An in silico evaluation of lorlatinib as a potential therapy for novel amino acid substitutions in the tyrosine kinase domain of the ALK protein associated with cancer

Abstract

The anaplastic lymphoma kinase (alk) gene on chromosome 2 encodes a receptor tyrosine kinase protein essential for key signaling pathways regulating cell proliferation and differentiation. Mutations in alk have been implicated in multiple cancers, including non-small cell lung cancer (NSCLC) and anaplastic large cell lymphoma. While ALK inhibitors have demonstrated efficacy in targeted therapies, resistance due to specific amino acid substitutions requires the development of novel therapeutic strategies. This study aims to identify ALK tyrosine kinase domain mutations using data from the Cancer Genome Atlas and to evaluate the potential of lorlatinib, a third-generation ALK inhibitor, in overcoming these mutations. Using the SIFT and Polyphen-2 algorithms, we identified 53 deleterious ALK mutations associated with different newly recognized cancer types. These mutations were subjected to in silico molecular docking with lorlatinib. Our results indicate strong binding affinities (ranging from −9.4 to −10.8 kcal/mol) across all identified mutations, suggesting a significant interaction between lorlatinib and mutated ALK variants. Furthermore, protein-ligand interaction analysis revealed critical hydrophobic interactions, hydrogen bonds, and essential halogen bonds reinforcing lorlatinib as a potential utility in treating a broader spectrum of ALK-positive tumors beyond NSCLC. This research underscores the importance of repurposing in silico drugs and highlights the need for continued exploration of ALK mutations in cancer therapeutics.

Introduction

The anaplastic lymphoma kinase (alk) gene, located on the short arm of chromosome 2 (2p23), encodes a receptor tyrosine kinase that plays a critical role in regulating signaling pathways involved in cell proliferation, survival, and differentiation, particularly during nervous system development (Huret and Dessen, 2024; Yao et al., 2013).

The ALK fusion gene was first identified in anaplastic large-cell lymphoma (ALCL), and has since been implicated in several types of cancer, including colorectal carcinoma (CRC), inflammatory myofibroblastic tumor (IMT); B-cell lymphoma (BCL); non-small-cell lung cancer (NSCLC); non-Hodgkin’s lymphoma (NHL); neuroblastoma, and other less common malignancies (Holla et al., 2017).

Targeted therapies using tyrosine kinase inhibitors (TKIs) have shown efficacy against ALK-positive cancers, with crizotinib being the first TKI approved by the U.S. Food and Drug Administration (FDA) for NSCLC. However, resistance to crizotinib has emerged due to point mutations or substitutions such as G1269A (Shaw et al., 2011; Doebele et al., 2012). Second-generation TKIs, including ceritinib, were developed to target mutations such as L1196M, G1269A, I1171T, and S1206Y to overcome this resistance. Despite these advancements, novel mutations continue to confer resistance, necessitating the development of more effective therapeutic options (Friboulet et al., 2014). Third-generation TKIs, such as alectinib, brigatinib, and lorlatinib, have demonstrated improved efficacy against resistant ALK variants, with lorlatinib being the most recently TKI approved by the FDA (Wang et al., 2022) (See Supplementary Figure 1).

A significant challenge in treating ALK-positive NSCLC is the extensive mutational variability of the eml4-alk fusion gene and the diversity of amino acid substitutions within the tyrosine kinase domain of the ALK protein (Elshatlawy et al., 2023). However, these mutations may also be present in other types of cancer. Importantly, not all amino acid substitutions act as oncogenic drivers; some are merely passenger mutations or variants of unknown significance (Li et al., 2017). To predict the functional impact of these mutations, computational algorithms such as SIFT (Sorting Intolerant From Tolerant) and PolyPhen-2 (Polymorphism Phenotyping v2) have been developed (Li et al., 2017; Flanagan et al., 2010). SIFT predicts whether an amino acid substitution affects protein function, with scores between 0 and 0.05 indicating deleterious mutations (Ng and Henikoff, 2003). Similarly, PolyPhen-2 assesses the potential impact of amino acid substitutions on protein structure and function, with scores between 0.85 and 1 classified as damaging (Adzhubei et al., 2013).

Given this context, this research aims to analyze data from the Cancer Genome Atlas (TCGA) to identify novel amino acid substitutions within the ALK tyrosine kinase domain that may contribute to oncogenesis. Furthermore, we evaluate the in silico binding potential of lorlatinib against these amino acid substitutions to assess its pharmacological sensitivity and explore its possible therapeutic repurpose in other types of cancer.

Materials and methods

Collection of alk gene somatic mutations

We retrieve primary data from the Cancer Genome Atlas (TCGA) through its official portal https://www.cancer.gov/ccg/research/genome-sequencing/tcga (Center for Cancer Genomics at the National Cancer Institute, 2006). Using the search engine, we queried the alk gene using the search engine and extracted all relevant mutation data. The database was then filtered to include only mutations classified as “deleterious” by SIFT and “damaging” by PolyPhen-2. Additionally, we focused on amino acid positions between 1,117 and 1,392, corresponding to the tyrosine kinase domain of the ALK protein (Huang, 2018). As a result, 53 somatic mutations were identified.

ALK amino acid substitution structure preparation

The structure of the ALK protein in PDB format was obtained from AlphaFold (https://alphafold.ebi.ac.uk/) and identified as AF-Q9UM73-F1-v4. This structure was refined using PyMOL (academic version) (Rigsby and Parker, 2016), focusing on the tyrosine kinase domain. The range of amino acid residues was expanded (from 1,090 to 1,400 residues), and modifications included removing water molecules, adding hydrogen atoms, and incorporating each somatic mutation identified in the TCGA database through mutagenesis. Finally, energy minimization was performed using Swiss PDB Viewer, and all 53 resulting structures were saved in PDB format.

