2. The Genomic Profile of Lung Cancer in Never Smokers
Genome-wide studies (GWS) have clearly indicated that the underlying tumour biology of NS differs strikingly from that of smokers (CS). Distinct pathways are altered between NS and CS
[6][9][10][6,9,14], with NS patients having better outcomes
[5][6][5,6].
Passive smokers show a similar mutation profile to smokers, albeit with a lower mutational rate overall. Therefore, analysis of the genomic differences between NS and CS will aid in uncovering the cellular and molecular pathways of malignant transformation. Indeed, although NS patients have a lower number of mutations compared to CS, they seem to be conducive for the malignant transformation, whereas, in CS, the numerous mutations seem to be mostly passenger mutations
[13][21]. Even so, there are specific changes that occur in the TME and distinct driver genes, as well as genetic pathway alterations, in NS. In spite of that, NS patients are usually younger, have a better prognosis, and respond better to treatment than smokers, due to the occurrence of certain molecular subtypes
[6] including EGFR-TK domain mutations. Verily, EGFR-TK domain mutations were found to be the first statistically significant molecular changes to occur particularly in NS. Furthermore, they are more frequent in NS than in long-term smokers (51% versus 10%) and in LUAD, contrary to cancers of other histologies
[14][22].
The transformation from LUAD to small-cell lung cancer (SCLC) as an outcome of EGFR-TKI resistance has recently been of much interest. This is not as common in NSCLC patients without
EGFR mutations. The inactivation of
RB1 increases the expression of neuroendocrine markers and decreases the expression of
EGFR, which is usually detected in the transformed tissue. Furthermore, in 21 patients, this transformation following EGFR-TKI resistance required an inactivation of both
RB1 and
TP53 just before branching. Inactivating gene alterations of both
RB1 and
TP53 could possibly serve as predictive markers for the transformation, since it is common and seems to be a necessary event for both genes to be inactivated in all small-cell lung cancer SCC tissues
[15].
Zhang and colleagues conducted an in-depth genomic and mutational signature analysis (WGS) with the hope of providing a guide to the development of precise clinical treatments to benefit LCINS. They recruited 232 NSLSC NS patients and established three genetic subtypes unique to NS. TMB was almost sevenfold lower in NS than in smokers
[1][13][1,21] and significantly associated with tumour stage, histology, and age, but not tumour purity. A higher frequency of
EGFR mutations was found in females than in males. Additionally, EGFR signalling was increased due to
EGFR or
KRAS mutations, and mutations in
RBM10 and
TP53 were found to frequently co-occur with
EGFR. Moreover, co-occurring patterns were also found between
RBM10 and
PIK3CA, as well as between
TP53 and
ERBB2 and marginally significant enrichment of
SETD2 mutations, in samples with oncogene fusions, particularly in TP53-proficient tumours. A strong mutually exclusive distribution was observed across the genes in the RTK–Ras pathway which were altered in a total of 54.3% tumours.
EGFR was the most frequently altered, followed by
KRAS,
ALK,
MET,
ERBB2,
ROS1, and
RET [1].
Moreover, unsupervised clustering of arm-level somatic copy number alterations in the Sherlock-Lung study identified three distinct subtypes. Briefly, subtype 1 called ‘piano’ tumours had the least number of mutations growing slowly over the years, with the most frequently mutated gene being
KRAS. Oncogenic
KRAS mutations were involved in LUAD since they induced proliferation of bronchioalveolar stem cells. Subtype 2 ‘mezzo-forte’ tumours demonstrated mutations in
EGFR, a common mutation found in lung cancers exhibiting faster tumour growth. ‘Mezzo-forte’ was enriched with chromosome arm-level amplifications of 1q, 5p, 7p/q and 8q. Lastly, they identified the ‘forte’ subtype which grew the quickest, was most similar to lung cancers among smokers, and was dominated by whole-genome doubling (WGD). Forte also had a low TMB and a larger proportion of subclonal mutations indicative of extensive intra-tumour heterogeneity. Piano also had a small number of known driver gene mutations suggesting stem-like features, in addition to a new driver gene
UBA1, which acts as one of the main orchestrators of cellular DNA damage response. It is worth mentioning that they did not find major differences in the mutational signature or types between passive and non-passive smokers. However, they observed a few tumours with diesel exhaust signatures. Since smoking-related mutations from passive smokers were below the detection threshold of 15%, it is possible that second-hand tobacco smoke may also act through alternative tumourigenic processes and selective constraints. They also found alterations in
KRAS,
UBA1,
RET, and
ARID1A to be mutually exclusive in piano.
KRAS and
UBA1 are important hematopoietic and pluripotent stem-cell regulators, and
RET is involved in murine hematopoietic stem-cell regulation.
ARID1A could promote exit from a quiescent cell state, causing high inter-tumoural heterogeneity; thus, it could drive some of the tumours with no detected known cancer driver gene mutations or fusions. On the contrary, the forte and mezzo-forte subtypes were generally clonal with driver gene mutations and WGD or gross somatic copy number alterations (SCNAs) facilitating identification and possibly successful treatment
[1].
A recent comprehensive genomic and transcriptomic analysis (WES) of 160 tumour and normal LUAD samples from NS highlighted a number of prevalent clinically relevant alterations, allowing a better understanding of the risk factors involved. They categorised NS LUAD into relatively immune cold and hot subtypes. The immune cold subtype tumours, in comparison with immunologically hot tumours, appeared to lack expression of immune markers, such as PD-L1, but were depleted with immune cells. This suggests immune evasion mechanisms that are not very well characterised. Indeed, compared to smokers, NS LUAD tumours had relatively lower TMBs and a higher frequency of mutations in genes such as
CTNNB1.
