A rapid one-generation genetic screen in a Drosophila model to capture rhabdomyosarcoma effectors and therapeutic targets.

Rhabdomyosarcoma (RMS) is an aggressive childhood malignancy of neoplastic muscle-lineage precursors that fail to terminally differentiate into syncytial muscle. The most aggressive form of RMS, alveolar-RMS, is driven by misexpression of the PAX-FOXO1 oncoprotein, which is generated by recurrent chromosomal translocations that fuse either the PAX3 or PAX7 gene to FOXO1. The molecular underpinnings of PAX-FOXO1-mediated RMS pathogenesis remain unclear, however, and clinical outcomes poor. Here, we report a new approach to dissect RMS, exploiting a highly efficient Drosophila PAX7-FOXO1 model uniquely configured to uncover PAX-FOXO1 RMS genetic effectors in only one generation. With this system, we have performed a comprehensive deletion screen against the Drosophila autosomes and demonstrate that mutation of Mef2, a myogenesis lynchpin in both flies and mammals, dominantly suppresses PAX7-FOXO1 pathogenicity and acts as a PAX7-FOXO1 gene target. Additionally, we reveal that mutation of mastermind, a gene encoding a MEF2 transcriptional coactivator, similarly suppresses PAX7-FOXO1, further pointing toward MEF2 transcriptional activity as a PAX-FOXO1 underpinning. These studies show the utility of the PAX-FOXO1 Drosophila system as a robust one-generation (F1) RMS gene discovery platform and demonstrate how Drosophila transgenic conditional expression models can be configured for the rapid dissection of human disease.

Childhood cancer differs biologically from adult neoplasia: whereas most solid adult tumors are epithelial carcinomas, solid childhood malignancies often are mesenchymal sarcomas. Soft-tissue sarcomas account for 10% of all childhood malignancies, 50% of which are skeletal muscle-lineage rhabdomyosarcomas (RMS) (Gurney ; Scheurer ). Despite aggressive therapies, children with high-risk RMS suffer from a 3-yr event-free survival of 20%. Treatments for high-risk RMS have not improved for three decades, underscoring the need to elucidate the molecular underpinnings of the disease.

RMS is comprised of neoplastic myoblasts that fail to exit the cell cycle and are blocked from terminally differentiating into syncytial muscle. RMS typically is divided into two clinically distinct subgroups (Huh and Skapek 2010; Wexler ): embryonal RMS and alveolar RMS (A-RMS). Embryonal RMS is a genetically heterogeneous subtype, whereas A-RMS, which is notoriously more aggressive, is uniquely driven by the PAX-FOXO1 fusion oncoprotein.

The PAX-FOXO1 transcription factor is generated by chromosomal translocations that fuse a PAX3/7 gene (PAX3 on chromosome 2 or PAX7 on chromosome 1) to the 3′ end of the FOXO1 locus on chromosome 13 (Galili ; Shapiro ; Davis ). The encoded chimera contains intact PAX3/7 DNA-binding domains fused to the FOXO1 transcriptional activation domain (Mahajan ). Because PAX3/7 encode genetic regulators of skeletal muscle development (Lagha ; Buckingham and Rigby 2014), it is postulated that genes regulated by PAX3/7 underlie A-RMS pathogenesis. Despite notable advances with mammalian PAX-FOXO1 RMS models (Keller ; Naini ; Nishijo ; Cao ), our understanding of RMS pathogenesis remains opaque, indicating the need for new genetic tools to dissect RMS pathobiology and uncover new molecular therapeutic targets.

As Drosophila models successfully yield critical insights into human disease, including cancer pathobiology (Gonzalez 2013), we have generated a Drosophila model to interrogate in vivo PAX-FOXO1 pathogenicity. Expression of human PAX-FOXO1 in differentiating fly muscle causes myoblast fusion defects that result in larval lethality (Galindo ). Although tumorigenesis is not observed, in part due to quick lethality, PAX-FOXO1 cells act aggressively and infiltrate nonmuscle tissues. Although PAX7-FOXO1 RMS is less common and demonstrates better clinical outcomes than PAX3-FOXO1, PAX7-FOXO1 phenotypes exhibit better penetrance in flies due to slightly greater sequence identity between human PAX7 and Drosophila PAX3/7. Because expression of wild-type human PAX3 in flies phenocopies PAX-FOXO1, PAX3-FOXO1, PAX7-FOXO1, and wild-type PAX3/7 activity presumably overlap in vivo (Galindo ).

PAX-FOXO1 phenotypes are susceptible to dominant genetic suppression and enhancement (Galindo ; Avirneni-Vadlamudi ; Crose ). Thus, we have been exploiting this genetically tractable model to uncover new PAX-FOXO1 gene targets and cofactors. We have subsequently shown that genetic modifiers isolated from the Drosophila PAX7-FOXO1 system impact RMS oncogenesis and tumorigenesis (Avirneni-Vadlamudi ; Crose ). These findings establish that insights gleaned from this invertebrate model successfully uncover new RMS mechanisms, and new molecular targets for RMS therapy.

Here, we report a comprehensive deletion screen against the Drosophila autosomes to identify PAX-FOXO1 gene targets and effectors, as well as the methods used to configure the screen such that PAX-FOXO1 modifiers are quickly identifiable with only one genetic cross. We additionally report that mutation of Drosophila Myocyte Enhancer Factor-2 (D-Mef2), a critical regulator of both fly and mammalian myogenesis, dominantly suppresses PAX7-FOXO1 lethality and acts as a PAX-FOXO1 gene target. We further find that mutation of (), a gene encoding a MEF2 transcriptional coactivator, similarly suppresses PAX7-FOXO1, further pointing toward MEF2 transcriptional activity as a mediator of PAX-FOXO1 pathogenicity. These studies show the utility of the PAX7-FOXO1 Drosophila system as a robust one-generation (F1) RMS gene discovery platform and demonstrate how Drosophila models can be configured for rapid and effective dissection of human disease.

Materials and Methods

Genetics

In the screen for PAX7-FOXO1 genetic modifiers, the UAS-PAX7-FOXO1 and Myosin Heavy Chain-Gal4 transgenes were used, and lethality assessed, as previously described (Avirneni-Vadlamudi ). The Gal80-containing X-chromosome is from stock #5132 from the Bloomington Drosophila Stock Center. For each experimental cross, approximately three males from the master screening stock were mated to approximately five to seven wild-type, deficiency-, or gene mutation-containing females, and at least two independent crosses performed. Crosses were reared at 23°. Multiple crosses of the master screening stock to the wild-type line were performed to generate large populations of F1 PAX7-FOXO1 male and control female siblings, from which we established a baseline percentage (22%) (SEM = 1.0%) of F1 PAX7-FOXO1 males expected upon routine outcrossing of the screening stock (Supporting Information, Table S1). Each time screening crosses were performed, we included new control crosses to insure that PAX7-FOXO1−induced semi-lethality of F1 males did not significantly differ from the established baseline. All deficiency- and mutation-containing stocks were obtained from the Bloomington Drosophila Stock Center.

The Drosophila PAX7-FOXO1 microarray raw data sets have been previously described and are publically available (Avirneni-Vadlamudi ).

Embryo immunofluorescence

For Drosophila embryo whole-mount immunofluorescence, embryos were treated as described previously (Chen and Olson 2001), incubated in primary antibody overnight at 4° [1:1000 rabbit anti-green fluorescent protein (GFP); Molecular Probes], secondary at room temperature for 2 hr. (1:2000, Alexa-568 goat anti-rabbit; Invitrogen), and mounted in VECTASHIELD with 4′,6-diamidino-2-phenylindole. Microscopy was performed with either an LSM150-meta confocal or Zeiss Axioplan2 fluorescent microscope.

Statistics and study approval

For the genetic screen, suppressors and enhancers were identified by crosses that showed a percent-F1 male population 1 SD above or below the mean, respectively. Fold change is the % F1 males observed for each line tested divided by baseline (22%). Unpaired 2-tailed Student’s t-tests were used to calculate significance. A P value < 0.05 was considered significant.