Obtaining ALK inhibitor ligand

The lorlatinib ligand was retrieved from DrugBank (https://go.drugbank.com/) in PDB format for molecular docking and SMILES format for ADME (absorption, distribution, metabolism, and excretion) analysis using SwisssADME (http://www.swissadme.ch/).

Molecular docking of ALK amino acid substitution structures with lorlatinib

Each ALK amino acid substitution structure was uploaded into PyRx - Virtual Screening Tool^® (free version) to serve as the receptor (Dallakyan and Olson, 2015). The lorlatinib ligand was imported into PyRx using the Open Babel extension and underwent energy minimization. Molecular docking simulations were performed using the AutoDock Vina extension within PyRx, with the grid box parameters set to X = −4.01, Y = −1.59, and Z = −24.75 in all 53 ALK-mutated receptors. Finally, wild-type (WT), C1156Y, and L1196M ALK mutations docked with lorlatinib as a reference group were validated by measuring the root mean square deviation “RMSD” (rmsd reference: <2Å) (Mishra and Sharma, 2016), against crystal structures available in the PDB-RSCB repository, including 7R7R, 5A9U, and 4CLJ, respectively.

Binding energy measurements

Binding energy is a key parameter that reflects the affinity of a ligand for its binding site on substrate, with more negative values indicating stronger interactions. This study measured the binding energy (in kilocalories per mol, kcal/mol) for the ligand lorlatinib against each previously prepared ALK mutated structure. Binding energies threshold (−6.8 kcal/mol) were calculated by molecular docking of three PDB-RSCB repository: 7R7R, 5A9U, 4CLJ, and AlphaFold model with ATP ligand retrieved from pubchem (see Supplementary Table 1) (NIH, 2025). Molecular interactions were visualized using Discovery Studio Visualizer (free software) to interpret better the binding behavior (Dassault Systemes, 2019).

Characterization of the protein-ligand interaction profile

Each of the 53 ALK somatic mutation proteins complexed with lorlatinib was saved in pdb format and submitted to the Protein-Ligand Interaction Profiler (PLIP) tool (https://plip-tool.biotec.tu-dresden.de/plip-web/plip/index) (Adasme et al., 2021). The resulting interaction profiles, which classified hydrophobic, hydrogen, and halogen bonds, were downloaded as. pse files, analyzed, and visualized using Pymol.

Statistical analysis

All data were analyzed and charted in GraphPad Prism 10.03, SRplot (https://www.bioinformatics.com.cn/en) (Tang et al., 2023), PyMOL (Rigsby and Parker, 2016), and BIOVIA Discovery Studio Visualizer software (Dassault Systemes, 2019). Binding energies were performed using the AutoDock Vina parameters (Trott and Olson, 2010). Normality testing and nonparametric analysis were performed using the Kolmogorov-Smirnov test to identify statistically significant changes. Finally, the predictive SIFT and PolyPhen-2 scores were correlated with binding energy values using the Spearman correlation test.

Results

Missense mutations in the anaplastic lymphoma kinase (ALK) gene encodes for heterogeneous amino acids substitutions associated with different types of cancer

Given the diversity of point mutations in the ALK oncoprotein that contribute to therapy resistance, we explore the TCGA database. A total of 548 somatic mutations were identified, leading to 341 unique missense protein expressions. Of these, 137 mutations were predicted to be deleterious or damaging in tumorigenesis based on SIFT and PolyPhen-2 scores. Specifically, 53 mutations were located within the tyrosine kinase domain (residues 1,090–1,400) (See Supplementary Figure 2).

The most frequently affected position was 1,174, where phenylalanine was substituted by leucine, cysteine, or isoleucine (n = 4; 7.55%). Other recurrent mutations included position 1,202 (n = 2; 3.77%), where glycine was replaced by glutamic acid or arginine, position 1,209 (n = 2; 3.77%), where arginine was substituted by proline or glutamine, and position 1,212, where arginine was replaced by histidine or cysteine. The remaining point mutations were observed only once (n = 1; 1.89%).

These 53 ALK variants were identified in 77 patients (40 males, 37 females), with the highest frequency in corpus uteri (15.58%), followed by adrenal gland (11.69%), skin (10.39%), lung (9.09%), and colon cancer (9.09%). Some mutations were associated with multiple cancer types, such as P1357H (uterus and stomach), R1120W (uterus and colon), R1181H (colon and cervix), G1121D (colon and breast), R1275Q (lymph node and adrenal gland), F1174L (peritoneum, adrenal gland, and others), and F1174C (kidney and adrenal gland) (See Supplementary Tables 2, 3).

In general, these ALK variants are associated with multiple cancer types. All 53 ALK amino acid substitutions had tumor mutant allele frequency (MAF) values ranging from 0.034 to 0.568, classifying them as rare or less frequent alleles in the population, in the context of TCGA data, MAF or also known as “variant allele frequency” (VAF) refers to the proportion of reads at a specific genomic location that carry a somatic mutation, indicating the percentage of tumour cells harbouring that mutation. Despite their low frequency, all 53 ALK variants were predicted to be deleterious or damaging according to SIFT and PolyPhen-2 scores (Figure 1A). These findings suggest that ALK could be a therapeutic target for cancers beyond non-small cell lung cancer (NSCLC). Notably, lorlatinib, the latest FDA-approved ALK inhibitor for NSCLC (The Asco Post, 2018), has demonstrated favorable pharmacokinetics properties, including ADME (absorption, distribution, metabolism, and excretion) parameters. These properties support its potential repurpose as a therapeutic option for other ALK-driven malignancies (See Figures 1B,C; Supplementary Table 4).