CTNNB1 participates in wingless WNT signalling, and WNT signalling activation facilitates immune evasion and contributes to immunotherapy resistance, possibly providing an explanation for the relatively lower response rates associated with immunotherapy targeting PD-L1 in NS LUAD patients compared with smokers. Moreover, they identified a small subset of the samples from NS to have pathogenic and likely pathogenic germline alterations in DNA repair genes. DNA repair genes exclusively mutated in NS included
BRCA1,
BRCA2,
FANCG,
FANCM,
MSH6, and
POLD1. Furthermore, some tumour cells exhibited mutational signatures characteristic of ETS. In fact, they found the genomic features of smokers and NS to be comparable with the activation of RTK/RAS/RAF signalling, a hallmark feature of LUAD. Nevertheless, NS LUAD demonstrated a higher prevalence of targetable driver alterations in RTK/RAS/RAF signalling pathway than smokers
[3].
Likewise, the genetic pathway of driver mutations in NS lung cancer are different to those in smokers. An evaluation of therapeutically targetable mutations such as
EGFR,
ALK, and
KRAS, as well as chromosomal rearrangement and fusion of
EML4, identified increasing odds of presenting these mutations in adenocarcinoma and NS compared to NSCLC and smokers, respectively. Additionally, the mutations of
EGFR were more prevalent in Asian women in comparison to women of Caucasian/mixed ethnicity
[16][17][16,23], further highlighting ethnicity as a risk factor. In support of this was a recent Australian study evaluating the risk factors in LCINS. They explored demographic, lifestyle, and health-related factors in NS and found growing evidence that ethnicity could be considered when assessing potential risk factors in LCINS
[18][24]. Furthermore, methylation profiles of LCINS are different from smokers, with 16p chromosomal aberration gain being more frequent in NS
[19][25].
A protogenomic study in Taiwan focused on exploring differences in NS LUAD, with a cohort of 83% non-smokers. They included the matched early-stage tumours and their normal adjacent tissues. Their study revealed different mutational profiles of previously explored driver genes in NS with significantly different mutational frequencies. Moreover, in their NS cohort, genes including
EGFR,
RBM10, and
RNF213 were amid the top-ranking mutations. Other somatic mutations prevalent in the NS cohort included
ATP2B3 and
TET2.
RBM10 is an LUAD tumour suppressor, and its loss, due to mutations, could impact its interactions and lead to impaired RNA splicing. Moreover, they observed co-regulated phosphorylation of MAPK pathway proteins, which distinguished patients with high activation to associate with
EGFR and
KRAS mutations, while low activation with
TP53 mutations, especially in later stages. A characteristic observation in this study is that there were no significant differences in C>A transversions between smokers and NS, suggesting that other factors aside from smoking contribute to the genomic landscape of NS. C>A transversions were significant in the smoker cohort and C>T transitions were significant in the NS cohort. Around 85% of the patients had
EGFR mutations followed by
TP53 (33%) and
RBM10 (20%)
[16].
A study by Paik et al. attempted to explain the survival advantage that NS with LUAD with stage IIIB/IV have over former/current smokers, living 50% longer. In line with previous studies, they also found that NS had a significantly higher proportion of
EGFR mutations and
ALK rearrangements, whereas smokers had a higher proportion of
KRAS mutations. Notably, they did not observe significant overall survival (OS) differences in both groups with identical genotypes. The authors concluded that both smoking groups are not homogeneous, with each group’s individuals having a set of disparate mutations that additively generate an overall prognosis. They went on to note that, although NS have a higher frequency of
EGFR mutations, EGFR-TKIs should not be the immediate route of treatment. This is because, following treatment with EGFR-TKIs, smokers with
EGFR mutations exhibit similar survival outcomes to NS. Instead, regardless of smoking history, LUAD patients should initially undergo testing for
EGFR mutations and rearrangements in
ALK in an effort to match patients with appropriate targeted therapy, while those who do not harbour mutations in
EGFR/KRAS or rearrangements in
ALK should be stratified by smoking history
[20][17].
Hormonal factors are another risk factor in lung carcinogenesis. The link between driver oncogenes and hormonal receptors in LUAD was supported by a study conducted by Mazieres and colleagues examining 140 females with LUAD, which included 63 never smokers. They found female NS to be characterised by older age and a higher frequency of lepidic features compared to smokers. Additionally, they observed differential genetic alterations to be prevalent in NS, including a higher mutation frequency of
EGFR but a lower frequency of
KRAS and a higher percentage of oestrogen receptor alpha (ERα). Furthermore, ERα expression correlates with the presence of
EGFR mutations, associated with both mutational and hormonal biomarkers
[21][18].
3. Immunological Changes: Tumour Microenvironment in Never Smokers
Although not widely studied, the TME in NS is distinctly different from that of smokers. Moreover, the immune system plays a vital role in cancer development progression. It is thought that specific immunological features contribute to lung cancer development irrespectively of tobacco smoking. When NS are exposed to ETS, immune cells are initially recruited to minimise the damage by carcinogenic substances. Signalling cascades including MAPK, which bridges the switch from extracellular signals to intracellular responses, are involved in not only cell proliferation but also immune escape, contributing to cancer progression
[22][27]. However, when a tumour arises, it might also cause harmful pro-inflammatory and immune reactions, and partake in the harmful TME, contributing to tumour growth invasion and metastatic spread. Therefore, the immune system could protect against cancer progression or enhance tumour growth by influencing the TME and weakening the surrounding normal cells
[6]. Moreover, cancer and autoimmune disorders are frequently encountered in elderly patients possibly due to the ageing process, which could also affect changes in innate and adaptive immune function, i.e., immunosenescence. Immunosenescence also causes dysfunctional maturation and function of natural killer (NK) cells and insufficient neutrophil migration, probably due to increased constitutive PI3K activation. In turn, this decline in NK comprises a slower response to inflammatory conditions and affects the adaptive immune system via immunosenescence, altering the function of B and T lymphocytes
[23][28].
