TP53 co-mutations in advanced lung adenocarcinoma: comparative bioinformatic analyses suggest ambivalent character on overall survival alongside KRAS, STK11 and KEAP1 mutations

Background Recently, we could show that the co-mutations of KRAS + KEAP1, STK11 + KEAP1 and KRAS + STK11 + KEAP1 lead to a significantly shorter median overall survival (mOS) in patients with lung cancer across treatments by analyzing multiple dataset. TP53, a tumor suppressor gene, plays a crucial role in regulating cell cycle progression. Its mutations occur in approximately 40-50% of non-small lung cancer (NSCLC). Co-occurrence of all four mentioned mutations has been a matter of debate for years. The aim of this study was to assess the distribution of these four mutations and the influence of the different co-mutational patterns on survival. Methods We present a comparative bioinformatic analysis and refer to data of 4,109 patients with lung adenocarcinoma (LUAD). Results Most of the mutations within the LUAD belong to TP53-only (29.0%), quadruple-negative (25.9%) and KRAS-only (13.4%). Whereas TP53-mutations seem to have protective effects in the context of further KEAP1- and KRAS + KEAP1-alterations (improved mOS), their role seems contrary if acquired in an already existing combination of mutations as KRAS + STK11, KRAS + STK11 + KEAP1 and STK11 + KEAP1. TP53 co-mutations had a negative influence on KRAS-only mutated LUAD (mOS reduced significantly by more than 30%). Discussion These data underline the need for complex mutational testing to estimate prognosis more accurately in patients with advanced LUAD.


Introduction
Lung Cancer is the leading cause of cancer death worldwide, with non-small lung cancer (NSCLC) representing the largest group.Lung adenocarcinoma (LUAD) belongs to the most common and best studied histological subgroups (1).Besides the common treatment strategies consisting of surgery, radiation and chemotherapy, the development and approval of targeted therapies and immune checkpoint inhibitors (ICI) explicitly improved therapy options and patients' outcome within the past decade.However, treatment responses still vary in a wide range even for the personalized treatment options available (2).Therefore, amongst others, one major need is to acknowledge the significance of genetic coalterations and their influence on therapy responses.
KRAS (Kirsten rat sarcoma viral oncogene homolog) plays a key role in the development and progression of various cancers, including NSCLC.Mutations occur in about 25% of cases, leading to the constitutive activation of KRAS signaling pathways, promoting uncontrolled cell growth and survival (3).KRASaltered NSCLC frequently show co-mutations within the genes Kelch-like ECH-associated protein 1 (KEAP1) and serine/ threonine kinase 11 (STK11), also known as liver kinase B1 (LKB1) (4).Recently, we could show, by analyzing multiple datasets, that the co-mutations of KRAS + KEAP1, STK11 + KEAP1 and KRAS + STK11 + KEAP1 lead to a significantly shorter median overall survival (mOS) across treatments.In contrast, patients with tumors harboring only KRAS mutations or being negative for all above-mentioned alterations show a significantly improved mOS in a multivariate analysis.Furthermore, triple co-mutated primary tumors showed a significantly increased frequency of distant metastases to bone and adrenal glands (5).Thus, analysis of the complex mutational network seems inevitable in the clinical routine setting.
TP53 (tumor protein p53), a tumor suppressor gene, plays a crucial role in regulating cell cycle progression, DNA repair, as well as apoptosis and is one of the most common alterations among all cancers, and among LUAD in particular (6).Its mutations occur in approximately 40-50% of NSCLC cases, leading mainly to a loss of function, allowing cells to evade normal regulatory mechanisms and promoting tumorigenesis (7).
Co-occurrence of all four mentioned mutations has been a matter of debate for years: Aredo et al. (8) described concurrent pathogenic mutations of KRAS with TP53 (39%), STK11 (12%) and KEAP1 (8%), discussing distinct molecular subtypes (study with a total of 186 patients).Furthermore, they could show that combined KRAS G12C and TP53 mutations predict benefit from immunotherapy.
Frost et al. focused on 119 patients with lung adenocarcinoma receiving pembrolizumab monotherapy as first-line palliative treatment.Here, rates for KRAS, TP53 and combined mutations were 52.1%, 47.1% and 21.9%, respectively.Whereas, TP53 mutations alone had no impact on response and survival, a subgroup (12 patients) with KRAS G12C + TP53 co-mutations defined long-term responders to immunotherapy (9).
Recently, Proulx-Rocray et al. described 100 patients with known KRAS status.They postulated that KRAS mutation in NSCLC might be associated with a favorable response to ICI therapy in the absence of a concurrent mutation in the STK11 and/or KEAP1 tumor suppressor genes (10).A survival advantage associated with TP53 mutation in NSCLC treated with ICIs has been reported in current literature (9,(11)(12)(13).
However, the above-mentioned studies only encompass a low number of patients.Furthermore, besides ICI, "classical" chemotherapy still presents the main cornerstone of therapy.Therefore, we here present a comprehensive bioinformatic analysis encompassing two datasets retrieved from cBioPortal and tested the influence of TP53 co-mutations depending on the KRAS, STK11 and KEAP1 status.
The aim of this study was to assess the distribution of the four mutations KRAS, STK11, KEAP1, and TP53 as single mutations as well as co-mutations in patients with advanced LUAD in a large dataset.Furthermore, we want to study the influence of the different co-mutational combinations on survival.

