Figures and data
Fmi is required in clonal tumors to outcompete wildtype tissue.
A) Schematic of the whole eye disc RasV12, scrib RNAi, RFP tumors.
B-H) Pupal RFP tumors imaged through the cuticle. All pupae are ey-Flp; act5C>CD2>Gal4, UAS-RFP with the indicated RNAi. N = 5 (no tumor), 10 (control RNAi), 13 (fmi RNAi), 17 (fz RNAi), 19 (dsh RNAi), 17 (vang RNAi) and 8 (pk RNAi).
C) Control UAS-RasV12, UAS-scrib RNAi, UAS-w RNAi tumors.
D-H) UAS-RasV12, UAS-scrib RNAi tumors co-expressing UAS-RNAi against the PCP genes indicated above each pupa.
B’-H’) Representative third instar larval brain and eye discs for each of the experimental groups from above. Arrowheads in B’ point to the eye-antenna imaginal disc (orange) and the brain lobes (green). N = 3 (no tumor), 10 (control RNAi), 13 (fmi RNAi), 12 (fz RNAi), 15 (dsh RNAi), 20 (vang RNAi) and 10 (pk RNAi).
I) Schematic of eye disc RasV12, scrib RNAi, RFP clonal tumors.
J-P) Third instar larval eye discs RFP tumor clones generated via ey-Flp; FRT42D Gal80 / FRT42D; act5C>CD2>Gal4, UAS-RFP.
J) Eye discs with non-tumor, control RFP clones.
K) Control RasV12, UAS-scrib RNAi, RFP clonal tumors.
L-P) RasV12, scrib RNAi, RFP clonal tumors carrying the PCP allele indicated above each panel. Scalebars: 100 μm
Winners require Fmi in developmental supercompetition.
A) Eye imaginal disc clones overexpressing UAS-Myc (>>Myc) and UAS-nGFP under control of the tub>CD2>GAL4 driver. Clones were generated with ey-Flp, so half of the cells are clones and the other half are wildtype twins.
B) Eye imaginal disc >>Myc, UAS-nGFP, fmiE59 clones, competing against wildtype twins.
C) Eye imaginal disc >>Myc, UAS-nGFP clones, competing against fmiE59 twins.
D) Ratio of RFP-labeled clone area versus unlabeled wildtype area. The unlabeled wildtype area was obtained by subtracting the RFP-labeled area from the total eye disc area. A ratio over 1 implies RFP-positive clones are over-represented, likely behaving as supercompetitors, while a ratio below 1 means the RFP-cells are under-represented, likely losers. N = 9 discs (>>Myc), 5 discs (fmiE59, >>Myc), 4 discs (fmiE59 vs >>Myc), groups were analyzed using multiple unpaired, two-tailed t test; p-values: <0.0001 (****), 0.6340 (ns)
E) Representative wing imaginal disc overexpressing >>Myc and UAS-nGFP under control of the tub-Gal4 driver. Clones were generated using hsp70-Flp, with a 15 min 37°C heat-shock. wildtype cells are labeled with tub-nRFP. Non-recombinant cells are heterozygous for nRFP, while twin spots are homozygous nRFP. n = 14 discs
F) Representative wing imaginal disc with >>Myc, UAS-nGFP, fmiE59 clones, over a wildtype background. Clones were generated in the same fashion as D. n = 19 discs
G) Representative wing imaginal disc with >>Myc, UAS-nGFP clones, competing against fmiE59 twin spots. n = 13 discs
H) log10 of the >>Myc/Twin spot cell ratio. nGFP total cells were divided by the number of twin spot homozygous nRFP cells, and then log10-transformed. A ratio over 0 indicates over-representation (likely supercompetition) of nGFP+ clones, and below 0 indicates nGFP+ cells are under-represented (likely behaving as losers). The difference between groups was analyzed using a one way ANOVA, with Dunnett correction for multiple comparisons; p-value: >>Myc vs fmiE59 >>Myc < 0.0001 (****); >>Myc vs fmiE59 = 0.9705 (ns).