For the microarray studies, Data represent mean ± SEM. Significance of differences was determined by unpaired 2-tailed Student’s t-test. A P value < 0.05 was considered significant. These studies did not include human tissue, and were exempt from institutional review board approval.

Results

PAX7-FOXO1 drives ectopic myogenesis in Drosophila embryos

Because PAX molecules and myogenesis show striking evolutionary conservation between Drosophila and vertebrates (Halder ; Xue and Noll 1996; Xue ; Daczewska ), we generated a genetically simple and efficient Drosophila PAX7-FOXO1 transgenic platform to dissect PAX-FOXO1 pathobiology. Our approach is based on the Gal4/UAS bipartite expression system (Brand and Perrimon 1993), where transgenic human PAX7-FOXO1 is expressed from the yeast UAS enhancer/promoter by driver lines that express the Gal4 transcriptional activator in tissue-specific patterns.

Toward validating the new fly PAX-FOXO1 system, we tested whether human PAX7-FOXO1 promotes myogenesis in Drosophila. We used the daughterless-Gal4 driver, which directs ubiquitous expression of UAS-transgenes, to express PAX7-FOXO1 during embryogenesis. We then probed for expression of a GFP-tagged Myosin Heavy Chain (MHC) reporter transgene, a marker specific for myogenesis and a reporter previously used in embryonic screens to successfully identify genes involved in Drosophila somatic muscle development and patterning (Chen and Olson 2001; Chen ). Drosophila embryos initiate native expression of MHC at embryonic stage 13—thus, we focused on embryos at stage 12 or younger for ectopic MHC expression. We observed robust expression of MHC-GFP in cells of all three germ layers, including nonmyogenic cells within the ectoderm and endoderm primordia (Figure 1), findings similar to PAX3-FOXO1 misexpression in mouse embryonic primordial cells (Scuoppo ). These results [as well as similar results described below (MEF2 as a PAX-FOXO gene target and putative RMS effector) (Figure 4C)] show that Drosophila precursors are vulnerable to the myogenic programming properties intrinsic to the PAX-FOXO1 chimera.

Figure 1

PAX7-FOXO1 drives myogenesis in Drosophila embryos. (A) Whole-mount wild-type and daughterless-Gal4;UAS-PAX7-FOXO1 (da>>PAX7-FOXO1) gastrulated embryos probed for expression of green fluorescent protein (GFP) from a Myosin Heavy Chain (MHC)-GFP reporter transgene. Because Drosophila embryos initiate native expression of MHC at embryonic stage 13, we focused on embryos at stage 12 or younger. Diffuse expression of MHC-GFP is only detected in the da>>PAX7-FOXO1 embryos. (B) Greater resolution images of embryo segments noted by the white bars in (A) MHC-GFP = GFP immunofluorescence from the MHC-GFP reporter; DAPI = 4′,6-diamidino-2-phenylindole nuclear staining.

Figure 4

Isolation of the myogenesis benchmark gene D-Mef2 as a PAX7-FOXO1 suppressor and gene target. (A) Smaller, overlapping chromosomal deletions reduce the PAX7-FOXO1 deletion suppressor Df(2R)X1 to chromosomal segments 46C1-46C7, which includes D-Mef2, the master regulator of Drosophila myogenesis. (B) D-Mef2 loss-of-function mutation dominantly suppresses PAX7-FOXO1 lethality. PAX7-FOXO1-expression is semilethal. In the presence of Df(2R)X1, which deletes D-Mef2, the population of PAX7-FOXO1−positive adults is increased 2.4-fold and is a PAX7-FOXO1 suppressor. Two smaller overlapping deletions, Df(2R)BSC152 and Df(2R)BSC298, also delete D-Mef2 and suppress PAX7-FOXO1, whereas Df(2R)eve neither deletes D-Mef2 nor acts as a PAX7-FOXO1 suppressor. The D-Mef2 null allele (n = 193 F1 adults scored) is a strong suppressor of PAX7-FOXO1 lethality (P = 0.0018), confirming that D-Mef2 genetically interacts with PAX7-FOXO1. Of note—although the Df(2R)BSC298 deletion showed a fold change of slightly less than 1.9, the increase in PAX7-FOXO1 adults (1.8-fold) was highly significant (P = 0.0004), and in this test we considered a suppressor. (C) PAX7-FOXO1 drives D-Mef2 expression. Whole-mount wild-type and daughterless-Gal4;UAS-PAX7-FOXO1 (da>>PAX7-FOXO1) gastrulated embryos (dorsal surface upper right corner, posterior surface, lower right corner) probed for expression of yellow fluorescent protein (YFP) from a D-Mef2-YFP embryonic reporter transgene. In wild-type embryos, D-Mef2 expression is limited to differentiating myoblasts within the mesoderm. In da>>PAX7-FOXO1 embryos, D-Mef2-YFP reporter expression is seen throughout the embryo, including ectodermal and endodermal derivatives. D-Mef2 is also detectably overexpressed in myoblasts, visible in a segmentally repeating pattern. The black lines note the posterior aspect of both embryos shown in the right-most, greater resolution images. Mef2-YFP = YFP immunofluorescence from the D-Mef2-YFP reporter; DAPI = 4′,6-diamidino-2-phenylindole nuclear staining. *P < 0.05, **P < 0.01, ***P < 0.001 vs. control.

A rapid, one-generation screen for PAX7-FOXO1 suppressors and enhancers

We next configured the Drosophila PAX-FOXO1 platform for unbiased forward genetic screening and RMS gene discovery. For these studies, we turned to an MHC>>PAX7-FOXO1 (MHC-Gal;UAS-PAX7-FOXO1) genetic background to hone in on PAX7-FOXO1 pathogenicity in differentiating Drosophila muscle lineage cells—a setting similar to a conditional PAX3-FOXO1 tumorigenic mouse model (Keller ).

Because MHC>>PAX7-FOXO1 expression is lethal, a typical forward genetic screen would normally require a multigenerational scheme to bring the MHC-Gal4 driver, UAS-PAX7-FOXO1 transgene, and candidate modifiers into the same genetic background, a cumbersome and lengthy process that would need repeating for every candidate mutation-containing chromosome to be tested. To bypass this issue, we configured a stable “master” stock that would allow for candidate modifiers to be efficiently tested with a simple, one-generation (F1) scheme (Figure 2). To generate this master stock, we incorporated a transgenic X-chromosome that ubiquitously expresses the potent Gal4 physical inhibitor, Gal80. Because Gal80 antagonizes MHC>>Gal4, homozygous MHC-Gal4;UAS-PAX7-FOXO1 animals are viable and stable. Upon outcrossing of this master stock, F1 female progeny inherit the Gal4-inactivating Gal80 X-chromosome (which serve as the control cohort), whereas all F1 male siblings express PAX7-FOXO1. Additionally, we exploited that fact that Gal4 activity is partially temperature dependent to identify a rearing temperature (23°) at which PAX7-FOXO1 phenotypes are semilethal.

Figure 2

A rapid unbiased one-generation (F1) screen to uncover dominant PAX7-FOXO1 genetic modifiers. Incorporating an X-linked Gal80 transgenic chromosome allows for the MHC-Gal4;UAS-PAX7-FOXO1 screen to be performed in a single generation. Using the Gal4 inhibitor, Gal80 (carried on the X-chromosome), a viable stable stock was generated that is homozygous for UAS-PAX7-FOXO1 on the second chromosome and Myosin Heavy Chain (MHC)-Gal4 on the third chromosome, which also contains a UAS-GFP transgenic reporter. With this stock, it is possible to screen against any mutant chromosome in one generation, where the number of PAX7-FOXO1-expressing F1 males are compared to control (Gal80-positive) female siblings. Without genetic modification, PAX7-FOXO1 expression is semilethal. Genetic suppressors rescue semilethality and thus increase the number of males in the F1 population, whereas enhancers decrease the percentage of F1 males (Table 1). In the scheme shown, the mutation tested is on the third chromosome, though an equivalent scheme is used for second chromosome mutations.

