Most lung and colon cancer susceptibility genes are pair-wise linked in mice, humans and rats

Abstract

Genetic predisposition controlled by susceptibility quantitative trait loci (QTLs) contributes to a large proportion of common cancers. Studies of genetics of cancer susceptibility, however, did not address systematically the relationship between susceptibility to cancers in different organs. We present five sets of data on genetic architecture of colon and lung cancer susceptibility in mice, humans and rats. They collectively show that the majority of genes for colon and lung cancer susceptibility are linked pair-wise and are likely identical or related. Four CcS/Dem recombinant congenic strains, each differing from strain BALB/cHeA by a different small random subset of ±12.5% of genes received from strain STS/A, suggestively show either extreme susceptibility or extreme resistance for both colon and lung tumors, which is unlikely if the two tumors were controlled by independent susceptibility genes. Indeed, susceptibility to lung cancer (Sluc) loci underlying the extreme susceptibility or resistance of such CcS/Dem strains, mapped in 226 (CcS-10 x CcS-19)F2 mice, co-localize with susceptibility to colon cancer (Scc) loci. Analysis of additional Sluc loci that were mapped in OcB/Dem strains and Scc loci in CcS/Dem strains, respectively, shows their widespread pair-wise co-localization (P = 0.0036). Finally, the majority of published human and rat colon cancer susceptibility genes map to chromosomal regions homologous to mouse Sluc loci. 12/12 mouse Scc loci, 9/11 human and 5/7 rat colon cancer susceptibility loci are close to a Sluc locus or its homologous site, forming 21 clusters of lung and colon cancer susceptibility genes from one, two or three species. Our data shows that cancer susceptibility QTLs can have much broader biological effects than presently appreciated. It also demonstrates the power of mouse genetics to predict human susceptibility genes. Comparison of molecular mechanisms of susceptibility genes that are organ-specific and those with trans-organ effects can provide a new dimension in understanding individual cancer susceptibility.

Figure 1. Schematic representation of the genetic composition of recombinant congenic (RC) strains.

The major donor-strain regions of the CcS RC strains that were used to map colon or lung tumor susceptibility genes are shown based on real genotypes.

Figure 2. Correlated lung and colon cancer susceptibility in the CcS RC strains.

A. Expected susceptibility to colon and lung tumors under different hypotheses. Concordant susceptibility or resistance to colon and lung tumors is expected when the majority of the susceptibility genes of the two cancers are closely linked or identical (upper panel); but not when the susceptibility genes of the two cancers are independent of each other (lower panel). B. Observed susceptibility to colon and lung tumors in the CcS RC strains with extreme susceptibility phenotype. Each dot represents a mouse. Mean tumor number of each strain is indicated. Upper panel: colon tumor numbers for CcS-19, CcS-11, CcS-10 and CcS-20 mice. Colon tumor number is directly proportional to colon tumor load, since in our experiments colon tumor sizes did not differ significantly among the CcS strains . Lower panel: lung tumor loads for CcS-19, CcS-11, CcS-10 and CcS-20 mice. The same extreme susceptibility or resistance to lung tumors observed here has been also seen in hybrid mice between CcS and (BALB/c×FVB)F1 mice (Figure S1, Table S1).

Figure 3. Linkage data and the estimated effects of lung cancer susceptibility (Sluc) loci in (CcS-10×CcS-19)F2 mice.

Eight main effects (A) and 17 interactions (B) were detected. The microsatellite markers with linkage are listed below each locus. Markers in pink represent donor chromosomal region from CcS-10 and markers in green represent donor regions from CcS-19. The data are presented as percent deviations from the means of tumor load, number and size for each genotype or genotypic combination (for interactions) of the corresponding loci in female, male or both sexes (all), respectively, adjusted for the remaining markers in the model (least-squares means from the ANOVA output). The means (±SEM) of tumor number in the whole cross are: females, 4.78(±0.37); males, 4.18(±0.29); all mice, 4.48(±0.24). The means (±SEM) of tumor size per mouse (mm³) are: females, 2.05(±0.25); males, 2.26(±0.28); all mice 2.15(±0.19). The means (±SEM) of tumor load per mouse (mm³) are: females, 10.57(±1.60); males, 10.04(±1.30); all mice, 10.30(±1.03). †Loci A and B are interacting; s/s, homozygous STS; c/c homozygous BALB/c; s/c heterozygous; C. Examples are shown that the STS alleles of the Sluc loci are susceptible when they are inherited from the CcS-19 parental mice (Sluc5) and resistant when they are inherited from the CcS-10 parental mice (Pas9). D. Example is shown that in interactions, the genotypic combinations that are similar to the CcS-19 parental mice (CcS-19-like) are susceptible compared to the genotypic combinations that are similar to the CcS-10 parental mice (CcS-10-like). In the example, one of the interacting loci Pas9 is inherited from CcS-10 and the other locus Sluc4 is inherited from CcS-19.

Figure 4. Sluc loci detected in (CcS-10×CcS-19)F₂ mice co-localize with previously detected Scc loci.

The seven Sluc loci in the chromosomal regions that have been tested for colon cancer susceptibility previously are shown. All the seven Sluc loci co-localize with Scc loci. Markers with linkage are highlighted in orange and the corresponding Sluc loci are listed. Additional markers tested on the same donor chromosomal regions that did not show linkage can help to limit the candidate regions and are shown in grey color. Previously detected Scc loci are highlighted in blue. Detailed information of each locus is listed in Suppl. Table 2 and Suppl. Table 3.

Figure 5. Co-localization between Sluc loci mapped in the OcB RC strains and Scc loci mapped in the CcS RC strains.

A. Schematic representation of overlapping donor chromosomal regions between the CcS and OcB RC strains (regions tested for both colon and lung cancer susceptibility loci). Such regions are informative and we used them to test whether Sluc and Scc loci are more frequently located together in the same donor chromosomal region. Part of a chromosome is shown as example. B. Frequent co-localization between Sluc and Scc loci identified independently in OcB and CcS RC strains, respectively (See also Table 1 for detailed locations). *Map locations of these Sluc loci are slightly different from the locations of the same loci shown in Figure 4, since they are mapped in different RC strains using different microsatellite markers.

Figure 6. Interspecies correlation between colon and lung cancer susceptibility loci.

A. Schematic representation of the part of genome used for the co-localization analyses. The Sluc loci analyzed here included also 2 Sluc loci identified in (CcS10 XCcS19)F2 mice and 1 CcS locus and 1 Par locus identified by other group. B. Interspecies correlation between colon and lung cancer susceptibility loci. This figure summarizes all 21 clusters of colon and lung cancer susceptibility loci mapped in mouse RC strains (orange for lung, blue for colon), human colon (green) and rat colon (purple). Clusters in which the lung and colon cancer loci mapped within 2.5cM of each other are highlighted in squares. Most colon and lung cancer susceptibility loci co-localize, with the exception of human 15q13 and 20p12.3 (colon), and rat rCcr6 and rCcr8 (colon). Orthologous regions of human 18q21, 11q23 (colon) and 5p15, 6p21 and 15q25 (lung), and rat rCcr1, rCcr4 and rCcr9 (colon) are not informative since they were not tested for lung or colon cancer susceptibility in mouse RC strains. †Pas1c has also been detected in our (CcS-10×CcS-19)F2 cross at D6Mit177. Human colon cancer locus 3q21-q24 is mapped to an 18Mb region and orthologous to two mouse chromosomal regions: Chr.6 (Pas1c) and Chr.9 (Sluc11), respectively. †† Two human colon cancer susceptibility loci co-localize with a Sluc locus.