ABNORMALITIES IN GROWTH-INHIBITORY PATHWAYS: THE TUMOR SUPPRESSOR GENES

The chromosomal, genomic, and epigenomic studies addressed
previously have revealed multiple changes involving tumor suppressor
genes and oncogenes in clinically evident lung cancers.
The tumor suppressor genes, also known as recessive oncogenes,
are inactivated by genetic mechanisms such as point mutations,
chromosomal rearrangements, and mitotic recombinations, and
by epigenetic events like hypomethylation or hypermethylation
of gene promoter regions. It is largely accepted that the inactivation
of tumor suppressor genes commonly occurs through a
combination of two or more events, the Knudson hypothesis.
Still, it is also recognized that the phenomenon in carcinomas
is more complex because of mutational instability and chromosomal
instability. 63 The major tumor suppressor genes involved
in lung cancer are TP53 (17p13.1), RB1 (13q14.11), CDKN2
( p16 INK4a or MST1 , 9p21), and several genes located at the
short arm of chromosome 3. The incidence of abnormalities in
each of these genes in lung cancer, their main role in the development
of the disease, and their contribution as prognostic or
predictive markers will be briefly summarized.
The TP53 gene is well-known for playing a key role on
the negative regulation of G1/S-phase transition of the cell
cycle 64 and for being the gatekeeper for apoptosis. 65 Mutations
and overexpression of TP53 are present almost universally in
SCLC and in approximately 50% of NSCLC. 66–69 Mutations
in TP53 have been associated with smoking 70 and more aggressive
tumors 66,71 ; nevertheless, some studies have failed to show
a prognostic role for this abnormality. 72 Physical and functional
loss of TP53 and p53 protein overexpression have been identified
in dysplastic bronchial epithelium as a highly predictive
marker for lung cancer. 65,73–76 TP53 is regulated upstream by
the oncogene MDM2 (12q13–q14), which is overexpressed in
25% of NSCLCs. 77 The p53 protein also interacts with BCL2 ,
which is a negative regulator of cell death prolonging survival of
noncycling cells and inhibiting apoptosis. 78 Positive immunostaining
for BCL2 was found in approximately 20% of NSCLC
patients and 80% of SCLC. 78–80
The G1/S transition checkpoint is also deregulated in
lung cancer cells by changes in RB1 , CDKN2 , CCND , and
CDK4 . The retinoblastoma gene ( RB1 ) controls the G1/S transition
through E2. 81,82 Loss of RB1 function by deletion and
nonsense mutation or splicing abnormalities, together with
loss of the wild-type RB1 allele, are very common phenomena
in SCLC while occurs in less than 30% of NSCL. 82–86
In NSCLC, a strong correlation between altered RB1 protein
expression and early stage has been documented. 87 However,
correlation between loss of RB1 and clinical outcome is still
controversial, with earlier findings of negative prognostic
impact on survival in early stage NSCLC 88 not confirmed in
later studies. 89,90
The CDKN2 gene encodes an inhibitor of the cyclindependent
kinase 4 and its inactivation occurs through homozygous
deletion, or hemizygous deletion coupled with
inactivation of the second allele by point mutation or promoter
hypermethylation. 91 Loss of 9p has been detected frequently
in NSCLC (16% to 100%) (Fig. 6.3A) but not in
SCLC. 92–96 Loss of 9p21 is also relatively frequent in very early
epithelial lesions such as hyperplasia or dysplasia 70,73,97–99 and
hypermethylation at this site was found to increase during disease
progression, from 17% in hyperplasia to 50% in CIS. 97
CDKN2 hypermethylation has been reported to predict poor
5-year survival rate in resectable NSCLC, 100 and early recurrence
in resected stage I NSCLC. 101
In lung cancer, partial deletion of the short arm of chromosome
3 (3p) has been one of the earliest and most common
genetic changes (Fig. 6.4). Chromosome 3p deletion occurs in
almost 100% of SCLC and 90% of NSCLC. 102,103 Searches
for tumor suppressor genes in this large region identified several
targets at multiple sites, including FHIT (3p14.2), RASSF1
(3p21.3), TUSC2 ( FUS1 , 3p21.3), SEMA3B (3p21.3),
SEMA3F (3p21.3), MLH1 (3p22.3), and RARB (3p24). FHIT
is one of the most extensively investigated suppressor genes in
lung cancer 104–106 ; allelic imbalance at FHIT was observed in
64% of NSCLC patients and loss of protein expression in 50%
of lung cancers. 105–107 Allelic imbalance is associated with
physical loss of the chromosomal region (Fig. 6.3B). RASSF1 is
inactivated by promoter hypermethylation in the large majority
of SCLC and almost half of the NSCLC, 108–111 while not
methylated in noncancerous tissues. 112 Expression of TUSC2
protein is absent or reduced in the majority of lung cancers
and premalignant lung lesions and restoration of its function
in 3p21.3-deficient NSCLC cells significantly inhibits tumor
cell growth by induction of apoptosis and alteration of cell
cycle kinetics. 113 Both SEMA3F and SEMA3B transcripts are
underrepresented in lung cancers, mainly squamous cell carcinomas.
