NOS3 and CTH gene mutations as new molecular markers for detection of lung adenocarcinoma

Human LUAD specimens

Thirty-one FFPE tissue samples of LUAD patients, of Kurdish nationality, were removed by microtome from patients who underwent surgical treatment. The microscopic morphology of tumor cells was used to classify lung cancer. The tumor node metastasis (TNM) staging system, 8th edition, was used in the staging method. Hematoxylin and eosin (H&E) staining, a regularly used method, was applied to analyze the structure and features of tissues. The T stage (size and extent of the main tumor), N stage (involvement of regional lymph nodes), and the existence of metastases (M stage) in lung cancer may all be confirmed by H&E staining (Lababede & Meziane, 2018). For this purpose, the ethical and regulatory issues related to human specimen collection for research purposes have been performed in compliance with the Declaration of Helsinki; approved by the Human Ethics Research Committee in the reference number (4/5/337) of the College of Science, Salahaddin University-Erbil. All patients were provided signed informed consent declaring to investigate the tissue materials. Patients with other co-morbid health problems which could introduce heterogeneity to the sample, such as additional acquired heart diseases, hypertension, diabetes mellitus, arthritis, ankylosing spondylitis, connective tissue diseases and other inflammatory diseases were excluded. Before the molecular investigation, tissue specimens were histologically confirmed. FFPE tissue samples were collected and stored at room temperature (20–22 °C) until subjected to DNA extraction and genotyping.

NOS3 and CTH mutation analysis by Sanger sequencing

The procedure of DNA extraction, amplification and sequencing has been previously described by Marangoni et al. (2008). In summary, using a commercial kit and following the manufacturer’s instructions, DNA was extracted and purified from 31 sectioned FFPE of LUAD tissue samples. Each sample block was 10 µm in diameter and divided into ten equal sections (ReliaPrepTM FFPE gDNA Miniprep System; Promega, Madison, WI, USA). The DNA content was then measured using a Thermo Scientific Nanodrop 1000 spectrophotometer, and the purity of the DNA was determined using an optimal A260/A280 absorbance ratio of 1.8. The study investigated two exons from the NOS3 gene (2 and 10) and two exons from the CTH gene (8 and 11). Purified DNA was amplified individually for each genetic variant using PCR on a Techne TC-5000 gradient thermal cycler (Cole-Parmer, Ltd.) using the primers listed below: NOS3 (first primer, exon 2) forward, AGG CCC TAT GGT AGT GCC TTT and reverse, TCT CTT AGT GCT GTG GTC AC; NOS3 (second primer, exon 10) AAG GCA GGA GAC AGT GGA TG and reverse, TGA AGG AAG AGT TCT GGT GGA; CTH (first primer, exon 8) forward, GGA CTT CTT GAG GAG TTG AAG C and reverse, ATT CTC ACC TCC TTC AGA GGC; and CTH (second primer, exon 11) forward, CTCCACTGACACAATATAGTGATCA and reverse, TGAAGGCAGCAGTAAGTCTTATT.

A ready-to-use master mix (Promega) has been used, containing Taq DNA polymerase, dNTPs, KCl, and reaction buffer. According to Abid, Qadir & Salihi (2021), the PCR thermocycling conditions consisted of the following: initial denaturation at 95 °C for 5 min, followed by 35 cycles of 95 °C for 30 s, annealing at 55 °C for 30 s, and elongation at 72 °C for 1 min, followed by a final extension step at 72 °C for 5 min. PCR products were separated using 2% agarose gel electrophoresis and compared with a 100 bp DNA marker (Froggabio, USA). The DNA marker included fragments of 400 bp from exon 8 of the CTH gene, 500 bp from exon 11 of the CTH gene, 300 bp from exon 2 of the NOS3 gene, and 600 bp from exon 10 of the NOS3 gene. Prior to casting into the tray, the gel was stained with safe dye (Safe DNA Gel Stain dye; Add Bio, Inc.). A gel documentation system (UV Transilluminator UST-20M-8K; Biostep GmbH) was used to visualise the gels.