Scalebars: 50 μm
Winners require fmi in scribble cell competition
A-L’) scrib RNAi clone analysis in the prospective eye. UAS-scrib RNAi was expressed with the act5C>CD2>Gal4 driver and clones were generated by ey-Flp. Clones are marked both with UAS-nGFP and tub-nRFP.
A) Representative early third instar disc with control UAS-nGFP clones.
B) Representative late third instar disc with control UAS-nGFP clones.
C-C’) Representative adult eye with control UAS-nGFP clones. C’ shows fluorescently labeled control clone cells.
D) Representative early third instar disc with UAS-scrib RNAi clones versus wildtype twin clones.
E) Representative late third instar disc with UAS-scrib RNAi clones versus wildtype twin clones.
F-F’) Representative adult eye with UAS-scrib RNAi clones versus wildtype twin clones. F’ shows fluorescently labeled surviving UAS-scrib RNAi cells.
G) Representative early third instar disc with UAS-scrib RNAi clones versus fmiE59 twin clones.
H) Representative late third instar disc with UAS-scrib RNAi clones versus fmiE59 twin clones.
I-I’) Representative adult eye with UAS-scrib RNAi clones versus fmiE59twin clones. This phenotype was lethal. The adult eye shown was from an escaper. Escapers had trouble eclosing and died within hours. I’ shows fluorescently labeled surviving scrib RNAi cells. tub-nRFP fluorescence was barely visible, so the stronger UAS-nGFP is shown.
J) Representative early third instar disc with fmiE59, UAS-scrib RNAi clones versus wildtype twin clones.
K) Representative late third instar disc with fmiE59, UAS-scrib RNAi clones versus wildtype twin clones.
L-L’) Representative adult eye with fmiE59, UAS-scrib RNAi clones versus wildtype twin clones. L’ shows fluorescently labeled surviving fmiE59, UAS-scrib RNAi cells.
M) Quantification of fluorescently tagged clone area for the phenotypes mentioned above. Each dot in the violin plot displays the ratio of RFP+ area versus total disc area. Of special significance is the fact that scrib RNAi clones competing with wildtype clones survive similarly to scrib RNAi clones competing with fmiE59. P-values are as follow: **** < 0.0001; *** < 0.0005; ns = 0.2634
Scalebars: 50 μm
Lack of Fmi increases cell death and reduces proliferation in would-be winners.
A-C) RasV12, scrib RNAi tumors stained for DAPI, Dcp1+ (A), and puc-lacZ (B). C) Merged channels
D) Representation of how Dcp1+ staining localizes in the wildtype cells at the boundary with the tumor.
E-G) RasV12, scrib RNAi tumors mutant for Fmi, stained with DAPI, Dcp1+ (E), and puc-LacZ (F). G) Merged channels.
H) Representation of how Dcp1+ staining localizes in the tumor cells in contact with the surrounding wildtype tissue.
Scalebar for A-G: 25 μm
I-K) Dcp1+ staining in >>Myc clones in eye discs. >>Myc clones are marked by GFP (I) and were stained against Dcp1+ (J). The arrows shows apoptotic WT (red arrow) and >>Myc (white arrow) cells at the clone boundary. K) Merged channels, showing apoptotic cells evenly distributed between WT and >>Myc cells.
L-N) Dcp1+ staining in >>Myc clones lacking Fmi in the eye disc. Eye disc >>Myc clones are marked by GFP (L) and were stained for Dcp1+ (M). N) Merged channels, showing apoptotic cells localized mainly in the >>Myc, fmiE59 clones.
Scalebar: 50 μm
O) Quantification of apoptotic cells in WT versus >>Myc clones in eye discs. Apoptosis occurs similarly in WT and >>Myc cells (two-tailed paired t-test; p-value = 0.6049). The left side of the graph shows the number of apoptotic WT and >>Myc cells. Each imaginal disc is displayed as a pair of dots, linked by a line, to easily visualize the Dcp1+ apoptotic cells in WT versus >Myc cells. Dots represent the number of apoptotic WT (left) or >>Myc (right) cells per disc. The right side of the graph displays the difference (>>Myc minus WT apoptotic cells). The dashed line indicates the mean difference between those values for all samples. N = 14 discs.
