Minimal contribution of ERK1/2-MAPK signalling towards the maintenance of oncogenic GNAQQ209P-driven uveal melanomas in zebrafish

Abstract

Mutations affecting Gαq proteins are pervasive in uveal melanoma (UM), suggesting they 'drive' UM pathogenesis. The ERK1/2-MAPK pathway is critical for cutaneous melanoma development and consequently an important therapeutic target. Defining the contribution of ERK1/2-MAPK signalling to UM development has been hampered by the lack of an informative animal model that spontaneously develops UM. Towards this end, we engineered transgenic zebrafish to express oncogenic GNAQQ209P in the melanocyte lineage. This resulted in hyperplasia of uveal melanocytes, but with no evidence of malignant progression, nor perturbation of skin melanocytes. Combining expression of oncogenic GNAQQ209P with p53 inactivation resulted in earlier onset and even more extensive hyperplasia of uveal melanocytes that progressed to UM. Immunohistochemistry revealed only weak immunoreactivity to phosphorylated (p)ERK1/2 in established uveal tumours-in contrast to strong immunoreactivity in oncogenic RAS-driven skin lesions-but ubiquitous positive staining for nuclear Yes-associated protein (YAP). Moreover, no changes were observed in pERK1/2 levels upon transient knockdown of GNAQ or phospholipase C-beta (PLC-β) inhibition in the majority of human UM cell lines we tested harbouring GNAQ mutations. In summary, our findings demonstrate a weak correlation between oncogenic GNAQQ209P mutation and sustained ERK1/2-MAPK activation, implying that ERK1/2 signalling is unlikely to be instrumental in the maintenance of GNAQQ209P-driven UMs.

Keywords: ERK1/2; GNAQ; p53; uveal melanoma; zebrafish.

Figure 1. Only oncogenic GNAQ^Q209P is sufficient to induce choroidal melanocyte hyperplasia

A. Schematic representation of elements in the Tol2-based transposon vector driving the expression of oncogenic GNAQ^Q209P under the control of zebrafish mitfa promoter in the melanocyte lineage and Venus fluorescent reporter under the control of cryaa promoter in the eye lens. Abbreviations: ITR, inverted terminal repeat. B. Example of a 5 dpf transgenic zebrafish embryo with a fluorescent eye lens. Scale bar, 100 μm. C. RT-qPCR data showing a 4.1 fold increase in GNAQ expression in the melanocytes of 2-month-old F₁ Tg (mitfa:GNAQ^Q209P) zebrafish, as compared to non-injected controls. Data represents mean ± SEM of triplicates of three independent experiments. *** P <0.05 using two-tailed, unpaired t test. D, E, F, G. H&E staining of transverse sections of formalin-fixed and paraffin-embedded eye specimens of control wild-type, Tg (mitfa:GNAQ^Q209P), Tg (mitfa:NRAS^Q61L), and Tg (mitfa:BRAF^V600E) zebrafish, respectively. Choroidal hyperplasia observed in the thickened choroid (E; black arrowhead) was only observed in transgenic animals expressing oncogenic GNAQ^Q209P. In contrast, as compared to control wild-type H. and Tg (mitfa:GNAQ^Q209P) I. hyperplasia of cutaneous melanocytes (black arrowhead) was only detected in transverse sections of the torso region of Tg (mitfa:NRAS^Q61L) J. and Tg (mitfa:BRAF^V600E) K. zebrafish. Abbreviations: RPE, retinal pigmented epithelium. Scale bars, 20 μm.

Figure 2. Oncogenic GNAQ^Q209P-mediated activation of ERK and YAP signalling in choroidal melanocytes at the junction between RPE and choroid

A-D. Transverse sections of formalin-fixed and paraffin-embedded eye tissues of F₁ generation, 5-month-old Tg (mitfa:GNAQ^Q209P) zebrafish. (A) H&E staining demonstrating choroidal hyperplasia (black arrowheads). White dashed box indicates the region of the choroid magnified in B-D. (B-D) Transverse sections of formalin-fixed and paraffin-embedded eye tissues were stained by IHC, visualized by ImmPact NovaRed peroxidase (HRP) substrate then counterstained with hematoxylin (blue). (B) Negative control: section incubated with 1x PBS instead of primary antibody. (C) Immunoreactivity to pERK1/2 (read-out of ERK activation; black arrowheads) in melanocytes at the interface between the RPE and choroid. (D) YAP-positive nuclei (read-out of YAP activation; black arrowheads) in the same cells. Scale bars, 20 μm.

