GENETIC MODELS OF LUNG CANCER

The generation of transgenic mouse strains able to develop
lung cancer similar to the human situation enabled the identification
of the genes that drive lung cancer development
and progression. Multiple genetic changes are involved in the
development and progression of lung cancer. 45 These genetic
changes were found to inactivate tumor suppressor genes, activate
oncogenes, cause loss of heterozygosity (i.e., deletion of
one of two copies of allelic DNA sequences in particular chromosomal
regions), or amplify specific chromosomal regions.
Three major tumor suppressor genes have been identified as
being inactivated in lung cancer. The p53 and retinoblastoma
(Rb) genes are frequently inactivated by genetic alterations
such as chromosomal deletions and loss-of-function mutations,
whereas the p16 gene is inactivated by genetic alterations as well
as by transcriptional silencing because of hypermethylation. 45
Among the oncogenes, K- ras represents the main member of
this family involved in the development of lung cancer and up
to 30% of adenocarcinomas contains activated K- ras . 46
To recapitulate the events of lung tumor development and
progression, mice harboring genetic mutations similar to that
observed in human lung cancer have been generated. Most of
these genetically manipulated mice are suitable for the study
of genetic alteration in NSCLC as they primarily develop pulmonary
adenomas and adenocarcinomas. Transgenic mice that
express SV40 large T antigen, under the control of promoters
specific to pulmonary epithelium, develop multiple lung
tumors that cause early death. 47 Moreover, transgenic mice
expressing the proto-oncogenic H-ras under control of its
own promoter have a high incidence of lung tumors, 48 and
transgenic mice carrying a dominant negative mutation in
the p53 gene placed under the control of its own endogenous
promoter 49 or under a Clara cell secretory protein promoter
to target p53 expression specifically in the lung 50 are more
susceptible to several potential lung carcinogens, including
nitrosamines and benzo(a)pyrene. The lung adenomas generated
in the dominant negative p53 transgenic mice exposed to
carcinogens possess activating mutations in the K- ras protooncogene.
51 Thus, the dominant negative p53 transgenic mice
not only offer a suitable model to study the additive effects
of genetic alteration and environmental factors in the development
of lung cancer, but also clearly demonstrate that the
progression of lung cancer is a multiple-hit genetic event.
Besides mutations in key genes such as oncogenes or
tumor suppressor genes, altered expression of various growth
factors can be used as prognostic marker for lung cancer.
Increased plasma levels of insulin-like growth factor-II
(IGF-II), for example, are associated with a poor prognosis in
human pulmonary adenocarcinoma 52 and specific transgenic
overexpression of IGF-II in lung epithelium induces spontaneous
lung tumors in 69% of mice older than 18 months
of age. 53 These tumors display morphological characteristics
of human pulmonary adenocarcinoma (i.e., epithelial origin,
tubulo acinar architecture), strongly suggesting a critical role
for IGF-II in NSCLC development.
In addition to transgenic mice, mice that have germline
disruptions of genes involved in lung cancer have been generated.
Mice heterozygous and homozygous deficient for the
tumor suppressor genes p53 54 and p16INK4a 55 are viable
but have shortened life span because of the predominant occurrence
of lymphomas and various sarcomas. Although the
incidence of lung cancer is not increased in the p53 - and
p16INK4a-null mice, most likely because of their short life
span, bronchiolar neuroendocrine cell hyperplasia has been
described in p53 -null mice. 56 Interestingly, similar histological
findings have been noted in some patients with benign obstructive
respiratory disorders or with carcinoid tumors of the
lung, an uncommon human pulmonary epithelial malignancy
that rarely metastasizes. 57
Thirty percent of human tumors carry Ras gene mutations.