P) Quantification of apoptotic cells in WT versus >>Myc, fmiE59 clones in eye discs. Apoptosis is found mainly in >>Myc, fmiE59 cells (two-tailed paired t-test; p-value = 0.0006). The left side of the graph shows the number of apoptotic WT and >>Myc, fmiE59 cells, side by side. Dots represent the number of apoptotic >>WT (left) or >>Myc, fmiE59 (right) cells per disc.The right side of the graph displays the difference (>>Myc, fmiE59 minus WT apoptotic cells). The dashed line indicates the mean difference between those values for all samples. N = 14 discs.
Q-R) Proliferation analysis performed by pHis3 staining in wing discs with either >>Myc clones (Q) or >>Myc, fmiE59 clones (R). Scalebar: 20 μm
S) Proliferative ratio of GFP cells in a non-competition Control (n=9 discs), >>Myc (n=9 discs), or >>Myc, fmiE59 (n=13 discs) clones. The proliferative ratio for each group was calculated as the ratio of pHis3 cells within the GFP+ clone versus the non-GFP wildtype tissue and the differences were analyzed as an ordinary ANOVA with a Tukey’s test for multiple comparisons, with all p-values < 0.0001 (****)
Vang is not required for >>Myc supercompetition
A) Representative wing imaginal disc showing the GFP-tagged, vangA3 >>Myc clones. Clones were generated using hsp70-Flp
B) RFP-labeled twin spots for the clones shown in A. Twin spot homozygous for RFP can be observed adjacent to the supercompetitor clones.
C) Graph showing the total GFP and homozygous RFP counts per disc, with each disc counts linked by a line and the log10 ratio of GFP/hRFP cells per disc (n=12 discs). Differences between GFP/hRFP clone size were analyzed using a two-tailed, ratio-paired t test. p-value<0.0001. When the ratio was compared with a two-tailed t test to >>Myc, the difference was not significant, with a p-value of 0.078.
Scalebar: 50 μm
The Fmi cadherin repeats are not required for cell competition.
A) Representative disc with nGFP-labeled, >>Myc (A), >>Myc, fmiE59 (B) or >>Myc, fmiE59, arm-fmiΔcad
(C) clones competing against wildtype twin spots in the wing disc. Twin spot clones are labeled with homozygous nRFP and clones were generated with hsp70-Flp.
D) Ratio of GFP versus RFP cells in the three groups, represented as the log10(GFP/hRFP) cell ratios. To evaluate the effect of the arm-fmiΔcad rescue (n=14 discs), the GFP/hRFP cell ratio was directly compared against the two other groups, already quantified and shown in Fig. 2F. Differences between the groups were analyzed using an unpaired, ordinary one-way ANOVA, which found a p-value <0.0001. Inter-group differences were analyzed with a Tukey multiple comparisons test, which found no differences between >>Myc and >>Myc, fmiE59, arm-fmiΔcad clones, whereas both groups strongly differed from >>Myc, fmiE59 clones, both returning a p-value <0.0001 (****) when compared directly against >>Myc, fmiE59.
Scalebar: 50 μm
The ey-Flp recombination system causes almost every cell in the eye disc to undergo chromosomal recombination.
A) GFP-marked clonal cells. Clones generated with ey-Flp; FRT42D tub-Gal80 / FRT42D tub-RFP; ac>Gal4 UAS-GFP, express GFP in one of two daughter cells undergoing recombination to create homozygous clones without tub-Gal80.
B) RFP-marked cells. Upon recombination, the same cells that express GFP also express two copies of tub-RFP, whereas cells that do not recombine express one copy of tub-RFP and lack GFP expression. We see few heterozygous RFP cells lacking GFP, indicating that few cells fail to undergo recombination in this system.
fmi RNAi co-expression decreases the size of clonal RasV12, scrib RNAi tumors.