A recent review 114 indicated that downregulation of
SEMA3B and SEMA3F is sustained by gene hypermethylation
in lung cancer cell lines. 115,116 Moreover, the loss of function
of these genes correlates inversely with grade and stage of lung
cancer. 98,117 Additionally, the SEMA3B and SEMA3F genes
were found to be targets of TP53 , 118,119 which suggests that
they could be activated during DNA damage or other stress responses.
Deregulation in MLH1 , a mismatch repair gene, was
detected in up to 78% of NSCLC specimens, predominantly
by promoter hypermethylation 120 and more recently has been
associated with poor prognosis in NSCLC. 121 RARB mediates
growth control responses 122,123 and its expression was found
reduced in about 50% of NSCLC and 70% of SCLC. 124
Promoter hypermethylation is the leading cause of silencing of
this gene. Conclusions have been controversial regarding the
prognostic role of RARB suppression in lung cancer 125,126 as
well as regarding the efficacy of retinoids as chemopreventive
agents for this disease. 127,128
ABNORMALITIES IN GROWTHSTIMULATORY
SIGNALLING PATHWAYS:
THE PROTO-ONCOGENES
The molecular events that lead to the cancer-initiating cell involve
critical mutations in genes regulating normal cell growth
and differentiation. There are numerous families of protooncogenes
that contribute to the tumorigenesis process when
constitutively activated. The most relevant molecular pathways
to lung cancer pathogenesis include members of the EGFR, 129
MYC, 130 RAS, 131 and STAT 132 families, and other related genes
such as PIK3CA , 133,134 CCND1, 135 and BCL2 . 136,137 Protooncogenes
are frequently activated by genetic mutations (KRAS,
EGFR , and PIK3CA ) and chromosomal rearrangements, such
as translocations and inversions that place these genes under
the regulation of constitutively activated genes ( MYC , BCL2 )
or create chimeric proteins ( EML4-ALK ). Other mechanism of
proto-oncogene activation in solid tumors is gene amplification,
which recently was shown to occur more commonly in solid
tumors than previously acknowledged. 138 Examples of genes
found amplified in significant subsets of lung cancers are MYC ,
EGFR , HER2 , CCND1 , PIK3CA , and NKX2-1 ( TITF-1 ),
and a summary of their incidence, role in disease initiation and
progression, and prognostic and predictive impact in lung cancer
will be presented.
Among the most important growth factors for lung tumor
growth and proliferation are the tyrosine kinase receptors of the
ERBB family, which are coded by EGFR ( erbB1 at 7p12), ERBB2
(HER-2/neu, 17q12), ERBB3 (12q13), and ERBB4 (2q33.3).
The EGFR protein is frequently overexpressed in lung carcinomas
(50% to 80%) 139,140 and also in metaplastic lung tissue
adjacent to malignant tumors and normal-appearing bronchial
epithelium of patients with lung cancer. 141,142 Phosphorylation
of EGFR can activate signaling to cell proliferation and survival
via RAS/MAPK and PIK3CA/AKT pathways. 143 The
EGFR gene has been proved a relevant marker in lung cancer.