Table 1

Deficiency enhancers, suppressors, and nonmodifiers of PAX7-FOXO1

Genotype	Breakpoints	Males (P-F)	Females (Control)	Total F1 Adults	% F1 Males	Fold Change	P value	Submapped
Df(3L)emc-E12	61A;61D3	0	115	115	0%	0.00	0	No
Df(3L)ZN47	64C;65C	0	53	53	0%	0.00	0	No
Df(3L)W10	75A6-7;75C1-2	0	217	217	0%	0.00	0	No
Df(3L)fz2	75F10-11;76A1-5	0	77	77	0%	0.00	0	No
Df(3R)crb-F89-4	95D7-D11;95F15	0	225	225	0%	0.00	0	No
Df(3R)crb87-5	95F7;96A17-18	0	194	194	0%	0.00	0	No
Df(2L)TW161	38A6-B1;40A4-B1	1	112	113	1%	0.04	0	No
Df(2R)AA21	57B19-C1;57E1-6	1	75	76	1%	0.06	0	Yes
Df(3R)p-XT103	85A2;85C1-2	3	189	192	2%	0.07	< 0.0001	Yes
Df(3L)fz-GF3b	70C2;70D4-5	1	50	51	2%	0.09	0	Yes
Df(2R)CX1	49C1-4;50C23-D2	2	88	90	2%	0.10	0	Yes
Df(2R)M60E	60E6;60E11	5	140	145	3%	0.16	0.02	No
Df(3L)GN24	63F6-7;64C13-15	4	97	101	4%	0.18	< 0.0001	No
Df(3R)23D1	94A3-4;94D1-4	3	71	74	4%	0.18	0	No
Df(2R)vg-C	49B2;49E2	6	128	134	4%	0.20	0	No
Df(3R)e-R1	93B6-7;93D4	5	98	103	5%	0.22	0.01	No
Df(2R)Egfr5	57D2-8;58D1	8	140	148	5%	0.25	0	Yes
Df(3R)ea	88E7-13;89A1	13	175	188	7%	0.31	0.01	No
Df(2R)BSC40	48E1-2;48E2-10	5	66	71	7%	0.32	0.2390	No
Df(2R)BSC161	54B2;54B17	5	66	71	7%	0.32	0.07	No
Df(2L)BSC32	32A1-2;32C5-D1	7	89	96	7%	0.33	0	No
Df(2L)TE35BC-24	35B4-6;35F1-7	9	114	123	7%	0.33	0	Yes
Df(2R)ED4065	60C8;60E8	30	332	362	8%	0.38	0.01	No
Df(2L)ast2	21E2;22B2-3	12	128	140	9%	0.39	0.1480	Yes
Df(2R)k10408	54B16,54B16	20	180	200	10%	0.45	0	No
Df(2R)BSC49	53D9-E1;54B5-10	12	112	124	10%	0.45	0.01	Yes
Df(3R)by10	85D8-12;85E7-F1	19	170	189	10%	0.46	0	Yes
Df(3R)D605	97E2;98A5	18	152	170	11%	0.50	0.01	No
Df(2R)M41A4	41A;41A	22	167	189	12%	0.53	−	−
Df(2R)Kr10	60F1;60F5	11	80	91	12%	0.55	−	−
Df(3R)Exel6144	83A6;83B6	19	136	155	12%	0.56	−	−
Df(2L)drm-P2	23F3-4;24A1-2	16	113	129	12%	0.56	−	−
Df(3L)66C-G28	66B8-9;66C9-10	19	129	148	13%	0.58	−	−
Df(2L)BSC30	34A3;34B7-9	19	129	148	13%	0.58	−	−
Df(3L)GN34	63E6-9;64A8-9	42	276	318	13%	0.60	−	−
Df(3R)Antp17	84A5;84D9	18	106	124	15%	0.66	−	−
Df(3R)e1025-14	82F8-10;83A1-3	34	196	230	15%	0.67	−	−
Df(3L)Aprt-1	62A10-B1;62D2-5	14	80	94	15%	0.68	−	−
Df(3R)BSC140	96F1;96F10	17	93	110	15%	0.70	−	−
Df(3L)ri-79c	77B-C;77F-78A	35	190	225	16%	0.71	−	−
Df(2R)or-BR6	59B;59D8-E1	18	94	112	16%	0.73	−	−
Df(2R)H3E1	44D1-4;44F12	38	196	234	16%	0.74	−	−
Df(2R)BSC19	56F12-14;57A4	14	71	85	16%	0.75	−	−
Df(3R)L127	99B5-6;99F1	51	254	305	17%	0.76	−	−
Df(3R)Exel6197	95D8;95E5	22	109	131	17%	0.76	−	−
Df(3R)B81	99D3;3Rt	41	197	238	17%	0.78	−	−
Df(2R)Exel7131	50E4;50F6	29	136	165	18%	0.80	−	−
Df(3R)Exel6202	96C9;96E2	32	150	182	18%	0.80	−	−
Df(3L)ME107	77F3;78C8-9	83	387	470	18%	0.80	−	−
Df(3L)BSC14	67E3-7;68A2-6	21	96	117	18%	0.82	−	−
Df(2R)en30	48A3-4;48C6-8	22	97	119	18%	0.84	−	−
Df(3R)Espl3	96F1;97B1	28	123	151	19%	0.84	−	−
Df(3R)Scr	84A1-2;84B1-2	26	114	140	19%	0.84	−	−
Df(3R)BSC47	83B7-C1;83C6-D1	32	140	172	19%	0.85	−	−
Df(2R)Jp1	51D3-8;52F5-9	27	118	145	19%	0.85	−	−
Df(3L)ED4978	78D5;79A2	93	390	483	19%	0.88	−	−
Df(2R)nap9	42A1-2;42E6-F1	28	116	144	19%	0.88	−	−
Df(3R)WIN11	83E1-2;84A5	19	78	97	20%	0.89	−	−
Df(2L)BSC41	28A4-B1;28D3-9	33	129	162	20%	0.93	−	−
Df(3L)brm11	72A3;72D5	72	278	350	21%	0.94	−	−
Df(3R)IR16	97F1-2;98A	46	176	222	21%	0.94	−	−
Df(2L)cl-h3	25D2-4;26B2-5	38	145	183	21%	0.94	−	−
Df(3L)Exel6087	62A2;62A7	103	389	492	21%	1.0	−	−
Df(2L)ed1	24A2;24D4	34	126	160	21%	1.0	−	−
Df(2R)robl-c	54B17-C4;54C1-4	34	126	160	21%	1.0	−	−
Df(3L)ri-XT1	77E2-4;78A2-4	44	163	207	21%	1.0	−	−
w1118 (control, No Df)	N/A	314	1123	1437	22%	1.0	−	−
Df(3L)81k19	73A3;74F	39	134	173	23%	1.0	−	−
Df(3R)Exel6203	96E2;96E6	37	160	164	23%	1.0	−	−
Df(3R)BSC137	95A2-4;95A8-B1	43	141	184	23%	1.1	−	−
Df(3R)Exel9012	94E9;94E13	37	118	155	24%	1.1	−	−
Df(2L)XE-3801	27E2;28D1	14	44	58	24%	1.1	−	−
Df(3R)Exel6196	95C12;95D8	52	155	207	25%	1.1	−	−
Df(2L)TE29Aa-11	28E4-7;29B2-C1	39	113	152	26%	1.2	−	−
Df(2L)FCK-20	32D1;32F1-3	32	91	123	26%	1.2	−	−
Df(2R)BSC44	54B1-2;54B7-10	47	133	180	26%	1.2	−	−
Df(2R)Px2	60C5-6;60D9-10	39	109	148	26%	1.2	−	−
Df(2L)ED611	29B4;29C3	45	125	170	26%	1.2	−	−
Df(3L)ZP1	66A17-20;66C1-5	23	62	85	27%	1.2	−	−
Df(3L)vin7	68C8-11;69B4-5	47	126	173	27%	1.2	−	−
Df(3R)M-Kx1	86C1;87B1-5	38	99	137	28%	1.3	−	−
Df(3L)rdgC-co2	77A1;77D1	58	150	208	28%	1.3	−	−
Df(3R)3450	98E3;99A6-8	140	358	498	28%	1.3	−	−
Df(3L)ED4782	75F2;76A1	31	79	110	28%	1.3	−	−
Df(2L)Prl	32F1-3;33F1-2	35	88	123	28%	1.3	−	−
Df(3L)eygC1	69A4-5;69D4-6	43	107	150	29%	1.3	−	−
Df(2R)X58-12	58D1-2;59A	40	92	132	30%	1.4	−	−
Df(2L)b87e25	34B12-C1;35B10-C1	43	96	139	31%	1.4	−	−
Df(2L)dp-79b	22A2-3;22D5-E1	48	106	154	31%	1.4	−	−
Df(2L)BSC36	32D1;32D4-E1	50	106	156	32%	1.5	−	−
Df(3R)ED5177	83B4;83B6	50	104	154	32%	1.5	−	−
Df(2R)PC4	55A;55F	71	147	218	33%	1.5	−	−
Df(3R)Exel9014	95B1;95D1	53	109	162	33%	1.5	−	−
Df(3L)BSC12	69F6-70A1;70A1-2	36	74	110	33%	1.5	−	−
Df(2R)P34	55E2-4;56C1-11	53	108	161	33%	1.5	−	−
Df(2L)BSC31	23E5;23F4-5	68	138	206	33%	1.5	−	−
Df(2R)BSC132	45F6;46B12	43	86	129	33%	1.5	−	−
Df(3R)Tl-P	97A;98A1-2	90	176	266	34%	1.5	−	−
Df(3L)pbl-X1	65F3;66B10	41	79	120	34%	1.6	−	−
Df(2R)BSC26	56C4;56D6-10	48	88	136	35%	1.6	−	−
Df(3L)st-f13	72C1-D1;73A3-4	58	106	164	35%	1.6	−	−
Df(2L)BSC4	21B7-C1;21C2-3	81	142	223	36%	1.7	−	−
Df(3L)R-G7	62B4-7;62D5-E5	51	89	140	36%	1.7	−	−
Df(2R)B5	46A;46C	50	86	136	37%	1.7	−	−
Df(2R)14H10Y-53	54D1-2;54E5-7	58	99	157	37%	1.7	−	−
Df(2R)w45-30n	45A6-7;45E2-3	83	140	223	37%	1.7	−	−
Df(2R)Exel7162	56F11;56F16	56	94	150	37%	1.7	−	−
Df(2L)BSC111	28F5;29B1	58	97	155	37%	1.7	−	−
Df(2R)ST1	42B3-5;43E15-18	56	91	147	38%	1.7	−	−
Df(2L)pr-A16	37B2-12;38D2-5	47	75	122	39%	1.8	−	−
Df(3L)XS533	76B4;77B	42	67	109	39%	1.8	−	−
Df(3L)BSC8	74D3-75A1;75B2-5	53	84	137	39%	1.8	−	−
Df(3L)vin5	68A2-3;69A1-3	38	60	98	39%	1.8	−	−
Df(2R)BSC39	48C5-D1;48D5-E1	67	102	169	40%	1.8	−	−
Df(2R)CB21	48E;49A	56	85	141	40%	1.8	−	−
Df(3R)3-4	82F3-4;82F10-11	59	87	146	40%	1.8	−	−
Df(3R)Tpl10	83C1-2;84B1-2	41	59	100	41%	1.9	0.08	No
Df(3R)BSC24	85B7;85D15	61	85	146	42%	1.9	0	No
Df(2L)JS17	23C1-2;23E1-2	90	125	215	42%	1.9	0.02	Yes
Df(2R)BSC22	56D7-E3;56F9-12	67	92	159	42%	1.9	0	No
Df(3L)BSC35	66F1-2;67B2-3	192	262	454	42%	1.9	0.01	No
Df(2R)BSC155	60B8;60C4	76	103	179	42%	1.9	0.01	No
Df(3L)BSC20	76A7-B1;76B4-5	61	81	142	43%	2.0	0.03	No
Df(3R)Exel6193	94D3;94E4	58	78	136	43%	2.0	0.01	No
Df(2L)BSC28	23C5-D1;23E2	98	129	227	43%	2.0	0	Yes
Df(3R)BSC42	98B1-2;98B3-5	129	169	298	43%	2.0	0.01	No
Df(2R)Np5	44F12;45DE3	48	60	108	44%	2.0	0.05	No
Df(3L)h-i22	66D10-11;66E1-2	51	63	114	45%	2.0	0.05	No
Df(3L)AC1	67A2;67D11-13	47	58	105	45%	2.0	0.02	No
Df(2L)BSC5	26B1-2;26D1-2	77	95	172	45%	2.0	0.02	Yes
Df(3R)Exel6195	95A4;95B1	45	55	100	45%	2.0	0.0020	Yes
Df(2R)vir130	59B;59D8-E1	59	72	131	45%	2.0	0	Yes
Df(2L)TW203	36E-36E3;37B10	46	55	101	46%	2.1	0	No
Df(2R)BSC18	50D1;50D2-7	101	120	221	46%	2.1	0.02	Yes
Df(2R)BSC29	45D3-4;45F2-6	64	75	139	46%	2.1	0	No
Df(2R)Exel7130	50D4;50E4	105	120	225	47%	2.1	0.03	No
Df(3R)mbc-R1	95A5-7;95D6-11	99	111	210	47%	2.1	0.0050	No
Df(2R)BSC3	48E12-F4;49A11-B6	79	88	167	47%	2.2	0.0180	No
Df(2R)BSC45	54C8-D1;54E2-7	98	109	207	47%	2.2	0	No
Df(2R)cn9	42E;44C	59	64	123	48%	2.2	0	Yes
Df(3L)BSC10	69D4-5;69F5-7	62	67	129	48%	2.2	0.01	No
Df(3L)Scf-R6	66E1-6;66F1-6	43	45	88	49%	2.2	0	No
Df(3L)XG5	71C2-3;72B1-C1	67	67	134	50%	2.3	0	Yes
Df(3L)BSC21	79E5-F1;80A2-3	32	31	63	51%	2.3	0.28	No
Df(2L)E110	25F3-26A1;26D3-11	85	80	165	52%	2.3	0	Yes
Df(3R)mbc-30	95A5-7;95C10-11	58	52	110	53%	2.4	0.01	Yes
Df(2L)spd[j2]	27B2-27F2	64	55	119	54%	2.4	0.02	No
Df(2R)X1	46C;47A1	84	72	156	54%	2.4	0.01	Yes
Df(2R)BSC11	50E6-F1;51E2-4	47	36	83	57%	2.6	< 0.0001	Yes

