Zebrafish rhabdomyosarcoma reflects the developmental stage of oncogene expression during myogenesis

Abstract

Rhabdomyosarcoma is a pediatric malignancy thought to arise from the uncontrolled proliferation of myogenic cells. Here, we have generated models of rhabdomyosarcoma in the zebrafish by inducing oncogenic KRAS(G12D) expression at different stages during muscle development. Several zebrafish promoters were used, including the cdh15 and rag2 promoters, which drive gene expression in early muscle progenitors, and the mylz2 promoter, which is expressed in differentiating myoblasts. The tumors that developed differed in their ability to recapitulate normal myogenesis. cdh15:KRAS(G12D) and rag2:KRAS(G12D) fish developed tumors that displayed an inability to complete muscle differentiation as determined by histological appearance and gene expression analyses. By contrast, mylz2:KRAS(G12D) tumors more closely resembled mature skeletal muscle and were most similar to well-differentiated human rhabdomyosarcoma in terms of gene expression. mylz2:KRAS(G12D) fish showed significantly improved survival compared with cdh15:KRAS(G12D) and rag2:KRAS(G12D) fish. Tumor-propagating activity was enriched in myf5-expressing cell populations within all of the tumor types. Our results demonstrate that oncogenic KRAS(G12D) expression at different stages during muscle development has profound effects on the ability of tumor cells to recapitulate normal myogenesis, altering the tumorigenic capability of these cells.

Keywords: Muscle; Rhabdomyosarcoma; Zebrafish.

Fig. 1.

cdh15 and mylz2 transgenic muscle promoters recapitulate endogenous expression patterns during development and turn on at different times during somite formation. Whole-mount RNA in situ hybridization was performed on endogenous genes in AB strain wild-type embryos (A-H,M-P), or the GFP (I-L) or mCherry (Q-T) transcript in transgenic embryos. dpf, days post-fertilization; hpf, hours post-fertilization; som, somite stage. Anterior is on the left. Arrowheads denote the location of the most posteriorly stained somite.

Fig. 2.

Zebrafish injected with rag2:KRAS^G12D, cdh15:KRAS^G12D or mylz2:KRAS^G12D develop muscle tumors that differ in histological appearance. (A-C) Brightfield images of fish bearing rag2 (A), cdh15 (B) and mylz2 (C) tumors. (D-G) H&E staining of tumors. (H-W) Tumors express diagnostic markers of RMS as assessed by RNA in situ hybridization. Histological sections of representative rag2 (D,H,L,P,T), cdh15 (E,I,M,Q,U) and mylz2 (F,J,N,R,V) tumors and normal muscle (G,K,O,S,W). Images in D-W are at the same magnification. (X) Tumors were assigned differentiation scores based on histological appearance. Scale bars: in A-C, 2 mm; in D-G, 50 μm.

Fig. 3.

rag2:KRAS^G12D, cdh15:KRAS^G12D and mylz2:KRAS^G12D tumors express myogenic markers at differing levels. Gene expression levels as determined by quantitative RT-PCR for the pax7b (A), myf5 (B), cdh15 (C), myog (D), desmin (E) and mylz2 (F) genes or the KRAS (G) transgene. Asterisks indicate statistically significant differences between groups (P<0.05, Student’s t-test). n=7 for each of the tumor models, and n=3 for normal muscle (AB strain). Error bars represent s.e.m.

Fig. 4.

Survival of tumor-bearing fish differs in rag2, cdh15 and mylz2 tumors. (A) Fish injected with rag2:KRAS^G12D, cdh15:KRAS^G12D or mylz2:KRAS^G12D transgenes developed tumors with differing latency and incidence (mylz2, n=19/117; rag2, n=50/144; cdh15, n=42/146 fish developed tumors; rag2 versus mylz2, P=0.0004; cdh15 versus mylz2, P<0.0001; rag2 versus cdh15, P=0.73). (B) mylz2 tumor-bearing fish displayed better survival over time compared with rag2 and cdh15 tumor-bearing fish (rag2, n=50; cdh15, n=48; mylz2, n=22; rag2 versus mylz2, P=0.001; cdh15 versus mylz2, P=0.009; rag2 versus cdh15, P=0.72). P-values for A and B were calculated using the log-rank test. (C-L) Rates of EdU incorporation did not significantly differ among the tumor models. Fluorescent microscopic images depicting DAPI-positive (C,F,I) and EdU-positive (D,G,J) nuclei. (E,H,K) Merged images. (C-E) rag2 tumor. (F-H) cdh15 tumor. (I-K) mylz2 tumor. (L) Quantification of the ratio of EdU-positive to total nuclei. n=4 for each tumor model; P>0.05, Student’s t-test. Ratios were calculated based on the average ratio for three separate high-power images per tumor. Error bars represent s.e.m.

Fig. 5.

Microarray analysis of zebrafish RMS driven by the rag2, cdh15 or mylz2 promoters. (A) Unsupervised hierarchical clustering of wild-type muscle (AB strain) or each of the three zebrafish RMS subtypes. The mylz2 tumors are transcriptionally the most distinct, whereas the rag2 and cdh15 tumors are more similar to each other. (B) Venn diagrams showing overlap of dysregulated genes in each RMS subtype. Each tumor was compared with wild-type muscle, and gene numbers represent the sum of both up- and downregulated genes. For breakdown of up- and downregulated genes separately, see supplementary material Table S1. (C,D) GSEA plots showing enrichment of the mylz2 signature (compared with rag2 or cdh15) in human well-differentiated (WD) RMS. For expanded GSEA analysis, see supplementary material Table S1.

Fig. 6.

Tumors in the myf5:GFP;mylz2:mCherry background differ in the degree of GFP fluorescence. (A-L) Brightfield (A,D,G,J), GFP (B,E,H,K) and mCherry (C,F,I,L) fluorescent images of uninjected fish (A-C) or tumor-bearing fish that had been injected with the rag2:KRAS^G12D (D-F), cdh15:KRAS^G12D (G-I) or mylz2:KRAS^G12D (J-L) transgenes at the one-cell stage. (M-U) Engrafted, secondary tumors in irradiated casper recipient fish display similar patterns of fluorescence to their respective primary, parent tumors. Brightfield (M,P,S), GFP fluorescence (N,Q,T), and mCherry fluorescence (O,R,U) images of casper fish transplanted with G+R- cells from a rag2 tumor (M-O), cdh15 tumor (P-R) or mylz2 tumor (S-U).

Fig. 7.

Tumors in the myf5:GFP;mylz2:mCherry background have differing proportions of GFP- and mCherry-positive cells. (A-O) FACS profiles of a primary rag2 (A), cdh15 (F) and mylz2 (K) tumor. G+R- (B,G,L), G+R+ (C,H,M), G-R+ (D,I,N) and G-R- (E,J,O) populations could be isolated with relative purity in rag2 (B-E), cdh15 (G-J) and mylz2 (L-O) tumors.