An integrative analysis study conducted by Li and colleagues included 11 lung cancer gene-expression datasets that provided data from 1111 LUAD patients and an adjacent 200 samples of normal tissue. They found distinct pathways altered between smokers and NS, with NS having a better outcome. In addition, the transforming growth factor beta (TGFβ) pathway has been identified to contribute to immune dysfunction and to be associated with immune checkpoint inhibitor (ICI) resistance. They identified the
TGFBR2 mutation to predict immunotherapeutic resistance, associated with increased JAK/STAT signalling and immune checkpoints including
CD274,
LAG3,
TIGIT,
PDCD1, and
PDCD1LG2 [10][14]. They also characterised the compositional patterns of 21 types of LUAD immune cells. Their study revealed complex and multi-layered associations between the composition of immune cell subtypes and clinical outcome among the smoking groups. In particular they found mast cells and CD4+ memory T cells with completely opposite associations with outcomes in resting and activated status. The number of resting mast cells, not having undergone degranulation, was found to be reduced in tumour samples, in comparison to the adjacent normal tissue and a predictor of favourable outcome. On the other hand, macrophages, activated mast cells following degranulation, and activated CD4+ memory T cells were enriched in the tumour samples, predicting a poor prognosis. NS had more resting mast and CD4+ memory T cells associated with a better outcome, whereas, in smokers, there were more activated mast cells and CD4+ cells, which correlated with a generally worse prognosis
[10][14].
In addition, oxidative stress can be a causative factor for lung carcinogenesis in NS due to the constant exposure to ambient air pollution. Ito et al. investigated the impact of oxidative stress in NSCLC patients who underwent surgical resection including 34 CS and 27 NS by examining oxidative damage on DNA. Immunohistochemistry was used to assess the oxidative damage by examining the accumulation of thymidine glycol (TG). TG is a specific marker for oxidative DNA damage since thymidine is not incorporated into RNA. The mean TG positive rate, indicative of oxidative DNA damage, was significantly higher in smokers compared to NS, and significantly higher in NS than surgical patients with benign lung disease. They also investigated the serum oxidative stress and antioxidant capacity (AOC). The mean level of AOC was found to be significantly lower in NS compared to smokers. Hence, the comparatively low antioxidant potential for NS could be a contributing factor to excessive oxidative DNA damage in lung tissue
[23][28].
Another study aimed to quantify the differences between smokers and NS LUAD patients by analysing immune infiltration and stemness amongst other variables. Owing to its crucial role in lung cancer outcomes, therapies against
NTS reduce tumour growth and metastasis; an important finding is that
NNAT and
NTS were upregulated in NS. Additionally, their overexpression in the NS cohort could be involved in LUAD development. Moreover, they identified mast cells, M2 macrophages, memory resting CD4 + T cells, and dendritic cells to be upregulated in NS, and both
TFF2 and
REG4 were downregulated in NS LUAD.
TFF2 is required by lung macrophages to promote epithelial proliferation; thus, its downregulation in NS could explain the weaker repair capacity and tumour development. On the other hand,
REG4 plays a role in
KRAS driven lung cancer pathogenesis. They also found a significant difference in the expression for programmed death-ligand 1 (PD-L1) between smokers and NS
[24][29].
Lastly, the mRNA expression-based stemness indices (mRNAsi) showed significant differences among the groups, with NS or reformed smokers exhibiting lower stemness than current and recently reformed smokers. All results provide an understanding of the causes of oncogenesis in NS LUAD and possible therapeutic approaches
[24][29]. This was also supported by an integrated multi-omics study exploring the possible underlying molecular mechanisms among NS, former smokers, and current smokers. In addition to tumour cell stemness, they found different immune content, genome stability, and sensitivity to chemotherapy drugs. Their results also indicated that NS had better OS and disease-specific survival (DSS) than smokers but were substantially more sensitive to multiple chemotherapeutic drugs than smokers. Additionally, leukocyte infiltration, intertumoural heterogeneity, and neoantigen levels were significantly higher in smokers compared to NS
[25][30].
4. Abnormalities in Growth-Stimulatory Signalling Pathways
Several oncogenic proteins are either members of cytoplasmic signalling cascades or interact with them, leading to transformations as a result of this deregulation. Recent studies have focused on investigating the carcinogenesis of LCINS. Inarguably, the functions of several pathways with major components are altered in LCINS, making it a separate entity. Some recent studies identified a list of DEGs in female NS LUAD patients enriched in p53, TGF-beta, and cell-cycle signalling pathways
[26][31] and nuclear division pathways
[27][32]. In NSCLC, several signalling pathways have been heavily implicated in both tumourigenesis and progression of the disease. Various specific inhibitors of PI3K, Akt, and mTOR are currently under development for NSCLC, at various stages of pre-clinical investigation and in early-phase clinical trials. Unfortunately, early evidence has not yielded promising results, but this could be due to the fact that these studies were performed on predominantly molecularly unselected populations. Selecting patients following patient enrichment strategies with a better understanding of the underlying molecular biology, including epigenetic alterations and guided combination approaches, will increase the likelihood of success
[28][33].