Figure 3. GNAQ^Q209P rescues the hypopigmentation phenotype of golden (slc24a5^−/−) zebrafish

A.I, B.I, C.I. Lateral views of wild-type, golden, and Tg (mitfa:GNAQ^Q209P;slc24a5^−/−) 5 dpf zebrafish embryos, respectively. A.II, B.II, C.II. Top views of wild-type, golden, and Tg (mitfa:GNAQ^Q209P;slc24a5^−/−) 5 dpf zebrafish embryos, respectively. Scale bars, 100 μm. D.I, E.I, F.I, G.I. Lateral views of 2-month-old adult zebrafish. (D.I) Wild-type illustrating normal melanin pigmentation in skin melanocytes. (E.I) Light-skinned golden mutant (slc24a5^−/−) with stripes comprising less melanin-rich melanocytes. (F.I, G.I) Two examples of Tg (mitfa:GNAQ^Q209P;slc24a5^−/−) zebrafish from independent founders, acquiring darker stripes as compared to wild-type (D.I) and golden mutants (E.I), but still conserving the normal striped pigment pattern. Scale bars, 0.5 cm. D.II, E.II, F.II, G.II. H&E staining of transverse sections of formalin-fixed and paraffin-embedded eye specimens of 2-month-old adult zebrafish. (D.II) Normally pigmented RPE located between the photoreceptors layer of the retina (rods & cones) and the choroid in a wild-type zebrafish. (E.II) Lightly pigmented RPE in a golden mutant. (F.II, G.II) Rescuing of hypopigmented RPE in both GNAQ^Q209P-expressing golden mutants, with no pathological findings yet observed in the choroidal melanocyte layer. Scale bars, 50 μm. H.I, I.I, J.I, K.I. Tail fins of wild-type, golden, and two Tg (mitfa:GNAQ^Q209P;slc24a5^−/−) zebrafish, respectively. These were dissected, fixed, and sectioned for TEM analysis. Black dashed boxes indicate representative regions of the tail fins examined by TEM. Scale bars, 0.5 cm. H.II, I.II, J.II, K.II. Transmission electron micrographs illustrating the structure of melanosomes in the melanocytes of dissected tail fins. Melanosomes in golden mutants (I.II) are less densely pigmented as compared to those of wild-type (H.II) and transgenic zebrafish (J.II, K.II). Scale bars, 200 nm. L. Relative to zebrafish EF1alpha (EF1α) expression and as compared to non-injected golden controls, qRT-PCR data show significant increase in mRNA expression levels of GNAQ in Tg (mitfa:GNAQ^Q209P;slc24a5^−/−) zebrafish. In contrast, no significant alterations were observed in the expression of mitfa or downstream differentiation genes including tyrosinase, tyrp-1, and dct. Results represent mean ± SEM of triplicates of three independent experiments. ***P <0.05, using two-tailed, unpaired t test. Abbreviations: tyrp-1, tyrosinase-related protein 1; dct, dopachrome tautomerase.

Figure 4. Co-operation of oncogenic GNAQ^Q209P and p53 loss-of-function results in UM development

Oncogenic GNAQ^Q209P Tol2 construct was injected into p53-deficient (p53^M214K/M214K) zygotes, animals were sacrificed at 2 and 5 months or sooner if ocular protrusion was prominent, then fixed eye specimens were processed and paraffin-embedded for transverse sectioning at a thickness of 5 μm, followed by H&E staining. Representative images are shown. A.I. A non-injected p53-deficient (p53^M214K/M214K) control zebrafish illustrating a single layer of choroidal melanocytes. B.I. Benign choroidal hyperplasia in an 8-week-old Tg (mitfa:GNAQ^Q209P;p53^M214K/M214K) zebrafish. Note the diffuse thickening of the entire choroid caused by the proliferation of choroidal melanocytes. Nuclear and plexiform layers of the retina, photoreceptors layer, and RPE are structurally normal. Also, sclera is respected, with no evidence of infiltration or perforation. C.I. An example of malignancy, with non-pigmented hyperproliferative atypical cells developing within an area that shows evidence of choroidal hyperplasia in a 20-week-old Tg (mitfa:GNAQ^Q209P;p53^M214K/M214K) zebrafish. D.I. A second example of a 20-week-old Tg (mitfa:GNAQ^Q209P;p53^M214K/M214K) zebrafish developing an eye malignancy infiltrating the ciliary body, iris, and cornea, also with evidence of choroidal hyperproliferation. A.II, B.II, C.II, D.II. Magnifications of the regions within the white dashed boxes in A.I, B.I, C.I, D.I, respectively, illustrating the changes within the choroidal melanocyte layer (B.II, C.II, D.II), as compared to a structurally normal choroid in a non-injected control (A.II). (C.III, D.III) Magnifications of transformed uveal melanocytes depicted in the black dashed boxes in C.I and D.I, respectively. Abbreviations: RPE, retinal pigmented epithelium. Scale bar lengths, as indicated.