Activating somatic mutations in its tyrosine kinase domain have
been identified, and prevail in lung cancer patients of East Asian
ethnicity, of whom 25% to 40% have mutations, 144–148 when
compared to American or European patients, of whom only 8%
to 15% of patients carry mutations. 146,147,149–152 EGFR mutations
also occurs more frequently in never-smokers, women, and
NSCLC with adenocarcinoma histology. 147 The EGFR gene is
amplified in approximately 10% to 15% of advanced NSCLC
(Fig. 6.3C). 139,152–156 Besides, there is a significant copy number
gain for the EGFR gene in lung cancer as consequence of
chromosomal aneusomy and nonbalanced translocation or
other rearrangements. 157
Results are still controversial regarding the prognostic
characteristic of EGFR protein overexpression and gene
amplification. Recent data support no significant impact of
these factors despite a trend toward poor prognosis. 139,140,158
The activating mutations have been associated with a better
prognosis and indolent disease but, thus far, there is no definitive
support for this role. 159–161 Moreover, the status of the
EGFR gene has proved a powerful predictive marker for target
therapy agents. Not surprisingly, patients with EGFR high copy
numbers and activating mutations are more sensitive to EGFR
TKIs such as gefitinib and erlotinib. 143,146,149,151,152,155,162–165
NSCLC patients with EGFR gene amplification or high level
of genomic gain by chromosomal aneusomy have also shown
higher sensitivity to the monoclonal antibody anti- EGFR
cetuximab. 154
Overexpression of ERBB2 in lung cancer is less common
than EGFR overexpression, ranging from 10% to 30% in
NSCLC. 166 ERBB2 overexpression was detected in early stages
of lung cancer 167 and is associated with poor survival. 168–170
FISH analysis documented occurrence of ERBB2 gene amplification
in 6% to 20% of NSCLC patients. 51,171–175 Activating
mutations in ERBB2 are very rare in lung cancer 176–178 and
have been associated with resistance to EGFR TKI in cases
with all clinical and biological features of sensitivity to such
treatment. 178
ERBB3 is unique within the EGFR family because of its
catalytically deficient kinase domain and its signaling relies on
heterodimerization with other EGFR family member, preferentially
HER-2. 179 ERBB3 expression was less investigated in
lung cancers but between 19% and 58% of tumor specimens revealed
an increased level of expression with the highest percentages
seen in squamous cell carcinomas 180 ; increased expression
was also found in association with shorter survival. 181 ERBB3
gene amplification investigated by FISH was detected in approximately
5% of patients and there was no association with
histology subtype. 182 The other member of the family, ERBB4 ,
is activated by binding with neuregulins, betacellulin, and
heparin-binding epidermal growth factor (EGF)–like growth
factor. Its activation leads to cellular proliferation, chemotaxis,
or differentiation via activation of specific signal transduction
proteins, such as PI3-kinase and Shc. 183 The status of ERBB4
in lung cancers is still poorly known and this gene seems to
infrequently harbor mutations ( 3%) in NSCLC. 184
The genes of the RAS family (H RAS at 11p15.1, KRAS at
12p12.1, and NRAS at 1p13.2) encode for highly homologous
G-proteins located at the inner surface of the cell membrane
that play an essential role in the signal transduction pathways
involved in differentiation, proliferation, and survival. The
inactive RAS proteins are bound to guanosine diphosphate
(GDP) and upon activation release GDP and bind to guanosine
triphosphate (GTP). Activated RAS transduces the EGFR
activation signal to multiple downstream pathways, including
BRAF, MAPK, and PI3K/AKT. 131,185 The intrinsic GTPase
activity of RAS terminates signaling by hydrolyzing GTP to
GDP, a reaction that is accelerated by the GTPase- activating
proteins (GAPs). In lung cancer, KRAS is more frequently
mutated than HRAS and NRAS . 186 Point mutations in KRAS
codon 12, which is prevalent in NSCLC, and in codon 13 result
in an increased affinity for GTP, while mutations in codon
61 confer resistance to GAPs. The mutant proteins permanently
switch to the active position and constitutively activate
the downstream signaling pathways. KRAS mutations occur
with variable frequency in the major types of lung tumors.