FIGURE 1

Novel somatic variants of the ALK structure exhibit strong binding energy to lorlatinib

Lorlatinib is the latest FDA-approved ALK inhibitor designed to overcome resistance in non-small cell lung cancer (NSCLC). It effectively targets mutations such as C1156Y, I1171 N/S/T, L1196M, and G1202R, which confer resistance to earlier-generation ALK inhibitors (Trott and Olson, 2010). We selected key ALK mutations sensitive to lorlatinib to establish a reference group, such as wild-type (WT), C1156Y, L1196M, and G1202R. This group, termed the “ALK protein group sensitive to lorlatinib,” was used for comparative analysis. Structural data for three of them were obtained from the Protein Data Bank (PDB-RSCB, https://www.rcsb.org/), including: 7R7R (ALK-WT-lorlatinib complex), 5A9U (C1156Y-lorlatinib complex), and 4CLJ (L1196M-lorlatinib complex) (Figure 2A).

FIGURE 2

Given that lorlatinib binds effectively to the ALK protein, we hypothesized that it could also interact with all 53 newly identified ALK variants. However, its efficacy against these novel mutations remained unknown. To address this, we performed molecular docking experiments using ALK protein structures obtained from AlphaFold, introducing the 53 somatic mutations as well as the reference mutations, followed by lorlatinib binding simulations (Figure 2B).

Our results demonstrated that the binding energy of lorlatinib for the ALK-sensitive reference group was: −9.6 kcal/mol (L1196M), −9.7 kcal/mol (ALK wild-type and C1156Y), and −9.9 kcal/mol (G1202R). Additionally, validation against available crystal structures (RSCB-PDB IDs: 7R7R, 5A9U, and 4CLJ) resulted in RMSD values of 0.462, 0.399, and 0.370, respectively, confirming that the docking models closely resemble experimentally determined structures. These findings suggest that strong binding energy correlates with sensitivity to lorlatinib (Figure 2C). Surprisingly, all 53 ALK variants also exhibited strong binding energy, ranging from −9.4 kcal/mol (P1355L) to −10.8 kcal/mol (A1252V) (AutoDock Vina analysis), indicating that lorlatinib maintains a high affinity for these mutations. Moreover, a nonparametric T-test comparison between the 53 ALK variants and the ALK-sensitive reference group showed no significant differences in binding energy (Figure 2D).

These results suggest that lorlatinib could effectively bind to and inhibit these 53 novel ALK somatic mutations, potentially regulating ALK signaling in multiple cancer types beyond NSCLC. The key amino acid residues mediating lorlatinib interactions will be described in the following section (Figures 2D–G).

Lorlatinib binds to ALK variants through strong hydrophobic, hydrogen, and halogen bond interactions

Molecular docking analysis revealed that each ALK variant exhibited a specific interaction pattern with lorlatinib, categorized into three main types: hydrophobic interactions, hydrogen bonds, and halogen bonds. These interactions play a crucial role in stabilizing the ligand within the protein’s active site and may contribute to the drug’s efficacy against novel ALK mutations (Figures 3A,B).

FIGURE 3

Hydrophobic interactions were the most prevalent type of interaction observed across the analyzed variants. Notably, valine at position 1,130 was involved in hydrophobic interactions in all ALK variants (100%), followed by leucine at positions 1,122 and 1,256, which were present in 98% and 84.9% of cases, respectively. Additionally, alanine at position 1,148 contributed to these interactions in 81.1% of the variants, while leucine at position 1,196 was involved in 54.7% of cases. Interesting, two substitutions at position 1,202 (glutamic acid and arginine) also maintained hydrophobic interactions with lorlatinib, suggesting that the overall binding environment of this region remains favorable for drug engagement despite the presence of mutations.

Hydrogen bonds, another key interaction type, were primarily observed at methionine 1,199, which formed hydrogen bonds in 62.3% of the variants. Additionally, leucine 1,122 participated in hydrogen bonding in 54.7% of cases, reinforcing its critical role in lorlatinib binding. Less frequently observed hydrogen bond interactions included histidine 1,124 (six cases), aspartate 1,270 (two cases), aspartate 1,203 (one case), serine 1,206 (one case), and arginine 1,253 (one case). These interactions suggest that lorlatinib maintains a stable binding conformation across a broad range of ALK variants, which could explain the strong binding affinity observed in molecular docking simulations.

In addition to hydrophobic and hydrogen bond interactions, sole ALK variants could also form halogen bonds with lorlatinib due to its fluorine atom. These halogen bonds were primarily established with glutamine at position 1,197, detected in variants A1234T, E1242K, F1245I, K1352N, and P1112Q. Additionally, aspartate at position 1,270 was involved in halogen bonding interactions in A1252V, R1212H, and V1338A. The ability of lorlatinib to form these halogen bonds may further contribute to its binding stability and potential inhibitory effects on mutated ALK proteins.

Taken together, these findings suggest that lorlatinib establishes a robust binding network across the analyzed ALK variants. In particular, leucine at position 1,122 appears to be a key residue, participating in both hydrophobic and hydrogen-bond interactions in most ALK variants.

This could play a critical role in maintaining drug sensitivity, even in the presence of somatic mutations classified as deleterious or damaging by tumor predictor algorithms such as SIFT and PolyPhen-2. These results provide valuable insights into the molecular basis of lorlatinib’s interaction with novel ALK mutations and support its potential therapeutic relevance in a broader range of cancer types beyond non-small cell lung cancer (NSCLC).

The discrepancy between binding affinity scores of ALK variants and somatic mutation predictor scores

All 53 ALK variants were manually generated using PyMOL academic software, following the methodology described in the Methods section. The structural modifications were based on the AlphaFold-derived ALK tyrosine kinase domain template (Figures 4A,B). Two computational predictors of tumorigenicity were used: SIFT, where a score close to or equal to 0 classifies substitutions as deleterious (indicating potential disruption of protein function), and PolyPhen-2, where a score close to or equal to 1 classified a substitution as damaging (suggesting structural and functional impairment of the protein). Additionally, molecular docking analyses using AutoDock Vina revealed binding energy values ranging from −9.6 to −10.8 kcal/mol (Figure 4C).