Of the three genes in this family, composed of K- ras ,
N- ras , and H- ras , K- ras is the most frequently mutated member
in human tumors, including adenocarcinomas of the pancreas
( 70% to 90% incidence), colon ( 50%), and lung
( 25% to 50%). Mice harboring the latent oncogenic mutation
G12D in the endogenous K- ras gene (K- ras LA) have
been generated. 58 Because early expression of oncogenic K- ras
causes mice to die very early during development, the K- ras LA
mice have been generated using the “hit-and-run” gene-targeting
procedure that involves two distinct steps of homologous
recombination. The first recombination event (insertion
event) is created in embryonic stem cells (“ hit ” step), whereas
the second recombination event (excision event) occur in vivo
only upon a somatic recombination event ( “run” step). All
mice, carrying this latent allele of K- ras , develop multiple lung
tumors, evident as early as 1 week after birth, with histological
features of human NSCLC. These tumors, unlike the human
situation, do not or rarely metastasize, most likely because of a
significant reduced life span of the mice. 58
The K- ras LA mice have been either crossed with various
transgenic mice, or treated with selective drugs to determine
the contribution of specific gene products in K- ras –mediated
lung tumorigenesis (Table 13.1). In this context, treatment of
K- ras LA1 mice with anti-CXCR2 neutralizing antibodies 59 or
the EGFR inhibitor gefitinib 60 has resulted in inhibition of
lung cancer progression and reduced number/expansion of alveolar
neoplasia, respectively. Similarly, the cross of K- ras LA2
mice with mice lacking the collagen receptor integrin- 1 1
resulted in prolonged survival as well as reduced number and
size of NSCLC, suggesting that collagen receptors and K-ras
might cooperate in lung cancer progression (Macias-Perez,
Pozzi, unpublished; Table 13.1 and Fig. 13.3 ) . In contrast,
the cross of K- ras -LA with TGF- heterozygote 61 or p53-null
mice 58 led to decreased survival rate with accelerated onset and
progression of lung cancer, suggesting that TGF- and p53
inhibit K- ras –mediated lung cancer initiation/progression.
Clearly, the generation of transgenic mice carrying the same
mutated genes observed in human lung cancer has enabled scientists
to better characterize the mechanisms by which these
genes drive lung cancer development. However, these genetic
models of lung cancer present the major disadvantages that
(a) mice develop primarily a subset of lung cancer; (b) tumors
usually do not metastasize, unlike in the human situation; and
(c) because of the short life span of tumor- bearing mice, it is
impossible to study the progression of lung cancer. To improve
mouse models of lung cancer initiation and progression, Jackson
and colleagues 62 generated a Lox-Stop-Lox K- ras conditional
mouse strain in which expression of oncogenic K- ras is controlled
by a removable transcriptional termination stop element.
Upon removal of the stop element, achieved by intranasal administration
of AdenoCre virus, the mice develop lung cancer.
Usually, tumors are visible within 2 weeks postinfection, and
they evolve from adenomatous hyperplasia to adenocarcinoma
within 16 weeks postinfection. Similarly, Meuwissen and colleagues
63 generated transgenic animals in which the chicken
-actin promoter drives the expression of GFP and oncogenic
K- ras V12 gene. As a polyadenylation signal behind the GFP cassette
prevents read-through into the K- ras V12 gene, expression
of K- ras V12 is dependent on Cre-lox–mediated deletion of the
GFP fragment. Within up to 56 weeks postinfection, all mice
developed multiple lesions that subsequently developed into
larger, papillary-like tumors within 8 weeks postinfection. 63
Without any doubt, the advances of these inducible
methods over existing models are that timing of tumor initiation
and location can be controlled and tumor multiplicity
can be adjusted by varying the administration of AdenoCre.
Recently, the K- ras –floxed mice described previously have
been crossed with either the Rac1 fl/fl mice 64 or the Lkb1 fl/fl
mice 65 and the incidence of lung cancer was followed upon
AdenoCre intranasal infection. Whereas expression of oncogenic
K- ras in the absence of small GTPase Rac1 led to prolonged
survival with reduced number, size, and incidence
of NSCLC 64 (Table 13.1), expression of oncogenic K- ras in
the absence of the serine/threonine kinase 11 Lkb1 increased
tumor multiplicity and metastasis, but decreased median survival
65 (Table 13.1). Thus, these studies establish Rac1 as a
critical key player in lung cancer progression, whereas LKB1
can be viewed as antipulmonary tumorigenic kinase, controlling
initiation, differentiation, and metastasis. Finally, Lyons
and colleagues 66 have crossed the K- ras –floxed mice with the
LucRep transgenic mouse that enables bioluminescence imaging
only upon AdenoCre treatment. Direct imaging of the
lungs from K- ras –floxed/LucRep mice treated with AdenoCre
revealed the in vivo detection of single lesion measuring
between 1 and 2 mm in diameter. 66 Thus, the LucRep mice can
be successfully used not only for noninvasive bioluminescence
imaging, but also for the analysis of tumors that cannot be easily
identified by traditional histology.
