XRCC4 suppresses medulloblastomas with recurrent translocations in p53-deficient mice

Yan et al. 10.1073/pnas.0601938103.

Supporting Information

Files in this Data Supplement:

Supporting Figure 5
Supporting Table 2
Supporting Figure 6
Supporting Figure 7
Supporting Figure 8
Supporting Figure 9

Fig. 5. Schematic of XRCC4 conditional targeting strategy. A targeting construct containing exon 3 of XRCC4 and Neo^R cassettes, flanked by loxP sites (arrowheads), were targeted into TC1 ES cells. Targeted cells containing only the floxed (Xc) or null (X–) allele and lacking the Neo^R cassette was generated in vitro by transient infection of Cre recombinase containing plasmid into targeted ES clones and injected for germ-line transmission.

Supporting Figure 6

Fig. 6. Loss of XRCC4 expression in the brain of NX^c/–p53^–/– mice. (A) Detection of XRCC4 expression by RT-PCR and Western blotting. For RT-PCR, cDNA obtained from reverse transcription of RNA was obtained from adult cerebellum, P2.5 brain, NX^c/– and NX^c/+ cerebellum. An X^–/–p53^–/– thymus control and two medulloblastomas, NXcP66 and NXcP167 are shown. GAPDH control is shown in Left Lower. XRCC4 primers: mXR3Fw, GAGAGCAATAGGATCCGGCTTTGTTAT; mXR4Rv: CTCCAGCTGGACTTCATTTAGCA. Western blot was probed with antibody to XRCC4 and actin as a control for protein loading. (B) Cryosections of NX^c/– and NX^c/+ brains immunostained with anti-XRCC4 (Cy5, red), anti-b3-tubulin antibody, TUJ1 (FITC, green), and counterstained with DAPI (blue, nuclear staining). Nuclear and light cytosolic staining of XRCC4 is observed in NX^c/+ brain, whereas XRCC4 is barely detectable in the brain of NX^c/– mice. (C) Hematoxylin/eosin staining of littermate NX^c/+ and NX^c/– brain sections. Sagittal brain section of E14.5 NX^c/+ (CA and CB) and NX^c/– (C and D). (Original magnification, A and C, ×100; B and D (boxed areas in A and C), ×400). Cx, cortex; IZ, intermediate zone; VZ, ventricular zone; V, ventricle.

Supporting Figure 7

Fig. 7. Histological analyses of NX^c/–p53^–/– medulloblastomas. Paraffin embedded brain sections containing medulloblastomas from three NX^c/–p53^–/– mice, NXcP55, NXcP71, and NXcP89 were immunostained by using antibodies to GFAP, NeuN, and Synaptophysin and were costained with hematoxylin/eosin. The NX^c/–p53^–/– tumors are strongly immunoreactive for GFAP and NeuN, an indication of prominent neuronal differentiation. All of the tumors have an overall high proliferative index (indicated by Ki67/MIB1 strong immunoreactivity; data not shown) and varying levels of pyknotic cells, an indication of ongoing apoptosis.

Supporting Figure 8

Fig. 8. Oncogene amplification and expression in NX^c/–p53^–/– medulloblastomas. (A) Detection of Cyclin D2 amplification in a subset of medulloblastomas. Southern blot analyses of 12 of the NX^c/–p53^–/– tumors and control samples were digested with EcoRI hybridized with Cyclin D2 probe and an LR8 probe as loading control. (B) Northern blot analyses of nine NX^c/–p53^–/– medulloblastomas as compared with postnatal day 4 (P4) neurospheres. Total RNAs were sequentially blotted with cDNA probes to Cyclin D2, N-myc, and GAPDH as loading control. (C) Western blot analyses of a subset of NX^c/–p53^–/– medulloblastomas probed with antibody to Cyclin D2 and Tubulin as control for protein loading. Expression of Cyclin D2 in NX^c/–p53^–/– tumors is comparable with P7 neurospheres and higher than that detected in the adult cerebellum.

Supporting Figure 9

Fig. 9. CGH analyses of NX^c/–p53^–/– Medulloblastomas. Array CGH is shown for tumors NXcP57 and NXcP71 plotted in Log₂ ratio. (Top) A function of genome length, with the chromosome numbers noted above. (Lower) Analyses of chromosome 6, 12, and 13. Both NXcP57 and NXcP71 exhibit N-myc amplification on chromosome 12 and loss of the distal arm of chromosome 13. By CGH, NXcP57 further exhibits amplification of Cyclin D2, which is not amplified in NXcP71 (verified by Southern blot, shown in Fig. 9A). Arrowheads indicate the positions of Cyclin D2 (in green), N-myc (in blue), and Patched (in purple). Mb, million base pairs. Comparison of XbaI and HindIII digests from genomic DNA from sample tumor, as compared with n = 5 reference DNA samples, were determined by using DCHIP to determine copy number alterations. The five nearest neighbor smoothing were used in total chromosomal analyses. In single-chromosome analyses, bar depicts the normal, amplification, and deletion levels detected across each chromosome.

Table 2. Tumor development in cohort nestin Cre, XRCC4 conditional, p53-deficient mice

	NX^+/+	Np53^–/–	NX^c/+	NX^c/–	NX^c/+p53^+/–	NX^c/–p53^+/–	NX^c/–p53^–/–
No. of mice	12	12	22	23	25	23	23
Medulloblastoma	0	0	0	0	0	0	20(104)
Thymic Lymphoma	0	5(130)	0	0	0	0	3(119)*
Sarcoma	0	2(118)	0	0	0	0	2(95)*
Uncharacterized	0	3(210)	0	0	3(180)	3(176)	3(138)^†

Notably, whereas NX^c/– and NX^c/+ mice are not predisposed to tumors, p53 deficiency invariably predisposes

NX^c/–p53^–/– mice to medulloblastoma development. Numbers of cohort mice and tumor development are shown. Numbers in parentheses indicate average day when mice become moribund and are analyzed.

^*Denotes additional tumors found in 5 of the 20 NX^c/–p53^–/– mice harboring medulloblastomas.

^†Unexpected death and autolysis precluded histological analyses of these mice.