FIGURE 4

The expected inverse correlation between SIFT and PolyPhen-2 scores was confirmed by Spearman’s correlation test (r = −0.3276, p = 0.0166), indicating a statistically significant negative association between these predictive metrics (Figure 4D). However, when comparing these tumorigenic predictor scores with binding energy values, the correlations were notably weak (r SIFT vs. binding energy = −0.04; r PolyPhen-2 vs. binding energy = −0.03) and statistically nonsignificant (p = 0.791 and p = 0.818, respectively) (Figure 4E).

These results suggest the relationship between computational predictors of oncogenic potential and molecular properties. Although SIFT and PolyPhen-2 scores predicted that these ALK substitutions could contribute to tumorigenesis, the strong binding affinity of lorlatinib indicates its potential to effectively target these ALK variants, possibly mitigating their oncogenic effects (See Supplementary 4).

Discussion

Amino acid point mutations in the ALK protein are critical in conferring resistance to ALK inhibitor therapies in cancer. Notably, mutations such as G2032R and I1151Tins are associated with resistance to the first-generation inhibitor crizotinib. Conversely, mutations such as C1156T/Y, L1198F, D1203N, and G1202R have been linked to resistance against second-generation inhibitors, including ceritinib, alectinib, and brigatinib. Furthermore, third-generation inhibitors like lorlatinib exhibit resistance to mutations such as L1198F, G2032R, D1203N, and G1123D (Poei et al., 2024). Despite these challenges, lorlatinib has demonstrated significant therapeutic efficacy in overcoming drug resistance (Poei et al., 2024; Johnson et al., 2014).

In this study, we identified 53 amino acid substitutions in the ALK protein using the ATCG database. We classified them as potential oncogenic drivers based on predictive tools such as SIFT and PolyPhen-2 scores. These mutations have been associated with various cancers, including NSCLC, anaplastic large-cell lymphoma, neuroblastoma, and colorectal carcinoma. Moreover, our findings suggest novel associations of these mutations with additional cancer types, including skin, retroperitoneal, peritoneal, pancreatic, kidney, esophageal, breast, bladder, and adrenal tumors (Holla et al., 2017; Adzhubei et al., 2010; Bresler et al., 2014).

Furthermore, lorlatinib has demonstrated the ability to target multiple ALK mutations, including C1156Y, I1171 N/S, F1174C, L1196M, L1198F, G1202R, D1203N, E1210K, and G1269A, showcasing sensitivity in NSCLC patients, even those resistant to prior-generation ALK inhibitors. This suggests that lorlatinib may potentially address a broader spectrum of ALK amino acid substitutions, positioning it as a promising candidate for repurposing therapies in various types of cancer (Bertacca et al., 2023). Our hypothesis was further validated by binding energy calculations for all 53 ALK variants revealed values comparable to those of experimentally validated ALK wild-type (WT), C1156Y, L1196M, and G1202R mutations, which are known to be sensitive to lorlatinib (Poei et al., 2024; Gainor et al., 2016).

Among the 53 identified ALK amino acid substitutions, three have been reported to interact with the active site of ALK protein. We identified two structures containing the F1174L mutation (PDB codes: 2YJR and 4FNW) and four structures for R1275Q (PDB codes: 4FNY, 4FNX, 4FNZ, and 4FNW), both implicated in neuroblastoma (Epstein et al., 2012; Sasaki et al., 2010). Additionally, a structure harboring the G1202R mutation (PDB code: 9GBE) was identified in complex with the NVL-655 ALK inhibitor for NSCLC, suggesting that lorlatinib may also target this variant (Lin et al., 2024). Furthermore, a case report in NSCLC demonstrated that the ALK I1171N mutation, along with other ALK amino acids substitutions identified in this study, conferred resistance to ensartinib but remained sensitive to lorlatinib, further supporting the efficacy of lorlatinib against multiple ALK amino acids substitutions (Ye and Guo, 2023). The remaining 49 ALK amino acids substitutions may also contribute to oncogenesis, as their oncogenic predictor scores suggest potential alterations in the ALK protein. Although these mutations have not yet been documented in the literature, they are available in the TCGA database and have been associated with different types of cancer, with some mutations occurring in more than one cancer type.

Given that lorlatinib has demonstrated efficacy against multiple ALK amino acids substitutions, our results suggest that it could target all 53 identified substitutions, as indicated by binding energy calculations and protein-ligand interaction profiles. Previous studies have reported binding energy values of −8.9, −8.6, and −8.4 kcal/mol for ALK-WT, F1174C, and F1174L, respectively, when interacting with lorlatinib. In contrast, our results for the same mutations yielded a stronger binding energy of −9.7 kcal/mol. Similarly, binding energy values ranging from −9.6 to −10.8 kcal/mol were observed for the remaining ALK amino acid substitutions analyzed in this study (Balasundaram and Doss, 2023), further supporting lorlatinib’s potential as a versatile therapeutic option.

The strong binding energy affinities observed for all 53 ALK amino acid substitutions can be attributed to key stabilizing interactions, including hydrophobic and hydrogen bond interactions, so they play an important role because they stabilize the protein-ligand energy (Varma et al., 2010). Specifically, lorlatinib exhibited strong hydrophobic interactions with Leu 1,122, Val 1,130, Ala 1,148, Leu 1,196, and Leu 1,256, as well as hydrogen bonds with Leu 1,122 and Met 1,199. Additionally, halogen bonds with Glu 1,197 and ASP 1270 were identified for substitutions lacking hydrogen bonds, which are critical in drug design as they enhance protein-ligand interactions. Fluorine and chlorine atoms are frequently incorporated into drug structures to improve physicochemical properties (Shinada et al., 2019; Margiotta et al., 2020), which may explain why ALK amino acid substitutions such as A1234T, A1252V, E1242K, F1245I, K1352N, P1112Q, R1212H, and V1338A exhibit strong binding to lorlatinib, as its molecular conformation includes a fluorine atom.

