A genetic basis for cancer sex differences revealed in Xp11 translocation renal cell carcinoma

Abstract

Xp11 translocation renal cell carcinoma (tRCC) is a female-predominant kidney cancer driven by translocations between the TFE3 gene on chromosome Xp11.2 and partner genes located on either chrX or on autosomes. The rearrangement processes that underlie TFE3 fusions, and whether they are linked to the female sex bias of this cancer, are largely unexplored. Moreover, whether oncogenic TFE3 fusions arise from both the active and inactive X chromosomes in females remains unknown. Here we address these questions by haplotype-specific analyses of whole-genome sequences of 29 tRCC samples from 15 patients and by re-analysis of 145 published tRCC whole-exome sequences. We show that TFE3 fusions universally arise as reciprocal translocations with minimal DNA loss or insertion at paired break ends. Strikingly, we observe a near exact 2:1 female:male ratio in TFE3 fusions arising via X:autosomal translocation (but not via X inversion), which accounts for the female predominance of tRCC. This 2:1 ratio is at least partially attributable to oncogenic fusions involving the inactive X chromosome and is accompanied by partial re-activation of silenced chrX genes on the rearranged chromosome. Our results highlight how somatic alterations involving the X chromosome place unique constraints on tumor initiation and exemplify how genetic rearrangements of the sex chromosomes can underlie cancer sex differences.

Fig. 1:. Etiology of TFE3 fusions revealed by tRCC whole-genome sequencing.

(A) Overview of tRCC samples profiled by WGS and RNA-seq in this study. (B) Summary of TFE3 fusions in the tRCC WGS cohort. Each line represents a TFE3 fusion in an individual patient and is colored based on the sex of the patient. The total number of instances of each fusion is listed in parentheses. (C) Rearrangement breakpoints at the TFE3 locus in the tRCC WGS cohort. (D) DNA copy number across chrX and chr17 in case TRCC14 with a TFE3-LUC7L3 fusion. The fusion partners (TFE3 on chrX and LUC7L3 on chr17) are marked by vertical lines; the completely balanced DNA copy number indicates balanced reciprocal translocations, which are also supported by rearrangements (red arcs). (E) Summary of different types of rearrangements giving rise to TFE3 fusions in the WGS cohort. From left to right: (1) intra-chrX inversion, including paracentric (i.e., inversion within the Xp arm) and pericentric (inversion spanning the centromere) inversion; (2) X:autosome translocations with autosomal partners on either the p-arm (e.g. SFPQ) or q-arm (e.g. PRCC) of an autosome; (3) chromoplexy, i.e., balanced translocations between three or more pairs of break ends (see table S4). (F) Sequence features of reciprocal translocations at TFE3 and partner loci illustrated by an example of TFE3-RBM10 fusion detected in TRCC7. Left: minimum DNA loss or gain at each partner locus. The example shows no DNA loss or gain at the TFE3 breakpoint and a 3bp deletion at the RBM10 breakpoint. Right: minimum homology or insertion at rearrangement junctions. The example shows a 2bp microhomology at the RBM10-TFE3 junction and a 3 bp insertion at the TFE3-RBM10 junction. (G) Summary of DNA loss/gain at fusion partner loci and junction homology of all cases in the tRCC WGS cohort. The x- and y-axes indicate the number of duplicated (+) or deleted (−) nucleotides at the TFE3 translocation partner (x) and the TFE3 breakpoint (y). The upper right and lower left triangles of each square show the classification of microhomology or insertion at reciprocal junctions (Partner-TFE3 or TFE3-Partner). (H) Mechanisms that result in no change (i), deletion (ii), or duplication (iii) between opposite breakpoints generated by canonical non-homologous end-joining of two break ends generated by a single double-strand break. (I) Histogram showing the length of microhomology (left) or insertion (right) at TFE3 fusion breakpoints. Both Partner-TFE3 and TFE3-Partner breakpoints are pooled for this histogram. An apparent outlier with 24bp insertion at the TFE3-MED15 junction detected in TRCC10 also shows a significant deletion (27bp) at the TFE3 locus (see panel G), both of which are more consistent with microhomology-mediated end-joining.

Fig. 2:. Sex differences in frequency of X:autosomal translocations underlies female predominance of tRCC.