Following the PCR procedure, the product has been sent to Macrogene, South Korea (using the same forward and reverse primer for each specified region in the gene) on an automated 3130 genetic analyzer (Applied Biosystems; Thermo Fisher Scientific, Inc.). Sequence data have been deposited to a curated data repository (GenBank database; http://www.ncbi.nlm.nih.gov/genbank/). The nucleotide sequence data stated are available in the GenBank database under the accession numbers OR750458.1 and OR757575. The Sanger sequencing data were analysed using the Mutation Surveyor software 5.1 (SoftGenetics, LLC) and compared to gnomAD, dbSNP, ClinVar, and COSMIC. The Mutation Surveyor programme results will be supplied as a file tree frame and a mutation report (output report table, ORT). A variety of ORTs can be generated using the “Reports” menu. An ORT indicates all of the detected mutations and connects them to their relevant sequence traces. One of the most crucial reports is the clinical report. It is a personalized report with choices for providing exclusively sample files with mutation calls or all locations with provided annotation information, including negative SNPs. For each identified mutation, the report shows one pair of sample files in a contig per page and provides the mutation code, Variant Percentage, and shared mutations in patients.

EGFR mutation and PD-L1 expression

For the EGFR mutation and PD-L1 expression tests, 110 FFPE samples of Kurdish nationality LUAD patients were used. Of these patients, 77 (70%) were men and 33 (30%) were women, respectively, with the following ages: Thirteen (11.82%) of the participants were under the age of fifty, followed by 29 (26.36%) in the 51–60 age range, 34 (30.91%) in the 61–70 age range, and 34 (30.91%) over the age of seventy.

According to the instructions in the product insert and as previously detailed by Angulo et al. (2012), the EGFR mutation was examined using the therascreen EGFR PCR kit (Qiagen, Inc., Tokyo, Japan) on the Rotor-Gene Q real-time instrument (Corbett Research, Qiagen GmbH). For each specimen, eight master mixes were prepared for testing on the Rotor-Gene Q machine. The real-time kit examines for deletions in exons 19 and 21 (it identifies the existence of deletions but is unable to differentiate between them). In addition to the mutation reactions, the kit provides a control gene reaction to assess the quality and quantity of the products. Real-time PCR testing takes around 4 hrs to complete. The EGFR gene testing methods have already been confirmed in the laboratory, and all reactions included negative and positive tissue controls.

As previously reported by Roach et al. (2016), automated staining techniques validated for the programmed death-ligand 1 (PD-L1) or cluster of differentiation 274 (CD274) immunohistochemistry (IHC) 22C3 pharmDx test were employed for IHC staining. In a 3-in-1 procedure, the PT Link (Dako PT100) was employed for deparaffinization, rehydration, and target retrieval. The specimens were first incubated with a monoclonal mouse anti-human PD-L1 antibody, clone 22C3, or a negative control reagent, mouse immunoglobulin G isotype control, before being exposed to an anti-mouse linker antibody specific to the primary antibody’s host species and a ready-to-use visualisation reagent composed of the secondary antibody and horseradish peroxidase molecules coupled to a dextran polymer back. Following the addition of 3,30-diaminobenzidine tetrahydrochloride chromogen, followed by 3,30-diaminobenzidine tetrahydrochloride enhancer, a visible reaction product precipitated at the antigen site. The specimens were then counterstained with hematoxylin and coverslipped. The data were examined using a light microscope. The sensitivity of the final test was optimised with the least amount of nonspecific staining by adjusting the primary antibody concentration and reagent incubation periods. Positivity was defined as entire circumferential or partial cell membrane staining of live tumour cells with 1+ to 3+ intensity to determine PD-L1 protein expression. Nonspecific staining was rated in increments of 0.25 on a 0 to 3 intensity scale. PD-L1 was not used to identify immune cells associated with tumours. Cytoplasmic staining was disregarded when computing the score if it was present. The score (TPS) was calculated using the percentage of PD-L1-positive tumour cells compared to total tumour cells in the denominator. The negative control reagent must result in no specific membrane staining and nonspecific (nonmembrane) staining of 1+ intensity in LUAD specimens.