In the TCGA database, 541 somatic ALK amino acid substitutions were identified; however, not all were predicted to be oncogenic drivers. Only 53 were classified as tumorigenic based on SIFT and Polyphen-2 predictor scores within the ALK tyrosine kinase domain. These predictors, which range from 0 to 1, assess the likelihood that an amino acid substitution will impact protein function. SIFT scores between 0 and 0.05 indicate deleterious effects, while PolyPhen-2 scores from 0.85 to 1 classify substitutions as damaging (Ng and Henikoff, 2003; Adzhubei et al., 2013). In our analysis, all 53 ALK amino acid substitutions exhibited SIFT scores close to 0 and PolyPhen-2 scores near 1, confirming their deleterious and damaging nature, respectively. Previous studies have proposed a combined model using both predictors to assess non-synonymous variants, showing a positive correlation between 1-SIFT scores and PolyPhen-2 scores (Wei et al., 2011). However, in our study, we observed a negative correlation, which may be attributed to the direct comparison of PolyPhen-2 scores with the original SIFT scores rather than their complementary 1-SIFT values.

While the SIFT and PolyPhen-2 predictors did not exhibit a strong correlation with binding energy, our analysis revealed a weak negative correlation (See Figure 4E), likely due to random variations. Nevertheless, the binding energies to lorlatinib suggest that lorlatinib remains effective against all 53 ALK amino acid substitutions classified as deleterious or damaging in cancer. These findings support the potential repurpose of lorlatinib to target additional ALK-driven cancers beyond NSCLC, as indicated by in silico simulations.

In this study, we acknowledge a limitation in relying solely on in silico tools such as SIFT and PolyPhen-2 for assessing the oncogenic potential of ALK genomic alterations. While these tools offer valuable predictive insights into variant pathogenicity, they do not encompass the full spectrum of clinical oncogenicity criteria. Future research should explore the broader context of ALK interactions and functional consequences to enhance understanding of their oncogenic roles.

To verify the predictive potential of pathogenicity, we conducted a manual search on the Alphamissense portal: https://alphamissense.hegelab.org/search (Author anonymous, 2025; Minton, 2023). Through this approach, we confirmed that all the 53 genetic variants we proposed were classified as likely pathogenic, with the majority exceeding a confidence score of 0.89 (See Supplementary Table 5). Additionally, we evaluated the amino acid changes associated with oncogenicity or treatment resistance using the OncoKB™ database (Suehnholz et al., 2024; Chakravarty et al., 2017), which is one of the most reputable platforms for genetic variant curation, adhering to rigorous standards of somatic variants classification (Horak et al., 2022).

On the other hand, while the exact impact on ATP-binding or kinase activation for these specific novel substitutions requires experimental validation, their classification as “likely pathogenic” suggests they may contribute to aberrant ALK signaling. Several identified mutations have known biological and clinical implications according to OncoKB™ (https://www.oncokb.org/gene/ALK) (Chakravarty et al., 2017; ALK, 2017). The ALK I1171N (p.Ile1171Asn) mutation is classified as “Likely Oncogenic” and is clinically relevant due to its resistance to first- and second-generation ALK inhibitors (crizotinib, ceritinib, and alectinib), while notably retaining sensitivity to brigatinib and lorlatinib (ALK, 2017; Ceccon et al., 2013). This highlights a critical mechanism of acquired drug resistance. Similarly, ALK F1174L (p.Phe1174Leu) is an “Oncogenic” mutation associated with resistance to crizotinib, ceritinib, and alectinib, but shows sensitivity to lorlatinib and brigatinib. The ALK F1174C (p.Phe1174Cys) mutation, also “Likely Oncogenic,” exhibits resistance to crizotinib and ceritinib but remains sensitive to alectinib, brigatinib, and lorlatinib. Conversely, ALK F1174I (p.Phe1174Ile), another “Likely Oncogenic” variant, displays a broader sensitivity to crizotinib, ceritinib, alectinib, and lorlatinib. These F1174 variants underscore the diverse therapeutic challenges and opportunities within a single amino acid position (see Supplementary 7) (ALK, 2017; Cheng et al., 2023).

Furthermore, the ALK G1202R (p.Gly1202Arg) mutation is categorized as a “Resistance” mutation, confirming its role in resistance to crizotinib, ceritinib, alectinib, and brigatinib, while being sensitive to lorlatinib. This well-characterized mutation is a key driver of acquired resistance to several ALK TKIs. In contrast, the ALK A1200V (p.Ala1200Val) and ALK E1242K (p.Glu1242Lys) mutations are both deemed “Likely Neutral” by OncoKB, suggesting they may not significantly impact ALK function or drug response. Lastly, the ALK R1275Q (p.Arg1275Gln) mutation is identified as “Oncogenic” and has demonstrated sensitivity to crizotinib and lorlatinib in both in vitro and in vivo studies, indicating its potential as a targetable alteration (see Supplementary 7) (ALK, 2017; Cheng et al., 2023).

The consistent sensitivity of several ALK mutations demonstrated in our in silico experiments, including I1171N, F1174L, F1174C, F1174I, G1202R, and R1275Q, to lorlatinib, as highlighted by our findings and supported by OncoKB classifications, strongly suggests the potential for therapeutic repurposing of this third-generation ALK inhibitor beyond its current primary indication in NSCLC. Lorlatinib’s known ability to overcome common resistance mutations that emerge from earlier-generation ALK TKIs, coupled with its excellent central nervous system (CNS) penetration, makes it a highly promising candidate for other ALK-driven malignancies, particularly those with a propensity for CNS metastases, such as neuroblastoma. This work influences future preclinical and clinical investigations by providing a rationale to explore lorlatinib’s efficacy in a broader spectrum of ALK-driven cancers.

