MOUSE MODELS OF LUNG CANCER − Rakyat Biologi

The mouse is the principal animal model for the study of lung cancer. Widespread adoption of mice as models for human lung cancer is a consequence of a long history of use focused around the breadth of genetic variation, ease of genetic manipulation, and the ability to induce lung cancer with molecular and histological similarities to human disease. Mouse lung cancer models are now frequently used in pre-clinical tests of therapy and prevention. Engineered models that replicate specific genetic lesions found in human tumors, such as expression of activated oncogenes (e.g. Kras) or inactivated tumor suppressor genes (e.g. p53) are often used to investigate the genesis of lung cancer. In addition, chemical carcinogen-induced models utilizing mouse strains with a predisposition to cancer have been used to successfully address the genetic complexity of human lung tumors. Our goal in this review is to provide a summary of the current state of these carcinogen-induced models. The reader is referred to several recent reviews addressing the use of engineered mouse models for further information on genetic models of lung cancer (2-4).

In humans, complex chemical mixtures, in particular cigarette smoke, are the predominant initiator of lung cancer. Because cigarette smoking is the primary cause of human lung cancer, individual cigarette smoke carcinogens are frequently used to induce lung tumors in mice. This is commonly performed by intraperitoneal or dietary administration of carcinogens of the polycyclic aromatic hydrocarbon (PAH) and nitrosamine class (5-8). PAHs are largely produced during the combustion of tobacco, while nitrosamines are already present in unburned tobacco and are formed as a consequence of the tobacco curing process. Benzo(a)pyrene (B(a)P), a PAH, and the nitrosamines, 4- (methylnitrosamino)-1- (3-pyridyl)-1-butanone (NNK) and N'-nitrosonornicotine (NNN), are strong inducers of lung adenomas and adenocarcinomas in mice. These chemicals are pro-carcinogens that require metabolic activation to electrophilic compounds that react with DNA and form adducts. Subsequent failure of repair or misrepair results in genetic mutations. Cytochrome P450 enzymes are central to bioactivation of B(a)P and NNK, and are encoded by a variety of Cyp genes that exhibit differences in tissue expression, and chemical specificity (9). P450s are responsible for activation of B(a)P to a diol-epoxide, that reacts with deoxyguanine to form a bulky adduct, C⁸-B(a)P-diolepoxide-guanine, which most often results in G-to-T nucleotide transversions. P450s also catalyze α-hydroxylation of NNK, which spontaneously decomposes to aldehydes and diazonium ions, subsequently forming O⁶-methyl-guanine adducts and ultimately G-to-T transversions (10). P450 enzymes are expressed most abundantly in the liver, but are also present in the peripheral and bronchial epithelia of the lung. Conditional, lung-specific deletion of NADPH-P450 reductase (the only mammalian P450 reductase gene and electron donor for many P450 reactions), demonstrated reduced NNK-induced tumor load concurrent with lower O⁶-methylguanine levels (11). In contrast, the authors demonstrated that conditional liver-specific deletion of NADPH-P450 reductase increased lung tumor burden after i.p injection of NNK. This result suggests that the overall function of liver P450 enzymes in response to NNK is detoxification and metabolism, while P450 enzymes in the lung bioactivate the NNK pro-carcinogen.

Inbred mouse strains, such as A/J and SWR, have a high incidence of spontaneous lung tumor development. A/J strain mice have roughly 80-100% incidence of spontaneous lung tumors after 24 months, and tumors are often detected within the first 6 months of age (5, 12). These strains are also very susceptible to carcinogen-induced lung tumors. Other strains such as C57BL6/J, C3H/J and DBA are very resistant to carcinogen-induced lung tumors, while strains such as O20 and BALB/c have intermediate susceptibility. A/J strain mice develop approximately 25 tumors per lung 14-16 weeks post-treatment with the carcinogen ethyl carbamate (urethane), while the C57BL/6J strain develops less than 1 tumor per lung on average (5, 7, 13-15) (Table 1). These strain-dependent differences have permitted several research groups to map genetic susceptibility loci that associate with both carcinogen-induced and spontaneous lung tumor development (16, 17).