4.1. RAS/MAPK Signalling Pathway
The mitogen-activated protein kinase (MAPK) signalling pathway is a vital aspect of NSCLC signalling and a predominant aspect in a wide number of cellular functions including cell survival, differentiation, proliferation, metastasis, and apoptosis. Targeting the Ras/Raf/MEK/ERK pathway can prove to be a promising therapeutic regimen for NSCLC patients
[29][34]. Other kinases involved in the Ras/Raf/MEK/ERK cascade include receptor tyrosine kinases (RTKs). Ras/Raf/MEK/ERK are involved in lung cell death, development, and pathogenesis
[29][34]. This family comprises the epidermal growth factor receptor (EGFR) and the fibroblast growth factor receptor (FGFR), which crosstalk with major tumour-promoting signalling pathways. Indeed, when a ligand binds to EGF it stimulates EGFR followed by its activation by an intracellular tyrosine kinase domain, thereafter resulting in its autophosphorylation and its overexpression, causing increased intracellular EGFR pathway activity. This atypical activity might be the reason that almost 40–89% of NSCLC patients have
EGFR deregulation
[29][34].
MAPK constitutes an evolutionarily preserved family of protein kinases acting as cytoplasmic mediators of signal transduction pathways critical for cellular proliferation and survival. In an LUAD study exploring differences between female and male never smokers, a number of genes relating to the MAPK/PI3K signalling pathway have shown a drastic difference in the prognosis of female and male patients. These include
ERBB4 and
NTF4, which showed different prognostic effects on LUAD progression in NS males and females. Therefore, sex should be taken into account when designing therapeutics for LUAD never smokers
[30][35].
In a recent study with 83% never smokers, downstream activation of the MAPK signalling correlated with EGFR-pY1197, MAP2K2-pT394, and its substrate MAPK3-pT198/pT202, and in turn with pMAPK1 and other downstream phosphoprotein (RSK2, cPLA2, and STMN1). They observed that the MAPK signalling pathway is commonly activated among both EGFR-WT (wild-type) and mutated patients with different degrees of activation. Indeed, patients without
EGFR-activating mutations corresponded with low MAPK signalling. Additionally, three of four EGFR-WT samples with higher
MAPK activity harboured
KRAS mutations, while samples that harboured both
EGFR and
TP53 mutations had low MAPK signalling, especially at the later stages. With specific regard to never smokers, variation of MAPK pathway activity was observed with patients having different
EGFR activating mutations and was also influenced by
TP53 mutations. Lastly, late-stage tumours with lymph-node metastasis seemed to have lower
MAPK activity
[16].
Extracellular regulated kinases (ERK1/2), c-Jun NH2-terminal kinases (JNK), and four P38 enzymes, p38α (MAPK14), p38β (MAPK11), p38γ (MAPK12), and p38δ (MAPK13), are well-characterised cytoplasmic mediators of the MAPK pathway
[31][36]. The expression level of
MAPK11 was found to be significantly higher in an Asian NS cohort
[32][37]. A large number of small-molecule p38 inhibitors have been developed and can theoretically be used to treat tumours that depend on
MAPK14 for progression. This is because
MAPK14 plays a role in cancer cell migration, tumour invasion, and metastasis, where the expression of matrix metalloproteinases (MMPs) and angiogenic factors is induced by p38 MAPK signalling
[31][36]. Activated
ERK and
JNK can result in increased proliferation and survival, whereas the P38 MAPK pathway is involved in suppressing tumorigenesis. In NSCLC,
ERK,
JNK, and
P38 are usually activated, but their activation degree is variable
[19][25].
In an attempt to explore whether the MAPK activation state differs according to smoking status, Mountzios et al. evaluated the expression of activated extracellular signal-regulated kinases including c-Jun and p38 enzymes using immunohistochemistry in LUAD. Following adjustment of any potentially confounding covariates, they found that 37 of 44 NS had higher levels of expression of pP38 compared to 45 of 104 smokers. They observed the P38 pathway to be ten times more activated in never smokers than smokers. Their results provide evidence that life-long non-smoking is associated with an activated P38 pathway and implies that higher
P38 levels are related to distinct molecular changes in never smokers. Since
P38 acts as a tumour suppressor among MAPKs, with higher activation levels observed in never smoker LUAD, this indicates that its action is different in the context of adenocarcinoma cells in never smokers, given their unique molecular and biological characteristics. They indeed confirmed this hypothesis by studying the effects of
P38 pharmacological inhibition on cell growth in the
EGFR mutant (delE746_A750) adenocarcinoma cell line (HCC827), which is derived from never smokers that do not harbour the
KRAS mutation. Indeed,
P38 activity contributed to HCC827 cell growth rather than inhibiting it. However,
P38 induced apoptosis or cell senescence in several models characterised by
RAS-induced proliferation, and contributed to cell growth in LUAD in never smokers. Therefore, it is speculated that the high levels of activated
P38 in never smokers could be explained by the lack of
KRAS mutations. Specific aberrations in MAPK or interacting pathways responsible for P38 pathway activation in never smoker LUAD are yet to be determined
[19][25].