Future steps should include: (Huret and Dessen, 2024): Preclinical validation: detailed in vitro and in vivo studies in diverse ALK-driven cancer models (e.g., specific lymphomas, inflammatory myofibroblastic tumors, or other solid tumors where ALK fusions or activating mutations are identified) harboring these specific sensitive mutations, to thoroughly characterize its anti-tumor activity and optimal dosing; (Yao et al., 2013); Biomarker-driven clinical trials: initiating basket or umbrella clinical trials specifically enriching for patients with these lorlatinib-sensitive ALK mutations, regardless of cancer type, to evaluate clinical response rates and safety in a real-world setting; and (Holla et al., 2017) Investigation of combination therapies: exploring rational combinations of lorlatinib with other targeted agents or conventional therapies to potentially overcome emergent resistance mechanisms or enhance therapeutic efficacy. By leveraging the comprehensive genomic and functional data presented here, a more precise and personalized approach to treating ALK-driven cancers can be developed, potentially improving patient outcomes in a wider range of malignancies.

In conclusion, our study identified novel amino acid substitutions in the ALK tyrosine kinase domain associated with cancers beyond hematological malignancies and NSCLC. Importantly, we demonstrated that lorlatinib retains efficacy against these mutations, suggesting its potential for therapeutic repurposing in other ALK-mutated cancers. This finding persists despite SIFT and PolyPhen-2 predictions of oncogenic progression, positioning lorlatinib as a promising candidate for broader clinical applications.

References

• 1

  AdasmeM. F.LinnemannK. L.BolzS. N.KaiserF.SalentinS.HauptV. J.et al (2021). PLIP 2021: expanding the scope of the protein-ligand interaction profiler to DNA and RNA. Nucleic Acids Res.49 (W1), W530–W534. 10.1093/nar/gkab294

  • CrossRef
  • Google Scholar
  • View reference in article

• 2

  AdzhubeiI.JordanD. M.SunyaevS. R. (2013). Predicting functional effect of human missense mutations using PolyPhen-2. Curr. Protoc. Hum. Genet.7 (6). 10.1002/0471142905.hg0720s76

  • CrossRef
  • Google Scholar
  • View reference in article

• 3

  AdzhubeiI. A.SchmidtS.PeshkinL.RamenskyV. E.GerasimovaA.BorkP.et al (2010). A method and server for predicting damaging missense mutations. Nat. Methods7, 248–249. 10.1038/nmeth0410-248

  • CrossRef
  • Google Scholar
  • View reference in article

• 4

  ALK (2017). OncoKB^TM. Available online at: https://www.oncokb.org/gene/ALK (Accessed May 22, 2022).

  • Google Scholar
  • View reference in article

• 5

  Author anonymous (2025). AlphaMissense. Available online at: https://alphamissense.hegelab.org/search (Accessed May 22, 2022).

  • Google Scholar
  • View reference in article

• 6

  BalasundaramA.DossG. P. C. (2023). A computational examination of the therapeutic advantages of fourth-generation ALK inhibitors TPX-0131 and repotrectinib over third-generation lorlatinib for NSCLC with ALK F1174C/L/V mutations. Front. Mol. Biosci.10, 1306046. 10.3389/fmolb.2023.1306046

  • CrossRef
  • Google Scholar
  • View reference in article

• 7

  BertaccaI.PegoraroF.TondoA.FavreC. (2023). Targeted treatment of solid tumors in pediatric precision oncology. Front. Oncol.13, 117679010.3389/fonc.2023.1176790

  • CrossRef
  • Google Scholar
  • View reference in article

• 8

  BreslerS. C.WeiserD. A.HuweP. J.ParkJ. H.KrytskaK.RylesH.et al (2014). ALK mutations confer differential oncogenic activation and sensitivity to ALK inhibition therapy in neuroblastoma. Cancer Cell26 (5), 682–694. 10.1016/j.ccell.2014.09.019

  • CrossRef
  • Google Scholar
  • View reference in article

• 9

  CecconM.MologniL.BissonW.ScapozzaL.Gambacorti-PasseriniC. (2013). Crizotinib-resistant NPM-ALK mutants confer differential sensitivity to unrelated Alk inhibitors. Mol. Cancer Res.11 (2), 122–132. 10.1158/1541-7786.MCR-12-0569

  • CrossRef
  • Google Scholar
  • View reference in article

• 10

  Center for Cancer Genomics at the National Cancer Institute (2006). Cancer Genome Atlas Program (TCGA). Available online at: https://www.cancer.gov/ccg/research/genome-sequencing/tcga (Accessed February 08, 2025).

  • Google Scholar
  • View reference in article

• 11

  ChakravartyD.GaoJ.PhillipsS.KundraR.ZhangH.WangJ.et al (2017). OncoKB: a precision oncology knowledge base. JCO Precis. Oncol.2017 (1), 1–16. 10.1200/PO.17.00011

  • CrossRef
  • Google Scholar
  • View reference in article

• 12

  ChengJ.NovatiG.PanJ.BycroftC.ŽemgulyteA.ApplebaumT.et al (2023). Accurate proteome-wide missense variant effect prediction with AlphaMissense. Sci. (80).381 (6664), eadg7492. 10.1126/science.adg7492

  • CrossRef
  • Google Scholar
  • View reference in article

• 13

  DallakyanS.OlsonA. J. (2015). Small-molecule library screening by docking with PyRx. Methods Mol. Biol.1263, 243–250. 10.1007/978-1-4939-2269-7_19

  • CrossRef
  • Google Scholar
  • View reference in article

• 14

  Dassault Systemes (2019). BIOVIA, discovery studio visualizer, release. San Diego: Dassault Systèmes. Available online at: https://www.3ds.com/products/biovia/discovery-studio/visualization February 08, 2025).