They are very scarce in SCLC and prevail in large cell carcinomas
and adenocarcinomas, of which 20% to 30% carry a
KRAS mutation. 187–189 KRAS and EGFR mutations are almost
completely exclusive. 147,176,190,191 KRAS mutation has been
reported as a negative prognostic factor in terms of survival
in NSCLC, especially in adenocarcinoma, in a metaanalysis
including 28 studies with NSCLC patients and in more recent
studies. 188,192 KRAS mutation is also a negative predictor in
NSCLC for response to EGFR TKIs. 193–196 The KRAS gene
was found amplified in lung cancer specimens (Fig. 6.3D),
although the frequency in which this phenomenon occurs
is unknown at this time. Other downstream effecters of the
RAS pathway, such as the BRAF gene, which encodes a serine–
threonine kinase activated by point mutation, are infrequently
( 5%) mutated in lung cancers and likely to have a lesser relevant
role in the pathogenesis of those carcinomas. 197,198
The MYC family of genes ( MYC or c-MYC at 8q24.1,
MYCN at 2p24, and MYCL1 at 1p34) encodes basic–
helix-loop-helix zipper (bHLHz) transcription factors that, after
dimerization with MYC -associated factor X (Max), binds to
E-box motifs (CACGTG, CANNTG) and stimulates the transcription
of various target genes relevant for cell growth, differentiation,
and apoptosis. 130,199 One of these target genes is E2F
and, interestingly, it was reported that cMYC activates expression
of a cluster of six microRNAs on chromosome 13, two of
which ( MIRN17-5p and MIRN20A ) negatively regulate E2F . 200
These findings reveal a tightly controlled mechanism for activation
of transcription and limitation of translation exert by MYC
on E2F . Additionally, there is increasing evidence that the MYC
genes bind very ubiquitously throughout the genome, apparently
to genomic sites of up to 15% of all cellular genes, which hints
at a potential nontranscriptional function for them. 201 The alternative
model for MYC role in cell growth and tumorigenesis
is corroborated by studies showing that MYC promotes DNA
replication via nontranscriptional mechanisms and its deregulation
causes DNA damage predominantly during the S phase. 202
Amplification and overexpression of the MYC family of oncogenes
occurs in 18% to 54% of SCLCs, being more common
in chemorefractory disease. 126,203 In advanced NSCLC, from
20% to more than 50% of tumors were found to have MYC
gene amplification and this phenomenon was associated with
tumor progression and worse prognosis. 204,205 A recent investigation
using high-resolution array platform showed that MYC
was the most frequently amplified oncogene in lung cancer cell
lines, 138 which 28% of the 53 investigated lung carcinoma cell
lines showing amplified levels of its genomic region.
The PI3K–PTEN-AKT signaling pathway overcomes
mechanisms that promote apoptosis by transmitting a strong cell
survival signal. Interactions between cell surface receptors, such
as IGF1R, PDGF, and EGFR , and extracellular ligands, such as
EGF and TGF- , result in activation of tyrosine kinases and recruitment
of class I PI3Ks, a family of heterodimeric complexes
composed of a p110 catalytic and a p85 regulatory subunit. 206
PI3K phosphorylates phosphatidylinositol, which recruits specific
intracellular proteins, such as phosphoinositide-dependent
kinase-1 (PDK-1) and Akt/PKB, to the cytoplasmic membrane,
through mechanisms regulated by the PTEN gene. 207 Akt is a
serine/threonine kinase that acts downstream of EGFR to regulate
numerous other proteins involved in growth, survival, and
movement of cells, and angiogenesis. Akt activation results in
inactivation of pro-apoptotic proteins, including members of
the Bcl2 and caspase families 208,209 and other proteins that
indirectly inhibit apoptosis, such as mdm-2 and the forkhead
transcription factors. 210 The p110- catalytic subunit of PI3Ks
is coded by the PI3KCA gene (3q26) and there is increasing evidence
that constitutive activation of the PI3K pathways in lung
cancer occur as a consequence of mutation or amplification of
the PIK3CA gene. PIK3CA genomic gain detected by FISH was
reported in 43% of lung cancers with prevalence in squamous
cell carcinoma. 133,134 High level of phosphorylated Akt expression
has been observed in premalignant and malignant human
bronchial epithelial cells 134,211,212 and in approximately 50%
of advanced NSCLC. 152,213
High-resolution genomic profiling of lung cancer cell lines
and tumors revealed new genes frequently involved in amplifications.