Based on Mendalian ratios, PAX7-FOXO1−positive males would be expected to represent 50% of all F1 adults. At baseline (w), PAX7-FOXO1 expression causes semilethality, with the percentage of F1 PAX7-FOXO1 males reduced to an average of 22% (SEM of 1.0%) (Please see Table S1.) Enhancers and suppressors decrease and increase, respectively, survival of PAX7-FOXO1 (“P-F”) F1 males 1 SD from the mean (mean = 26%) (SD = 15%). “Fold Change” = % of deficiency PAX7-FOXO1 F1 males observed divided by baseline (22%), with enhancers and suppressors showing a fold change value of ≤ 0.5 and ≥ 1.9, respectively. P values were calculated for the enhancers and suppressors. Three suppressors and three enhancers did not reach statistical significance.

Outcrossing of the screening stock to wild-type flies reared at 23° results in the PAX7-FOXO1-positive male cohort to comprise on average 22% of the F1 population (Table S1), significantly reduced from the rate of 50% that would otherwise be expected based on Mendelian ratios. When screening against this phenotype, PAX7-FOXO1 suppressors and enhancers are easily identified: suppressors and enhancers increase and decrease, respectively, the F1 percentage of males 1 SD from the mean. When compared with baseline and calculated as fold change, suppressors and enhancer show a ≥1.9-fold and ≤0.5-fold change in F1 male numbers, respectively (Table 1). A Student’s t-test is then used to confirm statistical significance for each modifier.