  • Google Scholar
  • View reference in article

• 15

  DoebeleR. C.PillingA. B.AisnerD. L.KutateladzeT. G.LeA. T.WeickhardtA. J.et al (2012). Mechanisms of resistance to crizotinib in patients with ALK gene rearranged non-small cell lung cancer. Clin. Cancer Res.18 (5), 1472–1482. 10.1158/1078-0432.CCR-11-2906

  • CrossRef
  • Google Scholar
  • View reference in article

• 16

  ElshatlawyM.SampsonJ.ClarkeK.BaylissR. (2023). EML4-ALK biology and drug resistance in non-small cell lung cancer: a new phase of discoveries. Mol. Oncol.17 (6), 950–963. 10.1002/1878-0261.13446

  • CrossRef
  • Google Scholar
  • View reference in article

• 17

  EpsteinL. F.ChenH.EmkeyR.WhittingtonD. A. (2012). The R1275Q neuroblastoma mutant and certain ATP-competitive inhibitors stabilize alternative activation loop conformations of anaplastic lymphoma kinase. J. Biol. Chem.287 (44), 37447–37457. 10.1074/jbc.M112.391425

  • CrossRef
  • Google Scholar
  • View reference in article

• 18

  FlanaganS. E.PatchA. M.EllardS. (2010). Using SIFT and PolyPhen to predict loss-of-function and gain-of-function mutations. Genet. Test. Mol. Biomarkers14 (4), 533–537. 10.1089/gtmb.2010.0036

  • CrossRef
  • Google Scholar
  • View reference in article

• 19

  FribouletL.LiN.KatayamaR.LeeC. C.GainorJ. F.CrystalA. S.et al (2014). The ALK inhibitor ceritinib overcomes crizotinib resistance in non-small cell lung cancer. Cancer Discov.4 (6), 662–673. 10.1158/2159-8290.CD-13-0846

  • CrossRef
  • Google Scholar
  • View reference in article

• 20

  GainorJ. F.DardaeiL.YodaS.FribouletL.LeshchinerI.KatayamaR.et al (2016). Molecular mechanisms of resistance to first- and second-generation ALK inhibitors in ALK -rearranged lung cancer. Cancer Discov.6 (10), 1118–1133. 10.1158/2159-8290.CD-16-0596

  • CrossRef
  • Google Scholar
  • View reference in article

• 21

  HollaV. R.ElaminY. Y.BaileyA. M.JohnsonA. M.LitzenburgerB. C.KhotskayaY. B.et al (2017). ALK: a tyrosine kinase target for cancer therapy. Mol. Case Stud.3 (1), a001115. 10.1101/mcs.a001115

  • CrossRef
  • Google Scholar
  • View reference in article

• 22

  HorakP.GriffithM.DanosA. M.PitelB. A.MadhavanS.LiuX.et al (2022). Standards for the classification of pathogenicity of somatic variants in cancer (oncogenicity): joint recommendations of clinical Genome resource (ClinGen), cancer genomics consortium (CGC), and variant interpretation for cancer consortium (VICC). Genet. Med.24 (5), 986–998. 10.1016/j.gim.2022.01.001

  • CrossRef
  • Google Scholar
  • View reference in article

• 23

  HuangH. (2018). Anaplastic lymphoma kinase (Alk) receptor tyrosine kinase: a catalytic receptor with many faces. Int. J. Mol. Sci.19 (11), 3448. 10.3390/ijms19113448

  • CrossRef
  • Google Scholar
  • View reference in article

• 24

  HuretJ.-L.DessenP. (2024). CAD/ALK (2p23). Atlas Genet. Cytogenet Oncol. Haematol.Available online at: https://atlasgeneticsoncology.org/solid-tumor/105197/cad-alk-(2p23 (Accessed October 29, 2024).

  • Google Scholar
  • View reference in article

• 25

  JohnsonT. W.RichardsonP. F.BaileyS.BroounA.BurkeB. J.CollinsM. R.et al (2014). Discovery of (10R)-7-amino-12-fluoro-2,10,16-trimethyl-15-oxo-10,15,16,17-tetrahydro-2H-8,4-(metheno)pyrazolo[4,3-h] [2,5,11]-benzoxadiazacyclotetradecine-3-carbonitrile (PF-06463922), a macrocyclic inhibitor of anaplastic lymphoma kinase (ALK) and c-ros oncogene 1 (ROS1) with preclinical brain exposure and broad-spectrum potency against ALK-resistant mutations. J. Med. Chem.57 (11), 4720–4744. 10.1021/jm500261q

  • CrossRef
  • Google Scholar
  • View reference in article

• 26

  LiM. M.DattoM.DuncavageE. J.KulkarniS.LindemanN. I.RoyS.et al (2017). Standards and guidelines for the interpretation and reporting of sequence variants in cancer: a joint consensus recommendation of the association for molecular pathology, American society of clinical oncology, and college of American pathologists. J. Mol. Diagnostics19, 4–23. 10.1016/j.jmoldx.2016.10.002

  • CrossRef
  • Google Scholar
  • View reference in article

• 27

  LinJ. J.HoranJ. C.TangpeerachaikulA.SwalduzA.ValdiviaA.JohnsonM. L.et al (2024). NVL-655 is a selective and brain-penetrant inhibitor of diverse ALK-mutant oncoproteins, including lorlatinib-resistant compound mutations. Cancer Discov.14 (12), 2367–2386. 10.1158/2159-8290.CD-24-0231

  • CrossRef
  • Google Scholar
  • View reference in article

• 28

  MargiottaE.Van Der LubbeS. C. C.De Azevedo SantosL.ParagiG.MoroS.BickelhauptF. M.et al (2020). Halogen bonds in ligand-protein systems: molecular orbital theory for drug design. J. Chem. Inf. Model60 (3), 1317–1328. 10.1021/acs.jcim.9b00946

  • CrossRef
  • Google Scholar
  • View reference in article

• 29

  MintonK. (2023). Predicting variant pathogenicity with AlphaMissense. Nat. Rev. Genet.24, 804. 10.1038/s41576-023-00668-9

