early stage carcinoma formation in UV-irradiated skin. A lack of collagen IV expression indicates the disruption of the BM (orange arrows) accompanied by downward migrating cells (black arrow). C) In the edge region of a UV-induced KSCC, the BM is lost and invading altered keratinocytes are detectable. D) In nKSCC, invasion of neoplastic cells is advanced as indicated by pan-cytokeratin staining. The discontinuous staining of the BM marks transition zones where invading neoplastic squamous cells acquire a spindle cell phenotype (Scale bars: 100 μm).
Fig 4.
Molecular analyses of tumor-bearing animals.
A) Viral load in tissue samples from UV-irradiated and control animals from the MnPV-infected colony analyzed by qPCR and normalized to a plasmid standard. Samples were grouped according to their origin as indicated (ctrl skin: skin from unirradiated animals; ui skin/UV skin: unirradiated or UV-irradiated skin from irradiated animals; KSCC/nKSCC: UV-induced SCCs; non-UV tumor: tumors from non-UV sites of irradiated animals and spontaneous tumors from unirradiated animals). UV+/- indicates whether the animal was UV-exposed or not (Kruskal-Wallis test, *p<0.05, ***p<0.001, nsp>0.05). B) Southern blot analysis of unirradiated and UV-irradiated skins, a KSCC and a non-UV tumor. DNA was digested with ApaI (no cleavage site in MnPV), XbaI (one site) or XhoI (two sites) as indicated (Form I: supercoiled; Form II: relaxed circular; Form III: linear form of MnPV). C) Semi-quantitative RT-PCR for the most abundant MnPV E1^E4 transcript in non-UV tumors and UV-induced SCCs or the control GAPDH. D) Semi-quantitative RT-PCR for MnPV E6, E7 and L1 transcripts in non-UV tumors and UV-induced SCCs or the control GAPDH.
Fig 5.
Spatial analysis of viral load in UV-induced SCCs.
Quantification and distribution of MnPV DNA in different microdissected areas (specified in the HE staining) of A) a KSCC and B) a UV-induced nKSCC. Asterisks indicate the position of the respective MnPV-specific in situ hybridization (ISH) (Scale bars: HE: 1 mm, ISH: 100 μm).
Fig 6.
Serological analyses of the animals.
A) Antibody responses of UV-irradiated and control animals against MnPV-L1-VLPs. Final sera of 55 to 75 week old animals were measured. Animals were grouped according to their origin (MnPV+ or MnPV- colonies) and treatment. Different groups represent distinct constellations of tumor types in the animals. Note that both MnPV- tumor-bearing animals were included in the last group. The cut-off for the assay is indicated by the red line (titer of 300) (Mean ± SEM; Kruskal-Wallis test, **p<0.01, ***p<0.001). B) Correlation of pseudovirion-based neutralization titers and antibody titers measured by VLP-ELISA. The non-linear fitting indicates a correlation of 99% between both assays.
Fig 7.
MnPV interferes with DNA damage repair.
A) Repair kinetics of CPDs in MnPV E6/E7-positive and -negative Mastomys keratinocytes (Mean ± SD; n = 2, measurements were performed in quadruplicates). B) Immunofluorescence staining of γH2AX foci in keratinocytes stably expressing MnPV E6/E7. Cells were irradiated with UVB and further incubated prior to detection and quantification of γH2AX foci (Ctrl: unirradiated, UV: irradiated; Red: γH2AX, blue: nuclei; scale bars: 50 μm). C) Quantification of γH2AX foci (Mean ± SEM; n≥242; 1way-ANOVA, *p<0.05, **p<0.01, ***p<0.001). D) Co-detection of CPDs and γH2AX in MnPV+/- skin harvested 24h after UV irradiation. Arrows point towards positive cells (Viral loads: animal 3: 13.68 ± 1.66 copies/cell, animal 4: 147.42 ± 14.62 copies/cell; Scale bars: 100 μm). E) Co-detection of CPDs and γH2AX in a KSCC harvested 24h after UV irradiation (Viral load: 611.88 ± 18.75 copies/cell; scale bars: 100 μm).
Fig 8.