We used a kit of minimally overlapping chromosomal deletions (a.k.a. “deficiencies”) (Table S1) to scan across the autosomes and identify genomic segments (or “hotspots”) that—when absent one copy—genetically modify PAX7-FOXO1 semilethality. Screening against ~95% of the Drosophila autosomes (~75% of the genome), we identified 33 suppressors and 28 enhancers (Table 1) (Figure 3), although three enhancers and three suppressors demonstrated P values above 0.05 and thus did not reach statistical significance. We next used smaller overlapping deletions to further delineate a subset of the hotspot regions, thereby significantly reducing candidate PAX7-FOXO1−interacting genes (Table 2). For the deficiency modifiers not submapped, candidate genes are provided in File S1.

Figure 3

Distribution of the genetic lines tested in the PAX7-FOXO1 Screen. Shown is the plotted distribution of the tested deficiencies and the average baseline wild-type (w) control score (blue line, noted by the arrow) based on the percentage of F1 PAX7-FOXO1 males observed for each line examined. The Mean F1 male percentage for the screen was 26%, with a calculated SD of 15%. Suppressors (green) rank one SD above the mean, whereas enhancers (red) rank one SD below the mean.

Table 2

Submapping of PAX7-FOXO1-modifying Deficiencies

Genotype	Breakpoints	Males (P-F)	Females (Control)	Total F1 Adults	% F1 Males	Fold Change	P value	Comment
Df(2R)CX1	49C1-4;50C23-D2	2	88	90	2%	0.10	0	1
Df(2R)Exel7123	49D5;49E6	6	202	208	3%	0.14	0
Candidate enhancers (49D5;49E6): Aats-aps, bic, CG3790, CG3814, CG13319, CG13321, CG17019, CG30487, Mdr49, NAT1, Nmda1, Psc, Sans, sug, vg
Df(3R)p-XT103	85A2;85C1-2	3	189	192	2%	0.07	0	1
Df(3R)Exel8143	85A5;85B3	27	322	349	8%	0.36	0
Candidate enhancers (85A5;85B3): CG8043, CG8112, CG8116, CG8136, CG8145, CG8159, CG8202, CG8223, CG8236, CG9773, CG9801, CG9837, CG9839, CG11755, CG11760, CG11762, CG11768, CG13318, Cks85A, hb, hng2, Ir85a, M1BP, mRpL19, Pif1A, ranshi, Tcp-1eta
Df(3R)by10	85D8-12;85E7-F1	19	170	189	10%	0.46	0	1
Df(3R)Exel6153	85D19;85E1	4	249	253	2%	0.09	0
Candidate enhancers (85D19;85E1): AP-1μ, bocksbeutel, by, CG8199, CG8273, CG8301, CG8312, CG8319, CG9386, CG9393, CG9396, CG9399, CG9427, CG16789, CG16790, Crc, Kap-a3, MBD-like, mRpL47, mura, P58IPK, Rib1, RnpS1, Vps45
Df(3L)fz-GF3b	70C2;70D4-5	1	50	51	2%	0.09	0	2
Df(3L)Exel6122	70D4;70D7	109	257	366	30%	1.4	—
Candidate enhancers (70C2-70D4): bru-3, CG43184, CG8757, CG8750, Tsp68C, Hml, CG8745, dysc, CG13737, Rgl, CG8833, Glued, Cg32137, Meics, Nxf3, ssp2, CG13738, Hsc70-1, CG17634, CG17632, CG9040, 26-29-p, CG17631, CG17359, CG8783, upSET, ptip, endos, CG6650, CG6661, Hsc70Cb, blue, CG6833, CG13484, CG32138, Pex1, breathless (FGFR), CG8100, Fbp1, Sox21a, Sox21b, Dichaete, nan, nuf, CG32141, CG7768, CG7924, CG34244, CG7906
Df(2L)E110	25F3-26A1;26D3-11	85	80	165	52%	2.3	0	1
Df(2L)BSC5	26B1-2;26D1-2	77	95	172	45%	2.0	0.02
Df(2L)BSC184	26B1;26B3	85	104	188	45%	2.0	0.01
Candidate suppressors (26B1;26B3): chickadee, eIF-4a, ifc, lid, Tsp26a, Gal, CG9098, H2.0, CG13996, CG9107, CG9109, mtm, CG9117, CG31643, ade2, mir-966, slowmo, CG34179, Cg12393
Df(2R)BSC18	50D1;50D2-7	101	120	221	46%	2.1	0.02	1
Df(2R)50C-36	50C19-23;50C21-D5	111	152	263	42%	1.9	0.01
Candidate suppressors (50D1;D5): mastermind, mir-4978, CG18371, Prosap, CG42287, CG42288
Df(2R)BSC11	50E6-F1;51E2-4	47	36	83	57%	2.6	< 0.0001	1
Df(2R)L48	50F6-F9;51B3	47	78	125	38%*	1.7*	0.04
Candidate suppressors (50F6;51B3): Shroom, CG8613, CG8617, Arc1, Arc2, Tfb1, CG34184, CG34442, CG4444, Obp50a, Obp50b, Obp50c, Obp50d, CG34185, Obp50e, CG30075, Dh44-R1, CG10104, CG17385, Sin1, CG17386, phyllopod, Oaz, Lobe, Cpsf160, Asx, Cpr51A, CG30197, tout-velu, CG30076, CG43919, LaminC
Df(2L)JS17	23C1-2;23E1-2	90	125	215	42%	1.9	0.02	2
Df(2L)Exel7015	23C5;23E3	59	143	202	29%	1.3	—
Candidate suppressors (23C1;23C5): CG8814, Prx6005, CG31950, betaggt-II, NTPase, lilliputian, Rbp9, Ts, Rrp1, gammaTub23C, CG9641, CG3165, CG9643, Chd1, Bem46, okra, CG3558, CG17265, CG17224, CG17264, alpha4GT1, CG3542, CG3605, CG17219,GABPI, CG17258, CG17259, CG17260, cnir, CG17221, CG17261
Df(2L)BSC28	23C5-D1;23E2	98	129	227	43%	2.0	0	2
Df(2L)Exel7015	23C5;23E3	59	143	202	29%	1.3	—
Candidate suppressors (23C5;23C5): CG17219, GABPI, CG17258, CG17259, CG17260, cnir, CG17221, CG17261
Df(2R)cn9	42E;44C	59	64	123	48%	2.2	0	2
Df(2R)Exel6053	43D3;43E9	75	138	213	35%	1.6	—
Candidate suppressors (42E1;43D3): CG3358, mim, CheB42b, CheB42c, Che42a, ppk25, Cyp6u1, CG30157, vimar, CG30156, CG17002, Tsp42E-(a-r), Cg30159, CG30160, CG43646, CG43647, CG33914, lbm, pgant3, CG12831, esn, Cyp9b1, Cyp9b2, Spn43Aa, CG12828, prickle, Spn43Ab, Spn43Ad, necrotic, Cg11060, Cg33140, Cg30385, CG30384, Or43a, Ady43a, Gadd45, CG1850, Br140, Incenp, pawn, CG12164, Dscam1, costa, CG11107, Gr43a, CG1707, Eaf, CG11112, CG11113, CG43123, CG43267, mir-4977, sine oculis, CG11145, Cg11123, sPLA2, CG30503, kappaB-ras, fa2h, CG11127, p47, Aldh-III, wech, Coop, CG1620, dpa, didum, CG12763, az2, CG1603, Cg1602, Cg2144, Orc1, Drat, CG2064, mRpL52, CG12107, U2A, CG1399, CG30493, CG4096, CG34216, torso, CG19421, mir-4909, CG1942, CG1946, CG18812, CG30497, CG45093, cn, CanB2, mir-4980, CG12825, Cg12824, Gapdh1, mus205, Nop171, saxophone, Cg1550, Cg1882, cathD, CG30383, phr, phosalpha1, Cg18853, CG30382, CG12822, Atg10, Dgk, CG30377, CG30377, CG12159, Cul1, Or43b, Kdm4a, CG8791, CG30381, rnh1, drosha, CG8728, CG30380, Cg30379, CG14764, CG34430, CG34431, CG11165, CG30378, CG2906, CG2915, Sep5, Nito, CH14763, CG8726, CSN4, ACC, Nup44a, Dic3, Hey, CG11191, Odc1, Odc2, CG14762, mir-4981, Optix, CG12769, CG17977, lig, Vps28, slv, sut1, sut2, sut3, CG8713, CG8712, CG11210, Cul4, udd, Asap, Nup50, coil, Socs44a, Pbp49, CG42516, Pabp2, Obp44a, Lpin, kermit, CG8708, RagC-D, Rs1, CG30373, Gasz, Mlh1, CG14757, LRP1