  • CrossRef
  • Google Scholar
  • View reference in article

• 30

  MishraG. P.SharmaR. (2016). Identification of potential PPAR γ agonists as hypoglycemic agents: molecular docking approach. Interdiscip. Sci. – Comput. Life Sci.8 (3), 220–228. 10.1007/s12539-015-0126-7

  • CrossRef
  • Google Scholar
  • View reference in article

• 31

  NgP. C.HenikoffS. (2003). SIFT: predicting amino acid changes that affect protein function. Nucleic Acids Res.31 (13), 3812–3814. 10.1093/nar/gkg509

  • CrossRef
  • Google Scholar
  • View reference in article

• 32

  NIH (2025). alpha,beta-Methyleneadenosine 5’-triphosphate | C11H18N5O12P3 | CID 91557 - PubChem. Available online at: https://pubchem.ncbi.nlm.nih.gov/compound/91557 (Accessed May 25, 2021).

  • Google Scholar
  • View reference in article

• 33

  PoeiD.AliS.YeS.HsuR. (2024). ALK inhibitors in cancer: mechanisms of resistance and therapeutic management strategies. Cancer drug Resist Alhambra, Calif.7, 20. 10.20517/cdr.2024.25

  • CrossRef
  • Google Scholar
  • View reference in article

• 34

  RigsbyR. E.ParkerA. B. (2016). Using the PyMOL application to reinforce visual understanding of protein structure. Biochem. Mol. Biol. Educ.44 (5), 433–437. 10.1002/bmb.20966

  • CrossRef
  • Google Scholar
  • View reference in article

• 35

  SasakiT.OkudaK.ZhengW.ButrynskiJ.CapellettiM.WangL.et al (2010). The neuroblastoma-associated F1174L ALK mutation causes resistance to an ALK kinase inhibitor in ALK-translocated cancers. Cancer Res.70 (24), 10038–10043. 10.1158/0008-5472.CAN-10-2956

  • CrossRef
  • Google Scholar
  • View reference in article

• 36

  ShawA. T.SolomonB.Mino-KenudsonM. (2011). Crizotinib and testing for ALK. J. Natl. Compr. Canc Netw.9 (12), 1335–1341. 10.6004/jnccn.2011.0115

  • CrossRef
  • Google Scholar
  • View reference in article

• 37

  ShinadaN. K.De BrevernA. G.SchmidtkeP. (2019). Halogens in protein-ligand binding mechanism: a structural perspective. J. Med. Chem. J Med Chem62, 9341–9356. 10.1021/acs.jmedchem.8b01453

  • CrossRef
  • Google Scholar
  • View reference in article

• 38

  SuehnholzS. P.NissanM. H.ZhangH.KundraR.NandakumarS.LuC.et al (2024). Quantifying the expanding landscape of clinical actionability for patients with cancer. Cancer Discov.14 (1), 49–65. 10.1158/2159-8290.CD-23-0467

  • CrossRef
  • Google Scholar
  • View reference in article

• 39

  TangD.ChenM.HuangX.ZhangG.ZengL.ZhangG.et al (2023). SRplot: a free online platform for data visualization and graphing. PLoS One18 (11 October), e0294236. 10.1371/journal.pone.0294236

  • CrossRef
  • Google Scholar
  • View reference in article

• 40

  The Asco Post (2018). FDA approves lorlatinib for second- or third-line treatment of ALK-positive metastatic NSCLC. Case Med. Res.Available online at: https://ascopost.com/issues/november-25-2018/fda-approves-lorlatinib-for-second-or-third-line-treatment-of-nsclc/(Accessed February 11, 2025). 10.31525/fda1-ucm625027.htm

  • CrossRef
  • Google Scholar
  • View reference in article

• 41

  TrottO.OlsonA. J. (2010). AutoDock Vina: improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading. J. Comput. Chem.31 (2), 455–461. 10.1002/jcc.21334

  • CrossRef
  • Google Scholar
  • View reference in article

• 42

  VarmaA. K.PatilR.DasS.StanleyA.YadavL.SudhakarA. (2010). Optimized hydrophobic interactions and hydrogen bonding at the target-ligand interface leads the pathways of Drug-Designing. PLoS One5 (8), e12029. 10.1371/journal.pone.0012029

  • CrossRef
  • Google Scholar
  • View reference in article

• 43

  WangL.ShengZ.ZhangJ.SongJ.TengL.LiuL.et al (2022). Comparison of lorlatinib, alectinib and brigatinib in ALK inhibitor–naive/untreated ALK-positive advanced non-small-cell lung cancer: a systematic review and network meta-analysis. J. Chemother. J Chemother34, 87–96. 10.1080/1120009X.2021.1937782

  • CrossRef
  • Google Scholar
  • View reference in article

• 44

  WeiP.LiuX.FuY. X. (2011). Incorporating predicted functions of nonsynonymous variants into gene-based analysis of exome sequencing data: a comparative study. BMC Proc.5 (Suppl. 9), S20–S26. 10.1186/1753-6561-5-S9-S20

  • CrossRef
  • Google Scholar
  • View reference in article

• 45

  YaoS.ChengM.ZhangQ.WasikM.KelshR.WinklerC. (2013). Anaplastic lymphoma kinase is required for neurogenesis in the developing central nervous system of zebrafish. PLoS One8 (5), 63757. 10.1371/journal.pone.0063757

  • CrossRef
  • Google Scholar
  • View reference in article

• 46

  YeZ.GuoJ. (2023). Acquired ALK G1202R-ALK I1171N-or EML4-ALK-mediated resistance to ensartinib in lung adenocarcinoma but responded to lorlatinib: a case report. Front. Oncol.13, 1082115. 10.3389/fonc.2023.1082115

  • CrossRef
  • Google Scholar
  • View reference in article