Mutation retrieval from databases and interaction between genes

Mutations in the selected genes across various cancer types were retrieved using the Genome Aggregation Database (gnomAD) versions 2.1.1 and 3.1.2. (https://gnomad.broadinstitute.org/). The gnomAD database is an important tool for genetic variation and cancer research. It contains global exome and genome sequencing data, and multiple large-scale sequencing programs give enormous variability in genomes among populations (Gudmundsson et al., 2022; Koch, 2020).

The entire reported mutations of selected genes have been retrieved from the Mutagene (developed at Panchenko Research Group, NCBI) (https://www.ncbi.nlm.nih.gov/research/mutagene/gene) and “catalogue of somatic mutations in cancer” (COSMIC v96, released May 31, 2022) (https://cancer.sanger.ac.uk/cosmic) in order to analyze the mutations commonly found in LUAD patients. As previously stated by (Goncearenco et al., 2017), Mutagene comprises 9,450 cancer samples from 37 projects containing 1,139,534 non-recurring mutations. Single base substitutions in protein-coding genomic loci from whole genome and whole exome-sequenced samples from ICGC projects, TCGA, and the Paediatric Cancer Genome Project are considered more frequently sequenced genes or mutations identified by genotyping. Cancer Drivers, Potential Drivers, and Passengers are predicted by Mutagene based on B_score thresholds defined using a comprehensive benchmark dataset comprised of multiple experimental studies. MutaGene adjusts the frequency of recurrent mutations in cancer patients according to their mutability concerning the predicted background mutagenesis. As a consequence, the computations will be affected by both the mutational model at baseline and the cohort. Catalogue of Somatic Mutations in Cancer (COSMIC; freely accessible to researchers worldwide) is a comprehensive database that covers cancer somatic mutations. It includes clinical and pathological data on mutation site, frequency, and function. This database may discover common mutations and other genetic variations that may contribute to cancer formation or progression and serve as therapeutic targets (Tate et al., 2019).

NOS3, CTH, EGFR, and PD-L1 interactions were predicted using the GeneMANIA prediction tool (https://genemania.org/). This online database provides gene interactions through the use of protein-protein interaction networks, co-expression patterns, and genetic connections. These four genes’ projected connections may reveal their involvement in cellular processes and disease pathways and lead to novel therapeutic targets or diagnostic biomarkers. The GeneMANIA prediction server is a major step in understanding gene interactions and their effects on human disease and health (Franz et al., 2018; Warde-Farley et al., 2010).

Patient characteristics

Fifteen of the LUAD patients are female (48.39%) and 16 are male (51.61%), with the following ages: Three (9.68%) were under 50, seven (22.58%) were between 51 and 60, eighteen (58.06%) were between 61 and 70, and three (9.68%) were older than 71.

Patients with LUAD were shown to have a very differentiated distribution, with several different subgroups. A greater number of cases were categorized as either poorly differentiated or moderately poorly differentiated (38.01%). These results suggest that aggressive and poorly differentiated tumors are significantly overrepresented among the patients. The frequency of moderately differentiated adenocarcinomas was lower than that of poorly differentiated subtypes, accounting for 14.28% of all cases. In addition, the study population had a lower incidence of primary adenocarcinoma and papillary adenocarcinoma (4.76%).

The extent of tumor invasion, lymph node involvement in the regional area, and the presence of distant metastases were all revealed by TNM staging. For 44.44% of cases, the TNM staging was determined to be pT3 N2 Mx. According to the results of this study, lymph nodes are affected by a significant proportion of locally advanced cancers. pT2 N1 Mx was the second most common staging pattern, occurring in 22.22% of patients and indicating tumors that were localised to the primary site but had spread to nearby lymph nodes. In addition, 11.11% of patients classified to each of the substages pT2 N2 Mx, pT2 Nx Mx, and T2 NO Mx. The differences between the phases are indicative of factors such tumor size, lymph node involvement, and the presence or absence of distant metastases, as shown in Table 1.