Df(2R)X1	46C;47A1	84	72	156	54%	2.4	0.01	1
Df(2R)BSC152	46C1;46D7	88	124	212	42%	1.9	0.01
Df(2R)BSC298	46B2;46C7	131	207	338	39%*	1.8*	0	1
Df(2R)eve	46C7;46C9-46C11	31	74	105	30%	1.4	—	2
Candidate genes (46B2;46C7): CG12744, CG1472, CG1513, CG12923, CG30008, CG30007, CG1441, FMRFa, Etf-QO, Mef2
Df(2R)vir130	59B;59D8-E1	59	72	131	45%	2.0	0	2
Df(2R)twi	59C3-4;59D1-2	123	308	431	29%	1.3	—
Candidate suppressors (59B1;59C4): CG42260, CG30270, blw, CycB, stall, CG30271, CG42284, CG30274, CG30272, CG30265, CG12490, CG9825, CG9826, CG3649, CG13531, RpL23, inaD, fd59A, CG13532, PIP5K59B, CG3501, CG3499, asrij, Gmer, MED23, CG3700, nahoda, CG30187, Nup214, CG42678, CG3788, CG3800, CG9849, CG3831, CG42694, CG32834, CG34371, CG13539, LS2, CG3092, uip3, RpL22-like, CG12782, Cg13540, ord, CG3124, CG13541, Prosbeta5R1, HP1Lcsd, CG0412, CG30416, CG9861,CG30417, CG30413, CG3502, CG9863, CG34210, CG30409, Rpi, Cg3500, CG9875, CG34423, CG34424, vir, Ice1
Df(3L)XG5	71C2-3;72B1-C1	67	67	134	50%	2.3	0
Df(3L)brm11	72A3;72D5	72	278	350	21%	1.0	—
Candidate suppressors (71C2;72A3): Best4, Best3, CG7255, Toll-6, CG33259, CG7804, Ran-like, CG12355, CG13455, CG7276, CG7275, Cg7272, CG7857, CG7841, Z600, gdl, gdl-ORF39, Eip71CD, CG13454, mex1, yellow-k, CG7945, CG33986, CG33985, CG42729, obst-H, CG42728, CG43248, CrebA, AGO2, CG7739, CG7427, dop, CG16979, mrn,CG12301, CG12304, CG7656, RhoGAP71E, CG7650, CG13449, comm3, CG7372, CG43083, CG43084, Eig71E-(a-k), CG43082, CG7304, CG7579, pgant8, CG34452, CG34451, comm2, CG42571, CG42570, comm, CG6244, CG13445, fwe, CkIIalpha-i1, DCP2, diablo, CG12713, CG18081, CG15715, CG32150, mir-263b
Df(3R)mbc-30	95A5-7;95C10-11	58	52	110	53%	2.4	0.01
Df(3R)Exel6195	95A4;95B1	45	55	100	45%	2.0	0.02
Df(3R)Exel9014	95B1;95D1	53	109	162	33%	1.5	—
Candidate suppressors (95A5;95B1): CG31145, GILT3, GILT2, eIF-3p66, CG1670, CG18754, SPE, CG10254, CG10252, prt, CG31468, CG31148, CG31413, CG31414, CG10301, CG10300, nautilus (Drosophila MyoD), CG10365
Df(2L)ast2	21E2;22B2-3	12	128	140	9%	0.39	0.1480**	3
Df(2L)Exel6004	21E4;21F1	60	80	140	43%	2.0	0.01
Candidate enhancers (21E2;21E4): CG2839, dachsous, Hsp60B, Eaat2, GABA-B-R3, CG12506, CG13946, CG13947, Gr21a, CG3544, Pkg21D, Nnf1b, Ddp21E2, Saf6, Pex12, CG15880, CG3867, clipper, CG3662, CG3862, dock, drongo, CG4291, kraken, CG13949, mir-375, CG13950, mir-375, aru, dbe, PNUTS, ninaA, CG15824, Lsp1beta, GluRIIC, CG4341, IA-2, Star
Candidate enhancers (21F1;22B2): Tango14, CG5080, IntS14, CG14341, Plap, CG31922, CG5118, CG4887, CG4896, CG5126, Tgt, CG5001, Cg5139, CG43348, CG43349, CG5011, CG14342, CG42329, CG5397, robo3, a5, CG5440, CG33923, CG33922, Cdkc2, CG5556, CG5561, CG31924, CG5565, CG31659, NLaz, CG14346, leak, CG43401, CG43402, CG31928, CG33128, CG31926, CG31661, CG18131, CG7420, CG18132, halo, Or22a, CG44072, Or22b, haf, CG10869, CG31935, CG14352, RFeSP, chinmo, cpb, CG17660, mRpL48, frtz, Rim2, Eno, Rrp40, CG31937, CG17652, CG17646, CG17712, CG17648, Gr22f, CG17650, Gr22-(e-a), CG31933
Candidate suppressors (21E4;21F1): asteroid, Atg4a, CG4692, MtRNApol, CG14339, CG14340, Pino, CG4552, Iris, CG4577, MFS3, CG4749, Tfb4, Vsp29, capulet
Df(2L)TE35BC-24	35B4-6;35F1-7	9	114	123	7%	0.33	0	2
Df(2L)TE35BC-7	35B2;35B10	55	123	154	36%	1.6	—
Df(2L)Exel7063	35D2;35D4	46	64	110	42%	1.9	0.03	3
Candidate enhancers (35C1;35D2): vasa, vig, CG15270, CG15296, stc, CG4168, Sfp35C, CG43230, ZnT35C, dao, Pol32, l(2)35Cc, yuri, Cul3, UK114, CG15263, CG15260, ms(2)35Ci, CG15262, nht, esgargot, CG15258, CG44869, worniu
Candidate suppressors (35D2;35D4): vasa, CG4161, snail, Tim17b2, lace, Skadu, CG15256, kek3, CG15255, Semp1, CG15254, CG15253, CG11865, Or35a, CG7631, CG18480, CG4578, CG44141, CG18477, CG18478, CG43923, CG44140, CG31780, CG1827, CG43924, CycE
Df(2R)BSC49	53D9-E1;54B5-10	12	112	124	10%	0.44	0.01	3
Df(2R)Exel6066	53F8;54B6	100	144	244	41%	1.9	0
Df(2R)BSC154	54B2;54B7	3	93	96	3%	0.14	0.01	1
Candidate enhancers (53D9;53F8): CG5522, CG15919, CG15615, CG5550, CG34459, CG34460, mir-8, Ugt37c1, IntS8, Fen1, Dek, Psi, Ef1beta, CG6426, CG6241, CG6429, CG6435, CG8910, CG6472, mir-990, inaC, Pkc53E, CG43788, CG43789, CG15614, CG43190, Vha16-4, mute, CG34191, PIG-V, CG6665, CG9010, Cbp53E, ste24c, CG30461, ste24b, ste24a, CG6796, NiPp1, CG6805, Ehbp1, CG8963, Dark, RhoGEF2, CG43327, CG43328, CG43371, CG9640, CG9642, CG9646, fat-spondin, tef, CG8950, CG6967, CG30460, Sply, CG6984
Candidate suppressors (53F8;54B2): GstS1, CG30456, CG15611, Amy-p, CG15605, Cda9, Acp54A1, CG11400, Gbp, Cg11395, CG43103, CG43107, CG17290, CG17287, CG30458, CG30457, CG10953, CG10950, CG43237, muscleblind, CG18469, CG12699, CG43272, CG43108
Candidate enhancers (54B2;54B7): Muscleblind, Sip1, CG6568, CG30101, Prosalpha5, cnk
Df(2R)AA21	57B19-C1;57E1-6	1	75	76	1%	0.06	0	2
Df(2R)Exel6072	57B16;57D4	30	169	199	15%	0.7	—
Df(2R)Exel6076	57E1;57F3	77	89	166	46%	2.1	0.03	3
Candidate enhancers (57D4-57E1): Rgk3, CG30391, CG30393, CG34023, MFS16, CG10505, CG30392, Sgf29, RpL29, CG9752, CG42672, CG9754, CG9485, CG33655, CG30394, dom, CG15666, CG9822, CG17974, cv-2
Candidate suppressors (57E1;57F3): Sdc, Sara, Fkbp14, TAF1c-like, MESK2, CG10494, CG30288, CG30289, EGFR, CG30286, CG30287, CG33226, CG30283, twz, CG30222, CG33225, CG10433, CG15673
Df(2R)Egfr5	57D2-8;58D1	8	140	148	5%	0.25	0	3
Df(2R)Exel6076	57E1;57F3	77	89	166	46%	2.1	0.03
Candidate enhancers (57D2;57E1): CG15661, ASPP, Rgk3, CG30391, CG30393, CG34203, MFS16, CG10505, CG30392, Sgf29, RpL29, CG9752, CG42672, CG957, CG9485, CG33655, CG30394, domino, CG15666, CG9822, CG17974, cv-2
Candidate suppressors (57E1; 57F3): CG10795, EfSec, Acox57D-p, Acox57D-d, Sdc, Sara, Fkbp14, TAF1C-like, MESK2, CG10494, CG30288, CG30289, EGFR, CG30286, CG30287, CG33226, CG30283, CG10440, CG30222, CG33225, CG10433