The expression of ERCC1 protein is a main predictor of the benefit of cisplatin-based chemotherapy in lung cancer, and its gene contains AP-1 sites bound by the transcription factors JUN and ATF2. Planchard et al. investigated whether p38 MAPK activity contributed to excision repair cross-complementation group 1 (ERCC1) mRNA expression and viability of cisplatin in lung cancer cell lines from light or never smokers. They found ERCC1 protein levels to be predicted by activated p38 MAPK in LUAD tissue; furthermore, cells from LUAD light or never smokers generally rely on p38 MAPK signalling for survival, with higher expression of ERCC1 in never smokers. Downregulation of
ERCC1 expression reduced cell viability and could account for the effect of p38 MAPK inhibition on cell viability. Inhibition of p38 MAPK with a specific inhibitor SB202190 that targets both
MAPK11 and
MAPK14 resulted in decreased cell viability in all never smoker cancer cell lines to different degrees.
MAPK11 downregulation reduced cell viability in all cell lines; however,
MAPK14 downregulation also reduced cell viability in a cell line derived from never smokers. This enunciates that
MAPK11 signalling is the main contributor to cancer cell survival in never smokers.
MAPK11 and
MAPK14 have opposite effects on cell differentiation and survival. In
MAPK14 knockout mice, the proliferation of immature lung stem cells also facilitates KRAS G12V tumorigenesis. Furthermore, pre-treatment of two cell lines from never smokers (H1793 and H1651) with SB202190 sensitised cells to cisplatin, which could provide insight into why not all lung cancer patients benefit from cisplatin-based chemotherapy. Additionally, sensitivity to cisplatin was higher following
MAPK11 downregulation of H1651, a cell line from never smokers, or
MAPK14 downregulation of H1650, a cell line from light smokers. This could clarify the cytotoxic effects observed in certain current treatments. Moreover, the crosstalk between p38 and JNK pathways should be investigated as they share many upstream regulators
[31][36].
4.2. Mutations in Other EGFR Signalling Pathway Genes
Although mutations in never smokers are lower than in smokers, they are perceived to cause malignant transformation, whereas, in smokers, they are mostly thought to be passenger mutations
[13][21].
EGFR mutations were observed in 40–60% of NSCLC NS patients, of which 17% accounted for LUAD
[33][38] and were more common in never smokers or light smokers
[34][39]. Yet, this does not mean that smokers do not have
EGFR mutations, but that only common
EGFR mutations were more frequent in never smokers, whereas smokers had more uncommon single and complex rare mutations
[35][40].
A significant revolution in NSCLC therapeutics is the identification of activating oncogenic aberrations such as
EGFR mutations.
EGFR tyrosine kinase inhibitors (EGFR-TKIs) are linked with superior efficacy in NSCLC patients with activating
EGFR mutations
[36][41] and never smokers
[37][42]. NSCLC patients with
EGFR activating mutations have an excellent response to EGFR-TKIs; however, approximately 20–30% of NSCLC patients with
EGFR mutations show de novo resistance to EGFR-TKIs. A possible explanation could be the presence of genetic alterations affecting genes downstream of
EGFR. The preliminary results of a recent clinical study NCT01405079 in NSCLC stage II–IIIa patients with
EGFR mutation treated with gefitinib versus vinorelbine/platinum indicated that patients on adjuvant gefitinib have a better disease-free survival (DFS) than those on chemotherapy (38.7 months vs. 18 months)
[36][41].
Nevertheless, identifying the
EGFR mutation is not sufficient to determine the patient’s response to TKIs due to the presence of secondary
EGFR mutation or any downstream or altered signal activation
[36][41]. Hence, comprehensive genotyping, specifically of interactions with
EGFR mutation, would be able to provide a better picture of any signal activations that may be present in the patient. In fact, several studies in NSCLC have reported that other signalling pathways mediate potential resistance to EGFR-TKIs; for example, activation of JAK2-related signalling upregulated
ROR1 via NKX2-1, resulting in the overexpression of
NOTCH1, leading to epithelial-to-mesenchymal transition (EMT). Moreover, EGFR-TKI resistance in patients with T790M mutation resulted from increased DNA repair due to high levels of
BRCA-1. Additionally, NFKB signalling presented TKI resistance to
EGFR-mutant NSCLC cells with smoking and never smoking history; however, inactivating
NFKB using TLR-9 agonist along with erlotinib did not increase PFS in comparison to using erlotinib alone
[36][41]. Indeed EGFR-TKIs have shown better responses in LUAD patients with no smoking history, in the female demographic, of Asian ethnicity, or with
EGFR mutation
[38][43].
4.2.1. Notch Signalling Pathway Genes
Notch signalling pathway is essential for embryonic lung development and tissue homeostasis. Activating mutations in NSCLC correlate with a worse prognosis
[38][43].
Notch1 contributes to EGFR-TKI acquired resistance in NSCLC. Moreover,
Notch3 expression is positively correlated with
EGFR expression, and
Notch3 overexpression is associated with poor prognosis in NSCLC
[39][44].
One study aimed to assess the impact of Notch signalling on survival by examining the expression of
Notch1, 2, 3, 4 in comparison with the adjacent normal tissues in resected NSCLC using RT-PCR. They found a higher expression of
Notch2 in females, as well as never smoker LUAD patients, than tumours of other histologies. They also found the expression of
Notch2 to positively correlate with more advanced lung cancer stages and a higher rate of recurrence or metastasis; ergo, it could be involved in EMT progression
[40][45]. However, this finding was not statistically significant, and its expression had no impact on DFS and OS in LUAD patients. Therefore,
Notch2 signalling plays a crucial role in mutation of LUAD, specifically in never smoker East Asian females. They also found LUAD patients with high expressions of both
Notch1 and
Notch3 to have poor DFS. In addition, Cox regression analysis showed that
Notch3 expression remained the leading predictive factor of DFS
[39][44].