A) RasV12, scrib RNAi clonal tumors generated by ey-Flp co-expressing w RNAi, imaged in wandering third instar larvae. N = 5 larvae
B) RasV12, scrib RNAi clonal tumors generated by ey-Flp co-expressing fmi RNAi, imaged in wandering third instar larvae. N = 6 larvae
C) RasV12, scrib RNAi whole eye third instar larva tumor generated from ey-Flp clones, where the WT cells were eliminated using GMR-Hid, l(2)CL. N = 8 larvae
D) RasV12, scrib RNAi, fmiE59 whole eye tumor generated from ey-Flp clones, where the WT cells were eliminated using GMR-Hid, l(2)CL. N = 7 larvae
Scalebars: 200 μM
Fmi by itself does not trigger cell competition
A) Eye disc GFP+ fmiE59 versus an unlabeled wildtype background. Clones were generated with ey-Flp as in Fig. 2A.
B) Eye disc GFP+, RFP control clones versus a wildtype background.
C) log of the GFP/WT clone ratio measured in either fmiE59 clones vs wildtype (n=11 discs) or RFP+ clones vs wildtype (n=15 discs). Differences were analyzed by a two-tailed, unpaired t test, rendering a p-value of 0.7655
D) Representative wing imaginal disc showing the GFP+ fmiE59 clone cells. Clones were generated with hsp70-Flp.
E) Twin spot RFP+ cells for the disc shown in D. Homozygous RFP+ twin spots can be observed adjacent to the RFP-fmiE59 clones.
F) fmiE59, GFP cells versus wildtype homozygous RFP twin spot cells counted in wing imaginal discs (n=10 discs). The differences were analyzed with a two-tailed, ratio-paired t test, p-value = 0.4815
Scalebar: 50 μm
Clone size during early third instar scrib competition
A) Percentage of RFP clone area in wildtype third instar discs. The X axes indicates the disc size measured as the total nuclear count in the eye imaginal disc. The line represents the best linear fit to the samples plotted.
B) Percentage of RFP, scrib RNAi clone area competing with wildtype twins in third instar discs.
C) Percentage of RFP, scrib RNAi clone area competing with fmiE59 twins in third instar discs.
D) Percentage of RFP, fmiE59, scrib RNAi clone area competing with wildtype twins in third instar discs.
Fmi-/- tumor cell debris is found in vesicles inside wildtype cells
A) Close-up of RFP-tagged fmiE59 RasV12, scrib RNAi tumors competing against unlabeled wildtype cells in eye discs. RFP+ tumor cells can be observed at the periphery, and RFP+ debris is detected in the unlabeled wildtype cells inside the annotated area.
B) Lysotracker staining marking acidic vesicles. Large, stained vesicles can be observed in the annotated area.
C) The merged channels show colocalization of the RFP+ cell debris with the lysosomal vesicles inside wildtype cells.
D) Closeup of the annotated region in panels A-C. Scale bar: 25 μm
Lack of fmi does not affect the activation of JNK signaling.
A-C) Closeup of late third instar eye discs showing scrib RNAi clones marked with RFP (A), puckered transcriptional activation using the puc-LacZ reporter (B) and a merge image showing both channels (C)
D-F) Closeup of late third instar eye discs showing scrib RNAi, fmiE59 clones marked with RFP (D), puckered transcriptional activation (E) and a merge image showing both channles (F).
Fmi levels are not affected by cell competition.
A/D) Overview of representative imaginal discs with >>Myc clones labeled with nGFP (green) and Fmi immunostaining (white), obtained from different larvae.
B) >>Myc clone (green), from the annotated box shown in (A). The clone boundary is delimited with a dashed line.
C) Fmi staining (white) from the annotated box shown in (A). The clone boundary is delimited with a dashed line.
E) >>Myc clone (green), from the annotated box shown in (D). The clone boundary is delimited with a dashed line.
F) Fmi staining (white) from the annotated box shown in (D). The clone boundary is delimited with a dashed line.
Scalebars: 50 μm
