The Drive to Acquire Biallelic Hits Inversely Correlates with the Functional Impact of the Primary TP53 Lesion: The Complexity of TP53 Role Assessment

While the clinical impact and the mechanistic contribution of TP53 mutations have been a subject of intense research, many questions about its role in myeloid neoplasia (MN) remain unanswered. Previous molecular studies have confirmed the assertion that biallelic inactivation confers less favorable prognosis as opposed to monoallelic hits. These evidences agree with the observation that carriers of Li-Fraumeni syndrome do not always exhibit a complete penetrance of the recessive TP53 lesion. Thus, the presence of a residual function of TP53 appears to be somehow protective until it is offset by additional damage of contralateral allele or compound heterozygous hits in synergistic pro-leukemogenic pathways.

TP53 can be affected by lesions of diverse configurations (e.g. biallelic, homo/hemizygous) targeting different locations [missense mutation (ms) in various hotspots vs truncations], and their assessment in terms of clinical consequences is complex. Only large cohorts of patients allow to discern the often discrete nuances of TP53 effects in individual inactivation patterns. We have compiled molecular and clinical data of a meta-analytic cohort (CCF and public datasets) of 1,011 patients with TP53 alterations, along with 3,419 cases found to be TP53 wild type (WT). A total of 1,258 TP53 mutations/deletions were found, 66% classified as biallelic and 37% as monoallelic hits (including single deletions).

We investigated the closest hotspot ms mutations, hypothesizing that lesions mapping sequences in proximity will have the same phenotypic impact. Next, we arranged ms mutations into 6 main sites with each one containing lesions mapping within 5 amino acidic positions from the canonical hotspot location. These sites were mutated in 58% of patients with presence of truncating hits in 27% of cases. When ms mutation sites were compared to each other, a less dismal survival was observed for only the R175H hotspot (p .03).

Most hotspots are known to exhibit dominant-negative effects (likely due to tetramer protein configurations) and thus, inhibit >50% of the TP53 activity as opposed to truncations which should inactivate ~50%. Consequently, one would expect that hotspot mutations produce a more aggressive phenotype. However, patients with ms hits had similar survival as those with truncating mutations (p=.6), likely because truncations were more often biallelic than ms mutation (81% vs 65%, p=.006). Indeed, we can stipulate that the strength (functional impairment) conveyed by a mutation will inversely correlate with the propensity to acquire biallelic hits. Therefore, we hypothesized that truncations (inactivating less TP53) would require an additional hit if compared to the stronger dominant-negative ms lesions. Notably, double hits were identified in 81% of cases carrying truncating mutations vs. 66% in those with ms canonical sites mutations (p<.009).

Carriers of biallelic mutations had worse prognosis than those with monoallelic hits in adjusted multivariate analysis (HR 2.2 95% CI 1.8-2.7 p<.001). However, unlike in previous reports, in our large cohort containing several MN types, monoallelic hits were not survival neutral, but worsened the prognosis as compared to WT patients (p<.001). This finding implies a strong driver effect for TP53 lesions, which are characterized by a rapid progression even in the monoallelic configuration. Similarly, monoallelic hits were associated with a higher mutational burden compared to biallelic ones (1.22 vs 0.91 co-mutations/patient, p=.02), which likely compensated the need for further TP53 inactivation.

When focusing on the accompanying genomic landscape of our cohort, we found that 45% of cases had TP53 mutations as the sole molecular lesion vs 55% of patients who also harbored co-occurring somatic events. In particular, complex karyotype was more frequent among patients without co-occurring mutations (79% vs 57%, p<.001). As of associations with disease subtypes, primary AML cases had a lower burden of co-mutations (p<.001) while the highest percentages were registered in LR-MDS (p=.005).

In summary, our study demonstrates the complexity of assigning a correct clinical impact to TP53 mutations, which are characterized by a high degree of genomic heterogeneity. In addition to the genetic context, TP53 role may also vary in different subtypes of MN (e.g., AML vs MDS) shaping in a different fashion individual patients' trajectories.