A comprehensive genomic analysis of classic SCC indicated that 25% of human SCCs are affected by genomic alterations of the NOTCH signalling family, and almost 77% of SCLCs with high expression of neuroendocrine markers show a gene expression pattern suggestive of low Notch signalling activity, including a high level of
ASCL1. Thus, inactivation of both
p53 and
RB1 is critical for tumorigenesis of SCLC, and the inactivation of Notch signalling causes neuroendocrine differentiation.
EGFR mutation-positive LUAD cells that harbour an inactive form of both
RB1 and
TP53 are more likely to transform
[15]. Moreover, the activation status of
Notch1 had a poor prognostic impact on NSCLC, and it was used in the subgroup of p53-negative NSCLC patients to predict overall survival
[39][44].
To identify key genetic changes defining histological transformation from LUAD to SCC, Kabo et al. recruited female never smokers with
EGFR-mutant LUAD patients. Patients harboured a deletion mutation in exon 19 of
EGFR. They then compared the gene alteration profile in the original LUAD prior to EGFR-TKI treatment and in the transformed SCC. They ensured that the samples were purely LUAD by surgical resection, and FFPE samples were micro-dissected. They also identified the inactivation of both
RB1 and
TP53 in SCC and LUAD. Additionally, they selected five genes as candidate key genes for the transformation from LUAD to SCC. All three cases had completely matching changes in the nucleotides with additive alterations common in all cases of the five genes identified:
MTOR,
JAK1,
NOTCH2, and
CSF1R; additionally
MAP2K2 was a lost alteration in all cases. Duplication alterations of
MTOR,
JAK1,
NOTCH2, and
MAP2K2 were located in the 3′ untranslated region (UTR), and the
CSF1R alteration was a single-nucleotide polymorphism in the intron. There was no common aberration in completely matching nucleotides for
PI3K and
AKT. They also focused on
NOTCH mutations as possible alterations that represent the transformation from LUAD with
EGFR mutations to SCC, as the expression of
NOTCH seems to be the key mechanism for the transformation.
NOTCH mutation and reduced expression at the protein level were detected only following the transformation
[15].
The expression levels of
ACSL1 and
MTOR were higher in LUAD, while those of
NOTCH1/2,
CSF1R, and
JAK1 were lower than in transformed SCC. Additionally,
EGFR expression disappeared in the transformed SCC compared to that in LUAD, and
Rb expression was not observed in either LUAD or the transformed SCC. They also focused on the expression of
ASCL1, a downstream gene of Notch signalling regulated through HES-1 and HEY-1, which act as ASCL1 transcriptional repressors.
ASCL1 expression was found to be higher in SCC transformed tissues.
TP53 alteration was detected in the paired samples of both LUAD and transformed SCC in all three cases. Aberrations in
TP53 were detected in all three cases as non-synonymous coding or frame-shift effects, which resulted in amino-acid changes. The authors hypothesised that the inactivation of
p53 and
RB1 emerged during carcinogenesis, and tissues that acquired
Notch inactivation subsequently transformed, since neuroendocrine differentiation did not occur in tissues lacking
NOTCH inactivation.
NOTCH2 expression was negative in the transformed SCC, but positive in LUAD.
ASCL1 expression was positive in the transformed SCC but negative in LUAD. They concluded that
NOTCH mutations were detected as additional alterations in all three cases. It is suggested that
Notch inactivation is one of the key conditions causing SCC under
RB1 and
p53 inactivation, indicating that the
NOTCH and
ASCL1-dependent pathway represents a key process in the transformation using actual tumour tissues from patients with the transformed SCC after becoming EGFR-TKI-resistant. Therefore, the NOTCH/ASCL1 axis could be a potential therapeutic target in transformed SCC from LUAD with oncogenic driver mutation, and AKT inhibitors could delay transformed neuroendocrine lung carcinoma
[15].
4.2.2. Abnormalities in Tumour Suppressor Gene Pathways: P53 and KRAS
The incidence of
TP53 mutation is higher in smokers than in never smokers and among patients with SCC compared to LUAD with a different biological impact among the smoking groups. Never smoker lung cancer patients show a totally different and random grouping of
p53 mutations. Moreover, a lower frequency of
TP53 mutations was identified among patients with
EGFR mutations who were never smokers with LUAD. Additionally, in never smokers,
TP53 mutations were identified as a significant independent negative prognostic factor
[41][46]. In lung cancer,
TP53 mutations are the most prevalent and often co-exist with driver mutations, being higher in SCC than LUAD
[42][47].
A study analysing the association among mutation status, clinicopathologic characteristics, and outcome in never smokers with LUAD identified never smokers to have a higher incidence of targetable mutations with a significantly longer survival than patients without mutations. The authors identified
EGFR mutations amid the most common encountered mutations—55.6% with deletions in exon 19 and associated with longer OS. They also found the frequency of
ALK rearrangements (12.3%) to be associated with ipsilateral mediastinal or subcarinal lymph-node metastasis (N2) and a better outcome compared to the wild-type (WT). They also found 14.3% of the tumours to harbour
TP53 mutation. Lastly, they observed significant differences in survival for patients with
EGFR mutations compared to EGFR-WT (wild type) and
EGFR pan-negative tumours, as well as
ALK rearranged versus WT and
ALK pan-negative tumours
[43][48].