Figure 5. Validation of GNAQ^Q209P-driven choroidal hyperplasia and uveal tumours in Tg (mitfa:GNAQ^Q209P;p53^M214K/M214K) zebrafish

Transverse sections of formalin-fixed and paraffin-embedded zebrafish eye tissues were stained by IHC, visualized by ImmPact NovaRed peroxidase (HRP) substrate then counterstained with hematoxylin (blue). Representative H&E and IHC images are shown for the pre-malignant A.I-A.V. and malignant B.I-B.V. stages. (A.I) H&E staining illustrating a 5x magnification of a benign choroidal hyperplasia in a 2-month-old Tg (mitfa:GNAQ^Q209P;p53^M214K/M214K) zebrafish. (A.II-A.V) Magnifications of the region depicted in the black dashed box in A.I. (A.II) H&E staining of a melanin-bleached section, revealing multiple cellular layers of the thickened choroid. (A.III) Negative control section for IHC staining. (A.IV) GNAQ expression in thickened choroid. (A.V) Hyperplastic choroidal melanocytes are positive for PCNA (dark brown nuclei), especially at the junctional zone between RPE and choroid. (B.I) H&E staining illustrating a 5x magnification of a uveal tumour developing in a 5-month-old Tg (mitfa:GNAQ^Q209P;p53^M214K/M214K) zebrafish. (B.II) Negative control section for IHC staining. (B.III) Malignant cells expressing GNAQ. (B.IV) Positive immunoreactivity to the melanocytic differentiation marker tyrosinase, indicating the melanocytic origin of malignant cells. (B.V) Numerous PCNA-positive nuclei (dark brown) in malignant melanocytes. Abbreviations: Tyr, tyrosinase; PCNA, proliferating cell nuclear antigen. Scale bars, 100 μm (A.I, B.I); 20 μm (A.II-A.V; B.II-B.V).

Figure 6. Sporadic ERK activation contrasted with ubiquitous nuclear YAP in malignancies in Tg (mitfa:GNAQ^Q209P;p53^M214K/M214K) zebrafish

Sections of formalin-fixed and paraffin-embedded zebrafish eye specimens were stained for GNAQ, pERK1/2, and YAP by IHC, visualized by ImmPact NovaRed peroxidase (HRP) substrate then counterstained with hematoxylin (purple nuclei). A.I-A.III. Representative IHC images of benign hyperproliferative choroid. (A.I) Uniform expression of GNAQ in hyperplastic choroidal melanocytes. (A.II) Only a few cells (black arrowheads) are immunoreactive to pERK1/2. (A.III) Comparatively, more cells displayed nuclear YAP (brown nuclei), although again, this was more noticeable for cells residing at the interface. B.I-B.III; C.I-C.III. Representative IHC images of malignant choroidal melanocytes in two independent GNAQ^Q209P-driven uveal tumours. (B.I, C.I) Transformed melanocytes expressing GNAQ. (B.II, C.II) Malignant cells showing only sporadic immunoreactivity to pERK1/2 (cells showing positive immunoreactivity are indicated with black arrowheads). (B.III, C.III) Ubiquitous YAP nuclear localization (brown nuclei) in transformed uveal melanocytes. Images are representative of sections from three animals. Abbreviations: RPE, retinal pigmented epithelium; RBC, red blood cells. Scale bars, 20 μm.

Figure 7. ERK1/2 activation in human UM cell lines harbouring GNAQ mutations is predominantly independent of GNAQ signalling

A. HEK 293 cells were transfected with vector alone (v) or plasmids encoding Myc-DDK tagged wild-type GNAQ or mutagenized GNAQ^Q209P/L. 48 hours later, expression of constructs was validated by immunoblotting whole cell lysates with anti-MycDDK antibody. Lysates were also probed for pERK1/2and total ERK1/2, revealing increased pERK1/2 levels upon expression of mutant, but not wild-type GNAQ or empty vector. GAPDH was used as a loading control. B. The indicated UM cell lines were transfected with control scrambled (Sc) siRNA or pre-designed and validated siRNA targeting human GNAQ. 48 hours post-transfection, cells were lysed and immunoblotting was performed for endogenous GNAQ, pERK1/2, and total ERK1/2. Knockdown of GNAQ in GNAQ^Q209P-expressing UM cell lines (Mel270, OMM1.3, and OMM1.5), GNAQ^Q209L-expressing UM cell line (Mel202), as well as OCM3 and OCM8 lines (expressing BRAF^V600E; negative controls) was not associated with any alterations in pERK1/2 levels, as compared to cell lines transfected with Sc siRNA. For 92.1 only (expressing GNAQ^Q209L), pERK1/2 levels were reduced upon GNAQ knockdown. C. As detected by pERK1/2 sandwich ELISA, knockdown of GNAQ in 92.1 cell line was associated with a significant decrease in pERK1/2 levels, but didn't affect pERK1/2 levels in Mel270 or OMM1.3. The corresponding immunoblot validates knockdown of GNAQ in the tested UM cell lines, as well as total ERK1/2 levels. Results represent the mean of absorbance readings at 450 nm ± SEM of three technical replicates. Data were analyzed for statistical significance using independent sample t-test at P <0.05. D. Treatment of the indicated UM cell lines withU-73122 (1μM), a PLC inhibitor, abolished ERK activation in 92.1 only, whereas the other cell lines showed steady levels of pERK1/2. Abbreviations: Sc, scrambled; Ctrl, control.