Original chromosomal deletions (a.k.a., Deficiencies, or Df) identified that genetically modify PAX7-FOXO1−induced male semilethality are noted in bold. Additional smaller Df’s tested to further delimit the critical modifying segments are shown directly below. Comments: 1) Df’s for which we were able to reduce the critical modifying chromosomal regions; 2) Df’s for which additional deletions tested showed no modification, indicating that the critical segments lie outside the smaller tested regions; 3) Df’s for which the smaller deletions showed opposite modifying behavior. Similar to Table 1, “Fold Change” = % of Deficiency PAX7-FOXO1 F1 males observed divided by the control baseline of 22%. “*” notes two smaller Df’s that, though with a fold change of slightly less than 1.9, showed a statistically significant increase in the male F1 population, and for this second-pass study we considered suppressors. “**” notes an original Df that did not reach statistical significance but was included in these submapping studies. “P-F” = PAX7-FOXO1.

MEF2 as a PAX-FOXO gene target and a putative RMS effector

We found that mutation of benchmark myogenesis genes modify PAX7-FOXO1. The D-Mef2 gene, which operates as a linchpin and critical nodal point in fly myogenesis (Lilly ), is cytogenetically located at 46C4-46C7 on chromosome 2 and is positioned within a hotspot region (46C1-47A1) initially uncovered by the deletion (Table 1). The hotspot region was further refined by smaller overlapping deletions to segments 46C1-46C7 (Figure 4, A and B and Table 2). Concomitantly, we found by mRNA expression profiling that D-Mef2 is overexpressed in PAX7-FOXO1 larval muscle (2.0-fold, P < 0.001, n = 3). No other gene in this region was reported as misexpressed 2.0-fold or more with statistical significance. Thus, we hypothesized that heterozygous deletion of the D-Mef2 locus might account for -mediated PAX7-FOXO1 suppression, and that D-Mef2 might act as a PAX7-FOXO1 target gene. We tested the D-Mef2−null mutation (Bour ), D-Mef2, which showed that heterozygous loss of D-Mef2 suppresses PAX7-FOXO1 (2.0-fold increase in PAX7-FOXO1 F1 males) (Figure 4B). Of note, these findings do not eliminate the possibility, however, that other genes in this region might also independently interact with PAX7-FOXO1.

Because D-Mef2 mutation suppresses PAX7-FOXO1 and D-Mef2 is overexpressed by PAX7-FOXO1 in our microarray analysis, we investigated whether D-Mef2 acts as a downstream PAX-FOXO1 target. We used the daughterless-Gal4 driver to ubiquitously express PAX-FOXO1 and then probed for expression of a YFP-tagged embryonic D-Mef2 reporter (Cripps , 1999). We found diffuse misexpression of the D-Mef2 reporter (Figure 4C) in ectoderm and endoderm derivatives. Additionally, D-Mef2 reporter overexpression was detected in mesodermal-derived myoblasts, visible in a segmentally repeating pattern. These studies corroborate our aforementioned MHC-GFP reporter expression studies, affirming that human PAX-FOXO1 promotes myogenic fate-specification in Drosophila. These studies further show that D-Mef2 acts as a PAX7-FOXO1 downstream target gene (direct or indirect) and PAX-FOXO1 genetic effector in vivo.

We next interrogated the 50D1-50D5 hotspot suppressor, which contains only five genes (Table 2 and Figure 5A), one of which is . In vivo studies in mammalian models have shown that the ortholog Mastermind-Like 1 (Maml1) encodes a transcriptional cofactor that physically interacts with Mef2 to augment Mef2-dependent promyogenic signaling (Shen ; Potthoff and Olson 2007). Similar to D-Mef2, loss-of-function mutation dominantly suppressed PAX7-FOXO1−induced lethality (Figure 5B). Of note, expression levels were not detectably altered in our PAX7-FOXO1 microarray studies, compatible with mam’s role as a cofactor vs. myogenesis gene target. Taken together, these Drosophila studies highlight a putative PAX-FOXO1→MEF2→RMS pathogenic axis, while also demonstrating that the one-generation (F1) genetic screen quickly uncovers dominant PAX-FOXO1 modifiers/effectors in an unbiased fashion.

Figure 5

Mutation of mastermind, which encodes a MEF2 transcriptional cofactor, is a dominant PAX7-FOXO1 suppressor. (A) Overlapping chromosomal deletions identify a small genomic region, 50D1-50D5, as a PAX7-FOXO1−suppressing hotspot. (B) mastermind loss-of-function mutation dominantly suppresses PAX7-FOXO1 lethality. Df(2R)BSC18 was isolated in our original screen as a PAX7-FOXO1 suppressor (Table 1), which deletes mastermind (mam). The overlapping deletion, Df(2R)50C-36 also suppresses PAX7-FOXO1 (Table 2). Two well-characterized, strong loss-of-function mam alleles, mam (n = 89 F1 adults scored) (P = 0.0044) and mam (n = 112 F1 adults scored) (P = 0.0043) suppress PAX7-FOXO1. Of note−although the mam allele showed a fold change of slightly less than 1.9, the increase in PAX7-FOXO1 adults (1.7-fold) was highly significant, and in this test we scored as a suppressor. *P < 0.05, **P < 0.01, ***P < 0.001 vs. control.

Finally, we surveyed MEF2 expression levels in a large collection of pediatric RMS cancer cell lines, xenograft tumors, and primary tumors using the Pediatric Tumor Affymetrix Database (http://home.ccr.cancer.gov/oncology/oncogenomics/) (Khan ). Four MEF2 orthologs (MEF2A, -B, -C, -D) are present in the mammalian genome, with -A and -C demonstrating greatest similarity to D-Mef2. Compared with normal tissues and non-RMS pediatric soft-tissue sarcomas, only MEF2C showed significant and consistent up-regulation in RMS samples (Figure 6; also shown are MYOD and MET, genes that associate with RMS), data that point toward MEF2C as influential in RMS. Similar to in our PAX7-FOXO1 Drosophila system, analysis of the three mammalian MASETRMIND orthologs (MAML1, -2, and -3) did not reveal a MAML overexpression pattern in these RMS data sets (Figure 6).