One of them is the homeobox transcription factor
NKX2-1 ( TTF-1 or TITF-1 ) mapped at 14q13.3. 60,62 NKX2-1
plays a master role in induction and maintenance of lung and
thyroid morphogenesis and in the differentiation of epithelial
cell lineages. 214 NKX2-1 gene amplification is accompanied by
increased expression at both the RNA and protein levels, and
knockdown with small interfering RNA (siRNA) in lung cancer
cell lines led to reduced cell cycle progression and increased
apoptosis. Gain at 14q13.3 was present in 7% to 33% of cell
lines and tumors and was significantly more frequent in adenocarcinomas
than squamous cell carcinomas. 215 Interestingly,
the NKX2-1 amplification was associated with the presence of
EGFR -activating mutations but not KRAS or TP53 mutations,
and its overexpression was highlighted as a good prognostic factor
in a metaanalysis. 216 The oncogenic role of a tissue-specific
transcription factor linked to lineage proliferation and survival
may look somewhat peculiar but it reflects the principle of the
oncology recapitulating ontogeny. This phenomenon has been
detected involving other genes and solid tumor combinations,
such as breast cancer and ESR1 , melanoma and MITF , prostate
cancer and AR , and was also detected previously in lung cancer
with the TP63 gene. 62,217,218 It has been postulated that genetic
alterations that directly interfere with transcriptional networks
normally regulating lung development may be a more common
feature of lung cancer than previously realized. 62
Genomic changes in proto-oncogenes may occur in association
with therapeutic strategies applied to the patients. An
interesting example of this phenomenon in lung cancer implicates
mesenchymal–epithelial transition (MET) (7q31.2),
the receptor for the hepatocyte growth factor (HGF). MET is
frequently deregulated in cancers via constitutive kinase activation,
paracrine/autocrine activation, mutation, gene amplification,
and epigenetic mechanisms. 219 Enhanced MET regulation
leads to oncogenic changes including cell proliferation, reduced
apoptosis, angiogenesis, altered cytoskeletal function, and metastasis.
MET overexpression occurs in a varied lung cancer histologies,
with stronger expression in NSCLC. 219 Mutations in
the tyrosine kinase domain of MET were detected in SCLC and
NSCLC 220,221 but are uncommon. MET gene amplification was
relatively frequent (20%) in few NSCLC cell lines tested, 222 but
appears to be infrequent ( 5%) in unselected clinical NSCLC
specimens. 223 Recently, amplification of MET was identified in
in vitro studies as a major mechanism by which lung tumors
overcome therapeutic inhibition of EGFR growth signals. 224 In
addition, assessment of tumor tissue from gefitinib or erlotinib
resistant NSCLC patients demonstrated MET amplification in
21% to 22% of patients, 223,224 a much higher frequency than in
the unselected patients. Mechanistically, it was shown that MET
protein regulates ERBB3-dependent activation of PI3K at the
same time that signals through ERBB3 in amplified cancers, and
this redundant activation of ERBB3 supports the downstream
signaling even in the presence of EGFR inhibitors. 223,224
The increased knowledge about the mechanisms leading
and maintaining oncogene activation in lung cancer has already
provided a striking contribution toward development of
new therapeutic approaches. It has also sustained the better understanding
of the complex signaling network in normal and
cancer cells. Observations that the inactivation of few or even
a single oncogene was sufficient to induce a sustained tumor
regression have supported the oncogene addiction hypothesis by
which tumors may become irrevocably addicted to the oncogene
that initiated tumorigenesis and a sudden interruption of its activity
balances toward proliferative arrest and apoptosis. 225–227
More recently, a compelling alternative has been raised to explain
those observations, the “oncogene amnesia” hypothesis.
The premise of this hypothesis is that the oncogene activation
initiates tumorigenesis by overriding essential mechanisms for
cellular mortality, self-renewal, and genomic integrity, thus inducing
a state of cellular amnesia. 228 The rationale behind the
oncogenic amnesia hypothesis is that the inactivation of a single
oncogene in a tumor that has acquired all oncogenic lesions required
to overcome the cellular safety mechanisms can restore
pathways leading to proliferative arrest, differentiation, cellular
senescence, and apoptosis. In this way, the oncogenes initiate
cancer inducing a cellular state of enforced amnesia in which,
only upon oncogene inactivation, the tumor becomes aware of
its transgression. 228 Oncogene addiction and oncogene amnesia
are not necessarily exclusive mechanisms, as the hypotheses are
proposed, they may coexist in complex tumors as carcinomas.