Another study aimed to investigate the impact of concomitant
TP53 mutations and their clinicopathological characteristics in
ALK-rearranged NSCLC patients, as well as the association of
TP53 with the effect of crizotinib in
ALK-rearranged patients. In their study, never smokers accounted for 76.6% of the patients and
TP53 mutations occurred in 23.4% of
ALK-rearranged NSCLC patients. They explored the correlation between
TP53 mutations and the outcome of
ALK-rearranged patients following crizotinib treatment. They found especially non-disruptive
TP53 mutations to negatively affect the response to crizotinib and correlate with shorter PFS in patients with
ALK-rearranged NSCLC patients. Non-disruptive
TP53 mutations, which cause partial loss of
p53 function having a retained functional property associated with gain of function (GOF) representing a heterogenous subgroup of
ALK rearranged NSCLC patients with inferior PFS
[42][47]. These results indicate the negative prognostic role of
TP53 mutations in ALK-rearranged NSCLC patients undergoing treatment with crizotinib.
In addition to the observed higher mutational burden and co-occurring mutations in smokers, they also have a more complex
KRAS mutation than that observed in never smokers
[43][48], suggesting a different mechanism of carcinogenesis in never smokers. A study analysed
TP53 and
KRAS mutations in lung cancer tumours of different smoking groups. Their results support the notion that tumorigenesis in lungs proceeds through different molecular mechanisms according to smoking status, and that the accumulation of N-Tyr in never smokers tumour cells is higher than in smokers, implying an aetiology involving severe inflammation. N-Tyr is a stable product of nitration of tyrosine residues and a biomarker of protein damage from peroxynitrite and other reactive nitrogen species (NOS), common in severe forms of inflammatory airway diseases such as chronic obstructive pulmonary disease (COPD) and asthma. However, they did not observe any correlation between N-Tyr and a particular
TP53 mutation type, which could indicate that mechanisms causing severe inflammation other than
TP53 mutation could contribute to carcinogenesis. Nevertheless, they also found that TP53 mutations were detected in 47.5% of never smokers with G:C-to-A:T transitions. It is also possible that the G:C-to-A:T transitions at non-CpG sites in never or former smokers might represent a DNA fingerprint for NNK in patients exposed to secondary smoke. Furthermore,
KRAS mutations were detected in 15.3% of the cases and were more frequent in LUAD than SCC and in former smokers than in other categories
[44][49].
The prognostic impact of
EGFR,
KRAS, or
TP53 mutations in LUAD indicated that these genes are not independently associated with prognosis in patients who underwent pulmonary resection, including never smokers. However, using univariate analysis, all except
KRAS have significant prognostic value, insinuating that the significance is possibly caused by confounding other prognostic factors including sex, smoking status, and tumour differentiation, which, following adjustment, lost their prognostic impact
[45][50].
4.2.3. PI3K–AKT–mTOR
The PI3K–AKT–mTOR pathway is involved in the regulation of several functions including adhesion, motility, invasion, and cell proliferation and differentiation. In NSCLC, abnormal activation of the PI3K–AKT–mTOR pathway seemed to generate resistance to EGFR-TKIs. Alterations can happen through the activation of tyrosine kinase receptors upstream of
PI3K and
PIK3CA amplifications, mutations in
KRAS,
PI3K,
AKT, and
TSC1/2, or loss of
PTEN.
PIK3CA and
AKT1 mutations and
PTEN loss, which are the prominent mutations leading to activation of the PI3K–AKT–mTOR pathway. mTOR inhibitors including everolimus, which targets mTORC1, and temsirolimus are approved for cancer treatment. Moreover, genetic mutations in the PIK3CA/AKT/mTOR pathway, one of the EGFR downstream pathways, might impact the response to EGFR-TKI in NSCLC with activating
EGFR mutations
[46][51].
Co-occurrence of
PI3K-related mutations with
EGFR-activating mutations leads to worse prognosis and shorter PFS with EGFR-TKIs. This is because such alterations may uncouple
EGFR from downstream signalling. Resistance to EGFR-TKIs would be evident by shorter survival outcomes and/or poor response rates. Patients with
PTEN mutations had a poor survival outcome, while those with
PIK3CA or
STK11 mutations revealed trends toward a poor survival outcome
[46][51]. Other proteins related to MAPK/PI3K signalling included
ERBB4 which seemed to show differential effects on LUAD progression in never smokers, with totally different prognostic effects in female never smokers compared to males. It had little effect on the prognosis in the whole LUAD population; however, its level of expression correlated with prognosis in never smokers.
ERBB4, which is a member of the
EGFR family, was reported to have abnormal activation via somatic mutations which could associate with tumour progression. Its higher expression in this female LUAD cohort seemed to correlate with poor prognosis in females, but a better prognosis in males
[30][35].
A recent study attempted to explore the genomic characteristics of the PI3K pathway activated in NSCLC patients following progression on EGFR-TKIs and the co-occurrence of common mutations through PI3K–AKT–mTOR. The study was further performed on six patients with a history of everolimus and EGFR-TKI treatment to estimate the anti-tumour activity. These stage IV NSCLC patients had specific mutations along the PI3K–AKT–mTOR pathway, and three of them were never smokers. Following progression on EGFR-TKIs, all patients acquired
PIK3CA mutations and
PTEN loss, and they achieved stable disease. Following several patient deaths, the authors inferred that EGFR-TKIs, along with everolimus, might not be enough to overcome EGFR-TKI resistance induced following abnormal PI3K pathway activation. Additionally, PI3K pathway alterations seem to serve as a common resistance mechanism, being present in 14.9% of EGFR-TKI resistance events. Nevertheless, the authors speculated that the failure of therapy was due to the specific targeting of
mTORC1 and not
mTORC2, since inhibition of
mTORC1 solely can activate an
mTORC1 negative feedback loop, resulting in
AKT activation via S6K-dependent upregulation of the IRS-1 and TGFR-1 pathways. Therefore, inhibiting
mTORC1 does not completely suppress the PI3K pathway. Moreover, there are other players involved since PI3K pathway activation interacts with other signalling pathways, including the MAPK pathway. Lastly, the safety and toxicity profile of PI3K pathway inhibitors remain unclear and pose issues. Additionally, combination therapy provided limited anti-tumour activity in patients with dysregulated PI3K–AKT–mTOR pathway
[47][52].