Materials and methods
For this study, the following two datasets from the MSK institute were retrieved from cBioPortal (14, 15): the "MSK-IMPACT Clinical Sequencing Cohort (MSKCC, Nat Med 2017)" (16) and the "MSK MetTropism (MSK, Cell 2021)" dataset (17).To create one dataset across treatments the two datasets were merged into one "MSK across treatments" dataset (N = 4,855 NSCLC patients, 4,109 LUAD patients, 542 lung squamous cell carcinoma (LSCC) patients).Therapy details were not annotated.Due to the more recent data, data from the "MSK MetTropism (MSK, Cell 2021)" were preferred over the "MSK-IMPACT Clinical Sequencing Cohort (MSKCC, Nat Med 2017)".All data comprised patients with advanced tumor stages (mainly stage IV).Analyses were performed on LUAD data.
While mutations in KRAS, STK11 and KEAP1 significantly cooccurred among themselves (q < 0.05), there was neither a significant co-occurrence nor a significant mutual exclusivity between mutations in one of the three genes and TP53 mutations.
In contrast, concurrent TP53 mutations to KEAP1-only and to KRAS + KEAP1 mutations showed an opposite effect and led to an improved mOS: 21.1 months for KEAP1-only improved to 52.2 months for KEAP1 + TP53 (p = 0.03; q = 0.085; Figure 3C) and 16.1 months for KRAS + KEAP1 improved to a mOS which did not reach the median for KRAS + KEAP1 + TP53 (p = 0.053; q = 0.1; Figure 3F).