Mutation analysis

At various locations on chromosome 7, a total of 189 mutations in the NOS3 gene were identified in LUAD patients. Among them, there were 174 substitution mutations, 155 heterozygous and 19 homozygous mutations, and eight deletions, five insertions and two duplications were not recorded, these mutations were changed 24 different types of amino acids. Furthermore, except for four variant mutations, the mutations were never recorded in databases. Six homozygous substitutions in the NOS3 gene account for the greatest percentage of variants (100%); amino acid changes have occurred place in five of them, but only one of them has been reported in external databases.

On chromosome 1, the CTH gene was examined and a total of 31 different mutations were identified in various locations. There were 27 heterozygous mutations, three homozygous mutations, and one deletion among (28658delA). Furthermore, except for one variant mutation (28400G < GT), the mutations have never been documented in the databases. Furthermore, 6 of the newly found mutant variants in the CTH gene are associated with a change in amino acid level. Twenty-two heterozygous and one homozygous substitutions were identified in the CTH gene, but no amino acid changes were detected in any of them, nor were any of them reported in external databases, as shown in Fig. 1 and Table S1, and as summarised in Table 2 and Fig. 2.

EGFR mutation and PD-L1 expression

There were 14 (12.73%) samples with PD-L1 positive that expressed 50% or more PD-L1 protein among the 110 LUAD patients. While eight (7.27%) of the samples had EGFR mutations, including two tumors with exon 19 deletions, two tumors with exon 19 point mutations, four tumors with exon 21 point mutation (c.2573T > G p. Leu858Arg), and two tumors with exon 21 point mutation (p.L858R), as shown in Table S5.

Database mutations and genes interactions

The total reported mutations of all cancer types acquired by gnomAD revealed that the NOS3 gene has 4012 mutations, while the CTH gene has 1214 mutations, as shown in Table S2.

The retrieved gene mutations in the Mutagene (including mutations in the selected cancer cohort, all genome wide-studies in ICGC, and all samples in COSMIC) and COSMIC databases in LUAD patients were further analyzed for the selected genes, as shown in Tables S3 and S4. The NOS3 gene had the greatest number of mutations (295 and 93 in both databases, respectively), while the CTH gene had 61 and 36 mutations. The most common type of mutations found in the NOS3 gene in LUAD patients in the Mutagene database is missense followed by silent mutations, and their status is mostly passenger mutations, followed by potential driver and driver mutations, whereas the mutations in the CTH gene are mostly missense followed by silent and nonsense mutations, and their status is mostly mutations, passenger mutations, and one potential driver. Substitution-missense mutations are the most common type of mutation found in LUAD patients in the NOS3 gene in the COSMIC database, they are followed by substitution-coding silent, unknown, and substitution-nonsense mutations, while substitution-missense, unknown, substitution-coding silent and substitution-nonsense mutations are the most common types of mutations in the CTH gene. The data retrieved from the gnomAD, Mutagene and COSMIC databases are summarized in Fig. 3.

According to the GeneMANIA prediction server, there are 10 genes (AKT1, CAV1, CDO1, EPS8, HSP90AB1, LYPLA1, NOSIP, NOSTRIN, WASL and ZDHHC21) associated with the NOS3 gene, eight genes (CAV1, CBS, CDO1, DARS1, EGFR, MTHFD1, PTPN1 and SQOR) associated with the CTH gene, 15 genes (AKT1, ANXA2, CAV1, CBS, CDO1, CSK, DARS1, EGF, EPS8, HSP90AB1, NOSIP, PIK3CA, PTPN1, WASL and ZDHHC21) associated with the EGFR gene, and five genes (CAV1, CSK, NOSTRIN, PDCD1 and PTPN1) associated with the PD-L1 gene via co-expression, co-localization, genetic interaction, pathway, physical interaction, predicted, or shared protein domain (Fig. 4 and Table S6). Only the CAV1 gene is shared between the four gene pathways, and eight genes (AKT1, CAV1, CDO1, EPS8, HSP90AB1, NOSIP, WASL, and ZDHHC21) are shared between the NOS3 and EGFR genes, indicating that these two genes are closely related.