Murine lung tumors bear similar morphology, histopathology, and molecular anomalies as those observed in human tumors. The majority of tumors observed in murine models are benign pulmonary adenomas that have clear borders and are comprised of well-differentiated cells. Although adenomas are rarely observed in humans, likely because they are asymptomatic and not frequently diagnosed, murine adenomas do exhibit histological similarity to non-small cell lung adenocarcinoma derived from airway type II cells. Murine adenomas are considered precursors to murine lung adenocarcinomas as they do progress to malignant adenocarcinomas of various subtypes (solid, papillary, brochiolo-alveolar) which show signs of nuclear atypia and invasiveness (18).

Table 1. Selected rodent models of chemically-induced lung carcinogenesis

Model	Strain	Carcinogen	Tumor
Mouse AD/ADC	A/J	B (a)P, i.p. 100 mg/kg	20 w: 8-10 tumors (AD), 100% incidence (20, 21)
		B (a)P, i.g. 100 mg/kg (3X)

A/J

NNK, i.p. 100 mg/kg

20 w: 6-8 tumors (AD),

100% incidence (20, 22, 23)

52 w: 15 tumors (95% AD, 5% ADC),

70-80% incidence (ADC)

A/J

Urethane, i.p. 1 g/kg

16 w: 20-25 tumors (AD) (21, 24-26)

A/J

Vinyl carbamate,

i.p. 60 mg/kg

24w: 25 tumors (AD), 12% incidence (ADC)

52 w: 30% incidence ADC (27, 28)

Swiss albino -newborn

Main-stream cigarette smoke,

120 days

26-33w: 6-14 tumors (AD), 80% incidence (AD), 5-20% incidence (ADC) (29)

A/J

Main- and side-stream cigarette smoke, 5 mos smoke + 4 mos air

3 tumors (AD) vs 1 spontaneous tumor (AD) (30)

B6C3F1

Mainstream cigarette smoke, lifetime

10X increase in hyperplasia, 4.6X AD and papilloma, 7.3X ADC, 5X metastatic pulmonary ADC (31)

Rat AD/ADC

F344

NNK, s.c. 1.5 mg/kg (3X, 20 w)

98w: 67% incidence (AD), (33% ADC) (32)

F344

Mainstream cigarette smoke, up to 30 months

Incidence increased from 0% in control to 6% (light smoke) to 14% (heavy smoke) (33)

Mouse squamous

Swiss – 8 w

NTCU, 3 mmol, 2x week (22 w)

24w: 50% hyper/metaplasia, 10% CIS/SCC (34)

i.p. = intraperitoneal, i.g. = intragastric, i.t. = intratracheal, AD = adenoma, ADC = adenocarcinoma

Murine models of lung carcinogenesis exist for adenocarcinoma and squamous cell subtypes (Table 1), while currently those for small cell carcinoma rely solely on genetic ablation of Rb and p53 genes, and large cell models have yet to be described. The majority of mouse lung adenoma/adenocarcinoma studies employ single intraperitoneal injection of the carcinogens B(a)P, NNK, urethane or vinyl carbamate in a susceptible strain such as A/J at 5-6 weeks of age. At 20 weeks post injection these carcinogens cause anywhere from 8-25 tumors of which almost 100% are histologically adenomas. At 52 weeks post-initiation, 1-2 adenocarcinomas are typically observed.

Most lung carcinogens are ‘complete’ and thus both initiate tumorigenesis and promote tumor progression through dysregulated proliferation of Clara cells and Type II pneumocytes. Soon after initiation, hyperplastic foci in the bronchioles and alveoli are observed. Which of these foci progress to adenomas, and which of these spontaneously regress is currently unknown. However, progression of adenoma to invasive adenocarcinoma may be infrequent (< 10%, 1 year post carcinogen) and metastasis to other organs is extremely rare in carcinogen-based models (19). Thus, the majority of tumors remain as benign adenomas even 1 year post administration of carcinogen. Immunohistochemical staining of tumors is typically positive for surfactant protein C (SPC, an immunohistochemical marker for type II cells), but not secretoglobin 1a1 (also known as Clara cell secretory protein (CCSP) or CC10, a marker for Clara cells), suggesting that most mouse lung adenoma/adenocarcinomas originate from Type II cells, or potentially through loss of expression of CC10 in a Clara cell transdifferentiation process.