(A) Summary of the aggregated tRCC cohort generated from 11 independent studies (including the current study). (B) Sex distribution of tRCC cases in the aggregate tRCC cohort (N = 203). (C) Classification of TFE3 fusions in the aggregate cohort based on the DNA copy-number status, sex, and rearrangement partner. Top (“Total”): Copy-number imbalance is rare in both female and male tRCCs and deletion is exclusive to female tRCCs. Middle (chrX intra-chromosomal inversions): No copy-number imbalance is detected in either female or male tRCCs. Bottom (X:autosomal translocations): There is female dominance in all autosomal TFE3 partners except PRCC and LUC7L3, and deletion is exclusive to female tRCCs. Copy number imbalance consistent with loss of the TFE3 reciprocal fusion is shown in blue while all other types of copy number imbalance involving either the TFE3 fusion or the reciprocal TFE3 fusion are in green. (D) An example of deletion of the reciprocal TFE3-SFPQ fusion t(1;X) in tRCC sample profiled via WES. Shown are the minor allele fraction (top) and copy ratio (tumor/normal depth, bottom) of both chr1 (left) and chrX (right). Data points in regions of allelic imbalance are highlighted in orange. (E) Constraints on each class of rearrangement leading to TFE3 fusions and the copy-number status of translocated chromosomes imposed by the genetic differences in chrX between males and females. Intra-chromosomal inversions can only lead to oncogenic fusions on the active (filled red) but not the inactive chrX (open red) due to the silenced epigenetic state chrXi; by contrast, chrXi:autosomal translocations may generate oncogenic TFE3 fusions if accompanied by partial activation of TFE3 (bottom). Importantly, deletion of the reciprocal translocation would only be tolerated in female tRCCs when the translocation is generated on the inactive X. (F) The frequency of intra-chromosomal inversions and X:autosome translocations in either male or female samples (N=182) indicates that the female tRCC dominance is largely caused by X:autosomal translocations with a near exact 2:1 female-to-male ratio. The slight enrichment of intra-chromosomal inversions in male patients (P=1.2e-2, binomial test against 1:1 female-to-male ratio) is largely attributed to the higher prevalence of RBM10-TFE3 fusions in male tRCCs; RBM10 shows a male bias in mutation frequency(51).

Fig. 3:. Genome evolution of tRCC reveals independent deletions of the reciprocal TFE3 fusion.

(A) Summary of systemic treatment history and anatomic sites of 14 tRCC samples from subject TRCC18. (B) Phylogenetic tree of the dominant tumor clones in 13 metastatic tRCC samples determined from single-nucleotide mutations (see Methods and fig. S9–S10). Ancestral branches (with more than one progeny clones) are shown as thick bars; terminal branches are shown as thin lines. Branch lengths are proportional to the number of somatic mutations. The gray bars preceding RPLN-2 and the Lung biopsies denote mutations on the translocated chromosome t(1;X) that were lost from these samples. Two samples inferred to have near tetraploid genomes (Liver-4 and MedLN) highlighted in underline (see panel E below). (C) Summary of ancestral copy-number alterations in all tumor biopsies using data from Liver-2. Left: CIRCOS plot showing haplotype-specific DNA copy number (blue and red), and rearrangements (green for intra-chromosomal rearrangements, purple for inter-chromosomal rearrangements) of chromosomes with copy number alterations. The reciprocal TFE3-SFPQ translocations are highlighted as a thick magenta line. Right: Distributions of haplotype-specific read depth data for chromosomes with altered DNA copy number (chrs.6,8,9,14,17,and 18). We reserve the A homolog for the haplotype with altered DNA copy number and the B homolog for the intact one; for chr9, the A homolog had focal deletions and the B homolog underwent complete loss. The integer copy-number states of each chromosome are determined by the allelic depths of the lost 9B homolog (CN=0), the gained 8A homolog (CN=2), and regions with normal DNA copy number (CN=1). There is a minor subclonal loss of the 4B homolog that is only present in the Liver-2 sample; subclonal gains of the translocated 17A and 18A are also observed in other samples and inferred to have arisen independently. All other alterations are shared by all metastatic lesions and the primary tumor. (D-F) CIRCOS and allelic-depth distribution plots of selected chromosomes showing downstream evolution after the ancestral changes as shown in C. (D) The lung metastases (using Lung-1 as the representative) show a near clonal deletion of most of the reciprocal translocation t(1;X) (see fig. S10), copy-number gains of 17A and 18A, and a near clonal deletion of 22A. (E) The MedLN and Liver-4 metastases (using MedLN as the representative) were inferred to be 4N based on the presence of 4A, 7A, and 11A at the median copy-number state between 22A (CN=0) and the basal copy-number state. Based on this inference, we further conclude that the loss of entire t(1;X) occurred after whole-genome duplication. (F) We infer the RPLN-2 to be polyclonal with many subclonal copy-number alterations (4A, 13A, 20A, 22A). Due to these alterations and the scarcity of point mutations, we cannot determine the number of subclones or their ploidy states. However, the deletion of t(1;X) shows the same clonality as the deletion of 4A and the p-arm of 20A, and these deletions cannot be accounted for by losses after whole-genome duplication as the copy-number states are lower than the median between complete deletion (CN=0 inferred from 9B) and the basal copy number state (CN=1 in 2N genomes or CN=2 in 4N genomes). (G) Top: Box plots of allelic fractions of somatic mutations on Xq in all metastatic biopsies. Each mutation is phased to either the retained (open boxes) or the deleted (filled boxes) homolog based on their allelic fractions in all samples. Each box plot shows the average variant allele fraction of mutations phased to each homolog in each sample. The Lung-1 and Lung-3 biopsies show complete deletion of mutations on the deleted homolog, whereas RPLN-2, Liver-4, MedLN show lower allelic fractions of these mutations consistent with incomplete losses. The Lung-2 sample is a polyclonal mixture. The mean allelic fraction of both groups of mutations in each sample is consistent with the clonal fraction of tumor cells estimated from the allelic depth of 9B (inferred to be deleted in the founding clone). The ratio between the average allelic fraction of mutations on the retained versus those on the deleted homolog in MedLN and Liver-4 is also consistent with a 2:1 copy-number ratio between the two homologs.