Administration of AdenoCre has also been used for the
establishment of the first mouse genetic model of SCLC 9. 67,68
Meuwissen et al. 67 have produced an animal model of SCLC
that closely resembles the human disease. Conditional mice
carrying combined floxed Rb and p53 genes were treated with
AdenoCre via intratracheal injection to delete these two tumor
suppressor genes specifically in the lung. Cre-mediated deletion
of all four conditional alleles resulted in the development of lung
tumors with the histological and immunohistochemical characteristics
of human SCLC. The originality of this model is that
lung tumors not only present the typical neuroendocrine features
as seen in human SCLC, but they also metastasize to sites
that mimic the human disease, including adrenal gland, ovaries,
and liver. Interestingly, in certain double floxed Rb and p53
mice treatment with AdenoCre led to conditional inactivation
of the p53 gene only, but not Rb gene, giving rise to NSCLClike
adenocarcinomas. 67 Thus, this model suggests that whereas
inactivation of a single tumor suppressor gene is sufficient for
the development of NSCLC, simultaneous inactivation of at
least two different tumor suppressor genes is necessary for the
establishment of SCLC. In addition, this model emphasizes the
importance of Rb during the development of SCLC.
The genetic models of lung cancers, similarly to the
chemical and orthotopic models, appear to be particularly
applicable not only for basic mechanistic studies, but
also for analyzing the efficacy of chemopreventive and/or
chemotherapeutic agents. This kind of study, however, is
limited to primary lung cancer, as in these genetic models
of lung cancer, primary tumors do not or rarely metastasize,
making the comparison with human tumors, and the evaluation
of potential antimetastatic drugs diff icult to evaluate.
Interestingly, it has been proposed that the effect of genetic
mutations on the development of metastatic tumors can be
influenced by the genetic background of the mouse. Using a
transgene-induced mouse tumor model that exhibits a high
incidence of pulmonary metastases and a breeding strategy
to vary genetic background, Lifsted and colleagues 69 found
signif icant differences in metastatic efficiency between the
original strains and f irst-generation hybrids, without altering
tumor initiation or growth kinetics. As all tumors are
initiated by the same oncogenic event, differences in the
metastasis are most likely because of genetic background
effects, rather than different combinations of oncogenic
mutations. Quantitative genetic mapping in the different
backcrossing has allowed the identif ication of at least three
loci that are associated with altered metastatic potential,
strongly reinforcing the notion that tumorigenesis and metastasis
are complex phenotypes involving not only cellular
responses to extrinsic stimuli, but also inherent genetic
components. 70 Thus, finding the ideal genetic background
for the development of primary as well as metastatic lung
cancers might enable researchers to study the different steps
involved in lung cancer progression, to screen potential antimetastatic
molecules, and to understand their mechanism
of action.

Cancer susceptibility is a complex interaction of an individual’s
genetic composition and environmental exposures.
Tremendous work has been done to understand cancer over
the past 100 years, from recognition of cancer as a genetic
disease to identification of specific carcinogens, isolation
of oncogenes, and recognition of tumor suppressors. Lung
cancer is most likely the result of an intricate interaction of
polymorphic susceptibility genes with many environmental
factors. Although genetic mutations in mice and humans do
not always lead to the same tumor spectrum, the underlying
molecular mechanisms are frequently relevant to both species.
The different mouse models of lung cancers described in
this chapter, mainly if used in combination with one another,
will facilitate the identification of genes that modulate an
individual’s susceptibility to cancer after a certain environmental
exposure, and will clearly allow researchers to gain
important insights into lung cancer development, treatment,
and, most importantly, prevention.
ACKNOWLEDGMENTS: I thank the National Cancer Institute
(NCI/NIH-CA94849) and the NIDDK (RO1-DK074359) for
their support. I apologize to many colleagues performing outstanding
research on mouse models of lung cancer whose work has not
been cited because of space constraints.