Activating mutations in Kras are a prominent early events observed in both human and mouse lung tumors induced by carcinogen (35, 36). The human precursor lesion to adenocarcinoma, atypical adenomatous hyperplasia (AAH), possesses Kras mutations at similar frequency to adenocarcinomas, suggesting that the Kras mutation is an important early event in tumorigenesis (37). Estimates are that 15-50% of all human lung adenocarcinomas possess mutations in Kras, most commonly in codon 12, and less often in codons 13 and 61 (38-40). The tumor suppressor genes Trp53, p16, Rb, Apc, and p16^INK4A (Cdkn2a) are also suppressed and/or inactivated in both human and mouse tumors, either through methylation or less frequently by mutation, suggesting that the molecular mechanisms between species are relatively well conserved (18, 41).

In mouse, factors influencing Kras mutational frequency and spectrum include the type of carcinogen, the type of adduct formed, age of the mouse (fetal or adult), dose of carcinogen and strain susceptibility. In the A/J mouse, activating mutations in Kras are observed in 100% of chemically-induced lung tumors and in greater than 80% of spontaneous tumors (21, 42). Kras mutations in humans are typically G-T transversions and correlate with smoking. Similarly, Kras mutations in mouse lung tumors are G-T transversions in codon 12 when initiated by the cigarette smoke carcinogens B(a)P or NNK. Alternatively, urethane and vinyl carbamate most often cause codon 61 A-T transversions in A/J mice.

Approximately 60 known carcinogens are present in both mainstream and sidestream cigarette smoke, in addition to several co-carcinogens and tumor promoting compounds. Initial attempts using cigarette smoke to produce pulmonary malignancies with high incidence in mice indicated cigarette smoke was weakly tumorigenic in mouse models (43). Subsequently, a protocol was developed by which tumors could be induced in A/J strain mice through a 5 month exposure of a combined mainstream and sidestream smoke followed by a 4 month exposure to normal air (30). This exposure to air was essential to the development of tumors. Multiple groups have used this protocol (generally in A/J and SWR strains) and have observed an increase in tumor multiplicity from approximately 1 to 3 tumors (adenomas) per lung. The effect on malignant lesions (adenocarcinomas) was inconclusive. Tests on nine different inbred strains demonstrated that strain sensitivity to smoke-induced lung tumors was similar to purified carcinogen-induced tumors (44). An alternative assay has since been developed using mainstream cigarette smoke for 920 days (6 h/day, 5 days/week) and demonstrated more robust enhancement of lung tumor incidence and multiplicity (31). This was observed through measurement of a variety of lesions (hyperplastic, benign, and malignant) in a mouse strain (B6C3HF₁) that normally has a very low basal level of pulmonary neoplasia. Incidence of lung adenoma (28% vs. 7%), adenocarcinoma (20% vs. 3%) and distant metastases (1.5% vs. 0.3%) was enhanced by smoke versus normal air. Exposure of newborn Swiss albino mice to mainstream cigarette smoke for 120 days induced lung tumor multiplicities of 6 and 14 tumors in males and females, respectively, when mice reach 180 to 230 days of age, demonstrating enhanced susceptibility in early life (29).

Squamous cell carcinoma is the second most common lung neoplasm, however, reproducible mouse models of this disease were lacking until the last 10 years. Early models using intratracheal inhalation of B(a)P and 3-methylcholanthrene (3-MCA) were technically difficult and not easily replicated (45-47). It was not until experiments involving topical application of nitrosoalkylureas to induce skin cancer in female Swiss mice revealed that such compounds could induce a broad spectrum of cancers that a reproducible model of squamous cell lung cancer was developed (48). Since then a protocol has been established using twice weekly application of N-nitroso-tris-chloroethylurea (NTCU) on a shaved dorsal patch of skin in female NIH Swiss mice for the duration of the experiment, typically 22-24 weeks. NTCU causes bronchiolar basal cell hyperplasia, and continues through squamous metaplasia, dysplasia, carcinoma in situ (CIS) and invasive squamous cell carcinoma, in a sequence of events comparable to the human condition. In contrast to mouse lung adenomas, SCC tumors are not nodular and appear as more opaque patches with indistinct boundaries. At 24 weeks after the initial NTCU dose there is a 50% incidence of hyperplasia/metaplasia and 10% incidence of CIS/SCC. Eight months of treatment results in an 80% incidence of squamous cell lung carcinoma in these mice (34). Further refinement of the timing and dose of NTCU administration in this model may be necessary to generate the full spectrum of pre-malignant and malignant lesions without overt toxicity (49).