Fig. 4:. Xi:autosome translocations are associated with Xi reactivation.

(A-C) Haplotype-specific transcription of chrX (A), chr1 (B) and chr7 (C) in the primary tumor from TRCC18. These chromosomes have segmental (chrX and chr1) or whole-chromosome allelic imbalance (chr7) in different metastases (first row), enabling the determination of haplotype phases of germline variants in colored regions (highlighted with orange and black dots corresponding to the allelic depths of parental homologs; allelic depths in regions in allelic balance are shown as gray dots). Shown in the 1^st and 2^nd rows are the allelic DNA depths at heterozygous variants in tumors with allelic imbalance (chrX: Lung-1, chr1 & chr7: MedLN) and in the primary tumor. All three chromosomes are disomic and in complete allelic balance in the primary tumor (second row). Haplotype-specific transcription (fourth row; shown is allelic transcription from expressed heterozygous variant sites) was assessed in the primary tumor, using phased variant genotype (third row; shown are allelic DNA fractions of reference and alternate genotypes at heterozygous sites). There is near complete monoallelic transcription on chrXq from the chrX homolog retained in lung metastases. The exception is a variant in XIST which is phased to the chrX homolog deleted in Lung-1, establishing that the inactive X (chrXi) was deleted in the lung metastases (and is also lost in Liver-4/MedLN and RPLN-2). Haplotype-specific transcription on chr1p also revealed allelic imbalance telomeric to the SFPQ locus, reflecting silencing of this region under the influence of XIC from the translocated chrXi. Chr7, disomic in the primary, has completely balanced allelic transcription. P-values calculated by Mann-Whitney U test between distributions of RNA minor allele fraction for heterozygous sites to the left and right of the TFE3 breakpoint (chrX, P = 1.05e-3) or SFPQ breakpoint (chr1, P = 1.14e-2). (D) Predicted transcriptional consequences of Xi:autosome translocations. (E) Single nucleus RNA sequencing of TRCC18 primary tumor sample reveals evidence for increased transcriptional output from the translocated Xp segment. Transcriptional output is assessed for regions telomeric to TFE3 breakpoint (left), chrXq (middle), or diploid autosomes (right) as indicated in the chromosome schematic. The distribution of normalized transcriptional output in each of the above regions is then plotted for malignant (N=4118) and normal cells (N=172) (see Methods). P-values calculated by Mann-Whitney U test (chrXp genes telomeric to TFE3: P = 6.3e-3 [**], chrXq genes: P = 0.48 [n.s.], diploid autosomes: P = 0.79 [n.s.]). (F) Cumulative fractions of biallelically expressed heterozygous variant sites (≥ 10 DNA reads and ≥ 10 RNA reads, minor allele fraction ≥ 0.2, excluding genes escaping XCI) on each chromosome (y-axis) compared to the cumulative fractions of heterozygous sites on each chromosome (x-axis). Only diploid autosomes and chrX are included. The fractions of heterozygous expression (y-axis) or heterozygosity (x-axis) are cumulated beginning at the p-terminus of each chromosome. Deviations from the diagonal (most obvious for chrX) indicate a positional divergence from biallelic expression. (G) Cumulative distributions of biallelic transcription (red) compared to cumulative distributions of DNA heterozygosity (black) on chrX (excluding genes escaping XCI) for selected female samples from the aggregate cohort in the following categories: female tRCC samples with autosome-TFE3 fusions and chrXp reactivation (P < 0.05 by Mann-Whitney U test between distributions of minor allele fraction for heterozygous sites to the left and right of the TFE3 breakpoint); female tRCC samples with no evidence of reactivation (P > 0.05); ccRCC; tRCC cancer-adjacent normal sample. See also fig. S11. (H) Donut plot of female tRCC samples with or without biallelic chrXp expression across the combined TCGA and DF/HCC cohorts. P-values calculated for each sample by Mann-Whitney U test between distributions of minor allele fraction for heterozygous sites to the left and right of the TFE3 breakpoint.