Discussion
Here, we presented a comparative bioinformatic analysis referring to data of 4,109 patients with the aim of analyzing the influence of TP53 mutations in KRAS, STK11 and KEAP1 (co-) mutated LUAD on the patients' overall survival.By employing this database approach, we were able to show that TP53 mutations had an influence on mOS for better or worse depending on the concurrent mutational pattern.Therapy and prognosis of NSCLC has changed in the last 15 years as several treatable targets have been detected within the concept of so-called personalized therapies.In daily practice, these targets encompass testing for rearrangements (ALK, ROS, RET, NTRK, MET) and mutations (KRAS, EGFR, BRAF, ERBB2) (18-20).For a long time KRAS-mutations were called "untreatable targets".Since early 2022 a specific (second-line) therapy for KRAS G12C has been available in Europe.However, the majority of NSCLC do not harbor the above-mentioned mutations and thus do not qualify for these treatment options.Thus, therapy is still based on different chemotherapy protocols with or without ICIs.
In our study, the mutation frequencies of the four genes (TP53: 45.6%, KRAS: 34.17%, STK11: 16.82%, KEAP1: 14.35%) correspond to the generally described values in LUAD (6).Further, as already shown (5), KRAS-only and KRAS + STK11 mutated patients have the longest mOS (56.5 and 53 months).This seems also true for the quadruple negative group as presented here (64 months).Nevertheless, it must be noticed that the overall survival of quadruple negative patients might be biased due to further common mutations or genetic rearrangements in genes like EGFR, ALK or ROS1 and their already approved targeted therapies.However, the comparable long survival times of these groups, especially for KRAS-only, further underlines the fact that the time has passed to describe KRAS as a prognostically unfavorable factor.Moreover, it seems more appropriate to analyze the complex surrounding mutational landscape, as patients mutated in KRAS + STK11 + KEAP1 + TP53, KRAS + STK11 + KEAP1 and STK11 + KEAP1 + TP53 show the shortest overall survival.So far, several studies described STK11 and KEAP1 alone or co-mutated with KRAS having a negative impact on OS, response to ICI-therapy (4,5,(21)(22)(23)(24)(25)(26) and also across treatment classes independent of immunotherapy (4,5,24,27).However, especially KEAP1 mutations seemed to be the driving factor being the only one significant in a multivariate model (4).This is also reflected in the current analyses of the CodeBreak 100 clinical trial.Here, KEAP1 mutations also appear to be a negative prognostic marker for sotorasib (28).
Therefore, it is particularly interesting that this role seems only true if TP53-mutations are absent, as patients with the combination of KRAS + KEAP1 + TP53 or KEAP1 + TP53 co-mutations show an improved mOS in comparison to KEAP1-only and KRAS + KEAP1 mutated patients.Thus, somewhat surprisingly, in this setting TP53 mutations seem to have a protective effect as long as STK11 is not comutated.So far, survival advantage of TP53 mutations could be shown under therapy with ICI (9, 11-13), however not across treatments.Regardless, it must be noticed that these studies did not include KEAP1 co-mutations into their analyses.
Vice versa, TP53 will have a negative influence on KRAS-only mutated LUAD.Here, the patients' mOS was reduced significantly more than 30% (from 56.5 to 38.3 months, see Table 1, Figure 3A).The negative influence of TP53 mutations were also visible for the following co-mutations: KRAS + STK11, KRAS + STK11 + KEAP1 and STK11 + KEAP1 (Figure 3).Therefore, the potential positive or negative impact of the altered tumor suppressor p53 seems to depend on the surrounding mutational network.This effect had already been described by Saleh et al. and Scalera et al., pointing out that molecular stratification of both alterations should be implemented for localized and advanced-stage NSCLC to optimize and modify clinical decisionmaking (29,30), even though both studies did not include KRAS and/ or STK11 in their investigations.
For KRAS G12C, sotorasib, a targeted therapy, is approved and has shown that its efficacy is influenced by the co-mutations of STK11 and KEAP1.While the co-mutation with STK11 leads to a slightly improved efficacy, KEAP1 and the co-mutation with both genes result in a reduced response (31).Therefore, we performed our analyses in the context of KRAS G12C.Comparable results were observed with a reduced overall survival when KRAS G12C (mOS = 85.7 month) is co-mutated with TP53 (mOS = 36.5;p = 0.02; q = 0.08).The benefits of TP53 co-mutations shown under ICI were subsequently not apparent across treatments (8,9).The reported tendency was not shown in the context of KRAS G12C + STK11 + TP53.Here again, a somewhat protective effect might be discussed for TP53 co-mutations in KRAS G12C + KEAP1 mutated patients, albeit group sizes are small and results not significant.This underlines the importance of the now available G12C-targeted therapy and the need for more druggable options in KRASmutated LUAD and should be kept in mind when interpreting the results of the recently published phase III CodeBreaK 200 trial (32) and the still ongoing phase III CodeBreaK 202 trial (NCT05920356) evaluating sotorasib for the second-line or firstline treatment, respectively.
Finally, we here present a bioinformatic analysis of merged data sets.This might be a limitation at first sight, as we did not refer to our own data.However, the sample sizes of the molecular subgroups, as summarized here, are too small within the single studies to obtain group sizes sufficient for robust statistical results.Thus, integrating database approaches, as presented here, are needed to draw first preliminary conclusions and to generate new hypotheses.These hypotheses must be then tested in future multi-center investigations.This study is further limited by the given annotations.Analyzing the mutation groups in correlation with further clinical variables like age, sex, smoking history, and in particular different treatment patterns is an important task for future studies.Another point of interest is the assessment of progression-free survival in addition to the overall survival.Nevertheless, our study gives important insights into the mutual influence of co-mutations and provides a starting point for future research approaches.
To conclude, the more mutations are analyzed to a greater extent, the greater will be the complexity of the mutational network of lung cancer and cancer in general.In the daily clinical routine setting, referring to panel-based sequencing (as e.g.suggested by the national network of genomic medicine/nNGM) seems mandatory and focusing on different combinations of mutations can help define different prognostic groups and might be the starting point for new treatment strategies.

FIGURE 1
FIGURE 1Distribution of mutation groups in the LUAD dataset (N=4,109).

FIGURE 2 Kaplan
FIGURE 2 Kaplan Meier curve showing overall survival of LUAD patients among all mutation groups.

TABLE 1
Distribution of co-mutations in the genes of KRAS, STK11, KEAP1, and TP53, and overall survival in patients with LUAD.

TABLE 2
Log-rank test comparing overall survival of KRAS G12C mutation groups with and without TP53 co-mutation., false discovery rate; mOS, median overall survival; N, number of patients; NR, not reached. FDR