Figure 6

MEF2C is overexpressed in Rhabdomyosarcoma. Shown are expression profiles for embryonal rhabdomyosarcoma (E-RMS), alveolar rhabdomyosarcoma (A-RMS), non-RMS soft-tissue sarcoma (Non-RMS STS), and Ewing sarcoma (EWS). Profiles are from cell lines (C), tumor xenografts (X), and primary human tumors (T). Three individual probes are shown for MEF2A, -B, -C (bordered in black), and -D. Probes are also shown for the three human Mastermind orthologs, MAML1, -2, and -3. Representative probes are shown for MYOD1 and MET—genes known to be up-regulated in RMS. mRNA Expression data sets are from the Pediatric Tumor Affymetrix Database (Oncogenomics; http://home.ccr.cancer.gov/oncology/oncogenomics/).

Discussion

The Drosophila PAX7-FOXO1 genetic model

Given the critical role that the PAX-FOXO1 fusion oncoprotein plays in RMS, we focused on PAX-FOXO1 as an entry-point for designing a transgenic Drosophila RMS-related model that would be amenable to forward genetic screening and RMS gene discovery. To bypass the issue of cumbersome multigenerational screening schemes that would normally be required, we incorporated a Gal80 X-linked chromosomal transgene to generate a viable screening Gal4/UAS-PAX-FOXO1 master stock that allows for the rapid identification of PAX-FOXO1 genetic modifiers in a single genetic cross.

With this platform, we have been probing for new PAX-FOXO1 pathogenesis underpinnings. Though very similar in molecular structure, PAX3-FOXO1− and PAX7-FOXO1−positive RMS demonstrate differing clinical behaviors, as PAX3-FOXO1 tumors are more common and notoriously aggressive (Kelly ). Consequently, PAX3-FOXO1 is the PAX-FOXO1 fusion most commonly investigated in vertebrate models. In our Drosophila system, we have focused on PAX7-FOXO1, which demonstrates phenotypes that are better penetrant and experimentally tractable due to the fact that human PAX7 demonstrates slightly greater sequence identity to fly PAX3/7 than does human PAX3. Additionally, as no other animal models of PAX7-FOXO1 presently exist, the fly PAX7-FOXO1 model also conveniently serves as a complement to vertebrate PAX3-FOXO1 models.

Initially unknown was the extent to which observations from the PAX7-FOXO1 fly model would impact the clinically more aggressive PAX3-FOXO1 RMS subtype, as well as PAX-FOXO1-negative (embryonal) RMS. Notably, our previous studies have shown that genetic modifiers identified from the Drosophila system impact PAX3-FOXO1 RMS oncogenesis and tumorigenesis (Avirneni-Vadlamudi ; Crose ). Furthermore, unpublished studies (U. Avirneni-Vadlamudi and R. L. Galindo, unpublished data) are demonstrating that fly PAX7-FOXO1 genetic modifiers are similarly involved in Embryonal RMS. These findings provide marked validation for the applicability and value of this genetic fly system to human RMS.

Interestingly, though PAX7-FOXO1 induces expression of the late myogenic differentiation marker MHC, PAX-FOXO1 RMS myoblasts in culture and in vivo demonstrate only partial differentiation with little-to-no MHC expression. In considering this discrepancy, we first note that PAX-FOXO1 is a relatively weak driver of RMS in culture and in vivo (Keller ; Naini ) and requires additional/sequential genetic aberrations to induce oncogenic transformation. Thus, secondary mutations might be necessary to force the strength of RMS myoblast differentiation-arrest seen in human RMS tumors; by contrast, our PAX7-FOXO1 model differs in that the system is free of any additional background mutations. Second, previous studies have shown that expression of PAX3-FOXO1 in mouse embryonic cultured cells induces the formation of MHC-positive myocytes and myotube formation (Scuoppo ), studies that are similar to those seen here in the Drosophila system, where the da-Gal4/UAS-PAX7-FOXO1 expression system targets undifferentiated embryonic primordia. Uncovering of the genetic/molecular sequence of RMS pathogenesis and the cell(s) origin will shed further insight into the underlying mechanisms that account for the myoblast differentiation arrest phenotypes seen in RMS in vivo.

MEF2 in myogenesis and RMS

The differentiation and fusion of myoblasts into postmitotic, syncytial muscle requires that the bHLH myogenic regulatory factors (MRFs: Myf5, Mrf4, MyoD, and Myogenin) interact with E-proteins, which drive and regulate critical aspects of myogenic fate determination (Braun and Gautel 2011). The MRFs subsequently interact with the MEF2 transcription factors that, although lacking intrinsic myogenic activity, cooperate with the MRFs to synergistically activate muscle-specific genes and the downstream myogenic terminal differentiation program (Molkentin and Olson 1996; Potthoff and Olson 2007).

Vertebrates possess four MEF2 family member genes (-A, -B, -C, -D), which demonstrate complex overlapping spatial and temporal expression patterns in embryonic and adult tissues, with greatest expression levels seen in striated muscle and brain (Potthoff and Olson 2007). Because of genetic redundancy and overlapping expression patterns of the MEF2 genes, interrogating individual MEF2 gene activity in mammals has been experimentally challenging, with loss-of-function mutation studies revealing only limited insights into MEF2 gene function in tissues in which the MEF2 genes do not overlap/compensate. Conveniently, flies possess only one gene (D-Mef2) and have served as an excellent model system to delineate MEF2’s critical role in myogenesis (Potthoff and Olson 2007). We speculate that the lack of redundancy in flies provided a marked experimental advantage in isolating D-Mef2 as a PAX7-FOXO1 effector. Similarly, the identification of was also likely facilitated by the fact that flies possess one gene, whereas mammals contain three orthologs (Saint Just Ribeiro and Wallberg 2009). Thus, we propose that the comparative lack of genetic compensation/redundancy is an attractive advantage to Drosophila as a disease model system.

Recent studies have made significant inroads toward dissecting MEF2 in myogenesis in vivo and RMS—most specifically, MEF2C and -D. Whereas global deletion of Mef2A or -D demonstrates little to no effect on embryonic myogenesis (Potthoff and Olson 2007), skeletal muscle-specific deletion of Mef2C causes neonatal lethality due to defective muscle integrity and sarcomere formation (Potthoff ,b). Regarding RMS, Zhang have found that RMS cells lack proper expression of MEF2D, and that exogenous expression of MEF2D promotes RMS cell differentiation, diminishes oncogenesis in culture, and blocks tumorigenesis in xenograft studies. Turning to adult muscle regeneration and satellite stem cells, which is likely at least one cell of origin for human RMS, Liu have now shown that Mef2A, -C, and -D are essential yet function redundantly in satellite cell differentiation. Lastly, our survey of published RMS microarrays (Khan ), as well as PAX-FOXO1-expressing myoblast cell lines (Avirneni-Vadlamudi ), shows a consistent pattern of MEF2C overexpression. Given the integrated and overlapping nature of the MEF2 genes, we hypothesize a potential mechanism in which overexpression of MEF2C feeds-back upon and down-regulates MEF2D, thereby preventing MEF2D from driving myoblast terminal differentiation.

We suggest that further interrogation of MEF2 in RMS will open new avenues for RMS chemotherapy, which for high-risk disease has not improved for decades. For example, since MEF2 activity is tightly governed by class IIa histone deacetylases (Haberland ; Potthoff and Olson 2007; Nebbioso ), histone deacetylase inhibitors are now ripe for preclinical testing as new anti-RMS agents. Additionally, we have found that the MEF2 cofactor Mastermind, which interacts with MEF2C and mediates crosstalk between Notch signals during myogenic differentiation (Shen ; Potthoff and Olson 2007), similarly influences PAX-FOXO1 pathogenicity in flies. Interestingly, Mastermind-specific, cell-permeable peptide inhibitors have been shown to block the progression of T-cell acute lymphoblastic leukemia in mice in vivo (Moellering ) and thus are also new agents available for RMS preclinical testing. Further characterization of MEF2 in RMS cell and mouse models will continue to refine both our understanding and the potential targeting of MEF2 activity in RMS.

In conclusion, we postulate that: 1) The Drosophila PAX7-FOXO1 model is uniquely configured for the quick uncovering of new RMS genetic effectors with one simple genetic screening cross; 2) a putative PAX-FOXO1→MEF2/MASTERMIND axis underlies A-RMS; and 3) Drosophila conditional expression models are an efficient and powerful gene discovery platform for the rapid dissection of human disease.