Induction of γH2AX foci by MnPV.
Co-detection of γH2AX and MnPV DNA in consecutive tissue sections. Left panel: γH2AX staining of a non-UV tumor correlates with high viral load detected by ISH with a MnPV-specific probe (Viral load: 12065.07 ± 1119.24 copies/cell). Middle panel: the same concurrence can be detected in UV-induced KSCCs (Viral load: 26592.94 ± 1823.92 copies/cell). Right panel: UV-induced nKSCCs are negative in both stainings (Viral load: 0.98 ± 0.1 copies/cell; scale bars: 100 μm).
Fig 9.
Transactivating capacity of Mastomys p53 in the presence of MnPV E6.
A) The capacity of p53 to transactivate a p53-responsive firefly luciferase gene measured in H1299 cells transfected with reporter plasmids and expression vectors for Mastomys p53 and MnPV E6 or human p53 and HPV16 E6 as a control. Transactivation activity was measured by luminescence (RLU, relative light units). Cells transfected only with p53 served as control and their RLU levels were arbitrarily set to 1 (Mean ± SEM; n = 7; 1way-ANOVA, ***p<0.0001). B) Western blots showing protein levels of p53 and E6 in the lysates of the transactivation assay. Actin served as an internal loading control.
Fig 10.
Analysis of the Trp53 status in UV-induced SCCs.
A) Schematic structure of Mastomys coucha p53 (based on [113]) matched to the position and the frequency of mutated residues (n = 43). Note that residue P271 was substituted by three different amino acids. B) Comparison of the number of Trp53 mutations in KSCCs (n = 16), nKSCCs (n = 27) and the total number of SCCs (n = 43) (unpaired t-test, p = 0.0227). C) Frequency of mutations at positions P145, R266 and P271 in KSCCs in comparison to nKSCCs. D) Locations of the hot-spot mutations in a 3D model of p53. Selected residues are highlighted (source: http://p53.iarc.fr/MakeJMol.aspx). E) Hot-spot mutants were cloned in an expression vector and tested for their capacity to transactivate a p53-responsive firefly luciferase gene. Transactivation activity was measured by luminescence (RLU, relative light units). Cells transfected with empty vector served as control and their RLU levels were arbitrarily set to 1 (Mean ± SEM; n = 7; 1way-ANOVA, ***p<0.001). F) Western blot showing protein levels of p53 mutants measured in the transactivation assay. EGFP was used as a control for transfection efficiency, actin as an internal loading control. G) Same as shown in panel E. Prior to the measurement of the transactivation, transfected cells were treated with 5 μM MG132 (Mean ± SEM; n = 4; 1way-ANOVA, ***p<0.001). H) Western blot of transfected cells after treatment with 5 μM MG132.
Fig 11.
Dedifferentiation correlates with positive p53 staining.
Consecutive sections of a poorly differentiated nKSCC were stained with antibodies against E-cadherin, vimentin, pan-Cytokeratin and p53. DAPI was used as nuclear counter stain. Note that in this tumor, only mutation R266C could be detected (Scale bars: 100 μm).
Fig 12.
Schematic overview of the mechanism suggested for UV-induced NMSC development in Mastomys coucha.
A) MnPV infects basal epithelial cells of the skin of young animals via small injuries. B) MnPV genome is amplified in stratified skin layers (pink and red nuclei) and new virions are released. C) UVB irradiation of the skin. D) UVB-irradiated skin is hyperproliferative, favoring viral replication and virion formation. UVB-induced photoproducts, e.g. in Trp53, occur in keratinocytes (altered nuclei). In uninfected cells, damages are repaired. In infected cells, MnPV-E6/E7 reduce chromosomal stability and inhibit DNA repair. Mutations can accumulate and altered cells become neoplastic. E) Neoplastic squamous cells (light blue) start forming a well-differentiated keratinizing SCC, still representing a permissive system that allows viral replication and formation of virions. F) When neoplastic squamous cells accumulate further mutations (dark blue), a spindle cell phenotype is acquired, forming a poorly differentiated SCC that may become ulcerated. MnPV cannot replicate in dedifferentiated cells and the viral DNA is subsequently lost.