Kim and colleagues aimed to investigate the relevance of
EGFR-downstream gene mutations including
PIK3CA,
AKT1,
PTEN, and
STK11, and of treatment outcomes of EGFR-TKIs in never smokers with activating
EGFR mutations. They aimed to prove the concept that the mutation in EGFR downstream genes may be related to EGFR-TKI resistance. Following screening of those patients, the frequency of genetic mutations related to
EGFR downstream signalling included 3 (4.4%) patients with
PIK3CA mutation (exons 9 and 20), 11 (16.1%) with
PTEN mutation (exons 1–9), 4 (5.9%) with
AKT2 mutation, and 9 (13.2%) with
STK11 mutation (exons 1–9). It is of importance to note that the frequency of
PTEN, which was 16.1% in this NS group, was found to be higher than previously reported studies, with mutations occurring in the phosphatase domain indicating alteration of gene function. They observed the EGFR-TKI treatment outcome of 55 patients including gefitinib (61.8%), erlotinib (36.3%), and a pan-HER inhibitor (1.9%). They found that 20% of the patients showed de novo resistance to EGFR-TKIs, and patients with mutations in the
EGFR downstream genes had a significantly higher resistance to EGFR-TKIs than those without mutations regarding objective response rate (ORR). Mutations in
PTEN or
STK11 were significantly associated with a low ORR to EGFR-TKIs; mutations in
PIK3CA and
AKT did not differ significantly in terms of treatment response but showed a trend toward poor responses. Of the 55 patients who were treated with EGFR-TKI, the median PFS and OS were 10.3 and 21.2 months, respectively. Patients with mutations in the genes downstream of
EGFR had significantly shorter median PFS and OS compared to patients without mutations (3.0 vs. 12.0 months,
p = 0.060; 18.9 vs. 25.0 months,
p = 0.048). Collectively, their study suggests that the presence of mutations in key
EGFR downstream genes (
PIK3CA,
AKT,
PTEN, and
STK11) could affect the treatment outcome
[46][51].
Another study aimed to investigate the association between ETS exposure and
EGFR mutations in never smoker NSCLC patients. They found ETS exposure to be associated with a lower frequency of
EGFR mutations with an inverse relationship. They inferred that the history of ETS prior to diagnosis could serve as a negative predictor for
EGFR mutations in a similar manner to tobacco smoke. Thus, exposure to ETS could result in a similar carcinogenesis mechanism in never smokers to that in smokers. Indeed, genotoxic and epigenetic changes in smokers such as DNA adduct formation and oxidative DNA damage, as well as an increased number of
p53 mutations, including sister chromosome exchange, were also found in never smokers exposed to ETS. This indicates that cumulative ETS exposure is a major risk for the development of lung cancer even in NS
[48][53].
Lastly, a study comparing the immunohistochemical expression of a panel of
EGFR-related biomarkers in LUAD smokers vs. NS indicated that
EGFR expression was higher in tumours from smokers, whereas
pAKT was mainly overexpressed in tumours from never smokers. Biomarkers included
EGFR,
pAKT, PTEN, ki-67, p27, and hTERT from 190 patients with completely resected LUAD, of which 43 were NS. The expression patterns of all biomarkers except
PTEN were different among the smoking groups, which confirmed that specific abnormalities, including changes in
EGFR signalling pathways, characterise lung carcinogenesis in never smokers. Tumours from never smokers had a higher expression level of
pAKT than those from smokers. Thus,
pAKT might be involved in the development of LUAD, and alternative mechanisms rather than
PTEN inactivation could be responsible for its increased expression
[49][54].
4.2.4. microRNAs
Expression levels of microRNAs (miRNAs) vary in different types of human cancers, with lung cancer being no exception. Additionally, miRNAs have been demonstrated to be diagnostic and prognostic markers in other types of cancer, including lung cancer. A recent study investigated global expression profiles of miRNAs in never smoker and smoker lung cancer patients with
EGFR mutation versus wild type. The study revealed
EGFR-mediated regulation of miRNA expression. They employed 29 matched pairs of lung cancer and their adjacent normal tissue from never smokers. The expression of miR-21 was upregulated in smokers in comparison to never smokers with remarkable changes in cases with
EGFR mutations. The correlation between phosphorylated EGFR (p-EGFR) and miR-21 levels was found to be significant. Furthermore, since miR-21 was found to be suppressed by EGFR-TKI inhibitor, this insinuates that EGFR signalling is a pathway positively regulating miR-21 expression. Moreover, in the never smoker LUAD cell line (H3255) with mutant
EGFR and high levels of pEGFR and miR-21, anti-sense inhibition of miR-21 enhanced EGFR-TKI-induced apoptosis. With another LUAD cell line (H441), also from never smokers having EGFR wild-type, the anti-sense inhibition of miR-21 not only enhanced the effect of EGFR-TKI, but also induced apoptosis on its own. The expression of miR-21 was identified to be a downstream effector of the activated EGFR signalling pathway with a major oncogenic role in lung carcinogenesis, which could be utilised to improve response to EGFR-TKI therapy
[50][55].